Efficacy and safety of remimazolam-based sedation for intensive care unit patients undergoing upper gastrointestinal endoscopy: a cohort study

Yuan-rui Zhao, Ke-sheng Huang, Guo Hou, Lan Yao, Li-ping Lu, Song Xu, Ying-tao Lian, Zhun Yao, Zhui Yu

Department of Critical Care Medicine, Renmin Hospital of Wuhan University, Wuhan 430060, China

KEYWORDS: Endoscopic sedation; Intensive care unit; Midazolam; Propofol; Remimazolam

INTRODUCTION

Emergency bedside endoscopic therapy, the best diagnostic and therapeutic procedure to diagnose and treat upper gastrointestinal bleeding (GIB), should be performed within 24 h under intensive monitoring in patients at a high risk of active bleeding and those with strong signs of recent re-bleeding, after appropriate resuscitation.[1,2]Abnormal coagulation, anemia, azotemia, and malnutrition caused by massive hemorrhage and the interaction of these conditions may increase the likelihood of cardiopulmonary failure, clotting anomaly, shock, and even death.[3,4]Thus, it is essential to unequivocally establish that sedation and analgesia do not further increase the risk of eliciting adverse events (AEs).[5]

Midazolam is frequently used during endoscopic procedures and in the intensive care unit (ICU), but unfortunately, it has several drawbacks, such as accumulation, a variable sedation time, and a long-acting metabolite that delays the recovery of neuropsychiatric functions.[6,7]While propofol has recently acquired popular clinical use due to its rapid onset and offset of actions, common AEs such as pain on injection, vomiting, aspiration pneumonia, and cardiopulmonary depression should not be neglected.[8,9]

Remimazolam is an ultra-short-acting benzodiazepine that acts on the gamma-aminobutyric acid type A (GABAA) receptor to induce sedation,[10]and is predominantly metabolized by plasma carboxylesterase into the inactive compound CNS 7054. This unique non-liver- or kidneydependent metabolism makes remimazolam an ideal sedative for critically ill patients.[11-13]Previous studies have demonstrated a decreased incidence of hypotension, respiratory depression, and injection pain, as well as no substantial drug accumulation and prolonged sedative latency after long-term administration.[14-19]It has mostly been evaluated in elective endoscopic therapy, but neither the efficacy nor AEs of remimazolam during upper endoscopy in the ICU setting have been thoroughly investigated. The aim of the present study was to compare the efficacy and safety of remimazolam to propofol and midazolam in patients with upper GIB in the ICU setting. It was hypothesized that remimazolam treatment would offer appropriate procedure sedation without increasing the incidence of AEs.

METHODS

Study design

This single-center, retrospective, observational study compared the efficacy and safety of remimazolam to propofol or midazolam coupled with analgesic administration (sufentanil or remifentanil) in patients undergoing upper gastrointestinal endoscopy. Among 118 upper GIB patients admitted to our center between January 1, 2021, and June 30, 2022, 30 patients were not enrolled in the study according to the exclusion criteria. The present study was reviewed and approved by the clinical research ethics committee of our hospital; patient written informed consent was not needed. The inclusion and exclusion criteria are summarized, and the overall study fl ow is shown in Figure 1 and supplementary Table 1.

Figure 1. Flowchart of the study. GIB: gastrointestinal bleeding.

Procedures

All patients were required to complete preoperative preparation according to the consensus on the emergency diagnosis and treatment of acute GIB.[20,21]Protective endotracheal intubation was applied in patients at high risk (age ＞65 years, active hemorrhaging, consciousness disturbances, respiratory or circulatory failure, confirmed or at risk of aspiration), and endoscopy was performed as soon as the Richmond Agitation-Sedation Scale (RASS) scores reached ≥ -3. Patients were treated with different drugs based on their vital signs, underlying diseases, and a comprehensive assessment by attending clinicians. Since the main concern after admission was to stabilize the vital signs of patients, the weight data were not been fully recorded. All test drugs were administered through continuous intravenous infusion at mean infusion rates of: (1) remimazolam, 0.27 mg/(kg·h); (2) propofol, 0.67 mg/(kg·h); (3) midazolam, 0.05 mg/(kg·h); (4) opioids: sufentanil, 0.3 μg/(kg·h); and (5) remifentanil, 0.1 μg/(kg·h). Bedside monitoring to guide dose titration included measurements of heart rate, blood pressure, oxygen saturation of blood and respiratory rate in real time, and the sedation effect was evaluated to maintain the RASS score between -3 and 0 during the procedure. Vital signs and symptomatic condition changes were used to determine when to stop the drug infusion and remove the endotracheal tube. All endoscopic procedures were carried out by a professional endoscopist in our hospital. Throughout the observation period, vital signs and the occurrence of AEs were monitored and carefully documented.

Assessments

During the observation period, treatment-related AEs, such as delirium, dysphoria, vomiting, hypoxemia, and hypotension, were the primary outcome measures. Secondary outcomes included the time to extubation, length of ICU stay, use of vasoactive agents, average cost of sedative per case, and other severe complications (details are provided in supplementary Table 2).

Key definitions of this study were: (1) observation period: 2 h after the last dose of sedative; (2) prolonged use of vasoactive drug: vasoactive agent support over 24 h after initiation of sedatives; (3) hypoxemia: oxygen saturation below 94% within 2 h of the procedure; (4) hypotension: blood pressure decreased by 20% from that pre-procedure, systolic blood pressure (SBP) ≤80 mmHg (1 mmHg=0.133 kPa) or mean arterial pressure (MAP) ＜60 mmHg, whether or not vasoactive agents administered; (5) bradycardia: a heart rate decrease of 20% compared with the baseline heart rate or if medication was needed; (6) basic disease level: patients with underlying diseases ≤3 were regarded as level I and ≥4 were regarded as level II.

Statistical analysis

IBM SPSS (ver. 25.0) was used to conduct all statistical analyses. Sample estimation was not conducted for the study design. Categorical data and ranked data are presented as numbers (percentages), and any differences between them were evaluated using a Chi-square, Fisher’s exact or Mann-WhitneyU-test. Continuous data are presented as the mean and standard deviation, and any differences were assessed using the Mann-WhitneyU-test. The Shapiro-Wilk test was used to establish the normality of the data distribution, and Student’st-test was employed to analyze normally distributed data groups. The Mann-WhitneyU-test was used in analyses of normally distributed data. The missing values were filled utilizing the means of the substitution method. AP-value＜0.05 was considered to be statistically significant.

RESULTS

Baseline characteristics of patients

Among the 88 patients enrolled, 47 were treated with remimazolam (group I) and 41 with propofol or midazolam (group II). The mean age was 61±11.76 years, and 39 (44.32%) were ＞ 65 years. The proportion of males was 69.32%, and the number of patients with a history of smoking and drinking was 29 (32.95%) and 24 (27.27%), respectively. According to the definition of basic disease level, 57 (64.77%) patients were in level II, with more than three underlying diseases, and 27 (30.68%) had a previous history of GIB. Demographics and baseline characteristics between the two groups were well balanced for age, sex, underlying diseases, basic vital signs, and laboratory test results (P＞0.05, except blood urea nitrogen [BUN]), except that more patients in the group I had a history of GIB (40.42% vs. 19.51%,P=0.034) (details are shown in supplementary Table 3).

Drug dose

The mean dose of remimazolam administered was 7.90±4.84 mg, propofol 21.19±17.98 mg, and midazolam 3.08±2.17 mg. The mean dose of sufentanil administered was slightly higher in the group I than in the group II, but no statistical significance was observed (15.06±12.20 μg vs. 12.40±6.01 μg,P=0.513). Similar results were found when the remifentanil dosage was calculated (0.38±0.16 mg vs. 0.31±0.17 mg,P=0.902). After converting the dose of analgesics to the morphine equivalent analgesic strength of opioids, no significant difference was found (22.75±19.91 mg vs. 17.64±13.35 mg,P=0.801) (details are shown in supplementary Table 4).

Procedure-associated details

The mean time from the onset of upper GIB symptoms to patients receiving endoscopic therapy was shorter in the group I than in the group II (27.76±29.32 h vs. 54.04±77.84 h,P=0.346). The average procedure duration appeared to be shorter in the group I, but no statistically significant difference was observed (35.89±13.37 min vs. 44.51±21.68 min,P=0.133). Peptic ulcer disease (PUD) and esophagogastric variceal bleeding (EGVB) were diagnosed in 38 (43.18%) patients and 50 (56.82%) patients, respectively (P=0.112). Except for 3 patients (2 in the group I and 1 in the group II) who did not receive endotracheal intubation, the time to extubation was 15.00±9.75 h with remimazolam, with a trend of a 5-hour decrease compared with 20.59±18.71 h with propofol or midazolam, but the results were not significantly different (P=0.211). The need for prolonged use of a vasoactive drug was 17 (36.17%) patients in the group I and 10 (24.39%) in the group II, but no significant difference was found between the groups (P=0.235). Patients in the group I were treated in the ICU for an average duration of 5.40±2.93 d, whereas patients in the group II stayed for 4.63 ± 3.31 d, but there was no statistically significant difference (P=0.072). The average cost of sedative per case was significantly lower in the group I compared to the group II (RMB 16.07±10.58 yuan vs. RMB 24.37±15.46 yuan,P=0.016) (details are shown in Table 1).

Treatment-related AEs

Treatment-related AEs during the observation period are listed in Table 2. A total of 72 treatment-related AEs were observed: 35 in the group I and 37 in the group II. The incidence of delirium was 17.02% with remimazolam and 17.07% with propofol or midazolam (P=0.995). Nausea was observed in 11 (23.40%) patients in the group I and 8 (19.51%) in the group II (P=0.658). The incidence of vomiting was identical in groups I and II, with 12 patients in each group having this existing symptom (P=0.695). Concerning severe AEs (such as cardiac arrest, shock, and stroke), 2 patients in the group II (1 treated with propofol and 1 with midazolam) had a stroke (P=0.214), and 1 patient in the group II given midazolam had a cardiac arrest during the endoscopic treatment (P=0.466). Shock was observed in 11 patients, 4 (8.51%) in the group I and 7 (17.07%) in the group II (5 with propofol and 2 with midazolam,P=0.226). However, they were finally diagnosed with hemorrhagic shock rather than drug-induced shock after reviewing the medical records. Hypotension, hypoxemia or bradycardia were not detected during the observation period.

Table 1. Comparison of procedure-related data between the two groups

Table 2. Comparison of treatment-related AEs occurred during the observation period between the two groups

DISCUSSION

This study evaluated the efficacy and safety of sedation between remimazolam and traditional sedatives in the target sedation range for ICU patients with upper GIB. Several findings based on our study were as follows. First, compared to propofol or midazolam, the mean duration of endotracheal intubation showed a non-significant trend of a 5-hour reduction with remimazolam treatment. Second, the length of ICU stay did not increase, but the average cost of sedative per case was lower with remimazolam treatment. Third, the overall treatment-related AEs were also similar.

Sedation is recommended for patients undergoing an endoscopic procedure, and optimal sedative medicines should have a rapid onset, short-acting time, complete metabolism, quick recovery, and produce minimal residual sedation.[4]Midazolam is a good option for procedural sedation for both anesthesiologists and nonanesthesiologists since it has the benefit of having less effect on hemodynamic stability and respiratory functions.[22-24]The strong cardiopulmonary depression, together with lipid accumulation and propofol infusion syndrome, has constrained the wide usage of propofol.[25,26]Remimazolam has been used for procedural sedation as well as for induction and maintenance of general anesthesia.[27]However, remimazolam and other classic sedatives are not suitable for comparison with each other in terms of dosage. Their safety, side effects, anesthetic effects, and health economic cost per case are the key concerns.

Our study focused on patients with severe upper GIB. Underlying diseases, especially liver or renal dysfunction, are likely to impact the pharmacokinetics and pharmacodynamics of the administered medicines. As previously reported, remimazolam is rapidly degraded into an inactive metabolite (CNS 7054) by plasma and tissue carboxylesterases, which accounts for its rapid removal even after prolonged infusions.[28]Due to its ultra-short action and organ-independent metabolism, remimazolam has been evaluated in American Society of Anesthesiologists (ASA) III or IV patients with no harm reported. This makes it an appropriate agent for neurological evaluations shortly after an infusion, and it possesses promising application value in the ICU,[29,30]although at present, no unequivocal conclusions have been reached. According to the results of our study, in which 57 (64.77%) patients had more than three underlying diseases, making them a group of patients prone to high-risk cardiopulmonary incidents, the AEs showed no significant differences between groups. The findings are in good agreement with a previous study that evaluated remimazolam safety in ASA IV patients undergoing highrisk colonoscopy, and no increased occurrence of AEs was found.[29]As far as we are aware, a prolonged duration of mechanical ventilation will increase the risk of ventilatorassociated pneumonia, which would have an impact on the survival rate of patients. The average duration of mechanical ventilation in the group I was 15 h, which was apparently 5 h shorter than that in the group II, but unfortunately, statistical significance was not observed. Future studies should provide more convincing findings about whether or not remimazolam can reduce tracheal intubation duration. Despite the lack of a significant difference, the nearly 1-day prolonged length of ICU stay following remimazolam treatment deserves moreattention. Future studies would be advised to substitute “time to meet criteria for ICU discharge” for “length of ICU stay” to obtain more convincing data. According to the price of sedatives used in our center (remizoram 52.8 yuan, propofol 178.3 yuan, midazolam 21.8 yuan), the average cost of remizoram sedation was significantly lower than that of classical sedatives. The remarkable advantages of health economics, combined with rapid onset and counteraction, independent metabolism of organs, sufficient sedation, and reduction of respiratory and circulatory inhibition, highlight the advantages of remimazolam in sedation in ICUs. Further research is needed to assess the risk benefit ratio of sedation in ICU.

The results from our study revealed no significant difference in terms of treatment-related AEs. Pastis et al[31]reported a rate of 34.7% related to treatmentrelated AEs in patients with remimazolam treatment compared to 25.4% in the placebo group and 31.9% in the midazolam group. However, ASA IV or very elderly patients were not included in their research, which limits its wider interpretation. The incidence of delirium or dysphoria was increased in our study. Several factors may have resulted in the elevated rate of delirium, such as decreased pre-existing consciousness, disease severity, depth and duration of sedation or mechanical ventilation. Mechanical ventilation and ICU admission were also associated with delirium and dysphoria,[5]but the precise mechanism has yet to be determined.[24]The severe AEs observed in our study included 11 (12.50%) patients with shock, 1 (1.14%) with cardiac arrest, and 2 (2.27%) with stroke. The incidence of serious AEs was comparatively higher than that in previous studies.[14,16,19]After careful examination of medical records, we found that the severe complications in these patients had a strong association with the underlying disease. Shock occurred in 4 (8.51%) patients in the group I and 7 (17.07%) in the group II during the observation period, but they were finally diagnosed as having hemorrhagic shock rather than drug-induced shock. Because of the severe underlying condition of the patients enrolled in our study, the results of severe complications likely do not accurately reflect the incidence of AEs caused by medication alone. In addition, the occurrence of symptoms related to dysfunction of the digestive tract cannot be simply attributed to adverse reactions to sedatives, which would lead to an overestimation of side effects. Large-scale randomized controlled trials are required to investigate this issue further in the ICU setting.

There were several limitations in the present study. First, the small sample size might have impacted the difference between groups. Large-scale trials are needed to further evaluate this concept to provide more convincing conclusions. Second, since the general applicability of current results is constrained by the specificity of the ICU patients and some baseline characteristics were not well balanced, potential confounding biases cannot be fully excluded. The fact that patients were not randomized into groups raises the possibility that different medications were selected according to non-measured variables, such as provider familiarity with medication or perceived patient stability.

CONCLUSION

Remimazolam did not increase the incidence of AEs and had lower average cost of sedative in high-risk patients with GIB during endoscopic therapy in the ICU, indicating that it may be a promising alternative for procedure sedation compared with classic sedatives. Further randomized controlled studies are required to provide more concrete conclusions.

Funding:This study was supported by China International Medical Foundation (Z-2017-24-2028-33).

Ethical approval:The study was reviewed and approved by the clinical research ethics committee of our hospital.

Conflicts of interest:The authors declare that they have no competing interests.

Contributors:YRZ contributed to data collection and analysis, wrote the initial draft of this manuscript, and performed further editing. KSH contributed to study design and data analysis. GH contributed to study design and data collection. LY contributed to writing assistance. ZY contributed to data collection. LPL contributed to language correction and writing assistance, and approved the final manuscript. SX contributed to language correction. YTL contributed to trial design and data collection. ZY contributed to study design and approved the final manuscript.

All the supplementary files in this paper are available at http://wjem.com.cn.