Refining management strategies for Lennox–Gastaut syndrome: Updated algorithms and practical approaches

Abstract

Lennox–Gastaut syndrome (LGS) is a severe developmental and epileptic encephalopathy (DEE) characterized by multiple types of drug‐resistant seizures (which must include tonic seizures) with classical onset before 8 years (although some cases with later onset have also been described), abnormal electroencephalographic features, and cognitive and behavioral impairments. Management and treatment of LGS are challenging, due to associated comorbidities and the treatment resistance of seizures. A panel of five epileptologists reconvened to provide updated guidance and treatment algorithms for LGS, incorporating recent advancements in antiseizure medications (ASMs) and understanding of DEEs. The resulting consensus document is based on current evidence from clinical trials and clinical practice and the panel's expert opinion, focusing on new ASMs with novel mechanisms of action, such as highly purified cannabidiol and fenfluramine. For a patient presenting with newly diagnosed LGS or suspected LGS, the recommended first‐line treatment continues to be valproate. If this is ineffective as monotherapy, adjunctive therapy with, firstly, lamotrigine and secondly, rufinamide, is recommended. If seizure control remains suboptimal, subsequent adjunctive ASM treatment options include (alphabetically) cannabidiol, clobazam, felbamate, fenfluramine, and topiramate, although evidence for these is more limited. Whenever possible, no more than two ASMs should be used together. Nonpharmacological treatment approaches should be used in conjunction with ASM therapy and include ketogenic diet therapies, vagus nerve stimulation, and corpus callosotomy. Patients with LGS that has evolved from another type of epilepsy who are not already being treated with valproate should be transitioned to valproate and then managed using the same algorithm as for newly diagnosed LGS. Older patients with established LGS should be reviewed at least annually by a suitably experienced neurologist. The revised guidance aims to improve seizure control and quality of life for patients with LGS through personalized, evidence‐based treatment strategies while addressing the challenges of accurate diagnosis and management in a rapidly evolving therapeutic landscape.

Plain Language Summary

Lennox–Gastaut syndrome (LGS) is a severe type of epilepsy that usually starts in childhood but continues into adulthood. It is characterized by a variety of different types of seizures (abnormal electrical activity in the brain), which are difficult to treat and often cause people with the condition to fall and injure themselves. Most people with LGS have learning difficulties and need a lot of support, often in residential care. The authors are experts in treating people with LGS and this article provides up‐to‐date guidance and advice on how best to care for those with the condition.

1. INTRODUCTION

In 2017, the authors published their guidance on the management of Lennox–Gastaut syndrome (LGS), including details of recommended treatment algorithms and other practical considerations. ¹ Since then, there have been important developments in the field, including the publication by the International League Against Epilepsy (ILAE) of updated classification and definitions of developmental and epileptic encephalopathies (DEEs) and other epilepsy syndromes ² , ³ and the approval of additional antiseizure medications (ASMs) with novel mechanisms of action (MoAs) for the treatment of seizures associated with LGS. ⁴ , ⁵ , ⁶ , ⁷ As a result of these developments, the authors reconvened to discuss and update their guidance, which is outlined in the current article.

The ILAE uses the term DEE to categorize conditions where the underlying pathology has developmental consequences that are partly independent of those related to the epileptic activity itself. ² , ⁸ LGS is a severe DEE that typically starts between the ages of 18 months and 8 years (most commonly 3–5 years), ³ but it persists through adolescence into adulthood and may also, rarely, have late onset after the age of 18 years. ³ , ⁹ Although LGS often occurs in patients without a history of other epileptic syndromes, it may also evolve from other severe infantile seizure disorders, such as infantile epileptic spasms syndrome (IESS), and early infantile DEE. ³ , ¹⁰ Overall, 3.6% of children with epilepsy, and 19% of those whose seizures start in infancy, evolve to LGS, ¹¹ and LGS represents approximately 1%–2% of all epilepsy cases. ³ LGS has a heterogenous etiology and has no biological marker. ¹ It may occur due to a brain abnormality (e.g., developmental malformation, tuberous sclerosis, brain insult, infection, and tumor), ¹² a genetic pathogenic variant (e.g., DNM1, ALG13, CDKL5, GABRB1, GABRB3, SCN1A, SCN2A, and STXBP1), ¹ , ¹³ , ¹⁴ , ¹⁵ or a metabolic or mitochondrial disease (e.g., Leigh syndrome and creatine metabolic disorders). ¹⁶ , ¹⁷ However, etiology is unknown in many cases. ¹⁸ , ¹⁹ LGS is associated with a great burden of illness and substantial healthcare costs, due to the need for both home‐based and inpatient care. ²⁰ , ²¹ Diagnosis is sometimes difficult because the characteristic seizure types and electroencephalogram (EEG) features are not individually pathognomonic when observed by themselves and evolve and change over time. ¹ However, early identification of LGS is crucial as it allows for timely intervention. This can potentially improve the clinical trajectory by stabilizing seizure activity, enhancing developmental trajectories, and ultimately improving the overall quality of life (QoL) for patients. ¹⁰ , ²²

2. DEFINITION OF LGS

In 2022, the ILAE Task Force on Nosology and Definitions published updated classifications and definitions of epilepsy syndromes with onset in childhood, including LGS. ³ The ILAE characterizes LGS by the presence of (i) multiple types of drug‐resistant seizures with onset before 18 years (one of which must include tonic seizures), (ii) diffuse, slow spike‐and‐wave (SSW), and generalized paroxysmal fast activity on EEG, and (iii) cognitive and often behavioral impairments, which may not be present at seizure onset. ³ The full complement of these clinical and EEG features is usually not present at onset, taking time to appear; therefore, young children presenting with characteristic seizure types but lacking all the aforementioned features should be closely monitored by a specialized medical team for evolution to LGS. ³ Importantly, repetitive assessment for LGS criteria can facilitate access to ASMs approved for LGS. ³ The presence of multiple seizure types (to include tonic seizures and at least one other seizure type) and the EEG features outlined above are mandatory for a definitive diagnosis of LGS. ³

Tonic seizures occur most often during sleep, typically detected using sleep EEG recording. ³ If tonic seizures occur while the individual is standing, they may lead to falls (drop seizures), often resulting in injury. ³ Additional seizure types associated with LGS include:

• Atypical absence seizures: challenging to identify in individuals with cognitive impairment without a video‐EEG recording. ³

• Atonic seizures: common and frequent, especially in younger patients with LGS. ³ Can cause drop seizures and result in injury. ³

• Generalized tonic–clonic seizures: may precede the core seizure types of LGS but usually occur later in the disease course. ¹⁰

• Epileptic spasms: in approximately 20% of cases, LGS evolves from IESS ²³ and epileptic spasms may persist during the progression to LGS. ²⁴

• Non‐convulsive status epilepticus (NCSE): occurs in 50%–75% of cases, typically consisting of sub‐continuous atypical absences, interspersed by recurring brief tonic seizures. ¹⁹

• Focal seizures (with or without bilateral involvement), and unilateral clonic seizures: may precede the core seizure types of LGS but usually occur later in the disease course. ¹⁰

• Myoclonic seizures: occur in approximately 10%–30% of patients and can lead to falls. ¹⁸ , ²⁵ , ²⁶ , ²⁷ , ²⁸ However, they are not a defining feature of LGS since they occur in many generalized epilepsies. ²² If myoclonic‐atonic seizures are present, the diagnosis of epilepsy with myoclonic atonic seizures (EMAtS) should be considered. ³

The interictal SSW pattern is characterized by spikes (<70 ms) and sharp waves (70–200 ms), followed by negative high‐voltage slow waves (350–400 ms), which occur at a frequency of ≤2.5 Hz and are bilaterally synchronous. ³ Generalized paroxysmal fast activity comprises short bursts of diffuse or bilateral fast activity (≥10 Hz), often occurring during sleep and usually lasting a few seconds or less. ³ The seizure and EEG characteristics outlined above are central features of LGS; however, additional characteristics may be seen. Some or all of these may be present before or at diagnosis, or they may evolve and change over time. In almost all cases, LGS persists into adulthood and seizures remain drug‐resistant. ³ Tonic seizures and atypical absence seizures continue to be frequent during adulthood, whereas atonic seizures often decrease in frequency. ³ The SSW pattern decreases in frequency over time and is sometimes absent after the age of 16 years. ³ , ²⁹ , ³⁰

In most cases, developmental impairment is present before seizure onset but the onset of frequent seizures may also result in further developmental stagnation or decline. ³ Over its disease course, LGS culminates in mild to profound intellectual disability in all patients, resulting from developmental slowing, plateauing, or regression over time. ³ Behavioral impairments are common in childhood, adolescence, and adulthood and include hyperactivity with or without attention deficit, autism spectrum disorder, aggression, and sleep disturbances. ³ , ⁹

3. DIAGNOSTIC CHALLENGES AND CONSIDERATIONS

LGS must be identified and diagnosed as early as possible to ensure that appropriate treatment is received. Searching for the underlying disease may be essential in choosing the most appropriate treatment strategy. Also, the choice of the most appropriate ASMs may avoid the worsening of seizure presentation. In some cases, targeted personalized treatment may be possible (e.g., the presence of SCN2A pathogenic variant ³¹ ). Given the challenges associated with accurate LGS diagnosis, several key investigations are recommended, including sleep video‐EEG to detect tonic seizures from sleep and paroxysmal fast rhythms ¹⁰ , ²⁹ ; magnetic resonance imaging (MRI) with an epilepsy protocol to identify potential structural abnormalities ³² ; genetic testing with a gene panel, or preferably whole exome sequencing but also including chromosomal microarray analysis, to detect chromosomal abnormalities and copy number variants ¹³ ; and metabolic testing, if an underlying etiology is not found with imaging or genetic investigation, since LGS can result from a neurometabolic disorder (e.g., Leigh syndrome). ³ , ¹⁶

None of the seizure types associated with LGS is pathognomonic, so a patient's seizure pattern must be assessed alongside their EEG features if an LGS diagnosis is suspected. ¹ The presence of tonic seizures is mandatory for diagnosis and highly suggestive of LGS if seen in the context of a specific EEG pattern. It is also important to bear in mind that the characteristic EEG features of LGS may take time to develop and may not be present initially in patients who transition to LGS from another epilepsy syndrome, such as IESS. ¹ However, it is important to try to identify these EEG features at an early stage, since they may also be transient. ³⁰ Indeed, neither the EEG features nor the seizure types associated with LGS remain static, but develop and change during the disease course, necessitating regular reevaluation and reassessment that might be done at least once in an expert center to ensure appropriate treatment as the syndrome evolves. ¹

Although the use of the aforementioned diagnostic investigations and careful consideration of a patient's clinical and EEG features should allow LGS to be identified, it may be misdiagnosed as other epilepsy syndromes, including, among others, EMAtS (formerly named Doose syndrome), Epileptic Encephalopathy with Spike‐and‐Wave Activation in Sleep (including Atypical Benign Partial Epilepsy of Childhood ³³ [previously called pseudo‐Lennox syndrome ³⁴ ]), and DEE with Spike Wave Activation in Sleep. ¹ , ³ , ²² As with early diagnosis, differential diagnosis is crucial to the successful management of LGS. The diagnostic challenges and considerations associated with LGS have been extensively reviewed elsewhere. ¹ , ³ , ²² , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹

4. REVIEW OF CURRENT EVIDENCE FOR TREATMENTS FOR LGS

4.1. Pharmacological treatment of LGS

Although the evidence base for pharmacological therapy in LGS is limited, ASM treatment options have expanded over recent years to include new agents with novel MoAs (highly purified cannabidiol [CBD] and fenfluramine [FFA]). ⁴⁰ Despite this, currently available epidemiological evidence indicates a very low likelihood of seizure freedom over the long term. ²⁹ , ⁴¹ , ⁴² , ⁴³ A Cochrane review of ASMs for LGS, published in 2021, included 11 studies of seven drugs utilized as adjunctive therapy: CBD, cinromide, clobazam (CLB), felbamate (FLB), lamotrigine (LTG), rufinamide (RUF), and topiramate (TPM). ⁴⁴ The authors concluded that there was high‐certainty evidence for overall seizure reduction with adjunctive LTG and RUF, but evidence for other adjunctive ASMs was low to very low. ⁴⁴ Since then, results of a randomized controlled trial (RCT) of adjunctive FFA in LGS have been published. ⁴⁵

4.2. ASMs licensed or widely used for LGS in Europe and the United States

A summary of ASMs licensed or widely used for LGS in Europe and/or the United States is presented in Table 1. ⁴ , ⁶ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ Since the evidence base for sodium valproate (VPA), LTG, RUF, TPM, CLB, and FLB has changed little since previously reviewed by the authors, ¹ the remainder of this section will focus on the ASMs that have been approved more recently for treatment of seizures associated with LGS; namely, CBD and FFA.

TABLE 1.

Summary of ASMs licensed or widely used for LGS in Europe and/or the United States.

ASM	MoA	Phase 3 efficacy in drop seizures	Phase 3 efficacy in other seizures associated with LGS	Effect on cognition a	Behavioral AEs a	Other considerations
VPA	Uncertain. Most likely MoA is potentiation of the inhibitory action of GABA 46	No Phase 3 study in LGS	No Phase 3 study in LGS	Common b AEs: confusional state, hallucinations, disturbance in attention, stupor, memory impairment 46 Rare c AEs: cognitive disorder, learning disorder 46 An increase in alertness may occur 46	Common b AEs: aggression, agitation 46 Rare c AE: abnormal behavior 46 Aggression, hyperactivity, and behavioral deterioration have occasionally been reported 46	Should not be used as first‐line treatment in female adolescents, in women of childbearing potential, and pregnant women unless alternative treatments are ineffective or not tolerated because of high teratogenic potential and risk of developmental disorders in infants exposed in utero 46
LTG	Sodium channel blocker 47 ; inhibits glutamate release and modulates voltage‐gated calcium channels 64	Phase 3 PBO‐controlled RCT in LGS (16‐week maintenance period) 48 : Significantly greater reduction in drop seizures d for LTG vs. PBO (−34% vs. −9%; p = 0.01) Significantly higher drop seizure RR e for LTG vs. PBO (37% vs. 22%; p = 0.04) Freedom from drop seizures not reported	Phase 3 PBO‐controlled RCT in LGS (16‐week maintenance period) 48 : Significantly greater reduction in all seizures for LTG vs. PBO (−32% vs. ‐9%; p = 0.002) Significantly greater reduction in tonic–clonic seizures for LTG vs. PBO (−36% vs. +10%; p = 0.03) Significantly higher all seizure RR e for LTG vs. PBO (33% vs. 16%; p = 0.01) Significantly higher tonic‐atonic seizure RR e for LTG vs. PBO (43% vs. 20%; p = 0.007) Seizure freedom not reported	Very rare f AEs: confusion, hallucinations 47	Common b AEs: aggression, irritability, agitation 47	VPA inhibits LTG glucuronidation, reducing its metabolism and increasing its half‐life nearly two‐fold; therefore, lower LTG doses are required when used concomitantly with VPA 47 Dose increase required following withdrawal of concomitant VPA therapy
RUF	Modulates activity of sodium channels, prolonging their inactive state 49	Phase 3 PBO‐controlled RCT in LGS (12‐week treatment period) 50 : Significantly greater reduction in drop seizures g for RUF vs. PBO (−42.5% vs. +1.4%; p < 0.0001) Significantly higher drop seizure RR e for RUF vs. PBO (42.5% vs. 16.7%; p = 0.002) Freedom from drop seizures g occurred in 4.1% RUF patients vs. 3.3% PBO patients (p = NS)	Phase 3 PBO‐controlled RCT in LGS (12‐week treatment period) 50 : Significantly greater reduction in all seizures for RUF vs. PBO (−32.7% vs. −11.7%; p = 0.0015) Significantly higher all seizure RR e for RUF vs. PBO (31.1% vs. 10.9%; p = 0.0045) Significantly greater reduction in absence and atypical absence seizures for RUF vs. PBO (−50.6% vs. −29.8%; p = 0.0222) Significantly greater reduction in atonic seizures for RUF vs. PBO (−44.8% vs. −21.0%; p = 0.0125) No patients were seizure‐free during the study Significantly higher percentage of RUF vs. PBO patients experienced an improvement in seizure severity on parental/guardian GE scale (53.4% vs. 30.6%; p = 0.0041)		Common b AE: anxiety 49	Common b AEs: weight decrease, anorexia, eating disorder, decreased appetite 49
TPM	Potentiates activity of GABA; may modulate GABAA receptors; inhibits some carbonic anhydrase isoenzymes 51	Phase 3 PBO‐controlled RCT in LGS (11‐week double‐blind treatment period) 52 : Significantly greater reduction in drop seizures g for TPM vs. PBO (−14.8% vs. +5.1%; p = 0.041) Higher (but not significantly) drop seizure RR e for TPM vs. PBO (28% vs. 14%; p = 0.071) During the double‐blind phase, 2.2% TPM vs. 0% PBO patients were free from drop seizures g During the maintenance period (last 8 weeks of double‐blind phase), 10.9% TPM vs. 0% PBO patients were free from drop seizures g	Phase 3 PBO‐controlled RCT in LGS (11‐week double‐blind treatment period) 52 : Significantly greater reduction in major seizures g for TPM vs. PBO (−25.8% vs. +5.2%; p = 0.015) Significantly higher major seizure h RR e for TPM vs. PBO (33% vs. 8%; p = 0.002) During the double‐blind phase, 2.2% TPM vs. 0% PBO patients were free from major seizures h Significantly more TPM vs. PBO patients experienced improvement in seizure severity on the parental GE scale (p = 0.037)	Common b AEs: bradyphrenia, expressive language disorder, confusional state, disorientation, disturbance in attention, memory impairment, amnesia, cognitive disorder, mental impairment, psychomotor skills impaired 51 Many uncommon i cognitive AEs 51	Very common j AE: depression 51 Common b AEs: aggression, mood altered, agitation, mood swings, depressed mood, anger, abnormal behavior 51 Many uncommon i behavioral AEs 51	Rare AE: Stevens–Johnson syndrome 51
CLB	Not fully understood but thought to involve potentiation of GABAergic neurotransmission resulting from binding at the benzodiazepine site of the GABAA receptor 53	Phase 3 PBO‐controlled RCT in LGS (12‐week maintenance period) 54 : Significantly greater reduction in drop seizures for CLB vs. PBO: average weekly rates decreased 12.1% for PBO vs. 41.2% (p = 0.0120), 49.4% (p = 0.0015) and 68.3% (p < 0.0001) for CLB 0.25, 0.5, and 1 mg/kg/day, respectively Significantly higher drop seizure RR e for CLB vs. PBO: 31.6% for PBO vs. 43.4% (p = NS), 58.6% (p = 0.0159), and 77.6% (p < 0.0001) for CLB 0.25, 0.5 and 1.0 mg/kg/day, respectively Rate of freedom from drop seizures was 3.5% for PBO vs. 7.5%, 12.1% and 24.5% for CLB 0.25, 0.5, and 1 mg/kg/day, respectively (statistical comparison not valid due to low patient numbers) Significantly lower frequency of seizure‐related injuries for CLB vs. PBO: 27.1% for PBO vs. 12.1% (p = NS), 4.8% (p < 0.001), 10.2% (p < 0.03) and 8.9% (≤0.05) for CLB 0.25, 0.5, 1 mg/kg/day and all CLB doses combined, respectively 55	Phase 3 PBO‐controlled RCT in LGS (12‐week maintenance period) 54 : Significantly greater reduction in total seizures for CLB vs. PBO: average weekly rates decreased 9.3% for PBO vs. 34.8% (p = 0.0414), 45.3% (p = 0.0044) and 65.3% (p < 0.0001) for CLB 0.25, 0.5, and 1 mg/kg/day, respectively	Common b AEs: disturbance in attention, slow speech/dysarthria /speech disorder 56 Uncommon k AEs: amnesia (may be associated with abnormal behavior), memory impairment, anterograde amnesia (especially at higher dose levels), confusional state 56	Common b AEs: irritability, aggression, restlessness, depression 56 Uncommon k AEs: abnormal behavior, anxiety, delusion, nightmare 56 Amnesia effects may be associated with inappropriate behavior 56	Tolerance and physical and/or psychic dependence may develop, especially during prolonged use 56 : in a long‐term, single‐center LGS study, tolerance developed in 48% of initial responders 57 Discontinuation may result in withdrawal or rebound phenomena 56 May change VPA plasma levels when used concomitantly; plasma levels of VPA should therefore be monitored 56 May decrease LTG plasma levels (expert opinion) Concomitant use of CLB and CBD increases the incidence of somnolence and sedation 56 , 58
FLB	Unclear. 59 Acts as non‐competitive GluN2B antagonist; inhibits Na+/Ca2+ channels with potentiation of GABA‐induced chloride currents, leading to inhibition of NMDA receptors 60	PBO‐controlled RCT in LGS (70‐day treatment period; 56‐day maintenance period) 61 : During the treatment period, significantly greater reduction in atonic seizures for FLB vs. PBO (−34% vs. −9%; p = 0.01); 3/28 FLB‐treated patients vs. 0/22 PBO‐treated patients had no atonic seizures During the maintenance period, significantly greater reduction in atonic seizures for FLB vs. PBO (−44% vs. −7%; p = 0.002); 5/28 FLB‐treated patients vs. 0/22 PBO‐treated patients had no atonic seizures	PBO‐controlled RCT in LGS (70‐day treatment period; 56‐day maintenance period) 61 : During the treatment period, significantly greater reduction in parental counts of all seizures for FLB vs. PBO (−19% vs. +4%; p = 0.002); no patients were seizure‐free During the maintenance period, significantly greater reduction in parental counts of all seizures for FLB vs. PBO (−26% vs. +5%; p < 0.001); 4/37 FLB‐treated patients vs. 1/36 PBO‐treated patients had no seizures During treatment period, reduction in generalized tonic–clonic seizures was −28% for FLB vs. +11% for PBO (p = NS); 2/16 FLB‐treated patients vs. 1/13 PBO‐treated patients had no generalized tonic–clonic seizures During the maintenance period, significantly greater reduction in generalized tonic–clonic seizures for FLB vs. PBO (−40% vs. +12%; p = 0.017); 7/16 FLB‐treated patients vs. 1/13 PBO‐treated patients had no generalized tonic–clonic seizures During the maintenance period, significantly higher parent/guardian GE scores for FLB vs. PBO from day 49 onward (p < 0.001)	In controlled pediatric LGS studies, 6.5% of patients reported abnormal thinking 59	Frequent AEs: agitation, psychological disturbance, aggressive reaction 59 Infrequent AEs: hallucination, euphoria, suicide attempt 59 In controlled pediatric LGS studies, 16.1% of patients reported nervousness and 6.5% of patients reported emotional lability 59	Black box warning over risk of aplastic anemia and acute liver failure 59 Only available for limited use in Europe on a patient‐by‐patient basis, due to the risk of aplastic anemia and liver failure
CBD	Unclear but does not exert its anticonvulsant effect through interaction with cannabinoid receptors. 4 Reduces neuronal hyper‐excitability through modulation of intracellular calcium via GPR55 and TRPV‐1 channels; modulates adenosine‐mediated signaling through inhibition of adenosine cellular uptake via ENT‐14	Two Phase 3 PBO‐controlled RCTs in LGS (GWPCARE3 and GWPCARE4; 14‐week treatment periods) 62 , 63 : GWPCARE3 62 : Significantly greater reduction in drop seizure l frequency for CBD vs. PBO: median frequency decreased 17.2% for PBO vs. 37.2% (p = 0.002) and 41.9% (p = 0.005) for CBD 10 and 20 mg/kg/day, respectively Significantly higher drop seizure l RR e for CBD vs. PBO: 14% for PBO vs. 36% (p = 0.003) and 39% (p < 0.001) for CBD 10 and 20 mg/kg/day, respectively Higher drop seizure l freedom rate during maintenance phase (12 weeks) for CBD vs. PBO: 1% for PBO vs. 4% and 7% for CBD 10 and 20 mg/kg/day, respectively (statistical comparison not performed); no patient was free of drop seizures during the entire 14‐week treatment period GWPCARE4 63 : Significantly greater reduction in drop seizure l frequency for CBD vs. PBO: median frequency decreased 21.8% for PBO vs. 43.9% for CBD 20 mg/kg/day (p = 0.0135) Significantly higher drop seizure l RR e for CBD vs. PBO: 24% for PBO vs. 44% for CBD 20 mg/kg/day (p = 0.0043) Significantly higher drop seizure l freedom rate during maintenance phase (12 weeks) for CBD vs. PBO: 0% for PBO vs. 6% for CBD 20 mg/kg/day (p = 0.0246); no patient was free of drop seizures during the entire 14‐week treatment period	Two Phase 3 PBO‐controlled RCTs in LGS (GWPCARE3 and GWPCARE4; 14‐week treatment periods) 62 , 63 : GWPCARE3 62 : Significantly greater reduction in total seizure frequency for CBD vs. PBO: median frequency decreased 18.5% for PBO vs. 36.4% (p = 0.002) and 38.4% (p = 0.009) for CBD 10 and 20 mg/kg/day, respectively Greater reduction in non‐drop seizure frequency for CBD vs. PBO: median frequency decreased 34.3% for PBO vs. 61.1% and 54.6% for CBD 10 and 20 mg/kg/day, respectively (statistical comparison not performed) GWPCARE4 63 : Significantly greater reduction in total seizure frequency for CBD vs. PBO: median frequency decreased 13.7% for PBO vs. 41.2% for CBD 20 mg/kg/day (p = 0.0005) Significantly greater reduction in non‐drop seizure frequency for CBD vs. PBO: median frequency decreased 22.9% for PBO vs. 49.4% for CBD 20 mg/kg/day (p = 0.0044)	Common b AE: lethargy 4	Common b AEs: irritability, aggression 4	Very common j AEs include somnolence, decreased appetite 4 CBD can cause dose‐related elevations of liver transaminases, particularly in patients taking concomitant VPA 4 Concomitant use of CLB and CBD increases the incidence of somnolence and sedation 58
FFA	Serotonin‐releasing agent that stimulates multiple 5‐HT receptor sub‐types through the release of serotonin; may reduce seizures by acting as an agonist at specific serotonin receptors, including the 5‐HT1D, 5‐HT2A, and 5‐HT2C receptors, and also by acting on the sigma‐1 receptor 6	Phase 3 PBO‐controlled RCT in LGS 45 : Significantly greater reduction in drop seizure m frequency for FFA vs. PBO: median frequency decreased 7.6% for PBO vs. 14.2% (p = NS) and 26.5% (p = 0.001) for FFA 0.2 and 0.7 mg/kg/day, respectively Significantly higher drop seizure m RR e for FFA vs. PBO: 10% for PBO vs. 28% (p = 0.005) and 25% (p = 0.02) for FFA 0.2 and 0.7 mg/kg/day, respectively GTCS were more responsive to FFA than other seizure types	Phase 3 PBO‐controlled RCT in LGS 45 : Significantly greater reduction in the frequency of all countable motor seizures n for FFA vs. PBO: median frequency decreased 8.4% for PBO vs. 11.8% (p = NS) and 26.3% (p = 0.003) for FFA 0.2 and 0.7 mg/kg/day, respectively Significantly greater proportion of patients rated as “very much improved" or “much improved” by site investigators on CGI‐I for FFA vs. PBO: 6% for PBO vs. 20% (p = 0.01) and 26% (p = 0.001) for FFA 0.2 and 0.7 mg/kg/day, respectively Significantly greater proportion of patients rated as “very much improved" or “much improved” by caregivers on CGI‐I for FFA vs. PBO: 5% for PBO vs. 27% (p < 0.001) and 34% (p < 0.001) for FFA 0.2 and 0.7 mg/kg/day, respectively	Common b AE: lethargy 6	Common b AE: aggression 6	Very common j AEs include decreased appetite 6

Abbreviations: AE, adverse event; ASM, antiseizure medication; Ca²⁺, calcium ion; CBD, cannabidiol; CGI‐I, Clinical Global Impression‐Improvement scale; CLB, clobazam; ENT‐1, equilibrative nucleoside transporter 1; FFA, fenfluramine; FLB, felbamate; GABA, gamma aminobutyric acid; GE, global evaluation; GPR55, G protein‐coupled receptor 55; GTCS, generalized tonic–clonic seizures; LGS, Lennox–Gastaut syndrome; LTG, lamotrigine; Na⁺, sodium ion; NMDA, N‐methyl‐d‐aspartate receptor; NS, not significant; PBO, placebo; RCT, randomized controlled trial; RR, responder rate; RUF, rufinamide; TPM, topiramate; TRPV‐1, transient receptor potential vanilloid 1; VPA, valproate.

^a Effects on cognition and behavioral AEs are based on each ASM's European Union Summary of Product Characteristics, with the exception of FLB, for which the United States Prescribing Information was used (since not approved in the European Union);

^b ≥1/100 to <1/10;

^c ≥1/10 000 to <1/1000;

^d Tonic, atonic, and major myoclonic seizures;

^e ≥50% seizure frequency reduction from baseline;

^f <1/10 000;

^g Tonic‐atonic seizures;

^h Drop seizures and tonic–clonic seizures;

ⁱ ≥1/1000 to <1/100;

^j ≥1/10;

^k ≥1/1000 to <1/100;

^l Atonic, tonic and tonic–clonic seizures;

^m Generalized tonic–clonic, focal to bilateral tonic–clonic, tonic, atonic, or tonic or atonic seizures;

ⁿ Generalised tonic–clonic, tonic, clonic, atonic, tonic or atonic, and clearly recognizable focal seizures.

4.2.1. Highly purified cannabidiol (CBD)

CBD is licensed by the European Medicines Agency (EMA) as adjunctive therapy, specifically in conjunction with CLB, for seizures associated with LGS in patients aged ≥2 years. ⁴ , ⁶⁵ In the United States, CBD is licensed for the treatment of seizures associated with LGS in patients aged ≥1 year. ⁵ The efficacy of CBD in treating seizures associated with LGS was demonstrated in two placebo‐controlled RCTs ⁶² , ⁶³ (Table 1). Patients completing these RCTs could continue into a long‐term open‐label extension study (median duration, 1090 days; n = 366). ⁶⁶ Reductions in drop seizure and total seizure frequency were maintained over the long term: median percent reductions from baseline ranged from 48% to 71% for drop seizures and from 48% to 68% for total seizures over 156 weeks. Responder rates (≥50% seizure frequency reduction from baseline) ranged from 49% to 68% for drop seizures and from 48% to 65% for total seizures, and seizure freedom rates ranged from 3% to 11% for drop seizures and from 1% to 7% for total seizures. Most AEs were mild or moderate in intensity, and the most common AEs were convulsion (39%), diarrhea (38%), pyrexia (34%), and somnolence (29%). ⁶⁶ CBD demonstrates many complex interactions with a broad range of hepatic drug‐metabolizing enzymes, which may result in significant drug–drug interactions. ⁴⁰ Clinicians should be aware of such interactions and use CBD wisely, in particular in patients with moderate or severe hepatic impairment. Given the interindividual response and the relationship between the dose administered and CBD blood levels, therapeutic drug monitoring of CBD and concomitant ASMs may be valuable in the clinical management of patients. ⁶⁷

In the real‐world setting, a Phase 4, retrospective, chart review study assessed the effectiveness and tolerability of adjunctive CBD, without concomitant use of CLB, in 92 patients with LGS and 15 patients with Dravet syndrome aged ≥2 years, who were treated for ≥3 months as part of a European Early Access Program. ⁶⁸ In patients with LGS, the median change from baseline in drop seizure frequency per 28 days over 3‐month intervals varied from −6.2% to −20.9%. A ≥50% reduction in drop seizure frequency was experienced by 18.8% (13/69) of patients after 3 months and 30.2% (16/53) after 12 months, and seizure freedom was achieved by 1.4% (1/69) of patients after 3 months and 3.8% (2/53) after 12 months. The mean number of seizure‐free days per 28 days in patients with LGS was 7.0 at baseline and this increased by 1.7 days after 12 months. The most common AEs (≥5% of patients in the overall LGS and Dravet population) were somnolence (5.6%) and diarrhea (5.6%). ⁶⁸ Similar results emerged from a large Italian Expanded Access Program study, in which CBD was administered to 93 individuals with highly refractory LGS (n = 63) or Dravet syndrome (n = 30) for 12 months. ⁶⁹ After 12 months, the responder rate (≥50% seizure frequency reduction from baseline) and seizure freedom rate were 49.0% and 3.9%, respectively, for total seizures, and 45.1% and 7.8%, respectively, for convulsive seizures (i.e., clonic, tonic, tonic–clonic, atonic, and focal to the bilateral tonic–clonic). The most frequently reported AEs (≥10% of patients) were somnolence (22.6%), diarrhea (11.8%), and transaminase elevation (10.7%). ⁶⁹

4.2.2. Fenfluramine (FFA)

FFA is licensed in Europe as adjunctive therapy for seizures associated with LGS in patients aged ≥2 years, ⁶ and in the United States for the treatment of seizures associated with LGS in patients aged ≥2 years. ⁷ The efficacy of FFA in treating seizures associated with LGS was established in a Phase 3 placebo‐controlled RCT ⁴⁵ (Table 1). Patients completing this RCT could continue into a long‐term open‐label extension study (currently ongoing; median duration, 364 days; n = 247 [October 2020]). ⁷⁰ An interim analysis of the open‐label extension study demonstrated that efficacy was maintained over the long term: the median reduction in monthly drop seizure frequency was 28.6% over the entire open‐label extension (p < 0.0001) and 50.5% at Month 15 (p < 0.0001); 31.1% of patients experienced ≥50% reduction in drop seizure frequency over the entire open‐label extension and 51.2% after approximately 1 year. ⁷⁰ The longest interval between drop seizures increased from a median of 2.0–7.0 days. ⁷⁰ , ⁷¹ Generalized tonic–clonic seizures and tonic seizures were most responsive to treatment, with median reductions over the entire open‐label extension of 48.8% and 35.8%, respectively (p < 0.0001 for both). ⁷⁰ The median reduction in non‐drop seizure frequency was 45.9% (p = 0.0038). Approximately one‐third of investigators (37.6%) and caregivers (35.2%) rated their patients as having “very much improved” or “much improved” on the Clinical Global Impression of Improvement at the last assessment. The most frequent treatment‐emergent AEs were decreased appetite (16.2%) and fatigue (13.4%). ⁷⁰ There were no cases of valvular heart disease or pulmonary arterial hypertension in either the initial Phase 3 RCT or open‐label extension study, ⁴⁵ , ⁷⁰ despite FFA being originally withdrawn from the market for this reason when previously developed as an appetite suppressant for the treatment of obesity. ⁴⁰ The EMA states that echocardiogram monitoring should be conducted every 6 months for the first 2 years of FFA treatment and annually thereafter, ⁶ and the US Food and Drug Administration states that echocardiogram assessments should be obtained before, during (every 6 months), and once after (3–6 months) the final dose of FFA. ⁷

In the real‐world setting, a Spanish, single‐center, retrospective, post‐marketing study assessed the effectiveness and tolerability of FFA when used in routine clinical practice to treat patients with Dravet syndrome or other developmental and genetic epileptic encephalopathies (n = 54) and LGS (n = 14). ⁷² Seizure frequency data were available for 13 patients with LGS. After 12 months, the sustained responder rate (≥50% seizure frequency reduction from baseline for ≥2 consecutive months) and seizure freedom rate in the patients with LGS were 50.0% and 8.3%, respectively. In the overall study population (i.e., patients with Dravet syndrome and LGS), the most frequently reported AEs (≥10% of patients) were decreased appetite (35.9%), somnolence (15.9%), and irritability (15.6%). No cases of pulmonary arterial hypertension or valvular heart disease were observed. ⁷²

4.3. Use of ASMs not approved for drop seizures in LGS

Soticlestat is a first‐in‐class selective inhibitor of cholesterol 24‐hydroxylase (CH24H) that is being investigated as an adjunctive ASM for DEEs, including LGS. ⁴⁰ , ⁷³ , ⁷⁴ A Phase 2 placebo‐controlled RCT assessed the efficacy and safety of soticlestat in patients aged 2–17 years with LGS (n = 88) and Dravet syndrome (n = 51). ⁷⁴ During the 12‐week maintenance period, the median placebo‐adjusted percent reduction in the frequency of drop seizures in patients with LGS was 17.1% (p = not significant). In the combined population, the most frequently reported treatment‐emergent AEs (≥5% greater with soticlestat vs. placebo) were lethargy and constipation. ⁷⁴ Phase 3 trials of soticlestat in patients with LGS are currently ongoing (ClinicalTrials.gov NCT04938427 and NCT05163314); however, it was recently announced that one of these (NCT04938427) failed to meet its primary endpoint (reduction in major motor drop seizures). ⁷⁵ A Phase 3 placebo‐controlled trial of carisbamate (which belongs to the alkyl‐carbamate family of drugs that also includes FLB ⁴⁰ ) in adult and pediatric patients with LGS is also currently ongoing (ClinicalTrials.gov NCT05219617).

A recent systematic review and meta‐analysis offers a comprehensive assessment of steroid efficacy beyond IESS. ⁷⁶ However, the level of evidence for LGS is low ⁷⁶ and clinicians should be cautious in extrapolating these findings directly to LGS due to the distinct pathophysiological mechanisms involved. There remains a notable gap between the absence of RCTs evaluating the efficacy of steroids in treating LGS and their widespread utilization in many medical centers. Although no RCTs have been conducted to date, ¹⁰ anecdotal evidence suggests that steroids may offer short‐term benefits in managing atypical absence seizures, drop seizures, and NCSE. ¹ However, it is important to acknowledge that relapse is frequent. ¹ Furthermore, prolonged usage of steroids carries the risk of serious adverse effects, such as hyperlipidemia, osteoporosis, and growth suppression. ⁷⁷ Consequently, it is advisable to reserve steroid administration for acute worsening periods, steering clear of prolonged usage, according to current recommendations. ¹

4.4. Real‐world evidence for the use of pharmacological agents in the treatment of LGS

Several ASMs (both approved and not approved for use in LGS) have been specifically assessed in patients with LGS in the real‐world clinical practice setting.

4.4.1. Real‐world evidence for ASMs approved for use in LGS

The greatest amount of real‐world evidence is for RUF (approximately 15 studies worldwide), and this has been reviewed elsewhere. ⁷⁸ , ⁷⁹ , ⁸⁰ In summary, clinical practice studies of RUF have confirmed evidence from clinical trials, demonstrating that it is effective and generally well tolerated in children as young as 1 year and adults, and that it may be particularly effective in treating drop seizures and generalized tonic–clonic seizures. ⁷⁹ They also indicate that a “low and slow” approach to RUF dosing and titration can improve tolerability while maintaining effectiveness. ⁷⁸

4.4.2. Real‐world evidence for ASMs not approved for use in LGS

Some broad‐spectrum ASMs may be useful in treating seizures associated with LGS, despite having limited evidence of use specifically in patients with LGS. These include levetiracetam (LEV), zonisamide (ZNS), and perampanel (PER). A Phase 3 trial of adjunctive PER in patients with LGS was terminated early due to recruitment challenges, exacerbated by the COVID‐19 pandemic, ⁸¹ and all other evidence for these three ASMs comes from the clinical practice setting. A multicenter, open‐label, observational study conducted in 55 patients with LGS assessed the efficacy and tolerability of adjunctive LEV over an 8‐week maintenance period. ⁸² A >50% seizure frequency reduction was observed in 32 (58.2%) patients and 15 (27.3%) achieved seizure freedom; seven of 12 (58.3%) patients with drop seizures experienced a >50% seizure frequency reduction. The most frequently reported AE was hyperactivity (12.7%). ⁸² A multicenter study conducted in Korea in 62 patients with LGS assessed the efficacy and safety of adjunctive ZNS over ≥12 months. ⁸³ A >50% seizure frequency reduction was observed in 32 (51.6%) patients, three of whom achieved seizure freedom. AEs included transient somnolence and anorexia. ⁸³ In a prospective study of 13 patients with LGS who were treated with adjunctive PER over a mean follow‐up duration of 10.8 months, nine (69.2%) were responders (≥50% reduction in total seizure frequency), and nine (69.2%) were rated by their clinician as “much improved” or “very much improved.” ⁸⁴ Four patients (30.8%) discontinued PER due to the lack of efficacy (n = 2) and seizure aggravation (n = 2), and six patients (46.2%) experienced AEs (decreased activity/social interaction, n = 3; behavioral disturbance with agitation, n = 2; fatigue, n = 2). Seven patients (53.8%) experienced improvements in cognitive function and/or behavior. ⁸⁴ A multicenter, retrospective, observational cohort study assessed the long‐term effectiveness of adjunctive PER in 87 patients with LGS. ⁸⁵ During a median follow‐up of 11 months, responder rates (≥50% seizure frequency reduction from baseline) were 41.4% for all seizure types and 61.1% for drop seizures. Seizure relapse occurred in 36.1% of patients over 36 months, but none of the patients experienced seizure worsening during PER treatment. The most common AEs (≥10% of patients) were behavioral disturbances (21.8%) and somnolence (12.6%). ⁸⁵ Another retrospective, open‐label study assessed the efficacy and tolerability of adjunctive PER when used to treat 71 adult patients with LGS (mean age, 40.1 years; 62 with “pure LGS”; nine with “LGS‐like epileptic encephalopathy”) over a median duration of 539 days. ⁸⁶ The overall responder rate (≥50% seizure frequency reduction) was 64.8%, with 16.9% of patients experiencing ≥90% seizure frequency reduction. Six patients (8.5%) experienced seizure aggravation with PER. Negative behavioral changes were observed in 32.4% of patients, but “positive side effects” (feelings of wellness and calm; improved contact) were observed in 5.6% of patients. ⁸⁶ Finally, a retrospective study assessed the 12‐month effectiveness and tolerability of adjunctive cenobamate in four adults with LGS (aged 27–37 years). ⁸⁷ At 12 months, the seizure frequency reduction from baseline ranged from 25% to 74%, with two patients experiencing ≥50% seizure frequency reduction. AEs included somnolence (n = 2) and ataxia, dizziness, and vomiting (n = 1). ⁸⁷

4.5. Non‐pharmacological treatment of LGS

Non‐pharmacological treatment options for LGS include ketogenic diet therapies (KDT), resective surgery, vagus nerve stimulation, corpus callosotomy, and deep brain stimulation. Since the evidence base for the use of all of these treatments, except deep brain stimulation, has changed little since previously reviewed by the authors, ¹ these are summarized in Table 2. ¹ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ , ⁹⁵ , ⁹⁶ , ⁹⁷ , ⁹⁸ , ⁹⁹ , ¹⁰⁰ , ¹⁰¹ , ¹⁰² , ¹⁰³ , ¹⁰⁴ , ¹⁰⁵ , ¹⁰⁶ , ¹⁰⁷ , ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ , ¹¹¹ , ¹¹² , ¹¹³ , ¹¹⁴

TABLE 2.

Summary of non‐pharmacological treatment options for LGS.

Treatment	Key details
Ketogenic diet therapies (Ketogenic diet; Modified Atkins diet)	Substantial evidence of efficacy and safety/tolerability in LGS: A 2012 literature review of 18 studies found that 47% of patients experienced >50% seizure frequency reduction and 16% achieved seizure freedom after 3–36 months 88 A 2022 literature review of seven studies found that 17%–73% of patients experienced ≥50% seizure frequency reduction and 1%–24% of patients achieved seizure freedom after treatment durations of 1 month to 19 years 89 Minimal AEs, generally alleviated by adjusting diet 1 Response is usually observed within 3 months, allowing effectiveness to be assessed relatively quickly 1 Requires a dedicated ketogenic diet team and commitment from carers 1
Resective surgery	All patients should be investigated for a potential localized brain abnormality (although not frequently found) since resective surgery can be curative in carefully selected individuals 1 In a retrospective review of 90 patients with LGS who underwent resective surgery, the most common pathological finding was malformation of cortical development 90 During a mean postoperative follow‐up duration of 6.1 years, 45 patients (50.0%) were seizure‐free and 15 (16.7%) reported infrequent seizures 90 Seizure‐free patients demonstrated significantly better adaptive behavior and social competence versus patients with persistent seizures (p < 0.05) 90
Vagus nerve stimulation	Some evidence of effectiveness in LGS 91 , 92 , 93 , 94 , 95 A meta‐analysis of seven prospective and 10 retrospective studies in patients with LGS (N = 480) demonstrated a responder rate of 54% 96 Recommended in AAN guidelines as a treatment option for LGS (Level C) 91 AEs are typically stimulation‐related, reversible, mild to moderate in intensity, and usually decrease over time; AEs seldom require device removal 97 Patient may experience continuing improvement over time, and improvement in alertness 98 Can be used with other pharmacological and non‐pharmacological therapies; does not interact with ASMs 1 Less invasive than callosotomy, but involves implantation of a device 1 Re‐evaluation of patients for the presence of a seizure focus may be appropriate following vagus nerve stimulation, as generalized epileptiform discharges may become more localized 1 , 99
Corpus callosotomy	Some evidence of effectiveness in LGS 100 , 101 , 102 , 103 , 104 Particularly effective for atonic seizures/drop seizures (more effective than vagus nerve stimulation) 100 , 105 Recommended for use when drop seizures are especially problematic 1 Other treatments can be used following corpus callosotomy if needed 1 Less invasive techniques, such as radiosurgical corpus callosotomy or stereotactic laser anterior corpus callosotomy, have obtained similar results in terms of efficacy in controlling drop seizures 106 , 107 A complete corpus callosotomy could be more efficacious than an anterior corpus callosotomy for some seizure types 105 Following complete corpus callosotomy, transient disconnection syndrome has been observed in 50% of patients but this resolved in all the patients following 2 years of follow‐up108 Re‐evaluation of patients for the presence of a seizure focus may be appropriate following corpus callosotomy, as generalized epileptiform discharges may become more localized 1 , 99
Deep brain stimulation	Several studies have reported favorable outcomes in patients with LGS when the centromedian thalamic nucleus has been targeted with deep brain stimulation, with seizure reductions of 25–100% reported 109 , 110 , 111 , 112 , 113 Cognitive effects and side effects of deep brain stimulation of the centromedian thalamic nucleus are currently unclear 114 Mood and memory problems have been reported following DBS of the anterior thalamic nucleus, but it is conceivable that different thalamic structures might vary in their role in cognition and mood 114 More research is required to establish the potential utility of deep brain stimulation in LGS114 More expensive than other non‐pharmacological treatment options, such as vagus nerve stimulation 114

Abbreviations: AAN, American Academy of Neurology; AE, adverse event; ASM, antiseizure medication; LGS, Lennox–Gastaut syndrome.

DBS is an emerging treatment for drug‐resistant epilepsies, including LGS. Given the heterogeneous nature of LGS, certain etiologies may be more responsive to DBS than others. A better understanding of anatomical targets, stimulation parameters, neurophysiological biomarkers, and appropriate patient selection would help clarify the potential utility of DBS in LGS. ¹¹⁴

5. GENERAL PRINCIPLES OF LGS TREATMENT

LGS is a DEE with a very poor prognosis in terms of long‐term seizure control and cognitive function. Despite a range of treatment options, the achievement of seizure freedom is very unlikely. ¹¹⁵ LGS adversely affects the QoL of both patients and caregivers ²⁰ , ¹¹⁶ , ¹¹⁷ , ¹¹⁸ ; seizure control (particularly drop seizure control) and days without seizures being the main determinants of QoL in LGS. ¹¹⁹ , ¹²⁰ From childhood to adulthood, management of LGS must carefully balance the burden of treatment against its potential side effects, always focusing on the patient's overall QoL ⁹ ; consequently, the goal of treatment should be optimal seizure control (particularly with regards to the more disabling seizure types), rather than seizure freedom. ¹ Treatment goals may change according to the patient's age and stage of disease, and this should be addressed as a key part of ongoing re‐evaluation and transition of care. ¹ Treatment goals and plans should be reassessed regularly by centers of reference with extensive experience in the treatment of rare and complex epilepsies, and plans should include a transparent explanation of potential AEs and how these should be managed. ¹ , ¹⁰

In the context of LGS, a personalized approach to treatment can be complex, due to the nature of the syndrome itself. LGS is characterized by a variety of different seizure types and a spectrum of cognitive impairments, along with a diverse range of etiologies that can contribute to its development. This heterogeneity makes it challenging to tailor treatments to individual patient profiles since what may be effective for one person may not be for another. Additionally, the syndrome often requires a multifaceted treatment strategy that includes a combination of medications, dietary treatment, and potentially surgical interventions. Due to the complexity of LGS, treatment strategies must be highly individualized and frequently adjusted to accommodate the unique and evolving needs of each patient. This tailored approach is essential for managing the multiple seizure types and comorbidities commonly associated with LGS, ultimately aiming to optimize therapeutic outcomes in this challenging patient population.

It is important to note, however, that while a personalized approach may be difficult, it is not entirely unfeasible. It requires careful consideration of each patient's unique circumstances, including the etiology of their condition, the specific characteristics of their seizures, their response to previous treatments, and their overall health status. A thorough evaluation by a team of specialists can help in creating a more tailored treatment plan, but it does require ongoing assessment and adjustments, which can be resource‐intensive. Moreover, there is preliminary evidence that computational methods for assessing multiple phenotypic features in specific epileptic disorders might help facilitate the development of future precision‐medicine approaches. ¹²¹

6. TREATMENT ALGORITHMS FOR LGS

The following recommendations and practical advice are the authors' expert opinions, based on the available evidence and their clinical experience, and include relevant updates to the 2017 guidance.

The clinical features diagnostic of LGS may evolve with time, and this must consequently remain a consideration when planning therapy in young children presenting with multiple seizure types. Since it may take time for patients to develop all the clinical and EEG features of LGS, it is recommended that—once all attempts to exclude other diagnoses have been undertaken—a patient presenting with features suggestive of LGS should be treated as though they have LGS until their full clinical/EEG profile becomes clear.

6.1. Newly diagnosed patients with LGS and patients with LGS already receiving ASM therapy

The treatment algorithm for a newly diagnosed patient with LGS or a patient with LGS already receiving ASM therapy is presented in Figure 1. ³⁶

FIGURE 1.

Treatment algorithm for a newly diagnosed patient with LGS or a patient with LGS already receiving ASM therapy. ^aCould be considered according to the most prevalent seizure type; ^bCLB is used without validation from an RCT or other high‐level evidence study; ^cBRV and CNB are broad‐spectrum ASMs without any specific evidence in LGS; ^dWith clinical goals; ^eIn combination with VPA and/or CLB. AE, adverse event; ASM, antiseizure medication; BRV, brivaracetam; CBD, cannabidiol; CBZ, carbamazepine; CLB, clobazam; CNB, cenobamate; EMA, European Medicines Agency; ESL, eslicarbazepine acetate; ETX, ethosuximide; FDA, Food and Drug Administration; FFA, fenfluramine; FLB, felbamate; LEV, levetiracetam; LGS, Lennox–Gastaut syndrome; LTG, lamotrigine; MRI, magnetic resonance imaging; NCSE, non‐convulsive status epilepticus; OXC, oxcarbazepine; PB, phenobarbital; PER, perampanel; PHT, phenytoin; RCT, randomized controlled trial; RUF, rufinamide; STP, stiripentol; TGB, tiagabine; TPM, topiramate; VPA, sodium valproate; ZNS, zonisamide.

6.1.1. Patients with newly diagnosed LGS or suspected LGS

For a patient presenting with newly diagnosed LGS or suspected LGS, the recommended first‐line treatment is VPA (Figure 1). VPA should generally not be used in women of childbearing potential, due to its high teratogenic potential, ⁴⁶ unless no suitable alternatives exist, and effective contraception is used. It is important to consider that female patients with LGS are generally unlikely to have children due to the severity of their condition. The risk–benefit ratio must be carefully balanced in each case. If VPA therapy does not provide adequate seizure control, LTG should be added as the first adjunctive therapy. Since VPA inhibits LTG metabolism, a decreased LTG dose with slow titration should be used. If VPA plus LTG does not provide adequate seizure control, RUF should be initiated as adjunctive therapy. Once RUF has been initiated, attempts should be made to discontinue either VPA or LTG, and, if VPA is discontinued, the LTG dose should be increased. When considering adding an adjunctive therapy, every attempt should be made to discontinue one of the two previous ASMs once the new ASM has been introduced, since there is no evidence for the effectiveness of more than two ASMs in combination, and the use of multiple ASMs unnecessarily raises the risk of side effects and/or drug–drug interactions.

If adequate seizure control is not achieved with the addition of RUF, the choice of the next adjunctive ASM should be discussed with the patient/parent/caregiver/clinical team, based on the patient's clinical profile and taking the caregiver's experience into consideration. When discussing further ASM options, the following points should be considered:

• Short‐term treatment with CLB may be considered, temporarily increasing the ASM load. Since there might be a risk of tolerance, dependence, and cognitive/behavioral AEs, CLB should preferably be used on an intermittent, short‐term (3–5 days) basis when acute episodes occur (expert opinion). Such episodes include sustained absence seizures (duration >1 day), cluster seizures, and NCSE. This aside, where tolerance is not an issue, CLB may prove a useful adjunctive ASM, particularly where drop seizures are troublesome. It should be noted that high‐dose benzodiazepines may increase sleepiness and consequently increase the risk of tonic seizures.

• There is increasing evidence supporting the use of CBD as adjunctive therapy in LGS patients, ¹²² although this requires confirmation in further trials and long‐term safety studies.

• Although considered a useful treatment option in LGS, particularly for drop seizures, ¹²³ FLB is not licensed by the EMA due to associated risks of aplastic anemia and liver failure. However, in some countries (e.g., France, Italy, Spain, and United Kingdom) it is authorized for use when prescribed by hospital‐based expert teams, and its use is therefore likely to be country‐specific. FLB could be considered as an early adjunctive treatment option in LGS, with careful consideration of its potential benefits against its risks. ¹²⁴ If deemed necessary, it is advisable to closely monitor the patient, including regular blood counts and liver function tests.

• TPM is licensed for LGS and may be used once other lines of treatment have been tried, but it has a greater potential for cognitive and behavioral adverse effects than other marketed ASMs.

• FFA has been a relatively recent addition to the options approved for the treatment of seizures associated with LGS. It may have particular efficacy in treating generalized tonic–clonic seizures. Real‐world evidence for the use of FFA in LGS is currently limited.

• ASMs used to treat LGS may be associated with behavioral and cognitive AEs, ¹²⁵ which should be considered when selecting an ASM at the individual patient level.

Based on the authors' clinical experience, several ASMs that do not have a specific license for LGS may nevertheless be considered as adjunctive therapy options. Given the variety of seizure types that may be present in the same patient, consideration should be given to the evidence for a particular ASM in the treatment of the specific seizure type(s) present (rather than the syndrome per se), with the caveat that the treating clinician is knowledgeable and up‐to‐date on the potential risks for drug–drug interactions. LEV, ZNS, and PER have demonstrated some evidence of effectiveness in LGS. All are broad‐spectrum in their modes of action and may therefore be useful in treating multiple seizure types. LEV particularly may be a useful adjunct as it has few interactions with other medications. Ethosuximide can be useful for the treatment of absence seizures but should always be combined with an ASM that is effective in treating generalized tonic–clonic seizures and tonic/atonic seizures since it is not effective for these seizure types. CLB may also be useful for the treatment of absence seizures. Benzodiazepines other than CLB may be considered for intermittent, short‐term use for acute episodes (as recommended for CLB), but should not be used in combination with each other. Although there is no published evidence to support the efficacy of stiripentol (STP) in treating LGS, it can be used in combination with VPA and/or CLB. The usual approach is to add STP to VPA and subsequently add a small dose of CLB if required. It should be noted that STP will increase VPA and CLB levels, so some dose adjustments might be required. Carbamazepine, oxcarbazepine, eslicarbazepine acetate, tiagabine, and phenytoin should only be used with caution, due to the potential risk of aggravation of drop seizures with a myoclonic component (expert opinion).

The use of non‐pharmacological treatment approaches should be considered alongside the use of ASMs and discussed from the outset as part of the patient's treatment plan. Some patients/parents/caregivers may propose trying KDT relatively early on in the patient's treatment pathway. If this is the case, KDT can be tried if VPA plus LTG does not provide an adequate seizure response, before RUF is initiated. Alternatively, KDT may be introduced later, when multiple ASMs have been tried. Since response to KDT is usually observed within 3 months, ¹²⁶ this therapeutic option can be explored relatively quickly.

An evaluation regarding resective surgery or disconnective surgery must be considered in all patients, particularly those with LGS with structural etiology who have lesions predominantly in one hemisphere or tuberous sclerosis. This should be conducted by a specialized team with extensive experience in pre‐surgical evaluation. As callosotomy involves surgery (except when a gamma knife is used, which is not widely available), its use will largely depend on patient/parent/caregiver choice, considering it is targeted at drop seizures. Complete callosotomy seems to be more effective than anterior callosotomy for drop seizures. ¹⁰⁸ , ¹²⁷ Callosotomy may be considered as an early treatment option in patients for whom drop seizures are particularly problematic (e.g., if the patient suffers repeated injury from drop seizures, or is wheelchair‐bound due to drop seizure frequency). ¹²⁸ Re‐evaluation with EEG and MRI is recommended before and after callosotomy to detect any changes resulting from the procedure (e.g., development of focal seizures). Although VNS is less invasive than callosotomy, it still involves a surgical procedure and its use will therefore largely depend on patient/parent/caregiver choice. It can take time for the effects of VNS to become apparent, with further improvement over time. The decision as to when VNS should be used will depend on a variety of factors, including age, time since LGS diagnosis, and whether the patient has been experiencing troublesome or intolerable AEs with ASM treatment. VNS can be used in conjunction with ASM therapy and with callosotomy.

6.1.2. Patients with LGS that has evolved from another type of epilepsy (e.g., IESS)

Many patients will either have been treated with ASM therapy to initially control seizures before LGS diagnosis or will have developed LGS having progressed over time from another epilepsy syndrome, such as IESS. ¹²⁹ Most patients presenting with apparently generalized seizures will already be receiving VPA. If this is the case, then the treatment algorithm shown in Figure 1 can be followed (i.e., adding LTG as the first adjunctive therapy if VPA does not provide adequate seizure control, etc.). If the patient is already being treated with another first‐line ASM (typically LEV), then VPA therapy should be initiated and the other therapy tapered off and discontinued. Thereafter, the treatment algorithm is the same as for a newly diagnosed patient (Figure 1).

If the patient is being treated with more than one ASM and neither is VPA, then VPA therapy should be initiated and one of the previous ASMs tapered off and discontinued. If seizure control is inadequate after introducing VPA and the second ASM is not LTG, then LTG should be initiated and the other non‐VPA ASM tapered off and discontinued. Thereafter, the treatment algorithm is the same as for a newly diagnosed patient (Figure 1).

6.2. Older patients with established LGS

It is recommended that patients with established LGS undergo review by a neurologist with established experience in the treatment of complex epilepsies and drug interactions on at least an annual basis, comprising a thorough reassessment of their diagnosis (in terms of epilepsy syndrome and etiology) and treatment plan. The diagnosis should be re‐evaluated by repeating investigations conducted at initial diagnosis and/or conducting investigations that were previously not undertaken (EEG [including sleep‐EEG, if possible], MRI, and genetic testing), to confirm the diagnosis and help elucidate etiology. Results of previous investigations should be reviewed alongside those of new investigations. Since the patient's clinical and EEG features continue to evolve, a diagnosis other than LGS may become apparent and treatment should be adapted accordingly. Clinicians should always be alert to the possibility that the diagnosis may change and be vigilant to the possibility of treatable etiologies. Clinicians should also be aware that the “classic” EEG features (SSW complexes) may evolve and/or disappear later in the disease course. ²⁹ , ¹³⁰ Loss of these features does not necessarily mean that the patient no longer has LGS, but this possibility must be considered alongside the reassessment of other clinical/EEG features. Genetic counseling should be offered, if appropriate. EEG should be repeated whenever there are any concerns over diagnosis, signs of deterioration, or suspected NCSE. It is mandatory to have a recent baseline EEG for comparison, in case the patient presents with suspected NCSE in the future.

6.3. Management of NCSE

Up to 75% of patients with LGS experience episodes of NCSE. ¹⁰ Its clinical presentation can vary from a mild confusional state to a coma. ¹³¹ However, presentation in this population may be quite subtle; by definition, NCSE is a change in behavior and EEG from baseline. ¹³² The diminished responsiveness may be insidious in its onset, and therefore may be missed. A high index of suspicion is therefore required on the part of the physician, particularly if a previous baseline EEG is not available for comparison. ¹³² Patients with LGS should be regularly assessed for the development of possible NCSE, including EEG, and, where possible, EEG results should be compared with a baseline recording. ¹³¹ Expert advice may be required to confirm or refute an NCSE diagnosis.

Treatment for NCSE is less urgent than for convulsive status epilepticus and overtreatment should be avoided since most patients do not require aggressive treatment (expert opinion). Patients should not be admitted to intensive care to induce anesthetic treatment, since this could potentially be more harmful to the patient than the condition itself. NCSE should be treated with CLB and/or steroids (expert opinion), although there is currently no evidence for any agent or dosing regimen, ⁷⁶ and the use of steroids varies according to local protocol. High‐dose intravenous VPA can also be effective, ¹³³ and there is emerging data that FFA can be used as a treatment option for NCSE in patients with LGS. ¹³⁴ The goal of treatment is to return the EEG to its pre‐NCSE baseline pattern, and treatment should be optimized with ongoing review. Patients with NCSE should be referred for specialist advice and/or EEG monitoring.

7. OTHER CONSIDERATIONS FOR LGS MANAGEMENT

7.1. Comorbidities

The lives of patients with LGS are affected not only by frequent and disabling seizures but also by a range of comorbidities, including physical disability, cognitive impairment, behavioral problems (e.g., hyperactivity, aggressiveness, and autistic traits), and sleep disturbances. ¹ , ⁹ , ³² Mobility is further restricted by the protective equipment (e.g., wheelchair, helmet, and faceguard) that is often required to prevent seizure‐related injury. ¹ Management of comorbidities in LGS is a key aspect of care. This should include careful consideration when choosing ASM treatment, not only with regard to potential drug interactions with medications used to treat comorbidities but also because some ASMs may themselves cause or worsen certain comorbidities (such as cognitive impairment and depression). ¹

7.2. Impact on QoL

The QoL of patients with LGS is severely impaired throughout their lives. ¹ , ⁹ , ²⁰ , ¹¹⁶ The ability to participate in everyday activities is restricted by the physical impact of LGS and the need for protective equipment, and cognitive and behavioral problems often prevent mainstream school attendance. ¹ , ⁹ As individuals progress into adulthood, LGS affects their independence, ability to work, personal relationships and social participation. ¹ , ⁹ LGS also severely affects the QoL of parents, caregivers and other family members. ²⁰ , ¹¹⁶ , ¹¹⁷ , ¹¹⁸

7.3. Transition from childhood to adulthood

Transitioning from pediatric to adult care is difficult for both patients and families but provides an important opportunity for reassessment of all aspects of care, including re‐evaluation of a patient's etiology, EEG features and pharmacological and non‐pharmacological treatment options. ¹ , ⁹ The multidisciplinary needs of the patient and family should also be reassessed at this time (e.g., social care support, provision of community/residential care, and psychiatric support). ¹ , ⁹ The transition process can be facilitated by attendance at transition/teenager clinics. ¹ , ⁹ , ¹³⁵ , ¹³⁶ All patients should be reviewed by a neurologist at least annually. ¹

7.4. Need for multidisciplinary care

LGS requires a multidisciplinary, individualized approach to care throughout the course of life, which addresses the patient's medical, psychological, educational, and social needs, as well as the needs of their caregiver/family. ¹ , ¹³⁷ Ideally, these multidisciplinary needs should be re‐assessed annually, and include the patient's health and social care needs, their potential need for institutionalization (particularly in adulthood), and support for the caregiver/family. ¹

8. FUTURE DIRECTIONS

Despite difficulties in identifying and characterizing LGS, future efforts should be made to formalize clinical trial strategies to provide clearer evidence of the potential benefits of therapeutic interventions, including non‐pharmacological treatments (e.g., cell therapy and gene therapy) in addition to ASMs. Eligibility criteria for study entry should be consistent with the current definition of LGS, ³ and EEG features should be clearly defined and standardized between study centers. ¹⁰ Inclusion of patients with non‐specific DEEs with stratification based on DEE type might also be explored. Duration of epilepsy prior to enrollment should be taken into account since this can impact the presentation of characteristic LGS features. ¹⁰ Improved early identification of LGS may allow subgroups of patients to be defined, such as those transitioning to LGS from IESS and those with early advanced neurophysiological changes; similarly, early elucidation of etiology (e.g., genetic pathogenic variants and structural abnormalities) may allow better understanding of the potential benefits of an intervention in different etiology subgroups. In terms of study design, the use of an active comparator would be preferable to comparing with a placebo, ¹⁰ and the choice of active comparator could be based on the next standard ASM in the treatment algorithm presented in Figure 1. Treatment duration should ideally be at least 6 months and studies could be designed to assess the effects of an intervention in subgroups with different LGS etiologies. The number and type of concomitant therapies and rescue therapies allowed during the trial should be specified. ¹⁰ The types of seizures assessed in the trial should also be specified and, in addition to seizure reduction, assessment of the severity of seizures, duration of seizures, number of days without seizures, and improvement/worsening of specific seizure types could be considered as potential study endpoints. ¹⁰ , ¹³⁸ , ¹³⁹ Specific EEG biomarkers that may serve as indicators of treatment response or disease progression in LGS could be explored; these might involve changes in spike–wave index, reduction in nocturnal paroxysmal fast activity, spectral power, or other quantitative measures. ¹⁰ Additional endpoints should also be considered, such as the effects of an intervention on QoL, cognition, and behavior. ¹⁰

Future directions in LGS management are likely to focus increasingly on the use of personalized/precision medicine; for example, the potential utility of genetic testing, including pharmacogenomics to convey information for patients sensitive to severe adverse drug reactions. ¹⁴⁰ , ¹⁴¹ , ¹⁴² If LGS is known to be caused by an underlying disease then this disease should be specifically treated whenever possible; for example, using everolimus to treat tuberous sclerosis complex ¹⁴³ and GAP activity toward RAGs (GATOR) complex 1 (GATOR 1) epilepsies ¹⁴⁴ ; resective surgery for patients with cortical dysplasia due to disheveled EGL‐10 and pleckstrin domain‐containing protein 5 (DEPDC5) variants ¹⁴⁵ ; and the use of KDT and avoidance of VPA in patients with mitochondrial diseases. ¹⁴⁶ , ¹⁴⁷ , ¹⁴⁸ , ¹⁴⁹ , ¹⁵⁰ , ¹⁵¹ It is also likely that existing pharmacological agents may be repurposed for use in LGS, as has been the case for FFA. ¹⁵²

CONFLICT OF INTEREST STATEMENT

Stéphane Auvin is Deputy Editor for Epilepsia. He has served as a consultant or received honoraria for lectures from Angelini Pharma, Biocodex, Biomarin, Eisai, Encoded, Grintherapeutics, Jazz Pharma, Neuraxpharm, Nutricia, Proveca, Stoke, Supernus, Takeda, UCB Pharma, and Xenon. He has been an investigator for clinical trials for Eisai, Marinus, Proveca, Takeda, and UCB Pharma. Alexis Arzimanoglou is Editor‐in‐Chief Emeritus for Epileptic Disorders, educational journal of the ILAE, and Associate Editor for the European Journal of Paediatric Neurology. He has served as a consultant or received honoraria for lectures from Biocodex, Biomarin, Eisai, Jazz Pharma, Neuraxpharm, Nutricia, Takeda, and UCB Pharma. He has been an investigator for clinical trials and/or DMC member for Eisai, Jazz Pharma, and UCB Pharma. His Institution received educational grants from UCB Pharma, GW, and the Caixa Foundation. In his mission as Coordinator of the ERN EpiCARE, his Institution received funding from the European Commission. Mercè Falip has served as a consultant or received honoraria for lectures from Angelini Pharma, Eisai, Esteve, Jazz Pharma, Livanova, Neuraxpharm, and UCB Pharma. She has served as an investigator for clinical trials for UCB Pharma, Jazz Pharma, and Angelini Pharma. Pasquale Striano is the Associate Editor for Epilepsia Open and Epileptic Disorders. He has served as a consultant or received honoraria for lectures from Angelini Pharma, Biocodex, Biomarin, Eisai, Encoded, Jazz Pharma, Neuraxpharm, Proveca, Takeda, and UCB Pharma. He has been an investigator for clinical trials for Longboard, Takeda, and UCB Pharma. He is supported by #NEXTGENERATIONEU (NGEU) and funded by the Ministry of University and Research (MUR), National Recovery and Resilience Plan (NRRP), project MNESYS (PE0000006)—A Multiscale Integrated Approach to the Study of the Nervous System in Health and Disease (DN. 1553 11.10.2022). This work was also supported by the Italian Ministry of Health, RICERCA CORRENTE 2024. J Helen Cross has served as an investigator for clinical trials for UCB, Jazz, Stoke Therapeutics, and Ultragenyx. She has served on advisory boards for Biocodex, UCB, and Jazz Pharmaceuticals and spoken in educational symposia for Biocodex, UCB, Nutricia, and Jazz; all remuneration has been paid to her department. She is in receipt of grants from Engineering and Physical Sciences Research Council (UK), Epilepsy Research UK, Great Ormond Street Hospital Children's Charity, National Institute for Health and Care Research. All research is supported by the National Institute of Health Research Biomedical Research Centre at Great Ormond Street Hospital.

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Auvin S, Arzimanoglou A, Falip M, Striano P, Cross JH. Refining management strategies for Lennox–Gastaut syndrome: Updated algorithms and practical approaches. Epilepsia Open. 2025;10:85–106. 10.1002/epi4.13075