104 elderly patients were enrolled in the First Affiliated Hospital of Nanchang University for laparoscopic radical resection of gastric cancer from June 2022 to December 2022. Non-participants in the study used a computer to generate random numbers, randomly divided into three groups in a 1:1:1 ratio. The generated random sequence numbers were placed in opaque envelopes that were randomly selected and opened by another anesthesiologist who was not involved in the study to reveal the groupings before the patients were admitted to the operating room. To exclude bias due to different surgical practices and habits all patients were from the same group of surgeons. With 34 cases in the remimazolam group (one case developed anaphylaxis intraoperatively and was excluded), 35 cases in the dexmedetomidine group, and 35 cases in the saline group.

Inclusion criteria: (a) aged individuals (65–80 years old) receiving elective laparoscopic radical excision for gastric cancer; (b) American Society of Anesthesiologists (ASA) classification II-III; and (c) BMI 18–24 kg/m².

Exclusion criteria: (a) preoperative Mini-mental State Examination (MMSE) or Montreal Cognitive Assessment (MoCA) cognitive dysfunction (preoperative scores ≤17 for illiterate, ≤20 for primary, ≤22 for secondary, and ≤23 for university); (b) History of cardiac surgery, cerebrovascular accident, and alcoholism; (c) presence of serious organic diseases such as liver and kidney dysfunction; (d) coagulation disorders; (e) severe visual and hearing impairment that prevents communication; and (f) history of psychiatric disorders or long-term sedation or depression medication.

Rejection criteria: (a) extended surgical scope resulting in prolonged surgery time (operative time > 4.5 h); (b) patients who have severe intraoperative bleeding or develop drug allergies; and (c) subjects requesting termination of the test.

The study was authorized by the Ethics Committee of the First Affiliated Hospital of Nanchang University, Ethics Number IIT 2022-089, and patients consented and signed a full informed consent form. The trial was registered at the Chinese Clinical Trials Registry (ChiCTR2300068036). The study protocol followed CONSORT guidelines and was conducted within the relevant guidelines.

A five-lead electrocardiogram (ECG), noninvasive blood pressure (NIBP), heart rate (HR), blood oxygen saturation (SPO2), and depth of anesthesia monitoring (BIS) were all routinely monitored once the patient was brought to the room. Peripheral venous access was opened and radial artery puncture was performed under local anesthesia with 0.8 ~ 1% lidocaine.

Induction of anesthesia: Patients in all three groups were induced with cisatracurium 0.15 mg/kg (Batch Number H20060869, Jiangsu Hengrui Pharmaceutical Co., Ltd.), sufentanil 0.4–0.6 μg/kg (Yichang Renfu Pharmaceutical Co. Ltd., National Drug Standard H20054171) and propofol 1.0–1.5 mg/kg (Xi’an Libang Pharmaceutical Co., Ltd., National Drug Standard H20123318). On this basis, remimazolam mesylate 0.2 mg/kg (Batch Number H20190034, Jiangsu Hengrui Pharmaceutical Co., Ltd.) was added in group R. Patients in group D were given dexmedetomidine 200 μg (Yangzijiang Pharmaceutical Group Co., Ltd., National Drug Standard H20183219) diluted to 50 mL with saline (prepared concentration 4 μg/mL) and pumped 0.5 μg/kg over 10 min before induction of anesthesia. Saline 5 mL was added in group C. After the analgesic, sedative, and muscle relaxant had taken effect, endotracheal intubation was conducted. The PETCO₂ was maintained at 35–45 mmHg by adjusting the tidal volume and the respiratory rate.

Anesthesia maintenance: In all three groups, propofol 4–8 mg/kg/h + injectable remifentanil hydrochloride 4–8 μg/kg/h (Batch Number H20143315, Jiangsu Enhua Pharmaceutical Co., Ltd.) were pumped intravenously. On this basis, intravenous pumping of remimazolam 0.3–0.5 mg/kg/h was added in group R, stop pumping remimazolam 40 min before the end of surgery; intravenous pumping of dexmedetomidine 0.3–0.5 μg/kg/h was added in group D, and dexmedetomidine was halted 40 min before the conclusion of operation; intravenous pumping of saline 0.3–0.5 mL/kg/h was added in group C, stop pumping saline 40 min before the end of surgery. Cisatracurium 0.1 mg/kg was administered intravenously in all three groups to maintain muscular relaxation. Propofol was decreased to discontinuation before suturing skin, and remifentanil was discontinued at the end of surgery. Intraoperatively, vasoactive drugs were administered according to circulatory fluctuations, BIS values of 40–60 were maintained, and the total amount of propofol and remifentanil was recorded. Following anesthetic induction, all three groups of patients received bilateral transverse abdominal plane blocks (TAPB) combined with rectus abdominis sheath blocks (The Anesthesia flow chart is shown in Figure 1).

Postoperative analgesic treatment plan: Sufentanil 5 μg was given as an analgesic bridge 10 min before the end of surgery. When patients were transferred to the PACU, they were given intravenous self-administered analgesia for 48 h. At a rate of 2 mL/h and a PCA volume of 1 mL/time, all three groups of analgesic pumps included hydromorphone 6 mg, ondansetron 32 mg, and flurbiprofen axetil 200 mg.

The data were analyzed using SPSS25.0 statistical software and GraphPad Prism 8.0.2 was used for plotting. The pre-test indicated a 12% POCD incidence in the R group. Based on literature searches, expected the incidence of POCD in group D is 11%, group C is 39%. Assuming that Class I error α was assumed to be bilateral 0.05, the test efficacy (1-β) was assumed to be 0.8 and the lost visit ratio was set to 0.1, resulting in 35 cases in each group. Normal or near-normal measures were expressed as mean ± standard deviation ( X ¯  ± SD), and one-way analysis of variance was used for the comparison of data between groups, and the Bonferroni test was used for two-way comparison between groups. Comparisons of measures at different time points were analyzed using repeated measures ANOVA, and results of within-subjects effect tests were used for compliance with the sphericity test and multivariate tests for non-compliance with the sphericity test. Data that did not conform to normal distribution were described by median and interquartile spacing, and rank sum test was used for comparison of means between groups. Categorical data were reported as n (%), and the Chi-squared test was used to compare count data. A statistically significant difference is shown by p < 0.05.

The three groups did not differ in terms of age, gender, BMI, years of education, time of surgery, preoperative hemoglobin, or preoperative comorbidities (Table 1).

At T₂, T₃, and T₄ time points, group R had greater HR and MAP than group D and C (p < 0.05), and from T₁ to T₂ time points, HR and MAP dropped considerably in group D and C patients (p < 0.05); from T₂ to T₃ time points, HR rose in group D and C patients (p < 0.05). HR and MAP dropped in all three groups between T₃ and T₄ (p > 0.05). The incidence of intraoperative hypotension was highest in group D and lowest in group R (p < 0.05) (Tables 2–4 and Figure 2).

Extubation and PACU residence times were not significantly different between groups R, D, and C (p > 0.05), and neither were the propofol and remifentanil dosages between groups R and D (p > 0.05). However, the propofol and remifentanil dosages in groups R and D were lower than those in group C, and the difference was statistically significant (p < 0.05) (Table 5 and Figure 3).

S100β and TNF-α concentration between groups R and D at each time point were not statistically significant (p > 0.05). At times T₄, T₅, and T₆, group C had higher S100β and TNF-α concentrations than did groups R and D. These differences were statistically significant (p < 0.05) (Tables 6, 7 and Figure 4).

On the first preoperative (T₀) day for all three groups, there were no statistically significant changes in the MMSE and MoCA test results (p > 0.05). The MMSE and MoCA test results for groups R and D at 3 days postoperative (T₆) and 7 days postoperative (T₇) were not statistically significant (p > 0.05), although both groups scored higher than group C and the difference was statistically significant (p < 0.05) (Tables 8, 9 and Figure 5).

At 3 days postoperative (T₆) and 7 days postoperative (T₇), there was no statistically significant difference (p > 0.05) in the incidence of POCD between groups R and D, even though both were lower than the incidence of POCD in group C, which was statistically significant (p < 0.05) (Table 10).

There was no significant difference in VAS scores between groups R and D after 24 h postoperatively (p > 0.05), although both had lower scores than group C, and the difference was statistically significant (p < 0.05). The VAS scores between the three groups at 72 h (T₆) and 7 days (T₇) were not statistically significant (p > 0.05).In group R, there was no postoperative respiratory depression, hypotension, or bradycardia, and the incidence of postoperative adverse events was considerably lower (p < 0.05). The incidence of adverse events was greatest in group C and second highest in group D (Tables 11, 12 and Figure 6).

According to the findings of this investigation, remimazolam induced a slight rise in HR in group R at time T₂, which is consistent with the findings of Schüttler J in healthy male volunteers with continuous infusion of remimazolam (Schüttler et al., 2020). Among the three groups, group R had the least variation in HR and MAP, the lowest incidence of intraoperative hypotension, and used fewer vasoactive drugs intraoperatively, indicating that remimazolam has the least effect on circulation and can maintain more stable hemodynamics, which is more suitable for elderly patients. In patients undergoing flexible bronchoscopy, remimazolam was shown to have non-inferior efficiency, superior time metrics, and hemodynamic stability when compared to dexmedetomidine (Chen et al., 2022).

TNF-α has potent anti-tumor and anti-infection effects at normal concentrations, while excessive TNF-α production can lead to organ dysfunction or death (Boehm et al., 2014). It can cause the body to produce more cytokines (IL-6, IL-8), enhance the production of neurotoxic chemicals, and hasten the destruction of neurological tissue by prompting neutrophils to accumulate (Li et al., 2015). TNF-α was able to reach a peak shortly after the surgery began and subsequently fall, a change that preceded all other inflammatory factors (Bittar et al., 2006). TNF-α concentrations in groups R and C were considerably lower than those in the control group at T₄, T₅, and T₆ time points, owing to remimazolam and dexmedetomidine inhibiting the NF-κB signaling pathway and reducing TNF-α production (Shi et al., 2022), as well as attenuating the associated impaired synaptic plasticity and loss of connectivity in the hippocampus (Alam et al., 2018). The incidence of POCD has decreased.

With the modern notion of speedy recovery, laparoscopy is currently the primary project for treating gastric cancer. The CO2 pneumoperitoneum of laparoscopic surgery can cause a decrease in ventilation as well as a disruption in the internal environment’s acid–base balance, compromising the patient’s lung function and perhaps causing lung damage (Liu et al., 2019). Liu et al. (2021) showed that remimazolam has a role in perioperative lung protection, and it effectively reduced LPS-induced serum pro-inflammatory cytokines production and attenuated multi-organ damage, including lung and liver, which may be related to preventing MAPK and PI3K signaling pathway activation.

S100β protein is a calcium-binding protein, mainly produced by astrocytes. When the central nervous system is injured, S100β protein escapes from the damaged cells into the cerebrospinal fluid and then into the blood through the damaged blood–brain barrier (Michetti et al., 2019). Tan et al. (2021) discovered that POCD development was strongly connected with postoperative S100β expression, and hence S100β protein is frequently utilized as a diagnostic for brain damage (Hu et al., 2023). We discovered that the serum concentration of S100β in groups R and D was significantly lower than in group C at time points T₄, T₅, and T₆, because remimazolam and dexmedetomidine attenuated the inflammatory response, inhibited astrocyte and microglia activation, decreased S100β protein release, attenuated the interaction between neurons and glial cells, and protected brain cells and neural tissue from inflammatory factors, maintains intracranial homeostasis, and played a role in brain protection (Wang N. et al., 2022).

Because a single neurological scale test does not adequately indicate changes in the patient’s cognitive function. The MMSE is the most widely used cognitive test, but suffers from ceiling effects and has low discrimination and sensitivity for diagnosing moderate cognitive decline, especially in highly educated individuals (Ciesielska et al., 2016). The MoCA was developed as a rapid screening tool for moderate cognitive decline in order to get over the limitations of the MMSE (Wang et al., 2015). On postoperative days 3 and 7, substantial improvements in cognitive function were seen in groups R and D compared to group C. Among the subtests, there were significant differences in the results for visuospatial and executive functions, memory, and orientation in group C patients. The reason for the lack of differences in the other subtests may be related to the different cognitive aspects and the difficulty level of the test scales.

The findings of this experiment revealed no significant differences in the time to extubation and the length of stay in the PACU between groups R, D, and C. Remimazolam has a terminal half-life of 0.75 h, and the metabolite has no pharmacological effect, therefore adding remimazolam to anesthesia maintenance and discontinuing it 40 min before the end of surgery does not prolong anesthesia waking time. Choi et al. compared total intravenous anesthesia with remimazolam and propofol on quality of postoperative recovery (QoR), and discovered that remimazolam was not inferior to propofol as measured by the QoR-15 score, and that pain intensity and analgesic requirements were lower in the remimazolam group (Choi et al., 2022). Multiple randomized clinical trials have demonstrated remimazolam’s safety and efficacy as a sedative and general anesthetic (Chen et al., 2021; Lee et al., 2022).

Gastric cancer patients are often in a state of immune suppression during the disease, and immune suppression is more pronounced in the advanced stages of the disease (Hurwitz and Watkins, 2012). The stress response induced by surgical trauma and opioid usage further suppresses the immune system, and side effects such as delayed recovery, respiratory depression, nausea and vomiting raise the risk of anesthesia in patients. Opioids have also been associated with increased cognitive dysfunction, and preclinical trials have shown that opioids can trigger a strong neuroinflammatory response through the Toll-like receptor 4 (TLR4) from being activated, leading to POCD (Muscat et al., 2021). TLR4 expression is significantly increased in the hippocampus of aging rodents compared to young rats, so older adults are more likely to develop POCD (Fonken et al., 2016). Groups R and D dramatically lowered the dosage of propofol and remifentanil in this trial, which might alleviate patients’ immune suppression after surgery, and the incidence of postoperative adverse reactions was greatly reduced. After 24 h postoperatively, groups R and D had lower VAS scores than group C. Postoperative healing was aided by the decrease in postoperative pain. Dexmedetomidine produces analgesia by acting on presynaptic and postsynaptic α2 receptors in the posterior horn of the spinal cord and the nucleus of the central solitary tract, cutting off pain signaling pathways, reducing the release of substance P and other injurious sensory neurotransmitters. Remimazolam, on the other hand, can lower the dose of remifentanil and reduce postoperative pain by inhibiting pro-inflammatory factor production and modulating bradykinin B1 receptors and autophagy to relieve neuropathic pain (Xie et al., 2021). Furthermore, there was no postoperative hypotension or respiratory depression in group R. This might be because remimazolam has the ability to regulate hemodynamics while also providing fast anesthesia and arousal with a less inhibitory impact on breathing.

Remimazolam is similarly beneficial as dexmedetomidine in lowering the incidence of early POCD in aged patients after radical gastric cancer resection, probably due to reduced inflammatory response.

This study did not conduct long-term follow-up of patients, missing statistics on the incidence of POCD after 7 days following the operation, and only two indicators, S100β and TNF-α, were assessed. Also, larger sample sizes allow for the exclusion of more confounding variables, which makes the results appear more precise. Later the group will expand the sample size, extend the study period, evaluate more specific proteins and inflammatory factors associated with POCD, use a variety of combination scales to assess the occurrence of POCD, increase the sensitivity, and conduct in-depth research on the pathogenesis of POCD.