Skip to main content

Foundational Trials

KETASED

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Etomidate versus ketamine for rapid sequence intubation in acutely ill patients: a multicentre randomised controlled trial
  • Acronym: KETASED
  • Year: 2009
  • Journal published in: The Lancet
  • Citation: Jabre P, Combes X, Lapostolle F, Dhaouadi M, Ricard JD, et al. Etomidate versus ketamine for rapid sequence intubation in acutely ill patients: a multicentre randomised controlled trial. Lancet. 2009;374(9686):293-300.

Context & Rationale

  • Background
    • Emergency rapid sequence induction (RSI) for tracheal intubation is a high-frequency intervention in prehospital, emergency department, and ICU pathways, where small differences in safety or organ outcomes could have large population effects.
    • Etomidate was widely used because it is perceived to preserve haemodynamic stability during induction, but it inhibits adrenal steroidogenesis after a single bolus (via 11β-hydroxylase inhibition), raising concern for clinically important adrenal insufficiency in critical illness—especially sepsis.
    • Ketamine provides dissociative anaesthesia with relative cardiovascular stability (and potential sympathetic stimulation), and does not share etomidate’s adrenal enzyme inhibition, but comparative outcome data in acutely ill RSI populations were sparse pre-KETASED.
    • Before this trial, the debate was largely driven by observational associations and biochemical endpoints (cortisol/ACTH response), with limited pragmatic randomised evidence linking agent choice to organ dysfunction or mortality.
  • Research Question/Hypothesis
    • In acutely ill adults requiring emergency RSI, does etomidate (0.3 mg/kg) compared with ketamine (2 mg/kg) increase early organ dysfunction (maximum SOFA over the first 3 ICU days)?
    • The trialists anticipated that any clinically meaningful harm from etomidate’s endocrine effects would be most detectable in higher-risk subgroups (notably suspected sepsis or trauma).
  • Why This Matters
    • Induction agent choice is one of the few immediately modifiable components of emergency airway management, but must be justified by patient-centred benefit (not only procedural conditions or biochemical surrogates).
    • A robust pragmatic RCT could reframe practice away from pathophysiological plausibility alone and towards demonstrable differences in organ dysfunction, resource use, and mortality.

Design & Methods

  • Research Question: In acutely ill adults requiring emergency RSI, does etomidate (0.3 mg/kg) versus ketamine (2 mg/kg) affect early organ dysfunction (maximum SOFA score during the first 3 ICU days)?
  • Study Type: Multicentre, randomised, controlled, single-blind trial (ICU clinicians blinded); investigator-initiated; conducted in France across prehospital and emergency department settings with ICU follow-up (April 2007 to February 2008).
  • Population:
    • Setting: Emergency medical services/prehospital and emergency departments (12 participating centres), with subsequent ICU care across 65 ICUs.
    • Inclusion criteria: Adults (≥18 years) requiring sedation for emergency tracheal intubation with RSI.
    • Exclusion criteria: Cardiac arrest; contraindication to succinylcholine, ketamine, or etomidate; known pregnancy.
    • Consent: Emergency inclusion with deferred consent/confirmation; participants could withdraw consent (post-randomisation exclusions described in Internal Validity).
  • Intervention:
    • Induction agent: Etomidate 0.3 mg/kg IV bolus.
    • Neuromuscular blockade: Succinylcholine 1 mg/kg IV administered after induction.
    • Post-intubation sedation: Midazolam infusion 0.1 mg/kg/h plus fentanyl 2–5 μg/kg/h or sufentanil 0.2–0.5 μg/kg/h (target Ramsay sedation score 3–4).
  • Comparison:
    • Induction agent: Ketamine 2 mg/kg IV bolus.
    • Neuromuscular blockade: Succinylcholine 1 mg/kg IV administered after induction.
    • Post-intubation sedation: Same protocolised midazolam + opioid regimen as the etomidate group.
  • Blinding: Single-blind: ICU nurses and intensivists were blinded to the induction agent (study drug not recorded in ICU notes); the intubating emergency physician was not blinded.
  • Statistics: Planned 650 participants to ensure at least 130 evaluable patients with suspected sepsis or trauma, powered to detect a 2-point difference in maximum SOFA score (assumed SD 4) with 80% power at a 5% two-sided significance level, allowing for ~30% exclusions and ~5% missing data; primary analysis was pre-specified modified intention-to-treat for the primary endpoint (with sensitivity analyses in an intention-to-treat population).
  • Follow-Up Period: Primary endpoint assessed over the first 3 ICU days; clinical outcomes reported to day 28.

Key Results

This trial was not stopped early. Recruitment occurred April 2007 to February 2008; no interim stopping was reported and the planned sample size was achieved.

Outcome Etomidate Ketamine Effect p value / 95% CI Notes
Maximum SOFA score over first 3 ICU days 10.3 ± 3.7 9.6 ± 3.9 Mean difference (E−K) 0.7 95% CI 0.0 to 1.4; P=0.056 Primary endpoint; modified ITT population (n=234 vs 235)
28-day mortality 81/234 (35%) 72/235 (31%) Absolute difference (E−K) 4% 95% CI −4 to 12; P=0.36 Modified ITT population
Catecholamine support (any) 137/234 (59%) 120/235 (51%) Absolute difference (E−K) 7.5% 95% CI −1.5 to 16.5; P=0.10 Modified ITT population
Difficult intubation 24/234 (10%) 20/235 (9%) Absolute difference (E−K) 2% 95% CI −4 to 7; P=0.52 Modified ITT population; defined using an intubation difficulty scale threshold
Cardiac arrest during intubation 7/234 (3%) 4/235 (2%) Absolute difference (E−K) 1.3% 95% CI −1.5 to 4.0; P=0.36 Procedural adverse event; modified ITT population
Non-responder to ACTH test 93/116 (81%) 49/116 (42%) OR (E vs K) 5.8 95% CI 3.2 to 10.5; P<0.0001 ACTH testing subset (n=116 vs 116); test performed ~48 h after ICU admission
Adrenal insufficiency (study definition) 100/116 (86%) 56/116 (48%) OR (E vs K) 6.7 95% CI 3.5 to 12.7; P<0.0001 ACTH testing subset (n=116 vs 116)
    • Primary endpoint (modified ITT): maximum SOFA 10.3 ± 3.7 (etomidate) vs 9.6 ± 3.9 (ketamine) (mean difference 0.7; 95% CI 0.0 to 1.4; P=0.056); intention-to-treat sensitivity analysis reported mean difference 0.4 (95% CI −0.2 to 1.0; P=0.20).
    • 28-day mortality (modified ITT): 35% vs 31% (absolute difference 4%; 95% CI −4 to 12; P=0.36); subgroup odds ratios for death (etomidate vs ketamine): septic 1.4 (95% CI 0.5 to 3.5), trauma 0.8 (95% CI 0.4 to 2.0), non-trauma/non-septic 1.3 (95% CI 0.8 to 2.1).
    • Marked endocrine separation in those tested (~48 h): adrenal insufficiency 86% vs 48% (OR 6.7; 95% CI 3.5 to 12.7; P<0.0001), despite broadly similar intubation conditions and procedural adverse events.

Internal Validity

    • Randomisation and allocation: Computer-generated list (block size 4), stratified by centre; sequentially numbered identical drug boxes supported allocation concealment until drug administration.
    • Drop out / post-randomisation exclusions: 655 randomised; 650 analysed in the intention-to-treat population (325 per group) after 1 withdrawal of consent and 4 cases with missing data; primary endpoint analysed in a modified intention-to-treat population of 469 (234 etomidate vs 235 ketamine) after excluding patients discharged from ICU alive within 3 days (79 vs 75) and patients dying before reaching hospital (12 vs 15).
    • Performance/detection bias: Intubating clinicians were unblinded (potential performance bias), but ICU clinicians were blinded (mitigating detection bias for ICU-delivered outcomes and SOFA ascertainment).
    • Protocol adherence / separation: All randomised patients received allocated induction agent; procedural separation appeared good (intubation difficulty scale median 1 [IQR 0–3] in both groups; difficult intubation 10% vs 9%); endocrine separation was large in the tested subset (adrenal insufficiency 86% vs 48%; non-responder 81% vs 42%).
    • Baseline comparability: Broadly similar severity at inclusion (e.g., age 57 ± 18 vs 59 ± 19 years; SAPS II 51.2 ± 16.0 vs 50.5 ± 15.7); heterogeneous case-mix with pre-specified high-risk subgroup (suspected sepsis or trauma) (n=180 in subgroup analysis).
    • Heterogeneity: Mixed indications and settings (prehospital/ED), increasing generalisability but diluting mechanism-specific effects; subgroup analyses (sepsis; trauma) were underpowered for mortality and show wide confidence intervals.
    • Timing and dose: Induction agent administered immediately before intubation at standard doses (etomidate 0.3 mg/kg; ketamine 2 mg/kg); endocrine testing occurred ~48 h and was not universal (232/650 tested), limiting inference about the full cohort.
    • Outcome assessment: Primary endpoint (maximum SOFA) is objective and widely validated; neurological component used Glasgow Coma Scale just before RSI to reduce sedative confounding; mortality and major procedural complications are objective.
    • Statistical rigour: Planned sample size achieved; primary analysis (modified ITT) and sensitivity analyses (intention-to-treat; adjusted analyses) were reported; primary endpoint narrowly missed conventional statistical significance.

Conclusion on Internal Validity: Overall, internal validity is moderate: randomisation and blinded ICU outcome assessment support causal inference for ICU endpoints, but the pre-specified modified ITT primary analysis excluded a substantial post-randomisation fraction, weakening a strict intention-to-treat interpretation for a population-level emergency airway intervention.

External Validity

    • Population representativeness: Included a broad spectrum of critically ill adults undergoing emergency RSI in prehospital and ED pathways; excluded cardiac arrest and specific contraindications.
    • System and workforce context: Conducted in a French physician-led prehospital model and participating emergency departments; transferability to paramedic-led or anaesthetist-led systems may differ in timing, airway expertise, and co-interventions.
    • Co-interventions and contemporary practice: RSI used succinylcholine and protocolised post-intubation sedation; applicability may be limited where rocuronium is routine or where post-intubation sedation/analgesia strategies differ.
    • Endpoint population: Primary endpoint evaluated only in those remaining in ICU ≥3 days and excluding prehospital deaths, limiting direct applicability to patients with rapidly resolving illness or those dying very early after induction.

Conclusion on External Validity: Generalisability is moderate: findings are most applicable to similar emergency RSI pathways with ICU admission and comparable drug dosing, but should be extrapolated cautiously to different prehospital systems, paralytic choices, and patients at the extremes of early death or rapid recovery.

Strengths & Limitations

  • Strengths:
    • Large, pragmatic, multicentre randomised comparison of two commonly used induction agents in real-world emergency RSI, including prehospital enrolment.
    • Blinding of ICU clinicians (nurses and intensivists) for downstream outcomes, reducing detection bias for SOFA and ICU co-interventions.
    • High protocol fidelity with standardised paralytic dosing and post-intubation sedation targets, and objective outcome definitions.
    • Mechanistic endocrine testing (ACTH stimulation) provided clear biological separation between agents.
  • Limitations:
    • Single-blind design: intubating clinicians were unblinded at drug administration.
    • Primary endpoint analysed in a modified ITT cohort (234 vs 235), excluding patients with early ICU discharge (79 vs 75) and prehospital death (12 vs 15), which complicates population-level interpretation.
    • Not powered for mortality, and subgroup analyses (particularly sepsis) were imprecise with wide confidence intervals.
    • ACTH testing was performed in a subset (116 per group) and at ~48 h, limiting inference about endocrine effects across the full cohort and across earlier time windows.

Interpretation & Why It Matters

  • Implications for practice
    • In this pragmatic emergency RSI cohort, etomidate and ketamine yielded similar procedural conditions and no clear difference in early organ dysfunction or 28-day mortality within the analysed primary endpoint population.
    • Etomidate produced substantial adrenal suppression in those tested (adrenal insufficiency 86% vs 48%), but this biochemical signal did not translate into a definitive organ dysfunction or mortality signal in the overall trial.
    • Agent selection should therefore be integrated into a broader airway bundle (preoxygenation, haemodynamic optimisation, vasopressor readiness, and post-intubation sedation), with particular caution in biologically vulnerable phenotypes (e.g., sepsis) where confidence intervals remain wide.

Controversies & Subsequent Evidence

    • Modified ITT primary endpoint population: Excluding patients dying before hospital arrival (12 vs 15) and those discharged alive from ICU within 3 days (79 vs 75) improves completeness of a 3-day SOFA-based endpoint, but departs from classical intention-to-treat reasoning and may preferentially remove patients in whom early drug effects (benefit or harm) are most consequential.
    • Biochemistry–outcomes disconnect: KETASED demonstrated marked endocrine effects of etomidate without a definitive downstream organ dysfunction or mortality signal; the accompanying Lancet commentary emphasised the trial’s pragmatic achievement in the prehospital domain while underscoring unresolved clinical significance of single-dose etomidate-associated adrenal inhibition.1
    • Follow-up randomised evidence: The later EvK trial used 7-day survival as the primary endpoint and reported lower day-7 survival with etomidate (77.3%) than ketamine (85.1%) (difference −7.8%; 95% CI −13.0 to −2.4; P=0.005), while day-28 survival was similar (64.1% vs 66.8%; difference −2.7%; 95% CI −9.3 to 3.9; P=0.294), sustaining debate about whether any etomidate-associated harm is early and potentially context-dependent.2
    • Contemporary large-trial landscape: A large pragmatic ICU trial comparing ketamine and etomidate has been published subsequently, increasing emphasis on patient-centred outcomes in modern ICU airway practice rather than reliance on endocrine surrogates alone.3
    • Evidence synthesis: Recent meta-analyses (including Bayesian approaches) pool the limited RCT set and highlight imprecision and between-study heterogeneity (population case-mix, endpoints, and follow-up windows), which constrains definitive conclusions about net mortality benefit or harm from either agent.456
    • Guidelines: The 2023 Society of Critical Care Medicine rapid sequence intubation guideline treats both ketamine and etomidate as reasonable induction options in critically ill adults, aligning with persistent uncertainty in net clinical benefit and emphasising individualisation to haemodynamics and contraindications.7

Summary

    • In 655 acutely ill adults undergoing emergency RSI, etomidate (0.3 mg/kg) and ketamine (2 mg/kg) produced similar procedural conditions and no statistically significant difference in early organ dysfunction in the analysed primary endpoint cohort (SOFAmax mean difference 0.7; 95% CI 0.0 to 1.4; P=0.056).
    • 28-day mortality in the modified ITT cohort was not significantly different (35% vs 31%; absolute difference 4%; 95% CI −4 to 12; P=0.36), with wide confidence intervals in sepsis and trauma subgroups.
    • Etomidate caused markedly greater adrenal suppression at ~48 h in those tested (adrenal insufficiency 86% vs 48%; OR 6.7; 95% CI 3.5 to 12.7).
    • The trial’s pre-specified modified ITT approach (excluding early ICU discharge and prehospital death) improves SOFA completeness but complicates strict intention-to-treat interpretation for an emergency airway intervention.
    • KETASED remains landmark pragmatic evidence because it shifted the debate from endocrine surrogates towards hard clinical and organ dysfunction outcomes in real-world prehospital/ED RSI.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Across trials and meta-analyses, direct comparisons are complicated by differing primary endpoints (SOFA-based organ dysfunction versus survival at differing time horizons), enrolment settings (prehospital/ED/ICU), and varying completeness of endocrine testing.

Overall Takeaway

KETASED is landmark because it was among the first large, pragmatic, multicentre randomised trials to compare two ubiquitous RSI induction agents in real-world prehospital and emergency pathways with blinded ICU outcome assessment. It showed that etomidate produces striking adrenal suppression, yet did not demonstrate a clear increase in early organ dysfunction or 28-day mortality in the analysed primary endpoint cohort—shaping subsequent trials and guideline positions that remain permissive while acknowledging residual uncertainty in high-risk subgroups.

Overall Summary

  • Etomidate and ketamine produced similar early organ outcomes and 28-day mortality in KETASED’s primary endpoint population, but etomidate caused markedly more adrenal suppression—keeping the clinical relevance of this endocrine effect central to ongoing evidence synthesis and practice guidance.

Bibliography