Characterizing SCN1A-Related Disorders Using Real-World Data Across 681 Patient-Years
1Division of Neurology, Children’s Hospital of Philadelphia, Philadelphia, PA, USA, 19104
2The Epilepsy Neurogenetics Initiative (ENGIN), Children’s Hospital of Philadelphia, Philadelphia, PA, USA, 19104
3Department of Biomedical and Health Informatics, Children’s Hospital of Philadelphia, Philadelphia, PA, USA, 19104
4Center for Epilepsy and Neurodevelopmental Disorders (ENDD), Children’s Hospital of Philadelphia, Philadelphia, PA, USA, 19104
5Dravet Syndrome Foundation, Cherry Hill, NJ, 08034
6Department of Neurology, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA, 19104
*Correspondence to: Ingo Helbig, MD Children’s Hospital of Philadelphia, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, 19104 helbigi@chop.eduAbstract
SCN1A-related disorders are the single most common monogenic cause of epilepsy and represent a major focus of precision medicine efforts. In conjunction with existing prospective studies, the analysis of real-world data obtained during routine clinical care can expand upon the scale and duration of available data and contribute to the development of meaningful outcomes for clinical trials.
Here, we leveraged real-world data to delineate the longitudinal disease history of 100 individuals with SCN1A-related disorders using a systematic approach. We mapped a total of 671 unique clinical terms to a standardized framework in monthly increments across 681 patient-years, including 75 terms related to seizure types. Within this cohort, 89 individuals had presumed loss-of-function variants in SCN1A based on variant type and clinical diagnosis, including those with Dravet syndrome (N = 79) and genetic epilepsy with febrile seizures plus (N = 10). Ten individuals had a non-Dravet developmental and epileptic encephalopathy caused by gain-of-function variants in SCN1A.
By annotating seizure type and frequency in monthly time-bins, we assessed seizure burden. A median of 17 changes in seizure frequency and ten terms referring to seizure type were identified per participant. Myoclonic seizures occurred with high frequency (median >5 daily), whereas hemiclonic, focal impaired consciousness, and bilateral tonic-clonic seizures occurred more rarely (median monthly). Retrospective analysis of developmental histories showed a range of cognitive abilities. Neurodevelopmental differences were observed in 83% (83/100) of individuals, of whom 83% (69/83) demonstrated delayed language skills. Motor coordination impairments, including gait disturbance, ataxia, hypotonia, and imbalance were annotated in 69% (69/100) of participants. EEG findings varied with age; most were reported as normal before nine months of age, after which the prevalence of abnormal interictal findings increased. Individuals with different clinical syndromes had unique medication landscapes, with 554 prescriptions of 37 unique therapies. Changes in treatment coincided with the diagnosis of an SCN1A-related disorder, with an increase in cannabidiol, clobazam, and fenfluramine and reduction in sodium channel–blocker use following genetic diagnosis.
In summary, we reconstructed the longitudinal disease history of SCN1A-related disorders from electronic medical records using a standardized framework for the analysis of real-world clinical data. We refine existing natural history data of SCN1A-related disorders by providing a granular landscape of seizures, comorbidities, and treatment approaches over time.
Article notes
Competing Interest Statement
The authors have declared no competing interest.
Funding Statement
I.H. is supported by the National Institute for Neurological Disorders and Stroke (R01 NS131512 and NS127830). E.M.G. is supported by the National Institute for Neurological Disorders and Stroke (R01 NS110869 and NS137604). I.H. and E.M.G. are supported by the Dravet Syndrome Foundation (Research Grant).
Introduction
Disease-causing variants in SCN1A cause one of the most clinically recognizable and well-studied monogenic epilepsies with a prevalence of 1 in 15,700.1 A spectrum of disorders are associated with SCN1A, the most notable of which are genetic epilepsy with febrile seizures plus2 (GEFS+) and Dravet syndrome.3 GEFS+ was first characterized in large kindreds of individuals with fever-sensitive seizures, often recurring beyond the typical age range of febrile seizures and followed by unprovoked seizures.4 Epilepsy in Dravet syndrome onsets in the first year of life—most frequently four to six months of age—in a previously typically-developing infant, often consisting of prolonged, alternating hemiclonic seizures. Later, seizures may be provoked or unprovoked and include bilateral tonic-clonic (BTC), myoclonic, atypical absence, and focal semiologies. Seizure precipitants may evolve with age, and include fever, heat, stress, exercise, sleep deprivation, and visual/auditory stimuli.5 Epilepsy in Dravet syndrome is often refractory and is a major target of therapy development. Unlike GEFS+, Dravet syndrome has prominent neurodevelopmental co-morbidities, including hypotonia, ataxia, and developmental delay.6 90% of cases of Dravet syndrome are the result of de novo variants in SCN1A, with a small percentage attributed to parental mosaicism.7,8
The spectrum of SCN1A-related disorders also includes gain-of-function (GoF) variants, which can cause an early infantile-onset disorder with features distinct from both GEFS+ and Dravet syndrome. Existing literature refers to this disorder by several names, including neonatal developmental and epileptic encephalopathy (DEE) with movement disorder and arthrogryposis (NDEEMA) and early infantile DEE (EIDEE).9 Formal curation of SCN1A by the Epilepsy Gene Curation Expert Panel selected DEE as the gene-disease relationship. Given that not all individuals with this phenotype present with a movement disorder, arthrogryposis, or early infantile seizures, and to distinguish it from Dravet syndrome, we will refer to it as non-Dravet syndrome DEE (nd-DEE). For these individuals, seizures often onset before three months of age and are rarely provoked. Tonic seizures, unusual in GEFS+ and Dravet syndrome, are frequent. Similarly to Dravet syndrome, BTC, focal, and myoclonic seizures are common. Contractures and hyperkinetic movement disorders also distinguish this disorder from Dravet syndrome and GEFS+.10 Whole-cell voltage clamp studies, have demonstrated mixed effects with an overall gain of function, including a hyperpolarized shift of steady-state inactivation and a persistent sodium current.10
At the molecular level, SCN1A encodes the alpha subunit of the voltage-gated sodium channel NaV1.1, expressed predominantly in GABAergic inhibitory interneurons.11 Haploinsufficiency of this channel, caused by loss-of-function (LoF) variation in SCN1A, causes Dravet syndrome and GEFS+, whereas GoF variation causes the nd-DEE phenotype.10 The later onset of disease associated with LoF variants is presumed to be caused by a developmental switch. During early development, most SCN1A-encoded mRNA results in a nonfunctional protein product due to the inclusion of a naturally occurring poison exon. Later, this exon is skipped, and functional protein is made.12,13 SCN1A haploinsufficiency is therefore typical during early development but becomes pathologic with age. Alternatively, channel dysregulation caused by GoF variants is disruptive at all developmental stages, resulting in the earlier onset of disease.
Though much is known about the natural history of SCN1A-related disorders and their underlying pathomechanisms, a month-by-month analysis of seizures and comorbidities leveraging a standardized phenotypic vocabulary has not yet been performed. Natural history studies often employ differing methodologies, such as investigator-created surveys with pre-defined symptoms of interest reported at varying levels of detail. This renders comparisons between studies and across diseases difficult. Data collected clinically is similarly nonstandard. Vast amounts of information are available in the electronic medical record (EMR) of individuals with SCN1A-related disorders, though they are typically written in a form difficult to leverage for large-scale analysis. Efforts to use this information include translation into terms from a standardized biomedical dictionary. Resources such as the Human Phenotype Ontology (HPO) are now used in both clinical and research settings to communicate patient phenotypes in a standardized manner. The ontological structure of the HPO—wherein terms are related hierarchically—enables the systematic comparison of patient phenotypes both within and between genetic disorders.14–17
In addition to information collected in clinical trials, a systematic assessment of the natural history of SCN1A-related disorders using real-world data is a key step toward clinical trial readiness, particularly for a disease entity with multiple ongoing trials for targeted therapeutics. Not only can retrospective studies incorporate a larger study population for minimal cost, but they also permit the use of data obtained prior to a participant’s genetic diagnosis and enable the inclusion of individuals who would not otherwise be able to participate in a prospective study based on location or time required. Here, we performed a retrospective phenotypic analysis of 100 individuals to create a detailed phenotypic landscape across the lifespan of individuals with SCN1A-related disorders in a standardized language amenable to future comparisons.
Materials and methods
Cohort Identification
Our cohort consisted of 100 individuals with SCN1A-related disorders recruited at a tertiary care center and throughout the United States via the Dravet Syndrome Foundation. This sample size is comparable to similar studies of rare genetic epilepsies.16 Clinical documentation was gathered for each patient. Records were reviewed for the participant’s entire lifespan from birth until the date of review, which occurred from 2023–2024. SCN1A variants were converted into standard HGVS nomenclature. Clinical Genome Resource (ClinGen) Epilepsy Sodium Channel Variant Curation Expert Panel (ESC-VCEP) guidelines were used to curate each SCN1A variant to confirm genetic diagnosis and study eligibility. Evidence criteria applied to each variant are available in Supplementary Table 2.
Clinical Data Review and Standardization
EMR data was reviewed manually for each participant. Phenotypes were translated from free text into HPO terms. Age in months was recorded for each term based on the date of EMR entry. Epilepsy phenotypes were assigned a frequency score using a modified scale endorsed by the Epilepsy Foundation’s Epilepsy Learning Health System and Pediatric Epilepsy Learning Healthcare System.18 Scores range from 0–5 (0 = none, 1 = monthly, 2 = weekly, 3 = daily, 4 = 2–5 daily, 5 = >5 daily). In this way, seizure semiologies and their frequencies were reconstructed monthly at all ages where data was available. Missing data was assigned N/A and not included in analysis. Basic demographic and clinical information are reported in Supplementary Table 1.
EEG reports and treatment data were also extracted from the EMR. EEG reports were assessed for HPO terms as described above. Age in months and normal/abnormal status was recorded for each EEG. Clinical notes were reviewed for therapy use. The month of age of initiation and discontinuation of each treatment was recorded.
After extracting phenotypic data, HPO terms were propagated, meaning all higher-level terms in the ontological structure were assumed to be present. Seizure frequencies were propagated as well, assigning the highest frequency among more specific terms to their common ancestor. This methodology has been employed previously.14–16,19
Statistical Analysis
Statistical analyses were conducted using the R statistical framework.20 The frequency of each phenotypic term was calculated across the entire cohort and across subgroups. Fisher Exact and Wilcoxon Rank-Sum tests were used to assess clinical feature associations, compare symptom onset between subgroups, and assess the comparative effectiveness of epilepsy treatments as described previously.16
Ethics and Regulatory Approval
Participants or their guardians provided informed consent according to the Declaration of Helsinki. This study was approved by the Institutional Review Board of the Children’s Hospital of Philadelphia (IRB 15-012226).
Results
A quarter of diagnostic SCN1A findings are variants of uncertain significance
Despite increased emphasis on genetic testing, Dravet Syndrome and GEFS+ remain clinical diagnoses. Accordingly, our real-world dataset allowed us to contrast clinical diagnosis with variant evidence. In clinical practice, variants lacking sufficient evidence of pathogenicity are labelled variants of uncertain significance (VUS). When VUS are presumed to cause a patient’s symptoms, they are referred to as “explanatory” or “diagnostic.” Given that variant classification criteria were updated by the ClinGen ESC-VCEP in 2025, we assessed the evidence supporting each variant identified in this cohort.29 Copy number variants were assessed as outlined by the joint ClinGen/ACMG/AMP consensus recommendations.30
Though all variants were considered diagnostic, only 76% (61/80) of variants met criteria to be classified as pathogenic or likely pathogenic. 24% (19/80) of variants, including 41% (17/41) of missense variants, were therefore classified as VUS. 97% (75/77) of variants (excluding copy number variants) were present ≤1 time in population databases (gnomAD v4.1.0), providing supporting evidence of pathogenicity. Of variants where in silico predictors were available, 38/39 were deemed likely deleterious, contributing supporting (3) or moderate (35) evidence. Additional evidence was applied for each proband and for cases reported in the literature as observations of the variant in affected individuals. The full list of criteria met for each variant is reported in Supplementary Table 2. In summary, the spectrum of SCN1A-related disorders encompasses at least three clinical syndromes, two molecular mechanisms, and a range of established and emerging variants, contributing to a diverse and variable disease landscape.
Features of SCN1A-related disorders can be captured by longitudinal disease reconstruction
While prospective natural history studies are the gold standard to resolve disease trajectory and develop outcomes for clinical trials, the scalability of this approach is limited. Accordingly, recent efforts have focused on repurposing real-world clinical data to elucidate the natural history of genetic epilepsies, including those caused by STXBP1, SYNGAP1, and SCN8A.15,16,31 Here, we used a standardized biomedical ontology to transform natural language data from medical records into a standardized format. 8,171 patient-months of data were reviewed for the 100 participants. Phenotypes were recorded in one-month time bins, resulting in 671 unique clinical terms mapped to 2,220 distinct observations across the cohort.
A key feature of biomedical ontologies is their hierarchical structure, where terms more general than those assigned manually can be inferred. For example, if an individual is assigned the phenotype “Myoclonic seizure,” (HP:0032794) term propagation would assign both “Motor seizure” (HP:0020219) and “Seizure” (HP:0001250). Term propagation expanded the dataset to 1,174 unique features assigned 9,176 times. Mean phenotypic depth increased concomitantly—from 22 to 92 terms per individual. Of all terms, 75 referred to aspects of seizure semiology and 1,103 referred to non-seizure phenotypes, including developmental differences, behavioral abnormalities, and a range of additional symptoms. Well-recognized clinical features of SCN1A-related epilepsies were observed frequently, demonstrating that our method of disease reconstruction recapitulated known aspects of SCN1A-related disorders (Fig. 1A). Overall, the longitudinal phenotypic profiling of 100 individuals with SCN1A-related epilepsies yielded a dataset encompassing 681 patient-years and 9,176 clinical annotations, reconstructing the natural history of these disorders in a complementary manner to prospective studies.
Diagnostic variants in SCN1A cause distinct clinical syndromes
The distinct clinical syndromes caused by variation in SCN1A are united by the presence of multiple seizure types in all individuals, demonstrating a high burden of epilepsy. BTC seizures (96/100) and status epilepticus (96/100) stood out as near-universal features, followed by seizures precipitated by febrile infection (81/100), focal motor seizures (72/100), convulsive status epilepticus (69/100), and focal hemiclonic seizures (64/100; Fig. 1A), seizure types known to be associated with SCN1A-related disorders. Reflective of this high seizure burden, 76/100 individuals had at least one episode of respiratory desaturation associated with a seizure and 57/100 had at least one seizure cluster. Despite the initial identification of Dravet syndrome as severe myoclonic epilepsy of infancy, only 41/79 individuals with Dravet syndrome had myoclonic seizures.
We then compared phenotypic terms to clinical diagnosis. Though some features were common across clinical diagnoses, others distinguished them. Compared to individuals with Dravet syndrome, individuals with GEFS+ were less likely to have neurodevelopmental delay (P < 0.001), status epilepticus (P < 0.001), myoclonic seizures (P = 0.004), hypotonia (P = 0.038), or hemiclonic seizures (P < 0.001) (Fig. 1B). Conversely, these participants were more likely to have simple febrile seizures (P = 0.006) and depression (P = 0.016). The milder disease course of GEFS+ was also reflected in the number of terms assigned. Those with GEFS+ had roughly half the number of clinical annotations as those with Dravet syndrome (mean 54 vs. 98 terms per individual; P < 0.001). This pattern remained consistent when restricted to seizure phenotypes (mean 14 vs. 22 terms per individual; P < 0.001). Participants with the nd-DEE phenotype had a unique phenotypic landscape (Fig. 1C). Symptoms enriched in this sub-cohort compared to Dravet syndrome included joint contractures (P = 0.006), severe global developmental delay (P < 0.001), abnormal brain morphology (P = 0.002), dystonia (P = 0.027), and failure to thrive (P = 0.023). These individuals were less likely to have seizures triggered by fever (P = 0.003), myoclonic seizures (P = 0.004), or convulsive status epilepticus (P = 0.009). Diagnoses of classical and atypical Dravet syndrome, as assigned by providers, were difficult to disentangle. Only six clinical terms were either enriched or depleted in those with an atypical presentation, including a higher frequency of tonic seizures (P = 0.046) and cardiac arrhythmia (P = 0.028; Fig. 1D). In summary, the clinical diagnoses associated with SCN1A are distinguished by clinical features, including the number of terms assigned and the presence or absence of certain symptoms. Classical and atypical Dravet syndrome displayed minimal differences, supporting emerging evidence that they may not be distinct entities.21
Highly variable seizure burdens are seen in individuals with SCN1A-related disorders
We next focused on the longitudinal trajectory of seizures among participants. Although all individuals had seizures, onset and semiology varied widely (Fig. 2A). For individuals with Dravet syndrome, seizure onset occurred from 1–22 months, with a median of five months. Onset was later for GEFS+ (14 months, P < 0.001). Of those with the nd-DEE phenotype, 8/10 had seizure onset from 0–4 months of age. The other 2/10 were identical twins whose first definite seizures occurred at 11 months, though they displayed abnormal movements in early infancy that were not confirmed to be epileptic. Taken together, this sub-cohort had an earlier seizure onset (median 2.5 months, P = 0.014) than individuals with Dravet syndrome (Fig. 2B).
Across the cohort, seizure onset occurred within a narrow window. Between three and five months, incidence of seizures increased from <20% to >50% of individuals. The proportion of individuals with active epilepsy, defined as a seizure frequency greater than zero, remained at or above 50% between one and five years of age, peaking at 38 months when over 75% of individuals had seizures.
Semiology varied with aged. Early life was characterized by status epilepticus and hemiclonic seizures, which affected the highest proportions of individuals before one year of age (37% and 22% monthly incidence, respectively). Seizures triggered by fever were also common at early ages with an overall prevalence of 82% and a monthly incidence peaking at 28% at 14 months. After these early peaks, all three seizure types became less prominent. Alternatively, BTC seizures occurred frequently both in infancy and childhood. Between one and eight years of age, 21–49% (median 32%) of individuals had BTC seizures each month. Myoclonic seizures displayed a similar pattern, though they affected fewer participants (5–24% per month, median 12%). Atonic seizures, non-motor seizures, and focal impaired consciousness seizures affected smaller proportions of participants and often onset later in childhood, (median 6%, 4%, and 7% per month, respectively from age 1-8; Fig. 2C). Overall, our reconstruction of real-world data shows that childhood with an SCN1A-related disorder is characterized by persistent seizures despite treatment, though myoclonic seizures were not observed in most individuals.
Seizure frequency in SCN1A-related epilepsies is multidimensional
We next focused on seizure frequency across the cohort. Of those with Dravet syndrome, a range of frequencies were observed, from seizure freedom to many daily seizures. At all months before age five, at least one individual was seizure-free and one had daily seizures, highlighting the large variability in clinical presentations. Of those with seizures (frequency > 0) in a given month, the most common frequency was monthly, followed by weekly and then > 5 daily (Fig. 2A). 76/100 individuals had monthly seizures for more than six months during their lifetime, whereas only 45/100 had weekly or more frequent seizures for this length of time. All participants with GEFS+ were seizure-free for more than half the months surveyed. Seizure frequency was significantly lower than in Dravet syndrome (P < 0.001). For nd-DEE, though the temporal dynamics of seizures differed from Dravet syndrome, overall seizure frequency did not differ.
Seizure frequency also varied month-to-month. The average individual had a median of three seizure frequencies and 17 alterations in seizure frequency throughout their life (Fig. 2A, 3A). Frequencies were not distributed equally by seizure type. Fever-triggered seizures, status epilepticus, and BTC seizures occurred almost exclusively monthly, whereas myoclonic seizures often occurred daily. Semiologies of intermediate frequency included hemiclonic seizures and focal impaired consciousness seizures, which tended to occur weekly or monthly (Fig. 3B-D).
Behavior, development, and motor differences are key features of SCN1A-related disorders
Neurodevelopmental and behavioral phenotypes commonly reported in association with Dravet syndrome were also observed in this cohort. Behavioral disturbances, including maladaptive behavior (45/100), sleep disturbance (43/100), and attention difficulties (22/100), were reported in 68/100 individuals (Fig. 4A). Only 14/100 individuals carried a formal diagnosis of autism spectrum disorder, which is much lower than that reported in studies assessing rates of autism in Dravet syndrome using standardized tools (24–40%).32 The proportion of participants reported to exhibit autistic features, however, falls within this range (28/100). Aggressive behavior was observed in 25/100 participants. Most behavioral features onset between one and ten years of age, though attention difficulties were often first reported between 3–6 years of age (14/22). Cognitive differences affected 83/100 individuals (Fig. 4B). The most commonly impaired domain was speech and language (69/100). Delayed motor development was less common, affecting 31/100 individuals, including gross motor (20/100) and fine motor (16/100) delays. Despite the high rate of developmental delay, only 19% of those over five years of age had a diagnosis of intellectual disability. This value—obtained from real-world data as opposed to formal developmental assessments—is lower than would be suggested by findings of prospective natural history studies of Dravet syndrome, which suggest plateauing around the developmental age of two years.33 Although motor delays affected only one-third of participants, 69/100 individuals went on to develop impaired motor function (Fig. 4C). This category of symptoms included abnormal tone (42/100), incoordination (43/100), gait disturbance (41/100), and imbalance (33/100). Of these features, hypotonia was observed at the earliest ages, with a median onset of two years. Gait disturbances were characterized by broad-based gait (20/100), ataxia (10/100), unsteadiness (7/100), tiptoeing (6/100), and crouch gait (5/100).
EEG abnormalities in SCN1A-related disorders develop with age
Due to a high frequency of refractory epilepsy, individuals with SCN1A-related disorders often undergo frequent EEG monitoring. We leveraged reports of 569 EEGs performed on 96 individuals. 77% (440/569) of EEGs were performed before age four, during which time the average participant received approximately one EEG annually. 191 studies were normal, and many individuals (68/96) had both normal and abnormal EEGs (Fig. 5A). Most EEGs were read as normal before nine months of age (81/146; Fig. 5B). EEG abnormalities before three months were limited to focal epileptiform discharges, almost exclusively occurring in the pGoF cohort due to their earlier onset of epilepsy (Fig. 5C). These abnormalities occurred in multiple brain regions—central, temporal, occipital, and parietal—and consisted of focal sharp waves, spikes, and spike-waves. Multifocal epileptiform discharges were also enriched in this population (6/9 pGoF, 14/87 pLoF; P = 0.002). With age, focal epileptiform discharges became more common in the pLoF cohort (50/89), consistent with their later onset of epilepsy. Across all participants, EEG anomalies included generalized epileptiform discharges (53/96), slowing (69/96), spike-wave complexes (48/96), and focal sharp waves (26/96). Overall, we leveraged clinical EEG data to demonstrate differences in EEG findings between pGoF and pLoF SCN1A-related disorders, particularly in early life.
The diagnosis of an SCN1A-related disorder is correlated with medication management changes in clinical practice
Seizure therapies, including medications, surgeries, and dietary approaches, were available for 95 patients, encompassing 554 therapies used over 1,405 patient years (Fig. 6A). Levetiracetam was the most frequently used medication, prescribed to 91% of participants (87/95), followed by valproic acid (69%, 66/95), clobazam (67%, 64/95), cannabidiol (57%, 54/95), and fenfluramine (53%, 50/95). The ketogenic diet was trialed in 35% of individuals (33/95). Surgical approaches, such as vagus nerve stimulation (11%, 10/95), were also recorded. Therapies were used for a median of 18 months. The ketogenic diet, clobazam, and cannabidiol were often used for extended periods (median 32, 26.5, and 23.5 months, respectively), whereas sodium channel blockers (SCBs)—contraindicated in pLoF SCN1A-related disorders—were used for shorter periods, such as oxcarbazepine, lacosamide, and lamotrigine (median 3, 3.5, and 5 months, respectively).
Therapies also varied with age. In infancy, levetiracetam was used by most individuals with pLoF variants, a proportion that decreased sharply between 0–2 years of age due to the increased use of other therapies, including cannabidiol, clobazam, fenfluramine, the ketogenic diet, and valproic acid (Fig. 6A). Those with pGoF variants also had a high frequency of levetiracetam use in early life; however, the medications replacing its use were more often SCBs such as oxcarbazepine. In addition to age, the diagnosis of an SCN1A-related disorder coincided with changes in treatment. Diagnosis of a pLoF SCN1A-related disorder was often accompanied by the discontinuation of levetiracetam and SCBs and initiation of clobazam, cannabidiol, and fenfluramine. Genetic diagnosis did not affect all therapies; valproic acid and ketogenic diet use was similar before and after diagnosis (Fig. 6B). Among those with pGoF variants, levetiracetam was often discontinued, and clobazam initiated, following genetic diagnosis. Changes in SCBs—recommended for those with pGoF variants—were more ambiguous, with increased use of lacosamide but decreased oxcarbazepine use.
Treatment data were compared to seizure frequency to assess their comparative efficacies, a metric used in prior studies to identify therapies particularly efficacious in rare disease populations (Fig. 6C).16 For those with pLoF variants, stiripentol (P = 0.033) and vagus nerve stimulation (P = 0.003) were more associated with reduced seizure frequency. Levetiracetam (P < 0.001), valproic acid (P < 0.001), zonisamide (P < 0.001), the ketogenic diet (P < 0.001), topiramate (P = 0.006), and clorazepate (P < 0.001) were associated with an increased ability to maintain seizure freedom. The SCBs lacosamide (P < 0.001) and lamotrigine (P < 0.001) were associated with a lack of seizure freedom maintenance (Fig. 6C). Comparative effectiveness was not assessed for the pGoF cohort due to the small number of participants.
Discussion
Here, we reconstructed the clinical trajectories of 100 individuals with SCN1A-related disorders across the lifespan, leveraging real-world data collected during routine clinical care. These data represent an abundant yet underused resource with the ability to complement the findings of prospective natural history studies. Standardized, retrospective, longitudinal disease trajectories were assembled for each participant—a novel paradigm as applied to SCN1A. Detailed information regarding the natural history of SCN1A-related disorders is essential to the provision of clinical care and anticipatory guidance, evaluation of approved therapies, and development of trial outcome measures in the era of precision medicine. We find that SCN1A-related disorders encompass distinct clinical syndromes, the diagnosis of which correlates with medication management changes, even though almost a quarter of diagnostic variants remain VUS.
Limitations
A primary limitation of this study is its restriction to the real-world data recorded during clinical care. These data are often heterogeneous and may be incomplete, as factors influencing medical decision-making may not align with those important to a study of natural history. Using a controlled biomedical dictionary, these challenges are at least partially addressed by enabling the comparison of features recorded at varying levels of detail. Additionally, by incorporating a larger cohort at relatively low cost, this methodology compensates for missing information in the EMR.
Conclusions
In summary, real-world retrospective data was used to assess the natural history of SCN1A-related disorders, demonstrating a diversity of variants, clinical diagnoses, seizure trajectories, developmental outcomes, and medication response among affected individuals. Our results emphasize ongoing challenges in diagnostics and management of SCN1A-related disorders. Though improvements in genetic diagnostics and variant curation continue, 24% of identified variants remain variants of uncertain significance. Moreover, even with the approval of pharmacologic therapies trialed in individuals with SCN1A-related disorders, seizure burden in this population persists, with over 50% of individuals having seizures from age 1–5, and developmental features remain therapeutically unaddressed. This underscores the critical need for disease-modifying therapies in this population. Encouragingly, the real-world data compiled here contributes to a wealth of knowledge of the natural history of SCN1A-related disorders and demonstrates that their diagnosis is associated with recommended changes in clinical management, which are integral building blocks for the development and application of precision medicine approaches.
Data availability
The data that support the findings of this study are available on request from the corresponding author. Data are not publicly available due to their containing information that could compromise the privacy of participants.
Supporting information
Data Availability
All data produced in the present study are available upon reasonable request to the authors.
Acknowledgements
We would like to acknowledge the Dravet Syndrome Foundation and all participants for their ongoing support of this project.
Funding
I.H. is supported by the National Institute for Neurological Disorders and Stroke (R01 NS131512 and NS127830). E.M.G. is supported by the National Institute for Neurological Disorders and Stroke (R01 NS110869 and NS137604). I.H. and E.M.G. are supported by the Dravet Syndrome Foundation (Research Grant).
Competing interests
The authors declare no competing interests.
Supplementary material
Supplementary material is available online.