Objective: To compare the value of bispectral index as a monitor of sedation in critically ill children with a validated sedation scoring system.
Design: Prospective convenience sample.
Setting: Paediatric intensive care unit in a tertiary paediatric centre.
Patients and participants: Forty-three critically ill children receiving sedation and mechanical ventilation.
Measurements and results: Simultaneous recording of bispectral index (BIS) and assessment of depth of sedation using the Comfort score were performed at regular intervals. To determine if BIS could detect episodes of arousal, times of endotracheal suctioning and the corresponding BIS score were recorded. There was an overall moderate correlation between BIS scores and Comfort scores ( r=0.50, r(2)=0.25, p<0.0001). Children who had a neurological reason for their current admission ( n=25) showed a weaker correlation ( r=0.26, r(2)=0.06, p<0.007) than those ( n=15) with normal neurology ( r=0.51, r(2)=0.26, p<0.0001). There were no significant differences in the rise in BIS following endotracheal suctioning among any of the predefined depths of sedation. There was a correlation of r=0.84 ( r(2)=0.71) (SE of slope 0.49, CI(95) 1.79-3.88) for mean BIS values for each individual Comfort score from 8-23. Using Spearman's rank correlation of Comfort versus mean BIS, the correlation coefficient was r=0.92.
Conclusions: Bispectral index scores correlate with Comfort scores to a moderate degree. BIS is able to discriminate between light and deep levels of sedation, but not between deep and very deep levels of sedation. The BIS monitor may provide a useful method for assessing sedation in critically ill children, especially those receiving neuromuscular blockers.