Clinical Question
Among critically ill adults undergoing tracheal intubation, does ketamine reduce mortality compared to etomidate?
TL;DR
Ketamine did not reduce mortality compared to etomidate, and was associated with significantly more cardiovascular collapse during intubation—especially in septic shock patients. This large trial challenges the narrative that ketamine is the hemodynamically safer choice and suggests etomidate remains a reasonable option for critically ill patients.
The Deets
For years, we've debated which induction agent to use for intubating critically ill patients. Etomidate has been the go-to in U.S. EDs and ICUs because it doesn't tank blood pressure. But there's long been concern that its adrenal suppression (it inhibits cortisol production for up to 72 hours) might increase mortality, particularly in sepsis. This has led some countries to pull etomidate from their markets and guidelines to recommend ketamine instead.
The RSI trial enrolled 2,365 critically ill adults across 14 U.S. emergency departments and ICUs. Nearly half the patients had sepsis or septic shock, and 22% were already on vasopressors. The median patient age was 60, and baseline severity of illness was substantial (median APACHE II score 18).
Patients were randomized to receive either ketamine or etomidate for intubation induction. The median dose was 140 mg ketamine (1.6 mg/kg) or 20 mg etomidate (0.28 mg/kg). Adherence was excellent—over 99% in each group received their assigned drug. Nearly all patients also received a neuromuscular blocker (rocuronium 69%, succinylcholine 31%).
The primary outcome was in-hospital death by day 28. The secondary outcome was cardiovascular collapse during intubation, defined as systolic BP below 65 mm Hg, new or increased vasopressors, or cardiac arrest within 2 minutes post-intubation.
Mortality was essentially identical: 28.1% with ketamine vs. 29.1% with etomidate (adjusted risk difference −0.8 percentage points; 95% CI −4.5 to 2.9; P=0.65). This held true across all subgroups, including the 1,101 patients with sepsis.
But ketamine had significantly more cardiovascular collapse: 22.1% vs. 17.0% (risk difference 5.1 percentage points; 95% CI 1.9 to 8.3). Among septic shock patients specifically, collapse rates were 30.6% with ketamine vs. 20.9% with etomidate—a nearly 10 percentage point difference. Ketamine also caused more drops in systolic BP below 80 mm Hg (14.4% vs. 10.6%) and more ventricular tachycardia (1.0% vs. 0.2%).
This contradicts the prevailing wisdom that ketamine maintains hemodynamic stability better than etomidate. While ketamine does increase endogenous catecholamines, it's also a direct myocardial depressant and vasodilator-effects that appear to dominate in critically ill patients.
Design
Multicenter, unblinded, parallel-group, randomized trial
N=2,365
Ketamine (n=1,176)
Etomidate (n=1,189)
Follow-up: 28 days
14 sites (6 EDs, 8 ICUs) across 6 U.S. medical centers
Population
Inclusion Criteria
Adults ≥18 years old
Critically ill patients in ED or ICU
Undergoing tracheal intubation with use of an induction medication
Exclusion Criteria
Known pregnancy
Known prisoners
Primary diagnosis of trauma
Immediate need for intubation precluding randomization
Clinician determination that ketamine or etomidate was necessary or contraindicated
Baseline Characteristics
From the ketamine group
Demographics: Median age 60 years (IQR 45-69), 42% female, 58% White, 26% Black, 11% Hispanic
Location: ED 56%, ICU 44%
Conditions: Sepsis/septic shock 46%, acute respiratory condition 58%, acute cardiac condition 18%
Chronic conditions: Hypertension 46%, cirrhosis 14%, CHF 15%, cancer 19%, adrenal insufficiency or long-term glucocorticoids 12%
Severity: Median APACHE II 18 (IQR 13-24), median GCS 11 (IQR 7-15)
Hemodynamics: Median lowest SBP in prior hour 115 mm Hg (IQR 96-136), vasopressor use 21%
Interventions and Controls
Randomly assigned to a group:
Ketamine - Ketamine IV with dose selected by clinician based on weight-based nomogram. Clinicians chose full dose (2.0 mg/kg), intermediate dose (1.5 mg/kg), or reduced dose (1.0 mg/kg). Median dose administered: 140 mg (1.6 mg/kg, IQR 1.4-2.0).
Etomidate - Etomidate IV with dose selected by clinician based on weight-based nomogram. Clinicians chose full dose (0.3 mg/kg), intermediate dose (0.25 mg/kg), or reduced dose (0.2 mg/kg). Median dose administered: 20 mg (0.28 mg/kg, IQR 0.24-0.31).
99.2% of ketamine group received ketamine; 99.6% of etomidate group received etomidate
Neuromuscular blocking agent used in 99.7% of both groups (69% rocuronium, 31% succinylcholine)
Video laryngoscopy used in ~95% of first attempts
Median time from induction to intubation: 112 seconds (ketamine) vs. 103 seconds (etomidate)
Outcomes
Comparisons are ketamine vs. etomidate.
Primary Outcome
In-hospital death by day 28
28.1% vs. 29.1% (adjusted risk difference −0.8 percentage points; 95% CI −4.5 to 2.9; P=0.65)
Death in any location by day 28 (sensitivity analysis)
32.2% vs. 32.4% (adjusted risk difference 0.0 percentage points; 95% CI −3.9 to 3.9)
Secondary Outcome
Cardiovascular collapse during intubation
22.1% vs. 17.0% (risk difference 5.1 percentage points; 95% CI 1.9 to 8.3)
Component | Ketamine | Etomidate |
|---|---|---|
Systolic BP <65 mm Hg | 6.4% | 5.5% |
New or increased vasopressors | 21.3% | 15.9% |
Cardiac arrest | 1.0% | 0.8% |
Exploratory Outcomes
Outcome | Ketamine | Etomidate |
|---|---|---|
Median lowest SBP during intubation | 112 mm Hg | 118 mm Hg |
Lowest SBP <80 mm Hg | 14.4% | 10.6% |
SBP drop >30 mm Hg | 23.9% | 14.7% |
Highest SBP >180 mm Hg | 13.5% | 16.5% |
Lowest oxygen saturation <80% | 11.1% | 11.1% |
Successful intubation on first attempt | 85.7% | 86.7% |
Median ventilator-free days | 23 (IQR 0-26) | 23 (IQR 0-26) |
Median vasopressor-free days | 25 (IQR 0-28) | 25 (IQR 0-28) |
Median ICU-free days | 20 (IQR 0-24) | 19 (IQR 0-24) |
Subgroup Analysis
All subgroup analyses showed consistent results-no mortality benefit from ketamine in any patient population.
Subgroup | Finding |
|---|---|
Location (ED vs. ICU) | No significant interaction |
Sepsis or septic shock | Mortality: 38.8% vs. 38.2% |
Collapse: 30.6% vs. 20.9% | |
Vasopressor use before enrollment | No significant interaction |
Adrenal insufficiency or glucocorticoid use | No significant interaction |
Acute neurologic condition | No significant interaction |
Acute cardiac condition | No significant interaction |
APACHE II score (<20 vs. ≥20) | High severity: collapse 31.4% vs. 20.7% |
Safety Outcomes
Event | Ketamine | Etomidate |
|---|---|---|
Ventricular tachycardia (post hoc) | 1.0% | 0.2% |
Vasopressor use at 24 hours | 38.9% | 42.3% |
Median SBP at 24 hours | 114 mm Hg | 114 mm Hg |
Criticisms
The trial excluded trauma patients, limiting generalizability to that population—an important gap since trauma represents a significant proportion of ED intubations.
The trial was unblinded, which could have influenced observer assessments of intraprocedural outcomes or clinician treatment decisions after intubation.
Ventricular tachycardia was a post hoc outcome added after enrolling 567 patients, though the finding remains concerning and warrants further investigation.
The trial doesn't address other commonly used induction agents like propofol or benzodiazepines, which some clinicians prefer in specific scenarios.
While the trial was powered for mortality, it wasn't powered to detect small differences that might still be clinically meaningful at scale—and for medications this commonly used, even small differences could matter at the population level.
The confidence intervals for the primary outcome don't completely rule out a small mortality benefit (or harm) from ketamine, though any true effect is likely modest.
Clinicians could choose doses within a range, introducing variability that might have diluted treatment effects—though this also increases real-world applicability.
Dashevsky's Dissection
For patients: If you need to be intubated emergently in an ICU or ED, this trial provides reassurance that the choice between ketamine and etomidate won't meaningfully affect your chance of surviving to hospital discharge. Both drugs are safe and effective when used appropriately. The slightly higher risk of blood pressure drops with ketamine during the procedure is typically manageable and doesn't translate to worse long-term outcomes.
For pulmonary and critical care physicians: This trial fundamentally changes how we should think about induction agents for critically ill patients. For years, we've been taught that etomidate's adrenal suppression is dangerous, particularly in sepsis. Some of us switched to ketamine reflexively for any sick patient. This trial, with more patients than all prior randomized trials combined, shows that concern was overblown.
The mortality data is unequivocal. No difference overall. No difference in sepsis. No difference in any subgroup. The pathophysiologic rationale that etomidate blocks cortisol, cortisol is critical in shock, therefore etomidate kills patients—doesn't hold up clinically. This is a good reminder that elegant biochemical mechanisms don't always translate to bedside outcomes in the messy complexity of critical illness.
What this trial does show is more cardiovascular instability with ketamine. We've been taught that ketamine is the hemodynamically stable choice because it increases endogenous catecholamines. But ketamine is also a direct myocardial depressant and vasodilator, and in critically ill patients—especially those already in shock—those negative effects appear to dominate.
The practical implication: we should stop reflexively choosing ketamine for hemodynamically unstable patients. For patients already on pressors or with septic shock, etomidate may actually be the better choice. The decades-long concern about adrenal suppression appears to have been a red herring.
For the health system: This trial has immediate practice implications across thousands of U.S. hospitals. Etomidate is widely available, inexpensive, and familiar to clinicians. If we've been avoiding it unnecessarily—potentially causing more hemodynamic instability by defaulting to ketamine—that's a systems-level problem we can fix immediately with education and updated protocols.
The trial also raises questions about how we generate and disseminate evidence. The concern about etomidate in sepsis was based largely on observational data and pathophysiologic rationale. Some countries banned the drug. Guidelines recommended against it. Yet this large, well-designed trial shows the concern was unfounded. We need better mechanisms to rapidly test and update practice when clinical equipoise exists.
From a resource perspective, both drugs are similarly priced and available. The choice won't meaningfully affect hospital budgets. But the reduction in peri-intubation cardiovascular collapse with etomidate could reduce the need for additional vasopressor boluses, additional personnel at the bedside, and prolonged post-intubation hemodynamic support. Those downstream effects are harder to quantify but could represent real efficiency gains.
The broader lesson: in critical care, we're often choosing between imperfect options with different trade-offs. Etomidate suppresses cortisol but maintains hemodynamics. Ketamine maintains catecholamines but depresses myocardial function. Neither is perfect. But this trial clearly shows that for most critically ill patients, etomidate is at least as safe as ketamine-and for hemodynamically unstable patients, it may be safer.
Casey JD, Seitz KP, Driver BE, et al. Ketamine or Etomidate for Tracheal Intubation of Critically Ill Adults. N Engl J Med. Published online December 9, 2025. doi:10.1056/NEJMoa2511420