Are two crutches better than one? The ongoing dilemma on the effects and need for left ventricular unloading during veno-arterial extracorporeal membrane oxygenation

This article refers to ‘Concomitant implantation of Impella® on top of veno-arterial extracorporeal membrane oxygenation may improve survival of patients with cardiogenic shock’ by F. Pappalardo et al., published in this issue on pages 404–412.

The mortality rate in acute cardiogenic shock (CS) is high and, more importantly, has not changed remarkably during the last 20 years.1, 2 Recently, the IABP-SHOCK II trial, which included the gold-standard therapeutic management of acute myocardial infarction-related CS, has shown that, at 30 days, 40% of the patients, regardless of whether supported with or without an intra-aortic balloon pump (IABP), ultimately die. In the case of refractory CS, the in-hospital mortality is even higher regardless of the aetiology of severe cardiac dysfunction.3, 4 The use of partial cardio-circulatory support, such as an IABP, and inotropic drugs therefore appear insufficient to counteract severe and unresponsive acute cardiac impairment significantly. So, we must confess that today a substantial proportion of patients with refractory CS have a dismal prognosis with conventional therapy.5

The use of temporary, or so-called ‘short-term’, mechanical circulatory support devices with full cardio-circulatory assistance has been increasing significantly in recent years, indicating that physicians are apparently more prone to consider more aggressive approaches to sustain cardio-circulatory function in such circumstances.6

Extracorporeal membrane oxygenation (ECMO), which represents a high-tech refinement of the extracorporeal technology used in everyday cardiac surgery practice, is being more and more considered a promising tool to allow a bridge to decision or to myocardial recovery in this scenario.7, 8 During the last decade, industry has provided ECMO systems which are increasingly miniaturized, easy to set-up, with an enhanced peripheral approach thanks to improved cannula designs, as well as with more compatible, performant, and durable system components. Despite all the above-mentioned benefits, however, ECMO accounts for several adverse effects either linked to the ‘flow-related dynamics’ or to inherent features related to continuous contact between the blood and artificial surfaces, obviously not forgetting the critical conditions of the majority of treated patients (many almost moribund).

The use of an artificial heart/lung machine for several hours, days, or weeks therefore comes at a cost. ECMO action induces a systemic inflammatory reaction and needs anticoagulation to avoid massive thrombosis in the circuit, but also generates an increased LV afterload secondary to the flow travelling in a retrograde direction towards the heart, with either a central or a peripheral access. This condition usually induces a further decrease of the already severely reduced antegrade transaortic valve flow, particularly in patients with very poor contractility or with a heart almost at a standstill. Such a combination of factors may lead to a variable extent of blood stagnation in the left ventricle, left atrium, and pulmonary venous system, with subsequent increased pressure in the left cardiac chambers, leading to subendocardial ischaemia, thrombus formation, and pulmonary oedema. Avoiding these untoward conditions while guaranteeing adequate peripheral perfusion is necessary with ECMO. However, these negative consequences do not occur in all patients. At least, not to such an extent so as to cause deterioration of haemodynamics.

In this issue of the journal, Pappalardo and colleagues have described a two-centre experience of the combination of ECMO and a left heart unloading system, named Impella®, to avoid LV-related blood stasis and dilatation in patients with CS.9 Such an axial flow pump, implanted by means of a peripheral access and placed in the left ventricle through the aortic valve, generates a transaortic suction, thereby contributing to cardiac and pulmonary vascular unloading,10 in addition to that achieved by the ECMO system which diverts the blood from the right heart chambers directly into the arterial circulation after oxygenation.

By a propensity-score analysis, the authors showed that the association of such assist devices was beneficial in terms of better hospital survival, and with a higher rate of patients who ultimately achieved cardiac recovery or were successfully bridged to further therapy, as compared with individuals supported only with ECMO. In contrast, they reported a high incidence of haemolysis and continuous veno-venous haemofiltration in those patients with the LV unloading device.

The combination of these two types of mechanical circulatory support systems is not new. Several other authors in the last few years have reported about this association of devices.11-13 Other alternatives to achieve the same goal in ECMO patients have also been proposed,14, 15 and others will be developed (Table 1). The study of Pappalardo and co-workers, however, has some novel and relevant information, mainly by providing a comparison between two ECMO-based configurations, with one represented by ECMO and the Impella® device, indicating that the concomitant use of the two systems might be more advantageous in terms of patient outcome.

It seems, from this study, that two crutches (in this case ECMO and Impella®) might be better than one (ECMO alone) to help the ‘lame’ heart. Nonetheless, as is well known in orthopaedics, using two crutches might be cumbersome as compared with the use of only one. Adverse effects should also be taken into account when combining the two cardio-circulatory assist systems, or may be just strictly related to the Impella® device. Indeed, as also highlighted by the authors, haemolysis is an issue in Impella® patients,10 and this might further deteriorate as a result of blood cell damage sometimes seen in ECMO patients. Appropriate device position, i.e. at the aortic valve level, may not be consistently guaranteed, with a possible event being the dislodgement of the axial pump in the ventricle or in the ascending aorta.10 Additionally, aortic mitral valve injury may also occur in rare cases.10 Constant clinical and echocardiographic checks, as confirmed by the authors, are therefore mandatory, although the frequency, complexity, and management of such adverse events have not been or have only partially been addressed in this study.9 Furthermore, costs must also be taken into account, based on the association of two expensive systems, and not forgetting the lack of consistent scientific evidence to date about the benefits of the Impella® pump.

The need for LV unloading remains a controversial issue in ECMO patients. So far, no study was able to prove the relevance or the actual need for LV unloading in ECMO. Some physicians advocate the use of such an approach in almost all ECMO cases, whereas others claim the absence of or negligible rates of negative haemodynamic ECMO-induced conditions. Other less expensive and less sophisticated means, such as an IABP, have been claimed to be sufficient.

From a theoretical standpoint, LV unloading should be beneficial in veno-arterial ECMO, but at what price, and for all patients? Is the balance between potential complications related to implant as well as function of venting procedures and LV unloading-related benefits in favour of this approach? Is the need for a higher level of close patient monitoring and costs related to Impella® a heavier burden than the potential benefits? Should we select patients on ECMO in whom such an axial transaortic pump would make a certain and critical difference in terms of outcome? Are embolic events reduced in ECMO patients with concomitant LV unloading? Which approach is better for LV unloading or with a higher benefit/complication ratio? Does added LV unloading provide a better and more prompt myocardial recovery than ECMO alone? Although the authors should be commended for their efforts in addressing this issue, particularly with an elegant statistical analysis, their study provides some light in this respect, but obviously not in a conclusive way, and all these issues remain mainly unresolved. So, although this study provides some promising clues in this setting, a prospective randomized trial, collecting as much information as possible, and hopefully for all ‘ECMO comers’, is certainly needed to support this approach and theory conclusively, together with information on the type of ancillary device.

In conclusion, the study of Pappalardo and co-workers represents a meaningful contribution and a step forward in the dilemma regarding ECMO and LV unloading or ECMO alone. ECMO is becoming a more and more valuable tool in the treatment of patients with refractory CS thanks to refined and highly effective temporary mechanical circulatory support, but the ideal configuration is yet to be determined. Further research, therefore, is required prior to suggesting or recommending the use of such ‘two crutches’ and which kind of associated crutch would be advisable for a more effective cardiac support in acute and severe cardiac contractile impairment, particularly in refractory CS.

Conflict of interest: none declared.