Lennox-Gastaut Syndrome (LGS) is a severe developmental epileptic encephalopathy (DEE) defined by drug refractory epilepsy with multiple seizure types including tonic seizures, cognitive impairment, and signature EEG findings.1 Once deemed largely idiopathic, identification of increasing genetic causes of LGS is rapidly growing.2 Despite the rapid elucidation of underlying causes and improved phenotypic classification, successful treatment of LGS has been elusive. The dire need for more effective, well-tolerated therapies calls for further innovation and development in oral drug therapies, which remain the cornerstone of LGS management.

The purpose of this paper is to address three such aspects of treatment evolution for LGS: (1) data supporting the repurposing of existing drugs, (2) recent, ongoing and upcoming clinical drug trials, and (3) future challenges in clinical drug trial.

Adaptation of already prescribable medications for the treatment of epilepsy is often necessary to address critical needs. This is particularly extant in pediatric situations to stave off the risks of continued seizures, Sudden Unexpected Death in EPilepsy (SUDEP) and developmental depression/regression. Rare disease, often the cause of LGS, pushes the need for adaptation of mechanistically appropriate medications, which may include repurposing of available agents. Finally, seizure control with drug resistant epilepsy, a defining component of the LGS trifecta, calls for innovative anti-seizure medication (ASM) application of newer drugs, alongside drugs in clinical trial, dietary, stimulatory and surgical procedures.

Here we review two recent agents vis-a-vis their applicability in LGS.

Perampanel is a first-in-class, selective, broad spectrum, non-competitive AMPA receptor antagonist, passed in the United States by the Food and Drug Administration (FDA) for adjunctive treatment of partial onset seizures (POS) for patients 12 years and older in 2012, for adjunctive treatment for primary generalized tonic-clonic (GTC) seizures in patients 12 years and older in 2015, and for monotherapy for partial onset seizures in patients 12 years and older in 2017. As is generally the case with clinical trial development, the initial registrational trials involved adults with partial onset (POS) drug resistant epilepsy (DRE).3, 4, 5 Subsequent developments include passage for generalized epilepsy in adults.6 Overall, evidence suggested strong efficacy and good tolerability, with treatment emergent adverse events largely related to somnolence and dizziness (mild and transient) in both phase 2 and 3 trials and in observational studies.7

Early data supporting perampanel use in LGS emerged with a prospective pediatric study of 13 children and adolescents, mean age 12.8 years, female n = 6, all with demonstrated LGS.8 At mean follow up of 10.8 months (range 1-24 months), the proportion of patients reporting >50 % decrease in total seizure frequency (>50 % responder rate) was 69.2 %. 31 % discontinued due to lack of efficacy (n = 2) and increased seizures (n = 2). Adverse Events (AEs) were reported in 46.2 % with decreased activity/social interaction (n = 3), agitation (n = 2), and fatigue (n = 2); all AEs reversed with dose reduction. Reduction of dose or discontinuation of concomitant ASM was reported in 53.8 % (n = 7). Improvements in cognitive function and/or behavior were reported for 53.8 % (n = 7).8

A subsequent retrospective, open-label study of adjunctive perampanel in 71 adults with LGS, mean age 40 years, female n = 28.9 The mean treatment duration was 539 days (median 429). The >50 % responder rate for total countable seizures (excluding atypical absence and tonic seizures in sleep) was around 2/3, >75 % was 35.2 %, and >90 % was 16.9 %. AEs appeared at a daily dose of 6 mg in the majority of cases and were reported in 50 %, including irritability in 8.5 % and aggressive behavior in 7 %. However positive behavioral and psychological well-being were reported in 4 patients (5.6 %). In sum, the study supported efficacy and tolerability of perampanel in adults with LGS, with behavioral side effects in some patients limiting dosages ≥ 6 mg/d.9

On the heels of these data, a multicenter, placebo-controlled, randomized clinical drug trial for perampanel in LGS was planned (NCT02834793). However, due to the COVID-19 pandemic, recruitment was stymied and findings ultimately underpowered; recruitment was n = 71, at half the planned enrollment number. Findings included a numerically greater median percent reduction in drop (tonic, atonic, or myoclonic with a fall if unsupported) seizure frequency of 23.1 % with perampanel vs 4.5 % with placebo, using prespecified assessments (p = 0.107). Post hoc analyses looking at median percent reduction of all countable motor seizure showed a difference of 44.0 % in perampanel treated vs −0.6 % in placebo (p = 0.017). Main treatment emergent adverse events (TEAEs) in perampanel treated patients included somnolence (23.5 % vs 5.6 % placebo), irritability (14.7 % vs 2.8 %), upper respiratory infection (11.8 % vs 2.8 %), and decreased appetite (11.8 % vs 0 %). Study drug withdrawal occurred in 3 perampanel treated patients (13 % vs 0 % placebo).10

To date, the largest study for perampanel in LGS was conducted retrospectively, across 19 centers, n = 87, with adolescents and adults (15-34 years), female = 32, followed over 36 months.11 The maximum dose ranged from 4 to 8 mg. Treatment failure, defined as the addition of another ASM or perampanel discontinuation, was 60 % at a median of 12 mo, due to lack of efficacy (52 %), tolerability 27 %, or both (21 %). Seizure relapse, defined as a seizure occurrence in those seizure free for at least one month or a > 50 % increase in monthly countable seizure frequency, was 36 % at a median of 21 months. The >50 % responder rate for all countable seizure types including drop attacks was 41 % at 11 months and for drop attacks was 61 % at 11 month. 1/3 of patients were able to wean or discontinue at least one concomitant ASM. AEs were reported by 39 (44.8 %) with behavioral exacerbation in 19, somnolence in 11, fatigue in 6, dizziness in 6, ataxia in 3 and anorexia in 1. AEs were ameliorated with dosage decrease or discontinuation. Risk factors for AEs included shorter duration of epilepsy, high number of prior ASMs, and faster upward titration.11

Perampanel’s effects on cognition and behavior in the DEE population has been systematically reviewed, with emphasis on studies including neuropsychiatric testing via validated scales (18 studies, 3562 pediatric patients).12 The study found that there was a lack of impairment of general cognitive functions, no impairment of visual spatial skills, a slight improvement in verbal memory, but a decline in attentional power. The evidence suggested an overall favorable risk-benefit profile. Avoidance of high-dose and/or rapid titration and close monitoring recommended in cases of underlying personal or family history of neuropsychiatric disorders which could be risk factors for anger, aggression, and irritability.12

Cenobamate is an akyl-carbamate with two unique mechanisms of action (MOA): (1) blockade of the persistent sodium current at the voltage-gated sodium channel, in contrast to carbamazepine, phenytoin, and lamotrigine (fast inactivation), and lacosamide (slow inactivation), and (2) dual positive allosteric modulation (PAM) of the intra-synaptic (phasic) and extra-synaptic (tonic) GABA-A receptors, at sites independent of benzodiazepine binding. The selective inactivation of the persistent sodium current occurs exclusively at the excitatory principal neurons (Nav 1.2/SCN2A, 1.6/SCN8A) that generate seizure activity, importantly, without concomitant suppression of the inhibitory GABAergic interneurons (Nav1.1/Dravet Syndrome) that suppress seizure activity. PAM at the extrasynaptic GABA-A receptor is important for treatment of prolonged or clustering seizures; synaptic GABA-A receptors, the site of action for benzodiazepines, are endocytosed after 40 min of seizures, after which it is rendered unavailable for targeting.13

The two registrational trials of cenobamate for adults with POS (C013, C017) produced remarkable results of 21 % seizure freedom (vs 1 % placebo) and 93 % decrease in convulsive seizure frequency.14,15 10–36 % of the open-label extension (OLE) patients were seizure free for up to 4 years (median 45–48 months). The United States FDA waived the phase 3 efficacy study as the efficacy and tolerability data indicated that conducting further trials with placebo would be unethical.

Adjunctive cenobamate for LGS16 was initially evaluated in a retrospective study of 4 adults with LGS. At 12 months’ therapy, they found a 25-74 % seizure frequency reduction, with two patients achieving >50 % reduction. In a study of drug load reduction with cenobamate in developmentally delayed adults with epilepsy, of 5 patients with LGS, two achieved >50 % seizure frequency reduction, another 2 achieved >75 % seizure frequency reduction, and one had no response. All 5 LGS patients were able to reduce concomitant ASM drug load with cenobamate therapy.17 A subsequent study evaluated the efficacy, retention and tolerability of adjunctive cenobamate in 41 DEE patients aged 4-73 years, including 33 LGS patients, 57.6 % (19) who experienced >50 % seizure frequency reduction (further subgroup analysis for LGS not available) 18

Most recently, a dedicated study of the efficacy and tolerability of adjunctive cenobamate in 16 LGS children and adults retrospectively evaluated responder rates, seizure freedom, TEAEs, reduction in polypharmacy, and non-seizure outcomes, including available pre and post treatment EEG tracings.19 All patients were characterized by current, multi-platform genetic testing. Of the intention-to-treat population of 16, 3 patients discontinued due to ataxia, somnolence and increased seizures, all reversed with dose reduction. Median maintenance treatment duration was 244 days (mean 294 days, 51-648 days). All patients were treated for >6 months and 81 % (13/16) for >12 months. Responder rates in the 13 patients who reached maintenance dosing was ≥50 % in 92 % (12/13), ≥75 % in 69 % (9/13), ≥90 % in 46 % (6/13) and 100 % in 39 % (5/13). Seizure freedom duration ranged from 107 - 681 days, with 3/16 (19 %) seizure-free for >6 months, including 2 (13 %) for >1 year. TEAEs were reported in 44 % (7/16) with somnolence (n = 5, 31 %), ataxia (n = 2, 12.5 %), aggression/agitation (n = 2, 12.5 %), and one (6 %) each dizziness, headache, blurry vision and insomnia. Positive non-seizure outcomes: Six (38 %) reported improved cognition described as improved attention/alertness. One patient began first ever verbal communication with formation of simple sentences. 6/16 (38 %) caregivers noted improvements in mood. Improved sleep regulation (decreased insomnia) was seen in 4/16 (25 %). Pre and post- treatment EEG tracings (n = 5): no change in 1, partial improvement in 2, and complete remission of interictal epileptiform discharges (IEDs) in 2. Regarding reduction of polypharmacy, 69 % (9) patients were able to wean off ≥1 concomitant ASMs (mean 1.38) and 6 (46 %) ≥2 ASMs including two of the three patients with ≥6 months’ long seizure freedom, who were able to wean off 3 and 4 ASMs, respectively. Since publication, the latter patient has weaned off 5 ASMs and has maintained seizure freedom on cenobamate monotherapy with good tolerance.19 Please refer to Table 1 for details regarding further real-world studies of adjunctive cenobamate in Lennox Gastaut Syndrome.20, 21, 22, 23

Particular characteristics of cenobamate make it especially suited for LGS. Polypharmacy in LGS is common and often problematic; data exist to support weaning (down or off) concomitant ASMs with cenobamate treatment, in populations including DEE and LGS.17 With concurrent ASMs, exploration of efficacy with adjunctive cenobamate has produced strikingly uniform efficacy across various co-mediations, including lacosamide, levetiracetam, lamotrigine, clobazam and zonisamide.24 While drug-drug interactions do exist, preemptive or early down-titration of sodium channel agents such as lacosamide, lamotrigine, or benzodiazepines such as clobazam have been noted as mitigating factors.24

Mortality and SUDEP with drug resistant epilepsy is another area of concern for the LGS population. In the cenobamate trials, overall mortality in DRE adults decreased from 9 to 4/1000 patient years, similar to rates seen in the general population and in people with epilepsy who are seizure free. Similarly, SUDEP rates decreased from 9.3 to 13.4 to 0.88/1000 patient years. The mechanistic explanation for these effects is yet unknown.25

Findings in cognition with cenobamate treatment in DRE are preliminary and exploratory. Improvements in verbal or visual special episodic memory, executive functions, and processing speed have been noted,26 but may be linked to reduced pharmacological burden. Others have found no differences in cognitive, emotional or quality-of-life variables with cenobamate use in DRE.27

Clinical drug trials (CDT) are an important therapeutic option for patients with DRE as seen in DEE and LGS. Many patients with failure of two appropriately chosen and dosed antiseizure medications and ongoing seizures may be eligible for access to trial medications with novel mechanisms and potentially improved efficacy and tolerability.

Many of the CDT for LGS are currently in phase 3, just prior to application for FDA approval. Roughly, these trials follow the traditional model of a central sponsor, often a pharmaceutical company, that enrolls and maintains subjects through trial at a variety of outside epilepsy trial centers.

Participation in an epilepsy CDT begins with an in-person screening visit. The trial requirements and inclusion and exclusion criteria are reviewed to see if the patient is appropriate for trial inclusion. If all criteria are met and patient is interested in the study, informed consent is completed.

Screening is often followed by a 4-week baseline period, during which baseline ASMs are not changed. During this time, the subject/caregiver(s) is recording seizures in a recommended format. After these 4 weeks, randomization occurs, where the patient is assigned to placebo or study drug. Generally, the trials in LGS are randomized and double-blinded. After titration and a maintenance phase in sum lasting around 16 weeks, the option to continue on the study drug in an open label fashion (until FDA approval) is available in most LGS trials. Specific details as to the structure of individual trials can be found at clinicaltrials.gov and on sponsor websites.

This section will provide an overview of recent, ongoing and potential upcoming clinical drug trials for LGS patients.

Soticlestat selectively inhibits cholesterol 24-hydroxylase, a brain specific enzyme responsible for the breakdown of cholesterol to 24S-hydroxycholesterol (24HC), which contributes to neuronal hyperexcitability through its action at the glutamate receptor. A phase 2, randomized, double-blind, placebo-controlled study evaluated the efficacy and safety of soticlestat in pediatric patients with Dravet Syndrome and Lennox-Gastaut Syndrome, aged 2-17 years (NCT03650452). The placebo-adjusted median reduction in GTC and drop seizures in Dravet was statistically significant at 50.00 % (p = 0.002, n = 51), but not statistically significant in LGS 17.08 % (p = 0.1160, n = 88). TEAEs were mostly mild or moderate in severity, with lethargy and constipation reported >5 % more with treatment than placebo.28 There exist rare cases of complete, sustained seizure control, including after withdrawal of soticlestat.29 Ultimately, soticlestat was denied approval by the US FDA in 2024 as the results did not meet statistical significance.

Carisbamate (YKP509) is a mono-carbamate in the same alkyl carbamate family as felbamate, cenobamate and ezogabine. Its precise mechanism of action is not entirely clear but may include reduction of repetitive neuronal firing by inhibition of voltage-gated sodium channels, activation of presynaptic chloride conductance, and blockade of T-type calcium channels.30,31 Animal studies demonstrated broad anticonvulsant activity and early human studies in LGS showed linear and dose-proportional pharmacokinetics with single and multiple dosing. Tolerance was good with TEAEs including dizziness, drowsiness, and headache. Carisbamate is now being studied in a global, multicenter, randomized, placebo-controlled, double-blind phase 3 study in patients with LGS, aged 4-55 years (NCT05219617). The primary objective is to evaluate efficacy in reducing drop seizures. The secondary objectives include the evaluation of total seizure reduction, safety and tolerability, and pharmacokinetics.

Bexicaserin (LP-352) is a superagonist of the highly brain specific 5HT2c serotonergic receptor. It is differentiated from fenfluramine, the first serotonergic antiseizure medication passed for LGS which acts at 5HT1D, 2a and c, 4 and possibly 7,32 in its specificity and high affinity for the 5HT2c receptor specifically, thus limiting off-target side effects, particularly of serotonergic receptors located external to the CNS. Phase 1b/2a study (NCT05364021) and OLE in LGS/DEE revealed median observed countable motor seizure reduction across the DEE population (DS=3, LGS=20, DEE Other=18) at −56.1 %, sustained over approximately 6 months. Overall TEAEs included fatigue, seizure, agitation and change in seizure presentation. Fatigue during titration in one subject, leading to discontinuation, with resolution after withdrawal. There were no serious TEAEs.

Bexicaserin is currently being evaluated in two phase 3 studies for Dravet (NCT06660394) and DEE/LGS (NCT06719141) patients aged 2-65 years. The primary objective is to evaluate efficacy as assessed by countable motor seizures. Secondary objectives include safety, tolerability, and efficacy as assessed by total seizures.

Clemizole (EPX-100) is a histamine H1 receptor antagonist of the benzimidazole group with antihistamine, antipruritic, and sedative properties, discovered in the 1950s. Its anticonvulsant properties as a serotonin receptor agonist were not recognized until 2013, when Baraban et al. described suppression of spontaneous convulsive behavior and electrographic seizures in zebrafish models for Dravet Syndrome.33 EPX-100 is under phase 3 investigation for Dravet Syndrome with primary efficacy endpoint of mean percent change in 28-day countable motor seizure frequency (CMS-28), with interim analyses indicating that the drug appears safe and generally tolerated.34

For LGS, EPX-100 is under evaluation in a multicenter, phase 3, randomized, double-blind, placebo-controlled study designed to evaluate the efficacy and safety of clemizole HCL (EPX-100) as adjunctive therapy in children and adults, aged 2–55 years (NCT05066217). Primary outcome measures will include change in countable major motor seizures, with secondary outcome measures to include 50 % responder rate, change in seizure free days, and safety and tolerance.

Opakalim (BHV-7000) is a potent, highly selective Kv7.2/7.3 potassium channel activator that regulates the hyperexcitable state in epilepsy without off-target sedative, cognitive or mood effects from GABA-A activation. The phase 1 single and multiple ascending dose study used qualitative EEG changes in spectral power as the primary endpoint and showed BHV-7000 to be well-tolerated, without central nervous system (CNS) adverse effects often encountered with other ASMs.35,36 Phase 2/3 studies for adults with DRE and focal epilepsy (NCT06443463) and idiopathic generalized epilepsy (NCT06425159) are currently ongoing. Potential investigation for opakalim use in DEE/LGS is under investigation.

Please refer to Table 2 for summary of clinical drug trials for LGS.

The Advent of Decentralized Clinical Drug Trials: Pros and Cons

During the COVID-19 pandemic, the urgency to continue ongoing clinical trials in a remote fashion led to the emergence of decentralized trials (DCT). DCT utilize remote technologies, with or without local lab and imaging facilities and health care providers to recruit, enroll, collect data, and monitor subjects in trial. DCT have since burgeoned, and we are now witnessing the nascence DCT in epilepsy for the DEEs (e.g. relutrigine for SCN2A and SCN8A).

In theory, DCT offer several potential advantages to studies conducted via traditional clinical trial sites. Most apparent are lower costs, greater access and convenience, and greater geographic coverage which may enhance race and income-related diversity of participants thus bolstering generalizability of results.

However, DCT design and implementation present unique challenges related to their remote nature, many of which are related to ethics and safety.37 Studies for DEE/LGS are long, sometimes complicated, and depend on the reliability of the participant/caregiver to collect/record data and to understand when and how to report possible side effects. The traditional on-site screening visit with trained staff serves as a bell-weather on several counts: Does this trial represent the best fit for the patient? Would the subject be too ill to weather the baseline period with medications unchanged and/or the randomization period with potential placebo? Does the study participant have a reliable caregiver willing to perform the study tasks for the duration of the trial? Informed consent, the critical process by which a patient comes to understand risks, benefits and details of trial participation, is arguably more effectively and safely done face-to-face, in a setting where patient has access to trained staff, ample opportunity to ask questions or raise concerns, and where staff can verify understanding. Once enrolled, monitoring for patient safety, tolerability and efficacy, particularly in a complex population such as LGS, may be better with an in-person epilepsy physician who has a direct and longitudinal relationship with the subject, rather than a remote physician for whom the subject may be an unknown entity. Such areas of concern should be addressed to secure success in remote LGS trial implementation.

With the richness of recent trial development for LGS combined with the nascence of clinical trials for specific genetic epilepsies encompassing LGS in their presentation comes a new dilemma of competing studies. A significant portion of the current generation of potential clinical trial participants with LGS have a genetic diagnosis. In some cases, this will be a single gene mutation with a known mechanism of disease. Such patients will be increasingly faced with the option of participating in one of several generic LGS/DEE trials or awaiting trial designed for their specific genetic disease. Disease-specific therapeutic trials may offer treatments geared to rectify the underlying process, such as a channelopathy or haploinsufficiency. Some of these therapies may aim at disease modification with a view to decrease polypharmacy over time. In contrast, generic LGS/DEE trials aim at mechanisms that would putatively decrease seizures stemming from various causes, genetic or otherwise, and generally do not encompass goals of disease modification, although rare cases may occur. Although there is no single correct answer in this situation, consideration of underlying mechanisms and pathways through discussion with knowledgeable specialists including epileptologists, clinical triallists and geneticists, may provide a more sophisticated approach to such dilemmas.