Seizures and Epilepsy after Ischemic stroke: A Systematic Review of Prognostic Studies
Natanael Duarte1, Eugenio Roman Cruz2, Mauricio Megchún Hernández3, Miguel A. Arce-Huamani4, Kevin Quizhpi-Urgilez5, Camila Vásquez-Avila6, Oriana Rivera-Lozada7, Cesar Bonilla-Asalde8, Joshuan J. Barboza9
1Holy Name Medical Center, 2Universidad Católica del Cibao, 3Hospital de Especialidades Pediátricas en Tuxtla Gutiérrez, 4Programa de Medicina Humana, Facultad de Ciencias de la Salud, Universidad Privada Norbert Wiener, 5Unidad Académica de Salud y Bienestar, Universidad Católica de Cuenca, 6Departamento de investigación, Universidad Católica de Cuenca, 7Vicerrectorado de Investigación, Universidad Señor de Sipán, Chiclayo, 8Escuela de Postgrado, Universidad Señor de Sipán, Chiclayo, 9Escuela de Medicina, Universidad Señor de Sipán, Chiclayo
Objective:
To systematically review and meta-analyze prognostic factors and predictive models for seizures and epilepsy following ischemic stroke.
Background:
Ischemic stroke is the leading cause of acquired epilepsy in adults. Post-stroke seizures and post-stroke epilepsy (PSE) significantly worsen morbidity, mortality, and quality of life, yet prognostic models remain inconsistently validated.
Design/Methods:
A systematic search was conducted in PubMed, Embase, Web of Science, and Scopus through August 2025. Eligible studies included observational cohorts and case–control designs evaluating adult patients with ischemic stroke and reporting seizures or epilepsy as outcomes. Data extraction followed the CHARMS framework, and risk of bias was assessed with PROBAST. Random-effects meta-analyses with Hartung–Knapp adjustment were performed for prognostic factors and model discrimination, synthesizing odds ratios (OR) and area under the curve (AUC) values.
Results:
Eight studies met inclusion criteria, comprising multicenter registries and prospective cohorts. Several prognostic factors were consistently associated with PSE, including early seizures within 7 days (OR 4.04, 95% CI 1.80–9.05), higher NIHSS scores (OR 4.01, 95% CI 2.12–7.59), and biomarker abnormalities such as elevated Endostatin, reduced Hsc70, and low TNF-R1. Electroencephalographic alterations, including epileptiform activity (OR 2.00, 95% CI 1.32–3.04), also conferred increased risk. The meta-analysis of model discrimination showed a pooled AUC of 0.739 (95% CI 0.703–0.773), with negligible heterogeneity (I² = 0%).
Conclusions:
Prognostic evidence supports a multidimensional approach combining clinical severity, acute seizures, EEG abnormalities, and biomarkers for risk stratification. Current models demonstrate moderate predictive performance, highlighting the need for external validation and refinement through multimodal and machine learning approaches to improve personalized risk prediction in stroke survivors.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.