Continuity with caveats in anesthesia: state and response entropy of the EEG

Max Ebensperger; Matthias Kreuzer; Stephan Kratzer; Gerhard Schneider; Stefan Schwerin

doi:10.1007/s10877-024-01130-9

Continuity with caveats in anesthesia: state and response entropy of the EEG

J Clin Monit Comput. 2024 Apr 3. doi: 10.1007/s10877-024-01130-9. Online ahead of print.

Authors

Max Ebensperger¹, Matthias Kreuzer², Stephan Kratzer³, Gerhard Schneider¹, Stefan Schwerin¹

Affiliations

¹ Department of Anesthesiology and Intensive Care, School of Medicine and Health, Technical University of Munich, Ismaningerstr. 22, 81675, Munich, Germany.
² Department of Anesthesiology and Intensive Care, School of Medicine and Health, Technical University of Munich, Ismaningerstr. 22, 81675, Munich, Germany. m.kreuzer@tum.de.
³ Abteilung für Anästhesiologie, Intensiv- und Schmerzmedizin, Hessing Stiftung, Hessingstraße 17, 86199, Augsburg, Germany.

PMID: 38568370
DOI: 10.1007/s10877-024-01130-9

Abstract

The growing use of neuromonitoring in general anesthesia provides detailed insights into the effects of anesthetics on the brain. Our study focuses on the processed EEG indices State Entropy (SE), Response Entropy (RE), and Burst Suppression Ratio (BSR) of the GE Entropy^TM Module, which serve as surrogate measures for estimating the level of anesthesia. While retrospectively analyzing SE and RE index values from patient records, we encountered a technical anomaly with a conspicuous distribution of index values. In this single-center, retrospective study, we analyzed processed intraoperative electroencephalographic (EEG) data from 15,608 patients who underwent general anesthesia. We employed various data visualization techniques, including histograms and heat maps, and fitted custom non-Gaussian curves. Individual patients' anesthetic periods were evaluated in detail. To compare distributions, we utilized the Kolmogorov-Smirnov test and Kullback-Leibler divergence. The analysis also included the influence of the BSR on the distribution of SE and RE values. We identified distinct pillar indices for both SE and RE, i.e., index values with a higher probability of occurrence than others. These pillar index values were not age-dependent and followed a non-equidistant distribution pattern. This phenomenon occurs independently of the BSR distribution. SE and RE index values do not adhere to a continuous distribution, instead displaying prominent pillar indices with a consistent pattern of occurrence across all age groups. The specific features of the underlying algorithm responsible for this pattern remain elusive.

Keywords: Anesthesia; Burst suppression; Electroencephalogram; Response entropy; State entropy.