KETASED: Etomidate versus ketamine for rapid sequence intubation in acutely ill patients: a multicentre randomised controlled trial

Jabre & KETASED Collaborative Study Group. Lancet 2009; 374:239-300. doi:10.1016/S0140-6736(09)60949-1

Clinical Question

  • In critically ill adults, does a single induction dose of etomidate compared to ketamine affect organ failure scores during the subsequent first 3 days?


  • Randomised, controlled, prospective trial
  • Randomisation by computer in blocks of 4
  • Study drug in sealed, sequentially numbered boxes
  • Single-blinded: recruiting emergency physician aware, but nurses and intensive care physicians blinded
  • Modified Intention-to-treat analysis (mITT) without post-randomisation exclusions – see below
  • 28-day follow-up period
  • Aimed to recruit 650 patients after complex power calculation, to allow appropriate analysis of sub-group with final diagnosis of confirmed sepsis or trauma
    • Defined clinically relevant difference in maximum SOFA score was 2 points
    • Powered at 80% – 20% chance of false negative
    • Type-I error defined at 0.05 – 5% chance of false positive
    • Anticipated 30% exclusion post-randomisation (modified intention-to-treat analysis)
    • Estimated sub-group of interest would be 30% of total population


  • 12 French emergency medical services (EMS) or emergency departments (ED)
    • An EMS is an ambulance with driver, nurse and emergency doctor
  • 65 French intensive care units (ICU)
  • April 2007 to February 2008


  • Inclusion: 18 years or older; needed sedation for emergency intubation
  • Exclusion: cardiac arrest; contraindications to suxamethonium, ketamine or etomidate; pregnancy.
    • Post-randomisation exclusion of those:
      • discharged from ICU within 3 days, “to retain only the most severely ill patients”
      • died before reaching hospital, “because their death could not reasonably have been attributed to sedation use”
  • 689 patients screened, 655 randomised, and 469 in mITT analysis


  • Ketamine (Ketalar, Panpharma)
    • 2 mg/kg intravenous bolus


  • Etomidate (Lipuro, B Braun Medical)
    • 0.3 mg/kg intravenous bolus
  • Both groups received suxamethonium 1 mg/kg and then continuous infusions of midazolam (0.1 mg/kg/hr )with either fentanyl (2–5 mcg/kg/hr) or sufentanil (0.2–0.5 mcg/kg/hr)


  • Primary outcome: there was no difference in maximum sequential organ failure assessment (SOFA) scores between the two groups.
    • Mean SOFA scores were:
      • Etomidate: 10.3 (SD 3.7)
      • Ketamine: 9.6 (SD 3.9)
    • Absolute difference (AD) was 0.7 (95% CI 0.0 to 1.4)
    • p-value 0.056 by generalised linear model analysis
  • Secondary outcome: there were no differences in any of the secondary outcomes between the two groups
  • Sub-group analysis: there were no differences seen in the sub-groups of sepsis, trauma or sepsis and trauma for the primary outcome of maximum SOFA score but small sample size increase change of false-negative
    • Sepsis and trauma n=180
    • Sepsis n=76
    • Trauma n=104
  • Tertiary data: basal cortisol level were lower and the proportion of patients identified as having adrenal insufficiency were higher in the etomidate group
    • As a subgroup analysis, ACTH test non-responders vs responders showed no difference in mortality outcome

Authors’ Conclusions

  • “Ketamine is a safe and valuable alternative to etomidate for intubation in critically ill patients, particularly in septic patients”


  • State-funded study, with no input into design, analysis or manuscript by funding body
  • A priori publication of study methods and statistical plan, particularly regarding focus on sub-group analysis
  • Allocation concealment was maintained, so unblinded emergency physician could not interfere with recruitment bias
  • Both modified and full intention-to-treat analyses presented for sensitivity analysis


  • Since etomidate not available in some countries, findings are not applicable everywhere!
  • Not entirely clear what is control and what is intervention. Are the authors trying to prove ketamine better than usual care of etomidate, or that etomidate is more harmful? Perhaps non-inferiority design would have been better.
  • Due to non-normal distribution of the primary outcome, the authors used an advanced generalised linear model analysis. This makes the meaningfulness of the result hard to interpret, which at a p-value = 0.056 it is only just statistically non-significant.
    • Conventional Student’s t-test (which may not be applicable) gives p-value = 0.0467 in favour of ketamine.
  • Sub-group data in figure 3 missing negative sign and graphical plot does not match numerical data.

The Bottom Line

  • This study investigates an interesting clinical dilemma and does not demonstrate any important clinical difference in organ failure scores between ketamine and etomidate when used as a single bolus for induction
  • It is my impression that the likelihood of a type II error (false negative) is high in this study and there is a trend suggesting benefit from ketamine in patients with sepsis compared against etomidate

External Links


Summary author: @DuncanChambler
Summary date: 22 October 2014
Peer-review editor: @DavidSlessor

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.