Determination of ED90 of Remimazolam Besylate Combined With Alfentanil for Gastroscopy Sedation in Obese Patients

Obese patients undergoing painless gastrointestinal endoscopy usually exhibit a higher incidence of sedation-related respiratory depression, airway obstruction and hypoxemia. Given the pharmacological profiles of remimazolam besylate and alfentanil, the combined-agent sedation strategy is well-suited for obese patients undergoing gastroscopy. Nevertheless, individualized dosing of remimazolam combined with alfentanil for gastroscopic sedation in obese patients remains unclear. This is a randomized double-blind, biased-coin sequential allocation dose-finding trial. The trial consists of two groups (Group L and Group M), in which two independent sequential experiments are performed separately. The random sequence will be generated and stored by an independent statistician who will not participate in patient recruitment, anesthesia delivery, or outcome evaluation. Participants will be allocated sequentially according to the order of enrollment, and allocation concealment will be realized using opaque sealed envelopes to ensure unpredictable treatment assignments. During the study, alfentanil hydrochloride and remimazolam besylate will be separately diluted with 0.9% normal saline to volumes of 10 mL and 20 mL in syringes of identical appearance, respectively. The medications will be prepared in advance by an anesthesiologist who will not participate in anesthesia administration or outcome assessment. Study participants, endoscopists, and all other investigators will be blinded to group allocation. Only the drug preparer will have access to the group information. Remimazolam besylate and alfentanil hydrochloride doses are both calculated based on the recommended lean body weight (LBW) for obese patients. The LBW calculation formulas are shown below:

For male patients: lean body weight (kg) = (9270 × total body weight (kg)) / (6680 + 216 × BMI) For female patients: lean body weight (kg) = (9270 × total body weight (kg)) / (8780 + 244 × BMI). Eligible study participants not meeting any exclusion criteria are randomly assigned at a 1:1 ratio to Group L (alfentanil 5 μg/kg) or Group M (alfentanil 10 μg/kg). The initial remimazolam dose is 0.15 mg/kg in Group L and 0.1 mg/kg in Group M, the remimazolam dose for the next study participant is determined by the response of the preceding participant. In the event of sedation failure in a participant, the remimazolam dose is increased for the subsequent participant. When sedation is effective, the next participant is randomly allocated to receive either the same remimazolam dose (probability 0.89) or a reduced dose (probability 0.11). The dose step is 0.03 mg/kg. After administration of remimazolam, the sedation status of patients will be continuously evaluated by assessors using the Modified Observer's Assessment of Alertness/Sedation scale (MOAA/S). Gastroscopy will be performed by endoscopists with the same professional qualification once the MOAA/S score≤1. If the MOAA/S score is ≤ 1 within 3 minutes and there are no positive reactions such as coughing, swallowing or body movement during gastroscope insertion and within 1 minute after insertion, it will be defined as effective sedation. Otherwise, it will be defined as sedation failure. When sedation is ineffective, a single intravenous dose of 2.5 mg remimazolam will be administered for rescue sedation. The interval between single remimazolam doses will be more than 1 minute, and the number of supplemental administrations will not exceed 5 times within 15 minutes. If sedation remains ineffective after 5 rescue administrations, propofol will be subsequently used for rescue sedation to complete the examination, with rescue dosage determined by the anesthesiologist according to the clinical situation. The primary outcome is the 90% effective dose (ED90) of remimazolam for suppressing gastroscope insertion-related responses in obese patients under different alfentanil doses. Secondary outcomes include haemodynamic changes during gastroscopy, the incidence of adverse reactions, and three-party satisfaction scores.Hypotension is defined as a more than 20% decrease in mean arterial pressure (MAP) from baseline, and 0.5 mg metaraminol will be administered intravenously. A heart rate lower than 50 beats per minute will be defined as bradycardia, and 0.3-0.5 mg atropine will be injected intravenously. When hypoxemia occurs (75% ≤ SpO₂ \< 90% with duration \< 60 seconds), oxygen flow will be increased, and jaw thrust maneuver or nasopharyngeal airway insertion will be performed (size 7.0 or 7.5 for males; 6.5 or 7.0 for females). If severe hypoxemia occurs (SpO₂ \< 75% or 75% ≤ SpO₂ \< 90% with duration ≥ 60 seconds), the gastroscope will be removed, mask mechanical ventilation will be initiated, and tracheal intubation will be considered if necessary. When postoperative nausea and vomiting occurs, 8 mg ondansetron will be given intravenously.

Sample size calculation According to guideline recommendations on sample size for biased-coin design up-down sequential method, a sample size of 50-60 participants is preferable for reliable estimation of the ED90. Based on previous studies, at least 45 effective sedation cases are required for ED90 determination using a biased coin sequential design. Accordingly, this study will recruit a minimum of 45 participants defined as achieving effective sedation. Therefore, the sample size of this study will be determined based on the above two considerations. If the number of successfully sedated participants in each group is fewer than 45 among the planned 60 participants, the trial will be continued until the cumulative number of successfully sedated participants per group reaches no fewer than 45. If 45 successfully sedated participants have been accrued in each group while the total sample size per group has not yet reached 60, the trial will be continued until each group includes a total of 60 participants.

Statistical analysis

1. The normality of continuous variables will be tested in two steps. First, histograms and Q-Q plots will be plotted to preliminarily assess the distribution pattern of the data. Second, the Shapiro-Wilk test will be applied for formal normality testing. Data conforming to a normal distribution will be described as mean ± standard deviation (SD), while non-normally distributed data will be presented using the median and interquartile range (IQR).
2. Normally distributed continuous variables will be compared by the independent t-test, while non-normally distributed continuous variables will be compared using the Mann-Whitney U test. Categorical variables will be analyzed with the Chi-square test or Fisher's exact test. The selection of the final statistical analysis method will depend on the distribution characteristics of the experimental data.
3. Repeated-measures analysis of variance will be used to compare changes in hemodynamic parameters at different time points among patients in the same group. If the result is statistically significant, Bonferroni correction will be applied for further post hoc comparisons.
4. The ED90 of remimazolam will be calculated using centralized isotonic regression, and the 95% confidence interval (CI) will be obtained from 2000 bootstrap replicates.
5. P-value of less than 0.05 will be defined as the threshold for statistical significance in all hypothesis tests. The 95% confidence interval (95% CI) will be adopted for interval estimation.
6. All the above statistical analyses will be performed using SPSS and R software.