Evaluating the performance of the NarcotrendR EEG index during anaesthesia for cardiothoracic surgery: a
Max Ebensperger1, Matthias Kreuzer1, Stephan Kratzer2
1Department of Anesthesiology and Intensive Care, TUM School of Medicine and Health, Technical University of Munich, Munich, Germany.
Background:
Using neuromonitoring during general anaesthesia provides insights into the effects of anaesthetics on the brain. We focus on the performance of the processed EEG indices NarcotrendR (NCT) and Burst Suppression Ratio (BSR) of the Narcotrend-Compact-MR Module, which serve as surrogate parameters for the level of consciousness.
Methods:
In this single-centre retrospective study, we analysed processed electroencephalographic (EEG) data from 439 patients who underwent general anaesthesia for cardiac surgery. We employed data visualisation techniques, such as histograms and heat maps. We conducted statistical analyses using correlation coefficients, receiver operating characteristics, and linear regression to evaluate Narcotrend performance under various BSR conditions.
Results:
The NCT index demonstrated distinct "peak" values (37, 46, and 61), which occurred with a probability more than two standard deviations above the overall index distribution (BSR = 0). During steady-state anaesthesia, 70% [Q1, Q3: 67,72] of values were within the manufacturer-recommended range for adequate anaesthesia, 22% [Q1, Q3: 21,29] were below, and 8% [Q1, Q3: 6,12] were above. With the onset of BSR > 0, NCT decreased significantly (p < 0.001) but showed significant variability immediately before and after automated burst suppression detection. Approximately 13% [Q1, Q3: 9,24] of NCT readings were non-interpretable. These brief episodes increased significantly with patient age (p = 0.013) and were not attributable to concurrent burst suppression.
Conclusion:
The Narcotrend index remains within recommended ranges during steady-state anaesthesia in a predominantly male patient cohort undergoing cardiac surgery. However, index performance decreases with age, and the high incidence of non-interpretable readings in elderly patients highlights the need for cautious interpretation, despite their short duration. Automatically detected burst suppression (BSR > 0) leads to a near-instant decrease in NCT values, suggesting a technical link between the algorithms. "Peak" index values indicate an irregular scaling in the distribution of NCT index values.
Trial Registration:
This trial was retrospectively registered at ClinicalTrials.gov (NCT02976584) in October 2016.
