Remimazolam (REM) and midazolam (MID) are currently used interchangeably for sedation in digestive endoscopy, with available literature indicating comparable safety profiles. However, data on patient-reported procedural tolerability are heterogeneous and largely derived from studies in which tolerability was a secondary outcome. Moreover, most evidence comes from anesthesiologist-led settings. The primary aim of this monocentric, prospective observational study was to evaluate and compare patient-reported tolerability of REM versus MID (both administered with fentanyl) during outpatient diagnostic endoscopic ultrasound (EUS) performed in a non-anesthesiological setting.

Consecutive outpatients undergoing diagnostic EUS under moderate sedation and with American Society of Anesthesiologists class 1 to 3 were enrolled. The choice between REM and MID was made according to routine clinical practice, also considering drug availability. Demographic data were collected. The Modified Observational Alertness/Sedation Assessment (MOAA/S) scale was used to monitor sedation depth and determine the need for supplemental dosing. Vital signs were recorded every 5 minutes throughout the procedure. Recovery time was assessed using the Modified Aldrete score, while procedural tolerability was evaluated with the PRO-STEP scale. Operator satisfaction was measured using a Numeric Rating Scale. The study protocol was approved and endorsed by the anesthesiology department.

At the time of this preliminary analysis, 63 patients (46 biliopancreatic and 17 upper GI EUS) had been enrolled (median age 69 years, 44% female, 63% ASA class 2), with 31 (49%) receiving REM. Patients in the REM and MID groups, respectively, did not differ significantly in age (69 vs. 68 median years, p=0.99), sex (52% vs. 38% female, p=0.81), BMI (24.04 vs. 24.01, p=0.76), or ASA class distribution (p=0.54). Supplemental dosing was required in 74% of REM cases and 63% of MID cases (p=0.15). Total fentanyl dose did not differ significantly between groups. Median procedural time was also comparable (25 minutes vs. 18 minutes, p=0.14). Patient tolerability, as measured by the PRO-STEP score, was similar between REM and MID (1.50 vs. 1.67, p=0.77), as was operator satisfaction (NRS 9 vs. 9). Awakening and recovery times were significantly shorter in the REM group (p<0.001). No adverse events were recorded.

These preliminary findings suggest that REM and MID, when combined with fentanyl for moderate sedation in diagnostic EUS in a non-anesthesiological setting, provide comparable patient tolerability and operator satisfaction, with excellent safety profiles. REM is associated with a significantly faster post-procedure recovery.