Abstract
This study tested the hypothesis that the bispectral index (BIS) is reliable relative to clinical judgment for estimating sedation level during daily propofol spontaneous awakening trials (SATs) in trauma patients.
This was a prospective observational trial with waiver of consent conducted in the intensive care unit of Level I trauma center in 94 mechanically ventilated trauma patients sedated with propofol alone or in combination with midazolam. BIS, Richmond Agitation Sedation Scale (RASS), electromyography, and heart rate variability, as a test of autonomic function, were measured for 45 minutes during daily SATs. Data were evaluated with analysis of variance, linear regression, and nonparametric tests.
The BIS wave form coincided almost exactly with propofol on/off. Steady-state BIS correlated with RASS (p < 0.0001) and with propofol dose (p < 0.0001), but the strengths of association were relatively low (all r(2) < 0.5). BIS wave form was not altered by age, heart rate, or heart rate variability and was similar with propofol alone or propofol plus midazolam, but the presence of brain injury or the use of paralytics shifted the curve downward (both p < 0.001). The overall test characteristics for BIS versus RASS without neuromuscular blockade were sensitivity: 90% versus 77% (p = 0.034); specificity: 90% versus 75% (p = 0.021); positive predictive value: 90% versus 76% (p = 0.021), and negative predictive value: 90% versus 76% (p = 0.021).
In the first trial in trauma patients and largest trial in any surgical population, the (1) BIS was reliable and has advantages over RASS of being continuous and objective, at least during a propofol SAT; (2) BIS interpretation remains somewhat subjective in patients receiving paralytic agents or with traumatic brain injury.
