Etiological associations and outcome predictors of acute electroencephalography in childhood encephalitis

Objectives: To examine EEG features in a retrospective 13-year cohort of children with encephalitis. Methods: 354 EEGs from 119 patients during their admission were rated blind using a proforma with demonstrated inter-rater reliability (mean k = 0.78). Patients belonged to 12 etiological groups that could be grouped into infectious and infection-associated (n = 47), immune-mediated (n = 36) and unknown (n = 33). EEG features were analyzed between groups and for risk of abnormal Liverpool Outcome Score and drug resistant epilepsy (DRE) at last follow up. Results: 86% children had an abnormal first EEG and 89% had at least one abnormal EEG. 55% had an abnormal outcome, and 13% had DRE after median follow-up of 7.3 years (2.0-15.8 years). Reactive background on first EEGs (9/11, p = 0.04) and extreme spindles (4/11, p < 0.001) distinguished patients with anti-N-Methyl-D-Aspartate Receptor encephalitis. Non-reactive EEG background (48% first EEGs) predicted abnormal outcome (OR 3.8, p < 0.001). A shifting focal seizure pattern, seen in FIRES (4/5), anti-voltage gated potassium channel (2/3), Mycoplasma (1/10), other viral (1/10) and other unknown (1/28) encephalitis, was most predictive of DRE after multivariable analysis (OR 11.9, p < 0.001). Conclusions: Non-reactive EEG background and the presence of shifting focal seizures resembling migrating partial seizures of infancy are predictors of abnormal outcome and DRE respectively in childhood encephalitis. Significance: EEG is a sensitive but non-discriminatory marker of childhood encephalitis. We highlight the EEG features that predict abnormal outcome and DRE. (C) 2016 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.