PROMINE — Propofol vs Ketamine for Rapid Sequence Intubation

1. Publication

• Title: Propofol versus ketamine in rapid sequence intubation in critically ill patients: a prospective, randomized, controlled trial
• Acronym: PROMINE
• Year & Journal: Intensive Care Medicine, epublished March 23, 2026 (2026;52(3):476-489)
• Citation: Schmidt RC, Zampieri FG, Ramos FJS, et al; PROMINE Investigators. Propofol versus ketamine in rapid sequence intubation in critically ill patients: a prospective, randomized, controlled trial. Intensive Care Med. 2026;52(3):476-489. doi:10.1007/s00134-026-08351-3

2. Context & Rationale

Background: RSI in critically ill patients carries risk of life-threatening peri-intubation events (hypotension, hypoxemia, cardiac arrest) in up to 45% of cases, independently associated with mortality. Propofol is commonly used but can exacerbate hypotension; ketamine has a more favorable hemodynamic reputation and may offer a safer alternative, though this had not been definitively tested in a dedicated RCT for this indication.

Research Question/Hypothesis: Among critically ill patients undergoing RSI, does ketamine reduce the incidence/severity of hypotension in the first 10 minutes post-induction compared with propofol?

Why This Matters: Peri-intubation hemodynamic collapse is a major, potentially preventable source of harm in ICU airway management; identifying a safer induction agent has direct, immediate bedside relevance.

3. Design & Methods

• Study Type: Prospective, randomized, open-label, pragmatic, bicenter trial
• Setting & Centers: 2 ICUs, Brazil (Hospital São Paulo/UNIFESP, 49 beds; SEPACO, 44 beds); conducted October 2021 – October 2023
• Population:
• Inclusion: Critically ill ICU patients requiring orotracheal intubation
• Exclusions: Not detailed in available trial text
• Intervention: Ketamine as hypnotic agent for RSI
• Comparator: Propofol as hypnotic agent for RSI
• Randomization: 1:1, RedCap-based, variable block sizes, stratified by site and vasopressor use during intubation. 207 randomized; 175 in modified ITT analysis
• Blinding: Open-label
• Statistical Power & Follow-Up: Primary outcome: lowest MAP within 10 minutes post-induction. Secondary outcomes (within 1h post-induction): mortality, cardiopulmonary arrest incidence, severe hypotension (SBP <80mmHg), severe hypoxemia (SpO2 <85%), number of intubation attempts. Mortality followed to day 7 and hospital discharge.

4. Key Results

207 randomized (October 2021–October 2023); 175 in modified ITT analysis.

5. Internal Validity Assessment

• Randomization & Allocation: RedCap-based randomization, variable block sizes, stratified by site and vasopressor use — sound methodology.
• Protocol Adherence & Separation: Pragmatic, real-world bicenter design; 175/207 (84.5%) included in modified ITT, a reasonable retention rate.
• Blinding & Detection Bias: Open-label — a relevant limitation for an intervention where clinician expectation about hemodynamic effects could influence peri-induction management (fluid boluses, vasopressor timing) in ways not fully captured by the trial.
• Missing Data & Sensitivity Analyses: Not detailed in available trial text beyond the ITT population definition.
• Overall Internal Validity Conclusion: Moderate — sound randomization and stratification, pragmatic real-world design, but open-label conduct and only moderate sample size (175 analyzed) for detecting mortality differences in a population with >50% hospital mortality; the authors themselves note the MAP difference, while statistically present, was not clinically meaningful or sustained.

6. External Validity Assessment

• Population Representativeness: Critically ill Brazilian ICU patients requiring intubation; notably, per independent commentary, many patients were already in shock and intubated late in the course of illness, contributing to the very high (>50%) overall mortality observed in both arms.
• Practice Context: Two-center, single-country (Brazil) trial; both hypnotic agents (propofol, ketamine/esketamine) are widely available internationally.
• Overall External Validity Conclusion: Moderate — generalizable to critically ill, often late-presenting/already-shocked ICU intubation populations; the very high background mortality in this specific cohort suggests results may be most applicable to similarly late-presenting, high-severity intubation scenarios rather than earlier, less critically unstable RSI situations.

7. Strengths & Limitations

Strengths:

• Directly randomized comparison addressing a genuine, common bedside decision (induction agent choice) with real hemodynamic consequences
• Pragmatic, real-world bicenter design with stratification by vasopressor use (a key confounder)
• Assessed both physiological (MAP) and patient-centered (mortality) outcomes

Limitations:

• Open-label design
• Modest sample size relative to the very high observed mortality, limiting power to detect true mortality differences
• Two-center, single-country trial
• Authors themselves note the primary MAP difference, while measurable, was not clinically meaningful or sustained — tempering the practical significance of the "positive" primary finding

8. Interpretation & Practice Impact

• Clinical Implications: Per the authors' own conclusion, these findings "challenge sedative selection based solely on perceived risk of hypotension" — ketamine produced a smaller MAP drop, but this was not clinically meaningful or sustained, and mortality (though not statistically significant) numerically favored propofol in this cohort.
• Mechanistic Coherence: Independent commentary (Chacko's Critical Care Blogspot) suggests that in a population already in shock and intubated late in their disease course, the choice of induction agent may transiently influence physiology but have little impact on outcomes ultimately driven by underlying illness severity.
• Systems-Level Takeaway: Argues against reflexively choosing ketamine over propofol purely on the assumption of hemodynamic superiority; larger trials are needed to determine whether either agent offers a genuine outcome advantage.

9. Controversies & Subsequent Evidence

• Editorial Commentary: Independent commentary explicitly frames the counterintuitive numerically-higher mortality with ketamine (despite its smaller MAP drop) as likely reflecting the severity and lateness of illness in this specific cohort rather than a true harm signal from ketamine itself — a case study in why physiological surrogates (MAP) don't always predict patient-centered outcomes.
• Guideline Integration: Contributes to an active area of RSI induction-agent research; a related large trial (Ketamine or etomidate for tracheal intubation of critically ill adults, NEJM 2025) and network meta-analyses of induction agents for emergency intubation are part of the same evolving evidence landscape PROMINE sits within.

10. Summary & Executive Takeaway

Summary: PROMINE randomized 207 (175 analyzed) critically ill Brazilian ICU patients undergoing RSI to ketamine or propofol. Ketamine produced a smaller MAP decrease in the first 10 minutes (66 vs 60 mmHg lowest MAP), but the authors describe this difference as not clinically meaningful or sustained. Day-7 mortality (33% vs 23.8%, OR 1.38) and hospital mortality (60.4% vs 50.0%, OR 1.21) were numerically higher with ketamine but not statistically significant.

Overall Takeaway: PROMINE challenges the assumption that induction-agent selection for RSI should be driven primarily by perceived hypotension risk — ketamine's smaller MAP drop did not translate into a clinically meaningful or sustained hemodynamic advantage, nor into better survival, in a population where underlying illness severity likely dominates outcomes; larger trials are needed before changing standard practice based on hemodynamic profile alone.

11. Bibliography

• Russotto V, Tassistro E, Myatra SN, et al. Peri-intubation cardiovascular collapse in patients who are critically ill: insights from the INTUBE study. Am J Respir Crit Care Med. 2022;206:449-458.
• Ketamine or etomidate for tracheal intubation of critically ill adults. N Engl J Med. 2025. doi:10.1056/NEJMoa2511420.