Early outcome-prediction with an automated EEG background trend in hypothermia-treated newborns with encephalopathy
- PMID: 40523949
- DOI: 10.1038/s41390-025-04193-9
Early outcome-prediction with an automated EEG background trend in hypothermia-treated newborns with encephalopathy
Abstract
Background: Therapeutic hypothermia is an intervention that improves outcomes and alters early outcome-prediction in infants with moderate-severe hypoxic-ischemic encephalopathy (HIE). This study evaluated the early predictive accuracy of a fully automated continuous EEG background trend, Brain State of the Newborn (BSN) in a regional Swedish cohort of infants with presumed HIE.
Method: The BSN trend characterizes 1-min segments of EEG from zero (inactive) to 100 (continuous) and was generated from aEEG/EEG in 85 infants treated with hypothermia. BSN trajectories were compared in relation to clinical grading of encephalopathy and outcome. Receiver operating characteristics were computed for good (no/mild impairment) and poor outcomes (moderate/severe impairment or death).
Results: During the first 48 h, BSN levels differed significantly between moderate and severe HIE (typical median BSN levels >80 and <40, respectively). The predictive accuracy of BSN was high already at 6 h (AUC 0.84) and at 12 h (AUC 0.91), with corresponding positive predictive values (PPV) > 0.92 for good outcome (cutoff BSN > 80) and PPV > 0.95 for poor outcome (cutoff BSN < 40).
Conclusion: BSN gives a continuous and objective measure of EEG background activity, which is highly predictive of good and poor outcomes already from the first 6-12 h in hypothermia-treated infants with moderate-severe HIE.
Impact: Brain State of the Newborn (BSN) is a deep learning-based EEG trend displaying electrocortical activity as numerical values. Here we establish that BSN trends over the first 48 h differ between infants with moderate versus severe hypoxic-ischemic encephalopathy (HIE) and is highly predictive of long-term outcome. This is the first study applying BSN trends in a cohort of exclusively hypothermia-treated infants, demonstrating its value for outcome-prediction already from the first 12 h after birth. The BSN provides a continuous bedside evaluation of brain function that complements the visual aEEG/EEG review and can assist bedside assessments in neonatal intensive care units.
© 2025. The Author(s).
Conflict of interest statement
Competing interests: The authors declare no competing interests. Consent statement: Ethical approval including waiver of consent was granted by Uppsala Regional Research Ethics Review board (Dnr 2015/511).
Similar articles
-
Clinical outcome prediction with an automated EEG trend, Brain State of the Newborn, after perinatal asphyxia.Clin Neurophysiol. 2024 Jun;162:68-76. doi: 10.1016/j.clinph.2024.03.007. Epub 2024 Mar 15. Clin Neurophysiol. 2024. PMID: 38583406
-
Automated assessment of EEG background for neurodevelopmental prediction in neonatal encephalopathy.Ann Clin Transl Neurol. 2024 Dec;11(12):3267-3279. doi: 10.1002/acn3.52233. Epub 2024 Nov 14. Ann Clin Transl Neurol. 2024. PMID: 39543820 Free PMC article.
-
Prognostic value of quantitative EEG in early hours of life for neonatal encephalopathy and neurodevelopmental outcomes.Pediatr Res. 2024 Aug;96(3):685-694. doi: 10.1038/s41390-024-03255-8. Epub 2024 Jul 22. Pediatr Res. 2024. PMID: 39039325 Free PMC article.
-
Mild Hypoxic-Ischemic Encephalopathy: Can Neurophysiological Monitoring Predict Unfavorable Neurological Outcome? A Systematic Review and Meta-analysis.Am J Perinatol. 2023 Jun;40(8):833-838. doi: 10.1055/s-0041-1736593. Epub 2021 Oct 19. Am J Perinatol. 2023. PMID: 34666398
-
Predictive value of amplitude-integrated EEG (aEEG) after rescue hypothermic neuroprotection for hypoxic ischemic encephalopathy: a meta-analysis.J Perinatol. 2017 Jun;37(6):684-689. doi: 10.1038/jp.2017.14. Epub 2017 Mar 2. J Perinatol. 2017. PMID: 28252661 Review.
References
LinkOut - more resources
Full Text Sources