This observational cohort study was conducted on patients who underwent TAVR at a tertiary care center in South Korea. All patients who underwent TAVR between July 2020 and July 2022 were evaluated for eligibility. Patients who underwent emergent or valve-in-valve TAVR and those scheduled for general anesthesia were excluded. The study data was obtained from the institutional Aortic Valve Replacement Registry and a medical record review. The study was approved by the Institutional Review Board (Approval number: 2022-1098), and the requirement for informed consent was waived owing to the study’s retrospective nature.

The primary exposure in the study was remimazolam, and the comparative exposure was dexmedetomidine. TAVR procedures were typically performed under MAC unless the patient’s overall condition was unstable or transapical TAVR was performed. Before July 2021, dexmedetomidine and remifentanil were the agents used for MAC, with dexmedetomidine dosages ranging from 0.3 to 0.7 μg/kg/h after loading of 1 μg/kg for 10 min and remifentanil target-controlled infusion (TCI) dosages ranging from 0.3 to 0.7 ng/ml. After its introduction in July 2021, remimazolam became the primary sedative in most TAVR procedures in the center. Remimazolam was administered in conjunction with remifentanil, with remimazolam dosages ranging from 0.2 to 0.6 mg/kg/h after a bolus of 2.5 to 5 mg and remifentanil TCI dosages ranging from 0 to 0.3 ng/ml. The target level of sedation aimed for was a Modified Observer’s Assessment of Alertness and Sedation score of ≤ 3. When dexmedetomidine was administered, the use of rescue sedatives, such as 1 mg of midazolam, was allowed in cases where the intended level of sedation was not achieved. All sedatives were discontinued upon confirming the integrity of the prosthetic aortic valve. At the end of the procedure, remimazolam was reversed with 0.2 mg of flumazenil. Perioperative management adhered to institutional standards involving multidisciplinary risk stratification and optimal management planning through collaboration with the heart team. The interventionist adopted a minimalist approach by simplifying the procedure, enabling TAVR without using transesophageal echocardiography, and relying on meticulous computed tomography measurement. The anesthesiologist performed arterial cannulation for perioperative blood pressure monitoring and 18-gauge venous cannulation for massive bleeding. The elective TAVR procedures were all performed in the morning, with two cases typically performed per day, sometimes up to three cases, all of which were completed by 1 p.m. The postoperative care objective was to minimize the cardiac ICU stay duration to less than a day, followed by discharge on the third day. A cardiac rehabilitation program was implemented if deemed necessary.

The primary outcome of the study was timely recovery after TAVR, defined as ICU discharge within the first day following TAVR. The criteria for discharge from the ICU in our center included several factors: the patient should be alert and conscious, hemodynamically stable, and not require vasopressors or inotropes, or if needed, they should be on minimal doses. The secondary outcomes included factors that may affect the patient’s timely recovery, such as the duration to be fully awake (the duration from ICU admission until the first instance when the Richmond Agitation-Sedation Scale score reaches 0, evaluated by ICU nurses), duration of postoperative oxygen supplementation, need for intubation at the ICU, infusion of vasopressor/inotropes (inclusive of drug infusion initiated in the operating room and continued in the ICU, as well as instances where a new infusion was initiated in the ICU), need for temporary pacemaker (TPM), and occurrence of delirium assessed using the Confusion Assessment Method for the Intensive Care Unit. The ICU nurses assessed delirium immediately upon arrival and during each nursing shift. Tertiary outcomes included all-cause mortality within 30 days after surgery, the occurrence of stroke, the need for cardiopulmonary resuscitation/extracorporeal membrane oxygenation, and the need for permanent pacemakers.

We anticipated that approximately 90% of the patients undergoing TAVR would achieve timely recovery based on the data from our TAVR registry. Remimazolam would be considered non-inferior with a margin of −10%. Based on these assumptions, 142 patients per group would be required to demonstrate non-inferiority with an alpha of 0.025 and a power of 0.8. The study duration was expected to achieve this sample size.

The analysis employed propensity-score matching to compare remimazolam and dexmedetomidine. A multivariable logistic regression model was used for estimating the propensity score, incorporating potential confounders such as age, sex, body mass index, smoking history, New York Heart Association (NYHA) functional classification, hypertension, diabetes mellitus, myocardial infarction, atrial fibrillation, stroke, peripheral vascular disease, pulmonary disease, chronic kidney disease, previous cardiac surgery, left ventricular ejection fraction, hemoglobin, B-natriuretic peptide, troponin I, albumin, and the Society of Thoracic Surgeons score. A complete-case analysis was conducted in the logistic regression modeling because there were a few participants with missing variables. After determining the propensity score, 1 : 1 greedy matching was conducted using a caliper width of 0.1. Matching balance was assessed using the standardized mean difference (SMD), considering it well-balanced when the SMD was < 0.1. In the matched-cohort analysis, McNemar’s test was used to compare categorical outcomes, and a paired t-test or Wilcoxon signed-rank test, as appropriate, was used for continuous outcomes. Categorical outcomes were reported with risk differences and 95% CIs.

To enhance the robustness of our primary findings, we conducted three sensitivity analyses. First, to address missing values of baseline characteristics in the propensity model, we performed single-value imputation using the median or mode. Second, a multivariable logistic regression analysis was executed on the unmatched cohort. Third, recognizing the possible confounding effect of the date of TAVR that could not be balanced between the two groups in our study, we conducted sensitivity analyses to assess its impact on the study results. To investigate the potential time-dependent pattern of the timely recovery rate, we used logistic regression with a restricted cubic spline to plot the estimated timely recovery rate by the month of TAVR. This analysis aimed to determine whether there was an overall correlation between the timing of TAVR and the observed timely recovery rate. Furthermore, the multivariable logistic regression analysis was repeated, treating the month of TAVR as a continuous variable and the six-month interval as a categorical variable. This approach helped ascertain whether there was a discernible effect of time on the outcomes.

The results of the secondary and tertiary outcomes were not adjusted for multiple comparisons. Therefore, all results, apart from those of the primary outcome, should be viewed as exploratory. All analyses were performed using R version 4.1.0 (R Foundation for Statistical Computing).

The medical records of 492 consecutive patients who underwent TAVR under MAC between July 2020 and July 2022 were reviewed retrospectively. Patients who underwent emergent TAVR (n = 5), valve-in-valve TAVR (n = 7), or had planned general anesthesia (n = 16) were excluded. After exclusion, 464 patients were included in the analytic cohort (Fig. 1). The baseline characteristics of the unmatched and matched cohorts are summarized in Table 1. The median age (Q1, Q3) was 81 (77, 84) years, and 195 (56.7%) were women. Among them, 218 (47.0%) received remimazolam as a sedative for MAC. After 1 : 1 propensity-score matching, the analysis included a total of 328 patients (164 in each group). Propensity matching resulted in a well-balanced baseline characteristic between the two groups, with an SMD < 0.1.

Intraoperative characteristics are outlined in Table 2. Rescue sedative requirements were not observed in the remimazolam group but were observed in 33.5% of the patients in the dexmedetomidine group (P < 0.001). No significant difference was observed in the need for vasopressor/inotropes during TAVR between the remimazolam and dexmedetomidine groups (30.5% vs. 24.4%; P = 0.216).

The primary and secondary outcomes are shown in Table 3. Of the 464 unmatched patients who underwent TAVR, 91.7% (200 of 218) in the remimazolam group were discharged from the ICU within the first day, compared to 91.5% (225 of 246) in the dexmedetomidine group. The median (Q1, Q3) length of ICU stay was 25.5 (23, 27) h and 26 (24, 28) h in the remimazolam and dexmedetomidine groups, respectively. The timely recovery rate remained consistent in the propensity score-matched cohort, with 92.7% (152 of 164) in the remimazolam group and 93.3% (153 of 164) in the dexmedetomidine group (P = 0.827). The median (Q1, Q3) length of ICU stay was 25.5 (23, 27) h in the remimazolam group and 26 (25, 28) h in the dexmedetomidine group. Non-inferiority was observed, as the difference in the proportion (95% CI) of patients with timely recovery was −0.6% (−6.1, 4.9) that was within the prespecified non-inferiority margin of −10%. Regarding secondary outcomes, the remimazolam group exhibited a significantly shorter duration to be fully awake and a lower need for vasopressor/inotrope support and TPM support. The incidence of postoperative delirium showed no significant difference between the groups.

Regarding the tertiary outcomes, no significant differences were observed in all-cause mortality within 30 days after surgery, the incidence of stroke, the need for cardiopulmonary resuscitation/extracorporeal membrane oxygenation, or the need for permanent pacemakers (Table 4). The results of the sensitivity analyses were consistent with those of the primary analysis, as discussed in Supplementary Tables 1 and 2 and Supplementary Fig. 1.

A similar retrospective study comparing remimazolam and dexmedetomidine for TAVR was recently published [23]. Compared to our study, this study used higher doses of remimazolam and dexmedetomidine, suggesting a deeper level of sedation. As an additional rescue sedative, propofol (20 mg) was administered to the dexmedetomidine group as opposed to midazolam in our study. The results of both studies indicated that rescue sedatives were not necessary for the remimazolam groups, suggesting that remimazolam may be more effective as a standalone sedative. In both studies, flumazenil was administered at the end of the procedure, and faster arousal was observed in the remimazolam group. The longer arousal time observed in the dexmedetomidine group appears to be consistent regardless of the rescue sedative used, propofol or midazolam.

The recent study focused primarily on the benefits of remimazolam in terms of arousal time; however, our study adds to the literature by focusing on clinically relevant postoperative outcomes. Our study demonstrates that remimazolam is associated with a reduced need for inotropes/vasopressors and TPMs following TAVR. A noteworthy point is that the different rescue drugs may have influenced the incidence of delirium in the dexmedetomidine group. Owing to the absence of information regarding delirium in the recent study, we could not assess the impact of midazolam use on delirium in the dexmedetomidine group.

The retrospective design and the single-center setting at a tertiary university hospital may limit the generalizability of the results. In addition, although the sensitivity analyses in this study revealed no significant trend or effect on the primary outcome over time, we cannot completely rule out the possibility that the temporal difference in the use of remimazolam and dexmedetomidine could have acted as a confounding factor.

In conclusion, in patients undergoing TAVR, remimazolam demonstrated a non-inferior association with timely recovery compared to dexmedetomidine. Additionally, remimazolam was associated with a more favorable recovery profile, including a shorter duration to be fully awake and reduced postoperative requirements for vasopressors/inotropes and TPMs.