Acute myocardial ischemia complicated with cardiogenic shock is the leading cause for circulatory assistance. It has to date never been evaluated in randomized studies, as this condition is associated with a rapid fatal outcome. However, ECMO-assisted percutaneous coronarography was recently evaluated in a retrospective before/after study in 58 patients presenting with cardiogenic shock due to acute myocardial ischemia from 2004 to 2009 in Taiwan. Mortality at 1 year tumbled down from 76 to 37% after ECMO implementation, while patients remained unchanged regarding to demographic characteristics and disease severity [7]. This further challenges the timing for ECMO implantation in those patients. Percutaneous coronary angioplasty remains the cornerstone of the treatment in such patients, but some with profound cardiogenic shock will necessitate initiation of the ECMO first in the catheterization laboratory.

The second challenge in those patients is to predict the potential of myocardial recovery under ECMO to guide further clinical strategy. Particularly, patients with a poor potential of recovery should be rapidly switched to prolonged assistance devices such as left ventricular assist device (LVAD) or to cardiac transplantation to avoid ECMO complications.

Outcome after ECMO implantation for myocardial ischemia was recently studied in 77 patients. ECLS duration was 9.8 ± 7.1 days. Nineteen patients (24%) were finally weaned from ECMO; 40 (52%) died under ECMO; 5 (6.5%) were transplanted; 9 (11.6%) were switched to LVAD therapy; and 4 (5.2%) to biventricular mechanical assistance. Thirty-day and in-hospital survival rates were respectively 38.9% and 33.8% in this cohort. Multivariable analysis identified preimplantation serum lactate level, preimplantation serum creatinine level, and previous cardiopulmonary resuscitation as independent predictors of 30-day mortality [8].

Refractory cardiogenic shock following cardiac surgery was historically the main area of development for ECMO assistance. It concerns from 0.5 to 2.9% of cardiac procedures with cardiopulmonary bypass. The rational for ECMO implantation is the potential of recovery from myocardial stunning after surgery. However, results appear quite disappointing, mainly due to age, previous medical condition, and previous cardiac damages of operated patients. In larger cohorts, patients referred for postcardiotomy ECMO had a mean age around 64 years, a mean euroscore around 21%, and a LVEF around 46% [9, 10]. More than half of the patients could be weaned from the ECMO, but only 24–33% were discharged home, and survival at 1 year varied from 17 to 29%. Age >70, diabetes, obesity, preoperative renal insufficiency, preoperative LVEF, and preimplantation acidosis were independently associated with a poor outcome, while isolated coronary artery bypass grafting appeared to be protective. Interestingly, neither cardiopulmonary bypass duration nor aortic clamping time seems to be associated with the outcome.

Primary graft failure is a frequent complication after heart transplantation, ranging from 4 to 24%, mostly depending on the local politics on marginal heart allografts. Several studies reported the successful use of transient mechanical support with an ECMO in this indication. Explantation rates varied from 60 to 80% and long-term survival from 50 to 82%. Interestingly, cumulative survival did not differ in patients who survived the ECMO period from non-ECMO patients [11]. Again, results appear far better when ECMO is implanted early during the time course of the disease, with a survival rate of only 14% when used as a salvage therapy [12].

Myocarditis is a disease that may progress rapidly to refractory cardiogenic shock and death. Considering the prompt myocardial recovery in most of the patients and its easy and rapid implantation and explantation, peripheral venoarterial ECMO has become the elective first-line assistance device for those patients.

Several large cohorts reported the favorable outcome of this otherwise fatal condition using ECMO support. In a cohort of 41 fulminant myocarditis with refractory cardiogenic shock patients assisted with VA-ECMO, survival to discharge was 70% in the ECMO group [13]. This high survival rate contrasted with the high severity of the patients before ECMO implantation, reflected by a mean SAPS-II score at 56. The median duration of assistance was quite short, 10 days, highlighting the rapid myocardial recovery in these patients. Mean LVEF was 57% at 18 months. Four patients who did not recover needed a heart transplantation and were alive at discharge. Patients needed however a high degree of healthcare resources: 88% required mechanical ventilation, 54% dialysis, and the mean hospitalization duration was 59 days. Importantly, 63% of the patients exhibited at least one major complication related to the ECMO. Ten developed hydrostatic pulmonary edema and necessitated a switch for a central assistance. Other complications comprised major bleeding at the cannulation site (46%), deep vein thrombosis (15%), arterial ischemia (15%), surgical wound infection (15%), and stroke (10%). After a mean 18 months of follow-up, patients reported a highly preserved mental health and vitality. However, they still reported physical and psychosocial difficulties, and anxiety, depression, and/or post-traumatic stress disorder symptoms were present in respectively 38, 27, and 27% of them. Ten patients presented also long-term paresthesia or neurological defect in the leg of the ECMO, and one necessitated a major amputation due to arterial ischemia. In this cohort, SAPS-II >56 and troponin >12 μg/L were the only independent predictors of poor outcome.

In another cohort of 75 pediatric and adult patients with myocarditis complicated with a refractory cardiogenic shock, PVA ECMO as first-line therapy gave comparable results. Sixty-four percent of the patients could be discharged home with a mean LVEF of 57%. Nine patients did not recover and were switched to long-term ventricular assist device (six patients) and heart transplantation (three patients). Thirty percent necessitated left ventricule drainage for refractory pulmonary edema under VAP ECMO. This condition was associated with a poorer weaning rate from the ECMO (39%) and a poorer survival (48%). Again, dialysis and a persistent elevation of troponin levels were independent predictors of a poor outcome [14].

PVA-ECMO is routinely used in daily practice during refractory cardiogenic shock following drug intoxication, and is recommended during cardiac arrest in this condition (grade IIb, level of evidence C) [15].

Interest for ECMO assistance during drug poisoning comes from the reversibility of the cardiac dysfunction observed. Experimental studies demonstrated a clear benefit of the technique in several models, and many single case studies reported its successful use in humans [16]. The largest cohort on that subject concerned 62 patients in a single center, mean age 48, presenting a refractory cardiogenic shock following drug intoxication [17]. Ten of them deteriorated to a refractory cardiac arrest, and fourteen of the patients were implanted with a PVA ECMO, three during cardiopulmonary resuscitation. Survival was strongly increased in ECMO-implanted patients (86% vs. 48%, p = 0.02). Particularly, none of the refractory cardiac arrest patients survived without ECMO, whereas all of the three ECMO-implanted patients during this condition survived. It was also noticed that none of the ECMO-implanted patients died during intoxication with membrane-stabilizing agent, whereas 65% of the nonimplanted patients died. The particularity of ECMO assistance during this condition comes from the vasoplegic properties of many toxic agents, making high flow rates difficult to achieve. Thus, although PVA-ECMO seems very useful in drug intoxication, its exact timing and efficacy for each agent remains to be better determined.

PVA ECMO has become the reference technique for rewarming patients presenting a deep hypothermia (<28 °C) complicated with a cardiac arrest after the description of several case reports and 15 survivors with favorable long-term neurological outcomes in a cohort of 32 patients [15, 18, 19]. In this last cohort, the mean temperature was 21.8 °C, and the mean interval between discovery to ECMO support was 141 min. ECMO allows indeed the fastest rewarming and assures an immediate adequate circulatory support. It also prevents the shock due to peripheral vasodilatation during rewarming, as the central body is rewarmed before its peripheral parts. As low temperatures considerably increase the ischemic tolerance of the brain, many efforts are employed by clinicians to resuscitate those hypothermic patients, with the universal idea that “nobody is dead until warm and dead.” However, the mortality rate after a cardiac arrest associated with deep hypothermia is still high, even in ECMO-assisted patients, ranging from 30 to 80% [20–23]. The major problem is to determine which occurred first between hypoxia and hypothermia. Asphyxia before hypothermia developed is associated with an extremely poor survival (ranging from 0 to 6%) and a poor neurological outcome in survivors [20–22]. On the contrary, patients in whom cooling preceded the cardiac arrest have a very good survival rate (from 60 to 100%) and satisfactory long-term neurological outcome when assisted with an ECMO [19, 20, 22]. Major causes of accidental hypothermia are avalanches, drowning, drug intoxication, and exposure to cold air. In clinical settings the determination of the exact sequence of clinical events is usually very difficult, although asphyxia is usually associated with avalanches, accidents, and drowning and is more unusual in drug intoxication and exposed to cold air patients.

Successful rescue therapy with an ECMO has been described in several cases of life-threatening pulmonary embolism [24, 25], even in patients in cardiac arrest [26]. The technique seems very promising in this indication, as it allows an immediate right ventricule and pulmonary support, whereas medical management of a severe cardiogenic shock due to right ventricular failure remains challenging. Further studies will have to precise the place of ECMO in the therapeutics available during severe pulmonary embolism, particularly regarding thrombolysis therapy. The other main challenge in this field will be to determinate whether an adjunctive treatment must be performed in ECMO-treated patients. One could advocate that a complementary treatment with catheter-guided thrombectomy or surgical embolectomy might allow a faster recovery from the right ventricule dysfunction [26, 27]. Others may argue that, once in place, ECMO allows to buy some time to ensure the natural lysis of the thrombus, as it was demonstrated in several patients [24, 28, 29]. Future studies will help us to better determinate the short-term and long-term outcomes of these different strategies.