ECLS-SHOCK Trial
Details
ECLS-SHOCK randomised patients with acute myocardial infarction and cardiogenic shock undergoing planned early revascularisation to early extracorporeal life support (VA-ECMO) plus usual medical treatment versus medical treatment alone. It is the largest randomised test of early VA-ECMO in AMI-CS and is negative on mortality. (sources/cardiogenic-shock-acc-2025, rating: very high) (sources/cardiogenic-shock-nejm-2026, rating: very high)
Key Facts
- n = 420, multicentre, randomised, open-label; AMI-CS with planned early revascularisation. (sources/cardiogenic-shock-acc-2025, rating: very high)
- 30-day mortality 47.8% with ECMO vs 49.0% control — RR 0.98 (95% CI 0.80–1.19; P = 0.81). (sources/cardiogenic-shock-nejm-2026, rating: very high)
- No difference at 1 year. (sources/cardiogenic-shock-nejm-2026, rating: very high)
- A 4-trial individual-patient-data meta-analysis of VA-ECMO in AMI-CS found no mortality benefit and consistently higher complication rates. (sources/cardiogenic-shock-nejm-2026, rating: very high)
- Mechanistic concern: retrograde aortic flow from VA-ECMO increases LV afterload (the "Windkessel effect"), so the device supports systemic perfusion without unloading the failing LV. (sources/cardiogenic-shock-nejm-2026, rating: very high)
Contradictions / Open Questions
- Enrolment of a substantial cardiac-arrest phenotype is the authors' proposed explanation for the neutral result — arrest-related hypoxic brain injury may dilute a benefit in shock without brain injury. This is plausible but not established, and it cannot be repaired post hoc. (sources/cardiogenic-shock-nejm-2026, rating: very high)
- Unloading strategies remain unproven. Adding a microaxial pump or IABP to VA-ECMO has observational support, but a randomised trial of a transseptal LA cannula versus VA-ECMO alone showed no mortality effect. Which patients, if any, benefit from unloading is unresolved. (sources/cardiogenic-shock-nejm-2026, rating: very high)
- The negative trial and positive observational signals cannot both be right; the discrepancy is usually attributed to confounding by indication in registry data. (sources/cardiogenic-shock-nejm-2026, rating: very high)
Connections
- Related to concepts/Cardiogenic-Shock — AMI-CS phenotype and management
- Related to concepts/Cardiogenic-Shock — VA-ECMO position and LV unloading
- Related to entities/IABP-SHOCK-II-Trial and entities/DanGer-Shock-Trial — the other pivotal tMCS trials
- Related to entities/Acute-Coronary-Syndrome — the enrolled population