RSI: Ketamine or Etomidate for Tracheal Intubation of Critically Ill Adults

Emergency Medicine, Intensive Care Medicine Leave a comment

Ketamine or Etomidate for Tracheal Intubation of Critically Ill Adults

Casey. NEJM. 2025. doi: 10.1056/NEJMoa2511420

Clinical Question

  • Among adults with non‑traumatic critical illness undergoing emergency tracheal intubation in the ED or ICU, how does induction with ketamine (1–2 mg/kg IV, dosed by actual body weight) versus etomidate (0.2–0.3 mg/kg IV, dosed by actual body weight) affect 28‑day mortality?

Background

  • Peri-intubation cardiovascular instability is common worldwide in critically ill patients, and is associated with worse mortality
    • In the INTUBE study, cardiovascular instability occurred in 43.4% of cases. Patients with peri-intubation cardiovascular instability/collapse were at a higher risk of ICU mortality with an adjusted OR of 2.47 (95% CI, 1.72–3.55)
    • The choice of induction agent influences the risk of cardiovascular instability. Among available options, ketamine and etomidate are widely regarded as the most hemodynamically stable
  • Existing evidence: 
    • Etomidate is an inhibitor of 11-β-hydroxylase, which is required for corticosteroid synthesis, and leads to transient adrenal suppression (up to 72 hours). The clinical significance of this remains uncertain
    • Existing meta-analyses (e.g. Koroki et al (2024) and Greer et al (2025)), including the KETASED and EvK trials, provide mixed conclusions, some suggest worse outcomes – especially in sepsis – with etomidate compared with ketamine
    • Etomidate, although commonly used for rapid sequence intubation in the USA, is not available in many countries due regulatory concerns about the drug’s safety

Design

  • Multicentre pragmatic randomized, parallel-group, unblinded trial
    • Investigator-initiated; IRB-approved at Vanderbilt; FDA approval IND 141424; conducted under exception from informed consent (EFIC); independent data safety monitoring board (DSMB); protocol and statistical analysis plans (SAP) published before completion of enrolment
    • Registered at ClinicalTrials.gov prior to initiation
  • Randomisation and allocation concealment
    • 1:1 allocation to ketamine or etomidate using permuted blocks of variable size
    • Stratification by trial site
    • Sequentially numbered, opaque envelopes; concealed until after enrollment
  • Open-label
    • Clinicians and research personnel aware of assignment post-randomization
  • Data collection:
    • Peri-intubation observation: Trained observer (not performing intubation) recorded intubation duration, systolic BP, and vasopressor administration from induction to 2 minutes post-intubation
    • Immediate clinician report: Arrhythmia or cardiac arrest occurrence between induction and 2 minutes post-intubation
    • Medical record review: Baseline characteristics, periprocedural care, clinical outcomes.
    • Mortality ascertainment: Hospital records, public vital statistics, and phone follow-up planned for 3 and 12 months
  • Sample size and power:
    • Assumptions: 30.0% incidence of in-hospital death by 28 days in etomidate group; alpha 0.05; 80% power
    • Calculated size: 2,308 patients to detect an absolute difference of 5.2%
    • Planned enrollment: 2,364 total (1,182 per group) allowing up to 3% missing data
    • Interim analysis: Single interim after 1,182 enrolled; stopping threshold P ≤ 0.001 for primary outcome difference
  • Statistical analysis
    • Primary analysis: Intention-to-treat using generalized linear mixed-effects model (random effect for site; fixed effect for group; no covariate adjustment)
    • Sensitivity analyses: Unadjusted chi-square; covariate-adjusted analysis; missing-data extremes; death in any location by 28 days; Kaplan–Meier survival to day 28
    • Effect modification: Subgroup analysis with interaction terms; baseline risk-modelling using validated models; individualized treatment effect via machine learning
    • Secondary/exploratory outcomes: Chi-square for categorical; Wilcoxon rank-sum for continuous/ordinal; report point estimates and 95% CIs; no multiplicity adjustment
    • Software: R version 4.4.2

Setting

  • 14 sites (6 EDs, 8 ICUs) across 6 US medical centres
  • April 2022 – August 2025

Population

  • Inclusion criteria: 
    • Adults ≥18 years; critically ill; undergoing tracheal intubation with anaesthetic induction medication
  • Exclusion criteria: 
    • Known pregnancy
    • Prisoner status
    • Primary trauma diagnosis
    • Immediate need for intubation precluding randomization
    • Clinician determination that ketamine or etomidate was necessary or contraindicated
  • Screening process:
    • Enrolment at point of care prior to induction; randomization feasible only with equipoise and sufficient time
  • Participant flow:
    • Screened/enrolled: 3,439 assessed; 2,367 enrolled (68.8%)
    • Analyzed: 2,365 (2 excluded post-enrollment as prisoners)
    • Groups: 1,176 ketamine (49.7%); 1,189 etomidate (50.3%)
    • Setting: 55.7% ED; 44.3% ICU
  • Baseline characteristics were similar between groups (ketamine vs etomidate): 
    • Demographics
      • Age: 60 vs 60
      • Female sex: 42.3% vs 41.4%
      • Non‑Hispanic White: 58.3% vs 59.4%
      • Non‑Hispanic Black: 25.5% vs 24.1%
      • Hispanic: 11.1% vs 11.1%
      • Other race/ethnicity: 5.1% vs 5.4%
      • Weight (kg): 78.9 vs 78.5
      • BMI: 26.9 vs 26.7
      • ED intubation: 56.4% vs 55.1%
      • ICU intubation: 43.6% vs 44.9%
    • Chronic conditions
      • Adrenal insufficiency / chronic steroids: 11.6% vs 10.8%
      • Cirrhosis: 14.0% vs 14.0%
      • Congestive heart failure: 14.9% vs 13.3%
      • Coronary artery disease: 12.0% vs 12.9%
      • Hypertension: 45.6% vs 44.8%
      • Malignancy: 19.3% vs 18.1%
    • Acute conditions
      • Acute cardiac condition: 18.4% vs 18.8%
      • Acute respiratory condition: 57.7% vs 57.4%
      • Acute neurologic condition: 10.6% vs 10.2%
      • Sepsis or septic shock: 45.8% vs 47.5%
    • Severity of illness
      • GCS: 11 vs 11
      • APACHE II: 18 vs 18
    • Pre‑enrollment physiology
      • Highest HR: 107 /min vs 108 /min
      • Lowest SBP: 115 mmHg vs 114 mmHg
      • Vasopressor use: 20.9% vs 23.0%
      • Median SBP at induction 127 mmHg vs 127 mmHg
      • Vasopressor bolus or increased infusion rate at induction 17.6% vs 19.7%
    • Primary reason for intubation
      • Encephalopathy 36.2% vs 35.5%
      • Hypoxemic respiratory failure 32.5% vs 34.7%
      • Hypercarbic and hypoxemic respiratory failure 6% vs 5.8%
      • Hypercarbic respiratory failure 3.9% vs 4.0%
      • Emergency procedure 7.6% vs 6.6%

Intervention

  • Ketamine:
    • Clinician-selected dose using nomogram: full 2.0 mg/kg, intermediate 1.5 mg/kg, reduced 1.0 mg/kg (dosing based on actual body weight)

Control

  • Etomidate: 
    • Clinician-selected dose using nomogram: full 0.3 mg/kg, intermediate 0.25 mg/kg, reduced 0.2 mg/kg (dosing based on actual body weight)

Management common to both groups

  • Co-interventions and other therapies were discretionary and similar in both groups (ketamine vs etomidate):
    • Pre-oxygenation
      • Non-rebreather mask: 46.2% vs 45.8%
      • NIV: 27.5% vs 25.7%
    • Intubation
      • Number of intubations performed by intubator (median) 55 vs 53
      • Video laryngoscopy used on first attempt 95.% vs 94.8%
      • Stylet 57.3% vs 58.1%
      • Bougie 38.8% vs 37.6%
      • First pass success 85.7% vs 86.7%
      • Failure to intubate 0% vs 0.3%
    • Neuromuscular blocker choice
      • Rocuronium 69% vs 69%
      • Suxamethonium 30.8% vs 30.7%
    • Medications administered after intubation and before 28 days
      • Corticosteroids 45.9% vs 49.5%
  • Protocol adherence
    • Induction agent received:
      • Ketamine 1,167/1,176 (99.2%); etomidate 1,184/1,189 (99.6%)
    • Dosing:
      • Ketamine median 140 mg (IQR 100–150), 1.6 mg/kg (IQR 1.4–2.0)
      • Etomidate median 20 mg (IQR 20–25), 0.28 mg/kg (IQR 0.24–0.31)
    • Co-induction meds:
      • ~2% in each group received propofol/benzodiazepines/opioids
    • Neuromuscular blocker: 99.7% in both groups received an NMBA

Outcomes

  • All are comparing ketamine vs etomidate
  • Primary outcome
    • In-hospital death by 28 days:
      • 28.1% ketamine (330/1,173) vs 29.1% etomidate (345/1,186)
      • Effect: Absolute risk difference (site-adjusted): −0.8 percentage points (95% CI −4.5 to 2.9; P=0.65)
    • No difference when adjusted for death in any location by 28 days: 32.2% vs 32.4% 
      • Effect: Absolute risk difference (ARD):  0.0 percentage points (95% CI −3.9 to 3.9)
    • Robustness:
      • Similar across sensitivity analyses and subgroups, including sepsis (38.8% vs 38.2%; ARD 1.0 [95% CI −4.8 to 6.7])
    • Heterogeneity:
      • No treatment-effect heterogeneity by risk- or effect-modeling
  • Secondary outcome – cardiovascular collapse during intubation:
    • Definition:
      • Occurrence of any of the following between induction and 2 minutes after intubation: (1) systolic blood pressure < 65 mmHg; (2) new or increased vasopressors; (3) cardiac arrest* not resulting in death; (4) cardiac arrest* resulting in death.
    • 22.1% ketamine (260) vs 17.0% etomidate (202)
      • Effect: Absolute risk difference 5.1 percentage points (95% CI 1.9 to 8.3).
      • Components of this composite outcome:
        • SBP < 65 mmHg: 6.4% vs 5.5%
        • Cardiac Arrest: 1.0% vs 0.8%
        • New or increased vasopressors: 21.3% vs 15.9%
    • Key subgroups:
      • Sepsis: 30.6%  vs 20.9%
        • ARD 9.7 (95% CI 4.6 to 14.9)
      • APACHE II ≥20: 31.4% vs 20.7% 
        • ARD 10.7 (95% CI 5.5 to 16.0)
  • Exploratory outcomes
    • Lowest SBP (induction to +2 min):
      • Median 112 (IQR 92–138)  vs 118 (IQR 98–141)
      • Median difference −6 (95% CI −9 to −1)
    • SBP <80 mm Hg:
      • 14.4% vs 10.6%; ARD 3.8 (95% CI 1.1 to 6.5)
    • SBP drop >30 mm Hg:
      • 23.9% vs 14.7%; ARD 9.2 (95% CI 6.0 to 12.5)
    • Time to successful intubation:
      • Median 112 s (IQR 86–155)  vs 103 s (IQR 80–134)
      • Median difference 9 s (95% CI 5 to 14)
    • Resource/trajectory metrics:
      • Ventilator-free, vasopressor-free, and ICU-free days appeared similar between groups
  • Safety outcomes
    • Ventricular tachycardia (post hoc, early subset):
      • 1.0%  (9/884) vs 0.2% (2/905); ARD 0.8 (95% CI 0.1 to 1.5)
    • Vasopressors at 24 hours:
      • 38.9% (420) vs 42.3% (458); ARD −3.4( 95% CI −7.5 to 0.7)

Authors’ Conclusions

  • Among critically ill adults undergoing tracheal intubation, the use of ketamine to induce anaesthesia did not lead to a significantly lower incidence of in-hospital death by day 28 than etomidate

Strengths

  • Pragmatic multiple center design 
  • Clear end points
    • Patient-centred 28-day in-hospital mortality and tightly timed peri-intubation collapse definition
  • Methodological rigor was high
    • Prepublished protocol/SAP; EFIC; site-stratified randomization; mixed-effects primary analysis; comprehensive sensitivity and heterogeneity analyses
  • Innovative effect-modeling approach: used a machine learning model to predict the effect of ketamine versus etomidate on the primary outcome for each patient based on his or her individual characteristics (individualized treatment effect)
    • Trial data and statistical modelling was used to estimate: “For a specific patient, is ketamine or etomidate safer?”
  • Adherence to intervention arms was excellent (>99%)
  • Follow up completion was excellent (>99%)
  • Baseline balance was excellent
  • Intention to treat analysis performed
  • Allocation concealment preserved randomisation
  • Power calculation is rational based on accurately predicted event rate and pre-existing studies suggesting harm from etomidate
  • Detailed and comprehensive description of procedures
  • No concerns regarding regulatory oversight, funding, or conflicts of interest

Weaknesses

  • Unblinded design
    • However, ascertainment bias was unlikely as observers were recording objective measures
  • Under-powered
    • powered for an absolute mortality difference of 5% and smaller mortality differences (e.g. 1%) could still be very important at the population level for commonly used medications
  • Composite secondary outcome
    • largely driven by non-patient centered components (hypotension, vasopressor use). However, hypotension is associated with mortality and expert clinicians actively strive to avoid hypotension when intubating critically patients
  • Limits to external validity
    • US-only centers
    • exclusion of primary trauma may limit applicability in mixed resuscitation settings. 
    • High use of NRB for pre-oxygenation (CPAP/NIV for pre-oxygenation may have different hemodynamic effects)
    • Potential selection bias (likely minor effect on external validity): of the 3439 patients that met inclusion criteria, the following were not randomised:
      • 347 who were intubated too urgently to complete trial procedures
      • 244 had clinician decline enrollment because a specific induction medication was determined to be required
  • Ketamine dosing
    • Actual bodyweight was used for dosing rather than ideal body weight, which is better supported by pharmacokinetic principles (Erstad and Barletta, 2020) and expert consensus (Morgan et al, 2021). This is relevant as half of patients were overweight and over a quarter were living with obesity based on BMI 
    • Furthermore, 25% of patients received higher than the maximum dose indicated on the nomogram (2mg/kg actual body weight) 
    • Many expert conditions use <1 mg/kg when intubating patients at risk of cardiovascular collapse (e.g. 0.5 mg/kg IV IBW)
    • Analysis of dose-dependence of effects in the trial was not performed and would have limited data to analyse lower dose ranges

The Bottom Line

  • The RSI trial demonstrated no difference in 28‑day mortality between etomidate (0.2–0.3 mg/kg IV, actual body weight) and ketamine (1–2 mg/kg IV, actual body weight) when used for induction during rapid sequence intubation
  • Cardiovascular collapse occurred frequently in both groups, but was more common with ketamine (22%) than with etomidate (17%) (NNH = 20). Among patients with sepsis or an APACHE II score >20, the excess risk associated with ketamine was even greater (NNH = 10)
  • These findings warrant reconsideration of current practice, although the trial does not clarify whether reduced‑dose ketamine, etomidate, or delayed sequence intubation offers a safer alternative for critically ill patients at high risk of cardiovascular collapse

External Links

Metadata

Summary author: Chris Nickson
Summary date: 14th December 2025
Peer-review editor: George Walker

Picture by: Tho-Ge/Pixabay

 

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.