Extracorporeal support, in the form of cardiopulmonary bypass, is used frequently in the operating room for physiologic support during cardiac surgery. However, in patients with severe respiratory and/or cardiac failure, such support may be required for prolonged periods of time in the intensive care unit setting. Typically, extracorporeal membrane oxygenation (ECMO) is applied in one of two configurations. For acute respiratory failure, a venovenous configuration is used in which deoxygenated venous blood is drained from the body, oxygenated, cleared of carbon dioxide, and pumped back into the venous system near the right atrium, at which time it enters the pulmonary and then systemic circulations. Although much more complex, a venoarterial configuration can be used for refractory cardiac failure, often with associated respiratory failure; this approach also involves drainage of blood from the venous circulation, which is then oxygenated, cleared of carbon dioxide, and pumped back into the arterial system, either through a femoral artery (peripheral) or directly into the ascending aorta (central). Whereas venovenous ECMO requires intact cardiac function, venoarterial ECMO can unload both the left and right heart and allow for rest and recovery—in theory.

The use of ECMO as salvage therapy for acute decompensated heart failure is becoming more popular despite a dark past and only a small improvement in outcomes over the years. An initial trial sponsored by the National Institutes of Health between 1974 and 1977 was discontinued early because of poor outcomes. These poor outcomes were attributed to a lack of established guidelines and the fact that few centers had significant experience in managing patients with this very complex, resource-intensive and labor-intensive therapy. However, despite this initial setback, those dedicated to the concepts of using extracorporeal support for the acutely failed heart and/or lungs to allow them time to recover, particularly when applied to dying patients with no other options, continued to push the technology and science. After worldwide successes with ECMO for pandemic influenza 2009 H1N1 and publication of the results of the Conventional Ventilatory Support Versus Extracorporeal Membrane Oxygenation for Severe Adult Respiratory Failure trial that showed a survival advantaged for ECMO over conventional ventilator management for severe respiratory failure, there has been renewed enthusiasm for venovenous ECMO for pulmonary support. Similarly, venoarterial ECMO for acute cardiogenic shock is also becoming a more commonly used tool in the early management of critically ill patients as part of a treatment strategy to bridge to recovery, ventricular assist device implantation, or transplantation. Algorithms that incorporate early venoarterial ECMO as part of the treatment for witnessed, in-hospital, cardiopulmonary arrest (ECMO-assisted cardiopulmonary resuscitation) are also becoming more common and have resulted in a twofold increase in neurologically intact patients surviving to discharge. Clearly, we are at the dawn of a new and exciting day for ECMO as indications for use become broader and contraindications such as neurologic pathology or inability to anticoagulate are challenged.

In this issue of JASE , an article by Aissaoui et al. takes advantage of a somewhat unique physiologic situation to assess the performance of several conventional as well as new markers of left ventricular (LV) function, namely, strain derived from velocity Vector Imaging. The authors used ECMO as a model to assess the load dependency of various indices of LV filling and contractility derived using echocardiographic techniques. By systematically decreasing the level of mechanical support and thus simultaneously increasing preload and decreasing afterload, they assessed the impact of changing loading conditions on mean arterial blood pressure, transmitral E velocity, E/e′ ratio, LV ejection fraction, aortic valve velocity-time integral, tissue Doppler lateral Ea, Sa, and parameters derived from Velocity Vector Imaging, including lateral systolic velocity, strain, and strain rate. In a relatively small number of subjects, they demonstrated a load dependence of the Velocity Vector Imaging–derived parameters of strain, a finding that is somewhat different from the results of previous animal model work. As ECMO support was weaned, LV ejection fraction, lateral E/Ea ratio, and velocity-time integral all increased, as did lateral strain and strain rate. In contrast, lateral Sa did not change. As a secondary finding, the authors addressed the role of these parameters in predicting the potential for weaning of ECMO. Patients who were successfully weaned from ECMO had more stable hemodynamics as ECMO support was reduced, higher ejection fractions and velocity-time integrals, as well as slightly (although not significantly) greater strain.

So what can we make from this? Is this just another report, among many over the years, exploring the changes in echocardiographic indices during acute and chronic changes in preload, afterload, and various pathologic interventions in both normal and diseased hearts? Or is there a greater clinical value to their findings?

Traditional and easily reproduced echocardiographic indices, such as ejection fraction, end-diastolic volumes, and even simple ratios of Doppler filling patterns such as the E/A ratio, continue to be used despite their well-known limitations. However, there is a clear need for more sophisticated methods to assess the complex physiology of systolic and diastolic function quantitatively and noninvasively, particularly in patients with severe cardiovascular disease in whom there is the need to monitor the effects of various therapies closely. Although a variety of different indices or combinations of indices have been proposed and validated under different physiologic and pathophysiologic conditions, their use clinically has been somewhat limited. Maybe they will find an application in evaluating the complex cardiovascular physiology of an unloaded and sick heart during ECMO support.

Clearly, challenges remain, as evidenced by the fact that outcomes for cardiac support (venoarterial ECMO) and pulmonary support (venovenous ECMO) have changed little over the past 10 years. Although single centers, with small series or unique patient populations, might report impressive outcomes, the reality is that survival after cardiac support is still on the order of 30% to 50%. Survival for pulmonary support is somewhat better at 50% to 60%. Even though there is much enthusiasm for newer technologies, such as portable pumps and advanced oxygenators, recent evidence has questioned the role of improvements in ECMO technology in improving outcomes. Clearer understanding of the indications and better patient selection and management by skilled centers have resulted in improved outcomes. However, further improvements will need to come not from patient selection but rather what we do with patients once therapy is initiated. Unfortunately, although large series have demonstrated risk factors for poor outcomes, there are few objective tools and data to guide clinicians in how to manage patients on ECMO in terms of determining if and when the heart (or lungs) will recover. In fact, given the nature of the precipitating event, such as an acute myocardial infarction or inability to wean from cardiopulmonary bypass after cardiac surgery, the final result is a heart that might be irreversibly damaged. But the problem is figuring that out definitively. For a variety of reasons, and given the crude tools available at the bedside, we still have great difficulties in determining which hearts will and which ones will not recover. This problem becomes significant because, unlike respiratory failure in which there are few options should the lungs not recover (even at very experienced centers, bridging patients with acute or chronic lung failure to lung transplantation is considered very high risk and is rarely performed), patients on venoarterial ECMO for heart failure fortunately have additional options. There is a growing body of evidence demonstrating the role of ECMO as a bridge to either a long-term ventricular assist device and subsequent heart transplantation or, less commonly, directly from ECMO to transplantation. The key to this decision making is understanding which patients will recover, not recover and require long-term heart failure options, or die from their disease and associated comorbidities. Given the expense, invasiveness, and limited availability of donor hearts, the decision to pursue either ventricular assist device implantation or transplantation is not taken lightly. Unfortunately, death is often an outcome that comes from waiting too long—waiting for recovery or some clue or clues as to which direction a patient will take. Furthermore, it is well known that the longer patients are treated with ECMO, the less likely they are to survive, illustrating the time constraints that constantly need to be evaluated when asking the question, Is this heart going to get better, or not?

The value of echocardiography in patients requiring ECMO was also recently reviewed in JASE . The quantitative assessment of ventricular function, beyond simple parameters such as ejection fraction, is a missing piece of the puzzle of how to care for these complex patients. The role of echocardiography is well established in ruling out mechanical complications, such as anatomic defects (aortic dissections, mitral regurgitation, aortic insufficiency, tamponade, etc) or ECMO-system complications (cannula position, thrombus, etc), but clearly the next step needs to be a sophisticated assessment of ventricular function to determined the degree of ventricular recovery. Hopefully then, with this further insight and experience, echocardiography can the play a critical role in decision making.

If we explore the literature, we quickly realize how few data there are on the topic of quantitative echocardiography and weaning mechanical support. Most of the data and experience come from patients with LV assist devices (LVADs). Typically, ejection fraction and ventricular dimensions were the only parameters used. Simon et al. used quantitative echocardiography and reported that successfully weaned patients had increases in stroke area and fractional area change, as well as greater maximal oxygen consumption, compared with LVAD-dependent patients. Khan et al. and colleagues used dobutamine echocardiography to predict successful LVAD explantation. They reported that successfully weaned patients had improved cardiac indices, improved LV ejection fractions, improved dP/dt, and decreased LV end-diastolic dimensions during dobutamine stress echocardiography. More recently, Estep et al. reviewed the role of echocardiography in patients with LVADs. In the assessment of myocardial function, echocardiographic parameters that may be predictive of recovery include an LV ejection fraction ≥45%, an LV end-diastolic diameter ≤55 mm, or a relative wall thickness ≥0.38, as well as the pattern of aortic valve opening (a measure of the ability of the left ventricle to generate sufficient force and output to support the circulation) in response to changes in the degree of LVAD support or with exercise. It is important to consider that the complex LV and right ventricular interactions with an LVAD are much different than those of an unloaded and acutely injured heart supported with ECMO. Nevertheless, the concepts of using advanced echocardiographic imaging modalities to assess for recovery in an unloaded (or partially loaded) ventricle are similar and hopefully will spark further studies.

The search continues for a robust, readily available tool to guide weaning, especially in the acute setting, for patients being supported by ECMO or LVADs. It remains to be determined if newer parameters of LV systolic function such as strain imaging provide incremental information compared with more conventional measures. In the interim, a combination of clinical judgment, hemodynamic parameters (heart rate, blood pressure, central venous pressures, pulmonary artery pressures, cardiac output and index), oxygenation status, as well as conventional (linear dimensions, ejection fraction, velocity-time integral of the aortic valve outflow, and even just the ability of aortic valve to open) and potentially more advanced echocardiographic techniques (including tissue Doppler velocities and strain imaging), should be used to aid in the management of these complex, critically ill patients.

Hopefully, combining the growing enthusiasm for ECMO with the interest in advanced quantitative echocardiographic imaging will continue to spark studies such as the one by Aissaoui et al. With any growing technology, once there is a critical application and useful tools, the growth and level of understanding increase rapidly, and better outcomes result. There is no doubt that the merging of ECMO and echocardiography will accelerate the fields of both by translating quantitative echocardiography into more of a mainstream clinical tool while bringing ECMO out from the shadows of an obscure salvage therapy.