Discharge Readiness After Propofol With or Without Dexmedetomidine for Colonoscopy

A Randomized Controlled Trial

Leonard U. Edokpolo, M.D.; Daniel J. Mastriano, M.D.; Joanna Serafin, Ph.D.; Jeremy C. Weedon, Ph.D.; Maryam T. Siddiqui, M.D.; Dennis P. Dimaculangan, M.D.


Anesthesiology. 2019;131(2):279-286. 

In This Article

Statistical Analysis

Our hypothesis was that the sedation regimen of combined propofol–dexmedetomidine would be noninferior to propofol–placebo. Based on institutional experience and implications from prior studies,[15,21–23] we projected that the probability of patients being discharged within 30 min would be 90% in the patients receiving propofol alone. We projected that the same probability would apply to patients receiving propofol–dexmedetomidine, but we asserted that if the latter number were as low as 70%, the extra inconvenience to all concerned would be minimal. In other words, this 20% differential represents a margin of clinical importance, and we aimed to demonstrate with reasonable certainty that the actual differential is smaller. This means that there would be no clinically important difference between both groups in the percentage of patients discharged within 30 min. Noninferiority would be shown by the probability that the percentage of patients discharged within 30 min from the end of the procedure would be within the 20% margin of clinical importance for the two groups. This margin was subjectively selected by the clinical anesthesiologists commonly involved in ambulatory endoscopy procedures at our institution.

Assuming equal numbers of subjects in each study arm, the power of the Pearson chi-square test to detect a differential as small as this is 90% for n = 37 per arm, given a two-tailed test and significance level of 0.05. However, that test is not always accurate for low prevalence, so we aimed instead for n = 50 per arm, which should yield an expected frequency of five or more in each cell of the 2×2 table that constitutes the principal analysis.

The primary outcome was evaluated with the Farrington–Manning method to compute a one-sided CI for the 30 min Modified Post Anesthetic Discharge Scoring System risks difference and the Fisher exact test. The method of Turnbull was used to estimate cumulative distribution of time to Modified Post Anesthetic Discharge Scoring System greater than 8 in the presence of both interval- and right-censoring. The secondary outcomes were compared between groups using the Mann–Whitney U test for nonparametric variables. Incidences of bradycardia and apnea were reported as total number of patients affected, n (percentage of total in the group). Statistical analyses were performed with SAS version 9.4 software (SAS Institute Inc., USA). Demographic characteristics were summarized with n (%) or median (quartiles) as appropriate. Results were expressed as median [quartiles], or number of patients, n (%). All analyses were planned prior to the investigation.