background & aims: Current practice guidelines strongly recommend differentiation of deep from moderate sedation during endoscopy. Standard methods of sedation monitoring are labor-intense. Bispectral index monitoring (BIS) is widely used during anesthesia, but its benefits during conscious sedation are controversial. Thus, we performed a prospective observational study to assess its ability for detecting deep sedation during endoscopy.
Methods: Patients presenting for elective outpatient endoscopy were monitored simultaneously with the Modified Observer's Assessment of Alertness and Sedation (MOAA/S) and BIS. A combination of a narcotic and benzodiazepine was used, with the target being moderate sedation and analgesia. Deep sedation was defined by MOAA/S score of 1-2 and BIS score of <or=75. With MOAA/S as the reference standard, the accuracy of BIS for detecting deep sedation was evaluated.
Results: A total of 775 simultaneous observations of BIS and MOAA/S scores were recorded on 76 patients. Deep sedation, defined by MOAA/S and BIS, was seen in 204 (26%) and 92 (12%) observations, respectively. BIS correlated poorly with deep sedation (rho, -0.02; 95% confidence interval [CI], -0.16-0.12). The sensitivity, specificity, and positive and negative predictive values (95% CI) for detecting deep sedation were 29.6 (23.4-36.3), 94.9 (92.8-96.6), 68.2 (57.4-77.7), and 78.6 (75.3-81.7), respectively.
Conclusions: BIS has a low accuracy for detecting deep sedation as a result of a considerable overlap of scores across the sedation levels. Further refinements in BIS are needed to differentiate deep from moderate sedation for future studies on conscious sedation.