Abstracts

Rationale: Variants in the SCN1A gene cause a wide range of epilepsy syndromes ranging from the milder genetic epilepsy with febrile seizures plus (GEFS+) to the severe Dravet syndrome (DS). The majority of variants in SCN1A-related epilepsies are de novo and when an individual presents with febrile seizures it is difficult to determine what phenotype they will develop. Our study aimed to investigate relationships between genotype and phenotype through examination of variant characteristics, clinical data and in-silico prediction tools.

Methods: We assessed the clinical and genetic data of a retrospective cohort of 1022 individuals with epilepsy attributed to SCN1A variants. We explored how variant type and position, in-silico (CADD and REVEL) and physico-chemical (Grantham) variant interpretation scores and seizure type related to the disease phenotype and seizure age of onset.

Results: DS has earlier median onset than GEFS+ (5.4 vs 12.0 months, p< 0.001). Truncating variants were associated with earlier median onset than missense variants (5.0 vs 6.0 months, p< 0.001). Mean Grantham (95.4 vs 72.5) and in-silico variant scores (CADD [30.8 vs 27.1], REVEL [0.91 vs 0.85]) were higher in DS vs GEFS+ phenotypes (p< 0.001) and were negatively correlated with age of onset (Grantham [-0.111, p=0.015], CADD [-0.169, p< 0.001] and REVEL [-0.197, p< 0.001]). Prescence of status epilepticus as first seizure had a 95.2% (76.2-99.9) specificity, 32.7% sensitivity (27.7-38.6), 98.9% (94.1-100) positive predictive value and 9.7% (6-14.5) negative predictive value as a diagnostic feature of DS vs GEFS+. Missense variants observably clustered in voltage sensor and pore-lining coding regions (S4, S5, S5-6 linker and S6). Missense variants in these regions had earlier median seizure onset than those found elsewhere in the gene (6.0 vs 7.0 months p< 0.001). Truncating variants associated with GEFS+ were more likely to be found in N-terminus or exon 26 coding regions than elsewhere in the gene (10.6% vs 0.8%, p< 0.001). We identified 14 identical variants that were found in 5 or more different patients often originating from the same family. Identical variant carriers exhibited less variability in age of onset compared to non-identical variant carriers (1.88 vs 2.9 months SD, p=0.001).

Conclusions: These findings strengthen and add to current understanding of genotype-phenotype associations in SCN1A-related epilepsies. Variant type and position are reflected in clinical presentation. In-silico and physico-chemical variant scores are significantly related to onset and phenotype but cannot be applied to individual variants. The identification of highly specific early disease features such as status epilepticus aids early diagnosis.

Funding: Please list any funding that was received in support of this abstract.: No funding.