This single-institution retrospective study was approved by the Ethics Committee of Hamamatsu University Hospital, Hamamatsu, Japan (approval number 23–062). Because it was a retrospective study, the requirement to obtain informed consent was waived by the Ethics Committee.

Patients who underwent noncardiac surgery under general anesthesia at our institution from September 2020 to March 2023 were retrospectively studied. Patients aged ≥ 20 years who underwent invasive arterial pressure measurements and in whom only remimazolam or sevoflurane was used for anesthetic maintenance were included. Patients who discontinued surgery, had a < 10-min observation period for arterial pressure measurement, or had incomplete data were excluded. The patients were divided into two groups according to the drug used for anesthetic maintenance: the sevoflurane group (Group S) and remimazolam group (Group R). The choice of anesthetic drug, intraoperative anesthetic depth, and hemodynamic control were at the discretion of the anesthesiologist in charge of the patient. Electroencephalography monitoring was used for patients with SedLine® (Masimo Corporation, Irvine, CA, USA) or Bispectral Index™ monitors (Medtronic Inc., Minneapolis, MN, USA) as needed.

Baseline characteristics including age, sex, body mass index (BMI), American Society of Anesthesiologists physical status (ASA-PS), comorbidities, and medication history were recorded for each patient. In addition, the surgical time, infusion volume, urine output, blood loss, anesthetic use (the mean infusion rate of remifentanil and remimazolam and the mean sevoflurane concentration), use of any vasopressors (ephedrine, phenylephrine, dopamine, or noradrenaline), and doses of vasopressors were investigated.

A catheter was inserted into the radial artery and the mean blood pressure was evaluated from the start to the end of the observation period for arterial pressure measurements. Invasive arterial pressure was recorded at 1-min intervals. Artifacts were removed using the following criteria: (1) out-of-range blood pressure as defined by (a) systolic blood pressure (SBP) ≥ 300 or ≤ 20 mmHg, (b) SBP ≤ diastolic blood pressure (DBP) + 5 mmHg, or (c) DBP ≤ 5 or ≥ 225 mmHg, and (d) mean arterial pressure ≤ 25 mmHg; and (2) abrupt changes in SBP ≥ 80 mmHg within 1 min in either direction or ≥ 40 mmHg within 2 min in both directions. Blood pressures between measurements were interpolated linearly.

The hypotensive time was recorded as the time for which the mean arterial pressure was < 65 mmHg and the IOH was defined as a cumulative hypotensive time of ≥ 10 min [7, 12]. The primary outcome was the difference in the incidence of IOH between the two groups. The secondary outcomes were the cumulative hypotensive time, incidence of vasopressor use, and intraoperative dose of vasopressors used (ephedrine, phenylephrine, dopamine, and noradrenaline).

Continuous variables are presented as the mean (standard deviation) or median (interquartile range), and categorical variables are presented as a number (percentage).

Continuous variables were tested for normality with the Shapiro–Wilk test. Comparisons between groups were performed with a t-test for data following a normal distribution, and asymmetric data were analyzed using the Mann–Whitney U-test. Categorical variables were compared with the chi-square test or Fisher’s direct probability test, as appropriate.

Propensity score matching was performed to match the study groups using logistic regression analysis including the following potential confounding factors as independent variables: age, sex, BMI, ASA-PS, diabetes, hypertension, coronary artery disease, cerebrovascular disease, angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, concurrent epidural anesthesia, scheduled or emergency surgery, surgical site, and scheduled operative time. The nearest-neighbor matching method (1:1 ratio) was applied, with a caliper width of 0.2 for the logit-transformed propensity score. Variables in a matched data set were considered balanced between the groups if the standardized mean difference was < 0.1. All statistical analyses were performed using R software version 4.3.0 (R Development Core Team, Vienna, Austria). The R package “Matching” was used for the propensity score matching. All p-values < 0.05 were considered statistically significant.