Genetic complexity in pediatric onset epilepsy‐movement disorder syndromes: Insights from a cohort of 97 subjects

Abstract

Objective

Conditions presenting with both epilepsy and movement disorders (EPIMDs) range from relatively benign cases to severe developmental encephalopathies. However, the full clinical and genetic spectrum still needs to be better defined. The aim of this study is to describe the presentation of EPIMDs in pediatric patients with known genetic etiologies, correlating these features with age at onset and underlying pathological mechanisms to identify patterns that could improve patient management.

Methods

We retrospectively analyzed the clinical records of pediatric patients who underwent genetic testing for EPIMDs at our institution between 2009 and 2022. Genetic testing included single‐gene Sanger sequencing, multigene next‐generation sequencing panels targeting epilepsy and/or MDs, and array comparative genome hybridization. Demographic, clinical, and electroencephalography (EEG) data were collected. Statistical analyses were conducted using multivariate and cluster analyses.

Results

A total of 97 subjects were included. The mean age ± SD at epilepsy onset was 3.6 ± 4.4 years, whereas the mean age at MD onset was 5.2 ± 4.9 years. Based on the mechanism of genetic dysfunction, we identified six groups: transportopathies (n = 18), synaptopathies (n = 11), channelopathies (n = 18), metabolic disorders (n = 15), intracellular trafficking defects (n = 8), and unknown/complex function disorders (n = 27). Cluster analysis identified three distinct groups: (1) early‐onset epilepsy and paroxysmal MDs with normal intelligence quotient (IQ) and mild intellectual disability, associated with transportopathies (n = 38); (2) early‐onset epilepsy with dystonia and myoclonus, low IQ, and developmental and epileptic encephalopathy, associated with channelopathies (n = 31); (3) epilepsy onset between 1 and 5 years of age with developmental delay, parkinsonism, and low IQ, associated with metabolic etiology (n = 26).

Significance

Our study demonstrates that, despite the phenotypic and genetic pleiotropy observed in EPIMDs, a precise characterization of semiological and cognitive features remains essential to support variant interpretation and to refine the diagnostic process and patient management in these rare conditions.

Key points.

• A wide range of genetic causes underlies epilepsy with movement disorders, reflecting strong clinical and genetic variability.

• In most patients, seizures appear before movement disorders, and early onset is often linked to more severe outcomes.

• Channelopathies and trafficking defects are more often associated with severe phenotypes.

• Cluster analysis reveals distinct profiles based on age at onset, symptoms, cognitive level, presence of developmental and epileptic encephalopathy, disease mechanism, and treatment response.

• Combining genetic, clinical, and neurophysiological data improves diagnostic precision in epilepsy and movement disorder syndromes.

1. INTRODUCTION

The relationship between epilepsy and movement disorders (MDs) during childhood is particularly intriguing, as these conditions can co‐occur in rare genetic syndromes. Their coexistence often complicates the clinical picture, and their paroxysmal nature requires specific clinical and neurophysiological expertise for an accurate differential diagnosis. Recent advances in genetic techniques have significantly improved our understanding of the etiology of both epilepsy and MDs, ¹ , ² suggesting that the two conditions may share common genetic factors. ³ This has led to the identification of several gene‐specific disorders caused by inborn errors of metabolism, single‐gene mutations, and chromosomal rearrangements, ⁴ all characterized by the co‐occurrence of epilepsy and MDs. Dysfunction in ion channels, synaptic transmission, transmembrane transport, as well as metabolic and neurodegenerative processes, may all contribute to the pathogenesis of both epilepsy and MDs. ⁵ , ⁶ Although the etiological spectrum of epilepsy and MDs (EPIMDs) is constantly expanding, many cases remain undiagnosed, and others are associated with ultra‐rare conditions, making it difficult to establish a clear genotype–phenotype correlation. Consequently, the clinical manifestations of genetic EPIMDs are highly heterogeneous, with significant phenotypic variability ranging from severe developmental and epileptic encephalopathy (DEE) to relatively benign conditions. ⁷ , ⁸ , ⁹ Moreover, EPIMDs frequently coexist with other neurodevelopmental disorders, such as intellectual disability (ID), cerebral palsy, autism spectrum disorder, and attention‐deficit/hyperactivity disorder, further complicating the clinical presentation. The aim of this study is to describe the clinical features of a large cohort of pediatric patients with EPIMDs and to explore their relationship with underlying genetic dysfunctions and supposed pathological mechanisms.

2. METHODS

2.1. Data collection

We retrospectively analyzed data from patients who underwent genetic testing for epilepsy or MDs at the Carlo Besta Neurological Institute between 2009 and 2022. Genetic testing included single‐gene Sanger sequencing, multigene next‐generation sequencing (NGS) panels (Agilent Sure Design, Santa Clara, CA, USA) targeted for epilepsy or MDs, and array comparative genome hybridization (aCGH) analysis (see ¹⁰ for details on the genetic testing methodology). Only class IV–V variants were included, according to the American College of Medical Genetics and Genomics (ACMG) guidelines. ¹¹ Written informed consent was obtained from parents/legal representatives. We included patients with EPIMDs with onset before the age of 18 years. We collected the following demographic and clinical data: family history, perinatal history, developmental milestones, neurological examination, age at onset, type and evolution of seizures/EPIMDs, neuropsychological assessment, history of status epilepticus/dystonicus, medications, and clinical outcome. Drug‐resistant epilepsy (DRE) was defined accordingly to Kwan et al. 2009. ¹² We also reviewed relevant diagnostic investigations, including neuroimaging data, video–electroencephalography and EEG–polygraphy studies, and post‐processing signal analysis (jerk‐locked back‐averaging and cortico‐muscular coherence analysis). According to the EEG terminology glossary, ¹³ EEG recordings were reviewed and classified based on background activity (normal, slowed, or disorganized) and the presence and topography of epileptiform abnormalities (focal, multifocal, diffuse, or generalized). Seizures were classified according to the recent International League Against Epilepsy (ILAE) classification system. ¹⁴ Patients presenting with more than one seizure type were grouped under “multiple seizure types.” In accordance with ILAE criteria, ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ we identified patients fulfilling the diagnostic features of DEE, and epilepsy syndromes were classified accordingly. The categorization reflects the recent operational definition, ¹⁹ which underscores the usual overlap of clinical features in DEE, and the ensuing difficulty in keeping them separate. Movement disorders were classified based on the pediatric movement disorders classification ²⁰ , ²¹ into: Paroxysmal (including kinesigenic [PKD], non‐kinesigenic [PNKD], exercise‐induced dyskinesias [PED], and alternating hemiplegia), mainly hypokinetic MDs (e.g., rigidity, bradykinesia), dystonia, myoclonus, and other hyperkinetic MDs (e.g., chorea, tremor, and undefined hyperkinetic movements).

2.2. Statistical analysis

A descriptive analysis was performed to summarize and characterize the clinical features of the patients (mean and standard deviation [SD] for continuous data and number and percentage for categorical data). To evaluate age‐specific phenotypes, we stratified subjects based on the first symptom (epilepsy first, MD first, EPI‐MD joint), age at onset of the first symptom (infancy: first year of life; early childhood: 1–5 years; childhood–adolescence: over ≥ 6 years), and supposed pathological mechanisms based on genetic dysfunction (transportopathies, synaptopathies, channelopathies, metabolic disorders, intracellular trafficking disorders, and unknown/complex mechanisms). Data were first checked for normality using the Shapiro–Wilk test. Comparison between groups for continuous data was performed using the non‐parametric Kruskal–Wallis test, whereas categorical data were compared using a chi‐square test with Bonferroni correction for multiple comparisons. Logistic regression analysis was used to identify the factors potentially associated with DEE presence, response to antiseizure medication (ASM) and intelligence quotient (IQ) impairment. Finally, a two‐step cluster analysis with Schwarz's Bayesian Information Criterion was undertaken to classify clinical groups according to age at onset, first symptom, etiology, IQ, and DEE. This analysis employed a hierarchical clustering algorithm for the pre‐clustering step and a k‐means algorithm for the final clustering. The differences between the resulting clusters were tested using Pearson's chi‐square test. All analyses were performed using SPSS (version 27, SPSS Inc., IBM). The significance level was assessed at p < .05.

3. RESULTS

We identified 97 patients (45 male, 52 female) with genetic proven etiology: 46 different genes and 6 copy number variants (CNVs). We reached genetic diagnosis through single‐gene Sanger sequencing in 41 patients (42.3%), multigene panels in 32 (33.0%) (including MD‐targeted panels in 4 patients), and aCGH in 22 (22.7%).

3.1. Genetic features

We identified pathogenic variants in 46 different genes and 6 microdeletion/microduplication syndromes (Figure 1). Pathogenic variants were most frequently detected in the following genes: SLC2A1 (n = 11), PRRT2 (n = 7), SCN1A (n = 6), ATP1A3 (n = 6), TBC1D24 (n = 5), and WDR45 (n = 5). In these cases where the pathogenic mechanism has been elucidated, the identified mutations predominantly result in loss‐of‐function effects. Most of the other genes were identified in fewer than three patients. Pathogenic CNVs were observed in six cases. According to the mechanism of genetic dysfunction we identified six groups: transportopathies (n = 18, 18.6%), synaptopathies (n = 11, 11.3%), channelopathies (n = 18, 18.6%), metabolic/degenerative (n = 15, 15.5%), intracellular trafficking (n = 8, 8.2%), and unknown/complex mechanism (n = 27, 27.0%) (including patients with chromosomal rearrangements and genes with complex/unknown function). Transportopathies and channelopathies were the most prevalent etiologies. Channelopathies were related to sodium and potassium channel dysfunction. Group 6 included several genes implicated in disorders of cell cycle regulation (i.e., FOXG1, MECP2) and degradation or turnover of cellular degradation (i.e., UBE3A).

FIGURE 1.

Etiologies and mechanisms of the 97 pathogenic variants described in our cohort; each one is followed by the number of patients affected.

3.2. Clinical and demographic data

Developmental delay was the first neurological symptom in 43 patients (44.3%). The mean age ± SD at onset was 3.58 ± 4.42 (range 0–18) years for seizures and 5.18 ± 4.94 (range 0–17) years for MDs with delta onset (mean individual difference) of 4.25 ± 4.09 (range 0–18) years. Age at first symptom onset (including developmental delay) was 2.29 ± 3.27. Seizures preceded MDs in 57 patients (58.8%), whereas MDs preceded seizures in 29 patients (29.9%). Seizures and MDs appeared at the same age in 11 patients (11%). At neurological evaluation, 49 patients (50.5%) had additional neurological signs including pyramidal signs (n = 28); axial hypotonia (n = 17); ataxia (n = 14); visual impairment, retinopathy, or optic atrophy (n = 8); and hearing impairment (n = 1). Thirty‐three patients had microcephaly (34.0%) and five had macrocephaly (5.2%). Formal cognitive testing was available for 95 of 97 patients. ID was present in most patients (n = 73, 76.8%), with severity ranging from mild (n = 11) through moderate (n = 22) to severe (n = 40). The 22 patients with normal IQ carried variants in the following genes: ATP1A3 (Rapid Onset Dystonia Parkinsonism phenotype), SLC2A1, PRRT2, CHRNA7, SCN1A (Generalized Epilepsy with Febrile Seizure + phenotype), KCNA2, KCNT1, POLG1, NKΧ2, and EFHC1. Epilepsy was refractory to anti‐seizure medications (ASM) in 54 patients (56.0%).

3.3. EPIMDs

The features of seizures and MDs are summarized in Table 1. Focal seizures were the most frequent (n = 33, 34.0%) followed by epileptic spasms (15%), absences (11%), and generalized tonic–clonic seizures (7%); 31 patients (31.9%) showed multiple types of seizures. Five children experienced status epilepticus and 15 a dystonic status. In 39 patients (40.2%) epilepsy featured as DEE, with onset of seizures in the first year of life in 21. A definition of epileptic syndrome was possible in 27 patients (28%), as most presented complex electro‐clinical and neurological phenotypes. In addition, most patients (45.4%) presented with a combination of different movement disorders. Based on the predominant MD types we classified MDs as either paroxysmal (26 patients) or chronic (71 patients). Paroxysmal MDs included alternating hemiplegia (n = 3, 11.5%), paroxysmal eye movements (n = 1, 3.9%), and paroxysmal dyskinesias (n = 22, 84.6%). Among paroxysmal dyskinesia, eight patients experienced PNKD (36.4%), seven PKD (31.8%), and seven PED (31.8%). Chronic MDs included dystonia (n = 21, 29.6%), myoclonus (n = 15, 21.1%), other hyperkinetic MDs (n = 22, 31.0%), and hypokinetic MDs (n = 13, 18.3%). EEG features comprised abnormal background activity in 50 patients (53.1%), with slowing observed in 25 and disorganization in the other 25, including 3 cases with hypsarrhythmia. Epileptiform abnormalities were present in 93.6% of patients (see Table 1). In three cases EEG were not available. In 37 patients with hyperkinetic MDs we performed additional video‐EEG and polygraphic recordings to better characterize the type of MD (see Figure 2 for examples of video‐EEG polygraphic studies): the electromyography (EMG) features were consistent with multifocal or action myoclonus in 23 (23.7%) and with tremor in 8 (8.2%) patients. In one patient (carrying a KCNA2 variant) the EEG polygraphic characteristics documented the coexistence of cortical myoclonus and tremor. An excessive startle response was recorded in three patients. When comparing DEE with non‐DEE patients, we did not observe statistically significant differences in seizure and MD type; however, we observed that absences and paroxysmal MDs were unlikely to be associated with DEE and spasm seizures are always present in DEE.

TABLE 1.

Demographics and clinical data.

	Gene (N)	Early developmental delay (N)	Other neurological signs (N)	Predominant MD (N)	Seizure types (N)	EEG (N)	Epilepsy syndrome (N)	DEE (N)	DRE (N)	Cognitive level (N)	Already published patients (N, [reference])
Transportopathies N = 18	ATP1A3 (6)	4	2	PD (2), Dystonia (4)	F (1), FIC (1), GTC (2), AAS (1), FPC/FIC /GTC (1)	N/MF (1), N/F (1), S/D (1), N/N (1), N/G (1), N.A. (1)	EMAts (1)	0	1	Normal (1), mild (1), moderate (2), severe (1), n/a (1)	2, [22]
	SLC2A1 (11)	7	6	PD (11)	AAS (6), AAS/GCT (1), AAS/GM (1), GTC/F (1), GM/GTC (1), FBTC/(1)	S/N (1), N/G (3), S/G (2), N/F (1), N/IRDA (1), N.D. (3)	0	0	5	Normal (4), mild (5), moderate (1), severe (1)	11, [23]
	SLC6A1 (1)	0	0	PD	AAS	N/G	EMAts	0	1	Moderate
Synaptopathies N = 11	PRRT2 (7)	3	3	PD (7)	FPC (1), FIC (4), GTC (1), TAS (1)	N/MF (4), N/N (1), N/G (1), N/F (1)	SeLFNIE (2), CAE (1), GEFS+ (1)	0	2	Normal (5), mild (1), moderate (1)	7, [24]
	GRIN2A (1)	1	1	PD	FIC (1)	N/D	DEE‐SWAS	1	0	Moderate
	CHRNA7 (2)	2	2	Myoclonus (1), other (1)	FIC (1), GES (1)	D/F (1), S/F (1)	0	1	2	Normal (1), moderate (1)
	DNM1 (1)	1	1	Other (1)	GES (1)	D/D	0	1	1	Severe
Channellopathies N = 18	SCN1A (6)	5	1	Hypokinetic (3), Myoclonus (2), other (1)	FIC/GM/GTC (1), FIC/GTC (2), FIC/GT/GTC (1), AAS/FIC/GTC (1), FIC (1)	S/MF (3), N/MF (2), D/G (1)	DS (5), GEFS+ (1)	5	5	Normal (1), moderate (4), severe (1)
	SCN2A (1)	1	1	Dystonia	GES	Hypsarrhythmia	EIDEE	1	1	Moderate
	SCN8A (2)	1	1	Myoclonus (2)	AAS/FIC (2)	S/MF (2)	0	0	1	Moderate (2)	1, [25]
	SCN1A, SCN2A, SCN9A (2)	2	1	Myoclonus (1), dystonia (1)	GT/GM/GTC (1), FIC (1)	N/G/IRDA (1), S/MF (1)	EIDEE (1)	2	2	Severe (2)
	KCNA2 (1)	0	0	Other	GM/GTC	D/MF	PME	0	1	Normal
	KCNB1 (1)	1	1	PD	FIC/GT	D/D	0	1	1	Severe
	KCNC1 (1)	1	1	Other	FBTC/FIC	S/D	0	0	1	Normal
	KCNQ2 (1)	0	1	PD	GM/GTC	N/G	GEFs+	0	1	Normal
	KCNQ3 (1)	1	1	Dystonia	FIC	N/MF	0	1	1	Moderate
	KCNT1 (1)	1	1	Myoclonus	AS	N/G	0	0	0	Normal
	HCN2 (1)	1	1	Dystonia	FIC/FBTC/ GTC	S/MF	0	1	1	Severe
Metabolic/degenerative N = 15	WDR45 (5)	5	1	Hypokinetic (5)	FIC (1), GES (3), GES/GT/FIC (1)	D/MF (1), N/MF (1), D/D (2), D/MF (1)	LGS (1)	4	5	Severe (5)	4, [26]
	PANK‐2 (3)	2	2	Dystonia (3)	GTC (2), FIC/GTC (1)	D/N (1),4, S/N (1), N/D (1)	0	0	0	Moderate (3)
	PLA2G6 (1)	1	0	Hypokinetic	GTC	S/MF	0	0	1	Severe
	HTT (2)	2	2	Hypokinetic (2)	GES (1), GTC (1)	D/D (1), N/D (1)	0	1	1	Normal (1), moderate (2)
	BRAT1 (1)	1	1	Myoclonus	F	D/D	EIDEE	1	1	Moderate
	POLG (1)	1	1	PD	FIC	N/MF	0	0	0	Normal
	NHLRC1 (1)	1	1	Hypokinetic	GM/FIC/GTC	D/G	PME	0	1	n/a
	ST3GAL5 (1)	0	0	Other	GES	Hypsarrhythmia	IESS	1	1	Severe
Intrac.traff./signaling N = 8	TBC1D24 (5)	5	3	Myoclonus (5)	FPC(EPC) (5)	D/MF (5)	0	5	5	Severe (5)	1, [27]
	GNAO1 (3)	3	2	Dystonia (3)	FIC (2), GES (1)	Hypsarrhythmia (1), D/MF (1), N/D (1)	EIDEE (1)	3	2	Severe (3)
Other or complex mechanisms N = 27	SLC16A2 (3)	3	3	Dystonia (3)	GES (3)	D/D (3)	0	3	2	Severe (3)
	CASK (2)	2	2	Dystonia (1), other (1)	GES/FIC (1), GES (1)	D/D (2)	0	2	1	Severe (2)
	COQ4 (1)	0	0	Other	FIP	N/MF	0	0	0	Moderate
	TITF1/NKX2 (1)	0	0	Other	FIC	N/N	0	0	0	Normal
	HNRNPU (1)	1	0	Other	GM/FIC/GTC	N/MF	0	0	0	Severe
	EFHC1 (1)	0	0	Other	AS/EMA/GTC	N/G	0	0	0	Normal
	MEF2C (1)	1	1	Other	FBTC/FIC/GTC	S/MF	0	0	0	Moderate
	FOXG1 (1)	1	0	Other	FIC	ND	0	0	0	Severe
	MECP2 (1)	1	0	Other	FIC	S/F	0	0	1	Severe
	GPR56 (1)	1	1	PD	GES	D/MF	LGS	1	1	Severe
	SOX2 (1)	1	1	Dystonia	FIC	N/MG	0	0	0	Severe
	NEXMIF (1)	1	0	Myoclonus	AAS	N/G	0	0	1	Normal
	UBE3A (1)	1	0	Myoclonus	FIC	N/MF	0	0	1	Severe
	SNRPN/UBE3; PMP22 (1)	1	0	Other	FIC	D/D	GEFS+	0	0	Mild
	CNVs (6)	4	3	Hypokinetic (1), myoclonus (1), other (4)	FIC/CGT (4), CGT (1), FIC (1)	S/F (1), N/D (1), N/MF (2), S/MF (1), S/D (1)	0	1	1	Normal (1), mild (3), severe (2)
	ASTN2 (1)	1	0	Other	GES	S/MF	IESS	1	1	Severe
	GNB5 (1)	1	1	Dystonia	GES	D/D	0	1	0	Severe
	SEMA6B (1)	0	0	Other	CGTC	N/G	GEFS+	0	0	Normal
	HUWE (1)	1	1	Other	FIC	N/MF	EE‐SWAS	1	1	Severe

Note: The causative genetic dysfunctions are listed according to hypothesized patho‐mechanism according to Guerrini et al. ⁵ (modified); CNVs (copy number variants) included are: MIM#607872, chromosome 1p36 deletion syndrome; MIM#614671, 16p11.2 duplication syndrome; MIM#615656, 15q11.2 deletion syndrome; MIM#611913, 16p11.2 deletion syndrome; MIM#610883, Potocki‐Lupski syndrome; MIM#308100, Ichthiosis X‐linked.

Abbreviations: EEG column: Background activity: D, disorganization; N, normal; S, slowing. Epileptiform Abnormalities: D, diffuse; F, focal; G, generalized; IRDA, intermittent rhythmic delta activity; MF, multifocal; N.A., not available. Seizure column: AA, atypical absence; F, focal; FBTC, focal‐to‐bilateral tonic–clonic; FIC, focal impaired consciousness seizure; FPC, focal preserved consciousness seizure; GES, generalized epileptic spasms; GM, generalized myoclonic; GMA, generalized myoclonic–atonic; GT, generalized tonic; GTC, generalized tonic–clonic; TA, typical absence. Epilepsy Syndromes: CAE, childhood absence epilepsy; DEE‐SWAS, DEE with spike and wave activation during sleep; DS, Dravet syndrome; EIDEE, early‐infantile developmental and epileptic encephalopathy; GEFS+, genetic epilepsy with febrile seizures plus spectrum; IESS, infantile epileptic spasms syndrome; PME, progressive myoclonic epilepsy; SelFNIE, self‐limited familial neonatal‐infantile epilepsy.

FIGURE 2.

Left panel: HNRNPU variant. At the age of 3 years, brief spike and wave (SW) discharges were associated with reduced muscle tone both at rest and during active moments (A); at the age of 5 years, there were rhythmic myoclonic jerks during active movements (B), the coherent and synchronous electromyographic (EMG) bursts are displayed in D (for details on methodology). ²⁸ Right panel. Micro‐duplication in chromosome 2 (SCN1A, SCN2A, and SCN9 genes). At the age of 2 years brief seizures with independent appearance on both hemispheres (E and F); at the age of 10 years, noise‐induced excessive startle reflex (G) and paroxysmal dyskinesia (H).

3.4. Correlations between epilepsy and MD age at onset and mechanisms

By comparing frequencies in the different age groups (Figure 3), we found that younger age at disease onset was associated mainly with DEE and seizures (χ ²(4) = 13.8, p = .008). In addition younger children exhibited significantly higher proportion of focal seizures and spasms (χ ²(8) = 27.0, p < .001) and DEE (χ ²(2) = 18.1, p < .001). Absences and generalized convulsive seizures were more frequent in patients who were older than 5 years. Moderate to severe ID was observed more frequently in children with onset in the first year of life and in those with onset between 1 and 5 years (χ ²(6) = 23.2, p < .001). Channelopathies and trafficking disorders were significantly more frequent in the group of children with onset in the first year of life, compared to children with later onset (χ ²(10) = 27.2, p = .002). Children 1 to 5 years of age were more likely to experience onset with MDs and developmental delay. No significant differences were found with regard to type of MD and age at onset, although dystonia and myoclonus were more represented in younger children. Furthermore, dystonia was significantly more frequent in children with MDs that began before seizures. By comparing etiologies (Table 2), we observed that both channelopathies and trafficking disorders exhibit early onset, typically with epilepsy in the former and MDs in the latter. Conversely, transportopathies and synaptopathies had a significantly older age at onset of MDs. Absence seizures were significantly more frequent in transportopathies, whereas paroxysmal MDs were more frequent in transportopathies and synaptopathies and parkinsonism in metabolic disorders. DEEs were more common in channelopathies, trafficking, and unknown/complex and never observed in transportopathies. ASM response was poor in chanelopathies. Transportopathies showed a normal‐mild cognitive level, whereas moderate to severe impairment was more frequent in channelopathies, trafficking, and the unknown group.

FIGURE 3.

Correlations between first symptom type, movement disorder (MD) type, seizure type, DEE presence, antiseizure medication (ASM) response, intelligence quotient (IQ) in relation with age. * indicates statistically significant differences between groups (p < .005).

TABLE 2.

Correlations between mechanisms and the collected variables.

	Transportopathies	Synaptopathies	Channellopathies	Metabolic	Intracellular trafficking or signaling	Other	Statistical test
Demographics
N (%)	18 (18.6)	11 (11.3)	18 (18.6)	15 (15.5)	8 (8.2)	27 (27.8)	H (5)=17.8 p = .003
Age onset, years	2.9 ± 3.8	2.4 ± 2.6	1.0 ± 1.7	2.8 ± 3.0	.1 ± .2	3.1 ± 4.1	H (5)=23.9 p < .001
MD onset, years	7.2 ± 5.0	7.7 ± 4.1	5.3 ± 4.3	3.2 ± 3.2	.2 ± .3	5.3 ± 5.9	H (5)=16.9 p = .005
Epilepsy onset, years	4.5 ± 5.8	2.8 ± 3.6	1.0 ± 1.7	4.5 ± 4.2	2.0 ± 3.5	4.9 ± 4.7	H (5)=12.4 p = .030
Δ onset (years)	6.2 ± 4.9	5.7 ± 3.9	4.4 ± 3.8	2.2 ± 2.8	2.0 ± 3.5	4.1 ± 3.9
First symptom N (%)							χ 2(10) = 24.3 p = .007
Movement disorder	4 (28.6)	1 (7.1)	0 (0)	5 (35.7)	0 (0)	4 (28.6)
Epilepsy	6 (15)	6 (15)	14 (35)	3 (7.5)	5 (12.5)	6 (15)
Developmental Delay	8 (18.6)	4 (9.3)	4 (9.3)	7 (16.3)	3 (7)	17 (39.5)
Movement disorder type N (%)							χ 2(20) = 111.6 p < .001
Paroxysmal	13 (52)	8 (32)	2 (8)	1 (4)	0 (0)	1 (4)
Parkinsonism	0 (0)	0 (0)	3 (23.1)	9 (69.2)	0 (0)	1 (7.7)
Dystonia	5 (22.7)	0 (0)	5 (22.7)	3 (13.6)	3 (13.6)	6 (27.3)
Myoclonus	0 (0)	1 (6.7)	5 (33.3)	1 (6.7)	5 (33.3)	3 (20)
Other	0 (0)	2 (9.1)	3 (13.6)	1 (4.5)	0 (0)	16 (72.7)
Seizure type N (%)							χ 2(20) = 64.5 p < .001
Focal	3 (9.1)	8 (24.2)	2 (6.1)	3 (9.1)	7 (21.2)	10 (30.3)
Generalized	2 (28.6)	1 (14.3)	0 (0)	2 (28.6)	0 (0)	2 (28.6)
Absences	8 (72.7)	1 (9.1)	1 (9.1)	0 (0)	0 (0)	1 (9.1)
Spasms	0 (0)	1 (6.7)	2 (13.3)	5 (33.3)	1 (6.7)	6 (40)
Multiple	5 (16.1)	0 (0)	13 (41.9)	5 (16.1)	0 (0)	8 (25.8)
DEE N (%)							χ 2(5) = 28.4 p < .001
No	18 (31)	8 (13.9)	7 (21.1)	8 (13.8)	0 (0)	17 (29.3)
Yes	0 (0)	3 (7.7)	11 (28.2)	7 (17.9)	8 (20.5)	10 (25.6)
Status N (%)							χ 2(15) = 56.4 p < .001
No	16 (21.6)	11 (14.9)	14 (18.9)	10 (13.5)	0 (0)	23 (31.1)
Epilepticus	0 (0)	0 (0)	0 (0)	4 (80)	1 (20)	0 (0)
Dystonicus	1 (6.7)	0 (0)	4 (26.7)	1 (6.7)	5 (33.3)	4 (26.7)
Both	1 (33.3)	0 (0)	0 (0)	0 (0)	2 (66.7)	0 (0)
DRE N (%)							χ 2(5) = 17.5 p = .004
No	10 (24.4)	6 (14.6)	2 (4.9)	5 (12.2)	1 (2.4)	17 (41.5)
Yes	7 (13)	5 (9.3)	16 (29.6)	9 (16.7)	7 (13)	10 (18.5)
ASM response N (%)							χ 2(5) = 21.3 p = .004
No	10 (23.8)	6 (14.3)	2 (4.8)	6 (14.3)	1 (2.4)	17 (40.5)
Yes	7 (13.2)	5 (9.4)	16 (30.2)	8 (15.1)	7 (13.2)	10 (18.9)
IQ N (%)							χ 2(15) = 47.3 p < .001
Normal	5 (22.7)	6 (27.3)	5 (22.7)	1 (4.5)	0 (0)	5 (22.7)
Mild	6 (54.5)	1 (9.1)	0 (0)	0 (0)	0 (0)	4 (36.4)
Moderate	4 (18.2)	2 (9.1)	8 (36.4)	6 (27.3)	0 (0)	2 (9.1)
Severe	2 (5)	2 (5)	5 (12.5)	7 (17.5)	8 (20)	16 (40)

Note: Statistical test: Kruskal‐Wallis test for continuous data and chi‐square test for categorical data.

Abbreviations: ASM, anti‐seizure medications; DEE, developmental and epileptic encephalopathy; DRE, drug‐resistant epilepsy; IQ, intelligence quotient; MD, movement disorder.

A logistic regression analysis was performed to ascertain the effects of age at onset, first symptom type, MD type, seizure type, and mechanism of DEE presence. The logistic regression model was statistically significant (χ ²(5) =33.50, p < .001). The model explained 39.5% (Nagelkerke R2) of the variance and correctly classified 76.3% of cases. The Hosmer–Lemeshow goodness‐of‐fit test indicated an acceptable fit of the model (χ ²(8) = 5.13, p = .743). The results indicated that DEE presence was significantly associated with etiology (odds ratio [OR] = 1.68, 95% confidence interval [CI]: 1.14–2.46, p = .008) and onset age group (OR = .14, 95% CI: .05–.36, p < .001). Specifically, the presence of DEE correlated with an early onset (χ ²(2) = 18.14, p < .001) and presence of channelopathies (χ ²(5) = 28.41, p < .001).

Another logistic regression analysis was performed to ascertain the effects of onset age, first symptom type, MD type, seizure type, mechanism, and DEE on IQ. The logistic regression model was statistically significant (χ ²(6) = 43.97, p < .001). The model explained 51.1% (Nagelkerke R2) of the variance and correctly classified 77.9% of cases. The Hosmer–Lemeshow goodness‐of‐fit test indicated that the model fit the data well (χ ²(7) = 7.80, p = .351). The results indicated that IQ was significantly associated only with the presence of DEE (OR = 32.11, 95% CI: 3.79–271.94, p = .001), which was correlated with a lower score (χ ²(1) = 15.77, p < .001).

Finally, a logistic regression analysis was performed to ascertain the effects of onset age, first symptom type, MD type, seizure type, mechanism, and DEE on ASM response. The logistic regression model was statistically significant (χ ²(6) = 33.74, p < .001). The model explained 40.1% (Nagelkerke R2) of the variance and correctly classified 76.8% of cases. The Hosmer–Lemeshow goodness‐of‐fit test indicated an acceptable fit of the model (χ ²(8) = 9.38, p = .312). The results indicated that ASM response was significantly associated with the presence of DEE (OR = 11.82, 95% CI: 3.21–43.55, p < .001), etiology (OR = .66, 95% CI: .44–.99, p = .045), and with the first symptom type (OR = 2.87, 95% CI: 1.27–6.46, p = .011). Drug resistance was associated with presence of DEE (χ ²(1) = 18.50, p < .001), epilepsy or developmental delay as first symptom (χ ²(2) = 14.15, p < .001), and presence of channelopathy (χ ²(5) = 17.07, p = .004).

A two‐step cluster analysis was carried out to identify distinct clinical groups based on age at onset, etiology, first symptom type, and IQ (Figure 4). The cluster analysis identified three clusters with distinctive features: the first cluster grouped 38 patients (40.0%), the second 31 patients (32.6%), and the third 26 patients (27.4%). Patients in Cluster 1 were characterized predominantly by an age at onset below 1 year. Epilepsy was usually the first symptom, most frequently presenting with focal seizures, absence seizures, or other seizure types. This clinical presentation was commonly associated with transportopathies. These patients had normal IQ or mild intellectual disability, and no DEE; most of them were drug sensitive, presenting with paroxysmal dyskinesia. In Cluster 2, all patients had early onset with epilepsy or developmental delay as first symptom and moderate or severe ID; the most represented etiologies were channelopathies, intracellular trafficking, or other, and most of the patients had DEE. The predominant types of chronic MDs were dystonia and myoclonus, seizures were mostly focal type or spasms, and almost all patients were drug‐resistant. Most of the patients included in Cluster 3 had onset between 1 and 5 years of age and moderate or severe ID. In most cases the first symptom was developmental delay followed by MD; the most frequent etiologic categories were metabolic or other, and a large proportion had no DEE. Most patients with parkinsonism were grouped in this cluster, whereas the predominant epileptic seizures were focal or other types. Around half of the patients were drug‐sensitive.

FIGURE 4.

Clinical characteristics of patients in the three clusters identified by cluster analysis. Statistical analysis of the differences in clinical variables between groups revealed significant differences in onset age (χ ²(4) = 82.22, p < .001), first symptom (χ ²(4) = 21,24, p < .001), predominant movement disorder (χ ²(8) = 42,47, p < .001), epilepsy seizure type (χ ²(8) = 31,35, p < .001), etiology (χ ²(10) = 68,88, p < .001), DEE (χ ²(2) = 57.83, p < .001), and intelligence quotient (IQ; χ ²(2) = 72,61, p < .001).

4. DISCUSSION

This study presents a comprehensive analysis of a large cohort of patients with genetically defined syndromes characterized by pediatric‐onset EPIMDs. The broad genetic landscape observed in our cohort, with 46 different genes and 6 pathogenic CNVs identified, reflects the growing complexity of genotype–phenotype correlations in these disorders. Our findings confirm significant heterogeneity in clinical presentations, genetic etiologies, and disease trajectories. This is in line with recent studies, ⁴ , ²⁹ , ³⁰ , ³¹ which have identified a large number of genes and patho‐mechanisms associated with epilepsy and DEES with co‐occurring MDs. In our cohort, only a few genes exhibited distinctive phenotypes in terms of seizure type, MD type, and symptom triggers. This includes SLC2A1, associated with paroxysmal dyskinesia and seizures triggered by fasting; SCN1A, linked to febrile‐induced seizures and myoclonus; and PRRT2, characterized by seizures and PKD, both responsive to sodium channel blockers. In most cases, we observed significant clinical heterogeneity in seizure and MD types and courses, as well as in associated neurodevelopmental disorders. This heterogeneity implies that, with rare exceptions, a molecular diagnosis can only be achieved through advanced NGS approaches such as multigene panels or whole‐exome/genome sequencing. ³² A notable finding of our study is that most patients with EPIMD present initially with epilepsy or developmental delay, with MDs typically emerging later, except in rare conditions associated with GNAO1 and TBC1D24 variants. This observation aligns with recent studies ³⁰ , ³¹ and is likely attributable to the high prevalence of early‐onset DEEs in both our cohort and previously reported series. The presence of several patients with early‐onset DEEs explains the high frequency of spasm seizures and multiple seizure types observed. In addition, the presence of DEEs emerged as the most significant predictor of poor cognitive outcomes and treatment resistance. On the other hand, the early onset age appeared as a predictor of DEEs. Channelopathy, the most represented etiology, was identified as predictor of the presence of DEEs and treatment resistance. In our cohort, paroxysmal dyskinesias, dystonia, and myoclonus were the most frequently observed MDs. Although we did not identify a clear association between MD type and age at onset, dystonia and myoclonus tend to occur earlier compared to other MDs. The pleiotropy and heterogeneity of the genetic background of EPIMDs are also evident from the various genetic dysfunction mechanisms observed. Although transportopathies, channelopathies, and metabolic/degenerative disorders were the most represented etiologies, we observed that 27% of patients harbored variants in genes with unknown or complex mechanisms, such as those involved in cell cycle regulation (e.g., FOXG1, MECP2), cellular degradation or turnover (e.g., UBE3A), and large DNA rearrangements detected by aCGH. Another goal of our study was to delineate potential associations between EPIMD presentations and their underlying etiologies. Notably, we identified key correlations between age at symptom onset, seizure characteristics, predominant movement disorder type, and underlying genetic mechanisms. The cluster analysis revealed three distinct subgroups. Patients in Cluster 1 (early‐onset epilepsy and paroxysmal MDs) often had transportopathies, relatively preserved cognitive function, and better treatment outcomes. These findings align with previous reports associating transportopathies with better cognitive outcomes and favorable treatment responses, namely dietary modifications in SLC2A1‐related disorders, ²³ , ³³ highlighting the importance of early and accurate genetic diagnosis for personalized treatment. Patients in Cluster 2 (early‐onset DEE with dystonia/myoclonus) were associated primarily with channelopathies and intracellular trafficking defects. Most patients experienced early‐onset seizures, severe developmental delay, and poor response to ASMs. Dystonia and myoclonus were the predominant MDs, aligning with previous studies on early‐onset DEEs with associated MDs. ²³ This finding also supports the results of Van der Veen et al., ³¹ who reported a predominance of sodium and potassium channelopathies in a series of children affected by DEE and MDs. Patients in Cluster 3 (developmental delay and parkinsonism) typically had metabolic etiologies. Symptoms emerged between 1 and 5 years, with developmental delay often preceding seizures and MDs. Parkinsonism was the predominant MD, and cognitive impairment was generally severe, consistent with previous reports. ⁶ , ³⁴

5. CONCLUSIONS

Our study highlights the increasing role of NGS in identifying genetic causes of EPIMDs, particularly in patients with atypical presentations. The aCGH analysis should not be overlooked, as it plays a significant role in achieving a diagnosis in a noteworthy percentage of cases. Timely genetic diagnosis can significantly influence clinical management. In our cohort, patients with well‐defined genetic etiologies often benefited from early, targeted interventions. Despite robust findings, our study has limitations. It is a retrospective analysis from a single center, which may limit the generalizability of our results. Moreover, as a tertiary referral center, there is a potential selection bias toward individuals with more severe phenotypes. Future research should focus on refining genotype–phenotype correlations, exploring new genetic mechanisms underlying EPIMDs, and developing personalized therapeutic strategies based on molecular pathophysiology. International collaborations and data‐sharing initiatives could further enhance our understanding of these rare disorders.

AUTHOR CONTRIBUTIONS

Davide Caputo and Roberta Solazzi contributed to conceptualization and original draft preparation. Elena Freri, Francesca Ragona, Federica Zibordi, Shari Gandelli, and Laura Canafoglia were involved in patient recruitment, characterization, and data curation. Barbara Castellotti, Cinzia Gellera, Celeste Panteghini, and Francesca Sciacca performed genetic analysis. Elisa Visani conducted statistical analysis and contributed to original draft preparation. Nardo Nardocci, Giovanna Zorzi, Silvana Franceschetti, and Tiziana Granata provided supervision and were involved in review and editing. All authors reviewed and approved the final version of the manuscript.

FUNDING INFORMATION

The present work was supported by the Italian Ministry of Health Project Ricerca Finalizzata RF‐2019‐12 370 491 to B.C.

CONFLICT OF INTEREST STATEMENT

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Caputo D, Solazzi R, Castellotti B, Panteghini C, Sciacca FL, Visani E, et al. Genetic complexity in pediatric onset epilepsy‐movement disorder syndromes: Insights from a cohort of 97 subjects. Epilepsia. 2026;67:299–314. 10.1111/epi.18669

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.