Volume 1 - Issue 1, 2016

Perampanel in generalised tonic-clonic seizures Back


Primary generalised tonic-clonic (PGTC) seizures are the most frequent and disabling seizure type among patientswith idiopathic generalised epilepsy. Uncontrolled seizures are associated with high morbidity and mortality. Sodium valproate is considered the antiepileptic drug (AED) of choice for PGTC seizures, while the newer AEDs, including perampanel, are effective for secondarily generalised tonic-clonic seizures and refractory PGTC seizures. In a phase 3 randomised controlled trial (RCT) of adjunctive perampanel in patients with refractory PGTC seizures, percent reduction in 28-day seizure frequency increased in the perampanel group in a dose-dependent manner compared with placebo. Quality of life improved and emergency room visits were reduced among patients receiving perampanel. Results for perampanel in a real-life clinical population (the FYDATA study) were similar to those of the RCTs. The median seizure reduction was 33% (last observation carried forward [LOCF]) and 58% (completer group) over 1 year. In secondarily generalised seizures, the median seizure reduction was 75% (LOCF) and 100% (completer group) over 1 year. There were fewer adverse events (AEs) and psychiatric AEs with titration of 2 mg every 3–4 weeks. Slower titration is recommended for patients who may be more sensitive to AEs or who have psychiatric comorbidities.


Perampanel (Fycompa®; Eisai Co. Ltd., Tokyo, Japan) is a non-competitive selective α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor antagonist. Glutamate binds to and activates the AMPA receptor, opening the channel and enabling sodium influx. In the presence of a competitive antagonist, glutamate cannot activate the receptor unless the glutamate levels are high, when it displaces the antagonist. Perampanel differs from the competitive antagonists in that it does not compete with glutamate.1 Irrespective of the amount of glutamate present, it does not displace perampanel from the binding sites. Thus, receptor antagonism is maintained and the channel remains closed.

Other receptors play a role in the development of seizures, including N-methyl-daspartate (NMDA) receptors. Compared with NMDA receptors, AMPA receptors mainly mediate fast excitatory synaptic neurotransmission. AMPA receptors do not impair synaptic plasticity and long-term potentiation. Therefore, AMPA receptor blockade does not affect memory. NMDA receptor antagonists may produce psychotomimetic effects, while AMPA receptor antagonists are not known to produce these effects. AMPA receptors are involved in the generation and propagation of epileptic discharges, while NMDA receptors have a limited role.

Primary generalised tonic-clonic (PGTC) seizures arise from widespread bilateral thalamocortical circuitry. Secondarily generalised tonic-clonic (SGTC) seizures spread bilaterally from the initial seizure focus to subcortical centres via thalamic projections. AMPA receptors are present in these pathways, including the thalamic neurons.

For partial-onset seizures (POS; focal seizures) with or without secondarily generalised seizures in patients aged 12 years or older, perampanel is approved in Japan, Hong Kong, South Korea, Malaysia, Thailand, Taiwan, Singapore, and Philippines. For PGTC seizures in patients aged 12 years or older, perampanel is approved in Japan, Hong Kong, Singapore, and Philippines, while approval is expected in Singapore, Malaysia, Thailand, and Taiwan in 2016. The Perampanel Asia-Pacific Advisory Board Meeting was held in Hong Kong on 12 May 2016 to discuss the burden of epilepsy in the region and the role of perampanel in the treatment of generalised tonic-clonic (GTC) seizures in Asia.

Disease burden and comorbidities of GTC seizures

Professor Patrick Kwan of the Department of Medicine, The University of Melbourne, Melbourne, Australia, discussed the burden of disease and comorbidities associated with GTC seizures. Among patients with idiopathic generalised epilepsy (IGE) and other generalised epilepsy syndromes, PGTC seizures are the most common type, and are considered to be the most disabling seizures.2,3 It can be difficult to differentiate PGTC seizures from secondarily generalised seizures, which have a focal onset with subsequent bilateral spread. However, correct differentiation of primary and secondarily generalised seizures is essential for ensuring appropriate treatment.4,5 The exact prevalence of PGTC seizures is difficult to determine, but one study estimated the prevalence of patients presenting exclusively with GTC seizures to be 20%.6

The frequency of PGTC seizures in IGE syndromes ranges from 30% to100%.Investigation of epilepsy syndromes in newly diagnosed patients shows that IGE mostly occurs in children and younger age groups, although the prevalence differs between countries.8 The distribution of seizure types varies with age. Data from Rochester, Minnesota, show that generalised seizures account for approximately half of all seizures in patients younger than 15 years and one-third of all seizures in adults (age 35–64 years).9 In adults, complex partial seizures, or focal seizures without impairment of awareness, are present in approximately 39% of new epilepsy cases. Absence seizures, which account for approximately 13% of seizures in patients younger than 15 years, are rare after adolescence. Many patients experience more than one seizure type.

There is variation in the reported frequencies of generalised seizures throughout Asia, both between countries (from 29% in Singapore to 92% in Malaysia) and within countries (from 29% to 69% in Singapore).10 These differences may be due to primary and secondarily generalised seizures being considered a single entity.

Figure 1: Consequences of uncontrolled epilepsy. Abbreviation: SUDEP, Sudden Unexplained Death in Epilepsy.

Figure 1: Consequences of uncontrolled epilepsy. Abbreviation: SUDEP, Sudden Unexplained Death in Epilepsy.


Treatment gap for patients with PGTC seizures

There are many drugs available for treatment of epilepsy, but the spectrum of efficacy varies between drugs, particularly those for generalised seizures. Drugs with a relatively narrow spectrum of activity, including carbamazepine, gabapentin, oxcarbazepine, phenytoin, and vigabatrin, can aggravate PGTC seizures and other generalised-onset seizure types.11-15 While carbamazepine and oxcarbazepine are included in guidelines as possible first-line treatments for PGTC seizures, there is potential for these drugs to exacerbate other generalised seizures such as myoclonic or absence seizures.16

A recent International League Against Epilepsy (ILAE) review of class I evidence of monotherapy effectiveness made recommendations for first-line treatment of POS, but there is insufficient high-level evidence to recommend treatment for GTC seizures, highlighting the limited treatment options available.11 Other guidelines such as those from the Scottish Intercollegiate Guidelines Network (SIGN) and National Institute for Health and Care Excellence (NICE) recommend sodium valproate and lamotrigine as first-line treatment for PGTC seizures, and levetiracetam is increasingly being used as first-line or add-on treatment.16,17 Other add-on options include topiramate, phenobarbital, and zonisamide.11,12

IGE and PGTC seizures are often considered to be easily treatable, but this is not always the case. IGE syndromes have the best treatment outcomes, with 66% of patients achieving seizure freedom compared with 57% of patients with cryptogenic (unknown) epilepsy and 56% with symptomatic (structural/metabolic syndrome) epilepsy. However, 27% of patients with IGE remain uncontrolled.18

Uncontrolled epilepsy is associated with considerable burden for patients, including unpredictability of seizures, increased morbidity and potential for injury, psychosocial consequences, and psychiatric comorbidities (Figure 1).19 Importantly, patients with GTC seizures have an increased risk of death, including sudden unexpected death in epilepsy (SUDEP).20-22 SUDEP is second only to stroke as a cause of life years lost to neurological diseases.23 Occurrence of one to three tonic-clonic seizures per year increases the risk for SUDEP,22,24,25 particularly in young people. For example, patients whose age at onset of seizures was <16 years who had three or more seizures/ year had a 19-fold risk compared with a 10-fold risk those whose age at onset of seizures was ≥16 years.24

While most of the mortality data come from western centres, data from Asia are emerging. Ding et al found that the risk of premature mortality in patients with convulsive seizures in rural China was 2.9 times higher in patients with epilepsy than in the general population.26 The highest proportional mortality ratios were for stroke (15%), drowning (14%), self-inflicted injury (13%), and status epilepticus (6%), with probable SUDEP at 1%. Uncontrolled seizures are associated with a higher risk of injuries, and the risk of minor injury in patients with epilepsy is up to 10-fold that of the general population.27,28 Frequent seizure occurrence can also impact cognition, and patients with uncontrolled PGTC seizures have greater cognitive decline than a healthy control group.29,30

Up to 15% of women will experience PGTC seizures during pregnancy, posing a potential risk to both the woman and the foetus. The SIGN and NICE guidelines recommend vigilance when treating pregnant women; while the long-term effects of tonic-clonic seizures in pregnancy are not well known, they could adversely affect the obstetric outcome.16,17 Pregnant women with GTC seizures should be informed about possible risks to the fetus during a seizure which, although low, may depend on seizure frequency.16

The socioeconomic burden of uncontrolled IGE is high. Patients with one or more seizures/month had worse health utility scores, and greater presenteeism, work impairment, activity impairment, health resource utilisation, indirect costs, and direct costs than those experiencing less than one seizure/year.31

In summary
PGTC seizures are the most common and most disabling seizure type among patients with IGE and other generalised epilepsy syndromes. However, treatment of these seizures is limited. Uncontrolled PGTC seizures are associated with cognitive impairment and psychological comorbidities, risk of seizure-related injury and SUDEP, and high socio-economic burden. Thus, more effective treatments for PGTC seizures are needed.

Management and treatment of GTC seizures: Asian perspective

Prof Byung-In Lee of the Department of Neurology, Inje University, Haeundae Paik Hospital, Busan, Korea, described the approach to the management of GTC seizures in Asia.

Neurobiology of epilepsy
To make an accurate diagnosis, it is necessary to understand GTC seizures. In 2010, the ILAE revised classification of epileptic seizures defined generalised seizures as those originating at some point within, and rapidly engaging, bilaterally distributed networks, which may not necessarily include the entire cortex. Although individual seizure onset can appear localised, the localization and lateralization are not consistent from one seizure to another. Focal seizures were defined as those originating within networks limited to one hemisphere. For each seizure type, ictal onset is consistent from one seizure to another, with preferential propagation patterns that can involve the contralateral hemisphere. Thus, the only differentiation between generalised and focal seizures is the distribution of the epilepsy network, which is different to the previous concept of an epileptogenic zone.

The epilepsy network was defined by Spencer et al as functionally and anatomically connected sets of cortical and subcortical brain structures and regions, where activity in any part affects activity in all others.32 In 2013, Bertram et al proposed a functional anatomy of an epilepsy network (Figure 2) consisting of four major components of.33

  • seizure focus: generating sparks for the transition from the interictal to ictal state
  • initiating circuit: a separate neuronal population that is necessary to support the start of a seizure – typically built around cortical and subcortical connections
  • pathways of spread or recruitment: routes that seizures follow to extend beyond the initiating circuit
  • modulatory centres: regions projecting to key components of the seizure circuits affecting the level of excitability of those components.

In the feline generalised penicillin epilepsy model of absence epilepsy, cortical unit activity seems to precede the thalamic unit, but during the spike and wave burst, thalamic unit activity seems to precede cortical unit activity, indicating the controversies around the hypotheses of generalised absence seizures.34 Several theories have been proposed for generalised absence epilepsy, among which the cortical focus theory by Meeren et al has the most support and is the most similar to Bertram et al’s concept

Figure 2: Functional anatomy of seizures. (a) Focal seizure and (b) generalised seizure [reprinted with permission from Elsevier © 2013].33 Abbreviation: NM, neuromodulatory centres.

Figure 2: Functional anatomy of seizures. (a) Focal seizure and (b) generalised seizure [reprinted with permission from Elsevier © 2013].33

Abbreviation: NM, Neuromodulatory Centres.

for the epilepsy network.35 During the first generalised spike and burst, the somatosensory cortex was the initiating circuit, but during the seizure the thalamus preceded the electrical discharges of the other structures. The ‘cortical focus’ theory may explain why focal features are frequently observed in IGE, and suggests that partial seizures and generalised seizures are on a continuum rather than discrete entities.

Diagnostic challenges
GTC seizures may be defined as part of PGTC seizures (IGE), secondary GTC seizures in focal epilepsy, or unclassified GTC seizures. There are frequent focal features in the semiology of GTC seizures (>50% of patients with IGE), and 64% of patients with IGE report auras. Thus, it is hard to differentiate secondarily generalised seizures from IGE. Frequent focal features on electroencephalography (EEG) have been found in 30–55% of patients with IGE. Other diagnostic challenges include GTC seizures without associated absence or myoclonic seizures, onset of GTC seizures after the age of 30 years, normal EEG at first consultation and persistent normal EEG at follow-up, and focal seizures mimicking IGE such as frontal lobe absence.

It is important to correctly differentiate between primary and secondary GTC seizures because of the selection of AEDs. Narrow-spectrum AEDs, including carbamazepine, oxcarbazepine, phenytoin, phenobarbital, lamotrigine, tiagabine, and benzodiazepines, may aggravate generalised seizures, especially tonic, absence, atonic, and myoclonic seizures.

AED treatment
There are no class 1 data for initial monotherapy for GTC seizures.11 In 1987, sodium valproate was reported to be highly effective in IGE, with 83% of patients becoming seizure free.36 Sodium valproate became the standard drug for IGE. In the SANAD (Standard and New Antiepileptic Drugs) study, sodium valproate was found to have better tolerability than topiramate and was more effective than lamotrigine in patients with PGTC seizures.37 KOMET (Keppra Versus Older Monotherapy in Epilepsy Trial), a large-scale open randomised multicentre trial, found that sodium valproate was more effective than levetiracetam in patients with IGE (12-month seizure freedom 61% vs. 68%).38 In the LaLiMo (Lamotrigine Versus Levetiracetam in the Initial Monotherapy of Epilepsy) trial, lamotrigine and levetiracetam were equally effective for GTC seizures (6-month seizure freedom 45.2% vs 47.8%).39 In a Cochrane review, comparison of eight drugs found that sodium valproate and phenytoin were the most effective AEDs, with equal effectiveness, in PGTC seizures.40 Topiramate, lamotrigine, levetiracetam, and perampanel have all been found to be effective as add-on therapy for drug-resistant PGTC seizures.

Carbamazepine, phenytoin, and phenobarbital were more effective than primidone as monotherapy in newly diagnosed patients with SGTC seizures.41 In a later study, carbamazepine was more effective than sodium valproate for control of complex partial seizures, although the two drugs were comparable for GTC seizures.42 Comparison of pregabalin and lamotrigine in patients with newly diagnosed SGTC seizures found that lamotrigine was more effective (6-month seizure freedom 80% vs 57%; p<0.0001).43 Perampanel has been found to be effective as adjunctive therapy in SGTC seizures (77% median percent change at 9 months and at 90% 2 years).44

In a survey of epilepsy experts from seven Asian countries (Malaysia, Philippines, Thailand, Hong Kong, Singapore, Taiwan, and Korea), the firstline therapy for IGE was sodium valproate (>50% of responders) and the second-line therapy was lamotrigine (50%) followed by levetiracetam. For add-on therapy to sodium valproate, lamotrigine was the preferred option, followed by levetiracetam. In women with epilepsy, the first choice was lamotrigine (50%) followed by levetiracetam. For SGTC seizures, there was no consensus for first- or second-line therapy or for women. Similar results were found in a survey of nine Korean epilepsy experts. For IGE, the preferred first-line therapy was sodium valproate, and for second-line therapy and women, the preferred choice was levetiracetam followed by lamotrigine. For SGTC seizures, there was no consensus for first- or second-line therapy or for women. For these specialists, the main concerns about using narrow-spectrum AEDs for patients with PGTC seizures were worsening of seizures, poorer efficacy, more adverse events (AEs), and lack of evidence.

In summary
Sodium valproate is the drug of choice for PGTC seizures in Asia, while the major AEDs seem to be effective and comparable for SGTC seizures. In an Asian experts survey, there was good consensus for the AEDs used to treat PGTC seizures, but no consensus for which AEDs to use for SGTC seizures. Concerns about using narrow-spectrum AEDs in PGTC seizures included poor efficacy and potential seizure aggravation. Perampanel was among the newer AEDs that are considered to be effective for SGTC and refractory PGTC seizures.

Results from the phase 3 perampanel study 332

Prof Eugen Trinka of the Department of Neurology, Paracelsus Medical University, Salzburg, Austria, presented the results for study 332, a randomised double-blind placebo-controlled study with an open-label extension phase done to evaluate the efficacy and safety of adjunctive perampanel in patients with PGTC seizures in IGE.45 Perampanel is licensed for focal epilepsy with or without secondary generalisation. Perampanel has demonstrated good efficacy in both PGCT and SGTC seizures, suggesting a broad spectrum of action.

Study design
The core study comprised a pre-randomisation phase (screening and baseline; 12 weeks) and randomisation phase (titration and maintenance; 16 weeks). Randomisation was 1:1 for perampanel or placebo. In the extension phase, all patients received perampanel.

The primary objective was to demonstrate efficacy of perampanel as adjunctive treatment compared with placebo in PGTC seizures. Key secondary objectives were to evaluate the safety and tolerability of perampanel in patients with inadequately controlled PGTC seizures and to evaluate efficacy of adjunctive perampanel on other types of generalised seizures (myoclonic, absence, and all seizures). Exploratory objectives included pharmacokinetics, physician-rated Clinical Global Impression of Change, time from first dose to the nth PGTC seizure event (where n is baseline  seizure frequency per 28 days plus 1), pharmacokinetic/ pharmacodynamic modelling of the relationship between perampanel concentrations and efficacy and safety, quality of life, and rates of hospitalisation and/or emergency-room visits.

Key inclusion criteria were age 12 years and older (in Germany and Austria age ≥18 years and in India age <65 years), clinical diagnosis of PGTC seizures with three or more PGTC seizures during the 8-week baseline period, and prior EEG consistent with PGTC seizures or IGE with no features of focal epilepsy. Patients were receiving a fixed dose of one to three concomitant AEDs of which only one was an inducer AED (carbamazepine, oxcarbazepine, or phenytoin). Vagal nerve stimulation was allowed.

The primary endpoints were percent change from baseline in PGTC seizure frequency per 28 days during treatment (non-EU registration) and 50% responder rate (≥50% reduction in PGTC seizure frequency during the maintenance period vs. baseline) [EU registration]. The key secondary endpoints included percent change from baseline in seizure frequency (EU registration), 50% responder rate (non-EU registration), safety and tolerability, and efficacy on other subtypes of generalised seizures and on all seizures.

Importantly, of 307 patients enrolled there were 143 screening failures despite enrolment being done by a high-quality consortium; 117 were due to misinterpretation of the clinical data as PGTC seizures when they were more likely to have been SGTC seizures. Overall, 164 patients were randomised and 163 were treated (82 placebo and 81 perampanel). Approximately 80–90% of patients in each group completed the study period.

There were only 18 adolescent patients and one patient older than 65 years. There
were no differences between the groups in baseline characteristics. All patients had GTC seizures, approximately half of the patients had additional absence seizures, and onethird had myoclonic seizures. Most patients were taking two concomitant AEDs (range, 1–3). The median baseline seizure frequency per 28 days was 2.5 and 2.6 for placebo and perampanel, respectively. This was a high seizure frequency given the risk of SUDEP, particularly for the placebo group.

For PGTC seizures, the change in seizure frequency per 28 days was 76.5% reduction for perampanel and 38.4% reduction for placebo (P<0.0001) in the full analysis set (FAS) [Figure 3]. The median difference

Figure 1: Efficacy of perampanel versus placebo in patients with PGTC seizures. (a) 50% responder rate; and (b) change in seizure frequency per 28 days [reprinted with permission from American Academy of Neurology © 2015].23. PGTC, primary generalised tonic-clonic.

Figure 3: Efficacy of perampanel versus placebo in patients with PGTC seizures. (a) 50% responder rate; and (b) change in seizure frequency per 28 days [reprinted with permission from American Academy of Neurology © 2015].23. PGTC, primary generalised tonic-clonic.

for perampanel compared with placebo was 30.8% (95% confidence interval [CI] 15.2–45.5%). The 50% responder rate was 64.2% for perampanel and 39.5% for placebo (p=0.0019).

For all seizures, including absence and myoclonic seizures, the change in seizure
frequency was 43.4% for perampanel and 22.9% for placebo (p=0.0018). The median difference for perampanel compared with placebo was 23.5% (95% CI 8.5–40.7%). The 50% responder rate was 45.7% for perampanel and 34.6% for placebo (p=0.18).

For absence seizures, the change in seizure frequency was 41.2% for perampanel and 7.6% for placebo, but this was not statistically significant (p=0.35). Similarly, for myoclonic seizures, there was no significant difference between the two groups (perampanel 24.5% and placebo 52.5%; p=0.61). The 50% responder rate for absence seizures was 48.1% for perampanel and 39.4% for placebo (p=0.46), and for myoclonic seizures was 41.7% for perampanel and 60.9% for placebo (p=0.37). The numbers of patients were too low to show meaningful statistical results. Importantly, perampanel did not increase the number of seizures in these groups.

For the exploratory endpoint of median time to pre-randomisation frequency per 28 days (nth +1) for PGTC seizures, there was a significant difference between perampanel and placebo (p<0.0001). The reason for including this endpoint is to avoid prolonged exposure to seizures in patients with a high seizure frequency.

For PGTC seizures in the FAS, seizure freedom (100% reduction in seizure frequency per 28 days) was 30.9% for perampanel and 12.3% for placebo. For patients who completed the maintenance phase, seizure freedom was 36.8% for perampanel and 13.9% for placebo. For all seizures in the FAS, seizure freedom was 23.5% for perampanel and 4.9% for placebo and, for all seizures in the completed maintenance phase, seizure freedom was 27.9% for perampanel and 5.6% for placebo. 

The rate of all treatment-emergent adverse events (TEAEs) was 82.7% for perampanel and 72.0% for placebo and the rate of serious TEAEs was 7.4% for perampanel and 8.5% for placebo. There was one death (1.2%) in each group, neither of which were drug related. The rate of TEAEs leading to study drug dose adjustment was 19.8% for perampanel and 12.2% for placebo. Study drug discontinuation occurred in 11.1% in the perampanel group and 6.1% in the placebo group, but dose reductions were similar in both groups at 9.9% and 7.3%, respectively.

For TEAEs experienced by ≥5% of patients, dizziness was common among the perampanel group at 32.1%, followed by fatigue (14.8%), headache (12.3%), somnolence (11.1%), and irritability (11.1%). Five patients in the perampanel group experienced non-fatal severe AEs of (Medical Dictionary for Regulatory Activities [MedDRA]-preferred term) convulsion, constipation, cholecystitis chronic and status epilepticus, suicide attempt, and suicidal ideation, and six patients in the placebo group experienced convulsion, fall, nausea, grand mal convulsion (twice in one patient), thermal burn, and status epilepticus.

When investigating psychiatric or cognitive issues, it is important to include both narrow and broad search terms to identify all possible cases, even though some may be inconsequential on closer inspection. Hostility- or aggression-related TEAEs were reported by 2.5% of patients in the perampanel group and 0% in the placebo group using the narrow MedDRA term, and by 18.5% of patients in the perampanel group and 4.9% of patients in the placebo group based on the broad MedDRA term. This was mainly due to higher reporting of irritability by patients in the perampanel group (11.1% vs. 2.4% in the placebo group). Serious hostility- or aggression-related TEAEs, were reported for one patient (1.2%) in the perampanel group (fatal drowning, possibly during a fight) and none in the placebo group. Three patients (3.7%) in the perampanel group and one (1.2%) in the placebo group discontinued treatment due to hostility-or aggression-related TEAEs.

For alertness- and cognition-related TEAEs, there was slightly more somnolence in the perampanel group (11.1% vs. 3.7% in the placebo group), but all other TEAEs were similar between the two groups. Psychosis- and psychotic disorder-related TEAEs were slightly more frequent in the perampanel group. Suicidal behaviour-related TEAEs were similar in the two groups. Status epilepticus and convulsion-related TEAEs occurred in three patients in the perampanel group and four patients in the placebo group.

Falls may be associated with either seizure or ataxia, so it is important to exclude those events that occur as part of a seizure. TEAEs related to falls were experienced by two patients (2.5%) in the perampanel group and one (1.2%) in the placebo group. None of the falls led to discontinuation from treatment.

Quality of life and healthcare resource utilization
There was a clinically meaningful change in quality of life in the domains of daily activities, cognition, and distress for patients receiving perampanel, which is likely to be related to the decrease in GTC seizures. For healthcare resource utilization, there were one or more emergency-room visits for 2.5% of patients in the perampanel group and 12.2% in the placebo group. Unscheduled physician visits were similar in the perampanel and placebo groups (6.2% vs. 4.9%), which could be due to AEs or lack of efficacy in the form of a seizure. The median number of visits resulting in hospital admission was 0 for both the perampanel and placebo groups.

In summary
Percent reduction in 28-day average PGTC seizure frequency from baseline increased in the perampanel group in a dose-dependent manner, with the probability of response predicted to increase with an increase in log exposure to perampanel. AEs related to hostility or aggression were associated with perampanel exposure. Quality of life improved and emergency room visits were reduced among patients taking perampanel compared with those taking placebo.

Real-world experience with perampanel in GTC seizures

Dr. Vicente Villanueva of the Neurology Service, Hospital La Fe, Valencia, Spain, summarised the data from the real-world experience with perampanel, with the focus on the FYDATA (Follow-up of 1 Year Data of paTients on perAmpanel) study.46 Evidencebased medicine is very important, but not everything can be demonstrated by evidence. Real-world studies can provide information about populations that are ineligible for inclusion in randomised clinical trials, including results for longer follow-up, dose titration, and AEs. As perampanel has a novel mechanism of action, real-life studies are important.

Table 1 shows the real-life experience with perampanel in focal epilepsy and secondarily generalised seizures.47 Most of the real-life studies have been presented as abstracts, while two have been published.48,49 Additionally, there are two studies in myoclonic seizures and one in Lennox-Gastaut syndrome, which showed that perampanel is efficacious in myoclonic seizures, particularly in patients with progressive myoclonic epilepsy.

Table 1: Real-world studies of perampanel [reprinted with permission from John Wiley and Sons © 2016].47

Germany/Austria2816 monthsCross-sectional observational
Germany (Kork)746 monthsProspective observational
Denmark2212 monthsProspective
Scotland546 monthsProspective observational
Canada93 yearsProspective special access programme
England (Bristol)6014 monthsRetrospective observational
England (Leeds)39NSRetrospective observational
England (Birmingham)167–285 daysProspective observational
England (Manchester)30NSRetrospective observational
Wales36134 (6–431) days3 epilepsy centres
UK/Ireland3101-22 monthsRetrospective observational
Spain46412 monthsRetrospective observational
NS, not specified.   

usually higher than those used in real-world clinical practice. In the clinical trials of
focal epilepsy, the perampanel dose ranged from 4 to 12 mg and that for IGE ranged from 2 to 8 mg. In the real-life studies, the perampanel dose was <8 mg for most of the patients. In the FYDATA study most patients were taking 4 to 8 mg (median 6 mg at 1 year). In the study from Denmark, the mean dose was 6 mg, with approximately 45% of patients taking 4 mg.50 According to the Summary of Product Characteristics, the recommended titration is to increase the dose by 2 mg every 1–2 weeks. This recommendation was followed for most patients in the FYDATA study. However, slower titration, which can be associated with better tolerability, was done for some patients.

In the real-world studies, the seizure-free rate ranged from 3% to 15% and the responder rate ranged from 15% to 89%, but this depended on the seizure type. In the FYDATA study, the median seizure reduction was 33% (last observation carried forward [LOCF]) and 58% (completer group) over 1 year. In secondarily generalised seizures, the median seizure reduction was 75% (LOCF) and 100% (completer group) over 1 year. This result is similar to the 76% reduction in GTC seizures in study 332.45 Thus, clinical trial results seem to be reproducible in real-world clinical practice. In the FYDATA study, according to the LOCF analysis, the rate for >50% reduction in seizures was 26.5%, and 7.2% of patients were seizure free at 1 year. In the completer analysis, the results were a little better at 40.2% for >50% reduction in seizures and 10.3% for seizure freedom. In the study from Germany and Austria, the response rate for secondarily generalised seizures was 57% and the seizure freedom rate was 32%.48 In the UK and Ireland audit, the responder rate for tonic-clonic seizures was 57.5%. Although only eight patients had IGE, six had some improvement and two achieved ≥50% seizure reduction.49 The Dianalund experience was similar, in that 9% of patients were seizure free and 27.2% had ≥50% seizure reduction.50 In a Scottish study, 5.6% of patients were seizure free and 14.8% had ≥50% seizure reduction.51

Enzyme-inducing AEDs
The results in patients who were taking non-enzyme inducing AEDs were better than those for patients taking enzyme-inducers (oxcarbazepine, carbamazepine, and phenytoin). Therefore, a lower dose of perampanel (<6 mg) was required for patients taking a non-enzyme inducer to achieve the same results as those taking an enzyme inducer who needed 6 mg. It has been suggested from the clinical trial results that there is a relationship between increasing perampanel efficacy and blood levels in these patients. However, this result was not reproduced in all the real-world studies.

Adverse events
The rate for AEs was 50–65% in the realworld studies. In the FYDATA study, 62% of patients had AEs. Most of the AEs were reported in the first 6 months of follow-up, possibly because those patients with the worst AEs withdrew from the study within 6 months. The most frequent AEs in the real-world studies were CNS related, and included dizziness, somnolence, and fatigue. Irritability occurred in 18% of patients in the FYDATA study and behavioural AEs occurred in 22% of patients. While this was higher than in the randomised controlled trials, it is likely that the real-life studies included patients with psychiatric comorbidities, many of who are ineligible for randomised clinical trials. Importantly, there were no unexpected AEs that were not reported in the randomised clinical trials. Interestingly, the German/Austrian study had lower psychiatric AEs at 2.8%, and the AEs did not appear to be dose related.

Behavioural AEs were related to prior psychiatric comorbidity, but were not related to concomitant treatment with levetiracetam or to the perampanel dose. However, there was a relationship between behavioural AEs and titration; the number of AEs and psychiatric AEs were reduced with slower titration of 2 mg every 3–4 weeks. Thus, slower titration is recommended for patients who may be more sensitive to AEs or who have psychiatric comorbidities.

Interestingly, the results were better among the elderly population than in patients younger than 65 years (responder rate 36% vs. 26% [p=0.291]; seizure freedom 20% vs 6.5% [p=0.011]). There was no difference in AEs between the older and younger groups. The perampanel dose was slightly lower in the elderly population (4–6 mg).

In summary
The results for perampanel in a real-life study population were similar to those of the randomised clinical trials. Interestingly, there was additional efficacy in SGTC seizures  and in older patients. The main AEs were central nervous system related, and prior psychiatric comorbidity was associated with higher psychiatric AEs. Slower titration can improve the AE profile and administration immediately before sleep can improve tolerability and minimise dizziness and somnolence.


1. Hanada T, et al. Epilepsia. 2011;52:1331–40.
2. Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised clinical and electrographic classification of epileptic seizures. In: Wyllie E, ed. The Treatment of Epilepsy: Principles and Practice. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2001:291–97.
3. Wheless JW, et al. Epilepsia. 2002;43(Suppl 3):33–52.
4. Zifkin B, et al. In: Engel J Jr, et al, eds. Epilepsy: a comprehensive textbook. Vol 1.
Philadelphia, Pa: Lippincott-Raven Publishers; 1997:567–77.
5. Fisch BJ, Olejniczak PW. Generalized tonic-clonic seizures. In: Wylie E, ed. The Treatment of Epilepsy: Principles and Practice. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2001:369–393.
6. Hauser WA, et al. Epilepsia. 1975;16:1–66.
7. Duron RM, et al. Epilepsia. 2005;46:34–47.
8. Wyllie E. Epidemiological aspects of epilepsy. Wyllie’s treatment of epilepsy.  Amsterdam; Wolters Kluwer: 2015.
9. Hauser WA, et al. Mayo Clin Proc. 1996;71:576–86.
10. Mac TL, et al. Lancet Neurol 2007;6:533–43.
11. Glauser T, et al. Epilepsia. 2013;54:551–63.
12. Rheims S, et al. Exp Opin Pharmacother. 2014;15:1417–26.
13. Perucca E, at al. Epilepsia. 1998;39:5–17.
14. Thomas P, et al. Brain. 2006;129:1281–92.
15. Shields WD, et al. Neurology. 1983;33:1487–9.
16. NICE clinical guideline 137. Updated 2016. www.nice.org.uk/guidance/cg137
17. SIGN national clinical guideline 70. Updated 2005. http://www.sign.ac.uk/guidelines/published/numlist.html
18. Mohanraj R, et al. Eur J Neurol. 2006;13: 277–82.
19. Laxer KD et al. Epilepsy Behav 2014;37:59–70.
20. Tomson T, et al. Epilepsia. 2005;46 (Suppl 11):54–61.
21. Monte CP, et al. Seizure. 2007;16:1–7.
22. Hesdorffer DC, et al. Epilepsia. 2012;53: 249–52.
23. Thurman DJ, et al. Epilepsia. 2014;55: 1479–85.
24. Hesdorffer DC, et al. Epilepsia. 2011;52: 1150–9.
25. Monté CP, et al. Seizure. 2007;16:1–7.
26. Ding D, et al. Epilepsia, 2013;54:512–7.
27. Tomson T, et al. Epilepsy Res. 2004;60:1–16.
28. Asadi-Pooya AA, et al. Seizure. 2012;21:165–8.
29. Henkin Y, et al. Dev Med Child Neurol. 2005;47:126–32.
30. Thompson PJ, et al. Epilepsia. 2005;46: 1780–7.
31. Gupta S, et al. Epilepsy Behav. 2016;55: 146–56.
32. Spenser SS. Epilepsia. 2002;43:219–27.
33. Bertram EH. Exp Neurol. 2013;244:67–74.
34. Avoli M. Epilepsia. 2012;53:779–89.
35. Meeren H, et al. Arch Neurol. 2005;62:371–6.
36. Collaborative Study Group. Epilepsia. 1987:28(Suppl 2):8–11.
37. Marson AG, et al. Lancet. 2007:369:1016–26.
38. Trinka E, et al. J Neurol Neurosurg Psychiatry. 2013:84:1138–47.
39. Rosenow F, et al. J Neurol Neurosurg Psychiatry. 2012:83;1093–8.
40. Tudur Smith C, et al. Trials. 2007:8:34.
41. Mattson RH, et al. NEJM. 1985;313:145–51.
42. Mattson RH, et al. NEJM. 1992; 327:765–71.
43. Kwan P, et al. Lancet Neurol .2011;10:881–91.
44. Krauss GL, et al. Epilepsia. 2014;55:1058–68.
45. French JA, et al. Neurology. 2015;85:950–7.
46. Villanueva V, et al. Epilepsy Res. 2016;126:201–10.
47. Trinka E, et al. Acta Neurol Scand. 2016;133:160–72.
48. Steinhoff BJ, et al. Epilepsy Res. 2014;108:986–8.
49. Shah E, et al. Seizure. 2016;34:1–5.
50. Juhl S, et al. Acta Neurol Scand. 2016. doi: 10.1111/ane.12558.
51. Brodie MJ, et al. Epilepsy Behav. 2016;54: 100–3