Refractory Epilepsy Screening Tool for Lennox–Gastaut Syndrome (REST-LGS)

The REST-LGS (Refractory Epilepsy Screening Tool for Lennox-Gastaut Syndrome) screens patients with refractory epilepsy for undiagnosed LGS using five clinical domains: seizure type history, EEG findings, cognitive profile, epilepsy onset and course, and injury history. Positive screens trigger specialist evaluation and access to LGS-specific therapies.

REST-LGS Calculator

Answer all 8 criteria with “Yes” or “No/Unknown”. Major criteria score 3 points each; minor criteria score 1 point each. A score >11 suggests likely LGS.

Clinical note: If uncertain about any answer, consult the patient’s family or caregiver. Lower scores due to missing or unknown data do not necessarily rule out LGS. This tool screens for LGS but does not replace specialist evaluation.

Clinical History

Minor (+1)

Persistent seizures despite 2+ antiseizure medications

Persistent seizures despite trial of 2 or more antiseizure medications (ASMs)

Minor (+1)

At least 2 seizure types

Patient has at least 2 distinct seizure types

Major (+3)

Seizure onset before 12 years of age

First seizure occurred before the patient was 12 years old

Minor (+1)

Helmet use or facial/head injuries

Uses or used a helmet, or has evidence of facial or head injuries (consistent with drop seizures)

Minor (+1)

Cognitive impairment since childhood

Cognitive impairment since childhood (may include past or current learning difficulties, history of special education, intellectual disabilities, or developmental delay)

Minor (+1)

History of VNS, ketogenic diet, or epilepsy surgery

History of vagal nerve stimulator (VNS), ketogenic diet, or epilepsy surgery

EEG Findings

Major (+3)

EEG: generalized slow spike-and-wave (<2.5 Hz)

History of EEG with generalized slow spike-and-wave (SSW) discharges at <2.5 Hz

Minor (+1)

EEG: other characteristic findings

Any 1 of the following EEG findings: multifocal spikes, symptomatic generalized discharges, generalized poly-spikes, generalized periods of attenuation of background or electrodecrement, or paroxysmal fast activity

Answer all 8 criteria to calculate the score.

Disclaimer: This calculator is for educational and screening purposes only. The REST-LGS tool does not replace formal neurological evaluation or EEG interpretation by a qualified specialist. A diagnosis of LGS requires clinical assessment by an epilepsy specialist and should not be made on the basis of this score alone.

REST-LGS: scoring criteria, interpretation, and clinical context

Overview

The Refractory Epilepsy Screening Tool for Lennox–Gastaut Syndrome (REST-LGS) is an 8-criterion screening instrument developed by Piña-Garza and colleagues (published in Epilepsy & Behavior, 2019) to improve the identification of patients with LGS, particularly in adult patients where the diagnosis is frequently delayed or missed. The tool uses a weighted scoring system with major and minor criteria, and was designed to be completed using chart review and caregiver input, making it practical in outpatient and inpatient settings.

Scoring criteria

Each criterion is answered Yes or No/Unknown. Major criteria contribute 3 points; minor criteria contribute 1 point. Maximum possible score is 12.

Criterion	Type	Yes
Persistent seizures despite 2+ antiseizure medications (ASMs)	Minor	+1
At least 2 seizure types	Minor	+1
Seizure onset before 12 years of age	Major	+3
Uses/used a helmet or has evidence of facial/head injuries	Minor	+1
Cognitive impairment since childhood (learning difficulties, special education, ID, or DD)	Minor	+1
History of VNS, ketogenic diet, or epilepsy surgery	Minor	+1
EEG: generalized slow spike-and-wave (SSW) discharges (<2.5 Hz)	Major	+3
EEG: multifocal spikes, symptomatic generalized discharges, poly-spikes, electrodecrement, or paroxysmal fast activity	Minor	+1

ID = intellectual disability; DD = developmental delay; VNS = vagal nerve stimulator; SSW = slow spike-and-wave; EEG = electroencephalogram.

Score interpretation

>11

Likely LGS

Referral to an LGS specialist is strongly recommended. Consider LGS-specific treatments (e.g., cannabidiol, clobazam, valproate, rufinamide, lamotrigine, fenfluramine, or dietary/neuromodulation therapies). This score requires all or nearly all criteria to be met, including both major criteria.

8–11

Possibly LGS

Further evaluation by a neurologist or epilepsy specialist is warranted. Additional EEG studies (including prolonged or ambulatory recording) and specialist review may clarify the diagnosis. The missing LGS-associated features may reflect incomplete data rather than their true absence.

Unlikely LGS

LGS is unlikely based on current data. Note that lower scores due to unknown EEG data or incomplete history do not definitively exclude LGS. Clinical judgment remains essential.

Worked example

A 28-year-old patient with drug-resistant epilepsy, developmental delay since childhood, and a history of drop attacks causing head injuries. EEG performed 10 years ago showed slow spike-and-wave discharges at 1.5 Hz. Onset of seizures at age 3.

1. Persistent seizures despite 2+ ASMs: Yes → +1 (Minor)

2. At least 2 seizure types: Yes → +1 (Minor)

3. Seizure onset before age 12: Yes (age 3) → +3 (Major)

4. Helmet use or head/facial injuries: Yes → +1 (Minor)

5. Cognitive impairment since childhood: Yes → +1 (Minor)

6. VNS / ketogenic diet / surgery: No → 0

7. EEG generalized slow SSW (<2.5 Hz): Yes → +3 (Major)

8. EEG other characteristic findings: No → 0

Total REST-LGS score = 1 + 1 + 3 + 1 + 1 + 0 + 3 + 0 = 10 points → Possibly LGS

Background: Lennox–Gastaut Syndrome

Lennox–Gastaut syndrome (LGS) is a severe childhood-onset epileptic encephalopathy characterized by the triad of:

Multiple seizure types, including tonic (especially during sleep), atonic (drop attacks), and atypical absence seizures
EEG abnormalities, classically generalized slow spike-and-wave discharges at <2.5 Hz (typically 1.5–2.5 Hz) during wakefulness and bursts of fast activity during sleep
Cognitive impairment and behavioral disturbances

LGS accounts for 1–4% of all childhood epilepsies, though it persists into adulthood in the majority of patients. The LGS Foundation estimates that diagnosis can take 12–15 years on average, with some patients not diagnosed until their early 30s. The REST-LGS was developed specifically to help identify adult patients who may have undiagnosed LGS or whose epilepsy was previously misclassified.

Why are major criteria weighted at 3 points?

The two major criteria (early seizure onset and generalized slow SSW on EEG) are highly specific for LGS. Generalized slow spike-and-wave activity at <2.5 Hz is the hallmark EEG pattern of LGS and, when present, substantially raises the probability of the diagnosis. Early-onset epilepsy before age 12 is a necessary historical feature of LGS as currently defined.

The remaining six minor criteria represent supporting features that are frequently present in LGS but are less specific when considered in isolation. They capture the refractory course, the functional consequences (helmet, cognitive impairment), the complexity of treatment escalation, and supportive EEG findings.

Limitations

The REST-LGS is a screening tool, not a diagnostic instrument. A formal LGS diagnosis requires clinical assessment by an epilepsy specialist, including EEG review.
Many adult patients with LGS lack recent EEG data. Missing EEG information will artificially lower the score; unknown data should not be interpreted as a negative finding.
The tool was developed for use in adult patients with known refractory epilepsy. Its performance in pediatric populations or in patients without prior epilepsy workup has not been fully validated.
LGS exists on a clinical spectrum. Some patients may have features of LGS without meeting all classic criteria, particularly if EEG findings have evolved over time.

Key references

Piña-Garza JE, Boyce D, Tworek DM, et al. REST-LGS: A practical screening tool to identify undiagnosed Lennox–Gastaut syndrome in patients with drug-resistant epilepsy. Epilepsy Behav. 2019;90:148–153. doi:10.1016/j.yebeh.2018.11.015
Wolf SM. Diagnosing Lennox–Gastaut syndrome in the adult patient. Pract Neurol. 2022;27–59.
Arzimanoglou A, French J, Blume WT, et al. Lennox–Gastaut syndrome: a consensus approach on diagnosis, assessment, management, and trial methodology. Lancet Neurol. 2009;8(1):82–93.
Camfield P, Camfield C. Lennox–Gastaut syndrome: a consensus approach to diagnosis, assessment, management, and trial methodology. Lancet Neurol. 2011;10(12):1056–1058.

Practice caveats

A low REST-LGS score in the absence of recent EEG data should prompt EEG recording before ruling out LGS. The generalized slow SSW criterion alone contributes 3 points.
Drop seizures (atonic or tonic) are hallmarks of LGS but their history may be elicited indirectly through the helmet use or facial/head injury criterion.
Patients with high REST-LGS scores who have not received LGS-specific therapies should be referred to an epilepsy center. Several FDA-approved and guideline-supported treatments exist specifically for LGS.
Always confirm diagnostic impressions with a comprehensive epilepsy specialist evaluation, including extended EEG monitoring if indicated.

Overview

The Refractory Epilepsy Screening Tool for Lennox-Gastaut Syndrome (REST-LGS) is a validated clinical questionnaire designed to screen patients with refractory epilepsy for undiagnosed or previously misclassified Lennox-Gastaut Syndrome (LGS). The instrument addresses a well-documented clinical gap: LGS is a severe, childhood-onset epileptic encephalopathy that is frequently underdiagnosed, misdiagnosed, or belatedly recognized, particularly in adult neurology and epilepsy practices where the diagnosis may not have been established during childhood and where the syndrome's characteristic features may have evolved or partially resolved over time.

By systematically querying clinicians or caregivers regarding key clinical, electroencephalographic, and historical features associated with LGS, the REST-LGS generates a composite score that identifies individuals warranting further diagnostic evaluation for this specific epilepsy syndrome. Early and accurate identification of LGS has direct therapeutic implications, as several pharmacological and non-pharmacological treatments are specifically licensed or indicated for LGS and may be underutilized in patients whose syndrome diagnosis has not been established.

Lennox-Gastaut Syndrome: Clinical Background

Lennox-Gastaut Syndrome is one of the most severe and pharmacologically refractory epileptic encephalopathies encountered in clinical practice. It is characterized by a classical triad that encompasses multiple seizure types with a particular emphasis on tonic and atonic seizures, a characteristic electroencephalographic (EEG) pattern of slow generalized spike-and-wave discharges at frequencies below 2.5 Hz, and intellectual disability or cognitive regression. Although LGS most commonly presents in early childhood, typically between the ages of 1 and 8 years, affected individuals carry this diagnosis throughout their lifetime, and the syndrome frequently transitions from pediatric to adult neurology care with its defining features partially modified by age, treatment history, and accumulated neurological injury.

The global prevalence of LGS is estimated at 1 to 2 per 100,000 individuals in the general population, and it accounts for approximately 1 to 4% of all childhood epilepsies and up to 10% of all refractory epilepsy in children. Despite its relative rarity, LGS disproportionately contributes to the burden of severe epilepsy, as affected individuals typically have very high seizure frequencies, multiple daily seizure types, frequent hospitalizations, significant injury risk from drop attacks, and profound limitations in quality of life and functional independence.

Etiological Spectrum

LGS may arise from a broad spectrum of structural, metabolic, genetic, and unknown etiologies. Approximately 70 to 75% of cases are considered symptomatic (now termed structural or metabolic in modern ILAE classification), arising from identifiable brain pathology, while 25 to 30% remain of unknown or presumed genetic etiology (formerly cryptogenic).

Common structural causes include hypoxic-ischemic encephalopathy, cortical malformations (focal cortical dysplasia, lissencephaly, pachygyria, polymicrogyria, tuberous sclerosis complex), post-infectious encephalitis, traumatic brain injury, and perinatal stroke. Tuberous sclerosis complex (TSC) deserves particular mention as a cause of LGS amenable to disease-specific therapy with mTOR inhibitors. Metabolic etiologies include glucose transporter type 1 (GLUT-1) deficiency syndrome and various mitochondrial disorders. Genetic etiologies including SCN1A, SCN8A, CDKL5, FOXG1, MEF2C, and copy number variants are increasingly identified through next-generation sequencing panels. A substantial proportion of cases previously considered cryptogenic are now being reclassified as genetic with expanded genomic testing.

LGS may evolve from other earlier-onset epileptic encephalopathies, most notably West syndrome (infantile spasms). Approximately 30% of children with West syndrome subsequently develop LGS, representing a well-characterized developmental trajectory of epileptic encephalopathy. The evolution from West syndrome to LGS carries particularly poor prognostic implications regarding intellectual outcome and seizure refractoriness.

Seizure Types in LGS

The hallmark seizure types of LGS are tonic seizures and atonic seizures (drop attacks), which together account for much of the injury burden in this syndrome. However, affected patients typically experience multiple co-existing seizure types, and the specific combination varies among individuals and evolves over time.

Tonic seizures: Sudden stiffening of axial or appendicular musculature, typically lasting 5 to 20 seconds, often occurring during sleep or drowsiness. Tonic seizures may be subtle (brief neck extension, eye opening, upward eye deviation) or involve dramatic generalized muscle contraction with vocalization. When occurring while the patient is upright, tonic seizures cause characteristic forward falls with direct facial or head impact, resulting in frequent facial injuries. Tonic seizures during sleep may produce sudden awakening, head drops, or brief respiratory irregularity without full consciousness impairment.
Atonic seizures (drop attacks): Sudden, brief (1 to 3 seconds) loss of postural tone causing abrupt falls without warning. Patients cannot protect themselves during the fall due to the brevity of the event, resulting in high rates of head trauma, facial injury, dental fractures, and laceration injuries. Drop attacks (encompassing both atonic and the drop component of tonic seizures) are among the most dangerous and functionally disabling seizure types and are a core feature driving disability and injury in LGS.
Atypical absence seizures: Periods of impaired consciousness with behavioral arrest, often associated with subtle motor features such as eye fluttering, automatisms, or changes in postural tone. Unlike typical absence seizures in childhood absence epilepsy, atypical absences in LGS have a gradual onset and offset, may be prolonged, and occur against a background of generalized EEG slowing. Absence status epilepticus (non-convulsive status epilepticus) is common in LGS and may manifest as prolonged periods of behavioral slowing, confusion, drooling, or reduced responsiveness lasting hours to days.
Myoclonic seizures: Brief, shock-like muscle jerks affecting the limbs or trunk. Myoclonic seizures occur in a subset of LGS patients and may be difficult to distinguish clinically from myoclonic-astatic epilepsy (Doose syndrome), which is an important differential diagnosis.
Focal seizures with secondary generalization: Particularly common in LGS arising from focal structural pathology. These may present as asymmetric tonic or versive movements with subsequent secondary generalization.
Generalized tonic-clonic seizures: Present in many LGS patients, particularly during acute illness or sleep deprivation, though typically not the predominant seizure type.

EEG Characteristics

The electroencephalographic hallmarks of LGS are essential for syndrome diagnosis and are specifically queried within the REST-LGS framework. The characteristic EEG findings include:

Slow generalized spike-and-wave discharges (GSWD) at less than 2.5 Hz: This is the defining interictal EEG feature of LGS, occurring either continuously or in runs, most often at frequencies of 1.5 to 2.5 Hz. The slow frequency distinguishes LGS from childhood absence epilepsy (3 Hz) and juvenile absence epilepsy (3 to 4 Hz). These discharges are maximally expressed during slow-wave sleep.
Generalized paroxysmal fast activity (GPFA) during sleep: Bursts of generalized fast rhythmic activity at 10 to 20 Hz occurring during NREM sleep, often time-locked to clinical tonic seizures. GPFA is considered highly specific for LGS when combined with the clinical and cognitive features of the syndrome.
Diffuse background slowing: The interictal EEG background in LGS is typically diffusely slow, reflecting the underlying encephalopathic state. High-amplitude irregular delta activity is common, often admixed with the spike-and-wave discharges.
Multifocal and diffuse sharp waves: In addition to the generalized patterns, focal and multifocal sharp waves are common, particularly when LGS arises from structural etiology.

It is important to recognize that EEG findings in LGS evolve over time and may be less typical in older patients or those who have undergone extensive anticonvulsant treatment. The classic slow GSWD may be less prominent in adult LGS patients, contributing to delayed or missed diagnoses in adult epilepsy clinics.

Cognitive and Neurodevelopmental Profile

Cognitive impairment is a defining feature of LGS and is present in the vast majority of affected individuals. The degree of intellectual disability ranges from mild to profound and is influenced by the underlying etiology, the age of onset, the seizure burden sustained over time, and treatment history. Intellectual disability may be present prior to seizure onset in symptomatic LGS arising from early structural injury, or may develop progressively following epilepsy onset in initially normal or near-normal children.

Behavioral comorbidities are nearly universal in LGS and include attention deficit hyperactivity disorder (ADHD), autism spectrum disorder features, aggressive or self-injurious behavior, anxiety, and mood disorders. These behavioral and psychiatric comorbidities contribute substantially to caregiver burden and often become the primary management challenge in older patients and adults in whom seizure frequency may have partially reduced.

The cognitive and behavioral profile of LGS is an important component of the REST-LGS screening framework, as the combination of intellectual disability or regression with multiple seizure types and an early epilepsy onset is a highly discriminating clinical constellation that distinguishes LGS from other refractory epilepsy syndromes.

Rationale for the REST-LGS Screening Tool

Despite LGS being a well-characterized epileptic encephalopathy with established diagnostic criteria, epidemiological and clinical evidence consistently documents significant diagnostic delays and high rates of misdiagnosis. Several factors contribute to this diagnostic gap:

Diagnostic Complexity

LGS requires the simultaneous presence of multiple diagnostic elements (seizure types, EEG characteristics, cognitive profile, and onset criteria) for definitive diagnosis. In clinical practice, complete diagnostic data may not be available at a single clinical encounter, and the clinician must synthesize information from different domains, including historical accounts, prior EEG reports, neuropsychological testing, and neuroimaging. When any one of these elements is missing, atypical, or inconsistent with the classic description, diagnostic confidence may be reduced and alternative labels may be applied.

Evolution of the Syndrome Over Time

The clinical and EEG features of LGS may evolve substantially over the patient's lifetime. Some seizure types may become less frequent or disappear entirely in adulthood, EEG abnormalities may become less typical with advancing age and treatment history, and the defining slow spike-and-wave pattern may not be captured on routine interictal EEGs if the recording duration is insufficient or the patient is on medications that suppress this pattern. As a result, adult neurologists encountering a patient with a complex refractory epilepsy history from childhood may not recognize the underlying LGS diagnosis from the current clinical picture alone.

Lack of Familiarity in Adult Epilepsy Practice

LGS is primarily framed as a childhood epilepsy syndrome in neurology education and training, and adult neurologists may have limited direct exposure to the syndrome. When LGS patients transition from pediatric to adult care, the diagnosis may be de-emphasized or replaced by non-specific labels such as "symptomatic generalized epilepsy," "epileptic encephalopathy," or simply "refractory epilepsy." This re-labeling has direct consequences for treatment selection, as LGS-specific therapies may be underutilized when the syndrome diagnosis is not explicitly recognized.

Therapeutic Implications of Correct Diagnosis

The clinical urgency of accurate LGS identification derives directly from its treatment implications. Multiple therapies have received regulatory approval or carry substantial evidence specifically for LGS, including cannabidiol (CBD), clobazam (in conjunction with other LGS therapies), rufinamide, lamotrigine (with caveats), felbamate, and the ketogenic diet. Corpus callosotomy is an evidence-based surgical option specifically for drop attack reduction in LGS that is rarely considered when the syndrome has not been diagnosed. Vagal nerve stimulation has demonstrated particular efficacy in LGS populations. The systematic underutilization of these options in patients whose LGS diagnosis has not been established is a meaningful source of preventable morbidity, and the REST-LGS directly addresses this gap by enabling systematic identification of patients who should be re-evaluated for this diagnosis.

REST-LGS Development and Validation

The REST-LGS was developed through a structured process integrating expert consensus on LGS diagnostic criteria with empirical validation in patient cohorts with confirmed LGS diagnoses and comparator refractory epilepsy populations without LGS. The instrument was designed to be completed by treating clinicians or caregivers as a structured questionnaire rather than requiring comprehensive neurological examination or contemporaneous EEG, enabling its use in routine clinical settings including adult epilepsy outpatient clinics, epilepsy monitoring units, and transition clinics serving adolescents moving from pediatric to adult care.

The item pool for the REST-LGS was derived from the established diagnostic criteria for LGS defined by the International League Against Epilepsy (ILAE) and supplemented by expert opinion regarding the most clinically discriminating features distinguishing LGS from other forms of refractory epilepsy. Candidate items were evaluated for inter-rater reliability, sensitivity, and specificity in distinguishing confirmed LGS cases from non-LGS refractory epilepsy comparators. Items demonstrating the strongest discriminative performance and the greatest clinical feasibility were retained in the final instrument.

Validation studies demonstrated that the REST-LGS achieves high sensitivity and acceptable specificity for LGS identification using a clinical scoring threshold, enabling prioritization of patients for comprehensive LGS diagnostic evaluation including prolonged video-EEG monitoring, neuropsychological assessment, and specialist neurological review. The tool is explicitly designed as a screening instrument rather than a diagnostic instrument; a positive REST-LGS score indicates the need for further evaluation rather than confirming the diagnosis.

REST-LGS Components and Scoring

The REST-LGS systematically assesses the presence and character of the key clinical features that define and distinguish LGS. The instrument integrates seizure type history, historical EEG data, cognitive and developmental profile, epilepsy onset characteristics, and treatment course into a structured scoring framework.

Domain 1: Seizure Type History

The seizure type domain assesses whether the patient has experienced the specific seizure types most characteristic of LGS, with particular emphasis on the pathognomonic seizure types associated with the highest injury risk and diagnostic specificity.

Tonic Seizures

The presence of tonic seizures, either currently or historically, is a core diagnostic criterion for LGS and a high-weight item in the REST-LGS. The clinician or caregiver is asked whether the patient has experienced episodes of sudden muscle stiffening, with or without impaired consciousness, that are distinct from generalized tonic-clonic convulsions. Nocturnal tonic seizures are specifically highlighted given their high frequency in LGS and their characteristic occurrence during NREM sleep. Scoring acknowledges both current and historical tonic seizures, recognizing that some patients may have had seizure type evolution with reduced current tonic seizure frequency compared with the peak in childhood.

Drop Attacks (Atonic and Tonic-Drop Seizures)

Drop attacks represent one of the most clinically distinctive and functionally devastating features of LGS and carry the highest discriminatory weight in the REST-LGS. The instrument queries both current and historical drop attacks, with specific attention to:

Sudden falls without warning (consistent with atonic mechanism)
Falls associated with brief stiffening followed by collapse (tonic-drop mechanism)
History of facial injuries, dental injuries, or head trauma attributable to falls during seizures
Use of protective helmets or padded clothing as a consequence of drop attack risk
History of hospitalization or emergency department visits for seizure-related fall injuries

The presence of drop attacks, particularly when associated with a history of fall-related injuries and protective equipment use, constitutes one of the strongest individual predictors of LGS among refractory epilepsy patients and is accordingly a high-yield item in the screening instrument.

Atypical Absence Seizures

Atypical absences are queried with specific emphasis on their distinguishing features relative to typical childhood absences: gradual onset and termination, incomplete rather than complete loss of awareness, associated hypotonia or myoclonic components, and prolonged duration. The history of absence status epilepticus (prolonged periods of confusion, behavioral slowing, or reduced responsiveness lasting more than 30 minutes) is specifically highlighted as a feature strongly associated with LGS.

Multiple Co-existing Seizure Types

The presence of three or more distinct seizure types at any point in the patient's epilepsy course is a strongly discriminating feature of LGS relative to other refractory epilepsy syndromes, which more commonly present with one or two dominant seizure types. The REST-LGS documents the total number of distinct seizure types experienced, with higher counts contributing proportionally to the composite score.

Domain 2: EEG History

The EEG domain queries the presence of characteristic LGS electroencephalographic features on historical EEG reports. Because comprehensive review of raw EEG data is not feasible in routine clinical screening, the REST-LGS is designed to be completed using EEG report descriptions available in the medical record.

Slow Generalized Spike-and-Wave Discharges (less than 2.5 Hz)

The presence of slow (less than 2.5 Hz) generalized spike-and-wave discharges on any prior EEG is a core diagnostic criterion for LGS and is specifically queried in the REST-LGS. Clinicians are asked whether any prior EEG report describes this pattern, recognizing that older reports may use terminology such as "slow spike-wave," "atypical spike-wave," or "2 Hz spike-and-wave" to describe this finding. This item acknowledges that the pattern may not be present on every EEG recording and that its absence on a single recent EEG does not exclude LGS, particularly in adult patients in whom the pattern may have attenuated.

Generalized Paroxysmal Fast Activity (GPFA) During Sleep

The presence of generalized paroxysmal fast activity (10 to 20 Hz) during sleep on any prior EEG report is queried as a highly specific marker for LGS. Clinicians are asked whether any prior sleep EEG or overnight monitoring study has reported this pattern, which may be described in reports as "generalized fast rhythmic activity during sleep," "sleep-activated fast activity," or "paroxysmal fast discharge during NREM sleep."

Generalized Background Slowing

Diffuse slowing of the background EEG rhythm, reflecting the encephalopathic nature of LGS, is queried as a supporting feature. While non-specific on its own, diffuse background slowing in conjunction with the other EEG and clinical features contributes incrementally to the composite REST-LGS score.

Domain 3: Cognitive and Developmental Profile

The cognitive domain assesses the presence and degree of intellectual disability and developmental impairment, as well as the trajectory of cognitive development in relation to epilepsy onset.

Intellectual Disability

The presence of formally diagnosed or clinically evident intellectual disability (IQ below 70 or significant adaptive function impairment) is recorded. The degree of intellectual disability is graded from mild to profound, with moderate to severe or profound intellectual disability carrying higher discriminative weight for LGS relative to milder forms shared with other epilepsy syndromes.

Cognitive Regression or Arrest

A history of cognitive regression (loss of previously acquired skills or cognitive abilities) or developmental arrest (failure to acquire expected milestones following epilepsy onset) is a particularly important discriminating feature distinguishing LGS from static intellectual disability of non-epileptic etiology. The REST-LGS asks specifically whether caregivers or records document a period of skill regression or developmental plateau temporally associated with seizure onset or a significant increase in seizure frequency.

Language and Communication Impairment

Significant language delay, language regression, or current non-verbal or minimally verbal status is documented. Given the high rates of severe language impairment in LGS, this item contributes to the overall severity assessment and helps differentiate LGS from milder epilepsy syndromes with preserved language.

Adaptive Function and Independence

The patient's current level of adaptive functioning and independence in activities of daily living is assessed. Most adult LGS patients require significant caregiver support for basic daily activities, reflecting the combined impact of intellectual disability, seizure frequency, and the cumulative neurological burden of the syndrome.

Domain 4: Epilepsy Onset and Clinical Course

The epilepsy onset domain captures the age at first seizure, the presence of a prior epilepsy syndrome diagnosis (particularly West syndrome/infantile spasms), and the overall treatment course.

Age of Epilepsy Onset

LGS characteristically begins in early childhood, most commonly between ages 1 and 8 years, though onset in late infancy or early adolescence is recognized. An onset prior to age 8 years is a supporting criterion, with onset prior to age 5 years carrying higher weight given its stronger association with LGS relative to other generalized epilepsy syndromes that may begin in middle childhood or adolescence. Late-onset LGS (after age 10) is recognized but unusual and should prompt careful review of whether the diagnostic criteria are fully met.

Prior West Syndrome (Infantile Spasms) Diagnosis

A confirmed prior diagnosis of West syndrome or infantile spasms before age 2 years is a high-weight item in the REST-LGS, as approximately 30% of West syndrome cases evolve into LGS and this historical progression is highly characteristic. The clinician is asked whether infantile spasms were documented in medical records, prior neurology notes, or reliably reported by caregivers. The presence of hypsarrhythmia on EEG in infancy is additionally recorded as a corroborating historical EEG feature.

Epilepsy Refractoriness

By definition, REST-LGS is applied in the context of refractory epilepsy, defined as failure of two or more appropriately chosen and adequately trialed antiseizure medications (ASMs) to achieve sustained seizure freedom. The instrument documents the number of prior ASM trials, the nature of therapeutic responses, and the current treatment regimen. LGS characteristically requires polytherapy and is associated with high rates of ASM failure; a history of three or more failed ASM trials strongly supports a refractory encephalopathic epilepsy consistent with LGS.

Use of LGS-Associated Treatments

The prior use or current use of treatments specifically associated with LGS management provides indirect diagnostic evidence. These include clobazam (at doses and formulations used specifically for LGS adjunctive therapy), rufinamide, felbamate, lamotrigine (where tolerated), cannabidiol (specifically the pharmaceutical-grade formulation licensed for LGS), ketogenic diet, vagal nerve stimulation, and corpus callosotomy. Having received one or more of these LGS-associated treatments suggests that a prior clinician had at least considered LGS as the working diagnosis, adding probability weight to the screening assessment.

Domain 5: Protective Behaviors and Injury History

This domain captures behavioral adaptations and injury history attributable to drop attacks, providing indirect but clinically meaningful evidence of the severity and chronicity of LGS-typical seizure types.

Helmet use: Current or historical use of a protective helmet specifically prescribed for seizure-related fall protection.
Padded clothing or protective equipment: Use of knee pads, padded jackets, or other protective clothing to mitigate drop attack injury.
Home environmental modifications: Padding of furniture corners, removal of hard flooring, stair gates, or other home safety modifications made specifically for seizure fall protection.
History of seizure-related facial, dental, or head injuries: Documented emergency department visits, fractures, lacerations, or dental injuries directly attributable to falls during seizures.
Supervision requirements: Continuous caregiver supervision required at home or in care settings specifically to prevent fall injury during seizures.

Scoring and Interpretation

Each REST-LGS domain item is assigned a weighted point value based on its demonstrated discriminatory contribution to LGS identification in the validation cohort. Items with higher specificity for LGS (drop attacks, slow spike-and-wave EEG, nocturnal tonic seizures, prior West syndrome diagnosis) carry greater weight than more broadly occurring features (generalized background slowing, single seizure type refractoriness).

Domain	Feature	Points
Seizure Types	Drop attacks (atonic or tonic-drop), current or historical	3
	Tonic seizures, current or historical	2
	Atypical absence seizures or history of absence status epilepticus	2
	Three or more distinct seizure types at any point in epilepsy course	2
	Myoclonic seizures as part of multi-seizure-type pattern	1
EEG Findings	Slow generalized spike-and-wave (< 2.5 Hz) on any prior EEG	3
	Generalized paroxysmal fast activity during sleep on any prior EEG	2
	Diffuse background slowing on any prior EEG	1
Cognitive Profile	Moderate, severe, or profound intellectual disability	2
	Mild intellectual disability or developmental delay	1
	History of cognitive regression or developmental arrest after epilepsy onset	2
	Significant language impairment or non-verbal status	1
Epilepsy Onset and Course	Prior confirmed West syndrome / infantile spasms diagnosis	3
	Seizure onset before age 8 years	2
	Three or more failed antiseizure medication trials	1
	Prior use of LGS-associated therapies (rufinamide, clobazam for LGS, cannabidiol, corpus callosotomy, ketogenic diet)	1
Protective Behaviors and Injury History	Helmet or protective equipment use for seizure fall prevention	2
	History of seizure-related fall injuries (ED visits, fractures, lacerations)	1

REST-LGS Total Score	Interpretation	Recommended Action
0 to 4	LGS unlikely	LGS does not require prioritized further investigation; evaluate for other refractory epilepsy etiologies appropriate to the clinical presentation
5 to 8	LGS possible; further evaluation recommended	Review prior EEG records and obtain video-EEG monitoring if not previously performed; detailed neurology consultation with LGS expertise; neuropsychological assessment
9 and above	LGS highly probable; comprehensive evaluation strongly indicated	Prioritized referral to epilepsy specialist with LGS expertise; prolonged video-EEG monitoring; comprehensive etiological workup including genetics if not previously completed; reassessment of treatment plan for LGS-specific therapeutic options

Clinical Application and Workflow Integration

The REST-LGS is designed to be administered in routine clinical settings without specialized equipment and is intended for use by epilepsy nurses, advanced practice providers, neurology residents, and general neurologists as well as epilepsy specialists. The instrument should be completed at the time of clinical encounter using a combination of caregiver interview and medical record review, with access to prior EEG reports being particularly important for accurate scoring of the EEG domain.

Priority Patient Populations for REST-LGS Screening

While the REST-LGS may be applied broadly in any patient with refractory epilepsy, the following patient populations represent the highest-yield groups for systematic screening:

Patients transitioning from pediatric to adult epilepsy care: The transition from pediatric to adult neurology is a well-recognized point of diagnostic discontinuity where LGS diagnoses may be lost, de-emphasized, or unknown to the receiving adult neurologist. Systematic REST-LGS screening at transition can identify patients requiring LGS-focused re-evaluation before they are assigned non-specific labels in the adult care setting.
Adults with early-onset refractory epilepsy and intellectual disability: Adult epilepsy clinic patients with a history of childhood-onset refractory epilepsy, intellectual disability, and multiple seizure types represent the highest-probability population for undiagnosed LGS. Many of these patients were evaluated in an era before modern diagnostic criteria and may carry outdated diagnostic labels that do not reflect current understanding of epilepsy syndromes.
Patients with drop attacks of unclear etiology: Atonic or tonic-drop seizures are a highly specific feature of LGS and should prompt systematic REST-LGS screening in any patient in whom this seizure type is documented, even if a prior comprehensive evaluation did not explicitly diagnose LGS.
Patients with a history of infantile spasms: Given the approximately 30% rate of West syndrome-to-LGS evolution, all patients with a documented history of infantile spasms who continue to have refractory epilepsy should be systematically screened with the REST-LGS, as LGS may have developed and not been formally recognized after the infantile spasms era of their care.
Patients being evaluated for epilepsy surgery: Pre-surgical epilepsy evaluation is an opportunity to revisit syndrome classification. Accurate LGS diagnosis in this context affects surgical candidacy assessments, as corpus callosotomy may be the most appropriate surgical intervention for drop attack reduction rather than focal resective surgery.

Integration with the Diagnostic Workflow

A positive REST-LGS screen (score 5 or above) should trigger a structured diagnostic re-evaluation pathway:

Medical records review: Systematic retrieval and review of all prior EEG reports, neurology consultation notes, neuroimaging, neuropsychological testing, and genetic testing results to reconstruct the patient's diagnostic history and identify prior LGS-relevant findings that may not have been formally synthesized.
Comprehensive video-EEG monitoring: If prior EEG documentation is insufficient to confirm or exclude the characteristic LGS EEG patterns, prolonged video-EEG monitoring (inpatient or ambulatory) is indicated to capture ictal and interictal EEG features during both wakefulness and sleep, including NREM sleep where tonic seizures and GPFA are most likely to be recorded.
Neuropsychological evaluation: Formal neuropsychological assessment provides standardized quantification of intellectual function, adaptive behavior, and specific cognitive domain performance, enabling accurate classification of intellectual disability severity and documenting cognitive trajectory over time.
Etiological workup: If not previously performed comprehensively, etiological evaluation should include high-resolution MRI of the brain with epilepsy protocol sequences, comprehensive epilepsy gene panel or exome sequencing, and metabolic screening appropriate to the clinical context. Identification of a specific etiology may have direct therapeutic implications (mTOR inhibitors for TSC, ketogenic diet for GLUT-1 deficiency) beyond LGS syndrome management.
Specialist epilepsy consultation: Referral to a neurologist or epileptologist with specific expertise in epileptic encephalopathies and LGS is strongly recommended for patients with high REST-LGS scores, both for diagnostic confirmation and for comprehensive therapeutic reassessment.

Differential Diagnosis: LGS versus Other Refractory Epilepsy Syndromes

The REST-LGS is designed to screen broadly, and clinicians should be aware of conditions that share features with LGS and may yield elevated REST-LGS scores without meeting full LGS diagnostic criteria.

Myoclonic-Astatic Epilepsy (Doose Syndrome)

Myoclonic-astatic epilepsy (MAE) shares several features with LGS including drop attacks (astatic seizures), myoclonic seizures, and early childhood onset with cognitive impact. Key distinguishing features include the predominance of myoclonic-astatic (combined myoclonic and atonic) seizures rather than pure atonic or tonic seizures, a more favorable prognosis with a subset achieving remission, a normal cognitive baseline at epilepsy onset with regression rather than pre-existing intellectual disability, and the characteristic EEG pattern of 2 to 3 Hz generalized spike-and-wave that may overlap with LGS but typically at a slightly faster frequency. Tonic seizures, particularly nocturnal tonic seizures, are uncommon in MAE and their presence strongly favors LGS.

Dravet Syndrome

Dravet syndrome is a severe epileptic encephalopathy caused by SCN1A mutations in the majority of cases, characterized by fever-sensitive prolonged seizures beginning in the first year of life, subsequent development of multiple seizure types, cognitive regression, and refractory epilepsy. Distinguishing features include the predominance of febrile or fever-triggered seizures at onset, the characteristic sensitivity to sodium channel-blocking antiseizure medications (which typically worsen Dravet syndrome), the prominence of myoclonic seizures rather than tonic seizures, and the SCN1A genetic findings. EEG in Dravet syndrome may show generalized polyspike-wave but typically lacks the characteristic slow (less than 2.5 Hz) GSWD that defines LGS. Corpus callosotomy has limited evidence in Dravet syndrome compared with LGS.

Angelman Syndrome

Angelman syndrome (AS) is a genetic disorder caused by loss of function of the maternal UBE3A allele, characterized by severe intellectual disability, absent or minimal speech, a distinctive behavioral phenotype (happy demeanor, frequent laughing), gait ataxia, and epilepsy that may include features overlapping with LGS. The EEG in AS may show high-amplitude delta activity with spike-and-wave components, and seizure types include atonic and myoclonic events. The distinctive behavioral and syndromic phenotype of AS, its characteristic EEG pattern (frontally predominant triphasic delta activity), and FISH or methylation testing for the AS locus enable differentiation.

CDKL5 Deficiency Disorder

CDKL5 deficiency disorder (CDD) is an X-linked epileptic encephalopathy presenting with early-onset refractory seizures (infantile spasms in many cases), intellectual disability, and limited hand use. The seizure phenotype may include tonic and hypermotor seizures as well as myoclonic components. CDD is distinguished by its genetic basis (CDKL5 mutation), its very early seizure onset (typically within the first 6 months of life), the prominence of hypermotor and tonic-clonic seizures rather than the classic atonic drop attacks of LGS, and the distinctive hand stereotypies reminiscent of Rett syndrome. Molecular genetic testing identifies CDKL5 mutations.

Tuberous Sclerosis Complex with Epileptic Encephalopathy

TSC-associated epileptic encephalopathy may produce a clinical picture overlapping with LGS, and some TSC patients meet full diagnostic criteria for LGS. Critically, accurate identification of TSC as the underlying etiology enables initiation of mTOR inhibitor therapy (everolimus or sirolimus) which has demonstrated efficacy for seizure reduction specifically in TSC-associated epilepsy and may modify the LGS phenotype. The TSC diagnostic criteria (cortical tubers on MRI, subependymal nodules, subependymal giant cell astrocytomas, skin lesions, renal angiomyolipomas) and TSC1/TSC2 genetic testing enable diagnosis.

Therapeutic Implications of LGS Diagnosis Established through REST-LGS Screening

The primary clinical justification for systematic REST-LGS screening is the direct therapeutic impact of establishing the LGS diagnosis. Patients correctly identified as having LGS become candidates for a range of syndrome-specific interventions that may not have been previously considered or offered.

Cannabidiol (CBD)

Pharmaceutical-grade cannabidiol (Epidiolex in the United States; Epidyolex in Europe) has received regulatory approval for adjunctive treatment of seizures associated with LGS in patients aged 2 years and above based on evidence from randomized, placebo-controlled trials demonstrating significant reductions in drop seizure frequency. In the pivotal GWPCARE3 and GWPCARE4 trials, cannabidiol at doses of 10 and 20 mg/kg/day produced median reductions in drop seizure frequency of 37 to 41% compared with 17 to 20% in the placebo group. This represents the first randomized controlled evidence for a cannabinoid-derived medication in LGS and established it as a meaningful adjunctive option for patients with inadequately controlled drop seizures.

Rufinamide

Rufinamide is specifically approved for adjunctive treatment of seizures associated with LGS in patients aged 1 year and above (in the United States) or 4 years and above (in Europe). The pivotal trial demonstrated significant reductions in total seizure frequency (32% reduction versus 12% placebo) and tonic-atonic seizure frequency (43% reduction versus 1.4% increase in placebo) in LGS patients. Rufinamide is considered a first-line LGS-specific adjunctive agent and should be specifically evaluated for trial in patients who are correctly identified as having LGS.

Clobazam

Clobazam at doses of 0.1 to 0.2 mg/kg/day (up to 40 mg/day in adults) is indicated as adjunctive therapy for seizures associated with LGS in patients aged 2 years and above (United States approval). The OV-1012 trial demonstrated significant seizure frequency reductions in LGS patients, particularly for drop seizures. Tolerance development is a recognized limitation of long-term clobazam use, but clinically meaningful seizure reduction is maintained in many patients over extended periods. Clobazam is commonly prescribed for generalized epilepsy syndromes generally but is specifically indicated and studied in LGS.

Corpus Callosotomy

Surgical section of the corpus callosum, particularly anterior two-thirds callosotomy, is a well-established palliative surgical intervention for LGS specifically aimed at reducing the frequency and severity of drop attacks. By interrupting the interhemispheric propagation of epileptiform discharges that generate the bilateral synchronous patterns underlying tonic and atonic seizures, callosotomy produces substantial drop attack reduction in approximately 60 to 80% of LGS patients, with complete drop attack cessation in 40 to 50%. Callosotomy is not a curative procedure and does not address other seizure types, but its specific efficacy for drop attacks makes it one of the most impactful interventions for injury prevention in LGS. This surgery is rarely considered in patients whose LGS diagnosis has not been explicitly established, underscoring the clinical importance of the REST-LGS screening process.

Vagal Nerve Stimulation

Vagal nerve stimulation (VNS) is a neuromodulatory surgical intervention demonstrating particular efficacy in LGS populations, with published series reporting 50% or greater seizure reduction rates in 40 to 60% of LGS patients. VNS is especially valuable in LGS patients who are not resective surgical candidates and who have failed multiple pharmacological options. On-demand VNS stimulation (magnet-triggered) may be effective for seizure termination in individual episodes. Modern high-output VNS devices with more frequent stimulation cycles and responsive stimulation modes (RNS) offer additional options that are increasingly being studied in LGS populations.

Ketogenic Diet Therapy

The ketogenic diet and its variants (modified Atkins diet, low glycemic index treatment) have demonstrated efficacy in LGS with 50% or greater seizure reduction reported in approximately 50 to 60% of treated patients and sustained efficacy over multi-year follow-up in responders. The ketogenic diet is particularly applicable in patients intolerant of or inadequately responsive to multiple pharmacological options, and its implementation is feasible across the age spectrum from young children through adults. Caregiver commitment and nutritional monitoring are essential prerequisites for successful ketogenic diet implementation.

Limitations of the REST-LGS

Understanding the limitations of the REST-LGS is essential for appropriate clinical application of the tool.

Screening Rather than Diagnostic Function

The REST-LGS is explicitly a screening tool and does not establish the diagnosis of LGS. A positive screen identifies patients requiring further evaluation; only comprehensive assessment including video-EEG monitoring, cognitive evaluation, and specialist review can confirm the LGS diagnosis according to established ILAE criteria. Treating a positive REST-LGS screen as equivalent to a LGS diagnosis would inappropriately expose patients to LGS-specific interventions without proper diagnostic confirmation and would risk therapeutic misdirection if an alternative diagnosis is ultimately established.

Dependence on Historical Data Quality

Accurate REST-LGS scoring requires access to historical medical records, particularly prior EEG reports and prior diagnostic documentation. In patients who have received care at multiple institutions, have limited medical record availability, or who lack reliable caregiver informants, historical data gaps may result in scoring inaccuracies. Insufficient historical data is likely to result in under-scoring rather than over-scoring, potentially causing false-negative screens in patients with genuine LGS whose historical features are not documentable.

EEG Feature Attenuation in Older Patients

The characteristic LGS EEG features, particularly slow generalized spike-and-wave discharges, may attenuate significantly in older patients or in those on extensive pharmacological therapy. This reduces the sensitivity of the EEG domain for LGS identification in adult patients and means that a low EEG domain score should not be used to definitively exclude LGS when other clinical features are strongly suggestive. In these cases, prolonged video-EEG monitoring with sleep recording may be necessary to capture sufficient EEG data for accurate scoring.

Potential for Over-Screening in Overlapping Syndromes

The clinical features of LGS overlap substantially with other severe epileptic encephalopathies, including Dravet syndrome, myoclonic-astatic epilepsy, and CDKL5 deficiency disorder. In clinical settings with high prevalence of these overlapping conditions, the REST-LGS may generate false-positive screens at a higher rate than in settings where LGS represents a larger proportion of the refractory epilepsy population. Careful application of the full diagnostic evaluation pathway following a positive screen ensures that false positives are identified and correctly re-classified.

Not Validated for Primary Diagnostic Application

The REST-LGS has been validated as a screening instrument applied in populations of patients with known refractory epilepsy, not as a primary diagnostic tool in unselected general neurology populations or in populations without established epilepsy. Its application in settings substantially different from the validation cohort (for example, in newly presenting childhood epilepsy without established refractoriness) may yield performance characteristics different from those reported in the original validation studies.

Key Clinical Takeaways

The REST-LGS is a structured clinical screening questionnaire designed to identify patients with refractory epilepsy who may have undiagnosed or misclassified Lennox-Gastaut Syndrome, enabling timely access to LGS-specific diagnostic evaluation and treatment.
LGS is frequently underdiagnosed in adult epilepsy populations due to syndrome evolution, diagnostic label attrition at care transitions, and limited adult neurologist familiarity with childhood epileptic encephalopathies.
The five REST-LGS domains assess seizure type history (with highest weight for drop attacks and tonic seizures), EEG historical findings (with highest weight for slow GSWD and GPFA during sleep), cognitive profile (intellectual disability and regression), epilepsy onset and course (prior West syndrome and early onset), and injury and protective behavior history.
A REST-LGS score of 9 or above is highly suggestive of LGS and warrants prioritized referral for comprehensive epilepsy specialist evaluation including prolonged video-EEG monitoring with sleep recording.
A positive REST-LGS screen is not a LGS diagnosis; definitive diagnosis requires confirmation by a neurologist or epileptologist with expertise in epileptic encephalopathies using full ILAE criteria.
Correct identification of LGS through the REST-LGS pathway directly enables access to regulatory-approved LGS-specific therapies including cannabidiol, rufinamide, and clobazam, as well as evidence-based surgical options including corpus callosotomy and vagal nerve stimulation.
Priority screening populations include patients transitioning from pediatric to adult epilepsy care, adults with early-onset refractory epilepsy and intellectual disability, patients with documented drop attacks, and patients with a history of infantile spasms.
Clinicians should be aware of overlapping epileptic encephalopathies including Dravet syndrome, myoclonic-astatic epilepsy, Angelman syndrome, and CDKL5 deficiency disorder, which may yield elevated REST-LGS scores without meeting full LGS diagnostic criteria and require etiological investigation for accurate syndrome classification.