A comparison of EEG spectral entropy with conventional quantitative EEG at varying depths of sevoflurane anaesthesia

Authors

  • PR Bartel
  • FJ Smith
  • PJ Becker

Keywords:

EEG spectral entropy, conventional QEEG, sevoflurane anaesthesia

Abstract

Background and Aim: Recently an electroencephalographic (EEG) spectral entropy module (M-ENTROPY) for an anaesthetic monitor has become commercially available. We compared its performance as an indicator of the state of anaesthesia with that of an older conventional quantitative EEG (QEEG) module (M-EEG) by the same manufacturer (Datex-Ohmeda Division, Instrumentarium Corp., Helsinki, Finland). Methods: There were 40 ASA class I or II subjects, aged between 16-60 years, who underwent elective abdominal surgery. EEG data were collected from the printouts of the respective modules. The data presented here were related to four levels of anaesthesia: Pre-anaesthetic wakefulness (state A), 2% sevoflurane endtidal (ET) concentration after completion of surgery (state B), low ET sevoflurane concentrations (~ 0.5%) just prior to regaining responsiveness (state C), and post-anaesthetic responsiveness (state D).

Results: In terms of the prediction probability (Pk statistic), response entropy (RE) and state entropy (SE) produced higher values (0.95-1.0) than the best performing QEEG variable, frontal amplitude (0.86-0.95). Only RE scores did not overlap between states A and B or between B and D. The misclassification of subjects between states C and D was far lower for RE (28%) than for any of the conventional QEEG measures (>90%). Conclusion: In on-line monitoring spectral entropy is superior in distinguishing states of anaesthesia and is also easier to use than conventional QEEG. It is speculated that the artefact rejection strategies accorded spectral entropy might significantly benefit conventional QEEG analysis.

Author Biographies

PR Bartel

Department of Neurology, University of Pretoria and Pretoria Academic Hospital, South Africa

FJ Smith

Department of Anaesthesiology, University of Pretoria and Pretoria Academic Hospital, South Africa

PJ Becker

Biostatistics Unit, South African Medical Research Council, and Faculty of Health Sciences, University of Pretoria, South Africa

Downloads

Issue

Section

Original Research