Creation of a score to predict risk of high conscious sedation requirements in patients undergoing endoscopy

Background and Aims: The administration of intravenous conscious sedation to patients undergoing GI endoscopy carries a risk of cardiopulmonary adverse events. Our study aim was to create a score that stratifies the risk of occurrence of either high-dose conscious sedation requirements or a failed procedure. Methods: Patients receiving endoscopy via endoscopist-directed conscious sedation were included. The primary outcome was occurrence of sedation failure, which was defined as one of the following: (1) high-dose sedation, (2) the need for benzodiazepine/narcotic reversal agents, (3) nurse-documented poor patient tolerance to the procedure, or (4) aborted procedure. High-dose sedation was defined as >10 mg of midazolam and/or >200 mg of fentanyl or the meperidine equivalent. Patients with sedation failure (n = 488) were matched to controls (n = 976) without a sedation failure by endoscopist and endoscopy date. Results: Significant associations with sedation failure were identified for age, sex, nonclonazepam benzodiazepine use, opioid use, and procedure type (EGD, colonoscopy, or both). Based on these 5 variables, we created the high conscious sedation requirements (HCSR) score, which predicted the risk of sedation failure with an area under the curve of 0.70. Compared with the patients with a risk score of 0, risk of a sedation failure was highest for patients with a score >= 3.5 (odds ratio, 17.31; P = 2 x 10(-14)). Estimated area under the curve of the HCSR score was 0.68 (95% confidence interval, 0.63-0.72) in a validation series of 250 cases and 250 controls. Conclusions: The HCSR risk score, based on 5 key patient and procedure characteristics, can function as a useful tool for physicians when discussing sedation options with patients before endoscopy.