Background: Sedation practice, especially when non-anaesthesia personnel are involved, requires efficient anaesthetic depth monitoring. Therefore, we used prediction probability (PK) to evaluate the performance of the bispectral index (BIS) of the EEG and automated responsiveness test (ART) to predict sedation depth and loss of subject's responsiveness during propofol sedation, with and without N2O.
Methods: Twenty volunteers were studied during propofol administration with (N2O) and without (Air) N2O. The protocol consisted of sequential 15-min cycles. After a control period, propofol was infused to a target effect-site concentration of 0.25 µg/ml (N2O) or 1.5 µg/ml (Air), which was subsequently increased by 0.25 or 0.5 µg/ml, respectively, until loss of responsiveness was detected by loss of response to command [observer's assessment of alertness/sedation (OAA/S) score ≤ 2].
Results: Deeply sedated patients did not respond to ART indicating that the sensitivity was 1.0 with or without N2O. The specificity of ART was 0.24 with Air and 0.98 with N2O. The BIS was better than ART in predicting OAA/S score (PK = 0.84 vs. 0.77) and loss of responsiveness (PK = 0.87 vs. 0.69) during the Air trial. Nitrous oxide decreased the performance of BIS in predicting OAA/S score (PK = 0.76), but increased the performance of ART to predict loss of responsiveness (PK = 0.85).
Conclusion: BIS and ART comparably predict sedation and loss of responsiveness. However, ART, because of its resistance to false-normal responses, might prove to be more useful on an individual patient basis – especially in the presence of agents that impair BIS sensitivity.