Abstracts

Rationale: Non convulsive seizures in critically ill children are often challenging to identify and are associated with increased mortality and morbidity. Continuous EEG (CEEG) is the simultaneous recording of EEG and video of patient’s clinical behavior over extended periods of time for surveillance of electrographic brain activity and seizures. Analyzing such large volumes of EEG data requires years of formal training and is highly labor and time intensive, which can potentially delay identification of subclinical seizures. Quantitative EEG (QEEG) uses mathematical and analytical algorithms to transform and compress hours of raw EEG data into a simpler graphical representation which can be easily comprehended and analyzed by a bed-side, non neurophysiologist, health care provider for faster identification of non convulsive seizures thus expediting intervention. Methods: A single-centre, double-blinded observational study was carried out at the McMaster Children’s Hospital from August 2018 to April 2019, recruiting general pediatric residents and critical care nursing staff who have not previously received a formal training in EEG interpretation. Participants were trained through a 15 minute power point presentation to analyze and detect seizures utilizing 4 QEEG trends (amplitude integrated EEG, rhythmicity spectrogram, seizure detection and seizure probability markers). Each participant received an online test comprising of 45 slides of one hour QEEG epochs carrying a total of 186 seizures. QEEG data was retrieved retrospectively from patients aged 1 month to 18 years who have qualified for continuous EEG monitoring at McMaster PICU from June 2014 to November 2019. Neonatal EEGs, EEGs with excessive artefact and EEGs of less than 4 hours duration were excluded. Results: Ten participants (5 pediatric residents and 5 registered ICU nurses) completed the test with an overall sensitivity of 91.7% and a specificity of 87%. The overall positive predictive value was 86% and negative predictive value was 92%. Conclusions: Non-neurophysiologist health care providers of critically ill children who are at risk of sub clinical seizures can detect seizures with a reasonable accuracy utilizing a panel of quantitative EEG trends following receiving a basic training. This may enable detection of electrographic seizures in such predisposed individuals in a timely manner leading to expedited intervention and potentially better outcomes. Funding: McMaster University - Pediatric Neurology resident research award