Electroencephalographic Arousal Patterns Under Dexmedetomidine Sedation

BACKGROUND: The depth of dexmedetomidine-induced sedation is difficult to assess without arousing the patient. We evaluated frontal electroencephalogram (EEG) as an objective measure of dexmedetomidine-induced sedation. Our aims were to characterize the response patterns of EEG during a wide range of dexmedetomidine-induced sedation and to determine which spectral power best correlated with assessed levels of dexmedetomidine-induced sedation. METHODS: Sedline EEG sensor was positioned on the forehead of 16 volunteers. Frontal EEG data were collected at 250 Hz using the Sedline monitor. A computer-controlled infusion pump was used to infuse dexmedetomidine to four 15-minute target plasma concentrations of 0.3, 0.6, 1.2, and 2.4 ng/mL. Arterial blood samples for dexmedetomidine plasma concentration and sedation (self-reported numerical rating scale) and arousal were measured at baseline and at the end of each infusion step. The EEG signal was used to estimate spectral power in sequential 4-second data segments with 75% overlap for 3 power bands: delta = 0.5-1.5 Hz, alpha = 9-14 Hz, beta = 15-24 Hz. We quantified the relationships among the plasma concentrations of dexmedetomidine, level of sedation, and various EEG parameters. RESULTS: EEG data at the end of the dexmedetomidine infusion steps show progressive loss of high frequencies (beta) and increase in alpha and delta powers, with increasing dexmedetomidine concentrations. Beta prearousal spectral power was best in predicting dexmedetomidine-induced level of sedation (R = -0.60, 95% CI, -0.43 to -0.75). The respective values for delta and alpha powers were R = 0.28 (95% CI, 0.03-0.45) and R = 0.16 (95% CI, -0.09 to 0.38). When the beta power has dropped below -16 dB or the delta power is above 15 dB, the subjects show moderate to deep levels of sedation. When awakening the subject, there is a reduction in power in the delta and alpha bands at the 0.6, 1.2, and 2.4 ng/mL dexmedetomidine target levels (P < .001 for all). In beta band, there is a rapid awakening-induced increase in power (P < .001) followed by a slow return toward baseline values. After arousing the subjects, the EEG powers returned toward baseline values significantly slower than our clinical observation of the subjects' wakefulness would have suggested. CONCLUSIONS: Using a wide range of dexmedetomidine doses, we found that frontal EEG beta power of less than -16 dB and/or a delta power of over 15 dB was associated with a state of moderate to deep sedation and that poststimulus return of EEG powers toward baseline values took significantly longer than expected from observation of the arousal response. It is unclear whether these observations are robust enough for clinical applicability.