Background: Manufacturers recommend maintaining Bispectral (BIS) or Spectral Entropy (State Entropy, SE) indexes between 40 and 60 during the maintenance of anesthesia. We compared these indexes during this period.
Methods: Data were obtained from 58 patients receiving sufentanil-sevoflurane-nitrous oxide anesthesia. The anesthesiologist was blinded to BIS and SE. Artifact-free concurrent BIS and SE values (7792 pairs), automatically recorded at 1-min intervals, were compared using Bland-Altman analysis, Kappa coefficient for agreement and crude proportion of agreement. The occurrence of errors of judgment (Type 1 defined as one parameter <40 and the other >60, or Type 2 defined as BIS and SE values on different sides of a threshold [40 or 60]) was also counted.
Results: Bias was -2 with limits of agreement of -18 and 9. Kappa BIS/SE obtained from all patients was 0.537 +/- 0.147; crude agreement >0.80 was observed in 45% of patients. Type 1 number of errors of judgment corresponded to two instances. Median and interquartile values of Type 2 number of errors of judgment were 4.5 [3.0-6.0] when considering a difference between BIS and SE more than 5.
Conclusion: Although limits of agreement between BIS and SE were large, Kappa value moderate, and crude agreement <0.80 in more than half of the patients, making completely contradictory decisions (e.g., deepening the anesthetic based on one parameter and lightening it based upon the other) would have been exceptional. More common would have been a risk of error between no change versus increasing or decreasing anesthetic depth.