Abstract 39: Time to Detect Seizures Among Hemorrhagic and Ischemic Stroke Patients Undergoing Continuous EEG (cEEG) Monitoring

Abstract

Objective: Stroke patient have a 13-19% risk of Electroencephalographic (EEG) seizures and frequently undergo continuous EEG (cEEG) monitoring. It is unclear how long cEEG monitoring should be done in hemorrhagic and ischemic stroke patients to exclude subclinical seizures. Our objective is to identify stroke patients who are at risk of delayed seizure detection (>24 hours) and therefore may benefit from prolonged cEEG monitoring. Methods: We identified all patients who had cEEG monitoring at a single tertiary center in 2016 (n=2425). Data was analyzed using Wilcoxon Ranksum & chi square tests. Results: Among the 2425 patients who had EEG monitoring, 334 had seizures. 25 were excluded because of missing data. Among the 309 patients included, median time to seizure onset was 3 hours. Seizures were detected by 24 hours in 80% (n=248) and within 36 hours of cEEG monitoring in 90 % of all patients with seizures. Patients were divided into ischemic, hemorrhagic and non-vascular etiologies. 19% (n=60) had seizures due to hemorrhage and 13% (n=41) due to ischemic strokes. Among ischemic strokes patients with seizures, seizures were detected by 2 hours 17 minutes in 50%, by 18 hours in 80% and by 34 hours in 90%. However, for hemorrhage patients with seizures, time to detect seizures in 50 % of patients was delayed to 15 hours (p= 0.0004) and seizures were detected by 33 hours in 80% and by 41 hours in 90 % of patients. Hemorrhage patients were more likely to have their first seizure recorded after 24 hours of cEEG monitoring (p=0.0221). Hemorrhages were subdivided into subdural, subarachnoid, parenchymal, epidural and mixed types. Among these, those who had more than one type of hemorrhage were more likely to have their first seizure detected after 24 hours of cEEG monitoring (p=0.0002). Conclusions: Patients with hemorrhage particularly with more than one type of hemorrhage are more likely to have delayed seizure detection and require longer cEEG monitoring (>24 hours) to diagnose seizures.