INTRODUCTION
Rapid sequence intubation (RSI) is defined as the administration of an induction agent and neuromuscular blocker drug in quick succession to obtain an ideal intubating condition [1]. Use of the RSI technique forgoes bag and mask ventilation prior to the intubation attempt to decrease the risk of life threatening complications such as aspiration of the gastric contents [2]. However, the rapid or bolus administration of induction agents to critically ill patients can be associated with an increased incidence of hypotension, mechanical ventilation, duration of intensive care unit (ICU) stays, and short-term mortality [3]. Therefore, choosing an induction agent suitable for the clinical profile of the patient is vitally important.
Etomidate, a carboxylated imidazole anesthetic, is the induction agent most commonly used for RSI airway management in emergency departments. Its hemodynamic stability and rapid onset of action makes it an attractive induction agent [4,5]. However, etomidate has been associated with adrenocortical insufficiency due to transient inhibition of the 11β-hydroxylase enzyme, which might increase the risk of organ dysfunction and short-term mortality [6–8]. Although it is well established that etomidate is associated with adrenocortical insufficiency in critically ill patients, the possibility that a single dose could increase short-term mortality compared with other induction agents is controversial. A previous meta-analysis found a higher likelihood of short-term mortality when etomidate was used for RSI in critically ill patients [9].
Ketamine is another viable induction agent for patients with sepsis. It increases blood pressure and heart rate through catecholamine release [10]. As a result, tissue perfusion is maintained, and further organ dysfunction is prevented. However, several studies have demonstrated that ketamine use can be associated with hypotension and myocardial ischemia in patients with catecholamine depletion and elderly patients [11].
Recent randomized controlled trials (RCTs) suggest that ketamine decreases short-term mortality, compared with etomidate, but the guidelines recently published by the Society of Critical Care Medicine do not recommend the use of one agent over the other in terms of short-term mortality, hypotension, or first-pass success [12]. Therefore, we conducted a systematic review and meta-analysis to compare the effects of a single dose of etomidate for RSI with those of ketamine in critically ill patients.
METHODS
This study was registered in PROSPERO (No. CRD42023469386) and is reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines [13]. No ethical clearance or consent was required because all analyses were performed on already published data.
Literature search
We systematically searched PubMed, Embase, and the Web of Science from the starting date of each database until April 19, 2024. The following search terms were used with the Boolean operators AND or OR: Etomidate; Ketamine; Emergency; ICU; critically ill; intensive care unit; Intubation; airway management; and rapid sequence intubation. Two authors independently reviewed the identified studies to determine their relevance. Lists of referenced articles and citations from the trials included were also searched to identify any further trials that might have been missed in the primary search. The full texts of all articles found to be potentially relevant by both reviewers were evaluated thoroughly. Any disagreements or discrepancies were arbitrated by another author.
Inclusion criteria and study selection
We included RCTs that compared the effects of etomidate and ketamine for RSI in non–operating room settings. Case reports, observational studies (both prospective and retrospective), conference abstracts, review articles, and RCTs unavailable in the English language were excluded.
Study outcome
The primary outcome of our review was 28-day mortality; the first-pass success rate, postinduction hypotension, and postinduction cardiac arrest were the secondary outcomes. The studies included in the meta-analysis defined postinduction hypotension as a systolic blood pressure below 90 mmHg or a mean arterial blood pressure below 65 mmHg. To define postinduction cardiovascular collapse, the studies used the Vanderbilt definition: (1) new (postinduction) systolic blood pressure <65 mmHg; (2) immediate need for intravenous vasopressor bolus or dose escalation; or (3) cardiac arrest or death within 1 hour.
Data extraction and outcome measures
We created a data extraction sheet in Microsoft Excel (Microsoft Corp) to collect the following variables: author, year, number of participants in the ketamine group, number of events (death, hypotension, cardiac collapse, or failure) in the ketamine group, number of participants in the etomidate group, number of events (death, hypotension, cardiac collapse, or failure) in the etomidate group, study setting (emergency, ward, or ICU), and intubation conditions. Data extraction was done independently by two authors, and discrepancies were resolved and adjudicated by a third author.
Risk-of-bias assessment
The quality of each RCT included in the final analysis was assessed independently by two authors based on the version 2 of the Cochrane risk-of-bias tool for RCTs [14]. Our risk-of-bias table contained the following domains: randomization process, allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcomes assessed (detection bias), deviations from the intended intervention, missing outcome data (attrition bias), measurement of the outcomes and selection of the reported results (reporting bias), and miscellaneous bias. Two authors assessed the risk of bias individually, and any disagreements were adjudicated by a third author.
Quality of evidence assessment
In this systematic review, the overall rating for the certainty that the evidence can aid clinical practice was assessed by two authors using the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) metric [15]. The quality of evidence was rated as very low, low, moderate, or high for each of the outcomes, based on the overall risk of bias, imprecision, inconsistency, indirectness, and publication bias.
Statistical analysis
The categorical outcomes were 28-day mortality, first-pass success rate, postinduction hypotension, and postinduction cardiac arrest. They are presented as relative risks (RRs) with their 95% confidence intervals (CIs). We present the I2 statistic to quantify heterogeneity among studies. Because we anticipated considerable between-study heterogeneity, we used a random-effects model to pool the effect sizes. The restricted maximum likelihood estimator was used to calculate the heterogeneity variance, τ2. All the findings, including the prediction interval of the pooled estimate, are shown as forest plots. To further explore the heterogeneity, we performed a sensitivity analysis using the leave-one-(study)-out method and extracted the change in I2 upon repeating the pooled analysis. Due to the small number of studies, we could not assess publication bias using funnel plot asymmetry.
We also performed a trial sequential analysis (TSA) because it provides a cumulative frequentist approach to control both type I and type II errors and estimate when the effect is large enough that it is unlikely be affected by further studies. The calculated sample size in a TSA is related to the pooled effect estimated in a meta-analysis. The TSA was performed for the primary outcome (28-day mortality) and the secondary outcome of postinduction hypotension to obtain the relative RR, achieved information size (AIS), and heterogeneity-adjusted required information size (HARIS; also called the diversity-adjusted required information size [DARIS]). A cumulative z-statistic line was drawn on the TSA graphical output by adding the included studies with a chronological criterion, with the last study representing the end of the line and the area (“benefit,” “harm,” “inner wedge,” or “nonstatistically significant”). Statistical significance was considered to be P<0.05. All analyses were done in R ver. 4.3.2 (packages: meta, metabin, and RTSA; R Foundation for Statistical Computing).
RESULTS
Fig. 1 depicts the PRISMA flowchart. Four RCTs involving 1,663 participants were included in the final meta-analysis (Table 1) [1,10,11,16]. Three trials were single-center studies (two in the United States [1,16] and one in Thailand [11]), and one multicenter trial was conducted in France [10].
Primary outcome
All four studies reported mortality outcomes, with three studies reporting 28-day mortality [10,11,16] and one study reporting 30-day mortality [1]. Of the 830 patients who received ketamine, 246 died during the study period, and among the 833 patients who received etomidate, 263 died during the study period. The pooled analysis revealed no statistically significant difference in mortality rates for the ketamine and etomidate groups (RR, 0.95; 95% CI, 0.72–1.25). The heterogeneity I2 statistic was 39% (Fig. 2A) [1,10,11,16]. In the TSA, the required information size was 688 (which was reached), and against a HARIS of 1,621, the cumulative z-score line reached an AIS of 1,663 (103%). At that point, the line did not cross either of the benefit boundaries, remaining in the “nonstatistically significant” zone, but it did reach the futility zone. The latter provides strong evidence that further studies are unlikely to change the no-effect results for the “inner wedge” area (Fig. 2B).
Secondary outcomes
Postinduction hypotension was reported by three studies [1,11,16]. The incidence of postinduction hypotension was higher in the ketamine group than the etomidate group (RR, 1.30; 95% CI, 1.03–1.64), and the TSA-adjusted RR was 1.30 (95% CI, 0.97–1.74) (Fig. 3) [1,11,16]. In the TSA, the HARIS required to reject the null hypothesis without type II error was 1,703, and the cumulative z-score line reached an AIS of 1,190, which was 81% of the HARIS. At that level, the line did not cross either of the benefit boundaries, remaining in the “nonstatistically significant” zone, but it also did not reach the futility zone. Therefore, further studies are needed to determine whether postinduction hypotension does indeed differ between the ketamine and etomidate groups. The TSA-adjusted 95% CI also indicated a nonsignificant difference (Fig. 3B).
The first-pass success rate was reported by three studies, and the ketamine and etomidate groups did not differ significantly (RR, 1.00; 95% CI, 0.97–1.03) (Fig. 4A) [1,11,16]. Furthermore, no statistically significant difference was found in the incidence of postinduction cardiac arrest (RR, 1.10; 95% CI, 0.62–1.96) (Fig. 4B) [10,11,16].
A sensitivity analysis using the leave-one-out method was performed for all the outcomes to find which studies contributed most to the heterogeneity of the pooled analysis. The results are depicted in Supplementary Fig. 1. In terms of the primary outcome, the study by Srivilaithon et al. [11] contributed most to the heterogeneity because omitting it resulted in I2 of 0%.
Publication bias
Because only four studies met our inclusion criteria, the Egger results might have limited sensitivity in detecting publication bias; therefore, we are not reporting them.
GRADE assessment
The certainty of the evidence according to our outcome parameters from the GRADE metric is summarized in Supplementary Table 1. The level of certainty is noted as moderate for all of the outcomes except postinduction hypotension (high).
DISCUSSION
This systematic review and meta-analysis has compared the effects of using ketamine and etomidate for RSI in critically ill patients. The pooled estimates show no statistically significant differences between the two agents in terms of 28-day mortality, first-pass success rate, or the incidence of postinduction cardiac arrest. The incidence of postinduction hypotension was higher in the ketamine group, but the TSA-adjusted 95% CIs indicate that the difference was not statistically significant. On the other hand, the TSA confirmed that for the primary outcome of 28-day mortality, the required information size had been reached, making it unlikely that future studies would change the no-effect results. The GRADE assessment suggests a moderate level of evidence for this outcome. Our study thus confirms that ketamine and etomidate offer comparable intubation conditions and enable prompt placement of a definitive airway in critically ill patients without increasing the risk of potentially fatal complications.
According to a meta-analysis by Chan et al. [6], using etomidate for RSI was linked to increased rates of 28-day mortality and adrenocortical insufficiency in sepsis patients. However, their meta-analysis included both RCTs and observational trials that compared etomidate with numerous sedative and hypnotic agents. A subsequent meta-analysis conducted by Gu et al. [9] found that although a single dose of etomidate raised the chance of adrenocortical insufficiency, it had no effect on overall mortality in sepsis patients. In the largest and most recent study, an anesthesia-based airway team treated 801 patients undergoing RSI in the ICU by randomly assigning them to receive etomidate or ketamine. The ketamine group had a better 7-day survival rate, but that difference did not persist until day 28, and no significant differences were seen in the secondary outcomes of length of ICU stay, amount of time spent on mechanical ventilation, Sequential Organ Failure Assessment (SOFA) scores, or the use of vasopressors [16]. Our pooled analysis of four studies supports that finding by showing that a single dose of etomidate was not associated with a significantly increased risk of 28-day mortality.
Hemodynamic status can be affected by the administration of a single dose of induction agent, particularly in patients who are critically ill and require emergency intubation. Although both etomidate and ketamine are considered to be hemodynamically stable induction agents, hypotension remains a concern in critical patients [5,6,17-19].
Previous observational studies comparing etomidate and ketamine have yielded inconsistent findings. In the US National Emergency Airway Registry of 6,806 patients, the use of ketamine was associated with a modest increase in hypotension (adjusted odds ratio, 1.4; 95% CI, 1.2–1.7) but it had no effect on the first-pass success rate [5]. In a retrospective analysis of 7,466 intubations performed by air medical teams, the use of ketamine was also associated with an increased incidence of hypotension without having an effect on the first-pass success rate [20]. Some retrospective studies have reported comparable hemodynamics between ketamine and etomidate, whereas others observed increased hypotension with ketamine [21,22]. Our investigation shows that ketamine increased the risk of hypotension but without statistical significance.
A meta-analysis comparing etomidate and ketamine for RSI in non–operating room settings was recently published by Sharda and Bhatia [23]. It also revealed that, although etomidate does not affect the first-pass intubation success rate, using it for induction during RSI is linked to a lower risk of postinduction hypotension than using ketamine. However, that meta-analysis included data from an observational study and non–peer-reviewed conference abstracts and was not conducted solely on RCTs. In their meta-analysis, Kotani et al. [24] demonstrated increased short-term mortality when etomidate was used as the induction agent in critically ill patients. However, their meta-analysis included studies in which the comparator arm was not only ketamine but also other induction agents such as midazolam, thiopentone sodium, and a ketamine/propofol mixture. Similarly, Koroki et al. [25] reported a moderate level of evidence that induction with ketamine was associated with a reduced risk of short-term mortality. However, their meta-analysis also included non–peer-reviewed, open-label, and propensity-matched studies, along with one RCT that used a propofol/ketamine mixture as the comparator. A 4-year interrupted time series analysis of an institutional change-in-practice from etomidate to ketamine during emergency intubation of trauma patients did not find a difference in the frequency of peri-intubation cardiac arrest [26]. We report similar findings in this meta-analysis.
Strengths and limitations
We estimated direct and indirect relative effects using data from high-quality RCTs. Our sensitivity analysis results indicate that the robustness of the evidence was adequate. In our research, we examined heterogeneity and provided the prediction interval. To determine the degree of evidence certainty for our outcome criteria, we also used the GRADE metric. We conducted a TSA to control for random errors and quantify imprecision in our meta-analyses.
Nonetheless, our meta-analysis also has some limitations. First, for most of the outcomes under study, few RCTs were found during our search. Second, certain outcomes, such as postinduction hypotension, lacked clear definitions in the research that evaluated it. Third, the choice of muscle relaxant was not standardized in the studies. Fourth, we included only studies done in non–operating room settings. Therefore the findings of this study cannot be extrapolated to patients being intubated for emergency surgical procedures because the addition of surgical insult to an already critically ill patient can lead to poor outcomes.
Conclusions
Our findings suggest that neither etomidate nor ketamine is superior to the other for use in emergency tracheal intubation. We found no significant differences with regard to 28-day mortality, first-pass success rate, or the incidence of postinduction cardiac arrest Ketamine showed a trend toward a higher risk of postinduction hypotension, but the association was not statistically robust. Clinicians can safely select either medicine, as they both seem to have sufficient efficacy for use in RSI. To clarify any possible variations in outcomes between ketamine and etomidate for use in emergency tracheal intubation, more large-scale randomized trials will be needed.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Supplement1
Print 
