Chapter 7. Extracorporeal Cardiopulmonary Membrane Oxygenation

Despite the advances in lung-protective strategies and cardiac assist devices, severe pulmonary and cardiac failure continue to be associated with high mortality. Acute respiratory distress syndrome (ARDS) still maintains a mortality rate as high as 30–40%, 50% for patients with cardiogenic shock.1 Patients with failing conventional or advanced methods of therapy or patients who are worsening clinically have few rescue therapy options. Early referral to an extracorporeal membrane oxygenation (ECMO) center might be the only option. This chapter will review the clinical indications, contraindications, types of ECMO, and the complications of ECMO use.

ECMO is a term used to describe a form of partial cardiopulmonary bypass used for temporary, albeit prolonged, support of respiratory and/or cardiac function. Cardiopulmonary bypass was initially developed for use in the operating room to provide short-term cardiopulmonary support (CPS) during cardiac surgery procedures. Extracorporeal membrane oxygenation is more commonly known as and referred to by the acronym ECMO. Other names synonymous with ECMO include extracorporeal lung assist (ECLA), extracorporeal CO2 removal (ECCOR), CPS, and extracorporeal cardiopulmonary resuscitation (ECPR).

Currently ECMO is utilized in the critical care setting for patients with acute, severe, potentially lethal but reversible respiratory or cardiac failure unresponsive to conventional medical management. It is occasionally used in patients with irreversible cardiac or respiratory disease under the circumstances that the patient is a candidate for a heart or lung transplant. In these cases, ECMO may be used as a “bridge” before or after transplantation. ECMO is reserved for patients with potentially reversible disease who are unlikely to survive with conventional medical management. It is important to remember that the primary medical indication for ECMO therapy must be a potentially reversible condition, with possibility for other treatments such as revascularization or transplant.

ECMO was developed in the 1970s from a modification of cardiopulmonary bypass. It uses a modified heart–lung bypass machine to provide gas exchange, and systemic perfusion if necessary, which can provide pulmonary and cardiac support. Unlike standard cardiopulmonary bypass, which is used for short-term support measured in hours, ECMO can provide longer support ranging from days to weeks in the intensive care unit (ICU). Additionally, the purpose of ECMO is to compensate for further harm or allow for intrinsic recovery of the heart and lungs, unlike standard cardiopulmonary bypass, which provides support during various cardiac surgical procedures.

ECMO functions by providing membrane oxygenation to temporarily take over the role of the lung or the heart. It provides for gas exchange while mechanical ventilation settings are adjusted to prevent a high-pressure environment. This minimizes ventilator-induced lung injury (VILI) and maximizes lung recruitment of the functional residual capacity. By providing support without reliance on mechanical ventilation for gas exchange, the native lung has time to heal and potentially recover. Additionally, some believe that the injured lung activates the release of inflammatory mediators, which may precipitate renal failure, liver failure, cardiac failure, and other systemic consequences. Release of these inflammatory mediators may be significantly decreased with ECMO support compared with high-pressure mechanical ventilation.2

There is clear evidence of the efficacy of respiratory ECMO for neonates. In 1996, a randomized controlled trial involving 185 neonates with respiratory failure showed a mortality reduction from 59% to 32%.3 Subsequently, ECMO has been a common medical therapy in neonatal ICUs around the globe for specific conditions including meconium aspiration syndrome, primary pulmonary hypertension of the newborn, myocarditis, congenital diaphragmatic hernias, and other reversible lung injuries. The use of ECMO in the aforementioned pathologies has yielded survival rates as high as 80%.4,5 Similar data have supported the use of ECMO in the pediatric population, with survival rates as high as 73% in pediatric patients suffering from respiratory failure.6 The data for cardiac failure in both the neonate and pediatric populations result in a much lower rate of survival than for lung disease with a quoted rate ranging from 38% to 43%. However, recently a very large retrospective review of ECMO for myocarditis in infants, children, and young adults with 255 patients found a 61% survival at discharge.7

Unfortunately, the efficacy and safety of ECMO for adults has been less clear and has been debated in the literature. ECMO has been subject to a wide range of opinions in the medical literature since the 1970s and has shown that use in adults with ARDS is invasive, expensive, and without any morbidity improvement as compared with mechanical ventilation.8 In 2004, one large study of 255 adults with severe ARDS who received ECMO revealed a 52% survival rate.9 The first multicenter randomized controlled trial in ARDS looking at ECMO versus conventional mechanical ventilation was evaluated in a recently published trial. Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR) trial.10 A total of 180 patients with ARDS were enrolled; they were either randomized to a tertiary care center or transferred and managed at a single ECMO center. Of the 90 patients who were due to transfer to the ECMO center, only 68 patients actually received ECMO therapy; there was an overall survival at 6 months of 63% for the ECMO group and 47% for the conventional group. This trial has, appropriately, come under scrutiny due to the lack of standardized treatment management in the control group and because the only ECMO center observed happened to be one of the most experienced in the world. One can conclude that early transfer to a specialized ECMO center might increase survivability but well-randomized controlled trials still need to be conducted.

Interestingly, due to the arrival of the H1N1 influenza epidemic, there has been a resurgence of small trials and case reports favoring the use of ECMO in adults11 (a large observational study of 68 patients suffering with the 2009 influenza [H1N1]–associated ARDS, who received ECMO therapy in Australia and New Zealand). The ECMO group had severe hypoxemia that was defined by a mean Pao2/FiO2 ratio of less than 60. The ECMO group had originally posted a survival rate of 79% but, in an update of their data, the authors quoted a survival rate of 75% at time of discharge.12

ECMO may have a role in certain disease processes, but it is not considered the standard of care. Recent studies suggest a mortality benefit when used early if proven ventilation techniques have failed. ECMO is best left to centers that are specialized with ideal resources.13

ECMO can be utilized for mechanical assistance during pulmonary or cardiac failure in newborn infants, children, or adults.

Neonatal

ECMO, as described above, is a routine part of neonatal care used in newborn infants for severe respiratory failure commonly associated with primary pulmonary hypertension, meconium aspiration syndrome, congenital diaphragmatic hernia, respiratory distress syndrome, group B streptococcal sepsis, and asphyxia.

Pediatric

Pediatric ECMO is used for respiratory distress syndrome and in low cardiac output situations such as right, left, or biventricular failure following repair of congenital heart defects and with pulmonary vasoreactive crises that can occur following these surgeries. Sometimes ECMO is used as a bridge to cardiac transplant or as a bridge to recovery in temporary cardiomyopathy secondary to renal failure, myocarditis, and burns.

Adult

Common etiologies of adult pulmonary and cardiac failure requiring ECMO support include:

1. Respiratory failure, characterized by the severity of hypoxemia or impaired ventilation:

   1. Adult respiratory distress syndrome due to:

      • Pneumonia—viral, bacterial, and aspiration
      • Sepsis
      • Multisystem trauma
      • Pulmonary contusion
      • Pancreatitis
      • Disseminated intravascular coagulation (DIC)
      • Intestinal infarction
      • Vasculitis

2. Cardiac failure:

   1. Acute cardiomyopathy

   2. Massive pulmonary embolus

   3. Congenital heart disease

   4. Right ventricular failure

   5. Biventricular failure

   6. “Bridge” to transplant

Unfortunately, patients who are sick enough to require ECMO but do not have access to this therapy often have a mortality rate approaching near 100% despite maximizing all other forms of available medical treatment. Early referral to a specialized ECMO center may provide a survival advantage in the adult.

Patient selection criteria for ECMO vary from center to center. The usual criteria include patients with a severe reversible process that would otherwise result in a very high predicted mortality with conventional medical support. See Table 7-1 for indications and Table 7-2 for contraindications.

ECMO is simply the use of a modified heart–lung bypass machine to provide gas exchange and systemic perfusion. The technique can be performed one of the following two ways: by venovenous (V-V) (Figure 7-1) or venoarterial (V-A) (Figure 7-2) bypass. V-A ECMO can support respiratory and cardiac function. A large (23–30 French) catheter is inserted into the inferior vena cava via the femoral vein or the right atrium via the right internal jugular vein. This is the drainage catheter. The ECMO circuit requires high flow, typically 100 mL/kg/min requiring placement of such large-diameter catheters.

The ECMO catheters, or cannulas as they are also called, are typically placed by direct cutdown on vessels or by percutaneous introduction of the cannula using sequential dilators and a guidewire similar to the technique used in standard central line placement.

A second catheter, from which blood is returned to the patient's arterial circulation, returns blood to the descending aorta via a cannula placed in the femoral artery or the aortic arch via a cannula placed in the right carotid artery. Since blood flow through the patient's native heart and lungs is diverted to the ECMO circuit, the patient's cardiac output is controlled by the amount of blood that travels through the circuit. In V-A ECMO, the ECMO flow determines the cardiac output and oxygen delivery. Thus, V-A ECMO can provide pulmonary and cardiac support.

V-V bypass is physiologically different from V-A ECMO. Whereas V-A ECMO supports both respiratory and cardiac function, V-V ECMO supports only respiratory function. V-V ECMO diverts the patient's blood from the venous circulation and returns it to the venous circulation. It utilizes two venous catheters: A drainage catheter is placed in the right atrium via the right internal jugular vein in the same manner as performed for V-A ECMO and a second cannula, the return venous catheter, is placed in the femoral vein. Additionally, a double-lumen cannula inserted into the right internal jugular vein and guided to the right atrium may be utilized; however, these catheters are smaller in size and may not be adequate for adult patients (Figure 7-3).

When catheter access has been obtained, the patient is connected to the ECMO circuit. Often, the circuit may be primed with blood to avoid hypotension from acute changes in hemoglobin that can occur with crystalloid primed circuits. Venous blood is removed from the patient via the drainage catheter and pumped through an artificial lung called a membrane oxygenator. It is in the oxygenator that oxygen diffusion occurs because of a pressure gradient between the partial pressure of oxygen in the patient's venous blood being pumped through the circuit and the partial pressure of oxygen perfusing the membrane oxygenator.

Venous saturation is the monitor used to assess the adequacy of oxygen delivery from the ECMO circuit. It is usually kept between 70% and 75%. This is accomplished by titrating the ECMO circuit's pump flow rate. Increasing the flow increases the oxygen delivery and directly affects the venous saturation.

Carbon dioxide diffusion across the membrane is a function of the gradient from the patient's blood to the ECMO circuit's gas, that is, the gas ventilating the membrane oxygenator. The PaCo2 is isolated by titrating the amount of gas used to ventilate the oxygenator (sweep gas). Increasing the sweep will decrease CO2 removal, while decreasing the sweep will increase the CO2.

After having been pumped through the oxygenator, the oxygenated blood, under pressure, is pumped through a heat exchanger that maintains the patient's body temperature at a set desired temperature, usually 37.0°C. The blood is pumped either to the arterial circulation via the aorta in V-A ECMO or back into the venous circulation via the right atrium in V-V ECMO.

Currently ECMO circuits require the use of systemic anticoagulation with heparin to keep the system patent. The surfaces of the ECMO circuit and devices are plastic and therefore thrombogenic. It is necessary to provide anticoagulant prophylaxis to the patient's blood with a continuous infusion of heparin. The level of anticoagulation is measured by whole-blood activated clotting times (ACT) performed at the patient's bedside by the ECMO specialist. The ACT is usually maintained at approximately 180–240 seconds.

During ECMO therapy, the patient's ventilator settings are gradually weaned to allow lung “rest,” keeping peak inspiratory pressures around 20 mm Hg. Patients are diuresed to dry weight. Hemoglobin levels are kept above 10 g/dL and platelet counts above 100,000/mL.

For summary of differences between V-A and V-V ECMO, see Table 7-3.

The average adult ECMO course can vary from days to weeks. During the first 24–48 hours, the condition of a patient's lungs will likely worsen as evidenced by increased radiographic opacification, which is thought to be due to the sudden decreasing airway pressure in response to diversion of pulmonary flow by ECMO. Additionally, it is thought that various vasoreactive substances are released and activated from the patient's blood reacting to the ECMO circuit surface. Improvement of lung function and compliance usually begin to occur within 1–3 days.

As lung function improves, the patients are weaned from ECMO by decreasing the circuit flow. When evidence of improved lung compliance and adequate gas exchange without excessive ventilatory support has been demonstrated, a “trial of” ECMO is performed. Good indicators of lung recovery include improving chest radiographs, increasing lung compliance, increasing PaO2, or decreasing PacO2 on resting ventilator settings. Eventually the patient is removed from ECMO support and the catheters are removed.

As with any invasive procedure, ECMO has many potential life-threatening complications that can occur. These can be categorized into mechanical complications and patient complications (Tables 7-4 and 7-5). Mechanical complications are related to cannula placement and the ECMO circuit itself. Patient complications may be attributed to physiologic complications that occur due to ECMO therapy.

Mechanical Complications

The required placement of large-bore ECMO cannulas can cause several complications. As with placement of any type of central line, pneumothorax, line infection, and bleeding may occur. In addition, due to the larger size of the cannula required with ECMO, direct damage to the internal jugular vein can cause massive mediastinal bleeding. Cannulation of the carotid artery can cause dissection of the carotid arterial intima, leading to aortic dissection. In addition, any potential for bleeding from the placement of the cannulas is increased due to the requirement of systemic heparinization to maintain the ECMO circuit. The cannulas may also serve as a nidus for thrombus formation and emboli.

The ECMO circuit has potential to cause numerous complications. Because the surfaces of the ECMO circuit and devices are plastic, it is necessary to provide anticoagulant prophylaxis to the patient's blood with a continuous infusion of heparin. The most common mechanical complication is the development of clots within the circuit. These develop due to the platelets adhering to the plastic surface of the circuit, becoming activated, recruiting more platelets, and growing into platelet aggregates. Eventually these platelet aggregates break off. These clots can cause failure of the ECMO circuit's oxygenator. Larger clots can cause pulmonary or systemic emboli. Thrombocytopenia and a consumptive coagulopathy may also occur due to a large clot burden in the circuit.

Air can enter the ECMO circuit from dislodgement of a cannula, which then sucks in air, a small tear in the membrane oxygenator, compromised integrity in any of the connections in the circuit tubing, or high partial pressure of oxygen in the blood. Small bubbles in the circuit can be easily removed and have low potential for harm. A large bolus of air can be fatal.

Malfunction of the circuit heat exchanger can lead to significant patient hypothermia that may cause or exacerbate any coagulopathy that already exists.

Patient Complications

Patients undergoing ECMO therapy can suffer complications in any organ system. Many of these complications are due to the need for systemic anticoagulation.

Neurologically, patients may have spontaneous intracranial bleeding due to anticoagulation. This is more commonly seen in the neonatal ECMO population. Infarction from emboli may occur, and seizures induced by bleeding, infarction, or hypoxemia are also a threat.

Hemolysis from clot development typically manifests itself as renal dysfunction and rising serum haptoglobin levels. Coagulopathy and thrombocytopenia occur due to platelet consumption from activation by the circuit's plastic surface. Moreover, a dilutional coagulopathy can occur. Hemorrhage at any surgical site or cannula site, or into the site of previous invasive procedures is a frequent complication because of the required systemic heparinization. Intrathoracic, abdominal, or retroperitoneal hemorrhage may also occur. Exsanguination from circuit disruption, while uncommon, can be fatal.

Pericardial tamponade can occur due to cannula placement in the face of systemic anticoagulation. Myocardial stunning, which is defined as a decrease of left ventricular ejection fraction by more than 25% with the initiation of ECMO, may occur, requiring further V-A-ECMO support or vasopressor and inotropic support. Fortunately, stunning is a temporary effect and cardiac ejection function returns to normal within 48 hours of ECMO initiation.

Pulmonary hemorrhage, and spontaneous and iatrogenic pneumothorax may occur, as well. Oliguria is common during early ECMO therapy, and acute tubular necrosis and renal failure may occur from hemolysis, hypovolemia, or decreased perfusion.

Gastrointestinal hemorrhage may occur due to physiologic stress response, ischemia, embolic, or systemic anticoagulation. Elevated direct bilirubin and the development of biliary calculi occur secondary to prolonged fasting, use of parenteral nutrition, hemolysis, and diuretics.

Last, due to the ECMO circuit functioning as a large intravascular foreign body, numerous metabolic complications of either acidosis or alkalosis in response to any electrolyte disturbances may develop.

Due to the highly invasive nature of ECMO and the potential for numerous complications, a trained ECMO technician is usually present at all times, 24 hours per day, at the patient's bedside to monitor the circuit and the patient for potential complications. This is in addition to the patient's usual nursing personnel.

ECMO is an effective, cutting-edge technology capable of providing pulmonary and cardiac life support in patients with severe respiratory failure. Often these patients have a high mortality risk despite optimal conventional medical care. While ECMO treatment is not without its risks and complications, it is a reasonable treatment modality that will allow for the recovery of injured native lung and improved survival in a patient population with a previously poor predicted outcome.