Abstract
Bispectral index (BIS) is used for monitoring anaesthetic depth with inhaled anaesthetic agents in the operating room but has not been evaluated as a monitor of sedation depth in the intensive care unit (ICU) setting with these agents. If BIS could predict sedation depth in ICU patients, patient disturbances could be reduced and oversedation avoided. Twenty ventilator-dependent ICU patients aged 27 to 80 years were randomised to sedation with isoflurane via the AnaConDa® or intravenous midazolam. BIS (A-2000 XP, version 3.12), electromyogram activity (EMG) and Signal Quality Index were measured continuously. Hourly clinical evaluation of sedation depth according to Bloomsbury Sedation Score (Bloomsbury) was performed. The median BIS value during a 10-minute interval prior to the clinical evaluation at the bedside was compared with Bloomsbury. Nurses performing the clinical sedation scoring were blinded to the BIS values. End-tidal isoflurane concentration was measured and compared with Bloomsbury. Correlation was poor between BIS and Bloomsbury in both groups (Spearman's rho 0.012 in the isoflurane group and -0.057 in the midazolam group). Strong correlation was found between BIS and EMG (Spearman's rho 0.74). Significant correlation was found between end-tidal isoflurane concentration and Bloomsbury (Spearman's rho 0.47). In conclusion, BIS XP does not reliably predict sedation depth as measured by clinical evaluation in non-paralysed ICU patients sedated with isoflurane or midazolam. EMG contributes significantly to BIS values in isoflurane or midazolam sedated, non-paralysed ICU patients. End-tidal isoflurane concentration appeared to be a better indicator of clinical sedation depth than BIS.