New England Journal of Medicine AUGUST 26, 2023

Extracorporeal Life Support in Infarct-Related Cardiogenic Shock

Holger Thiele, Uwe Zeymer, Ibrahim Akin, et al., on behalf of the ECLS-SHOCK Investigators

Bottom Line

The ECLS-SHOCK trial demonstrated that early implementation of venoarterial extracorporeal membrane oxygenation (VA-ECMO) in patients with acute myocardial infarction complicated by cardiogenic shock did not reduce 30-day all-cause mortality compared to standard medical therapy, while significantly increasing the risk of major bleeding and vascular complications.

Key Findings

1. The primary endpoint of all-cause death at 30 days occurred in 100 of 209 patients (47.8%) in the ECLS group compared with 102 of 208 patients (49.0%) in the control group (relative risk, 0.98; 95% confidence interval [CI], 0.80 to 1.19; P=0.81).
2. Moderate or severe bleeding occurred significantly more frequently in the ECLS group (23.4%) compared to the control group (9.6%; relative risk, 2.44; 95% CI, 1.50 to 3.95).
3. Peripheral vascular complications requiring intervention were more common in the ECLS group (11.0%) than in the control group (3.8%; relative risk, 2.86; 95% CI, 1.31 to 6.25).
4. Secondary outcomes, including time to hemodynamic stabilization, need for renal replacement therapy, and duration of intensive care unit stay, did not differ significantly between the two groups.

Study Design

Design
RCT
Open-Label
Sample
420
Patients
Duration
30 days
Median
Setting
Multicenter, Germany/Slovenia
Population Patients with acute myocardial infarction complicated by cardiogenic shock scheduled for early revascularization
Intervention Early implantation of venoarterial extracorporeal membrane oxygenation (VA-ECMO) plus usual medical treatment
Comparator Usual medical treatment alone
Outcome All-cause death at 30 days

Study Limitations

The trial was open-label, which may introduce observer bias, although the primary outcome of all-cause death is objective.
A crossover rate of approximately 12% from the control group to the ECLS arm potentially reduced the separation between treatment groups and may have biased the study toward a null result.
The high prevalence of pre-randomization cardiac arrest (approx. 77% in both groups) indicates a very high-risk population, which may limit the generalizability to patients with less severe shock profiles.
The study design did not mandate a standardized protocol for neuroprognostication, potentially affecting the interpretation of neurological outcomes.

Clinical Significance

The results challenge the routine, unselected use of early VA-ECMO in patients with acute myocardial infarction and cardiogenic shock. Given the lack of survival benefit and the higher rate of significant bleeding and vascular complications, the study suggests that routine prophylactic ECMO should be avoided in this clinical setting, supporting a more conservative, selective approach to mechanical circulatory support.

Historical Context

Following the IABP-SHOCK II trial, which failed to show a mortality benefit for intra-aortic balloon pumps (IABP) in cardiogenic shock, clinical practice shifted rapidly toward the use of more potent mechanical circulatory support devices like VA-ECMO. This shift was largely driven by observational evidence despite a lack of robust randomized controlled trial data, leading to a decade of increasing device utilization that the ECLS-SHOCK trial sought to test rigorously.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

How does the retrograde flow of venoarterial extracorporeal membrane oxygenation (VA-ECMO) potentially increase myocardial oxygen demand in a patient with an acute myocardial infarction?

Key Response

VA-ECMO pumps blood from the venous system into the arterial system (usually the femoral artery) in a retrograde fashion toward the heart. This increases the afterload on the left ventricle (LV). In the setting of an acute MI, increased LV wall stress and afterload can increase myocardial oxygen consumption, potentially exacerbating ischemia if the LV is not properly 'unloaded' or 'vented.'

Resident
Resident

Given the results of the ECLS-SHOCK trial, what are the primary safety concerns you must monitor for if a patient is placed on VA-ECMO for cardiogenic shock?

Key Response

The trial demonstrated a significantly higher risk of complications in the ECLS group. Specifically, clinicians must monitor for major bleeding (RR 2.44) and peripheral vascular complications requiring intervention (RR 2.86), such as limb ischemia, which occurred much more frequently in the ECMO arm compared to standard medical therapy.

Fellow
Fellow

In the ECLS-SHOCK trial, the median time from randomization to ECLS initiation was 2 hours. How does this timing, combined with the lack of mandatory LV venting, influence your interpretation of the trial's neutral mortality benefit?

Key Response

Some experts argue that the benefit of ECMO is highly time-dependent and that 2 hours may be too late to prevent multi-organ failure. Furthermore, without mandatory LV venting (e.g., using an Impella or atrial septostomy), the increased afterload from ECMO may have neutralized any benefit from improved systemic perfusion by causing LV distension and pulmonary edema.

Attending
Attending

How should the ECLS-SHOCK findings change your 'bridge-to-decision' conversation with the families of patients presenting with SCAI Stage C or D infarct-related shock?

Key Response

The trial provides high-quality evidence that routine early ECLS does not improve 30-day survival and carries a high risk of iatrogenic harm. The conversation should shift from viewing ECMO as a standard-of-care 'rescue' to a highly selective intervention, emphasizing that standard medical therapy (vasopressors/inotropes) remains the baseline and that ECMO's role is likely limited to specific salvage scenarios.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The ECLS-SHOCK trial was powered to detect an absolute risk reduction (ARR) of 12% in mortality. Was this trial potentially underpowered for a more modest but clinically significant effect, and how did the 15% crossover rate from the control arm impact the intention-to-treat analysis?

Key Response

A 12% ARR is quite large for a critical care intervention. If the true benefit was smaller (e.g., 5%), the trial would be underpowered. Additionally, the crossover of 15.4% of control patients to ECLS (mostly due to refractory shock) potentially dilutes the treatment effect in an intention-to-treat analysis, making the two groups appear more similar than they might have been with strict protocol adherence.

Journal Editor
Journal Editor

As a reviewer, how would you address the lack of blinding in ECLS-SHOCK and its impact on the reporting of secondary endpoints like 'time to hemodynamic stability' or 'bleeding'?

Key Response

While blinding the intervention (ECMO) is impossible, the lack of blinding can introduce ascertainment bias. For example, clinicians might monitor for bleeding or limb ischemia more aggressively in the ECMO group than in the control group. A rigorous review would look for whether the adjudication of these events was performed by a committee blinded to the treatment assignment to mitigate this bias.

Guideline Committee
Guideline Committee

Current ESC and AHA/ACC guidelines often provide a Class IIa or IIb recommendation for VA-ECMO in refractory cardiogenic shock. Based on the ECLS-SHOCK results, should these guidelines be updated to a Class III (Harm) for 'routine' early use in MI-related shock?

Key Response

The trial suggests that 'routine' early implementation provides no benefit and increases harm (bleeding/vascular issues). While not necessarily Class III for all shock (as it remains a vital bridge for some), the guidelines should likely be updated to reflect that routine use is not supported and should perhaps be downgraded or more strictly reserved for refractory cases that fail standard medical management, moving away from the 'early implementation' strategy tested here.

Clinical Landscape

Noteworthy Related Trials

1999

SHOCK Trial

n = 302 · NEJM

Tested

Emergency revascularization

Population

Patients with cardiogenic shock complicating acute myocardial infarction

Comparator

Initial medical stabilization

Endpoint

All-cause mortality at 30 days

Key result: Early revascularization did not significantly improve 30-day mortality compared with medical stabilization, though a survival benefit was observed at 6 and 12 months.
2012

IABP-SHOCK II Trial

n = 600 · NEJM

Tested

Intra-aortic balloon pump

Population

Patients with cardiogenic shock complicating acute myocardial infarction

Comparator

Standard medical therapy

Endpoint

30-day all-cause mortality

Key result: The use of an intra-aortic balloon pump did not reduce 30-day mortality in patients with cardiogenic shock complicating acute myocardial infarction compared with medical therapy alone.
2017

CULPRIT-SHOCK Trial

n = 696 · NEJM

Tested

Culprit-lesion-only PCI

Population

Patients with multivessel coronary artery disease and acute myocardial infarction complicated by cardiogenic shock

Comparator

Immediate multivessel PCI

Endpoint

Composite of death or severe renal failure at 30 days

Key result: Culprit-lesion-only PCI resulted in a lower risk of the composite primary endpoint of death or renal replacement therapy compared with immediate multivessel PCI.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis