A2B试验：α-2激动剂（右美托咪定/可乐定）与丙泊酚用于ICU镇静的随机对照试验

原文图文 · 图 28 张

REBEL Rundown

Key Points

No Benefit: Neither dexmedetomidine nor clonidine reduced the time to successful extubation compared to standard propofol sedation.

Safety: Both alpha-2 agonists were associated with significantly higher rates of severe bradycardia and agitation compared to propofol.

Polypharmacy: Despite being randomized to alpha-2 agonists, patients in the intervention arms received propofol on ~76-77% of study days, highlighting the difficulty of using these agents as monotherapy.

Clonidine: This was the first major trial comparing clonidine directly to propofol; it showed no advantage over propofol and similar safety risks to dexmedetomidine

Introduction

Sedation is a cornerstone of ICU care, with propofol being the standard agent. However, heavy sedation with propofol or benzodiazepines is historically linked to delirium and prolonged mechanical ventilation. Alpha-2 agonists (dexmedetomidine and clonidine) have been championed as “propofol-sparing” agents that produce a calm, rousable state to facilitate faster extubation. Earlier trials set a highly promising stage for dexmedetomidine: the PRODEX and MIDEX trials demonstrated it was just as effective as propofol and midazolam for maintaining light-to-moderate sedation, while significantly improving a patient’s ability to communicate pain. Furthermore, the DahLIA trial proved dexmedetomidine was highly effective as a rescue therapy, significantly increasing ventilator-free hours and accelerating extubation in patients suffering from severe agitated delirium.

However, despite these promising targeted successes and the findings of the SPICE III trial, questions remained regarding the routine use of alpha-2 agonists as primary sedatives for all-comers. Moreover, clonidine—a vastly cheaper alternative with similar pharmacology—has critically lacked high-quality head-to-head data against propofol. The A2B trial sought to close this gap by directly comparing both dexmedetomidine and clonidine against standard propofol care.

Paper

Walsh TS, et al. Dexmedetomidine- or clonidine-based sedation compared with propofol in critically ill patients: the A2B randomized clinical trial. JAMA. 2025. • Correction Published: JAMA. 2026; PMID: 40388916

Previously Covered On REBEL

MENDS 2 Trial

SPICE III Trial

Post Intubation Sedation for pregnant patients

What They Did

CLINICAL QUESTION

Does primary sedation with dexmedetomidine or clonidine reduce the time to successful extubation in mechanically ventilated critically ill patients compared to propofol-based sedation?

STUDY DESIGN

Pragmatic, multicenter, open-label, randomized clinical trial (RCT). From Dec 2018 to Oct 2023, 1438 patients were randomized at 41 ICUs in the UK. 34 were excluded.

Randomization: 1:1:1 ratio into three groups (dexmedetomidine, clonidine, or propofol) using permuted blocks (sizes 3, 6, 9, 12), stratified by center, via a web-based system (allocation concealed).

Blinding: Unblinded (open-label) due to the distinct appearance and titration requirements of the drugs. Long-term outcomes (QoL, cognition) were assessed by blinded staff, but the primary outcome was ascertained by unblinded local teams.

Medication Tracking: The trial recorded the daily dosages for all intravenous sedative and opioid medications, as well as any rescue medications used to manage agitation or delirium

Incorporated Hierarchical Testing: To control for Type I error (false positives), they had to prove superiority against propofol before they were allowed to statistically compare clonidine vs. dexmedetomidine.

* Note: Since neither drug beat Propofol, the study could not formally evaluate if Clonidine is non-inferior to Dexmedetomidine.

Power Calculation: Originally 1737 patients; reduced to 1437 due to COVID-19. Remodeled to maintain 99% power for superiority testing of alpha-2 agonists vs. propofol (assumed effect: 2-day reduction in ventilation time).

POPULATION

Inclusion Criteria:

Adults (≥ 18 years of age)

Those receiving mechanical ventilation in the ICU

Currently sedated with propofol (with or without opioids)

Within 48 hours of commencing mechanical ventilation

Expected to require mechanical ventilation for at least another 24 hours (Total anticipated ventilation time ≥ 48 hours).

Exclusion Criteria:

Acute brain injury

Neuromuscular paralysis

Bradycardia (Heart rate <50 beats/min for ≥60 minutes)

Expected survival < 24 hours

Additional Exclusion Criteria:

Status epilepticus, Guillain-Barré syndrome, Myasthenia Gravis

Cardiac arrest (post-arrest), untreated heart block

Fulminant hepatic failure.

Pregnancy

Incarcerated individuals

Known allergy to study products

Patients receiving home ventilation

Previous enrollment in the trial

Transfer to ICU having already received mechanical ventilation for >6 hours (if outside the 48h inclusion window)

INTERVENTION & COMPARATOR

Intervention Group:

Group 1: Dexmedetomidine (n = 457)

Dosing: Starting dose of 0.7 μg/kg/h up to a maximum of 1.4 μg/kg/h.

Group 2: Clonidine (n = 476)

Dosing: Starting dose of 1.0 μg/kg/h up to a maximum of 2.0 μg/kg/h

* Note: Both titrated to a target Richmond Agitation-Sedation Scale (RASS) of -2 to +1. Propofol was permitted for rescue or if max doses were reached.

Comparator Group:

Group 3: Propofol-based sedation / Usual Care (n = 471)

Dosing: No specific dose guidance was provided. Clinicians were instructed to administer propofol according to standard local practice to maintain the target RASS score (-2 to +1)

OUTCOMES

Primary Outcome:

Time from randomization to successful extubation (defined as extubation followed by 48 hours of spontaneous breathing).

For patients receiving noninvasive ventilation, a 48-hour period of receiving no more than 5 cmH2O of continuous positive airway pressure was required to achieve the primary outcome

Secondary Outcomes:

All-cause mortality at 180 days

ICU length of stay

Time to first RASS score of −2 or greater (light sedation)

Time to the first day without agitation, unnecessary deep sedation, or pain behavior (defined as “overall optimum sedation”)

Rates of delirium or coma

Rates of delirium (when accessible without coma)

Safety Outcomes

Severe bradycardia (Heart rate <50 beats/min)

Cardiac arrhythmia

Cardiac arrest

Results:

PRE-DELIRIC Score: The median scores were between 72% and 74%, indicating a very high baseline probability for developing delirium across all groups.

Sedation Depth at Randomization: Approximately 62% of the overall study population (863 out of 1,403 patients) was already deeply sedated (RASS -4 or -5) prior to initiating the group-specific algorithms.

Critical Results

No Extubation Benefit (Primary Outcome): Replacing propofol with alpha-2 agonists did not significantly reduce ventilation time or how quickly patients left the ICU.

No Delirium Benefit: Despite the goal of “awake sedation,” neither dexmedetomidine nor clonidine reduced the rates of delirium or coma compared to standard care.

Increased Agitation: Both dexmedetomidine and clonidine resulted in significantly higher rates of agitation events compared to propofol alone.

Cardiovascular & Adverse Harms: Alpha-2 agonists caused an absolute risk increase of 13% for severe bradycardia. Additionally, dexmedetomidine specifically had the highest rate of serious adverse events (4.4%).

No Long-Term Benefit: At 180 days, there were no clinically important differences among the groups in mortality, cognitive function, psychological outcomes, or overall quality of life

Strengths

Robust Randomization: Used a centralized, concealed web-based system stratified by center, minimizing selection bias at enrollment.

Pragmatic Design: The study included a broad range of critically ill patients (medical, surgical, trauma) across 41 centers, increasing external validity.

High Power: Despite COVID-19 reductions, the study remained 99% powered to detect the assumed 2-day difference in extubation time.

Novel Comparison: This is one of the first large-scale RCTs to directly compare clonidine (a low-cost alternative) to propofol and dexmedetomidine.

Patient-Centered Outcome Design: Unlike many trials that rely solely on clinician-derived metrics, the study involved survivors of ICU stays in the choice of the primary outcome and the conduct of the trial, ensuring the endpoints were relevant to patients. The choice of the primary outcome (time to successful extubation) was clinically relevant to both physicians and patients

Statistical Handling of Mortality: The authors utilized a Fine and Gray proportional subdistribution hazards model for the primary analysis. This is a sophisticated statistical approach that correctly accounts for death as a “competing risk” (since patients who die cannot be successfully extubated), preventing the overestimation of the probability of extubation.

Systematic Safety Monitoring: This trial instituted daily recording of cardiovascular safety events. This rigorous methodology allowed for the clear identification of the significant bradycardia signal often missed or under-reported in prior literature.

Long-Term Follow-Up: The study did not stop at ICU discharge; blinded research staff collected long-term patient-centered outcomes post-hospitalization at 90 and 180 days including health-related quality of life (EQ-5D-5L), anxiety/depression (HADS), post-traumatic stress (IES-R), and cognitive function (T-MoCA), providing a comprehensive view of the drugs’ long-term impact

Standardized Scenario Guidance: The trial attempted to limit confounding variables by providing clinical guidance for managing cardiovascular instability and other specific scenarios (such as when to pause or decrease the dose of the alpha-2 agonists)

Additional Sedatives: The study explicitly discouraged the use of other sedatives—especially benzodiazepines—which is consistent with current international clinical practice guidelines

Transparency: The authors issued a correction regarding baseline variable inaccuracies in Table 1, ensuring the final data record was accurate. (results were corrected in Dec 2025)

Limitations

Open-Label Design: The lack of blinding is a major limitation. Clinicians knew which drug the patient was receiving, which inevitably influences titration decisions, the threshold for extubation, and the reporting of subjective adverse events like “agitation.”

High Cross-Over / Contamination: Propofol was allowed as a rescue. Consequently, patients in the Dexmedetomidine and Clonidine groups received Propofol on ~76% and 77% of study days, respectively. This effectively makes the trial a comparison of “Propofol vs. Propofol + Alpha-2 agonist” rather than pure monotherapy.

Rescue Polypharmacy: While the intervention arms were heavily contaminated with propofol (used on ~76-77% of study days), there was also minor cross-contamination in the control arm, with propofol patients receiving clonidine (~12% of the time) or dexmedetomidine (~5% of the time) as rescue for agitation. Additionally, a similar number of patients across all three arms (31-38%) received other rescue medications (such as antipsychotics and benzodiazepines). This polypharmacy muddies the ability to isolate the true effects of the primary sedatives.

Subjective Outcomes: The primary outcome (time to extubation), while patient centered, relies on clinician decision-making, which can be biased in an open-label trial.

Lack of Weaning Standardization: While guidance was provided, the weaning of mechanical ventilation and sedation discontinuation were not tightly protocolized. In an open-label trial, this lack of strict control introduces significant practice variation and allows clinician bias to influence the “time to extubation” endpoint.

Delayed Randomization: The median time from intubation to randomization was 21.0 hours. During this pre-enrollment period, nearly 100% of patients received propofol, and 62% were already deeply sedated (RASS -4/-5) at the time of randomization.

Lack of Comparator Standardization: The study explicitly states that “no specific dose guidance was given for propofol-based sedation” in the control group. The choice and dosing of opioid analgesia were allowed across all three groups at the complete discretion of the treating clinical team. Without a strict protocol, “usual care” becomes a “black box” of clinical practice variation across 41 centers, effectively meaning the intervention was compared against a moving target of propofol dosing.

Randomization and Operational Errors: Likely exacerbated by pandemic-related staffing pressures, the study faced operational lapses, including one patient being randomized twice and four others randomized in error. While these patients were excluded from the analysis, these errors highlight the strain on the research infrastructure during the trial.

Operational Impact of COVID-19: The pandemic caused a pause in recruitment and had a “major effect on staff numbers and experience.” The authors note that during this period, nursing staff may have lacked confidence administering alpha-2 agonists alone, potentially driving the high use of rescue propofol.

Clinician Inexperience: The authors suggest that the unexpectedly high rates of agitation in the intervention groups might reflect “less clinician experience using dexmedetomidine and clonidine for primary sedation.” This implies the results might differ in centers with established alpha-2 agonist protocols.

Limited External Validity: The trial was conducted exclusively across 41 ICUs in the UK (United Kingdom), a single country with a socialized healthcare system. This limits the generalizability of the findings and makes it difficult to replicate the robust data extraction portion of this study in places like the US, where healthcare systems and electronic medical records (EMRs) are highly fractionated. Additionally, the study utilized intravenous (IV) clonidine—a formulation that is not typically used or readily available in US (United States) ICUs—further restricting the real-world applicability of the clonidine arm’s findings for US clinicians.

Discussion

Efficacy: The A2B trial effectively closes the door on the concept of using alpha-2 agonists (dexmedetomidine or clonidine) as monotherapy replacements for propofol in the general ICU population. Despite the theoretical benefits of “awake sedation,” this study found no difference in the time to extubation.

Randomization Delay: A major unseen driver of these neutral results is likely the median 21-hour delay to randomization. By the time patients were enrolled, nearly 100% had already received propofol, and 62% were deeply sedated. This created a massive “carryover effect.” Propofol accumulates in lipid tissues; therefore, even if you switch to an alpha-2 agonist immediately at hour 21, the patient still has to “wash out” the previous day’s propofol. This potentially diluted any benefit of the intervention, biasing the study toward the null.

Illness Severity as a Confounder: The patient population was highly morbid with a mean APACHE II score of ~20, 66% presenting with sepsis, and over 75% having severe cardiovascular dysfunction. This raises a major confounding variable: if a patient is that sick and requires deep sedation early in their ICU course, perhaps the severity of the underlying pathology is a far more potent driver of extubation time than the specific sedative agent chosen. In these highly morbid states, delirium and prolonged ventilation are likely inevitable consequences of the disease process itself, meaning the choice of a “better” sedative simply cannot overcome the trajectory of the critical illness.

Lack of Control Standardization: The study design suffered from a lack of strict controls in the “Usual Care” arm. The methods explicitly state that “no specific dose guidance was given for propofol-based sedation”. Without a standardized protocol, the control group becomes subject to clinical practice variation. This introduces significant noise, as it is unclear if the “Usual Care” provided was optimized or simply reflected heterogeneous local habits, potentially biasing the results toward the null.

Propofol Contamination: Perhaps the most critical point regarding the intervention is that it was not a true test of monotherapy. Despite being randomized to alpha-2 agonists, patients in the dexmedetomidine group received propofol on 77% of study days, and those in the clonidine group received it on 76% of study days. The authors note that propofol doses in these groups were lower (about 25-30% of the control group dose). Consequently, this wasn’t a trial of “Alpha-2 Agonists vs. Propofol”; it was functionally a trial of “Propofol vs. Propofol + Alpha-2 Agonists”. Since the combination arm failed to show benefit but significantly increased adverse events, adding these agents to background propofol offers no advantage – other than maybe using a lower dose of propofol.

Confounded Delirium Outcomes: A major question left unanswered by this trial is whether the delirium outcomes would have been different if propofol had been strictly prohibited in the alpha-2 agonist groups. Because the study design lacked standardized propofol dosing and allowed clinicians to use it “according to local practice guidelines,” propofol remained heavily present across all three treatment arms. Therefore, it is impossible to know if dexmedetomidine and clonidine truly failed to reduce delirium, or if the continued, unregulated exposure to propofol in the intervention arms simply “washed out” any potential neurocognitive benefit. If the protocol had forced strict alpha-2 monotherapy and only allowed non-GABAergic rescue medications (like haloperidol or quetiapine), the delirium and agitation data might have looked different.

Safety: The safety data is an actionable takeaway. Both alpha-2 agonists caused a ~60% relative increase in severe bradycardia. Furthermore, the paradoxical increase in agitation suggests that in actual practice these agents may be difficult to titrate effectively as primary sedatives, leading to “under-sedated” and agitated patients. Agitation was common among all age groups but younger patients suffered from more cardiovascular and bradycardic events. This aligns with the findings from the SPICE III trial, which also identified age-related heterogeneity in outcomes with dexmedetomidine.

Bradycardia: While both dexmedetomidine and clonidine significantly increased the risk of severe bradycardia (HR <50), it is worth questioning the clinical relevance of this surrogate safety marker. In critical care, asymptomatic bradycardia with an adequate mean arterial pressure (MAP) often requires no intervention. Supporting this, the trial found that while bradycardia spiked, the rates of severe hypotension were actually similar across all three groups. However, this numerical “adverse effect” had a real-world impact on clinician behavior. The authors suggest that seeing the heart rate drop likely intimidated clinicians into abandoning alpha-2 uptitration, prompting them to rely heavily on propofol rescue instead. Therefore, the bradycardia may not have caused hemodynamic disaster, but it directly contributed to the failure of the monotherapy strategy.

Cost & Applicability: Clonidine is significantly cheaper than dexmedetomidine and propofol. However, “cheap” is irrelevant if the drug is ineffective and unsafe. Since clonidine showed no benefit in extubation time and carried the same bradycardia risks as dexmedetomidine, there is no economic or clinical argument to support its widespread adoption as a primary sedative.

Off-Label Duration & Clinical Reality: The manufacturer labeling for dexmedetomidine recommends use for ≤ 24 hours. However, in this trial, the median duration of use was 4 days. While this reflects common “off-label” ICU practice, this prolonged exposure likely drove the significant safety signal. Both alpha-2 agonists caused a ~60% relative increase in severe bradycardia. This data serves as a stark reminder that while prolonged use is common, it carries a predictable and significant hemodynamic cost without delivering the promised benefit of earlier extubation.

Secondary Outcomes: Proponents of alpha-2 agonists often look at secondary benefits, arguing that these agents reduce delirium or preserve long-term cognition. In the A2B trial there was no difference in delirium or coma rates compared to propofol, challenging the dogma derived from previous trials like SPICE III. Uniquely, this trial maintained robust follow-up for 180 days and found zero difference in Health-Related Quality of Life (EQ-5D-5L), Anxiety/Depression (HADS), PTSD scores (IES-R), or Cognitive function (T-MoCA). If the drug doesn’t reduce ventilator time, doesn’t prevent delirium, and doesn’t improve long-term cognition, the justification for its routine use—especially given the increased risk of bradycardia—disappears.

Clonidine vs. Dexmedetomidine: If we look at the survival curves—which are nearly identical—and ask, “Can I just use the cheaper clonidine instead of the expensive dexmedetomidine?” Unfortunately, the study’s statistical design prevents a formal answer. The authors used a “gatekeeper” hierarchy: they were only allowed to test if clonidine was non-inferior to dexmedetomidine (Stage 2) if one of the drugs first beat propofol (Stage 1). Because neither drug beat propofol, the “gate” to Stage 2 never opened. Therefore, while the drugs appear clinically indistinguishable in this trial, we lack the formal statistical confirmation of non-inferiority.

In Clinical Practice: A critical interpretation of this negative trial is that it attempted to use “weaning phase” agents for “acute phase” pathology. In my clinical practice, dexmedetomidine is often most effective as a weaning tool. Providing light sedation/anxiolysis — smoothing the transition to extubation, once the patient has recovered from their underlying metabolic or septic insult. However, this study enrolled patients early (median 21 hours), with high acuity (APACHE II ~20, 66% Sepsis). In this acute phase of “toxic metabolic encephalopathy,” the brain often requires the reliable suppression of a GABA-ergic agent (propofol). Attempting to use a lighter, adjunct sedative like an alpha-2 agonist during this high-adrenergic acute phase likely resulted in the observed failure of monotherapy (necessitating propofol rescue 77% of the time) and the paradoxical increase in agitation. The trial essentially asked the right drug to do the wrong job at the wrong time.

This study won’t change my practice in the use of propofol or alpha 2 agents. It does however bring to light the off label duration usage of alpha 2 agents and their side effects especially in at risk populations.

Author's Conclusion

“In critically ill patients, neither dexmedetomidine nor clonidine was superior to propofol in reducing time to successful extubation.”

Our Conclusion

The A2B trial provides compelling evidence that replacing propofol with alpha-2 agonists (dexmedetomidine or clonidine) as the primary sedative strategy does not accelerate extubation in general ICU patients. Furthermore, this strategy introduces significant risks of bradycardia (with increasing doses) and agitation (if patients are undersedated) . The high rate of propofol use in the intervention arms suggests that alpha-2 agonists are better viewed as adjuncts or transitional rather than standalone replacements for propofol.

Clinical Bottom Line

Propofol should remain the standard of care for primary sedation in mechanically ventilated ICU patients. There is no evidence to support switching to dexmedetomidine or clonidine to shorten ventilation time, and doing so increases the risk of severe bradycardia.

References

Walsh TS, et al.

Dexmedetomidine- or clonidine-based sedation compared with propofol in critically ill patients: the A2B randomized clinical trial. JAMA. 2025

PMID: 40388916

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

Jakob SM, et al.

Dexmedetomidine vs Midazolam or Propofol for Sedation During Prolonged Mechanical Ventilation: Two Randomized Controlled Trials (MIDEX & PRODEX trials). JAMA. 2012

PMID: 22436955

Reade MC, et al.

Effect of Dexmedetomidine Added to Standard Care on Ventilator-Free Time in Patients With Agitated Delirium: A Randomized Clinical Trial. The Dexmedetomidine to Lessen ICU Agitation (DahLIA) study JAMA. 2016

PMID: 26975647

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

Meet the Author

Dorian Alexander

MD

Associate Professor of Emergency Medicine & Critical Care Medicine

UTHealth, Houston, TX

Meet The Team

Your Deep-Dive Starts Here

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

The post The A2B Trial: Are Alpha-2 Agonists Better than Propofol for ICU Sedation? appeared first on REBEL EM - Emergency Medicine Blog.