Extracorporeal membrane oxygenation (ECMO) in adults

shenxiu2

Extracorporeal membrane oxygenation (ECMO) in adults
Authors Jonathan Haft, MD Robert Bartlett, MD
|

INTRODUCTION

— Mechanical cardiopulmonary support is most often applied intraoperatively to facilitate cardiac surgery (ie, cardiopulmonary bypass). However, cardiopulmonary support can also be delivered in a more prolonged fashion in an intensive care unit, although it is less common.

Prolonged cardiopulmonary support is called extracorporeal membrane oxygenation (ECMO), extracorporeal life support, or extracorporeal lung assist. There are two types of ECMO — venoarterial (VA) and venovenous (VV). Both provide respiratory support, but only VA ECMO provides hemodynamic support.

The impact of ECMO on clinical outcomes as well as patient selection, technical aspects, and complications will be reviewed here.

OUTCOMES — Clinical outcomes in patients undergoing ECMO can be categorized according to the indication for the ECMO: severe acute respiratory failure or cardiac failure.

Acute respiratory failure — The evidence indicates that ECMO may improve survival in patients with severe acute respiratory failure. Initial observational studies and uncontrolled clinical trials of patients with severe acute respiratory failure reported survival rates from 50 to 71 percent among patients who received ECMO . These survival rates were better than historical survival rates.

The best data come from the Conventional ventilatory support versus Extracorporeal membrane oxygenation for Severe Acute Respiratory failure (CESAR) trial in which 180 patients with severe, but potentially reversible, acute respiratory failure were randomly assigned to either undergo conventional management or be evaluated for ECMO . The group evaluated for ECMO had significantly increased survival without disability at six months compared to conventional management (63 versus 47 percent).

The trial had two limitations with conflicting impact on the results that need to be considered when deciding how to apply the trial to clinical practice:

• Among the patients evaluated for ECMO, only 75 percent actually received the intervention. Thus, if ECMO is truly beneficial, the trial may have underestimated the magnitude of the benefit.
• Patients who were evaluated for ECMO were more likely to receive interventions with proven benefit, such as low tidal volume ventilation. This provides an alternative explanation for the benefit seen among patients who were evaluated for ECMO.

Severe acute respiratory failure was defined in the CESAR trial as hypercapnic respiratory acidosis with an arterial pH <7.20 or a Murray score greater than 3.0. The Murray score quantitates the severity of lung disease on the basis of the ratio of arterial oxygen tension to the fraction of inspired oxygen (PaO2/FiO2), positive end-expiratory pressure (PEEP), lung compliance, and chest radiograph. Important exclusion criteria included an age <18 years or >65 years, as well as contraindications to anticoagulation.

The trial was too small to use subgroup analyses to identify predictors of outcome, such as etiology of the respiratory failure, age, comorbidities, and duration of mechanical ventilation.

Cardiac failure — The use of ECMO for cardiac failure has been less extensively studied than ECMO for severe acute respiratory failure. Observational studies and case series have reported survival rates of 20 to 43 percent among patients who received venoarterial (VA) ECMO for cardiac arrest, severe cardiogenic shock, or failure to wean from cardiopulmonary bypass following cardiac surgery . VA ECMO has also been used as a bridge to cardiac transplantation or placement of a ventricular assist device.

Long-term survivors of ECMO performed for cardiogenic shock appear to have worse general health, physical health, and social functioning than healthy controls . However, they perform better on these same measures when compared to patients who require chronic hemodialysis, have advanced heart failure, or have recovered from ARDS.

PATIENT SELECTION

Indications — Criteria for the initiation of ECMO vary widely from institution to institution. In our practice, we initiate ECMO as temporary life support for patients with potentially reversible severe acute respiratory failure or cardiac failure . Examples of clinical situations that may prompt us to begin ECMO include the following:

• Hypoxemic respiratory failure with a ratio of arterial oxygen tension to fraction of inspired oxygen (PaO2/FiO2) of <100 mmHg despite optimization of the ventilator settings, including the tidal volume, positive end-expiratory pressure (PEEP), and inspiratory to expiratory (I:E) ratio
• Hypercapnic respiratory failure with an arterial pH less than 7.20
• Refractory cardiogenic shock
• Cardiac arrest
• Failure to wean from cardiopulmonary bypass after cardiac surgery
• As a bridge to either cardiac transplantation or placement of a ventricular assist device

Relative contraindications — ECMO may not be initiated if anticoagulation is contraindicated (eg, bleeding, recent surgery, recent intracranial injury), if the cause of the respiratory or cardiac failure is irreversible, or if the patient is not a potential candidate for an implantable ventricular assist device.

Additional considerations include the following:

• For patients with respiratory failure, ECMO might not be initiated if the patient has been mechanically ventilated for longer than seven days because outcomes may be poor in this population .
• For patients with cardiac failure, ECMO might not be initiated if a ventricular assist device or transplantation is contraindicated (eg, the patient has preexisting renal failure, preexisting hepatic failure, significant aortic valve insufficiency, or inadequate social support).

Other characteristics that may exclude some patients from receiving ECMO include advanced age, morbid obesity, neurologic dysfunction, or poor preexisting functional status.

TECHNIQUE

— ECMO should only be performed by clinicians with training and experience in its initiation, maintenance, and discontinuation.

During ECMO, a large volume of blood is extracted from the native vascular system and circulated outside the body by a mechanical pump. While outside the body, the blood passes through an oxygenator and heat exchanger (figure 1). In the oxygenator, hemoglobin becomes fully saturated with oxygen, while carbon dioxide (CO2) is removed. Elimination of CO2 is efficient and can be controlled by adjusting the rate of countercurrent gas flow through the oxygenator. Finally, the blood is reinfused into the native vascular system.

ECMO can be venovenous (VV) or venoarterial (VA). During VV ECMO, blood is extracted from a large central vein and returned to the venous circulation (figure 2). VV ECMO cannot be used for cardiac failure, because it provides respiratory support without hemodynamic support.

During VA ECMO, blood is extracted from a large central vein and returned to the arterial system, bypassing the heart and lungs (figure 3). VA ECMO can be used for respiratory or cardiac failure because it provides both respiratory and hemodynamic support. The additional benefit comes with additional risks, which are discussed below. (See 'VA ECMO-specific complications' below.)

Initiation — Once it has been decided that ECMO will be initiated, the patient should be anticoagulated with intravenous heparin and then the cannulae should be inserted. ECMO support can be initiated once the cannulae are connected to the appropriate limbs of the ECMO circuit.

Cannulation — The surgeon should select the largest cannula that can be inserted safely. For VV ECMO, venous cannulae are usually placed in the right common femoral vein (for drainage) and right internal jugular vein (for infusion). The tip of the femoral cannula should be maintained near the junction of the inferior vena cava and right atrium, while the tip of the internal jugular cannula should be maintained near the junction of the superior vena cava and right atrium. Alternatively, a double lumen cannula is available that is large enough to accommodate 4 to 5 L/min of blood flow . It is available in a variety of sizes, with 31 French being the largest and most appropriate for adult males. The drainage and infusion ports have been engineered to minimize recirculation.

For VA ECMO, a venous cannula is usually placed in the right common femoral vein (for extraction) and an arterial cannula is usually placed into the right femoral artery (for infusion). The tip of the femoral venous cannula should be maintained near the junction of the inferior vena cava and right atrium, while the tip of the femoral arterial cannula is maintained in the iliac artery.

Femoral access is preferred for VA ECMO because insertion is relatively easy. This is important because most patients are at high risk for bleeding since they are anticoagulated. The main drawback of femoral access is ischemia of the ipsilateral lower extremity. The likelihood of this complication can be decreased by inserting an additional arterial cannula distal to the femoral artery cannula and redirecting a portion of the infused blood to the additional cannula for "reperfusion" of the extremity. Alternatively, a cannula can be inserted percutaneously into the superficial femoral artery for antegrade flow to the extremity . This technique requires additional imaging.

Occasionally, the femoral vessels are unsuitable for cannulation for VA ECMO (eg, patients with severe occlusive peripheral arterial disease or prior femoral arterial reconstruction). In such circumstances, the right common carotid artery or axillary artery can be used. In our experience, there is a 5 to 10 percent risk of a large watershed cerebral infarction when the right common carotid artery is used. Use of the right axillary artery offers the advantage of allowing patients on ECMO to ambulate .

Cannulation is unnecessary for patients who require postcardiotomy ECMO. The intrathoracic cannulae employed for cardiopulmonary bypass can be transferred from the heart lung machine to the ECMO circuit, with blood extracted from the right atrium and reinfused into the ascending aorta.

Titration — Following cannulation, the patient is connected to the ECMO circuit and the blood flow is increased until respiratory and hemodynamic parameters are satisfactory. Reasonable targets include:

• A central venous oxyhemoglobin saturation (ScvO2, drawn from a central venous catheter) or a mixed venous oxyhemoglobin saturation (SvO2, drawn from a pulmonary artery catheter) of 75 to 80 percent

• An arterial oxyhemoglobin saturation of 100 percent for VA ECMO, or 85 to 100 percent for VV ECMO.

• Adequate tissue perfusion, as determined by the arterial blood pressure, venous oxygen saturation, and blood lactate level

Maintenance — Once the initial respiratory and hemodynamic goals have been achieved, the blood flow is maintained at that rate. Frequent assessment and adjustments are facilitated by continuous venous oximetry, which directly measures the oxyhemoglobin saturation of the blood in the venous limb of the ECMO circuit. Continuous venous oximetry does not measure the ScvO2 or SvO2, but it is a reasonable surrogate for these values. When the target ScvO2 or SvO2 are insufficient, interventions that may be helpful include increasing one or more of the following: blood flow, inotropy, intravascular volume, or hemoglobin concentration.

Anticoagulation is sustained during ECMO with a continuous infusion of unfractionated heparin, titrated to an activated clotting time (ACT) of 210 to 230 seconds. The ACT target is decreased if bleeding develops. ACT is the preferred measurement because it is easily determined at the point of care.

Platelets are continuously consumed because they are activated from both the sheer forces of extracorporeal flow and exposure to the foreign surface area. Platelet counts should be maintained greater than 100,000/microL, which may require several transfusions daily.

Ventilator settings are reduced during ECMO in order to avoid barotrauma, volutrauma (ie, ventilator-induced lung injury), and oxygen toxicity. Plateau airway pressures should be maintained less than 30 cm H2O and FiO2 less than 0.5. Reduction of ventilator support is usually accompanied by increased venous return and cardiac output, which may permit vasopressors to be weaned.

We perform early tracheostomy to reduce dead space and improve patient comfort. Patients typically require light sedation during ECMO.

Special considerations

— VV ECMO is typically used for respiratory failure, while VA ECMO is used for cardiac failure. There are unique considerations for each type of ECMO, which influence management.

• Blood flow — Near maximum flow rates are usually desired during VV ECMO to optimize oxygen delivery. In contrast, the flow rate used during VA ECMO must be high enough to provide adequate perfusion pressure and ScvO2 or SvO2, but low enough to provide sufficient preload to maintain left ventricular output.

• Diuresis — Since most patients are fluid overloaded when ECMO is initiated, aggressive diuresis is warranted once the patient is stable on VV ECMO. Ultrafiltration can be easily added to the ECMO circuit if patients are unable to produce sufficient urine for diuresis. In contrast, aggressive diuresis is often delayed during VA ECMO because preservation of renal function is essential to maintain candidacy for cardiac transplantation or ventricular assist device implantation.

• Left ventricular monitoring — Left ventricular output must be more rigorously monitored during VA ECMO because left ventricular output may worsen. The cause is usually multifactorial, including the underlying left ventricular dysfunction, increased afterload due to retrograde VA ECMO flow, and insufficient unloading of the distended left ventricle due to ongoing blood flow to the left ventricle from the bronchial circulation and right ventricle. Progressive impairment of left ventricular output results in left ventricular distension, left atrial hypertension, and pulmonary hemorrhage.

Left ventricular output can be closely monitored by identifying pulsatility in the arterial line's waveform and by frequent echocardiography. Interventions that can improve left ventricular output include inotropes (eg, dobutamine, milrinone) to increase contractility and intra-aortic balloon counterpulsation to reduce afterload and facilitate left ventricular output.

Immediate left ventricular decompression is essential to avoid pulmonary hemorrhage if left ventricular ejection cannot be maintained despite intra-aortic balloon counterpulsation and inotropic agents. This can be accomplished surgically or percutaneously. Methods of percutaneous left ventricular decompression include transatrial balloon septostomy and insertion of a left ventricular drainage catheter (figure 4).

Discontinuation — A patient's readiness for discontinuation of ECMO should be evaluated frequently. For patients with respiratory failure, improvements in radiographic appearance, pulmonary compliance, and arterial oxyhemoglobin saturation indicate that the patient may be ready to be liberated from ECMO. For patients with cardiac failure, enhanced aortic pulsatility correlates with improved left ventricular output and indicates that the patient may be ready to be liberated from ECMO.

One or more trials of taking the patient off ECMO should be performed prior to discontinuing ECMO permanently.

• VV ECMO trials are performed by eliminating all countercurrent sweep gas through the oxygenator. Extracorporeal blood flow remains constant, but gas transfer does not occur. Patients are observed for several hours, during which the ventilator settings that are necessary to maintain adequate oxygenation and ventilation off ECMO are determined.
• VA ECMO trials require temporary clamping of both the drainage and infusion lines, while allowing the ECMO circuit to circulate through a bridge between the arterial and venous limbs. This prevents thrombosis of stagnant blood within the ECMO circuit. In addition, the arterial and venous lines should be flushed continuously with heparinized saline or intermittently with heparinized blood from the circuit.

VA ECMO trials are generally shorter in duration than VV ECMO trials because of the higher risk of thrombus formation. Once the decision has been made to discontinue ECMO, the cannulae are removed. Hemostasis is achieved by compressing the insertion site. For patients who received VA ECMO:

• At least thirty minutes of compression is required for the arterial site. With proper technique and patience, retroperitoneal bleeding or pseudoaneurysm formation are infrequent. Low doses of protamine can be administered to reverse the remaining systemic heparin and facilitate hemostasis.
• The reperfusion cannula placed in the posterior tibial artery must be removed. This requires that the wound be opened at the bedside and the proximal posterior tibial artery ligated as the cannula is removed. The artery can be reconstructed, but this is rarely performed because the likelihood of long term patency is poor for patients supported for extended periods of time and the absence of ischemia of the distal extremity during VA ECMO confirms its viability despite occlusion of the artery.

COMPLICATIONS

Bleeding — Bleeding is frequent and can be life threatening. It is due to both the necessary continuous unfractionated heparin infusion and platelet dysfunction. The latter results from contact and sheer stress associated activation. Meticulous surgical technique, maintaining platelet counts greater than 100,000/mm3, and maintaining the target ACT appear to reduce the likelihood of bleeding.

Intervention is usually necessary when bleeding occurs. Bleeding from surgical wounds often requires prompt exploration with liberal use of electrocautery. Hemorrhage into body cavities (eg, abdomen, pleural space) may require surgical exploration to achieve hemostasis, after which vacuum-assisted closure is recommended because it allows removal and measurement of the blood. Plasminogen inhibitors (eg, aminocaproic acid) can be infused or heparin can be discontinued for several hours, but these actions may increase the risk of circuit thrombosis . Infusion of activated factor VII has been reported with mixed results and should only be considered for life threatening hemorrhage after all other options have failed .

The target ACT is usually reduced once bleeding occurs. As an example, the target ACT may become 170 to 190 seconds, instead of 210 to 230 seconds.

Thromboembolism — Thromboembolism due to thrombus formation within the extracorporeal circuit is an infrequent complication that can be devastating. Its impact is greater with VA ECMO than VV ECMO because infusion is into the systemic circulation. Heparin infusion that achieves its target ACT and vigilant observation of the circuit for signs of clot formation successfully prevents thromboembolism in most patients.

Observation of the circuit for signs of clot formation includes routine inspection of all connectors and monitoring the pressure gradient across the oxygenator. A sudden change in the pressure gradient suggests that a thrombus had developed. Large or mobile clots require immediate circuit or component exchange. Primed circuits are usually kept at the bedside if the target ACT has been reduced due to bleeding because the risk of thrombus formation is greatest in this situation. Having a primed circuit available facilitates urgent exchange, if necessary.

Cannulation-related — A variety of complications can occur during cannulation, including vessel perforation with hemorrhage, arterial dissection, distal ischemia, and incorrect location (eg, venous cannula within the artery). A skilled and experienced surgeon is important to avoid or address such complications.

Heparin-induced thrombocytopenia — Heparin-induced thrombocytopenia (HIT) is increasingly common among patients receiving ECMO. When HIT is suspected, the heparin infusion should be replaced by a non-heparin anticoagulant . We favor argatroban because its half life is short and a similar ACT target range is effective. (See "Heparin-induced thrombocytopenia".)

VA ECMO-specific complications

• Pulmonary hemorrhage — Pulmonary hemorrhage can occur in patients who receive VA ECMO. This complication was described above. (See 'Special considerations' above.)
• Pulmonary infarction — Pulmonary infarction can complicate VA ECMO because most pulmonary blood flow bypasses the lungs. However, this is rarely seen because pulmonary tissue has a low metabolic rate and bronchial flow is preserved during VA ECMO.
• Aortic thrombosis — There is retrograde blood flow in the ascending aorta whenever the femoral artery and vein are used for VA ECMO. Stasis of the blood can occur if left ventricular output is not maintained, which can result in thrombosis (figure 5).
• Coronary or cerebral hypoxia — During VA ECMO, fully saturated blood infused into the circulation from the ECMO circuit will preferentially perfuse the lower extremities and the abdominal viscera. Blood ejected from the heart will selectively perfuse the heart, brain, and upper extremities. As a result, the oxyhemoglobin saturation of the blood perfusing the lower extremities and abdominal viscera may be substantially higher than that perfusing the heart, brain, and upper extremities. Cardiac and cerebral hypoxia could exist and be unrecognized if oxygenation is monitored using blood from the lower extremity.

To avoid this complication, arterial oxyhemoglobin saturation should be monitored in both the upper extremity (eg, blood gases from the radial artery or pulse oximetry on a finger) and the lower extremity (eg, blood gases from the femoral artery). Poor arterial oxyhemoglobin saturations measured from the upper extremity should prompt adjustments to the mechanical ventilator that optimize pulmonary oxygenation or augmentation of blood flow through the ECMO circuit.

FUTURE

— Applications for ECMO may expand in the future to include percutaneous temporary left ventricular assistance and arteriovenous carbon dioxide removal without a pump . In addition, new technologies are expected to greatly impact ECMO in the future, including new oxygenators, pumps, and surface coatings.

New oxygenators that use hollow fibers constructed of polymethyl-pentene are available for short term use . Advantages include lower priming volume, rapid priming time, diminished plasma leakage, and low blood flow resistance, which may reduce platelet activation and consumption. The silicone oxygenators will likely become obsolete if these oxygenators prove durable and safe.

Most ECMO centers depend upon servoregulated (ie, automatic) roller pumps for blood flow generation, which require extensive priming, extensive tubing length, and continuous observation by trained personnel. Centrifugal pumps without servoregulation have replaced roller pumps in some centers. However, these pumps can generate heat, inducing microcavitation and hemolysis. Newer centrifugal pumps have hydrodynamically or magnetically levitated rotors, which reduce heat generation and microcavitation . Future ECMO circuits will likely incorporate these pumps, thereby facilitating miniaturization, reducing priming volume, and decreasing the need for continuous observation.

Surface coatings that mimic the endothelial lining of blood vessels and reduce blood cell activation are being developed
. In an ECMO circuit, such coatings could reduce thrombogenicity, obviate the need for continuous anticoagulation, and reduce the incidence of related complications.

SUMMARY AND RECOMMENDATIONS

• Extracorporeal membrane oxygenation (ECMO) is a type of prolonged mechanical cardiopulmonary support that is usually delivered in the intensive care unit. (See 'Introduction' above.)
• ECMO may improve survival in patients with severe acute respiratory failure. However, its impact on patients with cardiac failure is uncertain. (See 'Outcomes' above.)
• We suggest that patients with severe, but potentially reversible, acute respiratory failure be evaluated for ECMO if it is available within the medical center (Grade 2B). For patients who are in a medical center that does not provide ECMO, the decision to transfer a patient to another medical center to be evaluated for ECMO must be made on a case-by-case basis, since there are substantial risks associated with transferring a patient who has severe hypoxemic respiratory failure. (See 'Indications' above.)
• For patients with cardiac failure, the decision to evaluate a patient for ECMO should be made on a case-by-case basis. Types of cardiac failure for which ECMO has been used include cardiac arrest, severe cardiogenic shock, and failure to wean from cardiopulmonary bypass after cardiac surgery. (See 'Indications' above.)
• There are two types of ECMO, venoarterial (VA) and venovenous (VV). VV ECMO is used in patients with respiratory failure, while VA ECMO is used in patients with cardiac failure. (See 'Technique' above.)
• Once it has been determined that ECMO will be initiated, the patient should be anticoagulated with intravenous unfractionated heparin. Cannulae are then inserted and the patient is connected to the ECMO circuit. The blood flow is increased until respiratory and hemodynamic parameters are satisfactory. Once the initial respiratory and hemodynamic goals have been achieved, blood flow is maintained. Frequent reassessment and adjustments are usually necessary. (See 'Initiation' above and 'Maintenance' above.)
• The patient's readiness for discontinuation of ECMO should be evaluated frequently. Prior to discontinuing ECMO permanently, one or more trials should be performed during which the patient is off ECMO.

Such trials give the clinician to opportunity to determine whether conventional supportive care is sufficient for the patient. (See 'Discontinuation' above.)

• Complications are frequent during ECMO. They include bleeding, thromboembolism, heparin-induced thrombocytopenia, and cannula-related complications. VA ECMO-specific complications include pulmonary hemorrhage, pulmonary infarction, aortic thrombosis, cardiac hypoxia, or cerebral hypoxia. (See 'Complications' above.)

摘自"与时并进"

zeng1zhan

Such a good  Anesthesia Article