Skip to main content

Foundational Trials

RSI

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Ketamine or Etomidate for Tracheal Intubation of Critically Ill Adults
  • Acronym: RSI (Randomised Trial of Sedative Choice for Intubation)
  • Year: 2025
  • Journal published in: The New England Journal of Medicine
  • Citation: Casey JD, Seitz KP, Driver BE, et al. Ketamine or etomidate for tracheal intubation of critically ill adults. N Engl J Med. 2025 Dec 9. Online ahead of print.

Context & Rationale

  • Background
    • Emergency tracheal intubation is common in critical care and emergency medicine, with high downstream morbidity and mortality; peri-intubation hypotension/cardiovascular collapse is frequent and prognostically adverse.
    • Etomidate is widely used because of rapid onset and relative haemodynamic stability, but a single dose inhibits adrenal 11β-hydroxylase and suppresses cortisol synthesis for up to ~72 hours, raising concern for worse outcomes (particularly in sepsis).
    • Ketamine is an alternative with analgesia and sympathetic stimulation, but in catecholamine-depleted shock its direct myocardial depressant effects may predominate, plausibly increasing hypotension/collapse.
    • Pre-RSI evidence (small randomised trials and heterogeneous meta-analyses) did not establish whether the net clinical effect favours ketamine (avoid adrenal suppression) or etomidate (haemodynamic stability).
  • Research Question/Hypothesis
    • In critically ill adults undergoing emergency tracheal intubation, does ketamine (vs etomidate) reduce 28-day in-hospital mortality?
    • Hypothesis (trialist intent): ketamine would lower mortality by avoiding etomidate-associated adrenal suppression while maintaining acceptable peri-intubation haemodynamics.
  • Why This Matters
    • Induction agent choice is a ubiquitous, time-critical intervention in a high-risk population; even modest effects could translate to large absolute numbers of lives and complications.
    • Clinical equipoise persisted despite strong physiological arguments on both sides; practice variation and guideline uncertainty made a pragmatic mortality-powered trial methodologically and clinically important.

Design & Methods

  • Research Question: Among critically ill adults undergoing emergency tracheal intubation, does ketamine (vs etomidate) reduce death by day 28 while in hospital?
  • Study Type: Pragmatic, multicentre, parallel-group, investigator-initiated randomised trial; 1:1 allocation; stratified by site with variable block sizes; unblinded; conducted in emergency departments and intensive care units in the United States.
  • Population:
    • Setting: 14 US sites (6 emergency departments; 8 intensive care units); enrolment April 2022 to August 2025.
    • Inclusion: Adults (≥18 years) undergoing emergency tracheal intubation with an induction agent, where the treating clinician judged both ketamine and etomidate to be reasonable options.
    • Key exclusions: Trauma-related intubation; immediate need precluding randomisation without delaying care; clinician-determined requirement for (or contraindication to) one of the two study drugs; known pregnancy; prisoners.
    • Consent model: Enrolment under US exception from informed consent for emergency research; patients/surrogates approached subsequently for ongoing participation as feasible.
  • Intervention:
    • Ketamine as the induction agent for intubation (intravenous), with clinician-selected dose within a protocol-specified range (1–2 mg/kg).
    • Other peri-intubation care (preoxygenation strategy, fluids/vasopressors, neuromuscular blockade choice, laryngoscopy device/approach, post-intubation sedation) at clinician discretion.
  • Comparison:
    • Etomidate as the induction agent for intubation (intravenous), with clinician-selected dose within a protocol-specified range (0.15–0.30 mg/kg).
    • All other peri-intubation care at clinician discretion, as in the intervention arm.
  • Blinding: Unblinded (treating clinicians, patients, and data collectors); mortality outcome objective; peri-intubation haemodynamic outcomes based on recorded blood pressure/vasopressor use.
  • Statistics: Power calculation: 2308 patients required to detect a 5.2 percentage-point absolute reduction in 28-day in-hospital mortality (30.0% to 24.8%) with 90% power at two-sided α=0.05; inflation for withdrawal yielded a target sample size of 2364; analysis was intention-to-treat, with the primary outcome analysed using a generalised linear mixed-effects model with random site effect and marginal standardisation to estimate adjusted risk difference; one interim analysis used a Haybittle–Peto efficacy boundary (two-sided P<0.001). 1
  • Follow-Up Period: Primary endpoint through day 28 while in hospital (censored at discharge); peri-intubation endpoints from induction through 2 minutes after intubation; additional safety/organ support outcomes through 24 hours and to discharge.

Key Results

This trial was not stopped early. Recruitment continued to the planned sample size with a single prespecified interim analysis (approximately 50% enrolment) and ongoing DSMB oversight.

Outcome Ketamine Etomidate Effect p value / 95% CI Notes
Death by day 28 while in hospital (primary) 348/1176 (29.6%) 367/1189 (30.9%) Adjusted RD −0.8 pp 95% CI −4.5 to 2.9; P=0.65 Intention-to-treat; adjusted for site (random effect) with marginal standardisation
Cardiovascular collapse between induction and 2 min after intubation (secondary) 260/1176 (22.1%) 202/1189 (17.0%) RD 5.1 pp 95% CI 1.9 to 8.3 Composite: new/increased vasopressor, SBP <65 mmHg, or cardiac arrest
New or increased vasopressor use (component) 252/1173 (21.5%) 192/1186 (16.2%) RD 5.3 pp 95% CI 2.1 to 8.6 Component of cardiovascular collapse
Systolic BP <65 mmHg (component) 93/1173 (7.9%) 71/1186 (6.0%) RD 1.9 pp 95% CI −0.1 to 3.9 Component of cardiovascular collapse
Cardiac arrest between induction and 2 min after intubation (component) 22/1176 (1.9%) 17/1189 (1.4%) RD 0.4 pp 95% CI −0.5 to 1.4 Component of cardiovascular collapse
First-attempt intubation success (exploratory) 1028/1176 (87.4%) 1034/1189 (87.0%) RD 0.4 pp 95% CI −2.3 to 3.2 No clinically meaningful difference in procedural success
Time from induction to successful intubation (sec), median (IQR) 79 (56–115) 70 (50–100) Median difference 9 sec 95% CI 3 to 14 Exploratory; small absolute delay with ketamine
Lowest systolic BP between induction and 2 min (mmHg), median (IQR) 86 (75–98) 90 (79–102) Median difference −6 mmHg 95% CI −8 to −3 Physiological separation consistent with higher collapse rates in ketamine group
Systolic BP <80 mmHg between induction and 2 min (post hoc threshold) 294/1167 (25.2%) 254/1183 (21.5%) RD 3.7 pp 95% CI 0.2 to 7.2 Post hoc analysis (threshold not part of primary collapse definition)
Vasopressor use at 24 hours (safety) 457/1176 (38.9%) 503/1189 (42.3%) RD −3.4 pp 95% CI −7.5 to 0.7 No clear between-group difference at 24 hours
Cardiac arrest requiring CPR between induction and hospital discharge (safety) 88/1176 (7.5%) 63/1189 (5.3%) RD 2.1 pp 95% CI 0.0 to 4.2 Safety outcome reported through hospital discharge
Ventricular tachycardia between induction and 2 min (post hoc safety signal) 9/884 (1.0%) 2/905 (0.2%) RD 0.8 pp 95% CI 0.1 to 1.5 Collected after an early event prompted systematic ascertainment; denominators reflect those with data
  • Mortality: No evidence that ketamine reduced 28-day in-hospital mortality (29.6% vs 30.9%; adjusted RD −0.8 pp; 95% CI −4.5 to 2.9; P=0.65), including in the prespecified sepsis/septic shock subgroup (38.8% vs 38.2%; adjusted RD 1.0 pp; 95% CI −4.8 to 6.7).
  • Haemodynamics: Ketamine increased cardiovascular collapse (22.1% vs 17.0%; RD 5.1 pp; 95% CI 1.9 to 8.3), driven primarily by higher new/increased vasopressor use (21.5% vs 16.2%; RD 5.3 pp; 95% CI 2.1 to 8.6) and a lower nadir systolic BP (86 vs 90 mmHg; median difference −6 mmHg; 95% CI −8 to −3).
  • Heterogeneity and harms: Cardiovascular collapse differences were larger in sepsis/septic shock (23.5% vs 17.0%; RD 6.4 pp; 95% CI 2.0 to 10.8) and with higher severity (APACHE II ≥20: 27.9% vs 20.9%; RD 7.0 pp; 95% CI 3.0 to 11.1), while exploratory individualised treatment effect modelling did not identify meaningful treatment effect heterogeneity (C-for-benefit 0.51; 95% CI 0.46 to 0.55); a post hoc safety signal for ventricular tachycardia favoured etomidate (1.0% vs 0.2%; RD 0.8 pp; 95% CI 0.1 to 1.5).

Internal Validity

  • Randomisation and allocation: Site-stratified randomisation with variable block sizes and allocation concealment via sequential opaque envelopes; randomisation occurred after the clinician decided to intubate and to use an induction agent but before administration.
  • Dropout/exclusions: 2367 patients randomised; 2365 included in the primary analysis after post-enrolment withdrawals (very low proportion); primary outcome ascertainment was complete for analysed participants.
  • Performance/detection bias: Unblinded design; mortality is objective, but the secondary composite includes clinician-driven vasopressor initiation (potentially behaviour-sensitive). Objective blood pressure separation (nadir SBP) supports that the haemodynamic signal was not solely definitional.
  • Protocol adherence and separation: High adherence (99.2% received induction medication consistent with assignment); dose separation was clinically substantial (median ketamine 1.6 mg/kg [IQR 1.2–1.9] vs etomidate 0.28 mg/kg [IQR 0.23–0.30]).
  • Baseline balance: Groups were similar in key prognostic variables (e.g., median age 61 vs 60 years; median APACHE II 20 vs 20; sepsis/septic shock 37.1% vs 37.4%; vasopressors within 1 hour before induction 22.0% vs 22.9%), supporting successful randomisation.
  • Timing and delivery: Time from induction to successful intubation was slightly longer with ketamine (79 vs 70 sec; median difference 9 sec; 95% CI 3 to 14), but first-pass success was similar (87.4% vs 87.0%; RD 0.4 pp; 95% CI −2.3 to 3.2).
  • Statistical rigour: Primary analysis was prespecified, intention-to-treat, and incorporated site as a random effect; sample size targets were met; interim analysis used a conservative efficacy boundary.

Conclusion on Internal Validity: Strong overall: large sample size, robust randomisation with high protocol adherence, and objective primary outcome; the main internal validity caveat is the unblinded nature of haemodynamic co-interventions embedded in the secondary composite.

External Validity

  • Population representativeness: Broad adult, non-trauma critically ill population across emergency department and ICU settings, including a high prevalence of sepsis/septic shock and respiratory failure; enrolment under emergency research consent processes mirrors real-world urgency.
  • Applicability: Pragmatic co-intervention flexibility and dosing ranges aligned with common practice support translation to similar high-resource ED/ICU environments; generalisability is more limited for trauma, prehospital settings, paediatrics, and healthcare systems where etomidate is unavailable or dosing/RSI bundles differ materially.
  • Induction-agent scope: The trial answers a two-agent question (ketamine vs etomidate) and does not directly inform settings where propofol or midazolam are the dominant induction agents.

Conclusion on External Validity: Good generalisability to adult non-trauma emergency intubations in ED/ICU practice; extrapolation to trauma, prehospital, and markedly different resource contexts should be cautious.

Strengths & Limitations

  • Strengths:
    • Large, multicentre pragmatic randomised trial powered for a patient-centred endpoint (28-day in-hospital mortality).
    • High treatment adherence (99.2%) with clear physiological separation in haemodynamics.
    • Clinically relevant secondary endpoint (peri-intubation cardiovascular collapse) aligned with prior airway trials and implementation priorities.
  • Limitations:
    • Unblinded design, with potential influence on clinician-triggered components of the cardiovascular collapse composite (vasopressor initiation).
    • Exclusion of trauma and situations where clinicians required a specific agent may restrict inference in those high-stakes subgroups.
    • No direct measurement of adrenal function/cortisol response; the mechanistic pathway for etomidate-related harm could not be adjudicated.
    • Some safety signals (e.g., ventricular tachycardia) were collected after early events prompted additional ascertainment, and should be interpreted as hypothesis-generating.

Interpretation & Why It Matters

  • Clinical implications
    • In adult non-trauma emergency intubation, ketamine did not improve 28-day in-hospital mortality compared with etomidate, but increased peri-intubation cardiovascular collapse and lowered nadir systolic blood pressure.
    • The absence of a mortality benefit (including in sepsis/septic shock) challenges the assumption that avoiding single-dose etomidate adrenal suppression yields a clinically important survival advantage in this setting.
    • Where ketamine is used (e.g., etomidate unavailable or contraindicated), the results support proactively anticipating haemodynamic deterioration (early vasopressor readiness, careful dose selection, RSI bundle optimisation).

Controversies & Subsequent Evidence

  • Reconciling prior randomised evidence with RSI: Earlier smaller trials comparing ketamine and etomidate suggested mixed short-term morbidity signals and were not mortality-definitive, leaving persistent uncertainty that RSI was designed to resolve. 23
  • Meta-analytic synthesis pre-RSI: Recent meta-analyses reached differing conclusions depending on inclusion criteria and analytic frameworks (frequentist vs Bayesian), reflecting small trial sizes, heterogeneity, and outcome definition variability—limitations addressed by RSI’s scale and pragmatic design. 4567
  • Observational signals and confounding: Large observational analyses have associated etomidate use with worse outcomes in some cohorts, but confounding by indication and practice-pattern effects remain plausible; RSI provides randomised evidence that any net mortality disadvantage of etomidate (relative to ketamine) is not large in the studied population. 8
  • Composite haemodynamic endpoint interpretation: The cardiovascular collapse endpoint includes clinician-triggered vasopressor escalation; in an unblinded trial this can embed practice behaviour. RSI mitigates (but does not eliminate) this concern by also demonstrating objective blood pressure separation (median nadir SBP 86 vs 90 mmHg; median difference −6 mmHg).
  • Guideline implications: The most recent RSI-specific guideline from the Society of Critical Care Medicine (pre-RSI) recommended either ketamine or etomidate as reasonable options, emphasising individualisation; RSI provides large randomised evidence favouring etomidate for peri-intubation haemodynamic stability without showing a ketamine mortality advantage, and should inform future updates. 910
  • Positioning within evolving airway guidance: Contemporary Difficult Airway Society guidance (post-RSI publication window) addresses rapid sequence induction and difficult intubation management; RSI’s findings are directly relevant to haemodynamically fragile patients undergoing RSI within these frameworks. 11
  • Post-publication commentary emphasis: Expert commentary has highlighted that induction-agent choice should prioritise outcomes that matter (mortality and major complications) while acknowledging that peri-intubation haemodynamics are a near-term, modifiable mediator; RSI meaningfully shifts the evidentiary balance towards etomidate in many critically ill intubations. 12

Summary

  • RSI was a large pragmatic multicentre randomised trial comparing ketamine vs etomidate as induction agents for emergency tracheal intubation in critically ill adults.
  • Ketamine did not reduce 28-day in-hospital mortality compared with etomidate (29.6% vs 30.9%; adjusted RD −0.8 pp; 95% CI −4.5 to 2.9; P=0.65).
  • Ketamine increased peri-intubation cardiovascular collapse (22.1% vs 17.0%; RD 5.1 pp; 95% CI 1.9 to 8.3), with lower nadir systolic BP (86 vs 90 mmHg; median difference −6 mmHg; 95% CI −8 to −3).
  • No meaningful differences were observed in first-pass success, while time to intubation was slightly longer with ketamine (median difference 9 sec; 95% CI 3 to 14).
  • A post hoc safety signal for ventricular tachycardia occurred more often with ketamine (1.0% vs 0.2%; RD 0.8 pp; 95% CI 0.1 to 1.5).

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Most widely used RSI/airway guidelines were published before the RSI trial; future updates are likely to incorporate RSI’s mortality and haemodynamic findings.

Overall Takeaway

RSI is a landmark pragmatic trial because it definitively tested (at scale, in real-world ED/ICU practice) whether ketamine improves clinically meaningful outcomes over etomidate for emergency intubation. The trial found no mortality advantage with ketamine and demonstrated higher peri-intubation cardiovascular collapse, shifting the evidentiary balance toward etomidate as the default induction agent for many haemodynamically vulnerable critically ill adults, while reinforcing the need for proactive haemodynamic optimisation regardless of drug choice.

Overall Summary

  • Ketamine did not reduce 28-day in-hospital mortality compared with etomidate.
  • Ketamine increased peri-intubation cardiovascular collapse and lowered nadir systolic blood pressure.
  • Practice implication: etomidate is favoured for haemodynamic stability in many critically ill intubations; if ketamine is used, anticipate and mitigate hypotension.

Bibliography

  • 1DeMasi SC, Imhoff B, Lewis AA, et al. Protocol and statistical analysis plan for a multicenter randomized trial of ketamine vs etomidate for emergency tracheal intubation. CHEST Crit Care. 2025;3:100177. Link
  • 2Jabre P, Combes X, Lapostolle F, et al. Etomidate versus ketamine for rapid sequence intubation in acutely ill patients: a multicentre randomised controlled trial. Lancet. 2009;374(9686):293-300. Link
  • 3Matchett G, Gasanova I, Riccio CA, et al.; EvK Clinical Trial Collaborators. Etomidate versus ketamine for emergency endotracheal intubation: a randomized clinical trial. Intensive Care Med. 2022;48(1):78-91. Link
  • 4Greer A, Hewitt B, Khazaneh M, et al. Ketamine versus etomidate for rapid sequence intubation: a systematic review and meta-analysis of randomized trials. Crit Care Med. 2025;53(2):e374-e383. Link
  • 5Koroki T, et al. Ketamine vs etomidate for rapid sequence intubation: a Bayesian meta-analysis. Crit Care. 2024;28(1):48. Link
  • 6de Morais CAC, et al. Readdressing rapid sequence induction and intubation using ketamine or etomidate: a systematic review and meta-analysis. Medicine (Baltimore). 2025;104(9):e42207. Link
  • 7Kotani Y, et al. Etomidate as an induction agent for endotracheal intubation of critically ill patients: a systematic review and meta-analysis. J Crit Care. 2023;77:154317. Link
  • 8Wunsch H, Bosch NA, Law AC, et al. Evaluation of etomidate use and association with mortality compared with ketamine among critically ill patients. Am J Respir Crit Care Med. 2024;210(10):1243-1251. Link
  • 9Acquisto NM, Mosier JM, Bittner EA, et al. Society of Critical Care Medicine clinical practice guidelines for rapid sequence intubation in the critically ill adult patient. Crit Care Med. 2023;51(10):1411-1430. Link
  • 10Higgs A, McGrath BA, Goddard C, et al. Guidelines for the management of tracheal intubation in critically ill adults. Br J Anaesth. 2018;120(2):323-352. Link
  • 11Ahmad I, et al. Difficult Airway Society 2025 guidelines for management of unanticipated difficult intubation in adults. Br J Anaesth. 2026;136(1):36-63. Link
  • 12Miner JR. Ketamine, etomidate, and the outcomes that matter in critical care. JAMA Netw Open. 2025;8(12):e2547966. Link