Reducing the Shock of Acute Respiratory Distress Syndrome: Is Veno-Venous ECMO a Viable Therapy in the Hemodynamically Unstable Patient?

Over the past decade, the use of veno-arterial (VA) and veno-venous (VV) extracorporeal membrane oxygenation (ECMO) has increased by 2- to 4-fold, and these patients are even requiring cardiac and noncardiac surgical procedures.1–3 With its increased utilization, ECMO has emerged to become a valuable therapeutic technique for a variety of conditions. VA ECMO is being increasingly deployed in patients with refractory cardiogenic shock and during cardiac arrest, whereas VV ECMO is being utilized in acute respiratory distress syndrome (ARDS), respiratory failure secondary to polytrauma, and as a bridge to lung transplantation in patients with end-stage chronic respiratory diseases. In this issue of Anesthesia and Analgesia, Gutsche et al4 present a retrospective case series of 17 patients with hemodynamic instability and refractory hypoxemia who were managed with VV ECMO. After VV ECMO deployment, the need for vasoactive support, as measured in norepinephrine equivalents, decreased at both 6- and 24-hour time points. No patient required rescue VA ECMO. Of the 35 patients screened for the study, 18 were excluded because they did not require vasopressor or inotropic support, leaving an arguably “sicker” subset of patients on VV ECMO for analysis. Important limitations to the study are present such as selection bias, in which VA ECMO could be deployed at practitioner discretion, a lack of standardized criteria for vasopressor management, and provider bias resulting from the nonrandomized, nonblinded retrospective nature of this series. Although patients were placed on VA ECMO if they demonstrated “severe” ventricular dysfunction on predeployment echocardiography, no specific quantitative echocardiographic criteria were listed in the definition of this classification. Finally, although this case series is novel, this is potentially a product of publication bias, where other centers may not have had similar results with VV ECMO. Nevertheless, 64.7% of patients in the series survived to discharge, comparing favorably with other studies reporting survival to discharge of 63.2% to 73.7%.5–7 Although VA and VV ECMO utilize the same mechanical system, they are by no means interchangeable therapies. VA ECMO, which is used in cardiogenic shock, has inherent risks including left atrial and ventricular thrombosis that can result in stoke, arterial injury, and limb ischemia.3 Patients requiring VA ECMO also require a greater degree of anticoagulation than patients on VV ECMO.8 Given that cerebral hemorrhage is a potentially fatal ECMO complication, minimizing anticoagulation is valuable, particularly because viral illnesses capable of causing severe respiratory failure such as H1N1 can also cause encephalitis and alter the mental status.9 In patients who have severe respiratory failure and cardiogenic shock, especially those with viral myocarditis, VA ECMO will usually be the most supportive therapy; however, VV ECMO is preferred for patients with respiratory failure alone. If the heart is still able to eject blood, patients on VA ECMO cannulated via the commonly used configuration of the femoral vein and artery may develop mixing of both anterograde blood flow from the left ventricle and retrograde blood flow from the arterial ECMO cannula (also known as “north–south syndrome” or “Harlequin syndrome”). Such blood mixing causes the relatively oxygen-poor blood from central circulation to perfuse the brain and coronary arteries, whereas the oxygen-rich post-ECMO blood perfuses more distal distributions. Recognition of hypoxemia in the brachiocephalic artery using a right-sided arterial line or pulse oximeter is critical, because management often requires surgical intervention such as addition of a third cannula to oxygenate blood circulating through the heart and lungs or transitioning to a central cannulation strategy to decrease cardiac blood flow. Importantly, addition of a third (arterial) cannula can also be performed in hemodynamically unstable patients on VV ECMO, which results in a transition to veno-venous-arterial ECMO. Hybrid VA-VV ECMO techniques effectively treat the mixed state of hypoxemia and cardiogenic shock by oxygenating and decarboxylating blood downstream of the pulmonary artery, preventing the occurrence of north–south syndrome and providing mechanical circulatory support.10 Gutsche’s results demonstrate that VV ECMO can be successfully utilized despite hemodynamic instability and that the therapy itself may decrease patients’ requirements for vasoactive drugs. Although the cause of mortality was not described in patients who died, conversion to VA ECMO was permitted but not utilized in the study. Although further studies are necessary to clarify the indications for VV versus VA ECMO, Gutsche et al’s data4 suggest that VV ECMO should be considered as an alternative over VA ECMO in patients who require vasoactive support in the absence of severe myocardial dysfunction. Specifically, VV ECMO should not be withheld on the basis of a pre-ECMO need for vasopressors in patients who would otherwise be candidates as a result of severe respiratory failure. Although the authors were unable to identify the physiologic basis for the observed reduction in vasoactive requirements they observed, one possibility was the improvement in pH after ECMO deployment. An increase in pH 2 hours after ECMO deployment was statistically associated with a reduction of vasopressor requirements at 6 but not 24 hours; however, baseline pH values were not reported in the study, so the clinical significance is uncertain. Improvements in PaO2 or positive end-expiratory pressure (PEEP) were not associated with the need for vasoactive support. Each of these variables can independently impair cardiac output—because hypoxemia, hypercarbia, and acidemia may increase pulmonary vascular resistance (PVR) potentially impairing right ventricular (RV) function, whereas significant necessary PEEP, given the level of hypoxemia, may reduce effective preload. Acidemia also reduces the efficacy of vasoactive drugs, and the observed decreases in norepinephrine requirement in the study by Gutsche et al4 may reflect an increase in drug efficacy rather than an improvement in hemodynamics.11 In patients with ARDS, a norepinephrine requirement may be a surrogate for hypovolemia. The authors stated that they performed echocardiography and “normalized volume status,” leaving hypotension attributable to either myocardial dysfunction or vasoplegia. Both possibilities are inherent to many common causes of ARDS. Impaired ventricular function is present in up to 72% of patients with H1N1, and the virus can even cause fulminant myocarditis.12,13 Myocardial dysfunction also occurs in up to 50% of patients in septic shock.14,15 One probable mechanism by which VV ECMO may improve hemodynamics in such patients is through its ability to reverse acidemia and hypercapnia leading to a PVR decrease and subsequently improving RV function. The potential combination of cardiomyopathy resulting from critical illness and increased ventricular afterload caused by pulmonary artery hypertension likely results in impaired cardiac output and hypotension. In a series of 13 patients with pneumonia, Miranda et al16 noted that the deployment of VV ECMO led to immediate reduction in mean pulmonary artery pressure and suggested that right ventricular unloading resulted in a “significant, but not clinically relevant” increase in cardiac index by thermodilution without a change in vasopressor requirements. The authors proposed that a reduction in PVR, secondary to a decrease in PaCO2 and increase in SvO2, caused RV unloading, leading to an increase in cardiac index.16 This pathophysiology could also be a mechanism for the reduction in vasopressor requirements in Gutsche’s series. In addition, although hypoxemia increases pulmonary artery pressure, it did not affect vasopressor requirements in the study by Gutsche et al.4 This lack of relationship could be because the level of hypoxemia before ECMO deployment was insufficiently severe to cause increased PVR or that the study was underpowered to detect a difference. The level of PEEP also did not correlate with vasopressor need in this study. Meanwhile, the supportive effects of inotropes are blunted in patients with systemic hypoxemia and acidemia giving physiologic support to the association between increased post-ECMO pH and the reduction in vasoactive medications identified by the authors.11 Although this case series is relatively small, it identifies patients who were successfully managed with VV ECMO who otherwise may had been treated with VA ECMO or not offered ECMO at all. Based on current understanding of ECMO pathophysiology, the patients in whom the greatest hemodynamic benefit of VV ECMO might be expected are those with profound acidemia, in whom vasoactive drugs may be ineffective and RV failure may be exacerbated. A prospective data set could help better specify the relative roles of VA and VV ECMO in critically ill patients with cardiorespiratory failure. DISCLOSURES Name: Michael A. Fierro, MD. Contribution: This author helped write the manuscript. Name: Raquel R. Bartz, MD, MMCi. Contribution: This author helped write the manuscript. This manuscript was handled by: Avery Tung, MD, FCCM.