Abstracts

Rationale: Tuberous sclerosis complex (TSC) is caused by variants in TSC1 or TSC2. Up to 90% of individuals with TSC develop epilepsy in the first year, with many becoming drug-resistant by age 2. TSC2 mutations often result in more severe disease, but predicting severity and response to early interventions remains challenging due to variability even among TSC2 variants


Methods: We retrospectively reconstructed the epilepsy histories of 54 patients with clinical diagnosis of TSC and known variants in TSC2, capturing seizure types and categorical seizure severities in one-month time increments using the framework championed by the Epilepsy Learning Health System (ELHS) and Pediatric Epilepsy Learning Health System (PELHS). We assessed anti-seizure medication (ASM) effectiveness by comparing seizure improvement intervals during medication use with those without medication. We harmonized longitudinal phenotypic annotations across this cohort, analyzing 4,177 phenotypic annotations across 340 patient years. We analyzed the epilepsy phenotype within two patient sub-groups 1) those with drug-resistant epilepsy (DRE) by 24 months (according to the ILAE definition) versus those with controlled epilepsy, and 2) those with protein truncating variants or deletions (PTV/del) versus those with missense variants


Results: Among 54 patients, 80% developed epilepsy with a median onset age of 4 months. Infantile spasms occurred in 48%, and focal seizures in 78%. By 24 months, 41% of patients with epilepsy met ILAE criteria for DRE. TSC2 variants were distributed across the entire gene: 73% were PTV/del variants, and 26% were missense. Variants were localized to the Hamartin domain (16%), GAP domain (21%), and the region between these domains (62%). The median time to first seizure was 4 months for DRE patients and 5 months for those with controlled epilepsy. The probability of not developing epilepsy by 6 months was 45% for the controlled epilepsy group and 10% for the DRE group, with all DRE patients experiencing seizures by 1 year [χ²(1) = 5.614, p = 0.0178]. Patients with TSC2 PTV/del had earlier seizure onset (median 9 months vs. 15 months for missense variants), but this was not statistically significant. At 24 months, DRE patients were more likely to have severe developmental delay (p< 0.001, OR 50.4), tonic seizures (p< 0.001, OR 9.8), autistic behavior (p< 0.001, OR 6.8), and sleep disturbances (p< 0.01, OR 5.35). PTV/del patients more often had abdominal symptoms (p< 0.01, OR Inf) and SEGA (p< 0.04, OR 4.8) compared to missense variants. ASM effectiveness analysis showed that treatments with Vigabatrin (p< 0.001, OR 1.2), mTORi (p< 0.001, OR 1.31), Valproic Acid (p< 0.001, OR 1.31), and surgical resection (p< 0.001, OR 1.26) were most effective in maintaining seizure freedom