Automated EEG Background Analysis and 2-Year Outcomes in Neonatal Hypoxic-Ischemic Encephalopathy

Area of Science:

• Neonatal neurology
• Neurophysiology
• Clinical informatics

Background:

• Hypoxic-ischemic encephalopathy (HIE) is a major cause of neonatal mortality and morbidity, even with therapeutic hypothermia.
• Accurate early prognostication is crucial for managing HIE and stratifying patients for clinical trials.

Purpose of the Study:

• To assess the feasibility and predictive accuracy of automated electroencephalogram (EEG) background analysis using the Brain State of the Newborn (BSN) score for neurodevelopmental outcomes in neonates with HIE.
• To compare the prognostic ability of the BSN score with clinical variables and expert EEG interpretation.

Main Methods:

• Secondary analysis of the HEAL trial cohort, including 203 neonates with HIE.
• Raw EEG data collected within 24 hours of birth were analyzed using the BSN score.
• Generalized linear mixed models assessed the area under the receiver operating characteristic curve (AUROC) for predicting death or severe neurodevelopmental impairment (NDI) at 2 years.

Main Results:

• The BSN scores correlated well with expert EEG classifications (Pearson's r = 0.69).
• The BSN score significantly improved the prediction of severe NDI or death compared to clinical variables alone (AUROC 0.90-0.93 vs. 0.79).
• The prognostic accuracy of the median BSN score was comparable to expert EEG assessments (AUROC 0.90).

Conclusions:

• Automated EEG background analysis with the BSN score offers an objective method for early prognostication in neonates with HIE.
• The BSN score aids in risk stratification and its accuracy rivals expert interpretation.
• Further research is needed to evaluate the BSN score's utility in resource-limited settings lacking expert EEG interpretation.