Electroencephalogram Correlates of Delayed Emergence After Remimazolam-Induced Anesthesia Compared to Propofol

BACKGROUND: Delayed emergence from anesthesia presents clinical challenges, including prolonged stays in the postanesthesia care unit (PACU). The neurobiological mechanisms underlying delayed emergence, particularly in remimazolam-induced anesthesia, remain poorly understood. This study aimed to explore patterns of brain electrical activity of delayed emergence in remimazolam-induced anesthesia by comparing dynamic changes in electroencephalogram (EEG) activity under various anesthesia states of remimazolam and propofol, focusing on the prefrontal region. METHODS: Forty-eight patients (age >18) who underwent laparoscopic cholecystectomy randomly received remimazolam- or propofol-induced general anesthesia. Power spectrogram analysis and functional connectivity measures, phase lag entropy (PLE) and phase lag index (PLI), were used to the prefrontal EEG data collected at baseline, unconsciousness, and emergence. Correlation between EEG measures and Patient State Index (PSI) at PACU, as well as time to Aldrete 9, were compared. RESULTS: During emergence from anesthesia, EEG power revealed that the remimazolam group had higher powers than the propofol group in theta band during eyes-open (EO) (mean of 2.933 [standard deviation of 5.762] vs -2.342 [4.869]; P-value of 0.018 with independent 2-sample t test), and in the alpha band during eyes-closed (EC) (5.821 [7.35] vs -2.399 [4.53]; P < .001) and EO (4.84 [6.411] vs -3.613 [4.556]; P < .001). Conversely, the functional connectivity result showed lower PLE in the alpha band during EC (0.619 [0.0338] vs 0.684 [0.0392]; P < .0001) and EO (0.651 [0.0358] vs 0.692 [0.0428]; P = .015), and in the beta band during EC (0.682 [0.0308] vs 0.712 [0.0236]; P = .016) and EO (0.695 [0.0236] vs 0.725 [0.0195]; P < .001). In line with this, the remimazolam group had lower PSI values at PACU during EC (65.10 [14.67] vs 82.40 [6.678]; P < .0001) and EO (72.35 [12.55] vs 83.53 [6.632]; P = .006) and were slower to reach Aldrete score of 9 (median difference of 17.5; interquartile range of [0.0-21.0]; P < .001). Delayed consciousness recovery (time to Aldrete 9) under remimazolam was significantly correlated with PLE (Pearson's correlation = -.78, P < .0001) and PLI (Pearson's correlation =.69, P = .028) in the alpha band during deep anesthesia. CONCLUSIONS: Dynamic changes in prefrontal EEG during recovery and the correlation analyses show the potential of EEG in reflecting distinct consciousness recovery profiles between 2 drugs-slower recovery under remimazolam anesthesia. This suggests an association of EEG parameters with a unique behavioral profile of remimazolam, especially reflecting progressive changes in cerebral activity during recovery.