Extracorporeal membrane oxygenation (ECMO) use in adults has gained traction in recent years in the setting of technological advances that have made it safer, easier, and more effective.1,2 Typical ECMO configurations are venovenous (VV) and venoarterial (VA) which provide respiratory and combined respiratory and circulatory support, respectively. In VV ECMO, blood is withdrawn from and returned to the central venous system. This can occur via either a dual-site or single-site configuration (Figures 1 and 2). Venovenous extracorporeal membrane oxygenation augments systemic oxygenation in patients with preserved cardiac function; however, it does not provide circulatory support.2 In patients receiving VV ECMO who subsequently develop hemodynamically significant cardiac dysfunction, an additional reinfusion cannula may be inserted into an artery to direct the return of oxygenated blood to both the venous and arterial circulation, thereby providing both respiratory and circulatory support. The addition of an arterial cannula to an existing VV circuit is referred to as venovenous-arterial (VVA) ECMO (Figure 3).
Although a femoral VA ECMO configuration may provide both respiratory and circulatory support, in the setting of severely compromised gas exchange with recovering native cardiac function, the ability to deliver oxygenated blood to the coronary and carotid arteries may be compromised.3,4 Reinfused oxygenated blood from the ECMO circuit flows retrograde through the abdominal aorta and creates a “mixing point” as it meets antegrade blood flow from the left ventricle (Figure 4). To overcome this limitation of femoral VA ECMO, an additional venous reinfusion cannula may be branched from the arterial cannula line and inserted into the internal jugular vein, thereby improving delivery of oxygenated blood to the coronary and carotid arteries following the path of standard two-site VV ECMO. This configuration of venous drainage with combined arterial and oxygenated venous return is referred to as venoarterial-venous (VAV) ECMO (Figure 3). The configuration and physiology of the circuit in VAV and VVA ECMO are ultimately the same; however, the order in which the ECMO is implemented determines the nomenclature.
Such hybrid ECMO configurations, combining the cardiac support of standard VA ECMO with the gas exchange support of standard VV ECMO, offer differential support for various forms of combined cardiopulmonary failure (e.g., interstitial lung disease with pulmonary hypertension or stunned myocardium with concomitant lung injury). With multiple possible hybrid configurations and the ability to adjust the proportion of venous and arterial flow, ECMO support can be tailored to a patient’s specific physiologic needs (Table 1). By altering drainage and reinfusion pressures and flows with different tubing sizes or vascular clamps, the degree of cardiac and respiratory support can be appropriately adjusted to achieve a new physiologic homeostasis (Figure 5). We report our center’s experience using hybrid ECMO in patients with combined respiratory and cardiac failure.
All adult patients who received VAV or VVA ECMO at our institution from January 1, 2012, to December 31, 2013, were included in this retrospective study. Pertinent demographic, clinical, and laboratory data were obtained from our institution’s electronic medical record.
The decision to place a patient on ECMO, as well as initial configuration choice, was made by our institution’s ECMO team, which consists of cardiothoracic surgeons, pulmonary and critical care-trained intensivists, and critical care nurse practitioners. Indications for VV ECMO include severe hypoxemia (PaO2 to FiO2 ratio < 80) despite high levels of positive end-expiratory pressure in patients with potentially reversible respiratory failure, uncompensated respiratory acidosis (pH < 7.15) despite optimal ventilator management, excessively high end-inspiratory plateau pressure despite optimal ventilator management, and as a bridge to lung transplantation in select patients with end-stage lung disease.2
Indications for VA ECMO included extracorporeal cardiopulmonary resuscitation, cardiogenic shock as a bridge to either recovery or long-term circulatory device implantation, right heart failure in patients awaiting lung transplantation, or respiratory failure in patients with excessively high (50% of systemic) pulmonary artery pressure. For patients whose initial cannulation configuration was a hybrid (VVA or VAV) they were all placed on either VV or VA ECMO in an operating room or emergently at the bedside. They immediately demonstrated suboptimal support on the VV or VA configuration, whether it was hemodynamic compromise for the VV patients or persistent upper body hypoxemia for the VA patients. This is primary reason for placing patients on a VVA configuration at the outset of the ECMO run.
Indications for conversion to VVA ECMO from VV ECMO included right ventricular failure, cardiogenic shock, or progressive nonseptic shock. Indications for conversion to VAV ECMO from VA ECMO included differential upper and lower body oxygenation where coronary and carotid oxygenation was poor, as demonstrated by arterial blood gases drawn from the right radial artery.
Our institution’s ECMO circuit consists of either the combination of a Rotaflow centrifugal pump (Maquet Cardiovascular, Rastatt, Germany), Quadrox i or D oxygenator (Maquet), Cobe E Pack tubing (Sorin, Milan, Italy), and heat exchanger, or the CARDIOHELP support system (Maquet).
Patients were cannulated peripherally via Seldinger technique with portable ultrasound or an open cutdown approach after intravenous administration of 3,000–5,000 unit of heparin.4–6 Cannula size and locations are listed in Table 2. All patients received a heparin infusion to target an activated partial thromboplastin time between 40 and 60 s for the duration of ECMO support.
Ventilator strategies vary based on the patient’s indication for ECMO. In those patients in whom we anticipate recovery of lung function we employ a ventilator strategy that is intended to minimize ventilator-associated lung injury. In patients receiving ECMO as a bridge to lung transplantation who require mechanical ventilation, we target patient comfort and ventilator synchrony, and titrate ventilator and ECMO parameters as needed to allow for participation in pretransplant physical therapy. Outcomes reported include survival to ECMO decannulation, survival to hospital discharge, duration of ECMO support, and ECMO circuit-related complications. This study was approved by the Columbia University Institutional Review Board and performed in accordance with accepted ethical standards.
Twenty-one patients received hybrid ECMO during the study period (Tables 3 and 4). Arterial blood gas measurements pre- and post-ECMO cannulation are listed in Table 5.
Initial Venovenous-Arterial Extracorporeal Membrane Oxygenation
Eleven patients were initially cannulated with VVA ECMO because of demonstrated or anticipated severe, combined cardiopulmonary failure. Indications for ECMO included postpartum amniotic fluid or venous thrombus-related pulmonary embolism (n = 2), prelung transplantation-planned VV ECMO complicated by cardiac arrest (n = 2), pulmonary hypertensive crisis (n = 2), acute respiratory distress syndrome (ARDS) complicated by cardiogenic shock (n = 3), ARDS complicated by non-ST segment elevation myocardial infarction (n = 1), and left ventricular assist device failure complicated by cardiac arrest (n = 1).
Venovenous to Venovenous-Arterial Extracorporeal Membrane Oxygenation Conversion Group
Eight patients were converted from VV ECMO to VVA ECMO. Initial indications for VV ECMO included ARDS from either aspiration pneumonitis or pneumonia (n = 3), end-stage cystic fibrosis (CF) awaiting lung transplantation (n = 2), interstitial pulmonary fibrosis (IPF) awaiting lung transplantation (n = 1), progressive hypoxemia from IPF ineligible for transplantation (n = 1), and hypertrophic obstructive cardiomyopathy complicated by acute respiratory failure (n = 1).
Reasons for conversion to VVA ECMO included the development of outflow restriction in the bicaval dual-lumen cannula in two patients with ARDS. One of these patients had significant clot burden in the right internal jugular vein as well as in the dialysis and central venous catheters. The second patient had the tip of the cannula in the right ventricle precluding adequate outflow. A third patient with ARDS developed acute right ventricular failure most likely caused by pulmonary embolism. One pretransplant patient with CF developed a clot in the membrane oxygenator leading to massive pulmonary embolism while on VV ECMO. The second CF patient had progressive hypoxemia and secondary pulmonary hypertension, requiring not only an arterial reinfusion limb but also an additional venous cannula to improve venous drainage, effectively creating a venovenous-arterial-venous (VVAV) circuit. The pretransplant patient with IPF developed worsening hypoxemia and secondary pulmonary hypertension. The patient with IPF who was ineligible for transplantation developed progressive right ventricular failure, and the patient with hypertrophic obstructive cardiomyopathy had progressive acidosis and shock.
Venoarterial to Venoarterial-Venous Extracorporeal Membrane Oxygenation Conversion Group
Two patients were initially placed on VA ECMO, but were ultimately converted to VAV ECMO. One patient underwent lung transplantation on VA ECMO and was briefly converted to VAV ECMO to transition to VV ECMO when hemodynamic support was no longer required. The second patient was placed on VA ECMO emergently via femoral cannulation, but had differential oxygen saturation with persistent upper body hypoxemia requiring a venous reinfusion cannula in the internal jugular vein.
Of the 21 patients, eight (38.1%) died during ECMO support, four (19.0%) died after decannulation but before hospital discharge, and nine (42.9%) survived to hospital discharge. Four of 11 (36.4%) initially cannulated with a VVA ECMO configuration survived; four of the eight (50%) converted from VV to VVA ECMO survived; and one of two (50%) converted from VA to VAV ECMO survived (Table 6). The mean duration of ECMO support was 6.5 ± 5.5 days (range: 1–41). The average APACHE II Score was 31.1 ± 8.4 (Table 3).
Complications included oxygenator failure (n = 3), visible oxygenator clot (n = 4, one of which led to pulmonary embolism), cannula thromboses, and cannula malrotation requiring repositioning. Seven patients had bleeding complications, including hemoptysis (n = 1), gastrointestinal hemorrhage (n = 1), and cannula site-related hemorrhage (n = 5).
Both postpartum patients, each of whom had undergone Caesarean section, experienced disseminated intravascular coagulopathy leading to persistent bleeding at their surgical sites necessitating re-exploration. Limb ischemia occurred in three patients with one developing a lower extremity compartment syndrome from a left iliac artery bleed, ultimately requiring an above-the-knee amputation. There were no cases of pump malfunction, heat exchanger malfunction, air entrainment, intracerebral hemorrhage, cardiac arrhythmia, or heparin-induced thrombocytopenia.
We report our center’s experience with hybrid ECMO configurations in patients with concomitant cardiac and respiratory failure requiring extracorporeal support. For severe respiratory failure without cardiac dysfunction, VV ECMO is the configuration of choice, whereas cardiac failure with preserved pulmonary function is well managed by VA ECMO.2 Although these configurations are often sufficient, patients with severe, combined cardiopulmonary failure may require both respiratory and hemodynamic mechanical support.7 The VAV or VVA ECMO may provide sufficient physiological support when traditional VV or VA configurations are inadequate. In particular, patients on VA ECMO who have residual cardiac function and severely compromised native lung function may have inadequate oxygen delivery to the coronary and carotid arteries, potentially exacerbating neurologic and cardiac insults.8
Although VV ECMO does not provide direct hemodynamic support, many patients with respiratory failure, even in the setting of severe shock, may still be adequately supported. This is particularly true in younger patients without underlying cardiac dysfunction or pulmonary hypertension. In these patients, the improvement in oxygen delivery to the coronary arteries and the correction of respiratory acidosis that occurs after initiation of VV ECMO may improve both cardiac function and responsiveness to vasopressors, thereby reducing or eliminating shock. Patients who are older, have underlying pulmonary hypertension, or are cannulated during a cardiac arrest may initially require VA or VAV ECMO. Some patients who start with VV ECMO subsequently develop cardiogenic shock from septic cardiomyopathy, pulmonary emboli, worsening pulmonary hypertension, or fulminant myocarditis, making the placement of an arterial reinfusion cannula the next logical step in the treatment algorithm.
More complex is the physiology of patients initially placed on peripheral VA ECMO via femoral cannulation, or those initially presenting with both cardiac and respiratory failure. Femoral VA ECMO may provide adequate hemodynamic support,9 but may be insufficient to oxygenate the upper body, including the coronary and cerebral arteries, when there is residual native cardiac output and impaired native gas exchange.3,4 The reinfused oxygenated blood via the femoral artery competes with antegrade flow from the left ventricle to supply blood to the coronary and carotid circulations, referred to as the “mixing point.”3,4 The location of this interface is not physiologically significant when native gas exchange is satisfactory. However, in the setting of severe respiratory failure with residual native cardiac output, the “mixing point” will be lower in the aorta, and the coronary and cerebral vasculature may receive poorly oxygenated blood from the native circulation. The precise location of the mixing point is difficult to determine in any given patient and is dynamic over time, varying with the relative contributions of the native cardiac output and the arterial flow from the device as well as peripheral vascular resistance. Insertion of an additional reinfusion limb into the superior vena cava with outflow directed into the right atrium ensures adequate upper body oxygen delivery, while still providing the cardiac support of VA ECMO.
Although there is a paucity of literature on the subject of hybrid ECMO configurations, it is a strategy that, when used properly, can adequately support the physiology of complex patients. The VAV approach should be considered in all patients with combined severe cardiac and respiratory failure. Although an upper body cannulation using the subclavian artery is sufficient to support both circulation and oxygenation to the head vessels and below, the coronary arteries may still receive blood from the native cardiac output, and in select patients with such poor native lung function, the VAV approach is preferred.
For patients initially on VV ECMO, we recommend considering the addition of an arterial reinfusion limb when severe cardiogenic shock ensues. In an emergent setting this can be done via the femoral artery, as the upper body remains well oxygenated from the initial VV circuit. In a less urgent setting, the arterial limb can be added to the subclavian artery via an end-to-side graft, providing cardiac support and adequate oxygenation to the head vessels and beyond.5
For patients initially receiving VA ECMO support, we recommend using only the level of support that is required to maintain adequate blood pressure and systemic perfusion. We rarely use a blood flow that exceeds 75% of the patient’s calculated cardiac output, targeting a cardiac index of 2.4 L/min/m2. Because of this approach, we rarely see pulmonary edema in the setting of impaired left ventricular emptying caused by the retrograde arterial ECMO blood flow. However, if pulmonary edema does occur, or if the patient’s native lung function independently worsens, we prefer to insert an internal jugular oxygenated blood reinfusion limb. This has the benefit of providing the coronary and carotid circulation with oxygenated blood while obviating the need to place an intracardiac vent to decompress a distended left ventricle.
For all hybrid configurations we are able to titrate the arterial and venous reinfusion flows with the use of vascular clamps to preferentially divert blood flow toward one of the limbs (Figure 6). At the onset of VAV or VVA ECMO, we direct two-thirds of the reinfusion blood flow toward the arterial limb. We then employ a dynamic management strategy based on the patient’s physiologic needs as intrinsic cardiac and pulmonary function recovers. Regardless of total flow, we maintain at least 1 L/min of blood flow through each limb to avoid thrombus formation within the cannulae and tubing.
Our survival and complication rates are likely related to the high severity of illness of these patients. Nonetheless, our initial experience suggests that hybrid ECMO configurations can be effective, and it serves patients well to maintain a flexible and adaptable approach to ECMO configurations.
1. Paden ML, Conrad SA, Rycus PT, Thiagarajan RRELSO Registry. . Extracorporeal Life Support Organization Registry report 2012. ASAIO J. 2013;59:202–210
2. Brodie D, Bacchetta M. Extracorporeal membrane oxygenation for ARDS in adults. N Engl J Med. 2011;365:1905–1914
3. Stulak JM, Dearani JA, Burkhart HM, Barnes RD, Scott PD, Schears GJ. ECMO cannulation controversies and complications. Semin Cardiothorac Vasc Anesth. 2009;13:176–182
4. Field ML, Al-Alao B, Mediratta N, Sosnowski A. Open and closed chest extrathoracic cannulation for cardiopulmonary bypass and extracorporeal life support: Methods, indications, and outcomes. Postgrad Med J. 2006;82:323–331
5. Javidfar J, Brodie D, Costa J, et al. Subclavian artery cannulation for venoarterial extracorporeal membrane oxygenation. ASAIO J. 2012;58:494–498
6. Javidfar J, Brodie D, Wang D, et al. Use of bicaval dual-lumen catheter for adult venovenous extracorporeal membrane oxygenation. Ann Thorac Surg. 2011;91:1763–1768 discussion 1769
7. Bartlett RH, Roloff DW, Custer JR, Younger JG, Hirschl RB. Extracorporeal life support: The University of Michigan experience. JAMA. 2000;283:904–908
8. Abrams D, Combes A, Brodie D. Extracorporeal membrane oxygenation in cardiopulmonary disease in adults. J Am Coll Cardiol. 2014;63(25 Pt A):2769–2778
9. Tang GH, Malekan R, Kai M, Lansman SL, Spielvogel D. Peripheral venoarterial extracorporeal membrane oxygenation improves survival in myocardial infarction with cardiogenic shock. J Thorac Cardiovasc Surg. 2013;145:e32–e33