Efficacy of perampanel: a review of clinical trial data
F. Kerling, Department of Neurology, University of Ulm, Oberer Eselsberg 45, 89081 Ulm, Germany
Tel.: +49 9128 50 48460
Fax: +49 9128 50 43154
The efficacy of adjunctive perampanel has been investigated in an extensive clinical development program across a broad, multinational population of patients with refractory partial-onset seizures. Further to the results of two Phase II dose-finding studies, perampanel was evaluated in three large Phase III registration studies at the predicted no-effect dose of 2 mg/day and the predicted effective doses of 4, 8, and 12 mg/day. In all three studies, perampanel 4, 8, and 12 mg/day consistently provided significant reductions in the frequency of partial-onset seizures compared with placebo. Improvements in responder rates and seizure freedom rates were also observed. In addition, data from recent interim analyses of extension studies have indicated that these efficacy outcomes may be maintained with long-term treatment. Overall, these studies form a solid evidence base to support the efficacy of adjunctive perampanel in the treatment of refractory partial-onset seizures.
The α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor antagonist perampanel has been evaluated as an adjunctive treatment for refractory partial-onset seizures in an extensive clinical development program that has included two Phase II dose-finding studies and three multicenter, randomized, double-blind, placebo-controlled Phase III registration studies. In addition, two extension studies are currently ongoing, with the results of interim analyses recently reported. This review article will summarize and discuss the key efficacy data from these studies of perampanel. Safety data are reviewed in the accompanying article by Serratosa et al. .
Phase II dose-finding studies
Two randomized, double-blind, placebo-controlled Phase II studies evaluated a range of doses of adjunctive perampanel for the treatment of refractory partial-onset seizures (studies 206 and 208) . While the studies were primarily designed to evaluate safety and tolerability, they were also intended to guide the selection of the minimally effective and maximum tolerated doses of perampanel for subsequent Phase III clinical trials and to examine pharmacokinetic (PK) and pharmacodynamic (PD) relationships.
Both studies enrolled adults aged 18–70 years with a diagnosis of epilepsy and uncontrolled partial-onset seizures (with or without secondary generalization), despite treatment with ≥3 different antiepileptic drugs (AEDs) within the previous 2 years. In study 206 (proof-of-concept study), patients were randomized to receive placebo (n = 51), perampanel once daily (QD; n = 51), or perampanel twice daily (BID; n = 51). Perampanel doses were increased to a maximum of 4 mg/day (4 mg QD or 2 mg BID) over an 8-week titration period, with the highest tolerated dose continued for an additional 4 weeks (maintenance period). Study 208 was subsequently designed to explore a higher dose range: patients were randomized to receive placebo (n = 10) or perampanel QD (n = 38). Perampanel doses were increased up to a high dose of 12 mg/day over a 12-week titration period, which was then followed by a 4-week maintenance period .
While these Phase II studies were not powered to statistically assess efficacy, there was an indication of efficacy against partial-onset seizures according to data obtained from patient diaries. In study 206, the responder rate (percentage of patients with a ≥50% reduction in seizure frequency; maintenance vs baseline) was 30.7% with perampanel (QD and BID combined) and 21.6% with placebo. The median reduction in seizure frequency per 28 days was 25.7% for perampanel-treated patients and 19.5% for placebo-treated patients. In study 208, the responder rate (overall treatment phase [dose titration plus maintenance] vs baseline) was 39.5% with perampanel and 22.2% with placebo. The median reduction in seizure frequency per 28 days was 39.6% for perampanel-treated patients, while a median increase of 2.1% occurred in the placebo group .
A pooled PK/PD analysis of data from these two Phase II studies also showed that seizure frequency decreased and the probability of response increased with increasing exposure to perampanel . This, combined with the preliminary signal of efficacy at low doses in study 206 and tolerability data for doses up to 12 mg/day in study 208 [see accompanying article by Serratosa et al. ], guided the selection of dose ranges for further evaluation in the subsequent Phase III registration studies of perampanel.
Phase III registration studies
The efficacy of adjunctive perampanel at the predicted no-effect dose of 2 mg/day and the predicted effective doses of 4, 8, and 12 mg/day were further evaluated in three multinational, multicenter, randomized, double-blind, placebo-controlled Phase III registration studies (studies 304 [NCT00699972] , 305 [NCT00699582] , and 306 [NCT00700310] ). These studies enrolled refractory patients aged ≥12 years, who were experiencing simple partial seizures with motor signs or complex partial-onset seizures (with or without secondary generalization) despite receiving 1–3 AEDs. Following a 6-week baseline period, patients were randomized to receive once-daily placebo, perampanel 8 or 12 mg (studies 304 and 305), or placebo, perampanel 2, 4, or 8 mg (study 306) over a 19-week double-blind treatment phase (comprising a 6-week titration period [weekly dose adjustments in 2 mg/day increments] and 13-week maintenance period). Patients used a daily diary to record partial seizures and their types (simple partial seizures with/without motor signs, complex partial seizures with/without secondary generalization [CP+SG seizures], or complex partial seizures with secondary generalization [SG seizures]).
In total, 1478 patients were included in the intent-to-treat (ITT) populations of studies 304 (n = 387), 305 (n = 386), and 306 (n = 705). Across all three studies, randomized doses of perampanel 4, 8, or 12 mg/day were associated with significantly greater median reductions in the frequency of all partial-onset seizures (primary endpoint outside the EU in all studies) and also CP+SG seizures per 28 days compared with placebo (double-blind treatment phase vs baseline; Table 1) [4-6]. Furthermore, responder rates (maintenance vs baseline; primary endpoint in the EU) were also significantly greater with perampanel 4, 8, or 12 mg/day than with placebo in studies 305 (8, 12 mg) and 306 (2, 4, 8 mg) (Table 1). While there were no statistically significant differences in overall responder rates in study 304, a subanalysis indicated that this may have been related to regional variations: significant differences were observed in patients from North America (perampanel 8 mg/day, 40.5%; perampanel 12 mg/day, 40.0%; placebo, 21.9%; P < 0.05 for both comparisons), but not in those from Central/South America (perampanel 8 mg/day, 33.9%; perampanel 12 mg/day, 30.2%; placebo, 33.3%) . While these differences remain unexplained, the responder rate with placebo was substantially higher than expected in the Central/South American patients, while responder rates for patients receiving perampanel were approximately equivalent across the regions.
Table 1. Summary of key efficacy data from the randomized, double-blind, placebo-controlled Phase III perampanel registration studies 304, 305, and 306
|Median % change in frequency of all partial-onset seizures per 28 days (baseline vs double-blind treatment phase)a||−21.0||−26.3*||−34.5*||−9.7||−30.5***||−17.6*||−10.7||−13.6||−23.3**||−30.8***|
|Median % change in frequency of CP+SG seizures per 28 days (baseline vs double-blind treatment phase)||−17.9||−33.0**||−33.1**||−8.1||−32.7***||−21.9**||−17.6||−20.5||−31.2**||−38.7***|
|Responder rate (baseline vs maintenance period), %b||26.4||37.6||36.1||14.7||33.3**||33.9***||17.9||20.6||28.5*||34.9***|
|Seizure freedom rate (completer analysis), %,c,d||0||2.6||2.0||1.7||2.8||6.5||1.2||1.9||4.4||4.8|
|Seizure freedom rate (pragmatic analysis), %c,e||0||2.2||1.5||1.5||2.3||5.0||NR||NR||NR||NR|
The percentages of patients who were seizure free during the maintenance periods of the studies were also generally greater with perampanel 4, 8, or 12 mg/day than placebo in protocol-specified analyses of patients who completed the maintenance period (Table 1) [4-6]. A more conservative analysis of seizure-free patients, who completed the maintenance period, using the total ITT population as a larger denominator , demonstrated similar results (Table 1). In the two high-dose studies including perampanel doses up to 12 mg/day, there appeared to be some benefit of the 12 mg/day dose over the 8 mg/day dose in terms of seizure freedom in study 305 (statistical significance not evaluated), but not in study 304 (although data from this study may again have been affected by the relatively high placebo responder rates in Central and South American populations).
Pooled analyses of Phase III registration studies
The efficacy of perampanel has also been explored in post hoc analyses of pooled data from the Phase III registration studies by randomized dose [8, 9]. This approach resulted in a large data set, based on a pooled population of 1480 randomized and treated patients.
In general, the pooled analysis supported the individual study data [8, 9]. Median change in partial-onset seizure frequency per 28 days was significantly improved with perampanel 4, 8, or 12 mg/day, compared with placebo (perampanel 4 mg/day, −23.3%; 8 mg/day, −28.8%; 12 mg/day, −27.2%; placebo, −12.8%; P < 0.01 for all doses vs placebo) . Similarly, responder rates were significantly improved with perampanel 4, 8, or 12 mg/day (perampanel 4 mg/day, 28.5%; 8 mg/day, 35.3%; 12 mg/day, 35.0%; placebo, 19.3%; P < 0.05 for all doses vs placebo) . These improvements remained consistent across subgroups of patients receiving any of the four most commonly administered concomitant AEDs (carbamazepine, valproic acid, lamotrigine, levetiracetam), although the effects were diminished in the presence of the inducer carbamazepine, compared with the non-inducer medications [8, 9]. This is presumably due to a reduction in perampanel exposure in the presence of inducer AEDs, as indicated by PK/PD analyses of the pooled Phase III data .
In addition, a pooled analysis of data from studies 304 and 305 based on actual dose received has indicated that the 12 mg/day dose may confer additional efficacy benefits beyond those achieved with 8 mg/day in terms of seizure-free status and change in the frequency of SG seizures . This is consistent with pooled PK/PD analyses of studies 304, 305, and 306, which demonstrate a linear exposure–efficacy relationship across an actual dose range of 2–12 mg .
Extension studies are providing insight into the long-term efficacy of perampanel. Studies 207 (NCT00368472)  and 307 (NCT00735397)  were long-term safety extension studies with a secondary objective of evaluating the maintenance of efficacy of adjunctive perampanel in the treatment of refractory partial-onset seizures.
Study 207 enrolled 138 patients who had completed either of the two Phase II perampanel dose-finding studies . Following a 12-week titration period, patients entered a maintenance period, during which they could receive their optimum dose of perampanel (up to 12 mg/day), and concomitant AEDs, for up to 424 weeks (approximately 8 years). At the time of an interim analysis, performed at approximately 4 years after the study start, the mean (standard deviation) duration of perampanel exposure was 116.4 (74.9) weeks, 89.1% of patients had achieved a maximum dose within the effective dose range (4–12 mg/day), and 38.4% of patients remained on perampanel. The overall median change in the frequency of all partial-onset seizures per 28 days was −31.5% (entire duration of open-label treatment vs pre-perampanel baseline). For patients who received at least 1 (n = 89), 2 (n = 66), 3 (n = 52), or 4 (n = 18) years of open-label perampanel treatment, the median percentage change from pre-perampanel baseline in 28-day seizure frequency over 1, 2, 3, or 4 years was −43.7%, −52.0%, −49.7%, or −48.4%, respectively. The responder rates for these patients at the end of 1, 2, 3, or 4 years were 43.8%, 51.5%, 49.0%, and 50.0%, respectively.
In study 307, 1218 patients who had completed the double-blind treatment period of any of the three Phase III registration studies were titrated to their maximum tolerated dose of perampanel (up to 12 mg/day) over a 16-week blinded conversion period . These patients then entered an open-label maintenance period with a planned duration of 256 weeks (approximately 5 years). An interim analysis was performed at approximately 2 years after the study start, when the median (range) duration of perampanel exposure was 51.4 (1.1–128.1) weeks. At this time, the majority of patients (91.4%) had been titrated to perampanel 10 or 12 mg and 70.8% of patients remained on perampanel. Median changes from pre-perampanel baseline in the frequency of all partial-onset seizures per 28 days were −39.2% over weeks 14–26 (n = 1114), −46.5% over weeks 40–52 (n = 731), and −58.1% over weeks 92–104 (n = 59).
An analysis of patients who had their actual dose increased from 8 mg/day, in studies 304, 305, or 306, to 12 mg/day in the blinded conversion period of extension study 307 has supported the additional efficacy benefits of the 12 mg/day dose over the 8 mg/day dose in terms of percentage change in seizure frequency and responder rates . When perampanel dose was increased from 8 to 12 mg/day in these patients (n = 209), median change in seizure frequency per 28 days from baseline increased from −32.4% to −43.3% and responder rate increased from 37.8% to 43.5%.
A range of well-designed clinical trials are providing a large and consistent evidence base for the efficacy of adjunctive perampanel in the treatment of refractory partial-onset seizures. These trials have followed a rational progression, with initial experience from Phase II studies used to guide the dose range for further evaluation in three randomized, double-blind, placebo-controlled Phase III registration studies. This has led to the demonstration of significant reductions in seizure frequency and improvements in responder rates, with adjunctive perampanel 4, 8, or 12 mg/day compared with placebo across large patient populations.
One inconsistency in the Phase III efficacy data was the lack of statistical significance for the improvements in responder rates with perampanel compared with placebo in study 304. As described, this may have been related to the unexpectedly high placebo response observed in patients from Central or South America, perhaps reflecting regional variations in patient selection or study conduct.
Studies 304 and 305 both evaluated perampanel doses of up to 12 mg/day, but only study 305 indicated benefit of 12 mg/day over the 8 mg/day dose (seizure freedom rates were greater with 12 mg/day vs 8 mg/day, although no statistical analyses were performed). However, analyses of randomized dose may underestimate the efficacy of higher doses of AEDs, as data may have been confounded by patients who did not achieve the target randomized dose. Pooled analyses of data from studies 304 and 305 based on actual dose received  and analyses of patients who had their actual dose increased from 8 mg/day, in studies 304, 305, or 306, to 12 mg/day in the extension study 307  have indicated that the 12 mg/day dose may indeed confer additional efficacy benefits beyond those achieved with 8 mg/day. This is consistent with PK/PD analyses, which demonstrate a linear exposure–efficacy relationship across an actual dose range of 2–12 mg . The perampanel dose should be individually titrated in each patient until an effective and tolerated dose is achieved. Some patients will benefit from an increase in dose from 8 to 12 mg/day, while others may experience additional AEs at higher doses. This may be affected by several factors such as the refractoriness of an individual patient's seizures to treatment, the presence of inducing drugs, or the variation in individual exposures at a given dose.
In summary, clinical trials have demonstrated that adjunctive perampanel 4, 8, or 12 mg/day can significantly reduce seizure frequency and improve responder rates in patients with refractory partial-onset seizures. Furthermore, interim analyses of extension studies have indicated that efficacy outcomes may be maintained over long-term settings. These studies also demonstrate how careful dose titration can be used to optimize treatment for each individual patient. When considered in conjunction with safety data [see accompanying article by Serratosa et al. ], such clinical experience supports the clear potential of perampanel as a valuable adjunctive treatment for patients with refractory partial-onset seizures.
David Squillacote, MD, of Eisai Inc. assisted in the preparation of the article by providing up-to-date data access and editorial support. Editorial support was provided by Hannah FitzGibbon, PhD, of Complete Medical Communications and was funded by Eisai Inc.
Conflicts of interest and sources of funding statement
The studies described were funded by Eisai Inc. F. Kerling has received honoraria from Desitin, Eisai, GlaxoSmithKline, Janssen Cilag, Pfizer, and UCB for participation in advisory boards or symposia. B. Kasper reports no disclosures.