The Use of Bispectral Index to Monitor Barbiturate Coma in Severely Brain-Injured Patients with Refractory Intracranial Hypertension

Abstract

Barbiturate therapy in severely traumatic brain-injured (TBI) patients is usually monitored by an electroencephalogram (EEG) with burst-suppression pattern as a target. The Bispectral Index (BIS) is derived from EEG and considers cortical silence. We sought to determine whether a BIS range could predict a specific burst-suppression pattern. Eleven TBI patients treated with barbiturate were included prospectively. EEG was recorded daily for 1 h. Every 5 min, the number of bursts and the suppression ratio (suppression ratio from EEG [SR(EEG)]: percentage of last 60 s in cortical silence) was calculated for 1 min on the raw EEG and compared to concomitant data from the BIS-XP (BIS and suppression ratio [SR(BIS)]). The optimal level of barbiturate coma was defined as 2-5 bursts/min in the EEG. A BIS range predictive of optimal level was determined from all data and its accuracy was studied for each examination. Agreement between SR(EEG) and SR(BIS) was high (interclass correlation coefficient 0.94 [95% confidence interval: 0.90-0.96]). There was a significant association between SR(EEG) and BIS. Significant disagreements were observed in some examinations. The best accuracy to predict optimal pattern was obtained with a BIS range from 6 to 15. The relationship between BIS and SR(EEG) was high in TBI patients treated with barbiturates. The rate of barbiturate infusion might be decreased if BIS is <6 or increased if BIS is >15. Correspondence between BIS and suppression pattern should periodically be checked by observation of the EEG analogical signal (as displayed by BIS-XP).