414 Comparison of Quantitative Measures and Neurophysiologist Assessment Using Cotside EEG Monitors to Predict Outcome in Preterm Infants

Abstract

Background Conventional electroencephalography (EEG) may help predict outcome in preterm infants, but requires specialist interpretation. Some cotside EEG devices provide quantitative analyses assisting non-specialists with interpretation.Aim to compare quantitative measures of continuity with interpretation by a neurophysiologist using early cotside EEG recordings to predict outcome of preterm infants.Methods in infants <32 weeks gestation 60 minute epochs of EEG recorded using research BRM cotside monitors (BrainZ Instruments Ltd, Auckland, New Zealand) were analysed for continuity (percentage of each minute with EEG amplitude >10μV) and categorised as normal (100%, 62% of cohort), moderate (80–99%) or low (<80%, lowest 10% of cohort). The same portion of EEG was graded by a neurophysiologist, blinded to infant outcome, as 'normal' or 'abnormal' using interburst interval and seizure activity. Surviving infants had Bayley-II examinations at 18 months. Outcome was 'poor' if the infant died or had a Mental or Psychomotor Developmental Index (MDI/PDI) <70, 'moderate' if MDI or PDI=70–84, and 'normal' if MDI and PDI>=85.Results Analyses were performed for 29 infants with median (range) gestation at delivery 26(24–31)weeks, birthweight 910(605–1250)grams and postnatal age at EEG 22(2–53)hours. Outcomes were 'poor' in 7 (3 died, 4 had PDI<70), 'moderate' in 8 and 'normal' in 14. Three infants with low continuity had 'poor' outcome. Moderate continuity was found in 2/7 infants with 'poor' outcome, 3/8 infants with 'moderate' outcome and 5/14 with 'normal' outcome. The remainder had normal continuity (Cramer's V=0.43, Chi-squared p=0.03). The neurophysiologist graded traces as abnormal in 6/7 infants with 'poor' outcome, 1/8 with 'moderate' and 0/14 with 'normal' outcome (Cramer's V=0.82, Chi-squared p<0.0001). Categorisation by continuity measures and neurophysiologist assessment were dissimilar, Chi-squared p=0.19.Conclusion Both techniques offer some potential, but continuity measures performed less well than neurophysiologist interpretation in predicting outcome from early cotside EEG recordings. MB is a consultant for BrainZ Instruments Ltd.