Extracorporeal membrane oxygenation in the therapy of cardiogenic shock (ECMO-CS): the need for a better definition of refractory cardiogenic shock.

Despite advances in the management of patients with myocardial infarction, including increased revascularization, cardiogenic shock (CS) remains the leading cause of death in these patients.1-3 As most attempts to reduce mortality failed to succeed in randomized controlled trials (RCTs), except for revascularization, there is a growing interest in mechanical circulatory support (MCS).3, 4 This craze for MCS is illustrated by a large increase in extracorporeal life support (https://www.elso.org/). It is unclear whether this increased utilization actually translates into improved outcomes in European cohorts.1, 2 However, temporal trends in the USA indicated increased use of MCS associated with decreased mortality.5, 6 Some other fragmental data add to improved outcomes when MCS is used in CS; however, there is no compelling evidence as RCTs are lacking.7 Those three expert centres that try to surmount the challenge of designing a RCT of extracorporeal membrane oxygenation in the therapy of cardiogenic shock (ECMO-CS) have to be commended.8 They judiciously decided to exclude obstructive shock (mainly pulmonary embolism), where ECMO definitely has a role in severe cases. However, high-risk pulmonary embolism carries a different prognosis and should be evaluated in another RCT. Moreover, whereas many patients with CS have experienced previous cardiac arrest, those who remain comatose bear a dismal prognosis and may be part of either another RCT or a pre-specified subgroup. Guidelines about percutaneous MCS call for ‘a phase III, multicentre, study comparing outcomes with and without (total or partial) MCS, with power to determine clinical outcome differences not only in short-term haemodynamics but also long-term survival’.9 However, they concede that ‘optimal timing of MCS insertion in CS remains unknown’. Thus, this study is up-to-date; however, several issues need to be underscored. First, CS may be the Cape Horn for clinical trials.4 Even if those three centres are large, well-known, CS referral centres, others failed to recruit such patients (e.g. NCT00314847: this study was stopped prematurely because of insufficient recruitment by 13 expert centres in France). Moreover, will the study be sufficiently powered? Is the 50% reduction of primary endpoint (death, resuscitated circulatory arrest, or implantation of another MCS) realistic? Clearly, for the first two endpoints, there would be little hope of 50% reduction. Such severe CS patients have an estimated mortality over 50%. Only chance or the Holy Grail may lead to decreased mortality of less than 25%. Thus, differences in the primary endpoint will probably be by implantation of other MCS devices that may obscure the interpretation of the results. Routine implantation of MCS is performed in patients receiving ECMO for CS in many ECMO centres, including intra-aortic balloon pump or Impella.10-12 Recently, Pappalardo et al. reported decreased mortality in patients receiving ECMO and Impella when compared with propensity-matched patients with ECMO only.13 Guidance for MCS in patients already receiving ECMO should be strictly defined and protocolized. On the other hand, MCS implantation including ECMO is allowed in the conservative arm ‘in case of shock progression, i.e. rapid deterioration of haemodynamic or metabolic status (defined as a need for increasing doses of vasopressor or a rise of serum lactate by 3 mmol/L within a minimum of 30 minutes after randomization)’.8 The deterioration is only vague and should be more strictly defined, in a way in which no one would dispute the imperative to implant ECMO unless letting the patient die. If epinephrine is allowed, how should the serum lactate level be handled? How much vasopressor (and/or inotropes) increase is considered significant, in how much time, leading to the qualification of rapid deterioration? As a way to assess cardiac reserve or catecholamine responsiveness, we developed a simple clinical score in CS:14 Inotropic score is calculated as follows: inotropic score (μg/kg/min) = dobutamine + dopamine + 100 × (noradrenaline + adrenaline) + 15 × (type-3 phosphodiesterase inhibitors) + 10 for levosimendan. As previously published, one could imagine a trial of MCS for patients with signs of tissue hypoxia (or arrhythmias) and CRASH score < 0.0375 (a way to reproducibly define refractory CS) in the intervention arm, vs. either no MCS (last resort conception), or very restrictive reproducible criteria (including CRASH score < 0.0300 for example).14 In their study, the Authors plan to include severely compromised patients8 (cardiac index <2.2 L/min/m2 or systolic blood pressure <100 mmHg, that may be evaluated at about cardiac power index <0.3 W/m2) who do not respond to moderate doses of catecholamine (inotropic score > 15 to >25). Thus they barely aim to maintain CRASH score ≥ 0.06. Further efforts are needed to better define refractory CS, and delineate the niche where MCS are undoubtedly helpful devices in the care of such severely ill patients. We await with great impatience the results of this study for paving the way to randomized evaluation of MCS in CS. Conflict of interest: none declared.