Intraoperative hypotensive burden with remimazolam versus propofol during ambulatory dental anesthesia in adults with intellectual disabilities: a retrospective within-subject study.

Abstract

Purpose: Intraoperative hypotension is common during general anesthesia and may be poorly tolerated by adults with intellectual disabilities, who often require general anesthesia for dental care. We compared intraoperative hypotensive burden during propofol- versus remimazolam-based maintenance in this population.

Methods: We retrospectively analyzed 150 ambulatory dental general anesthesia episodes from 47 adults with intellectual disabilities who received both propofol- and remimazolam-based maintenance. The primary outcome was the percentage of maintenance-phase time with mean arterial pressure (MAP) below 80% of the episode-specific baseline. Outcomes were analyzed using linear mixed-effects models with a patient-level random intercept. Sensitivity analyses included covariate adjustment, a patient fixed-effects model, an anesthesia-time-matched paired comparison, alternative outcome models, and analyses of temporal confounding.

Results: The percentage of maintenance-phase time with MAP below 80% of baseline was lower during remimazolam-based maintenance than during propofol-based maintenance (median, 0.0% vs. 23.3%; estimated difference, -21.4 percentage points; 95% confidence interval, -29.2 to -13.7; P < 0.001). Remimazolam-based maintenance was also associated with a higher mean maintenance MAP (+ 4.6 mmHg; P < 0.001). The remimazolam-with-flumazenil-reversal strategy was associated with a shorter postanesthesia care unit stay (- 14.4 min; P < 0.001). The primary association remained directionally consistent across sensitivity analyses but was attenuated after calendar-time adjustment. Time-weighted average measures did not differ significantly.

Conclusion: In adults with intellectual disabilities undergoing ambulatory dental anesthesia, remimazolam-based maintenance was associated with less time below the prespecified relative MAP threshold than propofol-based maintenance. Because residual temporal and procedure-related confounding precludes causal inference, this finding should be considered hypothesis-generating.

Trial registration: Clinical Research Information Service (CRIS), Republic of Korea; registration number KCT0011046; registered on September 19, 2025.