Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is increasingly being used worldwide to treat severe cardiogenic shock. The survival rate of these patients has increased in the last decade, reaching 45-50% for patients with acute myocardial infarction (AMI), the most frequent indication of the technique, 50-60% for patients with end-stage dilated cardiomyopathy bridged to heart transplantation or long-term left ventricle assist device (LVAD) surgery, 60-70% for fulminant myocarditis, while it remains lower for post-cardiotomy cardiogenic shock (25-35%) and after cardiac arrest (20-40%). However, peripherally inserted VA-ECMO increases left ventricular (LV) afterload, that may lead to poorer clinical outcomes by fostering left ventricular distension, blood stagnation, aortic valve closure, all of which increasing pulmonary congestion and the need for mechanical ventilation and compromising myocardial recovery whenever it is possible, or delaying a bridge to a heart transplantation or long-term left ventricle assist device (LVAD) surgery for patients with end-stage cardiac dysfunction. Several methods have been proposed to reduce afterload after VA-ECMO, including the use of an intra-aortic balloon pump (IABP), balloon atrial septostomy, transseptal left atrial cannula insertion, and use of the left-sided Impella device (Abiomed, Danvers, MA, USA). The clinical benefits of left ventricular unloading have been suggested by many retrospective case-control studies, including a study by our group that showed that associating an IABP with peripheral VA-ECMO was independently associated with a lower frequency of hydrostatic pulmonary edema under ECMO and more days off mechanical ventilation. More recently, unloading the left ventricle with an IABP was associated with the best survival rate and security profile as compared to no unloading or unloading with a microaxial pump in 12,734 VA-ECMO patients included in the Extracorporeal Life Support Organization registry. It should also be mentioned that another large registry study showed that the greatest benefit of LV unloading under ECMO was observed with early versus delayed insertion of the unloading device. Lastly, the EARLY-UNLOAD randomized trial in which a transseptal left atrial cannula was used for LV unloading yielded negative results. However, it is important to note that 50% of control patients were rapidly transitioned to LV unloading, thereby compromising the opportunity to demonstrate a mortality benefit. It was also underpowered for the primary outcome of D30 mortality since it included only 116 patients As a result, the recourse to systematic early LV unloading remains highly heterogeneous in clinical practice. For example, , while IABP was EULODIA - Protocol, version 1.0 dated 24/01/2025 Page 6 sur 54 This document is the property of DRCI/AP-HP. All reproduction is strictly prohibited.

Version no. 4.0 of 31/05/2019 associated to ECMO in >70% of the cases in our series of AMI CS patients, only 5.8% of the patients included in the ECMO arm of the recent ECLS-Shock trial received an unloading device, which may have contributed to the neutral result of the study and the only randomized trial to date was underpowered and flawed by a very high rate of early cross-over. Indeed, there is large heterogeneity in current clinical practice, where decisions on whether to add an additional mechanical unloading device during VA-ECMO support vary widely.

Therefore, a new and adequately powered trial comparing systematic early left ventricular unloading to a conventional approach, with rescue left ventricular unloading only in case of clear and urgent indication, i.e. if overt hydrostatic cardiogenic pulmonary edema occurs, is urgently needed. The EULODIA trial is designed to test the hypothesis that early preventive left ventricle unloading with an IABP improves clinical outcomes as compared to conventional care with delayed curative unloading in patients under VA-ECMO for refractory cardiogenic shock.

Inclusion Criteria:

1. Patient on peripheral femoro-femoral VA ECMO for severe Cardiogenic shock for ≤24h
2. Initiation of LV unloading possible within 12 hours after randomization.
3. Consent obtained from a close relative or surrogate. Should such a person be absent, eligible patients will be randomized according to the specifications of emergency consent and the patient will be asked to give his/her consent for the continuation of the trial when his/her condition will allow.
4. Social security registration (AME excluded)

Exclusion Criteria:

1. Age <18 years
2. Pregnancy
3. Onset of VA-ECMO >24 h
4. Overt pulmonary edema despite optimization of patient's management (titration of ECMO blood flow, inotropes, non-invasive mechanical ventilation with PEEP, diuretics or hemofiltration) requiring urgent LV unloading
5. ECMO for massive pulmonary embolism or primary RV failure
6. ECMO after heart transplant
7. ECMO after LVAD surgery
8. Resuscitation >30 minutes before ECMO (cumulative low-flow time), except if the patient has fully recover consciousness at the time of randomization.
9. ECMO for refractory cardiac arrest (E-CPR)
10. Grade 3-4 aortic regurgitation
11. Mechanical complication of acute myocardial infarction (massive mitral regurgitation, pericardium drainage required, septal ventricular defect)
12. Patient moribund on the day of randomization
13. Cerebral deficit with fixed dilated pupils or Irreversible neurological pathology
14. Other severe concomitant disease with limited life expectancy <1 year
15. Patient has a durable ventricular assist device, an IABP or another temporary mechanical circulatory support (other than ECMO) prior to enrollment.
16. Severe peripheral artery disease or previous aortic or ilio-femoral surgery precluding IABP or IMPELLA insertion.

The procedures carried out with the French data privacy authority (CNIL, Commission nationale de l'informatique et des libertés) do not provide for the transmission of the database, nor do the information and consent documents signed by the patients.

Consultation by the editorial board or interested researchers of individual participant data that underlie the results reported in the article after deidentification may nevertheless be considered, subject to prior determination of the terms and conditions of such consultation and in respect for compliance with the applicable regulations.

Beginning 3 months and ending 3 years following article publication. Requests out of these time frame can also be submitted to the sponsor

Researchers who provide a methodologically sound proposal