A Bedside Screening Tool for Acute Intracranial Hemorrhages in Intubated Children using Continuous Quantitative Electroencephalography Monitoring

Abstract

Abstract Significance Recognition of intracranial hemorrhage is challenging in children who require deep sedation to tolerate mechanical ventilation. The Correlate Of Injury to the Nervous System (COIN) index may enable real-time recognition of intracranial hemorrhage at bedside. Methods Retrospective analysis of electroencephalography (EEG) data from children with spontaneous intracranial hemorrhage while intubated and sedated in the pediatric intensive care unit. Patients were selected for having normal head imaging at time of EEG start and required demonstration of hemorrhage on repeat imaging following an uninterrupted period of EEG recording. Power spectrum data were analyzed to yield a COIN value and visualization for every 4 seconds of recording. EEG recordings were subdivided based on COIN-risk alarm states (low, medium, or high). Changes in COIN were compared with changes in commercially available quantitative EEG trending software. COIN values for each subdivision were compared within cases using the Wilcoxon Rank-Sum Test. Results Two children developed spontaneous intracranial hemorrhage while intubated. COIN shows transitions from low-to-medium (p < 0.001) and medium-to-high-risk (p < 0.001 in both cases) alarm states. Discrete transitions in COIN alarm state preceded clinical recognition of hemorrhage by several hours. COIN visualized focal power attenuation concordant with hemorrhage localization. In both cases, qualitative EEG was not reported to have focal abnormalities during the medium-risk alarm state. Conclusion COIN may assist in real-time recognition of intracranial hemorrhage in children at bedside. Further study and development are required for clinical implementation of COIN in several clinical settings where patients are at high risk of new or worsening intracranial hemorrhage.