Late Onset Epileptic Spasms: A Hidden Challenge in the Electronic Medical Record
Abstract number :
1.284
Submission category :
4. Clinical Epilepsy / 4A. Classification and Syndromes
Year :
2025
Submission ID :
550
Source :
www.aesnet.org
Presentation date :
12/6/2025 12:00:00 AM
Published date :
Authors :
Presenting Author: Eva Catenaccio, MD – The Children's Hospital of Philadelphia/University of Pennsylvania
Michael Kaufman, MS – Children's Hospital of Philadelphia
Peter Galer, PhD – The Children's Hospital of Philadelphia
Laura Lubbers, PhD – CURE Epilepsy
Elaine O’Loughlin, PhD – CURE Epilepsy
Ingo Helbig, MD – Children's Hospital of Philadelphia
Rationale: Epileptic Spasms with or without hypsarrhythmia that have onset after age 2 years are referred to as late onset epileptic spasms (LOES). LOES are a poorly understood with regards to etiology, clinical characteristics, neurodevelopmental outcomes, and treatment responses. Attempts to quantify the frequency of this syndrome or characterize this population are limited by patient heterogeneity making identification of a patient cohort difficult. Understanding this population is critical to developing new approaches for this difficult-to-treat seizure type.
Methods: We sought to identify patients with LOES using natural language processing to search electronic medical record data from a large pediatric healthcare network over from 10/1/2015 to 5/01/2025 including 1) neurology clinical notes, 2) electroencephalography (EEG) reports, and 3) diagnostic codes (ICD). We limited our query to patients whose first appearance of LOES features occurred after age 2 years but before age 8 years and refined the search criteria based on a manual review of a subset of patients.
Results: Recognizing the limitations of EMR-based identification of rare epilepsy syndromes, we developed a broad search strategy anticipating a high false positive rate. Accordingly, our initial query identified 1319 patients including 524 by ICD, 166 by EEG reports, and 628 by clinical notes. Manual review of 25 patients per category identified 5 patients with confirmed spasm onset after age 2 years. There was a high false positive rate, specifically with regards to determining the age of spasm onset. The main driver of false positive findings was the inability to correctly assess age of spasm onset through various EMR queries. Individuals with true LOES were mainly identified through standardized EEG reports (4/5) and showed a high frequency of genetic etiologies including one patient with lissencephaly, one with MECP2 duplication syndrome, and one with Angelman syndrome. Of the remaining 2 LOES patients, one patient had a focal cortical dysplasia, and one patient had spasms of unknown etiology.
Conclusions: Our findings support that LOES is a rare and overlooked epilepsy phenotype that is challenging to identify through medical record data. Individuals with true LOES can be identified through a combination of search strategies combined with manual clinical review allowing for in depth characterization of this population. The existence of LOES challenges the conventional understanding of infantile spasms and adds to our ability to understand the biology and trajectory of developmental epileptic encephalopathies, potentially paving the way for future treatment avenues.
Funding: This research was supported by a grant from the CURE Epilepsy Foundation.
Clinical Epilepsy