Bispectral analysis of the electroencephalogram does not predict responsiveness to verbal command in patients emerging from xenon anaesthesia

The bispectral index (BIS) is derived empirically from the electroencephalo gram database of patients receiving common anaesthetics, but it may not be valid for uncommon agents. Therefore, we investigated how xenon affects the BIS. Nine and I I patients were anaesthetized with 0.8 of the minimal alve olar concentration (MAC) of isoflurane (0.92%) and xenon (56%), respectivel y. After the end of operation, these concentrations were decreased in decre ments of 0.1 MAC (isoflurane 0.12% or xenon 7%) and each new concentration was maintained for 15 min. This was repeated until the patient first opened her eyes or squeezed the investigator's hand on command. Isoflurane and xe non at 0.8 MAC reduced the BIS to a median of 40 (range 36-53) and 36 (30-6 1), respectively. With decreasing concentrations of isoflurane, the BIS inc reased progressively and it reached a median of 96 (90-98) when the patient s awoke. In contrast, four patients receiving xenon responded to verbal com mand while the BIS was below 50 [median 45 (range 41-49)]. The remaining se ven patients in the xenon group awoke when their BIS was greater than 80 [m edian 96 (range 82-98)], but in four of them the BIS was no greater than 50 when the xenon concentration was only 0.1 MAC (7%) higher than that associ ated with awakening. We conclude that low BIS values (<50) do not guarantee adequate hypnosis during xenon anaesthesia.