Longitudinal Analysis of Amplitude-Integrated Electroencephalography for Outcome Prediction in Hypoxic-Ischemic Encephalopathy

Méder, Ünőke and Csekő, Anna Judit and Szakacs, Laszlo and Balogh, Dóra Csenge and Szakmár, Enikő and Andorka, Csilla and Kovács, Kata and Dobi, Marianna and Brandt, Ferenc and Szabó, Miklós and Szabó, Attila and Jermendy, Ágnes (2022) Longitudinal Analysis of Amplitude-Integrated Electroencephalography for Outcome Prediction in Hypoxic-Ischemic Encephalopathy. JOURNAL OF PEDIATRICS, 246. 19-25.e5. ISSN 0022-3476

Preview

Text MederetalJPed2022.pdf - Published Version Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

Objective: To investigate the prognostic accuracy of longitudinal analysis of amplitude-integrated electroencephalography (aEEG) background activity to predict long-term neurodevelopmental outcome in neonates with hypoxic-ischemic encephalopathy (HIE) receiving therapeutic hypothermia.Study design This single-center observational study included 149 neonates for derivation and 55 neonates for validation with moderate-severe HIE and of gestational age 35 weeks at a tertiary neonatal intensive care unit. Single-channel aEEG background pattern, sleep-wake cycling, and seizure activity were monitored over 84 hours during therapeutic hypothermia and rewarming, then scored for each 6-hour interval. Neurodevelopmental outcome was assessed using the Bayley Scales of Infant Development, Second Edition. Favorable outcome was defined as having both a Mental Development Index (MDI) score and Psychomotor Development Index (PDI) score >= 70, and adverse outcome was defined as either an MDI or a PDI <70 or death. Regression modeling for longitudinal analysis of repeatedly measured data was applied, and area under the receiver operating characteristic curve (AUC) was calculated.Resuits Longitudinal aEEG background analysis combined with sleep-wake cycling score had excellent predictive value (AUC, 0.90; 95% CI, 0.85-0.95), better than single aEEG scores at any individual time point. The model performed well in the independent validation cohort (AUC, 0.87; 95% CI, 0.62-1.00). The reclassification rate of this model compared with the conventional analysis of aEEG background at 48 hours was 18% (24 patients); 14% (18 patients) were reclassified correctly. Our results were used to develop a user-friendly online outcome prediction tool.Conclusions Longitudinal analysis of aEEG background activity and sleep-wake cycling is a valuable and accurate prognostic tool.

Actions (login required)

Edit Item