Efficacy and Safety of Vagus Nerve Stimulation in Lennox–Gastaut Syndrome: A Scoping Review

Lennox–Gastaut syndrome (LGS) is a severe developmental and epileptic encephalopathy characterized by drug-resistant seizures, cognitive impairments, and abnormal electroencephalographic patterns. Vagus nerve stimulation (VNS) is a widely used neuromodulation therapy for LGS, but its effects on seizure outcomes, different seizure types, non-seizure outcomes, and adverse events in this population have not been comprehensively reviewed. To conduct a scoping review on the use of VNS in LGS, a literature search was performed in PubMed, OVID, Web of Science, and Embase from inception to 9 June 2024, using relevant keywords and without restrictions on study design. The search yielded forty eligible studies (twenty-four retrospective cohorts, fourteen prospective cohorts, and two registry analyses) comprising 1400 LGS patients treated with VNS. No randomized controlled trials were identified. Across studies, the median seizure reduction ranged from 20.6% to 65%, with 0% to 100% of patients achieving a ≥50% seizure reduction. No consistent preoperative biomarker of VNS responsiveness was identified in LGS. Although inconsistent among different studies, tonic, atonic, and tonic–clonic seizures responded best, while focal seizures responded worst. Improvements in seizure severity, alertness, and quality of life were reported in some studies, but cognitive and adaptive functioning generally remained unchanged. Adverse events were mostly mild and transient, including hoarseness, cough, and paresthesia. Device-related complications and infections were uncommon. In conclusion, further research is needed to better understand VNS’s position in the evolving LGS treatment landscape and its cost effectiveness.


Introduction
Lennox-Gastaut syndrome (LGS) is a prevalent developmental and epileptic encephalopathy (DEE), with around 20% of patients with infantile epilepsy progressing to LGS [1,2].This syndrome is characterized by [1] the presence of various drug-resistant seizures beginning before 18 years of age [including tonic seizures and potentially atypical absence, atonic, myoclonic, focal impaired awareness (FIAS), generalized tonic-clonic seizures (GTCS), and epileptic spasms]; Ref. [2] cognitive and frequently behavioral impairments; and [3] diffuse slow spike-and-wave and generalized paroxysmal fast activity on electroencephalography (EEG) [3].LGS is associated with a high rate of seizure emergencies, including nonconvulsive status and high premature mortality [2,4].The high frequency of seizures, intellectual disabilities, and behavioral impairments result in a low quality of life and significant daily functional support needs for individuals with LGS, imposing a heavy burden on caregivers, which manifests as fatigue, sleep disturbances, social isolation, poor mental health, and financial challenges [5].Despite polytherapy, pharmacoresistance is common in LGS, leading to frequent use of non-pharmacological treatments such as epilepsy surgery, dietary therapy, and neuromodulation [4].
Vagus Nerve Stimulation (VNS) Therapy ® was the first FDA-approved neuromodulation therapy for drug-resistant epilepsy (DRE).The European Community approved VNS in 1994, followed by the FDA in the USA in 1997 for patients over 12 years old with pharmacoresisant focal seizures [6].A later evidence-based guideline from the American Academy of Neurology (AAN) recommended its use in children with either focal or generalized epilepsy, based on 14 Class III studies, pooling data from 481 children with a 55% responder rate [7].This guideline also provided a level C recommendation for VNS in patients with LGS based on four Class III studies [8][9][10][11].A pooled analysis of 113 patients with LGS showed a 55% (95% CI 46-64%) responder rate [7].
Currently, VNS is among the most frequently performed pediatric neurosurgical procedures, with LGS being the leading cause of pediatric epilepsy treated with VNS [12].Consequently, some systematic reviews and meta-analyses have calculated the seizure responsiveness rate of LGS following VNS, with or without comparisons to other surgical treatments like corpus callosotomy [13][14][15].However, there has been a lack of exploration into other important outcomes, specifically VNS's effect on various seizure types, nonseizure outcomes including quality of life, and adverse effects [1].A holistic exploration of these outcomes is crucial for clinical decision-making, as VNS and other nonpharmacological options do not render patients seizure-free, with approximately half of the patients being responders [16].Therefore, physicians and caregivers need to consider these additional outcomes.Additionally, emerging neuromodulation techniques with diverse attributes, such as deep brain stimulation (DBS) and responsive neurostimulation (RNS), are increasingly accepted for treating LGS [17].This evolution necessitates a thorough review of all attributes of VNS to determine the best neuromodulation treatment approach that aligns with the preferences of patients, families, and physicians.

Materials and Methods
To allow a comprehensive analysis of the efficacy and safety of VNS in LGS, we chose to undertake a scoping review approach rather than a systematic review.Our goal was to broadly examine the literature and address several previously unexplored research questions following a PICO framework (Table 1).Studies where LGS patient numbers are not specifically identified.Non-English language studies without available translations.
The review followed the steps described by the PRISMA guidelines [18], Supplementary Materials.(Figure 1) For this scoping review, we utilized multiple searchable databases, including PubMed, OVID, Web of Science, and Embase from inception to 9 June 2024, using relevant keywords and without restrictions on study design.Our search strategy was structured as follows: ((terms for Lennox-Gastaut syndrome OR Lennox-Gastaut OR LGS) AND (VNS OR vagal nerve stimulation OR vagus nerve stimulation)).We also checked the reference lists of eligible studies and relevant systematic reviews using a 'snowball sampling' approach.The databases were searched from their inception to 9 June 2024.We imposed no restrictions on study design.LGS and non-LGS data not separately provided (5), missing data (5) Studies included in review (n =40) Abstracts identified through the searches were screened using predefined inclusion and exclusion criteria (Table 1) tailored to address the research questions.Relevant studies were retrieved in full-text format and further screened for the outcomes of interest.The criteria for inclusion and exclusion remained the same as those used for abstract screening, with the additional exclusion criteria detailed in Figure 1.Studies were excluded if they included patients with LGS but did not provide a specific number for the LGS group or did not report specific outcomes for the LGS subgroups (at least one outcome related to seizures, non-seizure events, or adverse events had to be reported separately for LGS).Single case reports of adverse effects were not included in the assessment but were mentioned in the text.Given that the studies were expected to be small, heterogeneous, and largely non-comparative, no formal statistical synthesis or quality assessment was planned.All qualitative and quantitative data were summarized in a narrative synthesis.

Study Characteristics
The search yielded forty eligible studies (twenty-four retrospective cohorts, fourteen prospective cohorts, and two registry analyses) comprising 1400 LGS patients treated with VNS (Table 2).Multiple studies were reported from USA, Canada, Sweden, Italy, Brazil, Korea, and Norway.No randomized placebo-controlled trial (high vs. low stimulation) has specifically evaluated the efficacy of VNS in LGS.The largest study included a registry study with 564 LGS patients who were compared to 1128 non-LGS patients [19].Another large registry study included 552 patients (167 with outcomes data) with LGS from the VNS Patient Registry in the United States in 2001 [20].The largest single-center study included 62 patients [21].The largest multicenter study, a retrospective open-label study conducted at 11 centers across the European Union, consisted of 146 LGS patients [22].Many studies in this review included different epilepsy groups, with LGS often representing the largest subgroup [23].Some studies compared VNS efficacy with corpus callosotomy [9,24].Others compared LGS with other epilepsy groups, such as genetic generalized epilepsy, other DEEs, or severe epilepsy with multiple independent spike foci [19,25,26].One study reported seizure outcomes in LGS for a range of therapeutic modalities, including VNS [27].A total of 3 out of 14 had a >50% seizure reduction.Improvement in seizure frequency and quality of life in some patients.Compared to baseline, after one year and three years of stimulation, no significant changes in cognitive level and adaptive behavior scores were observed in group level but 4 had clinically relevant improvement in adaptive behavior (at least 5 point increase in standard scores) mainly due to increased alertness and a better social reciprocity.Six patients had better QoL; this was independent of seizure outcome.

Seizure Reduction
The seizure outcomes were reported as percentages of mean or median seizure reduction.Some studies only reported seizure responders, defined as patients with more than 50% seizure reduction [23].Few studies utilized the McHugh Classification, which is modeled on the Engel classification but specifically tailored for use in VNS, including subdivisions to capture changes in ictal or postictal severity [28].
One study reported differences in seizure responsiveness between day and night, with a greater effect during the daytime and complete abolition of daytime drop attacks [32].
The effectiveness of VNS in LGS patients was noted to be comparable to other DRE patients [41].One study showed a seizure responder rate of 57.5% in LGS compared to 61.5% in other groups after 24 months of VNS implantation [19].The median seizure reduction was 64.3% in the LGS compared to 66.7% in the non-LGS group [19].Another study showed median seizure reduction of 52.1% in LGS compared to 57.1% in the non-LGS group with no statistical difference between these groups [21].LGS was also compared with genetic generalized epilepsy (GGE).Suller Marti et al. showed that 41.7% of patients in the LGS group had an overall seizure reduction of 50% or more, compared to 64.7% in the GGE group, without statistical significance between the groups [26].One study suggested that VNS may have superior efficacy in treating focal epilepsy compared to LGS [42].Another study indicated a better seizure freedom rate and a greater than 75% seizure reduction rate in GGE compared to LGS following VNS [26].

Seizure Types and Responsiveness to VNS
The efficacy of VNS in reducing specific seizure types in LGS patients varies widely and inconsistently across different studies and among patients within a single study [43].Specific seizure types in LGS that showed significant reductions included atonic seizures (up to 64.6% median reduction) and tonic seizures (50%) [10,24,33].Dibue reported a 90% decrease in drop seizures (atonic and other seizure types leading to falls) over 24 months, with 78% of patients experiencing a 50% or greater reduction in drop seizures [19].In general, GTCS showed a less significant reduction than atonic or tonic seizures [14].One study reported only a 33.3% reduction in tonic-clonic seizures (TCS) [21].However, contrasting findings were noted in other studies [8,24].Cukiert reported no change in tonic and atonic seizures following VNS, but found significant improvement in GTC seizures [24].Improvement in TCS was also noted in other studies, with 74.1% and 60% of patients experiencing a 50% or greater reduction in focal to bilateral TCS and bilateral TCS, respectively.Other studies showed the most significant reductions in GTCS [33,44].However, efficacy against TCS might diminish during follow-up [19].Compared to tonic, atonic, and TCS, FIAS had a less robust reduction of 33% [21].Frost et al. showed FIAS decreased by only 20% after 1 month and 23% after 3 months [33].Some studies did not report absence seizures due to difficulty in counting [21].However, other studies reported excellent efficacy against absence seizures [24,33,44].Effects on spasms and myoclonic seizures were infrequently reported.One study showed spasms reduced by a median of 21.7% [21].Seven of twelve LGS patients reported clinically significant reductions in myoclonic jerks, and one reported becoming free of this seizure type [21].Significant reduction in myoclonic seizures was also reported in other studies [24,37].One study reported the effects of VNS on residual seizures of various types after corpus callosotomy [45].The best response was noted in tonic seizures, while no response was seen in atonic seizures [45].Residual myoclonic seizures after callosotomy responded better than atypical absence seizures [45].

Variables Affecting Responsiveness
Some studies examined the different variables associated with LGS and its responsiveness to VNS.In the LGS group, Cersosimo did not find any differences between symptomatic and cryptogenic patients regarding the severity of intellectual disability or seizure reduction [11,46].Additionally, there were no differences in outcomes between patients who had previously had West syndrome and those who had not [46].History of prior callosotomy did not alter the responsiveness [20].Age younger than 12 vs.older did not show any improved responsiveness [21,33].A change in heart rate variability (HRV) reported in a patient with LGS who initially had minimal HRV but showed improved variability following VNS [47].HRV was also proposed as a measure to detect VNS responsiveness.However, one study showed no difference in pre-ictal HRV parameters between VNS responders and non-responders, despite a significant increase in high-frequency (HF) power, which indicates parasympathetic tone, significantly increased in the pre-ictal period for both groups [48].

Immediate and Long-Term Efficacy
Several studies have reported an immediate effect before stimulation is turned on and also longitudinal follow-up over several years.Cukiert et al. observed an implantation effect lasting an average of 20.2 days in 14 out of 24 children [24].However, this has not been consistently reported, and most studies indicated greater efficacy between 3 and 12 months of stimulation, with progressive improvement over the first 24 months also reported.
In a retrospective multicenter study of 50 LGS patients treated with VNS, the median seizure reduction was 42% after one month, increasing to 58.2% after three months [33].Progressive improvement over 24 months was reported in a study that showed a mean reduction in seizure frequency of 8% at 3 months, 33% at 12 months, and 50% at 2 years [42].Another study reported seizure responder rates of 28.5%, 32.9%, and 39.1% at 6, 12, and 24 months, respectively [22].A separate study on closed-loop VNS showed seizure responder rates of 55%, 67.7%, and 65% at 6, 12, and 24 months, respectively [49].Longer follow-up durations of up to 5 years have also been reported.In one study, seizure reduction at 6, 12, and 60 months was 17.7%, 35.5%, and 30.6%, respectively [21].This study noted a statistically significant increase in seizure reduction from 6 to 24 months (p = 0.006) and persistence of efficacy but no further improvement since 24 months of follow-up.Additionally, a progressive increase in seizure responsiveness greater than 75% was reported, from 15% at 1 month to 35% at 6 months, and then 38% at 12 months [33].However, one study reported a reversal of response in bilateral TCS during long-term follow-up [19].

Impact on Seizure Severity and Status Epilepticus
Approximately half of the patients with LGS experience prolonged seizures, seizure clusters, and status epilepticus [50].VNS studies have reported reductions in seizure duration, ictal and postictal severity, and seizure clustering in a significant proportion of LGS patients [10,21,22,33].One study noted a decrease in seizure severity scores measured with the adapted Chalfont severity scales [8].Another study observed a significant decrease in hospitalizations and episodes of status epilepticus in some patients after VNS implantation [32].Another study showed seizure-related hospitalization decreased from 96.6% to 51.7% after VNS implantation [26].Most responders in the study experienced milder, shorter seizures with faster recovery times [32].Nagarajan et al. reported that four out of five patients had over 50% seizure reduction, with most experiencing reduced duration and severity of seizures, and two had reduced diazepam use [51].Lundgren et al. found that two out of four patients had decreased seizure severity scores [52].Frost et al. showed that around 30% of patients had reduced postictal severity and seizure clustering [33].Mikati et al. reported that three out of six patients had improvement in seizure severity, and seizure duration decreased in two patients [53].Seizure severity reduction may be more pronounced in LGS than in other groups.One study reported that DEE patients (62 out of 105 had LGS) showed greater improvement in ictal or postictal severity (43.8% vs. 28.3%,p = 0.006) compared to patients without intellectual disability, possibly leading to a higher rate of continuing use of VNS at 5 years [21].

Standard vs. Rapid Cycling, Magnet Stimulation, and Autostimulation
The optimal device parameters remain uncertain and do not show clear associations with treatment outcomes [33].Rapid cycling has not been shown to have additional benefit than standard stimulation [54].Although additional magnet-activated stimulation for seizures associated with LGS has not consistently shown efficacy [34,51], one study demonstrated a reduction in seizure duration or severity in 61.0% of 105 patients with DEE, including 62 with LGS [21].
A newer method employing closed-loop autostimulation, which delivers electrical stimulation promptly in response to ictal tachycardia, has been found effective without increasing the average charge per day [55].Winston compared closed-loop versus open-loop VNS devices and observed greater seizure reduction and more favorable clinical outcomes with closed-loop devices at 9 months, though not at 24 months, across the entire cohort [56].However, there were no significant differences noted in the LGS subgroup during either follow-up period.Conversely, Abdelmoity et al. reported that auto-stimulation may offer additional benefits over standard stimulation.In a study involving 71 children with LGS, 55% achieved >50% seizure reduction at six months, 67.7% at 12 months, and 65% at 24 months with autostimulation VNS [49].At 12 months, 11% were completely seizure-free, and at 24 months, 17% remained seizure-free [49].Moreover, when transitioning from traditional to auto-stimulation models, 60-70% experienced further reductions in seizure frequency from baseline [49].

Cognitive and Adaptive Functioning, Quality of Life Outcomes
Almost all individuals with LGS have cognitive and behavioral impairments of varying severity, which limits standardized assessment.The severity of impairment may also preclude formal testing.However, some VNS studies have attempted to use various cognitive and psychological/behavioral instruments.Hallbrook et al. assessed cognitive development using the Bayley Scales of Infant Development and the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), while Parker et al. utilized the Vineland Adaptive Behavior Scale and the Wellcome Quality of Life (QoL) Assessment, specifically designed for LGS [29,57].Parker also employed the British Picture Vocabulary Scale (BPVS), the Leiter International Performance Scale, and Conner's Parent/Teacher Rating Scales [57].Aldenkamp used Dutch scales for provider assessments, including the Dutch version of the Bayley Developmental Scale, the Dutch version of the McCarthy Developmental Scale, and other scales to measure independence, communication skills, task acceptance, social sensitivity, and behavior disorders [39].Mood was evaluated using a Dutch scale.Similar detailed evaluations were conducted by Majoie et al. [8].
Many LGS patients cannot complete age-appropriate test batteries.For example, one study reported that only one subject out of twenty-four could undergo formal psychological evaluation [24].Even composite scaled scores from tests suitable for developmental age may be less useful due to floor effects (test scores falling ≥ 3 standard deviations below normalized sample means), necessitating the tracking of raw scores.Due to these difficulties, several studies relied on indirect indicators such as parental or provider reports using simple scales or questionnaires.For example, Frost utilized provider assessments employing a 5-point scale, while other studies used various tools like the QOLIE-31 questionnaire and the Child Epilepsy Questionnaire Parental Form (CEQ-P) [24,33,46,53].Different studies used various scales, including the parental visual analog scale [29,52], a parental 3-point scale [51], a Likert scale [25], and a parental global evaluation score [40].
Most studies did not observe significant changes or deterioration in cognitive measures or adaptive functioning [8,25,31,39].However, one study reported a mild positive change in mental age during follow-up, with the treatment effect being most prominent in the group with the highest mental age at baseline [39].Some studies reported improvements in quality of life, alertness, social communication, and reductions in behavioral disturbances in certain patients [24,31,46,52,58,59].Frost et al. found improvements in alertness for 50-60% of patients, over 20% for verbal abilities, and less for memory/ambulation during 3-6 months of the follow-up period [33].Cukiert et al. observed improvement in attention (based on 18 items Swanson, Nolan, and Pelham Questionnaire IV) in 21 out of 24 children and reported quality of life/health improvements in up to 50% (mean 25%) of 20 children [24].Another study of closed-loop VNS showed at least a 50% improvement rate in one or more quality-of-life measures (alertness, sleep, development, academic performance, and attention) in most patients [49].Four studies indicated that changes in quality of life, cognition, and behavior were observed independently of seizure reduction outcomes [24,25,33,39].One study reported no significant changes in cognitive level and adaptive behavior scores at group level following 1 and 3 years of stimulation, but four patients had clinically relevant improvement in adaptive behavior (at least 5-point increase in standard scores) mainly due to increased alertness and a better social reciprocity [25].One study reported that six out of seven patients showed improvement on the Clinical Global Impression (CGI) improvement scale, with two patients each experiencing marked, moderate, and minimal improvements [59].Katagiri et al. reported that VNS responders had better conversational ability than non-responders [45].However, consistent improvements in quality of life were not observed across all studies.For instance, Aldenkamp et al. did not find significant changes in quality-of-life measures [39].One study reported mood decline in one out of twenty-four patients [33], and another study reported behavioral worsening in two out of ten patients [32].
One study demonstrated reduced direct healthcare (medication and hospitalization) and non-healthcare (traveling) costs following VNS [8].The number of days of reduced functioning of the child was significantly less compared with the baseline period [8].The assessed cost-effectiveness ratio was EUR 16.93 per reduction in one seizure [8].

Adverse Events
Adverse events were not specifically reported in all studies [39].However, most studies noted no major complications [9,24,25,29,31,32,45,46,51].One study did report four serious adverse events, including three cases of status epilepticus and one sudden death [44].Nevertheless, all-cause mortality and SUDEP rates for VNS therapy in LGS patients have not yet been reported.Some other studies mentioned side effects but did not distinguish between LGS and other epilepsy patients [21].
While the majority of patients experience some side effects [33], these are mostly mild and transient, such as hoarseness, cough, paresthesia (neck tingling), voice alterations, shortness of breath, pain, drooling, dyspepsia, decreased appetite, hiccups, and headache.Although some studies reported only 5-30% of these side effects [9,24,60], others reported 50-67% [10,32].These side effects are not excessively high in LGS compared to other populations.One study showed the frequency of side effects similar in LGS vs. non-LGS groups (62.1% vs 64.7%) [26].Side effects during VNS use are typically related to the "on" phase of stimulation and tend to diminish over time, usually disappearing within five years of implantation.Other unusual side effects include torticollis, inappropriate laughter, involuntary movement, urinary retention, aspiration, and lower facial paresis.A 16-year-old patient with LGS was reported to have symptomatic complete atrioventricular block due to VNS [61].
Children with cognitive impairments often pick at or touch wounds, increasing the risk of infection.However, less than 5% of LGS patients reported infections that required antibiotic treatment, and even fewer needed the external hardware removed and reinserted because of this [33].One large study showed that two out of sixty-four patients developed an infectious complication [46]; in one case, the device was moved to the right side.To help these children, doctors have implanted the device in alternate locations, including under the pectoral muscles.
LGS patients are particularly susceptible to sleep-disordered breathing.A study of thirteen patients with DEEs, including six with LGS, showed a median arousal index of 29 per hour, with moderate to severe obstruction in 53.8% of patients and central apnea in 23% [62].This is concerning because patients with VNS have been shown to have an increased incidence of obstructive and central apneas [63].Although rare, some LGS patients have reported breathing difficulties following VNS, but a systematic exploration of sleep-disordered breathing has not been conducted.No studies reported the use of the SenTiva™ model, the newest VNS device with programmable stimulation that can be adjusted at different times of the day (e.g., day and night settings), to see if decreased stimulation parameters at night can alleviate sleep-disordered breathing.
Some patients experienced the emergence of a new seizure type following VNS that was not reported at baseline [19].Status epilepticus and seizure worsening were reported in relation to higher settings.In four adult patients with LGS, increasing the output current to 2.00-2.25 mA and switching from standard stimulation (30 s on, 300 s off) to rapid stimulation (7 s on, 30 s off) caused worsened seizure frequency due to the increased total charge delivered per day [64].This issue was resolved when the total charge delivered per day was reduced [64].Additionally, one LGS patient experienced forced normalization (subacute psychosis) after turning off VNS, which had not provided significant benefit since its implantation [65].However, the patient became seizure-free after VNS was turned off, suggesting a possible seizure-worsening effect of VNS that resolved when the device was deactivated, leading to psychosis [65].

Discussion
This scoping review, comprising 1400 patients with LGS from 40 studies, showed a wide range of seizure reduction and responder rates.Super responder and seizure freedom rates were less than 10%.Progressive improvement may be seen up to 24 months after implantation.The effectiveness of VNS in LGS patients was noted to be comparable to other DRE patients.Age, prior history of infantile spasms or corpus callosotomy, and specific etiology did not have any prognostic importance regarding responsiveness.Although rapid cycling may not have additional benefits over standard stimulation, magnet stimulation and closed-loop stimulation may provide additional efficacy.VNS may reduce seizure duration, ictal and postictal severity, seizure clustering, and seizure-related hospitalizations.The review assessed the adverse effects of VNS and noted primarily stimulation-related mild transient effects, with rare reports of infection and device-related complications.
In this scoping review, we also attempted to explore VNS's effect on specific seizure types, as changes in total seizure counts might not reflect clinical improvement in LGS since different seizure types are not equally disabling.We noted that the efficacy of VNS in reducing specific seizure types in LGS patients varied widely and inconsistently across different studies.However, tonic, atonic, and tonic-clonic seizures responded best, while focal seizures responded worst [2,5,19].Many studies did not specify particular seizure types for subgroups and likely only considered seizure reduction for the most disabling seizure types.Nevertheless, understanding the specific effects on seizure types may help with clinical decision-making in LGS patients, as many have distinct and most disabling seizure types.For example, atonic and TCS are usually more disabling than myoclonic and absence seizures, although the latter might be more frequent [21].Although the most common seizure type in LGS is tonic seizures, TCS has specific significance regarding SUDEP, while atonic seizures are associated with severe craniofacial injuries due to falls.Although tonic, atonic, and tonic-clonic seizures were noted to respond best following VNS, the cause of the discrepancy between the effects of VNS on various seizure types in LGS is unclear.It may be related to study design, such as recall bias or errors in seizure classification or counting.In some studies, drop seizures were counted, which might capture either or both tonic and atonic seizures.Additionally, absence and myoclonic seizures are also difficult to count and thus prone to erroneous counting.Even some motor seizures, such as tonic seizures in LGS, are grossly underreported in the diary method compared to objective counting by VEEG [66].
The scoping review also explored VNS's effect on cognitive and adaptive functioning, and quality of life outcomes in individuals with LGS.Although the use of age-appropriate psychological and cognitive test batteries was rare, indirect indicators such as parental or provider reports showed improvements in alertness and some measures of quality of life.However, there are no LGS-specific cognitive or QoL instruments, and the commonly used batteries varied widely.
There are several limitations in this study.The first limitation is the heterogeneous study population and characteristics, with many outcomes preventing us from conducting a meta-analysis (meta-analysis of the seizure responder rate of VNS in this population has already been performed).The second limitation is that we did not find any randomized placebo-controlled trials (high vs. low stimulation) specifically evaluating the efficacy of VNS in LGS.All included studies were uncontrolled and open label, which may have led to an overestimation of the effectiveness of VNS.Although we did not perform a formal bias assessment, retrospective studies had issues with data quality, and prospective studies faced problems with attrition.The largest studies were based on observational registries with data collected prospectively but analyzed retrospectively.This design introduced limitations such as voluntary data collection by clinical staff without stringent monitoring, potentially affecting the consistency of seizure type classification and LGS diagnosis across different sites.Additionally, uncontrolled variability in anti-seizure medication (ASM) combinations and stimulation parameters may have introduced bias.The time frame for outcome assessment varies considerably among studies.In summary, most studies have small sample sizes and only meet the American Academy of Neurology Class III evidence criteria for therapeutic studies.
Despite these limitations, the study has several strengths.This report represents the most comprehensive review of VNS therapy in LGS patients so far and highlights potentially important differences across studies.Previous systematic reviews included 17 studies with 176-480 patients, whereas we included more than double the number of studies and approximately three times the number of individual patients [13][14][15].Rather than comparing VNS with corpus callosotomy, this scoping review also delved deeper into non-seizure outcomes and adverse effects of VNS for a comprehensive outlook.As the emergence of RNS and DBS use may decrease the use of callosotomy, this review sets the stage for a comparison of VNS with other neuromodulatory approaches [17].

Conclusions
This scoping review assessed VNS use in LGS, highlighting significant knowledge gaps.Seizure outcomes following VNS in LGS vary widely, and identifying ideal candidates remains challenging despite exploring various biomarkers.As the epileptic connectivity in LGS becomes better understood, LGS-specific connectivity studies are necessary to preoperatively identify the best candidates for VNS [67].The effectiveness of preoperative non-invasive VNS (e.g., transcutaneous auricular) in identifying candidates for invasive VNS has not been explored [68].Although this review concentrated on invasive VNS, non-invasive VNS (nVNS) has also been examined in the context of epilepsy treatment [69].nVNS presents several benefits, including simpler integration of control groups in clinical studies, reduced costs, fewer side effects, and potentially favorable effects on cognition [70].Nevertheless, additional studies specifically targeting LGS are required to comprehensively assess its efficacy.
Current treatment decisions for LGS are complex, involving ASM regimen modifications, callosotomy, or neuromodulation devices like DBS, RNS, or VNS, without specifically targeting seizure types [4].Atonic seizures may respond better to callosotomy than VNS, but callosotomy has a higher complication rate and is less cost-effective [71].The increasing availability of neuromodulation options complicates clinical decision-making, and the absence of comparative studies, especially for other seizure types, adds to this complexity.Conducting randomized head-to-head comparison studies is challenging, with only 21.3% of patients and caregivers willing to participate in randomization [72].Due to the growing acceptance of DBS and RNS, deciding whether to use these before VNS is a consideration; however, a discrete-choice experiment indicated that VNS might be preferred over DBS by most clinicians [73].For non-responders to VNS, DBS, or RNS may be effective alternatives, but further evaluation is needed to determine when these interventions should be used and which patients would benefit most from combined therapy [74,75].As non-pharmacological approaches primarily provide palliation with rare seizure freedom, VNS studies should also focus on non-seizure outcomes, such as quality of life, cognitive function, psychiatric issues, and daily living, using LGS-specific standardized measures (which should be developed through a consensus approach).Understanding the cost-effectiveness of VNS compared to other treatments is crucial given limited healthcare resources [71].More detailed economic evaluations can inform healthcare policy and reimbursement decisions.

Figure 1 .
Figure 1.PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Ana flowchart.

Figure 1 .
Figure 1.PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flowchart.

Table 1 .
PICO Framework for Evaluating VNS in LGS Patients.

Table 1 .
PICO Framework for Evaluating VNS in LGS Patients.
PICO Element Details Patient PopulationPatients diagnosed with Lennox-Gastaut syndrome (LGS), all ages.Exclusion: studies specifically identifying LGS patients for any outcome measures.Intervention Vagus nerve stimulation (VNS) therapy Control Not required Outcome Measures -Seizure response: reduction in seizure frequency, responder rate, efficacy of VNS in re ducing specific seizure types in LGS, variables affecting seizure responsiveness, shortlong-term efficacy -Other seizure outcomes: seizure severity, status epilepticus -Effects of different stimulation parameters Reports sought for retrieval (n = 115) Reports not retrieved (n =0) Reports assessed for eligibility (n =116; 1 additional report included through 'snowball sampling') Reports excluded (n=76): non original article (34), outcome of interest not reported (30), duplicate study participants (2),

Table 2 .
Characteristics and key findings of all 40 included studies.
Significant regression from bilateral tonic-clonic response in LGS group.Within the 24 months of follow-up, there were 4 deaths in the LGS group and 5 deaths in the non-LGS group.24Meanseizure reduction was 57.6% (43.0-72.2%).Median seizure reduction was 52.1%.Not different from the majority of the study group who had unidentified TRE causes (mean reduction: 57.0%).Non-seizure outcomes, etc., were not given for specific LGS groups.

Table 2 .
Cont.Median age at implant was 13 years.At 1 month, median seizure reduction was 42%.At 3 and 6 months, reductions were 58.2% and 57.9%, respectively.Drop-attack seizures significantly decreased.QOL assessment (simple, invalidated, 5-point rating scale, which featured ratings of much worse, worse, same, better, and much better) was performed at 3 and 6 months.Improvement in alertness and verbal abilities in more patients than memory and mood.The least effect was shown on ambulation.The most common adverse events reported were voice alteration and coughing during stimulation.Other uncommon adverse events included increased drooling and behavioral changes.In total, 4% of patients had superficial wound infection; 10% had transient pain at the incision site; 44% had voice alteration/hoarseness.Moreover, 30% reported coughing, neck tingling (8%), nonspecific pain sensation (8%), shortness of breath during exertion (4%), decreased appetite (4%), hiccups (4%), and dyspepsia (4%).One (2%) patient reported dysphagia and insomnia, and another (2%) reported vomiting after stimulation started.One patient reported both ear pain and jaw pain as stimulation.Increased salivation (8%), worsening behavior or hyperactivity (6%).13Responder, 38.5%, partial responder(e (≥50% seizure reduction in 2 or less study periods) 23%, non-responder 38.5% in 24 months.Hajnsek Single center in Croatia Retrospective 3 Mean age of 26.67 ± 5.77 13.34 ± 2.88; p = 0.02 (follow-up duration unknown); improvement of mood and the general quality of life was reported in all patients.A total of 2/7 patients had >50% seizure reduction; no difference in pre-ictal HRV parameters between VNS responders and VNS non-responders could be found, but high frequency (HF) power, reflecting the parasympathetic tone increased significantly in the pre-ictal epoch in both VNS responders and VNS non-responders (p = 0.017, p = 0.004).
over 1 year follow-up; 2 were seizure free.Tonic seizures were the most common seizure type that was residual after CC.Responses were observed in five of the ten (50%) patients after VNS.Atypical absence seizures were less likely to be controlled with VNS (25% response rate).Myoclonic seizure was observed in four patients, two of whom were responders.None of the patients with residual drop attacks caused by atonic seizures responded to VNS.Patients who responded to VNS after CC could have conversations with others, while non-responders could not.After VNS, transient hoarseness was observed in two patients, and coughing only with magnet activation was observed in one patient.p < 0.001) than patients without ID.
A total of 5 out of 6 had a >50% seizure reduction.Reduction in seizure duration and severity, but the magnet was not useful.Total group outcome (n = 16) was given.On a three point scale, 12 of the 16 parents (entire group rather than LGS specific subgroup) reported that quality of life was definitely better and this correlated with improved seizure control.epilepsy, with and without drop attacks, than in LGS.Mean reduction in seizure frequency in LGS: 8% at 3 months, 33% at 12 months, and 50% at 2 years (not significant).