Narrative Review: Efficacy, Safety, and Tolerability of Newer Third-Generation Antiseizure Medications for Focal Epilepsy: A New Benchmark in Treatment Outcomes?

Abstract

Introduction

Focal seizures are the most common form of epilepsy. This narrative review analyzes the efficacy, safety, and tolerability of five newer third-generation antiseizure medications (ASMs) approved for treatment of focal seizures: lacosamide, perampanel, eslicarbazepine acetate, brivaracetam, and cenobamate.

Methods

A search of PubMed and ClinicalTrials.gov identified phase 2/3 randomized controlled trials and long-term open-label extension (OLE) studies for the five most recently approved ASMs for focal seizures. Efficacy data were summarized, including median percent reduction in seizure frequency, ≥ 50% responder rates, ≥ 75% responder rates (when reported), and seizure freedom rates. Safety and tolerability were assessed based on treatment-emergent adverse events (TEAEs), serious adverse events, and discontinuation rates. Long-term efficacy, safety, and treatment retention rates were also summarized.

Results

All five ASMs had statistically significant efficacy vs. placebo in short-term trials, with cenobamate demonstrating higher median percent seizure reduction (35.5–100%), ≥ 50% responder rates (40–81.6%), and seizure freedom rates (up to 52.4%) compared to the other four third-generation ASMs. In long-term OLEs, cenobamate had ≥ 50% responder rates > 70% and the highest estimated treatment retention rates at 3 years (61–65%) compared to lacosamide (51–53%), perampanel (46%), and brivaracetam (48%). Common TEAEs across all ASMs included dizziness, somnolence, fatigue, and headache. Brivaracetam and lacosamide had favorable profiles in multiple network meta-analyses. Specific safety considerations include psychiatric adverse events for perampanel and brivaracetam, and a risk of drug reaction with eosinophilia and systemic symptoms (DRESS) for cenobamate (significant risk reduction achieved using a gradual titration schedule).

Conclusion

All third-generation ASMs were effective adjunctive therapies. Cenobamate was associated with higher responder and seizure freedom rates and higher long-term retention rates. Brivaracetam and lacosamide had favorable tolerability profiles. The notable seizure freedom rates with cenobamate suggest a potential new benchmark in focal seizure treatment. Comparator trials are needed to confirm possible efficacy and tolerability differences among third-generation ASMs.

Key Summary Points

Why carry out this study?

Seizure freedom rates with ASMs have stagnated for decades. Improved efficacy with cenobamate may signal a new treatment breakthrough.

What was learned from this study?

The efficacy and safety of newer third-generation antiseizure medications were reviewed.

Among the newer drugs, brivaracetam and lacosamide had favorable tolerability profiles in multiple meta-analyses.

Cenobamate was associated with higher responder and seizure freedom rates in clinical trials and high retention rates.

The substantially higher seizure freedom rates associated with cenobamate may suggest a new benchmark in the treatment of focal seizures.

Introduction

Epilepsy is a chronic neurological disorder characterized by the presence of seizures and seizure-associated comorbidities. The estimated prevalence is 5–10 per 1000 people, with approximately 70 million people affected worldwide [1]. Among newly diagnosed patients, focal seizures predominate, accounting for ≥ 60% of diagnoses [2, 3]. Despite the development of more than 20 new antiseizure medications (ASMs) over the past 30 years, the proportion of patients with epilepsy categorized as treatment-resistant had remained unchanged at roughly 35% [4, 5]. In the past decade, the approval of several third-generation ASMs has generated renewed optimism that a breakthrough in epilepsy drug treatment may be forthcoming. These ASMs (not including those limited to treatment of rare epilepsy syndromes) include lacosamide (LCM), eslicarbazepine acetate (ESL), perampanel (PER), brivaracetam (BRV), and cenobamate (CNB). In terms of treatment outcomes (including seizure freedom), CNB appears the most promising.

ASMs are frequently categorized into three generations of drugs. Different methods may be used to define the “generation” of ASMs, such as by discovery date of the molecular entity or by the ASM’s mechanism of action (MOA). In this review we will define each generation by the regulatory approval date, since the date of approval represents the time when the ASM became part of a clinician’s therapeutic toolkit [6, 7].

Many first-generation ASMs (e.g., phenobarbital, phenytoin, carbamazepine, and valproate), which were discovered largely by chance or through empirical phenotypic screening in animal seizure models, are still used for the treatment of focal seizures [8, 9]. Many second-generation ASMs that were approved in the 1990s–2000s were discovered either serendipitously (e.g., topiramate) or through traditional phenotypic seizure-model screening (e.g., felbamate, gabapentin, lamotrigine, oxcarbazepine, zonisamide, and levetiracetam), whereas tiagabine and vigabatrin were developed through mechanism-based strategies aimed at potentiating GABAergic transmission [9]. They generally are characterized by efficacy approximately similar to first-generation ASMs, though head-to-head comparator trials are rare. However, on the basis of real-world data, selected second-generation ASMs appear to have more favorable safety profiles, and in some cases, MOAs that differ from first-generation ASMs. Gabapentin and levetiracetam had new MOAs identified serendipitously—binding the alpha-2 delta receptor at presynaptic calcium channels and synaptic vesicle protein (SV2A), respectively [10, 11].

Third-generation ASMs are those approved for focal seizures in the mid-2000s and 2010s [12–14], and include rufinamide, pregabalin, LCM, ezogabine, PER, ESL, BRV, everolimus, and CNB. Third-generation ASMs for focal seizures, such as LCM, ESL, and BRV, were mostly developed using a mixture of strategies, including phenotypic screening and structural modification, with the dual aim of targeting specific MOAs that had demonstrated efficacy in first- or second-generation drug classes, but with generally improved safety and tolerability profiles. Some third-generation ASMs, including perampanel (an AMPA receptor antagonist), ezogabine (a Kv7 potassium channel opener), and everolimus (an mTOR pathway inhibitor) were guided by emerging pathophysiologic insights, but still advanced through conventional in vivo screening approaches [9, 15–17]. CNB, one of the most recently approved third-generation ASMs for focal-onset seizures, was likewise identified through preclinical anticonvulsant screening. Its dual mechanism of action, including preferential inhibition of persistent sodium current combined with extrasynaptic tonic inhibition mediated by positive allosteric modulation of gamma-aminobutyric acid type A (GABA_A) receptors, was characterized in subsequent mechanistic studies [18–20]. High seizure-free rates reported in clinical trials with newer third-generation ASMs, specifically CNB, have raised the question of whether we may have reached a new benchmark in the treatment of focal seizures.

The purpose of this review is to describe in detail the five newer third-generation ASMs approved as treatments for focal seizures (LCM, PER, ESL, BRV, and CNB), including their MOAs and safety and efficacy data from phase 2/3 randomized clinical trials ([RCTs] including seizure frequency reduction, responder rate, and seizure freedom rate), prevalence of treatment-emergent adverse events (TEAEs), and long-term efficacy and safety/tolerability. Meta-analyses of pooled studies are also discussed.

Methods

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors. We searched the US National Institutes of Health Clinical Trials Registry (http://www.clinicaltrials.gov) and Medline (accessed via PubMed). The search terms were (lacosamide OR perampanel OR eslicarbazepine OR brivaracetam OR cenobamate) AND (epilepsy OR seizure) and (focal OR focal onset OR partial OR partial onset). In PubMed we searched for article types “clinical trial” or “randomized controlled trial.” Articles were included that represented phase 2 or 3 RCTs of ASM adjunctive treatment. Randomized trials and open-label studies conducted in pediatric-only populations were not included. Additional searches on specific adverse events of interest, sudden unexpected death in epilepsy (SUDEP), and meta-analyses were performed. The literature review was last updated on April 30, 2025.

For short-term trials we report the median percent reduction in seizure frequency per 28 days, the proportion of patients achieving ≥ 50% and ≥ 75% (when available) reduction in seizure frequency, and proportion achieving 100% reduction in seizure frequency (seizure freedom/seizure-free). In long-term open-label extension (OLE) trials we report the proportion of patients achieving ≥ 50% reduction in seizure frequency at set intervals (e.g., year 1, year 3), and proportion achieving 100% reduction in seizure frequency at set intervals (e.g., 6 months, 12 months).

The safety and tolerability outcomes included the top five most common TEAEs, the proportion of patients reporting a serious adverse event (SAE), discontinuation due to TEAEs, adverse events of special interest, and number of deaths during both short-term and long-term treatment studies. Rates of long-term treatment retention are also reported.

Results

Third-Generation ASMs in Focal Seizures: MOA and Short-Term RCTs

Detailed efficacy and safety results are shown in Tables 2 and 3, respectively. All the newer ASMs demonstrated significant reductions in seizure frequency vs. placebo, as assessed by median seizure reduction and ≥ 50% responder rates.

Table 2.

Efficacy outcomes for RCTs of third-generation ASMs

ASM	Study#; NCT#	Na	Median % reduction in seizure frequency per 28 days	≥ 50% seizure frequency reduction	≥ 75% seizure frequency reduction	100% seizure frequency reduction
Lacosamide (LCM)	SP667 [24]	415	200 mg: 14.6% (NS) 400 mg: 28.4% (p = 0.0023) 600 mg: 21.3% (p = 0.0084) PBO: NA	200 mg: 32.7% (NS) 400 mg: 41.1% (p = 0.0038) 600 mg: 38.1% (p = 0.0141) PBO: 21.9%	Completer population, n = 312 200 mg: 11.2% (NS) 400 mg: 22.4% (p = 0.0020) 600 mg: 16.2% (p = 0.0334) PBO: 6.3%	Completer population, n = 312 200 mg: 1.1% 400 mg: 6.3% 600 mg: 1.6% PBO: 0%
	SP754; NCT00136019 [25]	402	400 mg: 37.3% (p = 0.008) 600 mg: 37.8% (p = 0.006) PBO: 20.8%	400 mg: 38.3% (p < 0.001) 600 mg: 41.2% (p < 0.001) PBO: 18.3%	Completer population, n = 317 400 mg: 20.4% (p = 0.005) 600 mg: 21.6% (p = 0.007) PBO: 7.7% (ITT)	Completer population, n = 317 400 mg: 2.5% 600 mg: 8.1% PBO: 0%
	SP755; NCT00220415 [26]	477	200 mg: 35.3% (p = 0.02) 400 mg: 36.4% (p = 0.03) PBO: 20.5%	200 mg: 35.0% (NS) 400 mg: 40.5% (p = 0.01) PBO: 25.8%	NR	Completer population, n = 403 200 mg: 3.6% 400 mg: 2.4% PBO: 2.1%
	EP0008: NCT01710657 [27]	544	200 mg: 29.4% (p < 0.001) 400 mg: 39.6% (p < 0.001) PBO: NAb	200 mg: 38.5%c 400 mg: 49.2% PBO: 19.7%	200 mg: 21.4%c 400 mg: 24.6% PBO: 8.7%	200 mg: 4.7% 400 mg: 5.4% PBO: 0%
Perampanel (PER)	304; NCT00699972 [28]	387	8 mg: 26.3% (p = 0.0261) 12 mg: 34.5% (p = 0.0158) PBO: 21.0%	8 mg: 37.6% (NS) 12 mg: 36.1% (NS) PBO: 26.4%	8 mg: 18.8% (p = 0.001)d 12 mg: 17.3% (p = 0.001) PBO: 5%	8 mg: 2.2% 12 mg: 1.5% PBO: 0%
	305; NCT00699582 [29]	386	8 mg: 30.5% (p = 0.0008) 12 mg: 17.6% (p = 0.0105) PBO: 9.7%	8 mg: 33.3% (p = 0.002) 12 mg: 33.9% (p < 0.001) PBO: 14.7%	8 mg: 15.5% 12 mg: 16.5% PBO: 4.4%	8 mg: 2.3% 12 mg: 5.0% PBO: 1.5%
	306; NCT00700310 [30]	705	2 mg: 13.6% (NS) 4 mg: 23.3% (p = 0.003) 8 mg: 30.8% (p < 0.001) PBO: 10.7%	2 mg: 20.6% (NS) 4 mg: 28.5% (p = 0.013) 8 mg: 34.9% (p < 0.001) PBO: 17.9%	NR	Completer population, n = 623 2 mg: 1.9% 4 mg: 4.4% 8 mg: 4.8% PBO: 1.2%
	335; NCT01618695 [31]	704	4 mg: 17.3% (NS) 8 mg: 29.0% (p = 0.003) 12 mg: 38.0% (p < 0.0001) PBO: 10.8%	4 mg: 23.0% (NS) 8 mg: 36.0% (p = 0.0005) 12 mg: 43.3% (p < 0.001) PBO: 19.4%	NR	4 mg: 2.9% (NS) 8 mg: 4.0% (NS) 12 mg: 4.4% (p = 0.037) PBO: 0.6%
Eslicarbazepine acetate (ESL)	301; NCT00957684 [32]	397	LS mean difference vs. PBO 400 mg: − 0.08 (NS) 800 mg: − 0.19 (p = 0.0028) 1200 mg: − 0.22 (p = 0.0003)	400 mg: 23% (NS) 800 mg: 34% (p = 0.0359) 1200 mg: 43% (p = 0.0009) PBO: 20%	NR	400 mg: 2% (NS) 800 mg: 4% (NS) 1200 mg: 8% (p < 0.05) PBO: 2%
	302; NCT00957047 [33]	245	LS mean difference vs. PBO 800 mg: − 1.6 (p = 0.048) 1200 mg: − 1.9 (p = 0.021)	800 mg: 34.5% (NS) 1200 mg: 37.7% (p = 0.020) PBO: 22.6%	NR	800 mg: 4.8% (NS) 1200 mg: 3.9% (NS) PBO: 1.2%
	303; NCT00957372 [34]	393	400 mg: 18.7% (NS) 800 mg: 32.6% (p < 0.001) 1200 mg: 32.8% (p < 0.001) PBO: 0.8%	400 mg: 17% (NS) 800 mg: 40.0% (p < 0.001) 1200 mg: 37.1% (p < 0.01) PBO: 13.0%	400 mg: 2.1% (NS) 800 mg: 16.0% (p < 0.001) 1200 mg: 16.5% (p < 0.001) PBO: 4%	400 mg: 1.0% (NS) 800 mg: 8.0% (p < 0.05) 1200 mg: 4.1% (NS) PBO: 1.0%
	304; NCT00988429 [35]	640	800 mg: 29.7% (NS) 1200 mg: 35.6% (p = 0.02) PBO: 21.8%	800 mg: 30.5% (NS) 1200 mg: 42.6% (p < 0.001) PBO: 23.1%	NR	800 mg: 2.0% (NS) 1200 mg: 2.2% (NS) PBO: 0.9%
Brivaracetam (BRV)	N01114; NCT00175929 [39]	157	Reporting seizure frequency reductions/week over PBO 50 mg: 14.7% (NS) 150 mg: 13.6% (NS)	50 mg: 39.6% (NS) 150 mg: 33.3% (NS) PBO: 23.1%	NR	50 mg: 9.4% 150 mg: 5.8% PBO: 1.9%
	N01193; NCT00175825 [40]	208	Reporting seizure frequency reductions/week over PBO 5 mg: 9.8% (NS) 20 mg: 14.9% (NS) 50 mg: 22.1% (p = 0.004)	5 mg: 32.0% (p = 0.047) 20 mg: 44.2% (p < 0.002) 50 mg: 55.8% (p < 0.001) PBO: 16.7%	NR	5 mg: 8.0% (NS) 20 mg: 7.7% (NS) 50 mg: 7.7% (NS) PBO: 1.9%
	N01252; NCT00490035 [36]	398	Seizure frequency reductions/28 days over placebo 20 mg: 10.2% (NS) 50 mg: 9.2% (NS) 100 mg: 20.5% (p = 0.010) Seizure frequency reductions/week 20 mg: 30.0% (p = 0.019) 50 mg: 26.8% (NS) 100 mg: 32.5% (p = 0.004) PBO: 17.0%	20 mg: 27.3% (p = 0.339) 50 mg: 27.3% (NS) 100 mg: 36.0% (p = 0.023) PBO: 20.0%	NR	20 mg: 2.0% 50 mg: 0 100 mg: 4.0% PBO: 0%
	N01253; NCT00464269 [37]	396	Seizure frequency reductions/28 days over placebo 5 mg: 2.6% 20 mg: 8.7% 50 mg: 22% Seizure frequency reductions/week 5 mg: 20.0% (p = 0.991) 20 mg: 22.5% (p = 0.386) 50 mg: 30.5% (p = 0.003) PBO: 17.8%	5 mg: 21.9% (NS) 20 mg: 23.2% (NS) 50 mg: 32.7% (p = 0.008) PBO: 16.7%	NR	5 mg: 1.1% 20 mg: 1.0% 50 mg: 4.0% PBO: 0%
	N01358; NCT01261325 [38]	760	100 mg: 37.2% (p < 0.001) 200 mg: 35.6% (p < 0.001) PBO: 17.6%	100 mg: 38.9% (p < 0.001) 200 mg: 37.8% (p < 0.001) PBO: 21.6%	NR	100 mg: 5.2% (p = 0.003) 200 mg: 4.0% (p = 0.019) PBO: 0.8%
Cenobamate (CNB)	C013; NCT01397968 [41]	221	CNB: 55.6% (p < 0.0001) PBO: 21.5%	CNB: 50.4% (p < 0.0001) PBO: 22.2%	CNB: 38.7% (p = 0.0019) PBO: 20.6%	CNB: 28.3% (p = 0.0001) PBO: 8.8%
	C017; NCT01866111 [42]	434	100 mg: 35.5% (p = 0.0071) 200 mg: 55.0% (p < 0.0001) 400 mg: 55.0% (p < 0.0001) PBO: 24.0%	100 mg: 40% (p = 0.0365) 200 mg: 56% (p < 0.0001) 400 mg: 64% (p < 0.0001) PBO: 25.5%	100 mg: 17% (NS) 200 mg: 31% (p = 0.0003) 400 mg: 46% (p < 0.0001) PBO: 10%	100 mg: 4% (NS) 200 mg: 11% (p = 0.0022) 400 mg: 21% (p < 0.0001) PBO: 1.0%
	C035; NCT04557085 [43]	446	100 mg: 42.6% (p < 0.001) 200 mg: 78.3% (p < 0.001) 400 mg: 100% (p < 0.001) PBO: 25.9%	100 mg: 40.7% (p = 0.046) 200 mg: 68.1% (p < 0.001) 400 mg: 81.6% (p < 0.001) PBO: 28.2%	100 mg: 26.5% (p = 0.009) 200 mg: 53.1% (p < 0.001) 400 mg: 74.8% (p < 0.001) PBO: 12.8%	100 mg: 12.4% (p = 0.004) 200 mg: 30.1% (p < 0.001) 400 mg: 52.4% (p < 0.001) PBO: 2.6%

All p values are vs. placebo

ASM antiseizure medication, BRV brivaracetam, CNB cenobamate, ESL eslicarbazepine acetate, LCM lacosamide, NA not applicable, NR not reported, NS not significant, PBO placebo, PER perampanel

^aNumber of patients evaluated for efficacy

^bIn this study, the primary outcome was the median change in seizure frequency per 28 days vs. placebo (− 1.22, − 3.33, and − 4.50 for placebo and LCM 200 and 400 mg/day, respectively)

^cIn this study, significance testing was only reported for the primary outcome measure

^dP values were not adjusted for multiplicity

Table 3.

Safety outcomes for RCTs of third-generation ASMs

ASM	Study #; NCT#	N	TEAEs, 5 most common	SAEs	Discontinuations due to TEAEs	Deaths
Lacosamide (LCM)	SP667 [24]	418	LCM overall (n = 321) Dizziness: 35% Headache: 16% Nausea: 14% Fatigue: 14% Ataxia: 13%	LCM 200 mg/day: 9% LCM 400 mg/day: 6% LCM 600 mg/day: 3%	17%	None reported
	SP754; NCT00136019 [25]	405	LCM overall (n = 301) Dizziness: 44.9% Nausea: 13.3% Diplopia: 13.0% Vision blurred: 12.6% Headache: 12.3%	LCM 400 mg/day: 5.9% LCM 600 mg/day: 3.1%	16.5%	0
	SP755; NCT00220415 [26]	485	LCM overall (n = 322) Dizzi ness: 13.0% Headache: 9.6% Diplopia: 9.0% Nausea: 6.8% Vertigo: 6.5%	LCM 200 mg/day: 8.0% LCM 400 mg/day: 9.4%	9.1%	0
	EP0008; NCT01710657 [27]	692	LCM overall (n = 363) Dizziness: 25.9% Nasopharyngitis: 14.3% Somnolence: 10.2% Headache: 9.4% Upper respiratory tract infection: 6.9%	LCM 200 mg/day: 1.1% LCM 400 mg/day: 5.0% Placebo: 1.6%	8.8%	0
Perampanel (PER)	304; NCT00699972 [28]	388	PER 12 mg/day (n = 134) Dizziness: 38.1% Somnolence: 17.2% Headache: 13.4% Fall: 12.7% Irritability: 14.2%	PER 8 mg/day: n = 8 PER 12 mg/day: n = 9	11.1%	1
	305; NCT00699582 [29]	386	PER overall (n = 250) Dizziness: 40.0% Somnolence: 15.2% Fatigue: 14.8% Irritability: 9.2% Nausea: 9.2%	PER 8 mg/day: 7.8% PER 12 mg/day: 9.9%	10.6%	0
	306; NCT00700310 [30]	705	PER 8 mg/day (n = 169) Dizziness: 26.6% Somnolence: 16.0% Headache: 10.7% Fatigue: 5.3% Gait disturbance: 5.3%	PER 2 mg/day: 3.3% PER 4 mg/day: 3.5% PER 8 mg/day: 3.6%	5.1%	0
	335; NCT01618695 [31]	710	PER overall (n = 531) Dizziness: 31.3% Somnolence: 17.1% Nasopharyngitis: 13.2% Upper respiratory tract infection: 6.2% Headache: 6.6%	PER 4 mg/day: 3.4% PER 8 mg/day: 4.0% PER 12 mg/day: 6.7% Placebo: 5.7%	6.2%	2
Eslicarbazepine acetate (ESL)	301; NCT00957684 [32]	402	ESL 1200 mg/day (n = 102) Dizziness: 13.7% Headache: 10.8% Diplopia: 10.8% Somnolence: 9.8% Vertigo: 5.9%	ESL 400 mg/day: 5.0% ESL 800 mg/day: 4.1% ESL 1200 mg/day: 5.9%	9.0%	1
	302; NCT00957047 [33]	252	ESL 1200 mg/day (n = 80) Dizziness: 30.0% Somnolence: 13.8% Headache: 10.0% Nausea: 10.0% Vomiting: 7.5% Coordination abnormal: 5.0%	ESL 800 mg/day: 0% ESL 1200 mg/day: 1.3%	8.7%	0
	303; NCT00957372 [34]	418	ESL 1200 mg/day (n = 98) Dizziness: 43.9% Somnolence: 21.4% Headache: 19.4% Nausea: 15.3% Coordination abnormal: 11.2%	ESL 400 mg/day: 4.2% ESL 800 mg/day: 5.9% ESL 1200 mg/day: 2.0%	15.2%	0
	304; NCT00988429 [35]	653	ESL 1200 mg/day (n = 210) Dizziness: 26.2% Somnolence: 17.0% Nausea: 15.2% Headache: 11.4% Vomiting: 11.0%	ESL 800 mg/day: 6.5% ESL 1200 mg/day: 1.4% Placebo: 3.1%	15.1%	2
Brivaracetam (BRV)	N01114; NCT00175929 [39]	157	BRV 150 mg (n = 52) Dizziness: 9.6% Headache: 7.7% Nasopharyngitis: 7.7% Nausea: 7.7% Fatigue: 5.8%	3.8%	3.2%	0
	N01193; NCT00175825 [40]	208	BRV 50 mg (n = 52) Dizziness: 7.7% Somnolence: 5.8% Fatigue: 5.8% Headache: 1.9% Influenza: 1.9%	0%	2.9%	0
	N01252; NCT00490035 [36]	398	BRV overall (n = 298) Headache: 13.8% Somnolence: 7.4% Dizziness: 5.7% Fatigue: 5.0% Nasopharyngitis: 3.7%	2.3%	4.5%	1
	N01253; NCT00464269 [37]	396	BRV 50 mg/day (n = 101) Somnolence: 16.8% Dizziness: 15.8% Headache: 12.9% Fatigue: 9.9% Insomnia: 6.9%	2.3%	5.1%	2
	N01358; NCT01261325 [38]	764	BRV overall (n = 503) Somnolence: 18.1% Dizziness: 12.3% Fatigue: 9.5% Headache: 7.4% Urinary tract infection: 3.0%	3.2%	6.3%	2
	N01254; NCT00504881 [44]	480	BRV overall (n = 359) Headache: 14.2% Somnolence: 11.1% Dizziness: 8.6% Fatigue: 7.8% Nausea: 5.6%	5.3%	5.8%	1
Cenobamate (CNB)	C013; NCT01397968 [41]	222	CNB (n = 113) Somnolence: 22.1% Dizziness: 22.1% Headache: 12.4% Nausea: 11.5% Fatigue: 10.6%	1.8%	3.6%	0
	C017; NCT01866111 [42]	437	CNB 400 mg/day (n = 111) Somnolence: 37% Dizziness: 33% Fatigue: 24% Diplopia: 15% Headache: 11%	CNB 100 mg/day: 9% CNB 200 mg/day: 4% CNB 400 mg/day: 7%	12.4%	0
	C035; NCT04557085 [43]	446	CNB 400 mg/day (n = 128) Dizziness: 70.3% Somnolence: 46.9% GGT increased: 17.2% Diplopia: 13.3% Vomiting: 12.5%	CNB 100 mg/day: 3.1% CNB 200 mg/day: 3.8% CNB 400 mg/day: 8.6% Placebo: 2.3%	10.1%	0

ASM antiseizure medication, BRV brivaracetam, CNB cenobamate, ESL eslicarbazepine acetate, GGT gamma-glutamyl transferase, LCM lacosamide, PBO placebo, PER perampanel, SAE serious adverse event, TEAE treatment-emergent adverse event

Lacosamide (LCM)

LCM is a functionalized amino acid whose MOA involves sodium channel inactivation. In vitro electrophysiological studies have shown that LCM selectively enhances slow inactivation of voltage-gated sodium channels, resulting in stabilization of hyperexcitable neuronal membranes and inhibition of repetitive neuronal firing [21].

In the USA, LCM is approved for the treatment of focal seizures in patients ≥ 1 month old and as adjunctive therapy in the treatment of primary generalized tonic–clonic (PGTC) seizures in patients ≥ 4 years old [22]. In the EU, LCM is indicated as monotherapy in the treatment of focal seizures in patients ≥ 2 years old with epilepsy; as adjunctive therapy in the treatment of focal seizures in patients ≥ 2 years old, and for the treatment of PGTC seizures in patients ≥ 4 years old with genetic generalized epilepsy [23]. Three double-blind phase 3 trials of LCM as adjunctive therapy in focal seizures were conducted, SP667 [24], SP754 (NCT00136019) [25], and SP755 (NCT00220415) [26], at daily doses of 200 mg, 400 mg, and 600 mg (Table 1) [24–43]. The results of these studies showed reductions in median 28-day seizure frequency from baseline ranging from 21.3% to 37.8% for the maximum LCM doses of each trial (Table 2) [24–43]. Seizure freedom rates up to 8.1% were observed among completers taking LCM 200–600 mg/day [24, 25, 27]. The efficacy of LCM 200 and 400 mg/day in Chinese and Japanese adults with uncontrolled focal seizures was evaluated in Study EP0008 (NCT01710657) [27]. At the 400 mg/day dose, the 28-day median seizure frequency reduction was 39.6% and seizure freedom was achieved in 5.4% of patients who completed the maintenance phase [27]. The most common TEAEs in each trial were dizziness (13.0–44.9%), headache (9.4–16%), and nausea (6.8–14%); see Table 3 for additional TEAEs [24–44].

Table 1.

Summary of randomized clinical efficacy trials of third-generation ASMs

ASM	Approval yeara	Study#; NCT#	Phase	Nb	Study drug dosing	Patient description
Lacosamide (LCM)	USA: 2008 EU: 2008 Japan: 2016	SP667 [24]	3	421	LCM 200 mg/day oral LCM 400 mg/day oral LCM 600 mg/day oral PBO	Patients were 18–65 years old and diagnosed with focal seizures with preserved or impaired consciousness, with or without evolution to bilateral tonic–clonic seizures
		SP754; NCT00136019 [25]	3	405	LCM 400 mg/day oral LCM 600 mg/day oral PBO	Patients were 16–70 years old with focal seizures, with or without evolution to bilateral tonic–clonic seizures
		SP755; NCT00220415 [26]	3	485	LCM 200 mg/day oral LCM 400 mg/day oral PBO	Patients were 16–70 years old with a diagnosis of focal seizures with or without evolution to bilateral tonic–clonic seizures
		EP0008; NCT01710657 [27]	3	692	LCM 200 mg/day oral LCM 400 mg/day oral PBO	Patients were 16–70 years with uncontrolled focal seizures, with or without evolution to bilateral tonic–clonic seizures, and taking 1–3 concomitant ASMs
Perampanel (PER)	USA: 2012 EU: 2012 Japan: 2016	304; NCT00699972 [28]	3	390	PER 8 mg/day oral PER 12 mg/day oral PBO	Eligible patients were ≥ 12 years old, diagnosed with focal seizures with or without evolution to bilateral tonic–clonic seizures, had failed ≥ 2 ASMs, had ≥ 5 focal seizures during baseline, and were taking stable doses of up to 3 approved ASMs
		305; NCT00699582 [29]	3	389	PER 8 mg/day oral PER 12 mg/day oral PBO	Patients ≥ 12 years old with focal seizures with preserved or impaired consciousness, with or without evolution to bilateral tonic–clonic seizures, were eligible
		306; NCT00700310 [30]	3	712	PER 2 mg/day PER 4 mg/day PER 8 mg/day PBO	Patients ≥ 12 years old diagnosed with focal seizures with preserved or impaired consciousness, with or without evolution to bilateral tonic–clonic seizures, were enrolled
		335; NCT01618695 [31]	3	710	PER 4 mg/day PER 8 mg/day PER 12 mg/day PBO	Patients were ≥ 12 years; diagnosed with uncontrolled focal seizures, with or without evolution to bilateral tonic–clonic seizures despite ≥ 2 ASMs in the past 2 years
Eslicarbazepine acetate (ESL)	USA: 2013 EU: 2009 Japan: Not yet approved	301; NCT00957684 [32]	3	402	ESL 400 mg/day oral ESL 800 mg/day oral ESL 1200 mg/day oral PBO	Patients were ≥ 18 years old with focal seizures with preserved or impaired consciousness, with or without evolution to bilateral tonic–clonic seizures for at least 12 months before screening
		302; NCT00957047 [33]	3	252	ESL 800 mg/day oral ESL 1200 mg/day oral PBO	Patients were ≥ 18 years old and diagnosed with focal seizures with preserved or impaired consciousness, with or without evolution to bilateral tonic–clonic seizures for ≥ 12 months prior to screening
		303; NCT00957372 [34]	3	395	ESL 800 mg/day oral ESL 1200 mg/day oral PBO	Patients were ≥ 18 years old with focal seizures with preserved or impaired consciousness for ≥ 12 months prior to screening
		304; NCT00988429 [35]	3	653	ESL 800 mg/day oral ESL 1200 mg/day oral PBO	Patients were aged ≥ 16 years and diagnosed with uncontrolled focal seizures despite treatment with 1–2 ASMs; with ≥ 3 seizures every 4 weeks and no seizure-free interval exceeding 28 consecutive days
Brivaracetam (BRV)	USA: 2016 EU: 2016 Japan: 2024	N01114; NCT00175929 [39]	2b	157	BRV 50 mg/day oral BRV 150 mg/day oral PBO	Patients were 16–65 years old with ≥ 4 focal seizures despite treatment with 1–2 concomitant ASMs during a 4-week prospective baseline
		N01193; NCT00175825 [40]	2b	208	BRV 5 mg/day oral BRV 20 mg/day oral BRV 50 mg/day oral PBO	Patients 16–65 years old with epilepsy experiencing ≥ 4 focal seizures during 4-week baseline despite treatment with 1–2 concomitant ASMs
		N01252; NCT00490035 [36]	3	399	BRV 20 mg/day oral BRV 50 mg/day oral BRV 100 mg/day oral PBO	Patients were 16–70 years old with focal seizures or epileptic syndrome uncontrolled with 1–2 ASMs ≥ 1 month prior to screening
		N01253; NCT00464269 [37]	3	400	BRV 5 mg/day oral BRV 20 mg/day oral BRV 50 mg/day oral PBO	Patients were 16–70 years old with well-characterized focal seizures uncontrolled on 1–2 concomitant ASMs ≥ 1 month prior to screening
		N01358; NCT01261325 [38]	3	768	BRV 100 mg/day oral BRV 200 mg/day oral PBO	Patients were ≥ 16–80 years old, with well characterized focal seizures or epileptic syndrome, uncontrolled with 1–2 concomitant ASMs ≥ 1 month before visit 1
Cenobamate (CNB)	USA: 2019 EU: 2021 Japan: Not yet approved	C013; NCT01397968 [41]	2	222	CNB 200 mg/day oral (target dose) PBO	Patients were adults 18 to 65 years old with a diagnosis of treatment-resistant focal seizures, uncontrolled with 1–3 ASMs within 12 weeks of randomization
		C017; NCT01866111 [42]	2	437	CNB 100 mg/day oral CNB 200 mg/day oral CNB 400 mg/day oral PBO	Patients were adults 18 to 70 years old with a diagnosis of treatment-resistant focal epilepsy, uncontrolled with 1–3 ASMs within 4 weeks of screening
		C035: NCT04557085 [43]	3	446	CNB 100 mg/day oral CNB 200 mg/day oral CNB 400 mg/day oral PBO	Patients were adults 18–70 years old with a diagnosis of focal seizures, uncontrolled with ≥ 8 seizures during an 8-week baseline, despite treatment with 1–3 ASMs. 18-week titration phase was used

Terminology for seizure types is presented using the updated ILAE classifications, which replace the older terms as follows: simple partial focal seizure is described as focal preserved consciousness, complex partial focal seizure is described as focal impaired consciousness, and secondary generalized tonic–clonic seizure is described as focal seizures with or without evolution to bilateral tonic–clonic seizure

ASM antiseizure medication, BRV brivaracetam, CNB cenobamate, ESL eslicarbazepine acetate, ILAE International League Against Epilepsy, LCM lacosamide, PBO placebo, PER perampanel

^aApproval dates from US Food and Drug Administration, Orange Book; European Medicines Agency, Medicines Data Table; and Pharmaceuticals and Medical Devices Agency (Japan), List of Approved Products [12–14]

^bNumber of patients randomized

Perampanel (PER)

PER is a first-in-class noncompetitive antagonist of the ionotropic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptor on postsynaptic neurons [45]. Excitatory postsynaptic AMPA receptors drive glutamate-mediated postsynaptic signaling [46]. PER shows no activity for other glutamate-dependent ion channels, like N-methyl-d-aspartate (NMDA), and without TEAEs typically associated with NMDA receptor antagonists such as ketamine, dextromethorphan, memantine, or amantadine [46, 47].

In the USA, PER is approved for treatment of focal seizures with or without secondary generalized seizures in patients ≥ 4 years old with epilepsy, and as adjunctive therapy in the treatment of PGTC seizures in patients ≥ 12 years old with epilepsy [15]. In the EU, PER is indicated for the adjunctive treatment of focal seizures with or without secondarily generalized seizures in patients ≥ 4 years old, and as adjunctive treatment of PGTC seizures in patients ≥ 7 years old with genetic generalized epilepsy [48]. Three double-blind phase 3 trials were conducted of PER in daily doses ranging from 2 to 12 mg as adjunctive therapy in focal seizures, Study 304 (NCT00699972) [28], Study 305 (NCT00699582) [29], and Study 306 (NCT00700310) [30] (Table 1). In these trials, at the maximum doses, PER reduced median 28-day seizure frequency from 17.6% to 34.5% and was associated with ≥ 50% responder rates from 33.9% to 36.1% (Table 2). The highest seizure freedom rates for PER in the studies ranged from 2.2% to 5.0%.

A further phase 3 trial (Study 335, NCT01618695) evaluated the efficacy, safety, and tolerability of PER 4–12 mg/day in Asian patients with focal seizures [31]. Statistically significant reductions in median 28-day seizure frequency occurred with PER 8 mg/day (29.0%), and 12 mg/day (38.0%; Table 2). The highest ≥ 50% and 100% responder rates (43.3% and 4.4% respectively) occurred with PER 12 mg/day.

Across the clinical trials, the most common TEAEs in patients treated with PER were dizziness 26.6–40.0%) somnolence (15.2–17.2%), and headache (6.6–13.4%); see Table 3. Psychiatric and behavioral adverse reactions, including aggression, hostility, irritability, anger, and homicidal ideation, were also reported across all studies [15, 28–31]. Pooled data from double-blind phase 3 clinical trials in focal seizures showed that psychiatric and behavioral TEAEs were reported in 12.3% and 20.4% of patients receiving 8 and 12 mg/day PER, respectively (compared with 5.7% for placebo) [49]. In the USA, PER is labeled with a black box warning as a result of these adverse events [15]. Both the USA and EU recommend monitoring when beginning treatment with PER, and recommend dose reductions or discontinuation if psychiatric or behavioral symptoms persist [15, 48].

Eslicarbazepine Acetate (ESL)

ESL is a compound of the dibenzazepine class; and like other members of the drug class, carbamazepine and oxcarbazepine, ESL is a sodium channel blocker that shows enhanced binding selectivity for the inactive state [50]. However, ESL enhances slow inactivation, unlike carbamazepine and oxcarbazepine, which enhance fast activation.

In the USA, ESL is indicated for the treatment of focal seizures in patients ≥ 4 years old [51]. In the EU, ESL is indicated as monotherapy for focal seizures in adults with newly diagnosed epilepsy, and as adjunctive therapy in patients ≥ 6 years old with focal seizures [52]. Four randomized trials were conducted, Study 301 (NCT00957684) [32], Study 302 (NCT00957047) [33], Study 303 (NCT00957372) [34], and Study 304 (NCT00988429) [35] (Table 1). ESL at doses of 800 and 1200 mg/day was associated with median percentage seizure frequency reductions per 28 days from baseline of up to 35.6% (Table 2). In each trial, at maximum dose (1200 mg/day), ESL showed ≥ 50% responder rates up to 42.6% (Table 2). Up to 8% of patients receiving 800 mg/day or 1200 mg/day ESL experienced 100% seizure reduction (Table 2). ESL was well tolerated in the randomized trials, and the most common TEAEs occurring in ≥ 5% of patients were dizziness (up to 43.9%), somnolence (up to 21.4%), headache (up to 19.4%) and nausea (up to 15.3%); see Table 3.

Brivaracetam (BRV)

BRV was first discovered as an analogue of the second-generation ASM levetiracetam. BRV’s antiepileptic activity is due to its ability to modulate SV2A activity and inhibit glutamate exocytosis [53]. SV2A activity has been associated with seizure activity in animal models and humans with epilepsy [54, 55] and is a known target for levetiracetam. Compared with levetiracetam, BRV displays > 30-fold higher binding affinity for SV2A [56, 57]. BRV is selective for SV2A, acting only to inhibit vesicle glutamate release, whereas levetiracetam, along with its SV2A selectivity, also acts on the AMPA glutamatergic postsynaptic receptor and at the presynaptic calcium channel [11, 58, 59]. BRV and levetiracetam also bind to SV2A at different binding sites or bind different conformational states of SV2A [60].

The mechanistic differences between BRV and levetiracetam have significant clinical implications. Treatment with BRV has shown fewer behavioral and psychiatric adverse events compared with levetiracetam, although these outcomes are mostly based on real-world data. Several studies have shown a clinically meaningful reduction in the intensity and prevalence of behavioral adverse events compared with levetiracetam [61–64]. Studies have also shown that switching treatment from levetiracetam to BRV may result in improvements in psychiatric and behavioral AEs [61–65]. The study by Zahnert et al. showed that treatment with BRV led to a statistically significant (p < 0.001) decrease in psychiatric or behavioral side effects compared to levetiracetam [65]. However, 28% (26/93) of patients discontinued BRV as a result of behavioral and nonbehavioral adverse events; 10 of those were switched back to levetiracetam and 70% (7/10) showed improvements in adverse events.

In the USA, BRV is indicated as monotherapy or adjunctive therapy for the treatment of focal seizures in patients ≥ 1 month old [66]. In Europe, BRV is indicated as adjunctive treatment of focal seizures in patients ≥ 2 years old with epilepsy [67]. The antiseizure efficacy of BRV was evaluated in daily doses ranging from 5 mg up to 200 mg in two phase 2b trials, N01193 (NCT00175825) [40] and N01114 (NCT00175929) [39], and three phase 3 trials, N01252 (NCT00490035) [36], N01253 (NCT00464269) [37], and N01358 (NCT01261325) [38] (Table 2). In the phase 3 studies, at maximum doses in each trial, BRV showed reductions in median 28-day seizure frequency ranging from 20.5% to 35.6% (Table 2). Across the phase 3 studies, at maximum BRV dose, up to 37.8% of patients experienced a > 50% reduction in seizure frequency (Table 2). Seizure freedom rates with BRV in the phase 3 studies ranged from 0% to 5.2%. A separate phase 3 study further evaluated the safety and tolerability of BRV in adults with uncontrolled focal epilepsy (≤ 20% could be patients with generalized epilepsy) with flexible 20–150 mg/day dosing [44]. BRV was well tolerated in each trial; the most common TEAEs in the phase 3 trials were headache (7.4–14.2%), somnolence (7.4–18.1%), dizziness (5.7–15.8%), and fatigue (5.0–9.9%; see Table 3).

Cenobamate (CNB)

CNB is a tetrazole alkyl carbamate derivative that reduces repetitive neuronal firing via a dual MOA involving preferential inhibition of persistent voltage-gated sodium currents (while sparing transient currents) [19] and positive allosteric modulation (PAM) of GABA_A receptor-mediated tonic currents [20]. Transient sodium currents, mediated by Nav1.1 channels, serve to modulate sodium current (and action potentials) of small inhibitory interneurons, which are crucial for GABAergic inhibition [68, 69]. Nav1.6 channels mediate sodium currents in large excitatory neurons such as the pyramidal output neurons in the cortex and in the hippocampus [70]. Sodium channel blockers such as carbamazepine, and lamotrigine primarily impact the transient voltage-gated sodium current, although they also exert a lesser effect on the persistent current [71]. Phenytoin targets the persistent and transient currents concurrently. In contrast, CNB inhibition likely targets Nav1.6 channels (predominant in excitatory neurons) while sparing Nav1.1 channels, which are essential for the function of fast-spiking inhibitory interneurons [18, 19, 71].

In the USA, CNB is approved for adjunctive and monotherapy treatment of focal seizures in adult patients [72]. In the EU, CNB is indicated for the adjunctive treatment of focal seizures with or without secondary generalization in adult patients with epilepsy who have not been adequately controlled despite treatment with ≥ 2 ASMs [73]. Two randomized, placebo-controlled phase 2 trials were conducted, the fixed-dose C013 trial with a 6-week maintenance treatment phase (CNB, 200 mg/day; NCT01397968) [41], and the multidose C017 trial (CNB 100, 200, or 400 mg/day; NCT01866111) with a 12-week maintenance treatment phase [42] (Table 1). In the fixed-dose study (C013), CNB at 200 mg/day dose reduced median 28-day seizure frequency by 55.6% (Table 2). In the multidose trial (C017), CNB reduced median 28-day seizure frequency by up to 55.0% in the 200 and 400 mg/day dose groups (Table 2). During the maintenance phase of study C017, rates of ≥ 50% seizure reduction were up to 64% for the 400 mg/day dose. Seizure freedom rates during the maintenance phases of the C013 and C017 studies were 28.3% (200 mg) and 21% (400 mg/day), respectively (Table 2).

The US Food and Drug Administration (FDA) felt that the efficacy of CNB was adequately demonstrated in the two phase 2 double-blind trials, and thus no efficacy phase 3 studies were required for regulatory approval. However, the FDA requested that a phase 3, open-label safety trial, C021 (NCT02535091) [74], be designed to assess safety concerns regarding drug reaction with eosinophilia and systemic symptoms (DRESS). In the first 953 participants exposed to a higher starting CNB dose (50–200 mg/day) and larger weekly (rather than biweekly) CNB dose increases, three individuals experienced DRESS (two human volunteers in phase 1 studies, including one death, and one RCT participant during Study C017) [42, 72]. To mitigate the risk of DRESS, the C021 phase 3 safety study used a lower starting CNB dose (12.5 mg/day) and a slower titration schedule (biweekly dose increases of 12.5–50 mg/day) [74]. No cases of DRESS were observed in 1339 patients who received CNB in the C021 trial [74]. This starting dose and titration schedule was approved by the FDA and European Medicines Agency. Since the approval of cenobamate with the lower starting dose and slower titration schedule, over 270,000 patients have received CNB worldwide with no confirmed cases of DRESS [75, 76].

A recently completed 24-week phase 3 study (Study C035, NCT04557085) evaluated the efficacy and safety of adjunctive CNB 100, 200, and 400 mg/day in Asian patients with uncontrolled focal seizures [43]. For CNB 100, 200, and 400 mg/day doses, median seizure frequency reductions during the maintenance phase were 42.6%, 78.3%, and 100%, respectively. Seizure-free rates across CNB 100, 200, and 400 mg/day doses were 12.4%, 30.1%, and 52.4%, respectively (Table 2). The most commonly reported TEAEs across all the RCT trials in CNB dose groups were dizziness (22.1–70.3% of patients), somnolence (22.1–46.9%), fatigue (10.6–24%), headache (11–12.4%), and diplopia (13.3–15%) (Table 3). In the C035 study in Asian populations, dizziness and somnolence occurred with greater frequency in the 200- and 400-mg/day CNB groups compared to previous randomized clinical studies (C013 and C017). This may have been influenced by differences in the study design (longer duration and dose restrictions) and patient population (higher population taking three or more concomitant ASMs and lower BMI among the Asian patient population) for C035 compared to other CNB randomized trials.

Third-Generation ASMs in Focal Seizures: Efficacy in Long-Term Open-Label Studies

Lacosamide (LCM)

In two phase 3 OLE studies, SP774 (NCT00515619) [77] and SP615 (NCT00552305) [78], median treatment duration for LCM ranged from approximately 2.8 to 3.2 years. In both the OLE studies, approximately 50% of patients achieved ≥ 50% reduction in seizure frequency (Table 4) [77–89]. In the SP774 and SP615 OLE studies, the ≥ 6-month seizure-free rates by completer cohorts ranged from 21.2% to 24.4% for 1-year completers and from 29.4% to 30.0% for 3-year completers. In the SP615 OLE study [78], in the 1-year and 3-year completer cohorts, 12.7% and 19.3% of patients were seizure-free for ≥ 12 months, respectively (Table 4).

Table 4.

Long-term efficacy outcomes of third-generation ASMs

ASM	Study#; NCT#	Na	≥ 50% responder rate	≥ 75% responder rate	Seizure freedom rate
Lacosamide	SP774; NCT00515619 [77]	376	Overall: 50.0% Range of 3-month intervals Year 1: 52.3–58.8% Year 3: 55.5–66%	Overall: 23.7% Range of 3-month intervals Year 1: 28.0–33.0% Year 3: 31.0–39.0%	6-month seizure freedom Year 1: 24.4% Year 3: 30.0%
	EP0009; NCT01832038 [79]	471	Overall: 57.1% Year 1: 62.4% Year 2: 67.0% Year 3: 68.0%	Overall: 29.7% Year 1: 28.3% Year 2: 33.3% Year 3: 37.7%	12-month seizure freedom Year 1: 3.5% Year 2: 3.4% Year 3: 2.0%
	SP615; NCT00552305 [78]	369	Overall: 51.2% Year 1: 48.8% Year 3: 57.2%	NR	6-month seizure freedom; 12-month seizure freedom Year 1: 21.2%; 12.7% Year 3: 29.4%; 19.3%
	SP756; NCT00522275 [80]	307	Overall: 48.2% Year 1: 52.8% Year 2: 56.5% Year 3: 58.7%	NR	Year 1: 3.0% Year 2: 3.1% Year 3: 1.8%
Perampanel	Study 307; NCT00735397 [81]	1217	Year 1: 42.8–56.4% Year 2: 53.5–61.5% Year 3: 59.6–65.4%	NR	12-month seizure freedom Year 1: 0.0–1.6% Year 2: 6.4–8.5% Year 3: 11.5–11.7%
	Study 207; NCT00368472 [82]	138	Overall: 37.0% Year 1: 43.8% Year 2: 51.5% Year 3: 49.0%	Overall: 10.9%	Overall: 2.9%
Eslicarbazepine acetate	301 OLE; NCT00957684 (NCT includes original study and OLE) [84]	312	Weeks 1–4: 41.0% Weeks 5–52 (range of 12-week intervals): 48.1–53.2%	NR	Weeks 5–16: 8.7% Weeks 41–52: 12.5%
	302 OLE [83]	325	Weeks 1–4: 36.6% Weeks 5–16: 38.2% Weeks 17–28: 38.5% Weeks 29–40: 39.7% Weeks 41–52: 41.5%	NR	Weeks 5–16: 4.6% Weeks 17–28: 8.0% Weeks 29–40: 8.9% Weeks 41–52: 10.8%
Brivaracetam	Toledo pooled analysis phase 2/3 and OLE studies [88]	1836	Overall: 48.7% Months 10–12: 58.5% Months 22–24: 65.8% Months 34–36: 69.4%	NR	6-month seizure freedom; 12-month seizure freedom; 24-month seizure freedom Overall: 16.9%; 10.4%; 5.5%
	N01199; NCT00150800 [85]	648	Overall: 55.6% Year 1: 58.0% Year 2: 64.8% Year 3: 68.1%	NR	6-month seizure freedom; 12-month seizure freedom Overall: 30.3%; 20.3% (n = 561) Year 1: 11.3%; 4.7% Year 2: 22.8%; 13.8% Year 3: 32.5%; 21.2%
	N01125; NCT00175916 [86]	729	Overall: 43.6% Year 1: 53.2% Year 2: 61.0% Year 3: 60.7%	NR	6-month seizure freedom; 12-month seizure freedom Overall: 22.2%; 15.8% (n = 595) Year 1: 10.6%; 4.2% Year 2: 18.8%; 11.3% Year 3: 22.5%; 14.5%
	N01379; NCT01339559 [87]	675	Overall: 51.7% Year 1: 60.0% Year 2: 63.6% Year 3: 63.4%	NR	6-month seizure freedom; 12-month seizure freedom Overall: 26.0%; 17.9% (n = 642) Year 1: 17.8%; 6.3% Year 2: 27.2%; 16.2% Year 3: 29.5%; 18.2%
Cenobamate	C017 OLE [89]; NCT01866111 (NCT includes original study and OLE)	354	Months > 12–24: 72.0% Months > 24–36: 73.0% Months > 36–48: 76.4%	Months > 12–24: 46.5% Months > 24–36: 53.2% Months > 36–48: 51.8%	12-month seizure freedom Overall: 18.4% Months > 12–24: 13.3% Months > 24–36: 14.3% Months > 36–48: 16.4%

ASM antiseizure medication, NR not reported, OLE open-label extension

^aNumber of patients in the efficacy population

A multicenter, phase 3 OLE (SP756; NCT00522275) of oral LCM as adjunctive therapy was conducted in patients with uncontrolled focal seizures who had completed the pivotal trial SP754 [80]. Of 307 patients assessed for efficacy, 48.2% achieved ≥ 50% seizure reduction (Table 4). Seizure freedom was achieved by 3.0%, 3.1%, and 1.8% of patients in the 1-year, 2-year, and 3-year completer cohorts, respectively.

LCM’s long-term efficacy and safety in Chinese and Japanese patients with epilepsy were evaluated in a multicenter, phase 3 flexible-dose OLE (EP0009; NCT01832038) that enrolled patients who had completed Study EP008 [79]. In 471 patients from the full analysis set, 57.1% achieved ≥ 50% seizure reduction (Table 4). Seizure freedom was achieved by 3.5%, 3.4%, and 2.0% of patients in the 1-year, 2-year, and 3-year completer cohorts, respectively (Table 4).

In a meta-analysis of 5 LCM RCTs and 11 real-world studies among 3191 patients with focal or generalized seizures, the pooled ≥ 50% responder rate after ≥ 1 year was 38% (95% CI 0.35–0.42) for randomized trials and 53% (95% CI 0.44, 0.62) for observational studies [90]. The pooled seizure-free rate was 4% (95% CI 0.06, 0.11) for randomized studies and 13% (95% CI 0.09, 0.18) for observational studies. The seizure-free rate after 6 months (6 studies, n = 1068) and ≥ 1 year (3 studies, n = 411) was 14% (95% CI 0.09–0.19) and 12% (95% CI 0.01–0.24), respectively.

Perampanel (PER)

Patients aged ≥ 12 years with focal seizures, with or without focal-to-bilateral tonic–clonic seizure (FBTC) seizures, who completed one of the three pivotal phase 3 PER studies were eligible to enter the OLE Study 307 (NCT00735397) [81]. Of 1217 patients in the intent-to-treat (ITT) population, the percentages of patients achieving ≥ 50% seizure reduction in the 1-year (n = 894), 2-year (n = 681), and 3-year (n = 436) completer cohorts were up to 56.4%, 61.5%, and 65.4%, respectively (Table 4). In the same completer cohorts, the percentage of patients achieving seizure freedom was up to 1.6%, 8.5%, and 11.7%, respectively. Patients who completed the phase 2 studies Study 206 (NCT00144690) or Study 208 (NCT00416195) were eligible to enroll in the OLE Study 207 (NCT00368472) [82]. In 138 patients enrolled in the OLE, mean PER exposure was approximately 2.2 years. The overall ≥ 50% responder rate was 37.0% (Table 4). By the OLE data cutoff, 10.9% of 138 patients experienced a ≥ 75% reduction in seizures, and 2.9% experienced seizure freedom.

Eslicarbazepine Acetate (ESL)

Patients who completed part 1 of Studies 301 and 302 were eligible to continue into OLE studies that assessed safety and tolerability, seizure control, and quality of life over a 1-year treatment period [83, 84]. Of 239 patients who entered the Study 301 OLE and completed 1 year of treatment, 41.0–53.2% achieved ≥ 50% seizure reductions during year 1 [84] (Table 4). The proportion of patients who were seizure-free over 12-week intervals ranged from 8.7% to 12.5%. Of 223 patients who entered the Study 302 OLE and completed 1 year of treatment, 36.6–41.5% achieved ≥ 50% seizure reductions during year 1 [83]. The proportion of patients who were seizure-free over 12-week intervals ranged from 4.6% to 10.8%.

Brivaracetam (BRV)

A pooled analysis of phase 2b, phase 3, and long-term follow-up studies reported long-term seizure outcomes for adjunctive BRV 50–200 mg/day in patients with focal seizures [88] (Table 4). Of 2051 patients who completed core BRV studies and continued into long-term studies, 54.4% received BRV for 60 months. Among 1836 patients with seizure outcome data, the overall ≥ 50% responder rate was 48.7%. Rates of any continuous 6-month, 12-month, or 24-month duration of seizure freedom were 16.9%, 10.4%, and 5.5%, respectively. An open-label, multinational, follow-up phase 3 trial (N01199 [85]; NCT00150800) was conducted to evaluate long-term safety, tolerability, and efficacy of BRV at doses up to 200 mg/day, in patients with treatment-resistant focal seizures who had completed one of the core trials: N01193 (NCT00175825) [40], N01252 [36], N01253 [37], or N01254 (NCT00504881) [44]. Of 667 patients enrolled in the OLE, 49.2% received BRV for ≥ 4 years. Of 648 patients with focal seizures included in the efficacy analysis, the overall ≥ 50% responder rate was 55.6% (Table 4). Among 561 patients measured for seizure freedom, overall, 6-month and 12-month seizure freedom rates at any time were 30.3% and 20.3%, respectively. A phase 3 OLE study (N01125, NCT00175916 [86]) enrolled patients with focal seizures from a previous BRV trial (N01114 [NCT00175929] [39], N01252, or N01254), or patients with generalized seizures or Unverricht-Lundborg disease. In 729 patients with focal seizures, the overall ≥ 50% responder rate was 43.6% over the course of the study [86] (Table 4). In 595 patients, overall, 6- and 12-month seizure-free rates were 22.2% and 15.8% [86]. An open-label, multicenter, single-arm, phase 3 trial (N01379 [NCT01339559] [87]) was conducted in patients who had completed studies N01258 (NCT01405508) or N01358 [38]. In 675 patients with focal seizures, the overall ≥ 50% responder rate was 51.7% (Table 4); in 642 patients analyzed, 26.0% and 17.9% of patients achieved 6-month and 12-month seizure freedom overall [87].

Cenobamate (CNB)

Patients who were enrolled in the CNB C017 phase 2 trial [42] were eligible to continue in an OLE study [89]. Median treatment duration for the OLE was 53.9 months. In the observed (active) efficacy population, from months > 12–48, 72.0–76.4% of patients achieved ≥ 50% seizure reduction (Table 4). In the modified ITT (mITT) population, overall 18.4% of patients achieved ≥ 12-month seizure freedom [89].

Additional long-term results (not included in the efficacy comparison tables) with CNB were reported in post hoc analyses of the Study C021 phase 3 safety study. Efficacy and safety data collected from US sites conducting Study C021 that enrolled ≥ 11 patients who had recorded high-quality seizure data (n = 240) were included in the post hoc analyses [91]. In the first post hoc analysis, 25.8% (62/240) of patients had 100% seizure reduction for ≥ 12 months at their last visit (mean duration 23.5 months). Any consecutive ≥ 12-month 100% seizure reduction occurred in 36.3% (87/240) of patients. A subsequent post hoc analysis of this n = 240 cohort examined outcomes in patients with < 3 vs. ≥ 3 seizures/28 days [92]. The retention rate at data cutoff was 73.8% (177/240 patients, median time on study of 32.9 months [range 22.1–43.0 months]). Among patients continuing CNB (completer population, 177/240), 33.9% (60/177) achieved 100% seizure reduction for ≥ 12 months at data cutoff, with 44.4% and 23.0% in patients with < 3 and ≥ 3 seizures/28 days at baseline, respectively.

Third-Generation ASMs in Focal Seizures: Long-Term Safety and Tolerability

The long-term safety and tolerability of LCM, PER, ESL, BRV, and CNB have been characterized in several OLE studies, with key findings summarized in Table 5 [77–89, 93, 94].

Table 5.

Long-term safety outcomes of third-generation ASMs

ASM	Study #; NCT#	Na	TEAEs	SAEs	Discontinuations due to TEAEs	Deaths
Lacosamide	SP774; NCT00515619 [77]	376	Any TEAE: 82.7% Dizziness: 24.2% Headache: 14.4% Diplopia: 13.8% Nasopharyngitis: 13.8% Convulsion: 8.5%	23.1%	8.8%	0.8%
	EP0009; NCT01832038 [79]	473	Any TEAE: 86.7% Nasopharyngitis: 32.8% Dizziness: 26.4% Upper respiratory tract infection: 21.1% Headache: 16.1% Somnolence: 8.7%	16.5%	10.4%	1.1%
	SP615; NCT00552305 [78]	370	Any TEAE: 71.4% Dizziness: 39.7% Headache: 20.8% Nausea: 17.3% Diplopia: 17.0% Fatigue: 16.5%	33.8%	12.7%	1.4%
	SP754 [80]	308	Any TEAE: 93.5% Dizziness: 50.0% Headache: 21.8% Contusion: 18.5% Nausea: 18.5% Convulsion: 17.2%	23.1%	10.7%	0.6%
Perampanel	Study 307; NCT00735397 [81]	1218	Range (min–max) of ≥ 1-, ≥ 2-, ≥ 3-, and ≥ 4-year TEAEs Any TEAE: 92.9% Dizziness: 50.0–51.8% Somnolence: 22.1–29.5% Headache: 21.1–25.6% Fatigue: 12.9–16.7% Weight increase: 15.2–26.9%	23.7%	19.0%	0.9%
	Study 207; NCT00368472 [82]	138	Any TEAE: 93.5% Dizziness: 41.3% Headache: 21.0% Somnolence: 19.6% Fatigue: 13.8% Contusion (bruising): 11.6%	15.2%	12.3%	0.7%
Eslicarbazepine acetate	302 OLE [83]; NCT00957047 (NCT includes original study and OLE)	325	Any TEAE: 83.1% Dizziness: 26.5% Headache: 15.7% Somnolence: 12.0% Diastolic BP decreased: 8.6% Coordination abnormal: 8.6%	Not reported	11.4%	3 (0.9%)
	301 OLE [84]; NCT00957684 (NCT includes original study and OLE)	314	Any TEAE: 51.0% Headache: 10.2% Dizziness: 10.2% Diplopia: 5.4% Nasopharyngitis: 5.1% Vertigo: 3.8%	6.1%	3.5%	1 (0.3%)
Brivaracetam	N01199; NCT00150800 [85]	667	Any TEAE: 91.2% Headache: 24.9% Dizziness: 21.4% Nasopharyngitis: 14.1% Somnolence: 13.6% Influenza: 12.6%	22.8%	14.8%	18 (2.7%)
	N01125; NCT00175916 [86]	FOS: 729	Any TEAE: 83.7% Headache: 21.7% Nasopharyngitis: 19.2% Convulsion: 16.0% Dizziness: 14.7% Fatigue: 9.9%	29.1%	10.4%	11 (1.5%)
	N01379; NCT01339559 [87]	766	Any TEAE: 83.9% Headache: 13.6% Dizziness: 13.1% Somnolence: 9.5% Nasopharyngitis: 8.5% Upper respiratory tract infection: 7.7%	18.3%	11.9%	5 (0.7%)
	Toledo et al., 2016 pooled analysis [88]	2186	Any TEAE: 84.5% Headache: 20.9% Dizziness: 17.5% Somnolence: 15.2% Nasopharyngitis: 13.2% Fatigue: 11.3%	18.3%	12.1%	28 (1.3%)
Cenobamate	French et al., 2021 [94]; NCT01397968 (NCT includes original study and OLE)	149	Any TEAE: 89.3% Dizziness: 32.9% Headache: 26.8% Somnolence: 21.5% Viral upper respiratory tract infection: 20.1% Upper respiratory tract infection: 16.1%	25.5%	10.1%	3 (2.0%)
	C017 OLE [89]; NCT01866111 (NCT includes original study and OLE)	355	Any TEAE: 88.2% Dizziness: 34.4% Somnolence: 24.5% Fatigue: 15.8% Headache: 15.2% Diplopia: 14.4%	20.3%	8.7%	6 (1.7%)
	Sander et al., 2022 pooled analysis (C013, C017, C021) [93]	1844	Any TEAE: 87.3% Somnolence: 28.1% Dizziness: 27.6% Fatigue: 16.8% Headache: 14.8% Diplopia: 8.5%	13.6%	10.9% at 2-year cutoff	13 (0.7%) at 2-year cutoff

ASM antiseizure medication, BP blood pressure, FOS focal-onset seizures, SAEs serious adverse events, TEAEs treatment-emergent adverse events

^aNumber of patients in the safety population

Overall Tolerability and Common Treatment-Emergent Adverse Events (TEAEs)

Across long-term OLE studies, a majority of patients experienced at least one TEAE. The reported incidence of any TEAE ranged from 51.0% (ESL) to 93.5% (LCM, PER) [77, 79, 80, 82, 84]. Central nervous system (CNS) adverse events (dizziness, headache, and somnolence) were the most common TEAEs reported across the five third-generation ASMs. Dizziness was reported in 24.2–50.0% of patients on LCM [77–80], 41.3–51.8% on PER [81, 82], 10.2–26.5% on ESL [83, 84],13.1–21.4% on BRV [85, 87, 88], and 27.6–34.4% on CNB [89, 93, 94] (Table 5). Headache was reported in 14.4–21.8% of patients on LCM [77–80], 21.0–25.6% on PER [81, 82], 10.2–15.7% on ESL) [83, 84], 13.6–24.9% on BRV [85–88], and approximately 27% on CNB [94] (Table 5). Somnolence was reported in 19.6–29.5% patients on PER [81, 82], 9.5–15.2% on BRV [85, 87, 88], and 21.5–28.1% on CNB [89, 93, 94] (Table 5). Other common non-CNS TEAEs reported in > 10% in at least one OLE study included nausea and contusion for LCM [78, 80], nasopharyngitis for ESL and BRV [85, 86, 88], and fatigue for CNB [89, 93] (Table 5).

Serious Adverse Events (SAEs) and Discontinuations

Rates of SAEs varied, ranging from 16.5% to 33.8% for LCM [77–80], 15.2–23.7% for PER [81, 82], 6.1% for ESL [84], 18.3–29.1% for BRV [85, 87, 88, 95], and 13.6–25.5% for CNB [89, 93, 94]. The most common SAEs reported were often related to the underlying seizure disorder, such as convulsion and status epilepticus [77, 78, 81].

Discontinuations due to TEAEs ranged from 8.8% to 12.7% for LCM [77–80], 12.3% to 19.0% for PER [81, 82], 3.5% to 11.4% for ESL [83, 84], 10.4% to 14.8% for BRV [85, 87, 88, 95], and 8.7% to 10.9% for CNB [89, 93, 94]. Dizziness was a common reason for discontinuation for LCM, PER, and ESL [77, 79, 81, 83].

Adverse Events of Special Interest

Specific safety considerations have been noted for PER, BRV, and CNB.

PER. In its large OLE study (N = 1216), psychiatric and behavioral TEAEs included irritability (~ 14%), insomnia (~ 8%), and depression (~ 8%), which remained stable over time. Aggression was reported as an SAE in 14 patients (1.2%), and irritability led to discontinuation in 1.2% of patients [81].

BRV. Pooled analyses and OLEs reported psychiatric TEAEs in 25.3–32.3% of patients overall [85, 87, 96]. In individual OLEs, common events included depression (6.0–10.6%), insomnia (6.2–7.3%), anxiety (4.9–6.7%), and irritability (5.2%) [85, 87, 88]. Discontinuations due to psychiatric TEAEs occurred in 3.4–5.1% of patients [85, 87]. Suicidal ideation led to discontinuation in 0.5–1.2% of patients, and two completed suicides were reported across studies (one of which was deemed treatment-related) [85, 88].

CNB. Initial trials of CNB identified a risk of DRESS associated with rapid titration. In long-term OLEs utilizing slower titration schedules (e.g., C017 OLE, N = 355), no cases of DRESS were reported [89]. A large pooled analysis (N = 1844) also reported no DRESS cases, though rash led to discontinuation in 5% of patients [93]. Post hoc analyses of OLE data also showed that exposure-adjusted rates of cognitive and psychiatric TEAEs were low and remained so (< 0.02 and < 0.04 events per patient-year, respectively) through 7 years of CNB exposure [97]. Despite being classified as a class Ib and Id antiarrhythmic at supratherapeutic concentrations, CNB at therapeutic concentrations has shown little effect on the cardiac sodium channel [98].

Mortality

Deaths occurred across the long-term ASM studies, with rates generally ranging from 0.3% to 2.7% [77, 79–81, 84, 85, 87–89, 93]. The vast majority of deaths were not considered treatment-related by investigators.

SUDEP was reported as a cause of death or potential cause in studies of PER, BRV, and CNB [81, 82, 88, 93, 94]. A retrospective post hoc analysis of CNB (N = 2132, 5515 patient-years) identified five cases of definite/probable SUDEP, a rate of 0.88 per 1000 patient-years [99]. For context, a pooled analysis of data from four RCTs of adjunctive ASMs (N = 9144, 13,617.1 patient-years) for treating refractory focal and generalized seizures and Lennox-Gastaut syndrome published in 2001 found the rate of SUDEP per 1000 person-years was 3.8 (95% CI 2.9–5.0) [100]. A separate pooled analysis also demonstrated substantially lower SUDEP rates among patients with Dravet syndrome treated with fenfluramine [101].

Long-Term Treatment Retention

Long-term treatment retention, reflecting sustained efficacy and tolerability, was evaluated in OLEs for all five ASMs, typically using Kaplan–Meier (KM) estimates [77, 78, 81, 83, 84, 88, 89, 93, 94, 96, 102]. One-year retention rates were relatively high across the medications: LCM, 74.5–76.8% [77, 78]; PER, 73.5% (OLE); 65–66% (real-world) [81, 102]; ESL, 68.6–75.5% [83, 84]; BRV, 69.3–79.8% [87, 88, 96]; and CNB, 73–83% [89, 93, 94].

Retention naturally declined over time but remained substantial for many patients, notably during treatment with CNB. At 3 years, estimated retention rates were approximately 51–53% for LCM [77, 78], 46% for PER [81], 48% for BRV [96], and 61–65% for CNB [89, 93, 94]. At 5 years, KM-estimated retention rates were 41% for LCM (Study SP774) [77], 37–54% for BRV [88, 96], and 59–61% for CNB [93]. Pooled analyses for CNB showed a 7-year retention rate of 59% [93], and OLE data from C013 showed a 59% retention rate at 6 years [94].

Meta-analyses Comparing Third-Generation ASMs

Lattanzi et al. performed a network meta-analysis of 16 randomized, placebo-controlled trials that assessed both efficacy and safety of BRV (n = 3 studies), CNB (n = 1), ESL (n = 4), LCM (n = 4), and PER (n = 4) [103]. Of 6753 patients, 4507 were randomized to active treatment, including BRV (n = 803), CNB (n = 221), ESL (n = 990), LCM (n = 1104), and PER (n = 1389). All treatments were significantly associated with a higher ≥ 50% response rate than placebo, although CNB showed the strongest association (odds ratio [OR] 4.40, 95% CI 2.59–7.48). Comparison between third-generation ASMs showed CNB was significantly more likely to provide a higher ≥ 50% response rate than BRV (OR 2.02, 95% CI 1.11–3.66), ESL (OR 1.93, 95% CI 1.07–3.48), LCM (OR 1.86, 95% CI 1.04–3.32), and PER (OR 2.07, 95% CI 1.16–3.70). Similarly, all treatments were significantly associated with a higher rate of seizure freedom than placebo, although CNB again showed the strongest association (OR 14.07, 95% CI 1.88–105.17). According to the surface under the cumulative ranking curve (SUCRA), CNB had the greatest likelihood of providing both ≥ 50% seizure reduction and seizure freedom. All treatments were associated with a higher rate of TEAEs leading to discontinuation than placebo and no statistically significant differences were observed across ASMs. SUCRA showed that BRV and LCM were most likely to be the best tolerated treatments among the third-generation ASMs analyzed.

Mulheron et al. analyzed 23 clinical trials that were included in a network meta-analysis comparing CNB efficacy and safety to BRV, ESL, LCM, PER, and zonisamide [104]. The authors found that CNB was significantly associated with a higher likelihood of ≥ 50% seizure reduction and seizure freedom than the other ASMs. CNB was associated with a higher likelihood of experiencing ≥ 1 TEAE and TEAE leading to discontinuation compared to BRV, LCM, PER, and zonisamide, although no comparison was statistically significant.

Privitera et al. analyzed 21 clinical trials that included CNB, BRV, ESL, LCM, PER, lamotrigine, levetiracetam, and topiramate [105]. At FDA-recommended daily maintenance doses of ASMs, CNB (OR 4.20, 95% CI 2.28–7.74; p = 0.044) was significantly more likely to achieve ≥ 50 seizure reduction than all other assessed ASMs (OR 2.20, 95% CI 1.92–2.53). CNB demonstrated a similar likelihood of discontinuations due to TEAEs as the other ASMs analyzed (p > 0.05).

Chen et al. performed a network meta-analysis using efficacy and safety data from 29 clinical trials that included rufinamide, BRV, CNB, ESL, LCM, retigabine (withdrawn from market), and PER [106]. SUCRA was used to rank ASMs by ≥ 50% responder rate, seizure freedom rate, tolerability, and safety. All ASMs included in the analysis significantly outperformed placebo in ≥ 50% responder rates, as measured by relative risk. CNB was more likely to provide ≥ 50% reductions in seizure frequency compared with the other third-generation ASMs. BRV was most likely to provide seizure freedom, followed by CNB. Seizure freedom rates for all ASMs except rufinamide were significantly better than placebo. BRV was the most tolerated according to SUCRA scores.

Discussion

This narrative review summarized the results of pivotal short-term studies and long-term OLE studies of the five third-generation ASMs approved for the treatment of focal seizures. While all five third-generation ASMs have demonstrated significant efficacy vs. placebo, these data suggest some clinically meaningful differences, particularly regarding the rates of seizure reduction and seizure freedom achieved with CNB.

Efficacy results suggest that adjunctive CNB was associated with markedly higher mITT responder rates and seizure freedom rates in both its short-term randomized trials and long-term OLEs compared to the other four third-generation ASMs, and for that matter, all second-generation ASMs [107, 108]. In phase 2 RCTs, CNB demonstrated responder rates (≥ 50% reduction) of 40.0–64% and seizure freedom rates up to 28.3% [41, 42]. During the phase 3 study conducted in an Asian population, CNB 100–400 mg/day demonstrated ≥ 50% reduction in seizure frequency of 40.7–81.6% and seizure-free rates of 12.4–52.4% during the maintenance phase [43]. These figures are higher than responder rates reported for LCM, PER, ESL, and BRV. These findings are consistent with several recent network meta-analyses [103–105], which despite methodological differences consistently identified CNB as having the highest likelihood of achieving a response using the ≥ 50% seizure reduction criterion. The efficacy profile of CNB, particularly the seizure freedom rates observed in patients with treatment-resistant focal seizures, suggests that seizure freedom may become a more attainable objective for a larger portion of this difficult-to-treat population, and even a potential therapeutic benchmark in ASM development moving forward. For CNB, this goal is supported by the post hoc analyses of mortality rates from CNB clinical studies, where the all-cause mortality standardized mortality ratio (SMR) for patients treated with CNB was 1.32 (95% CI 0.84–2.40), which is not statistically different than the general population [99].

In terms of safety and tolerability, all five ASMs share an approximately comparable profile of dose-dependent, CNS-related TEAEs, primarily dizziness and somnolence, which were also the most frequent reasons for discontinuation across studies. However, network meta-analyses suggest that BRV and LCM may possess the best tolerability profiles [103, 106]. This is consistent with OLE data showing PER was associated with the highest rates of discontinuation due to TEAEs (12.3–19.0%) [81, 82]. Furthermore, specific safety considerations distinguish these agents. PER carries an FDA black box warning for serious psychiatric and behavioral reactions [15], and BRV carries precautions for similar reactions [66]. It is also important to note that initial studies of CNB reported cases of DRESS; however, this risk was successfully mitigated by initiating treatment at a lower dose and utilizing a slower, biweekly titration schedule, resulting in no cases of DRESS in the subsequent large-scale phase 3 safety study (C021) [74]. Of note, CNB has been associated in the postmarketing setting with the potential for clinically significant liver injury [72]. As with other ASMs that carry a risk of hepatotoxicity [15, 22, 51, 66], current labeling advises baseline assessment of liver function, with periodic monitoring in patients receiving ASMs that carry higher hepatotoxic potential or in patients with underlying liver disease, polytherapy, or concomitant hepatotoxic drugs.

Treatment retention is an important composite benefit–risk outcome reflecting the balance of sustained efficacy and long-term tolerability. In this review, CNB demonstrated higher long-term retention rates, with KM estimates of 61–65% at 3 years and 59–61% at 5 years [89, 93, 94]. These rates are notably higher than those for LCM (51–53% at 3 years), PER (46% at 3 years), and BRV (48% at 3 years) [77, 78, 81, 96]. This retention rate for CNB suggests that, for many patients, the higher rates of seizure reduction, or seizure freedom, achieved with CNB treatment outweigh the burden of its CNS-related side effects, which may be improved with appropriate management of concomitant ASMs [109].

Limitations

This review has several limitations. As a narrative review, it relies on the indirect comparison of data from trials that were not designed to be compared head-to-head. The included studies differed in their designs, baseline patient populations, and titration schedules, which makes cross-study comparisons unreliable. This is particularly relevant for CNB, for which the phase 2 clinical trial efficacy data [41, 42] and long-term OLE data [89] utilized a faster titration schedule than the approved titration schedule. The CNB phase 3 safety study, and the CNB phase 3 efficacy and safety study conducted in an Asian population, both utilized the currently approved slower titration schedule [43, 74]. Compared to the phase 2 studies, higher responses to CNB were observed during the phase 3 study in Asian populations. This could be related to differences in baseline characteristics between the Asian populations and those who were not Asian, or to the shorter maintenance phase (6 weeks) used during study C035 vs. studies C013 and C017 (6 and 12 weeks, respectively). Studies of LCM and PER have reported higher treatment responses for Asian patients vs. those who were not Asian, even in the absence of pharmacokinetic/pharmacodynamic differences between ethnic groups [79, 110, 111]. In addition, it is still uncertain whether short-term seizure freedom achieved in pivotal clinical trials accurately predicts long-term seizure freedom in patients with treatment-refractory epilepsy [112]. Furthermore, OLEs are limited by their open-label, uncontrolled design, the risk of selection bias, since enrollment is typically limited to patients who tolerated the ASM during the initial double-blind trial. Because patients who are not responding and/or not tolerating study treatment tend to withdraw from long-term trials, OLE patient populations can be enriched with treatment responders and show increased efficacy over time. Some of the OLE studies only included interval analyses by completer cohorts, in which only patients who completed the treatment interval were evaluated. In other words, this method preferentially counts patients who both tolerated and responded to treatment, giving higher seizure-freedom and responder rates than a true population-level estimate. More conservative efficacy analyses that use the larger ITT or mITT population should also be included. The lack of head-to-head randomized controlled trials remains the primary gap in the literature, and indirect comparisons, such as those discussed here and in formal meta-analyses [103–106], must be interpreted with caution. Additionally, some authors of this review are former employees or consultants of SK Life Science, and SK Life Science also supported medical writing for this article.

Conclusions

Third-generation ASMs gained their regulatory approvals after demonstrating significant reductions in seizure frequency and tolerable safety profiles vs. placebo in pivotal phase 2 and 3 clinical trials. Because the treatment goal for focal seizures is to reduce seizure burden, data summarizing the long-term safety, tolerability, and efficacy of ASMs both in phase 2/3 randomized trials and OLE studies are useful for practitioners and patients alike. Long-term studies that assessed patients with focal seizures receiving a third-generation ASM showed no new safety signals compared with their respective pivotal trials. According to multiple meta-analyses, BRV and LCM are expected to be the most tolerable third-generation ASMs. For each third-generation ASM, long-term efficacy data showed a clinically relevant proportion of patients experienced seizure frequency reductions and seizure freedom. However, currently available data from long-term efficacy studies, confirmed by meta-analyses, suggest that CNB provides notably better seizure freedom rates than the other third-generation ASMs. Seizure freedom may be considered the gold standard therapeutic outcome, not the least because it can potentially change the course of the disease by significantly reducing both morbidity and mortality risks, and by altering the serious neuropsychological, cognitive, functional, social, and quality of life effects associated with continuing to experience seizures. Consequently, it is our opinion that newer ASMs that are characterized by sustained seizure freedom rates, such as CNB, may potentially be considered as a new benchmark in treatment outcomes.

Author Contribution

William E. Rosenfeld: Conceptualization, investigation, visualization, writing, review and editing. Pavel Klein: Writing, review and editing. Louis Ferrari: Writing, review and editing. Vicente Villanueva: Writing, review and editing. William E Rosenfeld, Pavel Klein, Louis Ferrari, and Vincente Villanueva approve the final version of the manuscript and agree to be accountable for the work.

Funding

Preparation of the manuscript and the journal’s Rapid Service Fee were funded by SK Life Science, Inc. This manuscript was prepared according to the International Society for Medical Publication Professionals’ Good Publication Practice (GPP) guidelines for company-sponsored biomedical research: 2022 update (DeTora LM, et al. Ann Intern Med. 2022;175(9):1298–1304).

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Conflict of Interest

William E. Rosenfeld: Consultant/advisor: SK Life Science, Inc.; Speaker: SK Life Science, Inc.; Research support: SK Life Science, Inc., UCB Pharma. Pavel Klein: Consultant/advisor and/or speaker: Abbott, Angelini, Aquestive, Arvelle, Aucta, Biogen, Dr Reddy’s, Eisai, Engage, Greenwich, GRIN Therapeutics, Jazz, Longboard, Lundbeck, Neurelis, Neurona, Paladin, SK Life Science, Inc., Stratus, Sunovion, UCB Pharma, UNEEG Medical, uniQure, Xenon; Research support: Citizens United for Research in Epilepsy (CURE), National Institute of Neurological Disorders and Stroke (NINDS); Scientific Advisory Board for NEUmiRNA, OB Pharma; CEO, PrevEp, LLC. Louis Ferrari: Employee, SK Life Science, at the time the manuscript was written. Vicente Villanueva: Consultant/advisor: Angelini, BIAL, Biocodex, Eisai, Esteve, Exeltis, Jazz, Novartis, Rapport, Takeda, UCB Pharma, Xenon; Speaker: Adium, Angelini, BIAL, Eisai, Esteve, Jazz, Paladin, UCB Pharma; Research support: Angelini, BIAL, Eisai, Jazz, UCB Pharma.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.