SPICE III试验：早期使用右美托咪定能否改善机械通气ICU患者的生存率？

原文图文 · 图 29 张

REBEL Rundown

Key Points

No Mortality Benefit: Using dexmedetomidine as the primary early sedative for mechanically ventilated ICU patients did not improve 90-day all-cause mortality compared to usual care.

High Failure as Monotherapy: Nearly 65% of patients in the dexmedetomidine arm required supplemental propofol to achieve their target sedation levels.

Marginal Delirium Benefit: Dexmedetomidine resulted in just 1 additional day free from coma/delirium and 1 additional ventilator-free day at 28 days.

Increased Adverse Events: The dexmedetomidine group experienced significantly more adverse events, specifically bradycardia and hypotension, with a Number Needed to Harm (NNH) of 43 for a serious adverse event.

Introduction

Sedation is a cornerstone of care for mechanically ventilated patients in the ICU, with propofol and midazolam traditionally serving as the go-to agents. Dexmedetomidine, an alpha-2 agonist, has gained popularity because it provides sedation without significant depression of the respiratory drive, potentially reducing the duration of mechanical ventilation and delirium. Despite these proposed benefits, direct, large-scale comparisons investigating whether dexmedetomidine actually improves patient-centered outcomes like mortality were lacking. The SPICE III trial sought to answer if early use of dexmedetomidine as a primary or sole sedative improves 90-day survival compared to standard care.

Paper

Shehabi Y, et al. Early Sedation with Dexmedetomidine in Critically Ill Patients. N Engl J Med. 2019;380(26):2506-2517. PMID: 31104004

PREVIOUSLY COVERED ON REBEL EM:

Post Intubation Sedation for pregnant patients

The Mends2 Trial

What They Did

CLINICAL QUESTION

In critically ill adult patients undergoing mechanical ventilation, does the use of dexmedetomidine as the primary sedative, compared to usual care with other sedatives, lead to improvements in 90-day mortality?

STUDY DESIGN

This was a multinational, open-label, randomized, controlled trial. Patients were randomized in a 1:1 ratio using block randomization with a variable block size via a secure website, stratified by site and the presence/absence of sepsis. The trial was powered at 90% to detect an absolute difference of 4.5% in mortality, requiring 4000 patients assuming a 26% baseline mortality. The trial was funded by national research councils in Australia, New Zealand, and Malaysia; the study drug was supplied by Pfizer and Orion Pharma.

Setting: 74 ICUs across 8 countries (Australia, Ireland, Italy, Malaysia, New Zealand, Saudi Arabia, Switzerland, and the UK).

Statistical Analysis: Analyzed using a modified intention-to-treat approach. The primary outcome was evaluated using logistic regression (to find odds ratios) and binomial regression (to find risk differences), with both models strictly adjusted for the patient’s baseline sepsis status and within-center clustering. Highly skewed continuous variables, such as ventilator-free days, were appropriately analyzed using median regression

POPULATION

Inclusion Criteria:

Adults age 18 or older

Receiving mechanical ventilation via endotracheal tube

Expected to continue receiving ventilatory support beyond the next full calendar day

Receiving sedatives for safety and comfort.

Exclusion Criteria:

Invasive ventilation in the ICU for longer than 12 hours prior to enrollment

Suspected or proven acute primary brain injury

ICU admission due to drug overdose or burns.

INTERVENTION & COMPARATOR

Intervention:

Dexmedetomidine was initiated at a dose of 1 mcg/kg/hr without a loading dose and titrated (maximum 1.5 mcg/kg/hr) to maintain a Richmond Agitation and Sedation Scale (RASS) score of -2 to +1. If the maximum dose was insufficient, propofol at the lowest possible dose was permitted, while midazolam was discouraged.

Comparator:

Usual care consisted of propofol, midazolam, or other sedatives directed by the treating physician with the intention of excluding dexmedetomidine. The target RASS was identical (-2 to +1).

Both groups received standard ACS management (hydration, pain control, antibiotics, transfusion per predefined criteria, oxygen/respiratory support)

OUTCOMES

Primary OutcomeSecondary Outcomes

Rate of death from any cause at 90 days after randomization.

180-day mortality

Institutional dependency at 180 days (e.g. transfer to a full-time nursing home or rehabilitation center)

Cognitive function at 180 days, evaluated using the Short IQCODE (Informant Questionnaire on Cognitive Decline in the Elderly) score

Patient-reported health-related quality of life at 180 days, assessed using the EQ-5D-3L questionnaire

Days free from coma/delirium at 28 days

Ventilator-free days at 28 days

Results

Baseline Characteristics

Baseline Characteristics

Critical Results

Strengths

Enrollment: 4,000 patients across 74 ICUs in 8 countries makes this study highly generalizable to diverse critical care settings and adds to external validity

Early Intervention Window: The median time from eligibility to randomization was fast (4.6 hours), accurately reflecting the initial resuscitation and sedation phase in the ICU.

Minimal Loss to Follow-Up: Only 2.4% of patients withdrew consent or were lost to follow-up, ensuring the integrity of the primary outcome data.

Relevance to real world practice: Clinically relevant question for practitioners in the ICU

Robust Randomization & Stratification Strategy: The trial utilized block randomization with variable block sizes implemented via a password-protected website, which effectively prevents clinicians from guessing the allocation sequence. Furthermore, they proactively stratified patients by study site and the presence of sepsis (a major independent driver of ICU mortality), ensuring the groups were highly balanced at baseline

Rigorous Statistical Analysis: The trial’s protocol and statistical analysis plan were published prior to completion, minimizing the risk of outcome reporting bias. The authors appropriately utilized a modified intention-to-treat (ITT) analysis to preserve the integrity of randomization, and they ran robust post hoc sensitivity analyses (including multiple imputations) to ensure that any missing data did not skew their primary outcomes

Limitations

Open-Label Design: The lack of blinding allowed treating clinicians to know the group assignments, which inherently introduces bias in subjective scoring (RASS, CAM-ICU) and thresholds for adding supplementary sedatives.

Intervention Contamination: Dexmedetomidine frequently failed as a standalone agent. Over 64% of the intervention group required propofol in the first two days, making this less of a “Dexmedetomidine vs. Usual Care” trial and more of a “Dexmedetomidine + Propofol vs. Propofol” trial.

Inappropriate Inclusion for the Drug: The trial did not exclude patients who had a clinical indication for deep sedation. Because dexmedetomidine is fundamentally a light sedative, these patients were essentially set up to fail and mandated crossover to propofol/midazolam.

Adverse Event Reporting: Adverse events were not systematically collected or independently adjudicated; they were reported by site investigators, which can lead to reporting bias (especially in an unblinded trial).

Usual Care used Dexmedetomidine: The study protocol allowed clinicians to administer dexmedetomidine as a “rescue” medication for uncontrolled agitation if standard conventional therapies failed. 11.5% of the patients assigned to the usual care group ended up receiving dexmedetomidine during the trial

Discussion

The SPICE III trial is a large study that helped to answer a common critical care question. For years, physiological rationale and smaller studies suggested dexmedetomidine could reduce brain dysfunction (delirium) and improve long-term outcomes. This trial plainly shows that reaching for dexmedetomidine first will have no impact on mortality. An identical 29.1% 90-day mortality is found in both groups.

The high crossover rate is the real challenge when using dexmedetomidine as a single agent in the study. Critically ill, newly intubated patients often require deep sedation for ventilator synchrony and stabilization. Because dexmedetomidine struggles to achieve deep sedation alone, 64.7% of patients received propofol anyway. This tells us that dexmedetomidine is a poor choice for the acute, phase post intubation when deeper sedation is required.

While dexmedetomidine failed as a standalone agent (with ~65% requiring rescue propofol), it did significantly lower the total doses of other sedatives. Patients in the dexmedetomidine group required roughly half the median daily dose of propofol (9.51 mg/kg vs. 17.9 mg/kg) and less midazolam (0.11 mg/kg vs. 0.31 mg/kg) than the usual care group. However, dexmedetomidine provided virtually no opioid-sparing effect—fentanyl was still required by 78.5% of the dexmedetomidine group compared to 80.7% of the usual care group. Ultimately, this reduction in GABA-ergic drug doses did not translate into a mortality benefit

While the dexmedetomidine group saw a 1-day median increase in ventilator-free and coma/delirium-free days, these secondary outcomes were not adjusted for multiple comparisons and should be interpreted with caution. More importantly, these minor early improvements did not translate into meaningful long-term patient-centered benefits—such as 180-day mortality, cognitive function, or institutional dependency—and are heavily outweighed by the significant increase in serious adverse events like bradycardia and asystole ; the NNH for a serious adverse event (like severe bradycardia or hypotension requiring intervention) was 43.

Prespecified subgroup analysis showed interesting age related data. Older patients (above the median age of 63.7 years) experienced slightly lower 90-day mortality with dexmedetomidine, whereas younger patients had higher mortality. While this requires confirmation in future trials, the authors hypothesize that age-related changes in how patients metabolize sedatives may play a critical role in drug safety and efficacy

This study reflects what I see clinically in both the Emergency Department and the ICU. As a sole agent for post intubation sedation it often fails and requires adjunctive medication or a switch to an alternate sedative.

If looked at from the alternate perspective however, given the “propofol sparing effect” when using demedetomidine, it fits perfectly into the weaning phase of sedation. As a patient’s clinical condition stabilizes and propofol dosing is being reduced, dexmedetomidine can be used as an adjunct as opposed to the primary agent.

Knowing that dexmedetomidine is generally more expensive than propofol and causes more hemodynamic compromise, its use should be focused and deliberate. It remains an excellent drug for facilitating extubation or treating agitated delirium, but it is not a universal, early-phase replacement for propofol or when deep sedation is needed

Author's Conclusion

“Among patients undergoing mechanical ventilation in the ICU, those who received early dexmedetomidine for sedation had a rate of death at 90 days similar to that in the usual-care group and required supplemental sedatives to achieve the prescribed level of sedation. More adverse events were reported in the dexmedetomidine group than in the usual-care group.”

Our Conclusion

The SPICE III trial demonstrates that using dexmedetomidine as the primary, early sedative in mechanically ventilated ICU patients does not improve mortality. While it offers a very modest reduction in days on the ventilator and days with delirium, it frequently fails as a standalone agent, requiring continued propofol supplementation. Furthermore, this practice carries a significant risk of hemodynamic adverse events like bradycardia. Dexmedetomidine should be utilized selectively based on patient-specific goals (e.g., transitioning to extubation) rather than as an indiscriminate first-line therapy.

Clinical Bottom Line

Dexmedetomidine as the primary early sedative in mechanically ventilated patients does not improve survival, often requires propofol supplementation, and significantly increases the risk of serious bradycardia and hypotension. Reserve dexmedetomidine for targeted clinical scenarios, such as facilitating weaning in patients with agitated delirium, rather than using it as a blanket first-line agent.

Guest Authors

Dorian Alexander

MD

Associate Professor of Emergency Medicine & Critical Care Medicine

UTHealth, Houston, TX

Meet The Team

References

Shehabi Y, et al.

Early sedation with dexmedetomidine in critically ill patients (SPICE III).

N Engl J Med.2019

PMID: 31112380

Hughes CG et al.

Dexmedetomidine or propofol for sedation in mechanically ventilated adults with sepsis (MENDS2).

N Engl J Med. 2021

PMID: 33528922

Post Peer Reviewed By: Mark Ramzy, DO (X/IG: @MRamzyDO) and Jad Dandashi, MD

Your Deep-Dive Starts Here

Rocuronium vs Succinylcholine for RSI: Awareness, Paralysis, and Post-Intubation Sedation

Does rocuronium delay post-intubation sedation and analgesia compared with succinylcholine after RSI? This ...

Procedures and Skills

Read More

REBEL Core Cast 142.0–Ventilators Part 2: Simplifying Mechanical Ventilation – Most Common Ventilator Modes

Mechanical ventilation can feel overwhelming, especially when faced with a sea of ventilator ...

Thoracic and Respiratory

Read More

REBEL Core Cast 141.0–Ventilators Part 1: Simplifying Mechanical Ventilation — Types of Breathes

For many medical residents, the ICU can feel like stepping into a pressure ...

Thoracic and Respiratory

Read More

The MENDS2 Trial: Dexmedetomidine vs Propofol for Sedation

Background: In patients requiring mechanical ventilation, sedative medications are used for patient comfort ...

Thoracic and Respiratory

Read More

Post Intubation Sedation for Pregnant Patients

You have just intubated a seizing eclamptic woman who is 34 weeks ...

Obstetrics and Gynecology

Read More

Showing Slide 1 of 5: Rocuronium vs Succinylcholine for RSI: Awareness, Paralysis, and Post-Intubation Sedation

The post SPICE III Trial: Does Early Dexmedetomidine Improve Survival in Mechanically Ventilated ICU Patients? appeared first on REBEL EM - Emergency Medicine Blog.