Katri Silvennoinen

Genetic testing performed to identify the underlying etiology of epilepsy has become increasingly common and is now being recommended as part of the presurgical evaluation for epilepsy surgery. This study aimed to characterize the types of genetic tests performed in patients evaluated for epilepsy surgery and assess how genetic findings have influenced clinical treatment decisions and outcomes during the last decade in a tertiary epilepsy center. Pediatric and adult epilepsy patients treated at the tertiary Kuopio Epilepsy Center in Finland are given the opportunity to participate in an EpiBiomarker study, which aims to identify epilepsy-related biomarkers and has established a comprehensive prospective database encompassing detailed clinical information including data of the possible preoperative evaluation and genetic testing results. We identified 448 patients who (or their guardian) had provided consent to the Epi Biomarker study out of the 563 patients who had been evaluated for epilepsy surgery between January 1, 2015, and December 31, 2024. Altogether 117/448 patients (26.1%) had undergone some form of genetic testing before or during their presurgical evaluation. Clinical preoperative genetic testing became increasingly common in the later years of the study, ranging from 6.1% of patients evaluated in 2016 to 40.5% in 2024. Genetic etiology was identified in 8.0% of patients (36/448) and the identified genetic etiologies increased from 1.1% (2015) to 13.5% (2024). Only 4/36 (11.1%) of these genetically diagnosed patients underwent resective epilepsy surgery and only two of these patients were completely seizure-free at 1-year post-operative follow-up. Germline genetic testing is increasingly incorporated into the presurgical evaluation of patients with epilepsy, leading to a higher detection of underlying genetic etiologies. Improved recognition of these etiologies may enhance clinical decision-making by providing additional context for treatment planning and prognostic assessment, while also helping to reduce unnecessary investigations in selected cases.

Objective: Variants in GABRG2 , encoding the γ2 subunit of the γ‐aminobutyric acid type A receptor, are linked to epilepsy phenotypes of varying severity, with gain‐of‐function (GoF) variants associated with the more severe phenotypes than loss‐of‐function variants. Here, we provide a comprehensive analysis of the early clinical features, motor, language, and eating abilities, and structural consequences of the recurrent GABRG2 p.(Ala106Thr) GoF variant, aiming to refine genotype–phenotype correlations and deepen the understanding of GoF‐associated GABRG2 disorders. Methods: Individuals were recruited through international collaborations, literature review, and patient advocacy groups. Clinical data were collected via standardized interviews and physician reports. Functional effects were assessed using electrophysiological recordings. Structural modeling was performed using homology‐based approaches to evaluate conformational changes. Results: We collected 23 unrelated individuals harboring the GABRG2 p.(Ala106Thr) variant. Symptoms began with early dysphagia (median onset age = 1 month), hypotonia, seizures (median onset age = 2 months), and hyperkinetic movement disorders (median onset age = 3 months). Epilepsy was present in 91% of the individuals, 71% featuring a developmental and epileptic encephalopathy; 33% of the individuals with epilepsy achieved seizure freedom during the follow‐up. All individuals exhibited moderate to severe cognitive/neurodevelopmental impairment, with profound deficits in motor, language, and other domains. Functional analysis confirmed a GoF effect of the p.(Ala106Thr) variant. By contrast, another variant affecting the same residue, p.(Ala106Pro), which is observed in population databases, was functionally characterized as neutral. Structural modeling indicated that the p.(Ala106Thr) variant may enhance receptor gating by facilitating a hydrogen bond between the extracellular and transmembrane domains, a mechanism not observed with the p.(Ala106Pro) variant. Significance: Although some individuals exhibited severe developmental impairment without epilepsy, our findings demonstrate that the recurrent GABRG2 p.(Ala106Thr) GoF variant is consistently associated with a severe neurodevelopmental phenotype. Despite its profound clinical impact, in silico tools consistently predict this variant as benign, highlighting a critical gap in current predictive models.

Although a striking female preponderance has been consistently reported in epilepsy with eyelid myoclonia (EEM), no study has specifically explored the variability of clinical presentation according to sex in this syndrome. Here, we aimed to investigate sex-specific electroclinical differences and prognostic determinants in EEM. Data from 267 EEM patients were retrospectively analyzed by the EEM Study Group, and a dedicated multivariable logistic regression analysis was developed separately for each sex. We found that females with EEM showed a significantly higher rate of persistence of photosensitivity and eye closure sensitivity at the last visit, along with a higher prevalence of migraine with/without aura, whereas males with EEM presented a higher rate of borderline intellectual functioning/intellectual disability. In female patients, multivariable logistic regression analysis revealed age at epilepsy onset, eyelid myoclonia status epilepticus, psychiatric comorbidities, and catamenial seizures as significant predictors of drug resistance. In male patients, a history of febrile seizures was the only predictor of drug resistance. Hence, our study reveals sex-specific differences in terms of both electroclinical features and prognostic factors. Our findings support the importance of a sex-based personalized approach in epilepsy care and research, especially in genetic generalized epilepsies.

Dravet syndrome is an archetypal rare severe epilepsy, considered ‘monogenic’, typically caused by loss-of-function SCN1A variants. Despite a recognizable core phenotype, its marked phenotypic heterogeneity is incompletely explained by differences in the causal SCN1A variant or clinical factors. In 34 adults with SCN1A-related Dravet syndrome, we show additional genomic variation beyond SCN1A contributes to phenotype and its diversity, with an excess of rare variants in epilepsy-related genes as a set and examples of blended phenotypes, including one individual with an ultra-rare DEPDC5 variant and focal cortical dysplasia. The polygenic risk score for intelligence was lower, and for longevity, higher, in Dravet syndrome than in epilepsy controls. The causal, major-effect, SCN1A variant may need to act against a broadly compromised genomic background to generate the full Dravet syndrome phenotype, whilst genomic resilience may help to ameliorate the risk of premature mortality in adult Dravet syndrome survivors.

Objective: Progressive myoclonic epilepsy type 1 (EPM1) is caused by biallelic alterations in the CSTB gene, most commonly dodecamer repeat expansions. Although transcranial magnetic stimulation (TMS)–induced long‐interval intracortical inhibition (LICI) was previously reported to be normal in EPM1, short‐interval intracortical inhibition (SICI) was reduced. We explored the association between these measures and the clinical and genetic features in a separate group of patients with EPM1.Methods: TMS combined with electromyography was performed under neuronavigation. LICI was induced with an inter‐stimulus interval (ISI) of 100 ms, and SICI with ISIs of 2 and 3 ms, and their means (mSICIs) were expressed as the ratio of conditioned to unconditioned stimuli. LICI and mSICI were compared between patients and controls. Nonparametric correlation was used to study the association between inhibition and parameters of clinical severity, including the Unified Myoclonus Rating Scale (UMRS); among patients with EPM1 due to biallelic expansion repeats, also the association with the number of repeats was assessed.Results: The study protocol was completed in 19 patients (15 with biallelic expansion repeats and 4 compound heterozygotes), and 7 healthy, age‐ and sex‐matched control participants. Compared to controls, patients demonstrated significantly less SICI (median mSICI ratio 1.18 vs 0.38; p < .001). Neither LICI nor SICI was associated with parameters of clinical severity. In participants with biallelic repeat expansions, the number of repeats in the more affected allele (greater repeat number [GRN]) correlated with LICI (rho = 0.872; p < .001) and SICI (rho = 0.689; p = .006).Significance: Our results strengthen the finding of deranged γ‐aminobutyric acid (GABA)ergic inhibition in EPM1. LICI and SICI may have use as markers of GABAergic impairment in future trials of disease‐modifying treatment in this condition. Whether a higher number of expansion repeats leads to greater GABAergic impairment warrants further study.

Objective: Epilepsy with eyelid myoclonia (EEM) has been associated with marked clinical heterogeneity. Early epilepsy onset has been recently linked to lower chances of achieving sustained remission and to a less favorable neuropsychiatric outcome. However, much work is still needed to better delineate this epilepsy syndrome.Methods: In this multicenter retrospective cohort study, we included 267 EEM patients from nine countries. Data on electroclinical and demographic features, intellectual functioning, migraine with or without aura, family history of epilepsy, and epilepsy syndromes in relatives were collected in each patient. The impact of age at epilepsy onset (AEO) on EEM clinical features was investigated, along with the distinctive clinical characteristics of patients showing sporadic myoclonia involving body regions other than eyelids (body‐MYO).Results: Kernel density estimation revealed a trimodal distribution of AEO, and Fisher–Jenks optimization disclosed three EEM subgroups: early onset (EO‐EEM), intermediate onset (IO‐EEM), and late onset (LO‐EEM). EO‐EEM was associated with the highest rate of intellectual disability, antiseizure medication refractoriness, and psychiatric comorbidities and with the lowest rate of family history of epilepsy. LO‐EEM was associated with the highest proportion of body‐MYO and generalized tonic–clonic seizures (GTCS), whereas IO‐EEM had the lowest observed rate of additional findings. A family history of EEM was significantly more frequent in IO‐EEM and LO‐EEM compared with EO‐EEM. In the subset of patients with body‐MYO (58/267), we observed a significantly higher rate of migraine and GTCS but no relevant differences in other electroclinical features and seizure outcome.Significance: Based on AEO, we identified consistent EEM subtypes characterized by distinct electroclinical and familial features. Our observations shed new light on the spectrum of clinical features of this generalized epilepsy syndrome and may help clinicians toward a more accurate classification and prognostic profiling of EEM patients.

Background Cannabidiol is a newly-available treatment option for seizures in Dravet syndrome (DS). Methods Adults in a cannabidiol early access programme at two epilepsy centres were included. Can- nabidiol was titrated up to 20mg/kg, as tolerated. The primary outcome measure was >30% reduction in convulsive seizures assessed at three months of treatment. Secondary outcome measures included later reduction in seizures and incidence of adverse drug reactions (ADRs). Results Eighteen individuals (seven female; median age 27.5 years) with SCN1A mutation-positive DS took part. Median follow-up was 166 days; nine individuals continued on maintenance treatment. Median maximum dose was 18.6mg/kg/day. Four individuals (22.2%) had a >30% reduction in convulsive seizures by three months; in two, this became evident after three months. ADRs occurred in 14 (77.8%). Commonest ADRs were somnolence and derangement of LFTs (all 18 took concomitant valproate), occurring in eight and seven individuals, respectively. ADRs led to discontinuation in three, including a seizure-free individual. Twelve people (66.7%) required dosage changes to other antiepileptic drugs. Conclusion Cannabidiol was effective in a fifth of patients. Two additional patients showed later reduction in seizure frequency, suggesting longer follow-up is warranted. Treatment was complicated by ADRs. Dose adjustments of other antiepileptic drugs may be needed. k.silvennoinen@ucl.ac.uk

Mesial temporal lobe epilepsy with hippocampal sclerosis and a history of febrile seizures is associated with common variation at rs7587026, located in the promoter region of SCN1A. We sought to explore possible underlying mechanisms. SCN1A expression was analysed in hippocampal biopsy specimens of individuals with mesial temporal lobe epilepsy with hippocampal sclerosis who underwent surgical treatment, and hippocampal neuronal cell loss was quantitatively assessed using immunohistochemistry. In healthy individuals, hippocampal volume was measured using MRI. Analyses were performed stratified by rs7587026 type. To study the functional consequences of increased SCN1A expression, we generated, using transposon-mediated bacterial artificial chromosome transgenesis, a zebrafish line expressing exogenous scn1a, and performed EEG analysis on larval optic tecta at 4 day post-fertilization. Finally, we used an in vitro promoter analysis to study whether the genetic motif containing rs7587026 influences promoter activity. Hippocampal SCN1A expression differed by rs7587026 genotype (Kruskal–Wallis test P = 0.004). Individuals homozygous for the minor allele showed significantly increased expression compared to those homozygous for the major allele (Dunn’s test P = 0.003), and to heterozygotes (Dunn’s test P = 0.035). No statistically significant differences in hippocampal neuronal cell loss were observed between the three genotypes. Among 597 healthy participants, individuals homozygous for the minor allele at rs7587026 displayed significantly reduced mean hippocampal volume compared to major allele homozygotes (Cohen’s D = − 0.28, P = 0.02), and to heterozygotes (Cohen’s D = − 0.36, P = 0.009). Compared to wild type, scn1lab-overexpressing zebrafish larvae exhibited more frequent spontaneous seizures [one-way ANOVA F(4,54) = 6.95 (P < 0.001)]. The number of EEG discharges correlated with the level of scn1lab overexpression [one-way ANOVA F(4,15) = 10.75 (P < 0.001]. Finally, we showed that a 50 bp promoter motif containing rs7587026 exerts a strong regulatory role on SCN1A expression, though we could not directly link this to rs7587026 itself. Our results develop the mechanistic link between rs7587026 and mesial temporal lobe epilepsy with hippocampal sclerosis and a history of febrile seizures. Furthermore, we propose that quantitative precision may be important when increasing SCN1A expression in current strategies aiming to treat seizures in conditions involving SCN1A haploinsufficiency, such as Dravet syndrome.

Background and objectives: Eyelid myoclonia with absences (EMA) is a generalized epilepsy syndrome whose prognosis and clinical characteristics are still partially undefined. We investigated electroclinical endophenotypes and long-term seizure outcome in a large cohort of EMA patients. Methods: In this multicenter retrospective study, EMA patients with ≥5 years of follow-up were included. We investigated prognostic patterns and sustained terminal remission (STR), along with their prognostic factors. Moreover, a two-step cluster analysis was used to investigate the presence of distinct EMA endophenotypes. Results: We included 172 patients, with a median age at onset of 7 years (interquartile range (IQR) 5-10) and a median follow-up duration of 14 years (IQR 8.25-23.75). Sixty-six patients (38.4%) displayed a non-remission pattern, whereas remission and relapse patterns were encountered in 56 (32.6%) and 50 (29.1%) subjects. Early epilepsy onset, history of febrile seizures (FS) and eyelid myoclonia (EM) status epilepticus significantly predicted a non-remission pattern according to multinomial logistic regression analysis. STR was achieved by 68 (39.5%) patients with a mean latency of 14.05 years (SD ± 12.47). Early epilepsy onset, psychiatric comorbidities, and a history of FS and generalized tonic-clonic seizures (GTCS) were associated with a lower probability of achieving STR according to a Cox regression proportional hazards model. Antiseizure medication (ASM) withdrawal was attempted in 62/172 patients, and seizures relapsed in 74.2%. Cluster analysis revealed two distinct clusters with 86 patients each. Cluster 2, which we defined as “EMA-plus”, was characterized by an earlier age at epilepsy onset, higher rate of intellectual disability, EM status epilepticus, generalized paroxysmal fast activity, self-induced seizures, FS, and poor ASM response, whereas Cluster 1, the “EMA-only” cluster, was characterized by a higher rate of seizure remission and more favorable neuropsychiatric outcome. Discussion: Early epilepsy onset was the most relevant prognostic factor for poor treatment response. A long latency between epilepsy onset and ASM response was observed, suggesting the impact of age-related brain changes in EMA remission. Finally, our cluster analysis showed a clear-cut distinction of EMA patients into an EMA-plus insidious subphenotype and an EMA-only benign cluster that strongly differed in terms of remission rates and cognitive outcomes.