Introduction

Remimazolam is a recently introduced benzodiazepine anesthetic that targets γ-aminobutyric acid type A (GABA_A) inhibitory neurotransmitter receptors in the central nervous system (CNS)1. Structurally similar to midazolam, remimazolam differs by featuring a distinctive carboxylic ester linkage, which allows for dose-dependent hydrolysis by non-specific tissue esterases, bypassing hepatic metabolism as required by midazolam1,2. Additionally, compared with the metabolites of midazolam, those of remimazolam have a 400-fold lower affinity for GABA_A receptors, resulting in pharmacological inactivity1.

The organ-independent metabolism of remimazolam, along with the production of inactive metabolites and its reversibility, results in a favorable pharmacokinetic profile characterized by an ultra-short half-life and a reduced risk of prolonged sedation, facilitating enhanced recovery after surgery (ERAS)1. These properties, combined with an increasing preference for outpatient surgical procedures, have driven the rapid global adoption of remimazolam, a trend expected to continue1.

Postoperative nausea and vomiting (PONV) is a prevalent and distressing complication, affecting 10–30% of surgical patients, with incidence rates as high as 80% in high-risk populations3. PONV is associated with postoperative complications, increased healthcare costs, and extended hospital stays, thus underscoring the importance of PONV prevention, particularly within ERAS protocols in contemporary surgical care4. In this context, the selection of anesthetics that can contribute to PONV prophylaxis is critical in clinical settings5.

Previous studies have demonstrated that benzodiazepine administration before anesthesia, during induction, or at the end of surgery can significantly reduce PONV4,6,7. In particular, midazolam has been shown to be effective in managing acute refractory emesis during chemotherapy4,6,7. However, compared with other benzodiazepines, the effect of remimazolam on PONV remains unclear. Therefore, in this study, we performed a meta-analysis of randomized controlled trials (RCTs) to evaluate the efficacy of remimazolam compared with other anesthetic agents in preventing PONV.

Methods

Search strategy

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and was registered in the PROSPERO database (ID CRD42024511038). Two independent investigators (S.J. and H.J.) systematically searched PubMed, Cochrane Library, and Embase for relevant studies published up to January 8, 2024, using the search strategy outlined in Supplemental Table S1−3.

Study selection

After removing duplicates, two reviewers (S.J. and H.R.) independently screened the titles and abstracts of identified articles to assess their eligibility. The inclusion criteria were: (1) trials comparing “remimazolam (remimazolam group)” with “other anesthetic agents (comparator group)” in the context of general anesthesia or sedation, (2) trials involving adult patients (≥ 18 years), and (3) trials reporting the incidence of postoperative nausea, postoperative vomiting, or PONV as primary or secondary outcomes. The initial search included both RCTs and observational studies; however, only RCTs were included in the final analysis to ensure greater result clarity. The exclusion criteria were: (1) studies that used a placebo as the control group and (2) articles not published in English. Subsequently, the reviewers (S.J. and H.R.) independently conducted full-text reviews and data extraction. Disagreements regarding study inclusion or exclusion were resolved through consensus or by consultation with a third reviewer (C.K.).

Outcome and risk-of-bias assessment

Data were extracted directly from the original text, tables, or figures of the articles. The following variables were collected: author, publication year, journal, country, study design, patient demographics, surgical procedure, type of anesthesia (sedation or general anesthesia), type and dosage of anesthetics, sample size, incidence of outcomes (postoperative nausea, postoperative vomiting, or PONV), and timing of outcome measurement.

The primary outcome was PONV incidence in the remimazolam group compared with the comparator group. Given that few patients experience vomiting without nausea, the incidence rates of postoperative nausea and PONV are generally comparable8. Therefore, when only nausea was reported without data on PONV, the incidence of PONV was regarded as equivalent to that of postoperative nausea8. In cases where only postoperative vomiting was reported, this was interpreted as the incidence of PONV9. Most studies did not distinguish between nausea and vomiting, instead reporting them collectively as PONV. However, in trials that separately reported postoperative nausea or vomiting, these outcomes were analyzed independently in subgroup analyses.

In one study, the total number of PONV events per patient was reported on a scale of 0 (none) to 310. An occurrence of 0 was regarded as the absence of PONV, whereas occurrences ranging from 1 to 3 were defined as the presence of PONV events for our analysis. For studies reporting outcomes at different time points, the earliest postoperative assessment was used9. Two researchers (S.J. and C.K.) independently assessed the risk of bias in the included studies using the Cochrane Risk of Bias 2 (RoB 2) tool for RCTs.

Data synthesis and subgroup analysis

All analyses were performed using R version 4.0.4 (R Foundation for Statistical Computing, Vienna, Austria), MedCalc version 18.11.6 (MedCalc Software bvba, Ostend, Belgium), and Comprehensive Meta-Analysis version 4.0 (Biostat Inc., Englewood, NJ, USA).

Data are expressed as the relative risk (RR) with 95% confidence interval (95% CI) and illustrated using forest plots. A random effects model was applied for overall group analysis to account for anticipated trial heterogeneity, including differences in the subjects, type of surgery, and use of anesthetics in the comparator group.

Subgroup analyses were conducted based on the following factors: (1) type of anesthetic used in the comparator group (propofol, inhalation agents, dexmedetomidine, midazolam, or etomidate), (2) type of anesthesia (sedation or general), (3) type of procedure (gastrointestinal [GI] endoscopy, bronchoscopy, or other surgical procedures), (4) participant age (adults [> 18 years] or elderly only), (5) participant sex (all or female only), (6) outcomes (postoperative nausea, postoperative vomiting, or PONV), and (7) remimazolam administration protocol (continuous infusion or intermittent bolus).

A meta-regression analysis was conducted to evaluate the effects of mean age, sex distribution, and coadministered opioids on the incidence of PONV. Coadministered opioids were classified according to the dominant agent; when both single-dose and continuous-infusion opioids were used, the infusion agent was used for classification. Sensitivity analysis was also performed on trials that designated PONV as the primary outcome, which was determined based on whether PONV was specified as the primary endpoint in the sample size calculation.

Heterogeneity was assessed using the I² statistic. For low heterogeneity (I² < 50%), a fixed effects model (Mantel-Haenszel method) was used. For moderate heterogeneity (50% ≤ I² < 75%), a random effects model was applied. Data were not pooled for high heterogeneity (I² > 75%). Publication bias was evaluated using funnel plots and Egger’s test. A P-value less than 0.05 was considered statistically significant.

Results

Literature search

An initial search yielded 319 articles from the Medline, Embase, and Cochrane Library databases. After removing 68 duplicates, the titles and abstracts of the remaining 251 unique studies were reviewed. Of these studies, 63 studies were subjected to a full-text review, resulting in the identification of 50 eligible RCTs9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58. Excluded studies after the full-text review included 7 retrospective studies, 3 studies with inappropriate designs, and 3 studies with inadequate endpoints (Fig. 1). The PRISMA checklist is provided in Supplementary Table S4.

Fig. 1
figure 1

PRISMA study selection flow diagram.

Full size image

Study characteristics

The eligible studies, published between 2018 and 2023, involved a total of 9,193 participants9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58. Table 1 presents a summary of the key characteristics of the included RCTs. Among these trials, 10 RCTs focused on geriatric populations, with the elderly cohort defined by a varied age range from over 60 to over 75 years19,20,22,24,34,41,45,47,53,56. Nine RCTs specifically included female patients undergoing breast, thyroid, obstetric, or gynecologic surgery15,23,25,35,36,48,49,54,55. The comparators used for evaluating remimazolam in the studies were as follows: propofol in 37 RCTs9,11,13,14,15,16,17,18,19,20,21,22,23,24,26,28,29,30,31,33,35,36,37,39,40,41,43,48,49,50,51,52,53,55,56,57,58inhalation agents in 4 RCTs25,42,45,54dexmedetomidine in 5 RCTs10,27,32,34,47midazolam in 3 RCTs12,44,46and etomidate in 1 RCT38. A total of 24 trials enrolled patients receiving general anesthesia9,13,14,23,24,25,29,30,34,35,36,38,39,41,42,43,45,50,51,52,53,54,55,58and 26 trials included patients receiving procedural sedation10,11,12,15,16,17,18,19,20,21,22,26,27,28,31,32,33,37,40,44,46,47,48,49,56,57. In addition, 17 trials focused on patients undergoing GI endoscopy11,12,16,17,18,19,20,21,22,26,28,29,31,33,37,56,574 trials involved patients undergoing bronchoscopy9,10,39,44and 29 trials included patients undergoing other surgical procedures13,14,15,23,24,25,27,30,32,34,35,36,38,40,41,42,43,45,46,47,48,49,50,51,52,53,54,55,58. Notably, none of the trials targeted patients undergoing cardiac surgery. Among the reviewed RCTs, 33 trials collectively reported PONV as an outcome. Moreover, 18 trials specifically focused on postoperative nausea11,12,13,16,17,19,23,31,33,36,40,43,44,46,50,53,54,57and 17 trials examined postoperative vomiting9,11,12,13,16,19,23,31,33,36,40,43,44,46,50,54,57.

Table 1 Study characteristics.
Full size table

Risk of bias

The risk of bias for individual studies is presented in Fig. 2 and Supplemental Figure S5. In bias assessment, “some concerns” were identified in 3 RCTs for randomization, 2 RCTs for deviations from intended interventions, 3 RCTs for outcome measurement, and 19 RCTs for selective reporting. All included RCTs demonstrated a low risk of bias for missing outcome data. Overall, 25 trials were classified as having “some concerns” regarding bias, and the remaining 25 trials were categorized as having a “low risk”.

Fig. 2
figure 2

Risk of bias percentage across domains.

Full size image

Overall analysis

A pooled analysis of 9,193 participants indicated that the effect of remimazolam on PONV was comparable to that of other anesthetics (random effects model; RR: 0.96, 95% CI: 0.80–1.13, P = 0.607, I2: 21%; Fig. 3). The analysis revealed no significant publication bias, as evidenced by the funnel plot (Fig. 4A) and Egger’s test (P = 0.323).

Fig. 3
figure 3

Forest plot for overall analysis. RR, relative risk; 95% CI, 95% confidence interval.

Full size image
Fig. 4
figure 4

Funnel plot. A: Overall analysis, B-E: Subgroup analysis—variations according to anesthetic comparators, F-G: Subgroup analysis—general anesthesia vs. procedural sedation, H-J: Subgroup analysis—variations according to surgical procedure types, K: Subgroup analysis—elderly patients only, L: Subgroup analysis—female patients only, M: Subgroup analysis—nausea only outcome, N: Subgroup analysis—vomiting only outcome, O: Subgroup analysis—continuous infusion, P: Subgroup analysis—intermittent bolus, Q: Sensitivity analysis—PONV as primary outcomes.

Full size image

Subgroup analysis 1: variations according to anesthetic comparators

In comparison with propofol, remimazolam demonstrated a similar incidence of PONV (fixed effects model; n = 7,340; RR: 1.03, 95% CI: 0.89–1.20, P = 0.676, I2: 14%; Fig. 5). However, compared with inhalation anesthetics, remimazolam was associated with a lower incidence of PONV (fixed effects model; n = 363; RR: 0.50, 95% CI: 0.34–0.73, P < 0.001, I2: 33%). No significant difference in PONV incidence was observed between remimazolam and dexmedetomidine (fixed effects model; n = 504; RR: 0.94, 95% CI: 0.75–1.19, P = 0.621, I2: 9%) or between remimazolam and midazolam (fixed effects model; n = 853; RR: 0.86, 95% CI: 0.32–2.28, P = 0.757, I2: 0%). Analysis of etomidate was not feasible due to the presence of only a single RCT, which reported no PONV cases. Subgroup analyses for the propofol, inhalation, dexmedetomidine, and midazolam comparator groups showed no significant publication bias, as determined by Egger’s test (P = 0.113, 0.309, 0.864, and 0.141, respectively). The corresponding funnel plots are presented in Fig. 4B-E.

Fig. 5
figure 5

Forest plot for subgroup analysis—variations according to anesthetic comparators. RR, relative risk; 95% CI, 95% confidence interval.

Full size image

Subgroup analysis 2: general anesthesia vs. procedural sedation

Subgroup analysis stratified by anesthesia type (general anesthesia vs. procedural sedation) revealed that the effect of remimazolam on PONV did not differ significantly from that of other anesthetics in patients receiving either general anesthesia (fixed effects model; n = 3,137; RR: 0.91, 95% CI: 0.74–1.12, P = 0.368, I2: 30%; Supplemental Fig. S6) or procedural sedation (fixed effects model; n = 6,056; RR: 0.96, 95% CI: 0.83–1.12, P = 0.591, I2: 15%). No significant publication bias was detected, as demonstrated by funnel plots (Fig. 4F and G) and Egger’s test results (general anesthesia: P = 0.986, procedural sedation: P = 0.157).

Subgroup analysis 3: variations according to surgical procedure types

Subgroup analysis stratified by procedure type showed that the effect of remimazolam on PONV was comparable to that of other anesthetics among patients undergoing GI endoscopy (fixed effects model; n = 4,724; RR: 0.91, 95% CI: 0.76–1.11, P = 0.358, I2: 15%; Supplemental Fig. S7), bronchoscopy (fixed effects model; n = 858; RR: 1.02, 95% CI: 0.80–1.30, P = 0.875, I2: 44%), and other surgical procedures (fixed effects model; n = 3,611; RR: 0.94, 95% CI: 0.77–1.14, P = 0.542, I2: 28%). No significant publication bias was identified in subgroup analyses focusing on GI endoscopy, bronchoscopy, and other surgical procedures, as indicated by both funnel plots (Fig. 4H-J) and Egger’s test results (P = 0.342, 0.334, and 0.914, respectively).

Subgroup analysis 4: elderly patients

A review of 10 RCTs involving 1,692 geriatric patients excluded a study from the analysis due to the absence of PONV. Among these geriatric patients, no significant difference in PONV incidence was observed between those treated with remimazolam and those treated with other anesthetics (fixed effects model; n = 1,692; RR: 1.14, 95% CI: 0.88–1.49, P = 0.322, I² = 0%; Supplemental Fig. S8). The funnel plot (Fig. 4K) and Egger’s test (P = 0.638) indicated no significant publication bias.

Subgroup analysis 5: female patients

An analysis of 9 RCTs involving 1,061 female patients found no significant difference in the risk of PONV between remimazolam and other anesthetics (fixed effects model; n = 1,061; RR: 0.89, 95% CI: 0.61–1.30, P = 0.541, I² = 37%; Supplemental Fig. S9). Egger’s test revealed no evidence of publication bias (P = 0.092), with the corresponding funnel plot shown in Fig. 4L.

Subgroup analysis 6: nausea-only outcome

Postoperative nausea was specifically reported by 18 RCTs (Fig. 6). When compared with propofol, remimazolam exhibited no significant difference in postoperative nausea incidence (fixed effects model; 14 RCTs, n = 3,750; RR: 0.92, 95% CI: 0.76–1.11, P = 0.363, I² = 18%). One RCT comparing remimazolam with inhalation anesthetics showed a lower risk of postoperative nausea for remimazolam (n = 60; RR: 0.44, 95% CI: 0.23–0.86). Three RCTs comparing remimazolam to midazolam reported no significant difference in postoperative nausea incidence (fixed effects model; n = 853; RR: 0.86, 95% CI: 0.32–2.28, P = 0.757, I² = 0%). Due to a lack of RCTs reporting postoperative nausea as an outcome, analysis of dexmedetomidine and etomidate was not feasible. No publication bias was detected in subgroup analyses for propofol (Egger’s test: P = 0.208) and inhalation agents (Egger’s test: P = 0.141; funnel plot: Fig. 4M).

Fig. 6
figure 6

Forest plot for subgroup analysis—nausea only outcome. RR, relative risk; 95% CI, 95% confidence interval.

Full size image

Subgroup analysis 7: vomiting-only outcome

Postoperative vomiting was specifically reported by 17 RCTs (Fig. 7). When compared with propofol in 13 RCTs involving 3,860 patients, remimazolam was associated with a higher incidence of postoperative vomiting (fixed effects model; n = 3,860; RR: 1.41, 95% CI: 1.05–1.90, P = 0.024, I² = 0%). However, a single RCT comparing remimazolam with inhalation anesthetics showed no difference in postoperative vomiting risk (n = 60; RR: 1.00, 95% CI: 0.28–3.63). Similarly, 3 RCTs involving 853 patients showed no significant difference in postoperative vomiting incidence between remimazolam and midazolam (fixed effects model; n = 853; RR: 0.82, 95% CI: 0.24–2.78, P = 0.750, I² = 0%). Egger’s test indicated no significant publication bias in subgroup analyses for propofol (P = 0.197) and inhalation agents (P = 0.890; funnel plot: Fig. 4N).

Fig. 7
figure 7

Forest plot for subgroup analysis—vomiting only outcome. RR, relative risk; 95% CI, 95% confidence interval.

Full size image

Subgroup analysis 8: continuous infusion vs. intermittent bolus of remimazolam

In subgroup analysis stratified by remimazolam administration protocol, the incidence of PONV did not differ significantly between remimazolam and other anesthetics in studies using either continuous infusion (fixed effects model; n = 4,795; RR: 0.94, 95% CI: 0.82–1.07, P = 0.478, I² = 21%) or intermittent bolus (fixed effects model; n = 4,398; RR: 0.87, 95% CI: 0.66–1.14, P = 0.483, I² = 24%; Supplemental Fig. S10). No evidence of publication bias was observed in subgroup analyses for continuous infusion and intermittent bolus, as demonstrated by funnel plots (Fig. 4O–P) and Egger’s test (P = 0.984 and 0.068, respectively).

Meta-regression analysis: effects of mean age, sex distribution, and coadministered opioids

In a random-effects meta-regression model using mean age as the moderator, the age coefficient was 0.01 (95% CI: − 0.01–0.03, P = 0.412), indicating no significant effect of age; the age-specific R² analog (proportion of between-study variance explained) was 0.03 (Supplemental Fig. S11).

When the proportion of male participants served as moderator, the coefficient was 0.01 (95% CI: 0.00–0.02, P = 0.236), demonstrating that sex distribution did not have a significant effect; the R² analog was less than 0.01 (Supplemental Fig. S11).

In a random-effects meta‐regression model with the type of coadministered opioids as a categorical moderator (reference: no opioid coadministration, 9 RCTs), the coefficients for remifentanil, alfentanil, fentanyl, and sufentanil were − 0.64 (95% CI: −1.40–0.12, P = 0.098, 22 RCTs), − 0.39 (95% CI: −1.15–0.37, P = 0.311, 8 RCTs), 0.33 (95% CI: −0.67–1.34, P = 0.516, 6 RCTs), and 0.03 (95% CI: −1.07–1.13, P = 0.959, 5 RCTs), respectively. None of them differed significantly from the reference, and the R² analog was 0.33.

Sensitivity analysis: PONV as the primary outcome

Three RCTs involving 309 patients specifically assessed PONV as the primary outcome. Sensitivity analysis demonstrated no significant difference in PONV risk between remimazolam and comparator anesthetics (random effects model; RR: 0.78, 95% CI: 0.39–1.59, P = 0.499, I² = 52%; Supplemental Fig. S12). Egger’s test revealed no evidence of publication bias (P = 0.385), with the corresponding funnel plot shown in Fig. 4Q.

Discussion

This meta-analysis of 50 RCTs involving 9,193 patients evaluated the efficacy of remimazolam compared with other anesthetic agents in preventing PONV. The overall analysis showed no significant difference in PONV prevention between remimazolam and other agents. However, substantial heterogeneity was observed across studies with respect to remimazolam dosage and administration protocol, coadministered medications, and patient populations. To address heterogeneity, subgroup analyses were performed. The results of the analyses revealed that remimazolam was more effective than inhalation anesthetics in mitigating PONV. Patients who received remimazolam instead of inhalation agents experienced reduced postoperative nausea; however, compared with propofol, remimazolam was associated with a high incidence of postoperative vomiting. The efficacy of remimazolam in preventing PONV did not differ according to patient age, sex, type of anesthesia, type of surgical procedure, or remimazolam administration protocol.

As ERAS protocols gain traction, there is a growing emphasis on the management of PONV. Previously regarded as a normal postoperative event, PONV is now recognized as a complication requiring proactive and strategic intervention4. This shift aligns with the overarching goal of enhancing perioperative care to facilitate more efficient recovery and improved patient outcomes. Established risk factors for PONV include female gender, non-smoking status, younger age, type and duration of surgery, history of PONV or motion sickness, and postoperative opioid use3,4,5. In our study, the statistically significant difference in PONV incidence observed between anesthetic agents suggests that anesthetic selection may warrant greater attention, particularly for patients with a high risk of PONV. Moreover, beyond statistical significance, this finding carries clinical relevance. Given that most established risk factors for PONV are non-modifiable, the choice of the anesthetic agent as a modifiable factor should be strategically optimized to support preventive efforts.

The antiemetic properties of benzodiazepines have been more extensively studied in chemotherapy populations than in surgical populations59. The most recent guidelines from the American Society of Clinical Oncology include a moderate-level recommendation for the use of lorazepam or alprazolam as an adjunct in managing chemotherapy-induced nausea and vomiting refractory to standard prophylaxis60,61.

Although the precise antiemetic mechanism remains uncertain, a prevailing hypothesis posits that benzodiazepines may inhibit dopaminergic signaling in the chemoreceptor trigger zone, thereby impeding the emetic reflex59,62,63. This inhibition may occur through the suppression of dopamine synthesis, release, and postsynaptic activity59,62,63. However, it remains unclear whether this dopamine reduction results directly from the effect of benzodiazepines or indirectly via inhibition of adenosine reuptake59,62,63. Additionally, benzodiazepines may attenuate serotonin release, and their anxiolytic properties are thought to contribute to antiemetic effects59,62,63.

Remimazolam, a novel benzodiazepine, is expected to have similar antiemetic effects1. Existing studies on the effect of benzodiazepines on PONV in surgical patients have been limited, often exploring their use as an adjunct or premedication59. Therefore, the use of remimazolam as a primary anesthetic agent may offer more comprehensive insights into the role of benzodiazepines in PONV management for these patients.

In the present study, the overall pooled analysis revealed no significant difference in the incidence of PONV between remimazolam and other anesthetics. However, a more detailed subgroup analysis of 4 RCTs involving 363 subjects demonstrated a significantly lower incidence of PONV with remimazolam compared with inhalation agents. Inhalation agents are well known as potent triggers of emetogenesis64,65. Regardless of the specific volatile anesthetic, such as isoflurane, sevoflurane, or desflurane, approximately 1 in 3 or 4 patients is estimated to experience PONV following general anesthesia with these agents64,65. In this study, 30.8% of patients who received inhalation agents experienced PONV, consistent with the findings of prior research. In contrast, only 15.5% of those treated with remimazolam reported PONV, showing a significantly lower incidence.

A proposed mechanism for inhalation agent-induced emesis involves the enhancement of serotonin type 3 (5-HT3) signaling in both peripheral and central regions66. Specifically, the area postrema/nucleus of the solitary tract (AP/NTS) in the brainstem, along with vagal afferent fibers from the gastrointestinal tract, are key players in this process66. Activation of these pathways by inhalation agents triggers the emetic reflex, resulting in nausea and vomiting. Additionally, inhalation anesthetics may impair gastrointestinal function, which could further contribute to emesis66. A subgroup analysis focusing solely on postoperative nausea identified inhalation agents as inferior to remimazolam. However, when assessing postoperative vomiting alone, no significant difference was observed between remimazolam and inhalation agents. This discrepancy may be attributed to the inclusion of only a single RCT with a small sample size of 60 subjects in this subgroup analysis. Additional large-scale studies are necessary to accurately delineate the effects of remimazolam and inhalation agents on nausea and vomiting separately.

In this study, although no significant difference was found in overall PONV or postoperative nausea only between propofol and remimazolam, subgroup analysis of postoperative vomiting only demonstrated that it was reduced with propofol compared with remimazolam. Propofol is considered an effective antiemetic even though its precise mechanism of action remains uncertain67. One hypothesis suggests that propofol may exert antiemetic effects by antagonizing serotonin activity, thereby inhibiting pathways responsible for nausea and vomiting67,68. Another proposed mechanism indicates that propofol reduces synaptic transmission in the olfactory cortex, leading to a decrease in the release of excitatory amino acids such as glutamate and aspartate, which may be associated with the emetic response67,68. Currently, large-scale prospective studies comparing the incidence of PONV as a primary outcome between remimazolam and propofol are limited. A recent meta-analysis, which included 11 studies published up to July 2023 and focused on patients receiving general anesthesia, found no significant difference in the incidence of PONV between remimazolam and propofol (odds ratio: 1.04, 95% CI: 0.70–1.56); however, compared with inhalation agents, remimazolam demonstrated a lower PONV rate (odds ratio: 0.25, 95% CI: 0.13–0.47)69. In a retrospective propensity-matched study involving 666 patients with a primary focus on PONV, Suzuki et al. found that PONV was significantly reduced with propofol compared with remimazolam70. Specifically, propofol was associated with a lower incidence of overall PONV (21% vs. 35%, P < 0.001), postoperative nausea (21% vs. 35%, P < 0.001), and postoperative vomiting (9% vs. 16%, P = 0.009). Our findings also suggest that the profile of propofol may be more favorable than that of remimazolam in the prevention of postoperative vomiting, indicating its potential superiority as an antiemetic. However, further research is needed to draw definitive conclusions regarding the relative effects of propofol and remimazolam on PONV.

Our study has several limitations. First, there was considerable heterogeneity in the design of the RCTs included in this meta-analysis, such as variations in PONV evaluation (timing and outcome, including nausea only, vomiting only, or overall PONV), PONV prophylaxis protocol, anesthetic administration method, flumazenil use, and patient population. Second, heterogeneity was observed in the administration protocol of remimazolam across the included trials. In particular, variations in the allowed rescue dosing and the duration of sedation made it difficult to accurately quantify and stratify the total amount of remimazolam administered. To approximate the impact of cumulative exposure, subgroup analyses were conducted according to the type of procedure (general anesthesia vs. procedural sedation) and the administration method (continuous infusion vs. single bolus with intermittent rescue dosing). However, this approach may not have fully accounted for differences in total remimazolam exposure. Further well-designed studies are warranted to clarify the dose-dependent effects of remimazolam on PONV. Third, various perioperative drugs coadministered during surgery could influence the incidence of PONV. In particular, opioid use is a well-established risk factor3,4,5. To account for these drugs, a meta-regression analysis was performed incorporating opioid use and the type of opioid administered. However, this analysis could not fully capture the effects of all coadministered perioperative drugs. Table 1 presents a summary of the coadministered medications reported in each trial. Fourth, only a limited number of studies reported PONV as the primary outcome, with the majority addressing it as a secondary outcome. This underscores the need for rigorously designed, large-scale RCTs that specifically focus on PONV as the primary outcome to establish more definitive conclusions. Lastly, most RCTs were conducted in East Asia, resulting in a predominance of Asian participants, which may restrict the generalizability of the findings to other ethnic groups. Future RCTs in more diverse regions are essential to provide more comprehensive insights applicable to various populations.

In conclusion, remimazolam was not significantly different from other anesthetic agents in overall PONV prevention. However, subgroup analysis revealed differing effects: remimazolam was associated with a lower incidence of PONV compared with inhalation anesthetics, and was associated with a higher incidence of postoperative vomiting compared with propofol.