Cardioprotective effects of remimazolam in elderly hypertensive patients undergoing gastrointestinal tumor surgery: A randomized controlled trial

Abstract

BACKGROUND

Elderly patients with hypertension undergoing gastrointestinal tumor surgery are at increased risk of perioperative cardiovascular complications, particularly myocardial injury after non-cardiac surgery (MINS). Anesthetic management plays a crucial role in maintaining hemodynamic stability and reducing cardiac stress in this vulnerable population. Remimazolam, a novel ultra-short-acting benzodiazepine, has shown potential advantages in cardiovascular stability compared with conventional anesthetic agents; however, its cardioprotective effects in high-risk surgical patients remain insufficiently explored.

AIM

To evaluate the cardioprotective effects and safety of remimazolam-based general anesthesia compared with conventional anesthesia in elderly hypertensive patients undergoing gastrointestinal tumor surgery, with a particular focus on myocardial injury markers, hemodynamic stability, postoperative recovery, and the incidence of MINS.

METHODS

A total of 80 elderly hypertensive patients who underwent gastrointestinal tumor surgery at Yizheng Hospital of Nanjing Drum Tower Hospital Group between April 2024 and August 2025 were selected. Patients were randomly divided into a control group (P group) and an experimental group (R group) using the envelope method. P group received conventional general anesthesia, while R group received remimazolam-based general anesthesia. Perioperative serum biomarkers, arterial blood gas indices, postoperative pain, incidence of postoperative complications, surgical indicators, and adverse events within 30 days of follow-up were compared between the two groups. Patients were further divided into non-MINS and MINS groups based on the occurrence of MINS. Differences in clinical indicators between the two groups were analyzed. Binary logistic regression was used to identify risk factors for MINS, and a predictive model was constructed and validated based on these factors.

RESULTS

At T0, creatine kinase (CK) levels were significantly lower and lactate levels significantly higher in P group compared with R group (P < 0.05). At T1, systolic and diastolic blood pressures were significantly lower in P group than in R group (P < 0.05). At T6, CK-MB levels were significantly higher in P group (P < 0.05). At T7, High-Sensitivity Troponin (hs-cTn), CK-MB, and aspartate aminotransferase were significantly higher in P group (P < 0.05). At T8, CK-MB and N-terminal pro-brain natriuretic peptide (NT-proBNP) were significantly higher in P group (P < 0.05). Visual Analog Scale scores at T7 and T8 were significantly higher in P group (P < 0.05). Postoperative agitation scores and incidence were significantly higher in P group (P < 0.05), as was the incidence of MINS (P < 0.05). Intraoperatively, colloid use was significantly lower, while crystalloid and norepinephrine use were significantly higher in P group (P < 0.05). There was no significant difference in the incidence of adverse events within 30 days postoperatively between the two groups (P > 0.05). Compared with the MINS group, the non-MINS group had significantly lower proportions of colon cancer and conventional general anesthesia, as well as lower levels of hs-cTn, CK-MB, NT-proBNP at T6-T8, and shorter operative time (P < 0.05). Multivariate analysis showed that tumor type, anesthesia method, NT-proBNP at T7, and NT-proBNP at T8 were independent risk factors for MINS. The area under the curve for NT-proBNP at T8 and the predictive model was > 0.7.

CONCLUSION

The use of remimazolam for general anesthesia in elderly hypertensive patients undergoing gastrointestinal tumor surgery can reduce myocardial injury and the incidence of MINS. It helps stabilize perioperative blood pressure, alleviates postoperative pain, reduces the risk of postoperative agitation, and decreases intraoperative norepinephrine requirements. It is safe and does not increase postoperative complications or adverse events within 30 days. Tumor type, anesthesia method, and postoperative NT-proBNP levels can help predict the risk of MINS, providing a reference for prevention and treatment strategies.

Key Words: Hypertension; Gastrointestinal tumors; Surgery; Remimazolam; Myocardial injury after non-cardiac surgery; Cardioprotection

Core Tip: This randomized controlled study demonstrates that remimazolam-based anesthesia provides significant cardioprotective effects in elderly hypertensive patients undergoing gastrointestinal tumor surgery. Compared with conventional anesthesia, remimazolam reduces cardiac biomarker levels, improves hemodynamic stability, alleviates postoperative pain and agitation, and lowers the incidence of myocardial injury after non-cardiac surgery. These findings suggest that remimazolam may be a safer anesthetic option in high-risk elderly populations and highlight potential strategies for preventing perioperative cardiac complications.

INTRODUCTION

With the aging population and ongoing advances in medical technology, the proportion of elderly patients undergoing surgery has steadily increased. Elderly patients with hypertension often exhibit reduced physiological reserve and are frequently complicated by multi-organ dysfunction. During the perioperative period, they are particularly susceptible to surgical trauma and stress responses, which can increase myocardial oxygen consumption and exacerbate relative myocardial ischemia[1]. In addition, decreased vascular elasticity and impaired autonomic regulation in this population place coronary reserve in a vulnerable state. Perioperative stress may further disrupt the balance between myocardial oxygen supply and demand, leading to metabolic disturbances in cardiomyocytes and even triggering apoptosis, thereby contributing to perioperative mortality and cardiovascular adverse events[2,3].

Myocardial injury after non-cardiac surgery (MINS) is a major contributor to perioperative cardiovascular adverse events. It refers to myocardial injury occurring after surgery that cannot be explained by non-ischemic causes, with elevated serum troponin serving as the primary diagnostic criterion. The incidence of MINS is approximately 16%, and its clinical presentation is often atypical about 60% of patients remain asymptomatic making it easily overlooked. MINS is strongly associated with cardiovascular complications and increased mortality within both 30 days and 1 year after surgery[4]. Patients who develop MINS have a 10.07-fold higher risk of 30-day mortality, along with a markedly increased incidence of cardiovascular complications. Age is an independent risk factor for MINS. With the rapid aging of the population in China, where individuals aged ≥ 65 years accounted for more than one-tenth of the population in 2015[5], the number of elderly surgical patients continues to rise, accompanied by a higher incidence of MINS 1.62-fold in patients aged 65-74 years and 2.66-fold in those aged ≥ 75 years[6]. Furthermore, patients with MINS face a substantial risk of long-term cardiovascular events, with a 5-year incidence of myocardial infarction exceeding 25%, significantly impairing long-term survival and quality of life. Therefore, early identification and prevention of MINS in elderly patients are crucial for reducing perioperative cardiovascular risk and improving prognosis.

A cohort study reported that approximately 12.6% of elderly patients undergoing abdominal surgery develop MINS[7]. Patients with gastrointestinal malignancies are often complicated by preoperative anemia, hypoalbuminemia, electrolyte imbalance, and a hypercoagulable state. Preoperative anemia and hypoalbuminemia increase the risk of early postoperative mortality, myocardial injury, and infection[8]. Without effective preventive measures, long-term survival and quality of life may be adversely affected, increasing both social and healthcare burdens. Propofol is a commonly used intravenous anesthetic but is associated with adverse effects such as hypotension, which may increase the risk of myocardial ischemia, particularly in elderly and hypertensive patients. Remimazolam, a novel benzodiazepine, has favorable safety and metabolic characteristics and may serve as an ideal sedative agent[9]. It has a rapid onset, minimal respiratory and circulatory depression, low accumulation, and can be rapidly reversed by flumazenil. It is primarily metabolized by CES1 into inactive metabolites, with a short elimination half-life (approximately 1.1-2.1 hours), making it suitable for elderly patients. In cardiovascular anesthesia, it exerts less cardiac suppression and has significantly less impact on myocardial contractility and heart rate (HR) compared with propofol[10]. However, there is currently no definitive evidence demonstrating whether remimazolam provides cardioprotective effects in elderly hypertensive patients undergoing gastrointestinal tumor surgery, particularly in terms of improving postoperative myocardial injury biomarkers and reducing the incidence of MINS.

MATERIALS AND METHODS

General information

Study subjects: This study recruited patients undergoing gastrointestinal tumor surgery at Yizheng Hospital of Nanjing Drum Tower Hospital Group. Written and verbal informed consent was obtained from all participants prior to enrollment. Among the included patients, there were 51 males and 29 females, aged 60-91 years, with a mean age of (72.36 ± 7.51) years. According to the American Society of Anesthesiologists (ASA) classification, 22 cases were grade II and 58 cases were grade III. Body mass index (BMI) ranged from 16.60 kg/m² to 30.60 kg/m², with an average of (23.55 ± 3.11) kg/m². Tumor types included 31 cases of colon cancer, 21 cases of rectal cancer, 27 cases of gastric cancer, and 1 case of gastrointestinal stromal tumor. Comorbidities included cerebral infarction (11 cases), diabetes mellitus (16 cases), coronary heart disease (3 cases), and others (8 cases).

Patients were randomly assigned into two groups using the envelope method. Specifically, 80 opaque envelopes containing numbered cards (numbers 1-326, non-repeating) were prepared. After admission, each patient was randomly assigned an envelope. Patients with odd numbers (1, 3, …, 79) were allocated to the control group (P group) (propofol-based intravenous general anesthesia), while those with even numbers (2, 4, …, 80) were assigned to the experimental group (R group) (remimazolam-based intravenous general anesthesia), with 40 patients in each group.

All intravenous anesthetic agents used in both groups were prepared in advance by anesthesia nurses. Syringes were wrapped with opaque paper, and extension tubes were also opaque. Other anesthetic medications were identical between the two groups. Anesthesiologists were responsible for anesthesia administration, while designated personnel recorded relevant data. All participants involved in the study remained blinded to treatment allocation throughout. The randomization sequence was kept confidential until completion of statistical analysis and drafting of the manuscript.

All patients in this single-center trial consented to participate and signed informed consent forms. The study was approved by the Medical Ethics Committee (No. YZYY-2024-06) and registered with the Chinese Clinical Trial Registry (No. ChiCTR2500112319), in accordance with the Declaration of Helsinki[11].

Grouping

A total of 80 patients were included in this study. Forty patients in the P group received conventional general anesthesia (P group), while 40 patients in the R group received remimazolam-based general anesthesia (R group). P group received conventional general anesthesia, while R group received remimazolam-based general anesthesia. All patients fasted for 8-12 hours preoperatively and continued antihypertensive therapy, with systolic blood pressure (SBP) maintained at 120-160 mmHg and diastolic blood pressure (DBP) at 80-90 mmHg before surgery. Upon entering the operating room, patients were monitored for blood pressure, HR, pulse oxygen saturation, electrocardiogram, and bispectral index (BIS). Oxygen was administered. Under local anesthesia, right internal jugular vein catheterization was performed to establish venous access, and left radial artery catheterization was conducted for invasive arterial pressure monitoring.

Anesthesia induction: In R group, intravenous injection included dexamethasone sodium phosphate (10 mg; 5 mg per vial; approval No. H20051748; Ma’anshan Fengyuan Pharmaceutical Co., Ltd.), remimazolam besylate (0.2-0.3 mg/kg; 25 mg per vial; approval No. H20200006; Yichang Humanwell Pharmaceutical Co., Ltd.), sufentanil citrate (0.4-0.5 μg/kg; 10 mL: 50 μg; approval No. H20237165; Yichang Humanwell Pharmaceutical Co., Ltd.), and cisatracurium besylate (0.15-0.2 mg/kg; 5 mL: 10 mg; approval No. H20183042; Jiangsu Hengrui Medicine Co., Ltd.) for intravenous induction.

In P group, intravenous induction consisted of dexamethasone (10 mg), sufentanil citrate (0.4-0.5 μg/kg), propofol emulsion (1.5-2.5 mg/kg; 10 mL: 200 mg; approval No. H20163406; Guangdong Jiabo Pharmaceutical Co., Ltd.), and cisatracurium besylate (0.15-0.2 mg/kg).

Anesthesia maintenance: R group received continuous intravenous infusion of remimazolam (0.4-1.0 mg/kg/hour) and remifentanil hydrochloride (0.05-0.2 μg/kg/minute; 2 mg per vial; approval No. H20030199; Yichang Humanwell Pharmaceutical Co., Ltd.). P group received propofol infusion (4-10 mg/kg/hour) combined with remifentanil hydrochloride (0.05-0.2 μg/kg/minute).

Both groups received cisatracurium besylate at 0.1 mg/kg/hour combined with sevoflurane (120 mL; approval No. H20213735; Shanghai Hengrui Medicine Co., Ltd.). BIS was maintained between 40 and 60, and drug dosages were adjusted according to anesthesia depth. Arterial blood pressure fluctuations were controlled within 20% of baseline values. Hypotension was managed with fluid infusion and/or vasopressors, and bradycardia was treated with atropine, with interventions determined by the anesthesiologist based on clinical judgment.

Sevoflurane was discontinued 30 minutes before the end of surgery, and cisatracurium was stopped 20 minutes before the end. All intravenous anesthetics were discontinued during skin closure. After recovery of spontaneous respiration, neostigmine (0.02 mg/kg; 1 mL: 0.5 mg; approval No. H20057097; Zhejiang Xianju Pharmaceutical Co., Ltd.) combined with atropine (0.01 mg/kg; 1 mL: 0.5 mg; approval No. H12020382; Beijing Jingfeng Pharmaceutical Group Co., Ltd.) was administered intravenously for antagonism. After meeting extubation criteria, airway suctioning and tracheal extubation were performed, and the patient was transferred to the post-anesthesia care unit (PACU).

For analgesia, flurbiprofen axetil (50 mg; 5 mL: 50 mg; approval No. H20183054; Grand Medical Nutrition Science Co., Ltd.) was administered intravenously before incision and 10 minutes before the end of surgery. Postoperative analgesia was provided using a patient-controlled analgesia pump with sufentanil citrate (3 μg/kg) and granisetron (3 mg; 3 mL: 3 mg; approval No. H10970243; Ningbo Tianheng Pharmaceutical Co., Ltd.).

Outcome measures

Data were recorded at the following time points: Before anesthesia induction (T0), after anesthesia induction (T1), during tracheal intubation (T2), at skin incision (T3), at skin closure (T4), immediately after extubation (T5), 1 hour after surgery (T6), 24 hours after surgery (T7), and 48 hours after surgery (T8).

Primary outcomes: (1) At T0, T6, T7, and T8, 3 mL of central venous blood was collected from both groups to measure High-Sensitivity Troponin (hs-cTn), creatine kinase (CK), CK-MB (CK-MB), lactate dehydrogenase (LDH), aspartate aminotransferase (AST), and N-terminal pro-brain natriuretic peptide (NT-proBNP), in order to evaluate whether remimazolam reduces the release of postoperative myocardial injury biomarkers in elderly hypertensive patients undergoing gastrointestinal tumor surgery; and (2) The incidence of MINS within 30 days postoperatively was recorded.

Secondary outcomes: (1) SBP, DBP, and HR were recorded at T0, T1, T2, T3, T4, T5, and T6; (2) At T0 and T4, arterial blood gas analysis was performed to measure arterial blood glucose, lactate, and partial pressure of carbon dioxide (PaCO₂); (3) Pain intensity was assessed at T6, T7, and T8; (4) Postoperative complications were recorded, including agitation, delayed recovery, delirium, hypoxemia, and nausea/vomiting; (5) Surgical-related indicators were recorded, including operation time, intraoperative fluid volume, colloid volume, crystalloid volume, intraoperative blood loss, blood transfusion volume, urine output, norepinephrine dosage, extubation time (from discontinuation of intravenous anesthetics to tracheal extubation), PACU stay duration, and postoperative hospital stay; and (6) All patients were followed up for 30 days, and adverse events during follow-up were recorded, including pulmonary infection, hypertension, incision infection, cerebral infarction, loss of appetite, and intestinal obstruction.

Scales and assessment methods

Pain was assessed using the Visual Analog Scale (VAS), which classifies pain intensity into 11 levels from 0 to 10, where 0 indicates no pain and 10 indicates unbearable severe pain; higher scores indicate more severe pain. The Ramsay Sedation Scale was used to assess the level of agitation during postoperative recovery, classifying agitation into six levels: 1 = anxious, agitated, and highly responsive to stimuli; 2 = cooperative, oriented, and tranquil; 3 = responds to commands but sluggish; 4 = brisk response to light glabellar tap or loud auditory stimulus; 5 = no response to light glabellar tap or loud auditory stimulus but responds to painful stimuli; 6 = no response to painful stimuli.

Statistical analysis

Statistical analysis was performed using SPSS version 27.0. The Kolmogorov-Smirnov test was used to assess the normality of continuous data. Normally distributed data were expressed as mean ± SD and analyzed using independent-samples t-tests for between-group comparisons and paired t-tests for within-group comparisons. Non-normally distributed data were expressed as median and interquartile range and analyzed using the Mann-Whitney U test for between-group comparisons and the Wilcoxon signed-rank test for within-group comparisons. Categorical data were expressed as n (%); the χ² test was used for non-ordinal data, and the Mann-Whitney U test was used for ordinal data. Binary logistic regression analysis was performed to identify risk factors for MINS in elderly hypertensive patients undergoing gastrointestinal tumor surgery. A logistic model was constructed to predict MINS, and receiver operating characteristic (ROC) curves were used to evaluate predictive performance by calculating the area under the curve (AUC). The DeLong test was used to compare AUCs between different indicators. ROC curves were plotted using GraphPad Prism 9.5. A P value < 0.05 was considered statistically significant.

RESULTS

Analysis of baseline characteristics and intraoperative indicators

There were no statistically significant differences between the two groups in baseline characteristics, including sex, age, ASA classification, height, weight, BMI, tumor type, and proportions of comorbidities (cerebral infarction, diabetes, coronary heart disease, and other conditions) (P > 0.05), as shown in Table 1.

Table 1 Comparison of baseline characteristics between the two groups, mean ± SD/n (%).

Item	P group (n = 40)	R group (n = 40)	χ2/t value	P value
Sex			2.650	0.104
Male	29 (72.50)	22 (55.00)
Female	11 (27.50)	18 (45.00)
Age (years)	73.90 ± 8.03	70.83 ± 6.70	1.859	0.067
ASA classification			2.257	0.133
II	8 (20.00)	14 (35.00)
III	32 (80.00)	26 (65.00)
Height (cm)	165.98 ± 6.73	164.73 ± 6.72	0.831	0.409
Weight (kg)	64.46 ± 9.59	64.74 ± 10.41	-0.123	0.903
BMI (kg/m2)	23.30 ± 2.99	23.80 ± 3.25	-0.713	0.478
Tumor type			5.137	0.162
Colon cancer	11 (27.50)	20 (50.00)
Rectal cancer	13 (32.50)	8 (20.00)
Gastric cancer	15 (37.50)	12 (30.00)
Gastrointestinal stromal tumor	1 (2.50)	0 (0.00)
Comorbidities
Cerebral infarction	7 (17.50)	4 (10.00)	0.949	0.330
Diabetes mellitus	6 (15.00)	10 (25.00)	1.250	0.264
Coronary heart disease	2 (5.00)	1 (2.50)	0.000	1.000
Others	5 (12.50)	3 (7.50)	0.139	0.709

P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia); ASA: American Society of Anesthesiologists; BMI: Body mass index.

Comparison of perioperative serum biomarkers of myocardial injury between the two groups

For intergroup comparisons, at T0, CK levels in the P group were significantly lower than those in the R group (P < 0.05). At T6, CK-MB levels in the P group were significantly higher than those in the R group (P < 0.05). At T7, levels of hs-cTn, CK-MB, and AST in the P group were significantly higher than those in the R group (P < 0.05). At T8, CK-MB and NT-proBNP levels in the P group were significantly higher than those in the R group (P < 0.05).

For intragroup comparisons, at T6, CK, CK-MB, LDH, and NT-proBNP levels in the P group were all significantly increased compared with those at T0, while hs-cTn levels in the R group were significantly decreased compared with T0 (P < 0.05). At T7, levels of hs-cTn, CK, CK-MB, LDH, AST, and NT-proBNP in the P group were all significantly increased compared with T0, and hs-cTn, CK, and NT-proBNP were also significantly increased compared with T6; in the R group, NT-proBNP was significantly decreased compared with T0, while CK, LDH, and NT-proBNP were significantly increased compared with T6 (P < 0.05). At T8, CK, CK-MB, LDH, and NT-proBNP levels in the P group were significantly increased compared with T0, and NT-proBNP was significantly increased compared with T6; in the R group, hs-cTn was significantly decreased compared with T0, NT-proBNP was significantly increased compared with both T0 and T6, and CK, CK-MB, LDH, and NT-proBNP were all significantly decreased compared with T7 (P < 0.05). Details are shown in Figure 1 and Table 2.

Figure 1 Comparison of trends in perioperative serum biomarkers of myocardial injury between the two groups. A: High-Sensitivity Troponin; B: Creatine kinase; C: Creatine kinase-MB; D: Lactate dehydrogenase; E: AST; F: NT-proBNP. hs-cTn: High-Sensitivity Troponin; CK: Creatine kinase; CK-MB: Creatine kinase-MB; LDH: Lactate dehydrogenase; AST: Aspartate aminotransferase; NT-proBNP: N-terminal pro-brain natriuretic peptide.

Table 2 Comparison of perioperative serum biomarkers of myocardial injury between the two groups, mean ± SD.

Item	Time	P group (n = 40)	R group (n = 40)	t value	P value
hs-cTn (ng/mL)	T0	11.73 ± 14.55	13.15 ± 5.83	-0.575	0.567
	T8	14.73 ± 6.73	10.25 ± 4.55a	3.483	< 0.001
CK (U/L)	T0	31.96 ± 18.41	56.40 ± 47.37	-3.042	0.003
	T8	68.87 ± 63.87a	63.24 ± 84.49	0.336	0.738
CK-MB (U/L)	T0	8.41 ± 5.09	6.69 ± 4.36	1.626	0.108
	T8	10.21 ± 5.83a	5.54 ± 3.27	4.417	< 0.001
LDH (U/L)	T0	186.87 ± 76.69	190.21 ± 59.56	-0.218	0.828
	T8	210.09 ± 88.78a	179.52 ± 52.75	1.873	0.065
AST (U/L)	T0	23.81 ± 14.79	26.54 ± 35.80	-0.445	0.657
	T8	27.09 ± 16.79	21.62 ± 9.54	1.794	0.077
NT-proBNP (pg/mL)	T0	147.60 ± 144.07	142.56 ± 152.48	0.152	0.880
	T8	490.16 ± 328.72a	285.00 ± 239.52a	3.190	0.002

^aP < 0.05 vs T0.

hs-cTn: High-Sensitivity Troponin; CK: Creatine kinase; CK-MB: Creatine kinase-MB; LDH: Lactate dehydrogenase; AST: Aspartate aminotransferase; NT-proBNP: N-terminal pro-brain natriuretic peptide; P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia).

Comparison of arterial blood gas parameters between the two groups

For intergroup comparisons, at T0, lactate levels in the P group were significantly higher than those in the R group (P < 0.05). At T1, SBP and DBP in the P group were significantly lower than those in the R group (P < 0.05).

For intragroup comparisons, at T1, SBP, DBP, and HR in both groups were significantly decreased compared with T0 (P < 0.05). At T2, SBP and DBP in the P group were significantly decreased compared with T0, while SBP, DBP, and HR in both groups were significantly increased compared with T1 (P < 0.05). At T3, SBP, DBP, and HR in the P group were significantly decreased compared with T0; SBP and DBP in both groups were significantly increased compared with T1; HR in both groups was significantly decreased compared with T2 (P < 0.05).

At T4, SBP, DBP, HR, and lactate in both groups were significantly decreased compared with T0, whereas glucose and PaCO₂ were significantly increased compared with T0; SBP in the P group was significantly increased compared with T1; SBP, DBP, and HR in both groups were significantly decreased compared with T2; SBP and DBP in both groups were significantly decreased compared with T3 (P < 0.05).

At T5, HR in both groups was significantly increased compared with T0; SBP and DBP in the R group were significantly increased compared with T0; SBP, DBP, and HR in both groups were significantly increased compared with T1, T2, T3, and T4 (P < 0.05).

At T6, SBP in both groups was significantly decreased compared with T0, and DBP in the R group was significantly decreased compared with T0; SBP, DBP, and HR in both groups were significantly increased compared with T1 and T4, but significantly decreased compared with T5; HR in the R group was significantly decreased compared with T2, while HR in both groups was significantly increased compared with T3 (P < 0.05). Details are shown in Figure 2 and Table 3.

Figure 2 Comparison of trends in arterial blood gas parameters between the two groups. A: Systolic blood pressure; B: Diastolic blood pressure; C: Heart rate; D: Glucose; E: Lactate; F: Partial pressure of carbon dioxide. SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HR: Heart rate; Glu: Glucose; Lac: Lactate; PaCO₂: Partial pressure of carbon dioxide.

Table 3 Comparison of arterial blood gas parameters between the two groups, mean ± SD.

Item	Time	P group (n = 40)	R group (n = 40)	t value	P value
SBP (mmHg)	T0	157.00 ± 19.34	153.58 ± 17.32	0.834	0.407
	T6	145.30 ± 18.10a	144.08 ± 17.91a	0.304	0.762
DBP (mmHg)	T0	79.88 ± 13.60	80.10 ± 10.80	-0.082	0.935
	T6	76.48 ± 10.29	73.58 ± 9.41a	1.316	0.192
HR (bpm)	T0	74.28 ± 11.33	72.43 ± 12.78	0.685	0.495
	T6	72.85 ± 13.01	72.05 ± 11.19	0.295	0.769
Glu (mmol/L)	T0	6.16 ± 1.36	6.28 ± 1.95	-0.321	0.749
	T4	6.91 ± 1.34a	7.32 ± 1.77a	-1.170	0.245
Lac (mmol/L)	T0	1.08 ± 0.53	0.84 ± 0.28	2.507	0.014
	T4	0.80 ± 0.36a	0.68 ± 0.29a	1.714	0.091
PaCO2 (mmHg)	T0	40.11 ± 6.27	42.29 ± 5.88	-1.608	0.112
	T4	48.34 ± 7.05a	46.22 ± 5.56a	1.488	0.141

^aP < 0.05 vs T0.

SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HR: Heart rate; Glu: Glucose; Lac: Lactate; PaCO₂: Partial pressure of carbon dioxide; P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia).

Postoperative pain comparison between the two groups

Intergroup comparison showed that VAS scores at T7 and T8 were significantly higher in the P group than in the R group (P < 0.05). Intragroup analysis demonstrated that, in both groups, VAS scores at T7 and T8 were significantly increased compared with those at T6 (P < 0.05), while at T8, VAS scores were significantly decreased compared with those at T7 (P < 0.05). See Figure 3 and Table 4 for details.

Figure 3 Comparison of trends in postoperative Visual Analog Scale scores between the two groups. VAS: Visual Analog Scale.

Table 4 Comparison of postoperative pain between the two groups, mean ± SD.

Time	P group (n = 40)	R group (n = 40)	t value	P value
T6	0.45 ± 0.88	0.43 ± 0.96	0.122	0.903
T7	3.38 ± 1.44a	2.65 ± 1.44a	2.247	0.028
T8	2.28 ± 1.11a,b	1.20 ± 0.88a,b	4.796	< 0.001

^aP < 0.05 vs T6 within the same group.

^bP < 0.05 vs T7 within the same group.

P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia).

Comparison of postoperative complication rates between the two groups

The postoperative agitation score in the P group was significantly lower than that in the R group (P < 0.05). However, the incidence of postoperative agitation and MINS was significantly higher in the P group than in the R group (P < 0.05). See Table 5 for details.

Table 5 Comparison of postoperative complication rates between the two groups, n (%).

Item	P group (n = 40)	R group (n = 40)	χ2/t value	P value
Postoperative agitation score, median (IQR)	1.00 (1.00-2.00)	2.00 (2.00-3.00)	-5.12	< 0.001
Postoperative agitation	24 (60.00)	4 (10.00)	21.978	< 0.001
Delayed recovery	0 (0.00)	0 (0.00)
Postoperative delirium	1 (2.50)	0 (0.00)	0	1
Hypoxemia	0 (0.00)	0 (0.00)
Nausea and vomiting	5 (12.50)	8 (20.00)	0.827	0.363
MINS	18 (45.00)	6 (15.00)	8.571	0.003

IQR: Interquartile range; MINS: Myocardial injury after non-cardiac surgery; P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia).

Comparison of intraoperative parameters between the two groups

The volume of colloid infusion was significantly lower in the P group than in the R group (P < 0.05), whereas the volumes of crystalloid infusion and norepinephrine consumption were significantly higher in the P group than in the R group (P < 0.05). See Table 6 for details.

Table 6 Comparison of intraoperative parameters between the two groups, mean ± SD.

Item	P group (n = 40)	R group (n = 40)	t value	P value
Operative time (minute)	176.25 ± 55.42	163.38 ± 57.07	1.024	0.309
Intraoperative fluid volume (mL)	1530.00 ± 316.19	1462.50 ± 327.92	0.937	0.352
Colloid volume (mL)	153.75 ± 230.77	287.50 ± 250.32	-2.485	0.015
Crystalloid volume (mL)	1376.25 ± 301.70	1175.00 ± 401.12	2.536	0.013
Intraoperative blood loss (mL)	56.00 ± 50.32	49.75 ± 25.77	0.699	0.487
Intraoperative transfusion (mL)	7.50 ± 47.43	15.00 ± 66.22	-0.582	0.562
Urine output (mL)	335.25 ± 146.71	316.25 ± 129.29	0.615	0.541
Norepinephrine consumption (μg)	354.00 ± 509.58	37.70 ± 80.77	3.877	< 0.001
Extubation time (minute)	16.35 ± 11.34	14.08 ± 7.56	1.056	0.294
PACU stay (minute)	45.00 ± 11.82	44.75 ± 11.09	0.098	0.923
Postoperative hospital stay (day)	12.53 ± 3.00	11.55 ± 2.02	1.702	0.093

PACU: Post-anesthesia care unit; P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia).

Comparison of 30-day adverse events between the two groups

During the 30-day follow-up, 5 patients in the P group and 4 patients in the R group experienced adverse events. There was no statistically significant difference in the overall incidence of adverse events between the two groups (P > 0.05), as shown in Table 7. However, due to the low event rate and limited sample size, the statistical power of this analysis is insufficient to detect small or moderate differences between groups. Therefore, the findings should be interpreted with caution, and larger studies are required to further validate the safety profile.

Table 7 Comparison of 30-day adverse events between the two groups, n (%).

Item	P group (n = 40)	R group (n = 40)	Fisher’s exact test	P value
Pulmonary infection	1 (2.50)	0 (0.00)
Incision infection	2 (5.00)	1 (2.50)
Cerebral infarction	1 (2.50)	0 (0.00)
Loss of appetite	1 (0.00)	1 (2.50)
Intestinal obstruction	0 (0.00)	2 (5.00)
Total adverse events	5 (12.50)	4 (10.00)	0.000	1.000

P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia).

Univariate analysis of MINS in elderly hypertensive patients undergoing gastrointestinal tumor surgery

Univariate analysis showed that, compared with the MINS group, the non-MINS group had significantly lower proportions of colon cancer and conventional general anesthesia, as well as lower levels of hs-cTn, CK-MB, and NT-proBNP at T6-T8, and shorter operative time (all P < 0.05). See Tables 8, 9, 10 and 11 for details.

Table 8 Comparison of baseline characteristics between non-myocardial injury after non-cardiac surgery and myocardial injury after non-cardiac surgery patients, mean ± SD/n (%).

Item	Non-MINS group (n = 56)	MINS group (n = 24)	χ2/t value	P value
Sex			1.878	0.171
Male	33 (58.93)	18 (75.00)
Female	23 (41.07)	6 (25.00)
Age (years)	71.96 ± 7.45	73.29 ± 7.74	-0.722	0.472
ASA classification			2.018	0.155
II	18 (32.14)	4 (16.67)
III	38 (67.86)	20 (83.33)
Height (cm)	165.27 ± 6.66	165.54 ± 6.98	-0.166	0.869
Weight (kg)	64.79 ± 9.57	64.17 ± 10.99	0.254	0.801
BMI (kg/m2)	23.70 ± 3.16	23.21 ± 3.06	0.634	0.528
Tumor type			13.582	0.004
Colon cancer	28 (50.00)	3 (12.50)
Rectal cancer	10 (17.86)	11 (45.83)
Gastric cancer	18 (32.14)	9 (37.50)
Gastrointestinal stromal tumor	0 (0.00)	1 (4.17)
Comorbidities
Cerebral infarction	10 (17.86)	1 (4.17)	1.626	0.202
Diabetes mellitus	10 (17.86)	6 (25.00)	0.182	0.669
Coronary heart disease	2 (3.57)	1 (4.17)	0.000	1.000
Others	5 (8.93)	3 (12.50)	0.007	0.935
Anesthesia strategy			8.571	0.003
Conventional general anesthesia (P group)	22 (39.29)	18 (75.00)
General anesthesia + remimazolam (R group)	34 (60.71)	6 (25.00)

MINS: Myocardial injury after non-cardiac surgery; P group: Control group (propofol-based intravenous general anesthesia); R group: Experimental group (remimazolam-based intravenous general anesthesia); ASA: American Society of Anesthesiologists; BMI: Body mass index.

Table 9 Comparison of perioperative serum biomarkers of myocardial injury between non-myocardial injury after non-cardiac surgery and myocardial injury after non-cardiac surgery patients, mean ± SD.

Item	Time	Non-MINS group (n = 56)	MINS group (n = 24)	t value	P value
hs-cTn (ng/mL)	T0	12.38 ± 12.40	12.58 ± 7.11	-0.077	0.939
	T8	10.80 ± 6.03	16.42 ± 4.42	-4.108	< 0.001
CK (U/L)	T0	44.52 ± 39.38	43.39 ± 34.49	0.121	0.904
	T8	58.66 ± 60.17	83.31 ± 99.74	-1.364	0.176
CK-MB (U/L)	T0	7.15 ± 4.40	8.51 ± 5.58	-1.167	0.247
	T8	7.07 ± 5.03	9.74 ± 5.39	-2.133	0.036
LDH (U/L)	T0	192.13 ± 73.07	180.15 ± 55.89	0.717	0.475
	T8	189.93 ± 75.86	206.18 ± 70.29	-0.897	0.372
AST (U/L)	T0	26.25 ± 30.89	22.66 ± 16.09	0.538	0.592
	T8	23.65 ± 10.89	26.00 ± 19.26	-0.692	0.491
NT-proBNP (pg/mL)	T0	139.27 ± 154.79	158.64 ± 130.65	-0.536	0.593
	T8	334.21 ± 286.67	512.11 ± 312.21	-2.477	0.015

hs-cTn: High-Sensitivity Troponin; CK: Creatine kinase; CK-MB: Creatine kinase-MB; LDH: Lactate dehydrogenase; AST: Aspartate aminotransferase; NT-proBNP: N-terminal pro-brain natriuretic peptide; MINS: Myocardial injury after non-cardiac surgery.

Table 10 Comparison of arterial blood gas parameters between non-myocardial injury after non-cardiac surgery and myocardial injury after non-cardiac surgery patients, mean ± SD.

Item	Time	Non-MINS group (n = 56)	MINS group (n = 24)	t value	P value
SBP (mmHg)	T0	155.93 ± 20.05	153.79 ± 13.73	0.476	0.636
	T6	143.61 ± 16.41	147.21 ± 21.15	-0.823	0.413
DBP (mmHg)	T0	79.61 ± 12.38	80.88 ± 11.99	-0.424	0.673
	T6	74.27 ± 10.48	76.79 ± 8.34	-1.045	0.299
HR (bpm)	T0	73.86 ± 11.17	72.17 ± 14.04	0.573	0.568
	T6	72.73 ± 11.95	71.79 ± 12.56	0.318	0.752
Glu (mmol/L)	T0	6.11 ± 1.54	6.47 ± 1.96	-0.891	0.376
	T4	7.16 ± 1.68	7.01 ± 1.32	0.401	0.689
Lac (mmol/L)	T0	0.99 ± 0.44	0.88 ± 0.44	1.022	0.310
	T4	0.74 ± 0.33	0.73 ± 0.33	0.117	0.907
PaCO2 (mmHg)	T0	41.14 ± 5.69	41.33 ± 7.22	-0.121	0.904
	T4	47.20 ± 6.24	47.46 ± 6.88	-0.163	0.871

SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HR: Heart rate; Glu: Glucose; Lac: Lactate; PaCO₂: Partial pressure of carbon dioxide; MINS: Myocardial injury after non-cardiac surgery.

Table 11 Comparison of surgical indicators between non-myocardial injury after non-cardiac surgery and myocardial injury after non-cardiac surgery patients, mean ± SD.

Item	Non-MINS group (n = 56)	MINS group (n = 24)	t value	P value
Operative time (minute)	159.64 ± 44.09	193.54 ± 73.40	-2.554	0.013
Intraoperative fluid volume (mL)	1484.82 ± 286.34	1522.92 ± 398.36	-0.483	0.631
Colloid volume (mL)	252.68 ± 250.35	145.83 ± 232.15	1.787	0.078
Crystalloid volume (mL)	1232.14 ± 315.71	1377.08 ± 456.83	-1.636	0.106
Intraoperative blood loss (mL)	50.71 ± 40.40	57.92 ± 38.90	-0.739	0.462
Intraoperative transfusion (mL)	10.71 ± 56.18	12.50 ± 61.24	-0.127	0.899
Urine output (mL)	310.89 ± 112.61	360.42 ± 181.77	-1.485	0.142
Norepinephrine consumption (μg)	172.89 ± 429.37	249.42 ± 305.35	-0.790	0.432
Extubation time (minute)	14.93 ± 9.15	15.88 ± 10.90	-0.400	0.690
PACU stay (minute)	43.93 ± 11.07	47.08 ± 12.06	-1.137	0.259
Postoperative hospital stay (day)	11.70 ± 2.21	12.83 ± 3.24	-1.824	0.072

PACU: Post-anesthesia care unit; MINS: Myocardial injury after non-cardiac surgery.

Multivariable logistic regression analysis of factors associated with MINS in elderly hypertensive patients undergoing gastrointestinal tumor surgery

MINS occurrence (yes = 1, no = 0) was defined as the dependent variable. Variables with P < 0.05 in the univariate analysis were included as independent variables in the multivariable logistic regression model. The results showed that tumor type, anesthesia method, NT-proBNP at T7, and NT-proBNP at T8 were independent predictors of MINS in elderly hypertensive patients undergoing gastrointestinal tumor surgery (Table 12).

Table 12 Multivariable logistic regression analysis of factors associated with myocardial injury after non-cardiac surgery.

Variable	B	SE	Wald χ2 value	P value	OR	95%CI
						Lower	Upper
Tumor type
Rectal cancer	4.281	1.675	6.534	0.011	72.348	2.714	1928.285
Gastric cancer	3.650	1.526	5.723	0.017	38.475	1.934	765.444
Gastrointestinal stromal tumor	24.031		0.000	1.000		0.000
Anesthesia method
General anesthesia + remimazolam (R group)	-3.188	1.305	5.972	0.015	0.041	0.003	0.532
hs-cTn
T6	-0.105	0.094	1.250	0.263	0.900	0.749	1.082
T7	0.285	0.176	2.624	0.105	1.330	0.942	1.877
T8	-0.041	0.181	0.050	0.823	0.960	0.674	1.368
CK-MB
T6	0.225	0.137	2.699	0.100	1.252	0.958	1.638
T7	-0.001	0.107	0.000	0.989	0.999	0.810	1.230
T8	-0.152	0.140	1.182	0.277	0.859	0.652	1.130
NT-proBNP
T6	-0.004	0.004	1.313	0.252	0.996	0.988	1.003
T7	0.011	0.004	8.340	0.004	1.011	1.004	1.019
T8	-0.009	0.004	5.380	0.020	0.991	0.984	0.999
Operation time	0.015	0.008	3.060	0.080	1.015	0.998	1.032
Constant	-9.040	2.599	12.102	< 0.001	0.000

OR: Odds ratio; CI: Confidence interval; hs-cTn: High-Sensitivity Troponin; CK-MB: Creatine kinase-MB; NT-proBNP: N-terminal pro-brain natriuretic peptide; R group: Experimental group (remimazolam-based intravenous general anesthesia).

Construction of a predictive model for MINS in elderly patients with hypertension undergoing gastrointestinal tumor surgery

Based on the independent predictors identified above, a multivariable logistic regression model was constructed to estimate the probability of MINS. In this model, the dependent variable remained the occurrence of MINS. The Y values represent the logit transformation (log-odds) of the predicted probability of MINS.

The predictive equations are as follows: Logit (P) = -9.040 + 4.281 × (rectal cancer) - 3.188 × (anesthesia method) + 0.011 × NT-proBNP at T7 + 0.011 × NT-proBNP at T7 - 0.009 × NT-proBNP at T8. Logit (P) = -9.040 + 3.650 × (gastric cancer) - 3.188 × (anesthesia method) + 0.011 × NT-proBNP at T7 + 0.011 × NT-proBNP at T7 - 0.009 × NT-proBNP at T8, where P represents the probability of MINS. Tumor type and anesthesia method are coded as binary variables (rectal cancer = 1, others = 0; gastric cancer = 1, others = 0; remimazolam-based anesthesia = 1, conventional anesthesia = 0).

Validation of the predictive model for MINS in elderly patients with hypertension undergoing gastrointestinal tumor surgery

ROC curve analysis was performed to evaluate the predictive performance of individual variables and the multivariable model for MINS. The occurrence of MINS (yes = 1, no = 0) was defined as the state variable. Tumor type, anesthesia method, NT-proBNP at T7, NT-proBNP at T8, and the predicted probability of MINS derived from the logistic regression model were included as test variables. The results showed that NT-proBNP at T8 [AUC = 0.704, 95% confidence interval (CI): 0.592-0.801, P = 0.001] and the MINS predictive model (AUC = 0.765, 95%CI: 0.656-0.852, P < 0.001) demonstrated acceptable discriminatory ability (AUC > 0.7), whereas tumor type, anesthesia method, and NT-proBNP at T7 showed relatively limited predictive performance (AUC < 0.7). Among all variables, the predictive model achieved the highest AUC, indicating superior overall discrimination compared with any single predictor. The optimal cut-off value for NT-proBNP at T8 was 412.6 pg/mL, yielding a sensitivity of 58.33% and specificity of 76.79%. The predictive model showed high sensitivity (91.67%) but moderate specificity (50.00%), suggesting it is more suitable as a screening tool for identifying high-risk patients. Detailed ROC parameters are presented in Table 13, and the ROC curves are illustrated in Figure 4.

Figure 4 Receiver operating characteristic curve of the myocardial injury after non-cardiac surgery prediction model in elderly hypertensive patients after gastrointestinal tumor surgery. NT-proBNP: N-terminal pro-brain natriuretic peptide; MINS: Myocardial injury after non-cardiac surgery.

Table 13 Receiver operating characteristic curve parameters of the predictive model for myocardial injury after non-cardiac surgery in elderly patients with hypertension undergoing gastrointestinal tumor surgery.

Variable	AUC	95%CI		aP value	Cut-off	Sensitivity (%)	Specificity (%)
		Lower	Upper
Tumor type	0.658b	0.543	0.760	0.012	Non-colon cancer	87.50	50.00
Anesthesia method	0.679	0.565	0.779	0.006	General anesthesia + remimazolam	75.00	60.71
NT-proBNP at T7	0.680	0.567	0.780	0.007	647.2 pg/mL	54.17	80.36
NT-proBNP at T8	0.704	0.592	0.801	0.001	412.6 pg/mL	58.33	76.79
MINS predictive model	0.765	0.656	0.852	< 0.001		91.67	50.00

^aP < 0.05 vs the reference line (area under the curve = 0.500).

^bP < 0.05 vs the myocardial injury after non-cardiac surgery predictive model.

AUC: Area under the curve; CI: Confidence interval; NT-proBNP: N-terminal pro-brain natriuretic peptide; MINS: Myocardial injury after non-cardiac surgery.

DISCUSSION

The results of this study showed that CK-MB levels at T6 were significantly higher in the P group than in the R group (P < 0.05); at T7, hs-cTn, CK-MB, and AST levels in the P group were significantly higher than those in the R group (P < 0.05); and at T8, CK-MB and NT-proBNP levels in the P group were significantly higher than those in the R group (P < 0.05). These findings indicate that, in elderly patients with hypertension undergoing gastrointestinal tumor surgery under general anesthesia, the use of remimazolam, compared with propofol, can effectively suppress the postoperative elevation of myocardial injury markers. In addition, the incidence of postoperative MINS in the R group was significantly lower than that in the P group (15.00% vs 45.00%, P < 0.05), confirming that remimazolam exerts cardioprotective effects, attenuates perioperative myocardial injury, and reduces the risk of MINS, which is consistent with previous international studies[12,13].

hs-cTn, CK-MB, and NT-proBNP are widely used serum biomarkers for the clinical assessment of myocardial injury. Among them, hs-cTn is a core diagnostic marker for myocardial infarction with high specificity. As hs-cTn is exclusively expressed in cardiomyocytes, even minor myocardial injury can result in a measurable increase in circulating levels[14,15]. CK-MB typically rises 6-12 hours after myocardial injury and peaks at approximately 24 hours, making it valuable for the early diagnosis of myocardial infarction[16]. NT-proBNP, a peptide secreted by ventricular myocardium, is an important biomarker for the diagnosis and monitoring of heart failure. It is released in response to myocardial stretch or increased wall stress, and elevated circulating levels reflect impaired cardiac function[17]. Elderly patients with hypertension are chronically exposed to elevated blood pressure, leading to structural and functional remodeling of the cardiovascular system and an increased risk of cardiovascular events. In addition, the stress response induced by gastrointestinal tumor surgery can enhance the generation of reactive oxygen species, directly damaging cardiomyocyte membranes and mitochondria, thereby promoting myocardial necrosis and postoperative myocardial injury, ultimately contributing to the development of MINS.

Several animal studies have demonstrated that remimazolam can alleviate myocardial ischemia-reperfusion injury by modulating relevant signaling pathways. Previous study[18] reported that remimazolam may reduce cardiomyocyte injury caused by ischemia-reperfusion through inhibition of NLRP3 inflammasome activation. Remimazolam may inhibit pyroptosis induced by myocardial ischemia-reperfusion via the HMGB1-receptor for advanced glycation end products signaling pathway[19]. Xu et al[20] showed that remimazolam can attenuate ischemia-reperfusion injury by inhibiting the nuclear factor kappa-B pathway in macrophages. In addition, a clinical study[21] reported that remimazolam used for anesthesia induction significantly reduced the levels of myocardial injury markers such as hs-cTn and CK-MB at 24 hours and 72 hours postoperatively, which is consistent with the findings of the present study.

As a novel benzodiazepine, remimazolam possesses unique pharmacokinetic properties that form the basis of its cardioprotective effects[22]. Compared with the conventional intravenous anesthetic propofol, remimazolam is rapidly metabolized by tissue esterases, has a short elimination half-life, and does not accumulate active metabolites. These characteristics reduce the likelihood of prolonged circulatory depression due to drug accumulation in elderly patients, thereby decreasing myocardial ischemia and hypoxia caused by sustained hypotension[23,24]. Moreover, the regulation of the perioperative stress response also contributes to its cardioprotective effects. The occurrence of MINS is closely associated with the stress response induced by surgical trauma. Under stress conditions, sympathetic activation leads to excessive catecholamine release and increased myocardial oxygen consumption. In elderly patients with hypertension, reduced coronary reserve predisposes them to an imbalance between myocardial oxygen supply and demand, ultimately resulting in myocardial injury[25]. Zhang et al[26] demonstrated that, compared with propofol, patients receiving remimazolam had significantly lower levels of cortisol and norepinephrine at the end of surgery, suggesting that remimazolam may attenuate the stress response by inhibiting the hypothalamic-pituitary-adrenal axis and sympathetic activity, thereby reducing myocardial oxygen consumption.

The arterial blood gas analysis in this study showed that lactate levels at T0 were significantly higher in the P group than in the R group (P < 0.05), while SBP and DBP at T1 were significantly lower in the P group than in the R group (P < 0.05). These findings suggest that remimazolam not only improves perioperative hemodynamic stability and tissue perfusion in elderly hypertensive patients but may also contribute to better postoperative recovery.

In this study, SBP, DBP, and HR in both groups exhibited an initial decrease followed by a gradual increase across the time points from T0 to T6. This pattern can be attributed to the pharmacological effects of anesthetic agents, which reduce myocardial contractility by inhibiting calcium influx into cardiomyocytes, thereby decreasing cardiac output[27]. In addition, anesthetics can induce vasodilation particularly of arterial vessels leading to increased vascular capacitance and subsequent reductions in blood pressure and HR. During surgery, however, fluid administration and vasoactive agents are used to maintain hemodynamic stability, resulting in a gradual return toward baseline levels. Notably, patients receiving remimazolam demonstrated more stable intraoperative hemodynamics. This may be related to its modulation of central nervous system activity. Remimazolam selectively binds to gamma-aminobutyric acid (GABA)-A receptors, enhancing chloride influx and promoting neuronal hyperpolarization, thereby suppressing neuronal excitability. This effect may attenuate activity in the locus coeruleus and medullary sympathetic centers, leading to reduced sympathetic outflow to peripheral vessels, the heart, and the adrenal medulla. In addition, stress responses require integration by the limbic system and hypothalamus to activate the sympathetic nervous system; remimazolam may dampen this process by suppressing excessive limbic excitation and reducing hypothalamic responsiveness to nociceptive stimuli, thereby limiting the downstream conversion of stress signals into sympathetic activation[28-30]. Several studies have confirmed that, compared with propofol, remimazolam is more effective in maintaining hemodynamic stability during non-cardiac surgery[31,32]. Lactate is a sensitive indicator of the balance between oxygen supply and demand in tissues, and elevated levels usually indicate tissue hypoxia or inadequate perfusion[33]. In this study, lactate levels at T0 were already higher in the P group than in the R group, which may be related to baseline circulatory status; however, more importantly, lactate levels remained consistently lower in the R group during the intraoperative and postoperative periods, whereas they remained higher in the P group despite some reduction. This suggests that remimazolam may improve tissue perfusion and reduce lactate production by maintaining stable blood pressure.

Stable blood pressure is particularly critical in elderly patients with hypertension. Owing to the frequent presence of coronary atherosclerosis and myocardial hypertrophy in this population, myocardial perfusion is more dependent on arterial pressure. When blood pressure falls below the lower limit of autoregulation, coronary perfusion pressure declines markedly, predisposing to myocardial ischemia[34]. Both domestic and international guidelines recommend maintaining perioperative SBP within ± 20% of baseline in elderly hypertensive patients, as exceeding this range increases the risk of myocardial ischemia by 2-3 fold[35,36]. In the present study, the R group exhibited smaller fluctuations in SBP and DBP from T0 to T6, without episodes of significant hypotension. In contrast, in the P group, SBP decreased to 74.19% of baseline at T1 (116.48/157.00), exceeding the safe fluctuation range, which likely contributed to the higher incidence of postoperative MINS. Furthermore, blood pressure recovery at T4 and T5 was faster in the R group, reflecting the rapid metabolism and recovery profile of remimazolam, which may attenuate hemodynamic instability during extubation and reduce the risk of myocardial ischemia.

Regarding postoperative pain management, VAS scores at T7 and T8 were lower in the R group than in the P group, with a greater reduction observed from T7 to T8. This pattern suggests that remimazolam may be associated with attenuation of acute postoperative pain and a more rapid decline in pain intensity over time. Although remimazolam does not possess intrinsic analgesic properties, this observation may be related to a potential synergistic interaction with opioids. It has been proposed that remimazolam may enhance opioid-mediated inhibitory effects at the level of spinal dorsal horn neurons, thereby improving overall analgesic efficacy and potentially reducing opioid requirements and related adverse effects[37]. However, the current evidence remains inconsistent. Guo et al[38] reported no significant difference in postoperative VAS scores between remimazolam and propofol in elderly patients undergoing gastrointestinal endoscopy, despite improved hemodynamic stability and reduced respiratory depression in the remimazolam group. Similarly, Liu et al[39] found that while remimazolam reduced the incidence of postoperative delirium in elderly patients undergoing radical colorectal cancer surgery, it did not significantly affect postoperative pain intensity. These discrepancies may be attributable to differences in study design, perioperative analgesic strategies, or sample size, and suggest that the potential analgesia-related benefits of remimazolam should be interpreted with caution.

In addition, the analgesic benefit of remimazolam may be partly mediated by a reduction in postoperative agitation. In this study, the incidence of postoperative agitation was markedly higher in the P group (60.00%) than in the R group (10.00%). Agitation is often associated with muscle tension, tachycardia, and hypertension, which can exacerbate pain perception[40]. It has been shown that agitation-induced stress responses activate peripheral nociceptors and amplify pain signal transmission[41]. By stabilizing neuronal excitability during emergence, remimazolam reduces agitation and thereby indirectly alleviates pain, which may partly explain the lower postoperative pain scores observed in the R group[42].

Notably, no significant differences in glucose or PaCO₂ were observed between the two groups at T4, suggesting that remimazolam does not exert additional effects on glucose metabolism or respiratory function. Its impact on perioperative physiology appears to be primarily related to improvements in circulation and tissue perfusion, as reflected by changes in lactate levels and blood pressure.

Analysis of surgical variables showed that norepinephrine consumption and crystalloid infusion were significantly higher in the P group, whereas colloid use was greater in the R group. These findings indicate that remimazolam can reduce the requirement for norepinephrine in elderly hypertensive patients undergoing gastrointestinal tumor surgery, thereby lowering the risks associated with vasoactive agents and optimizing intraoperative fluid management, supporting its clinical safety.

From the perspective of vasoactive drug requirements, elderly hypertensive patients are prone to refractory hypotension during general anesthesia due to reduced vascular elasticity and impaired sympathetic regulation, particularly in the presence of surgical trauma and blood loss, often necessitating the use of α-adrenergic agonists such as norepinephrine to maintain blood pressure[43]. While norepinephrine effectively increases arterial pressure via peripheral vasoconstriction, it also enhances myocardial contractility and HR, leading to increased myocardial oxygen consumption. It has been reported that each additional 100 μg of norepinephrine increases myocardial oxygen consumption by 15%-20%, which may readily precipitate myocardial ischemia or even MINS in patients with limited coronary reserve[44]. The reduced norepinephrine requirement observed with remimazolam is primarily attributable to its more favorable hemodynamic profile. As discussed, propofol frequently induces pronounced vasodilation and myocardial depression, necessitating substantial crystalloid loading and vasopressor support to maintain target blood pressure. In contrast, remimazolam produces only mild reductions in systemic vascular resistance without suppressing myocardial contractility, resulting in less pronounced hypotension and reduced need for both fluid resuscitation and vasoactive drugs[45]. Fechner et al[46] demonstrated that remimazolam used for both induction and maintenance of general anesthesia significantly decreased norepinephrine requirements, consistent with the present findings. Moreover, their study reported more stable fluid balance in the remimazolam group, with a 25% reduction in crystalloid use and an 18% increase in colloid administration, aligning with our observation of higher crystalloid and lower colloid use in the P group. These results suggest that remimazolam not only reduces vasopressor requirements but also facilitates more balanced fluid management, potentially minimizing tissue edema and cardiopulmonary burden associated with excessive crystalloid infusion. Another important mechanism involves modulation of sympathetic activity. Excessive perioperative sympathetic activation contributes to increased vasopressor demand, whereas remimazolam may attenuate sympathetic outflow via inhibition of the locus coeruleus, thereby reducing endogenous norepinephrine release and stabilizing peripheral vascular resistance[47]. Additionally, the rapid metabolism of remimazolam allows precise titration based on BIS monitoring, avoiding drug accumulation and excessive sedation-induced circulatory depression. Although both groups maintained BIS values between 40 and 60, the slower metabolism of propofol often necessitates higher plasma concentrations to prevent intraoperative awareness, resulting in sustained circulatory suppression and increased norepinephrine use. In contrast, flexible adjustment of remimazolam infusion enables maintenance of adequate anesthetic depth while minimizing hemodynamic compromise, which likely contributes to its lower norepinephrine requirement.

Although colloid administration was significantly higher in the R group than in the P group, this does not indicate an increased burden of fluid management. Owing to their superior and sustained volume-expanding effects, appropriate use of colloids can reduce crystalloid requirements and mitigate the risk of tissue edema.

The incidence of postoperative complications (except for agitation) and 30-day adverse events did not differ between groups (12.50% vs 10.00%). Notably, agitation scores were higher in the R group, while the incidence of agitation was markedly lower compared with the P group (10.00% vs 60.00%). These findings suggest that remimazolam is safe in elderly hypertensive patients undergoing gastrointestinal tumor surgery, without increasing postoperative complications or short-term adverse events, while substantially reducing postoperative agitation.

The favorable safety profile of remimazolam appears to be related to its distinct pharmacokinetic and pharmacodynamic properties. It is rapidly metabolized by tissue esterases into inactive metabolites and does not depend on the hepatic cytochrome P450 enzyme system, resulting in fewer drug-drug interactions and potentially more predictable pharmacological effects in elderly patients and those with hepatic or renal dysfunction[48]. However, dose adjustment may still be required in clinical practice depending on patient condition. Takaki et al[49] reported comparable rates of postoperative complications between remimazolam and propofol during anesthetic induction, with a lower incidence of hypotension in the remimazolam group. A meta-analysis[50] similarly suggested that both agents are safe for intravenous anesthesia in endoscopic procedures, while remimazolam may offer advantages in hemodynamic stability, which is broadly consistent with our findings. In the present study, no significant differences were observed between groups in delayed recovery, delirium, hypoxemia, or postoperative nausea and vomiting, suggesting comparable recovery profiles and gastrointestinal tolerance. These findings are in line with those reported by Kim et al[51], who found no significant differences in extubation time or postoperative delirium in elderly patients undergoing gastrointestinal tumor surgery. Importantly, the inactive metabolites of remimazolam lack sedative activity, reducing the likelihood of residual sedation and potentially lowering the risk of postoperative delirium, particularly in elderly patients[52].

Postoperative agitation is a common complication in elderly patients after general anesthesia, with an incidence of up to 30%-60%, and may lead to hypertension, tachycardia, wound dehiscence, and accidental catheter removal, thereby adversely affecting prognosis. In the present study, the incidence of agitation in the R group was only 10.00%, significantly lower than the 60.00% observed in the P group. This advantage may be attributed to the subtype selectivity of remimazolam for the GABA-A receptor. Among its subtypes, α1 is primarily associated with sedation, whereas α2 is linked to anxiolytic and anti-agitation effects. Remimazolam exhibits higher affinity for the α2 subtype than propofol, enabling more effective suppression of emergence-related anxiety and agitation while maintaining adequate sedation[53]. Regarding longer-term safety, no difference in 30-day adverse events was observed between groups, suggesting that remimazolam does not increase short-term postoperative risk. Although one case of postoperative delirium occurred in the P group and none in the R group, the difference was not statistically significant; nevertheless, this finding suggests a potential trend toward reduced delirium with remimazolam. This observation is consistent with previous reports[54] and may be related to its minimal impact on the cholinergic system. Propofol can inhibit central cholinergic activity, whereas cholinergic dysfunction is a key mechanism underlying postoperative delirium; in contrast, remimazolam exerts less inhibitory effect on this system, potentially reducing delirium risk[55]. Larger studies are required to confirm this trend.

Univariate analysis showed that the proportion of colon cancer was higher in the non-MINS group (50.00% vs 12.50%), whereas rectal cancer was less frequent (17.86% vs 45.83%) compared with the MINS group. Multivariate analysis further identified rectal cancer [odds ratio (OR) = 72.348, 95%CI: 2.714-1928.285, P = 0.011] and gastric cancer (OR = 38.475, 95%CI: 1.934-765.444, P = 0.017) as independent risk factors for postoperative MINS compared with colon cancer. These findings are closely related to anatomical characteristics, surgical invasiveness, and intraoperative hemodynamic perturbations. Rectal cancer surgery involves complex pelvic dissection, manipulation of autonomic nerves, and traction of vascular plexuses, which can provoke sympathetic activation, leading to abrupt increases in HR and blood pressure. In elderly hypertensive patients with reduced vascular compliance and heightened dependence of myocardial perfusion on blood pressure, such fluctuations can disrupt the myocardial oxygen supply-demand balance and precipitate ischemia. In addition, gastric cancer surgery was associated with longer operative duration (192.5 ± 68.3 minutes vs 156.2 ± 45.7 minutes for colon cancer), which prolongs anesthetic exposure and complicates fluid management. Consistently, operative time was longer in the MINS group in univariate analysis, and prolonged hemodynamic stress may cumulatively exacerbate myocardial injury, increasing the likelihood of MINS[56-59].

Multivariate analysis demonstrated that, compared with remimazolam-based anesthesia, conventional propofol-based anesthesia was an independent risk factor for postoperative MINS (OR = 0.041, 95%CI: 0.003-0.532, P = 0.015), consistent with the marked difference in MINS incidence between groups. This difference is primarily attributable to their distinct effects on perioperative hemodynamic stability and myocardial protection. Propofol exerts stronger suppression of myocardial contractility and peripheral vascular tone. In this study, SBP and DBP at T1 were significantly lower in the P group, and the duration of intraoperative hypotension was longer than in the R group. In elderly hypertensive patients with limited coronary reserve, prolonged hypotension compromises coronary perfusion, leading to myocardial ischemia and hypoxia. In contrast, remimazolam, with rapid metabolism via tissue esterases, produces milder circulatory depression, reduces the duration of hypotension, and lowers the risk of myocardial injury. Regarding stress response modulation, propofol provides weaker suppression of the hypothalamic-pituitary-adrenal axis, making patients more susceptible to stress-induced hypertension and tachycardia during surgical stimuli. Although HR at T3 did not differ significantly between groups, blood pressure variability was greater in the P group, increasing myocardial oxygen consumption. Remimazolam, through enhanced affinity for the GABA-A α2 subtype, more effectively attenuates sympathetic activation and limits stress-related myocardial injury. Consistently, postoperative levels of CK-MB, hs-cTn, and NT-proBNP at T6-T8 were significantly higher in the P group; for example, CK-MB at T8 was 1.8-fold higher than in the R group, indicating more severe myocardial injury under propofol-based anesthesia. By mitigating myocardial injury, remimazolam directly reduces the risk of MINS[60-64].

An increase in NT-proBNP at T7 and a decrease at T8 were independent risk factors for postoperative MINS. ROC curve analysis showed that NT-proBNP at T8 predicted MINS with an AUC > 0.7, indicating that postoperative NT-proBNP levels are a reliable dynamic indicator for assessing MINS risk. Biologically, NT-proBNP is released by ventricular myocytes in response to increased myocardial load and ischemic injury, serving as a sensitive marker of myocardial stress and damage. In this study, NT-proBNP at T7 was significantly higher in the MINS group than in the non-MINS group and remained elevated at T8, whereas in the non-MINS group it showed a downward trend at T8. This suggests that persistently elevated postoperative NT-proBNP indicates ongoing or unrepaired myocardial injury and warrants vigilance for MINS. From a clinical monitoring perspective, NT-proBNP testing is convenient, and its trend reflects myocardial status in real time. In this study, patients with NT-proBNP > 412.6 pg/mL at T8 had a significantly higher incidence of MINS (51.85%, 14/27) compared with those with NT-proBNP < 412.6 pg/mL at T8 (18.87%, 10/53). Therefore, monitoring NT-proBNP at 48 hours postoperatively can serve as an early screening method for MINS. For patients exceeding the cut-off value, further evaluation with electrocardiography and cardiac enzyme profiling is recommended, along with timely intervention to improve prognosis[65,66].

A predictive model incorporating tumor type, anesthetic technique, NT-proBNP at T7, and NT-proBNP at T8 demonstrated good discriminatory performance, with an AUC of 0.765 (95%CI: 0.656-0.852, P < 0.001), sensitivity of 91.67%, and specificity of 50.00%. This model has clear clinical utility. Preoperatively, combining tumor type (rectal/gastric vs colon cancer) with the planned anesthetic technique (propofol vs remimazolam) allows preliminary risk stratification; high-risk patients (e.g., rectal or gastric cancer with propofol anesthesia) may benefit from optimization of anesthetic strategy and preoperative management, including blood pressure control and correction of anemia. Postoperatively, targeted monitoring of NT-proBNP at 24 hours and 48 hours, integrated with model-based risk estimation, can facilitate early detection and intervention for MINS. The high sensitivity of the model enables identification of most high-risk patients, reducing missed diagnoses, while the relatively lower specificity can be complemented by clinical assessment and additional investigations such as electrocardiography to avoid overtreatment. Nevertheless, several limitations should be considered. First, certain potentially relevant variables, including preoperative cardiac function, intraoperative blood loss, and postoperative analgesia, were not incorporated. Although intraoperative norepinephrine use was higher in the MINS group in univariate analysis, it did not reach statistical significance but may still contribute to MINS risk; future models should integrate these variables to improve predictive performance. Second, the model relies on NT-proBNP measurements at 24 hours and 48 hours postoperatively and therefore cannot provide immediate preoperative prediction, limiting its applicability in urgent surgical settings; earlier predictive markers warrant further investigation.

CONCLUSION

In elderly hypertensive patients undergoing gastrointestinal tumor surgery, remimazolam-based general anesthesia demonstrates several clinically meaningful advantages. It attenuates myocardial injury and reduces the incidence of MINS, while maintaining more stable perioperative blood pressure. In addition, it is associated with lower postoperative pain scores and a markedly reduced risk of postoperative agitation. Remimazolam also decreases intraoperative norepinephrine requirements, thereby optimizing anesthetic management. Importantly, its use does not increase postoperative complications or 30-day adverse events, supporting a favorable safety profile. Furthermore, a predictive model integrating tumor type, anesthetic technique, and postoperative NT-proBNP levels enables early identification of patients at risk for MINS, providing a practical tool to guide perioperative management and improve clinical outcomes.