Perioperative Extracorporeal Cardiopulmonary Resuscitation in Adult Patients

Extracorporeal CPR is the rapid deployment of venoarterial extracorporeal membrane oxygenation (ECMO) in patients who have not achieved sustained return of spontaneous circulation with conventional CPR. While various approaches to extracorporeal CPR have been explored since the 1960s recent growth in its use may be attributed, in part, to continued poor survival after cardiac arrest, increasing familiarity with ECMO, and results from observational studies showing the potential benefit of extracorporeal CPR.  In venoarterial ECMO, venous blood is drained into an extracorporeal circuit through a venous (inflow) cannula, pumped through a membrane oxygenator where gas exchange (oxygenation and carbon dioxide removal) occurs, and returned to the arterial system through an arterial (outflow) cannula (fig. 1). Venoarterial ECMO can deliver near-normal levels of end-organ perfusion.  Its implementation in extracorporeal CPR provides additional time for clinicians to identify and treat the underlying etiology of cardiac arrest while stabilizing the patient’s hemodynamics and minimizing end-organ injury. Venoarterial ECMO support may also allow bridging to therapies such as durable left ventricular assist device implantation or heart transplantation.

Venoarterial ECMO can be implemented either centrally or peripherally.  Peripheral venoarterial ECMO is typically used for extracorporeal CPR because it can be established expeditiously, either percutaneously or by surgical cutdown. A multistage venous inflow cannula (21 to 25 French) is placed in a femoral vein and advanced through the inferior vena cava into the right atrium, while an outflow cannula (15 to 17 French) is placed in a common femoral artery and advanced into the common iliac artery or the abdominal aorta. Central venoarterial ECMO, on the other hand, is typically established in the operating room in patients who cannot be weaned from cardiopulmonary bypass. In rare cases, extracorporeal CPR can be implemented rapidly through a de novo sternotomy with right atrial and central aortic cannulation.

Several recent randomized controlled trials evaluating the use of extracorporeal CPR for out-of-hospital cardiac arrests have been published. In the Advanced Reperfusion Strategies for Refractory Cardiac Arrest (ARREST) single-center trial, 30 patients presenting with out-of-hospital cardiac arrest and initial rhythm of ventricular fibrillation or pulseless ventricular tachycardia were randomized to initial standard advanced cardiovascular life support (ACLS) or ECMO-facilitated resuscitation at a specialized ECMO center. Survival to hospital discharge was greater in the ECMO-facilitated group (43%; 95% CI, 21.3 to 67.7) compared to the standard ACLS group (7%; 95% CI, 1.6 to 30.2). The cumulative 6-month survival was also superior in the ECMO-facilitated group. However, the results of this trial should be interpreted with caution given its small size and the results of subsequent larger trials described below.

In a second single-center trial conducted in Prague, 256 patients presenting with out-of-hospital cardiac arrest were randomized to extracorporeal CPR versus standard ACLS. Patients randomized to extracorporeal CPR did not experience a significantly higher survival rate with good neurologic outcome at 180 days compared to those randomized to standard resuscitation (odds ratio, 1.63; 95% CI, 0.93 to 2.85; P = 0.09). However, this study was terminated by the data safety monitoring board before achieving full enrollment. It is possible that this trial was underpowered to show clinical benefit in the extracorporeal CPR group given the substantial effect size and the wide CI. In addition, the survival rate for patients in the “control” group was higher than what has traditionally been reported for patients with conventional CPR. A third pragmatic multicenter study (INCEPTION [Early Initiation of Extracorporeal Life Support in Refractory Out-of-Hospital Cardiac Arrest] trial), conducted in the Netherlands, randomized 134 patients presenting with out-of-hospital cardiac arrest to receive either extracorporeal CPR or conventional CPR. In this study, patients in both groups experienced similar rates of survival and good neurologic outcomes at 30 days (odds ratio, 1.4; 95% CI, 0.5 to 3.5; P = 0.52).  However, the INCEPTION trial had longer times from arrest onset to venoarterial ECMO initiation and a higher rate of cannulation failure than the other two trials. Also, cannulations were primarily performed in the emergency department compared to in the cardiac catheterization lab. There was no reporting on the ECMO experience of the cannulating physicians.

Although randomized controlled trials evaluating the use of extracorporeal CPR compared to conventional CPR for in-hospital cardiac arrest have not been performed, the use of extracorporeal CPR for in-hospital cardiac arrest has been examined in several observational studies. The findings from existing studies, including those highlighted below, should be interpreted with caution due to the limitations associated with nonrandomized observational studies. In a single-center prospective observational study of 172 patients with in-hospital cardiac arrest, patients who received extracorporeal CPR, compared to those who received conventional CPR, experienced higher survival at discharge (hazard ratio for in-hospital mortality, 0.51; 95% CI, 0.35 to 0.74; P < 0.0001), 30 days (hazard ratio, 0.47; 95% CI, 0.28 to 0.77; P = 0.003), and 1 yr (hazard ratio, 0.53; 95% CI, 0.33 to 0.83; P = 0.006).  A second single-center observational study (n = 120) reported improved survival with minimal neurologic impairment at discharge (odds ratio for in-hospital mortality, 0.17; 95% CI, 0.04 to 0.68; P = 0.012) and 6 months (hazard ratio, 0.48; 95% CI, 0.29 to 0.77; P < 0.001) in patients receiving extracorporeal CPR compared to conventional CPR.  A third single-center observational study (n = 353) also demonstrated improved survival with extracorporeal CPR compared to conventional CPR (hazard ratio for mortality, 0.57; 95% CI, 0.35 to 0.90; P = 0.02) for patients with in-hospital cardiac arrest.  Additional studies of extracorporeal CPR for in-hospital cardiac arrest are summarized in the appendix in table A1.

A 2023 systematic review updating a previous review by the International Liaison Committee on Resuscitation of randomized and nonrandomized trials comparing extracorporeal CPR to conventional CPR for in- and out-of-hospital cardiac arrests concluded that recent randomized controlled trials suggest potential benefit of extracorporeal CPR, while nonrandomized trials had a critical risk of bias because of confounding and selection bias. 

Most recently, the ECLS-SHOCK randomized controlled trial compared the use of early extracorporeal life support plus usual medical care to usual medical care alone. This study was performed in 420 patients with acute myocardial infarction complicated by cardiogenic shock for whom early revascularization was planned. In the main analysis, there was no significant difference in 30-day all-cause mortality between groups (relative risk, 0.98; 95% CI, 0.80 to 1.19; P = 0.81), and the extracorporeal life support group experienced increased rates of moderate to severe bleeding (relative risk, 2.44; 95% CI, 1.50 to 3.95) and peripheral vascular complications requiring intervention (relative risk, 2.86; 95% CI, 1.31 to 6.25). Furthermore, a post hoc analysis performed on the subgroup of patients who received CPR before randomization (77.7% of all patients) also found no significant difference in 30-day mortality between groups (relative risk, 1.00; 95% CI, 0.81 to 1.23).

Guideline recommendations for the use of extracorporeal CPR reflect the inconclusive nature of the available evidence. The 2020 American Heart Association (Dallas, Texas) Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care cites insufficient evidence for the routine use of extracorporeal CPR in cardiac arrest but states that it may be considered for select patients in whom the cause of cardiac arrest is reversible.  The 2021 European Resuscitation Council (Niel, Belgium) Guidelines recommend considering extracorporeal CPR in select patients with refractory cardiac arrest or as a means to facilitate interventions to treat the etiology of cardiac arrest.  Similarly, the International Liaison Committee on Resuscitation suggests that extracorporeal CPR may be considered a rescue option in cardiac arrest that does not respond to conventional CPR. This was graded as a weak recommendation with very low certainty evidence.

There are no guidelines or recommendations that specifically address perioperative cardiac arrest. However, an international writing group consisting of experts in the field of perioperative resuscitation has recently included extracorporeal CPR in the management algorithm for various etiologies of cardiac arrest in the operating room. Perioperative arrest is distinct from out-of-hospital and in-hospital cardiac arrests in other settings. In perioperative arrest, providers often witness the arrest, know the patient’s medical history and etiology of arrest, and can intervene rapidly. Also, the ECMO circuits and ECMO cannulation teams are readily available close to the operating rooms in many hospitals. These unique aspects of perioperative cardiac arrest allow for improved decision-making regarding the implementation of extracorporeal CPR and reduced time to initiation of conventional and extracorporeal CPR, which may lead to improved outcomes. Several studies have found more favorable outcomes after extracorporeal CPR for in-hospital cardiac arrest, including perioperative cardiac arrest, than out-of-hospital cardiac arrest. In one such study, Wengenmayer et al.  evaluated 133 patients who underwent extracorporeal CPR and found a higher survival rate in the in-hospital cardiac arrest group versus the out-of-hospital cardiac arrest group (18.9% vs. 8.5%, P < 0.042). Low-flow time, which is the time from initiation of conventional CPR to initiation of venoarterial ECMO support, was shorter in the in-hospital cardiac arrest group (49.6 ± 5.9 min vs. 72.2 ± 7.4 min; P = 0.001) and was an independent predictor of mortality. Prolonged no-flow time, which is the time from arrest to initiation of conventional CPR, has also been associated with worse outcomes. No-flow times of less than 5 min and combined no-flow and low-flow times of less than 60 min in the setting of witnessed arrest, two common features of perioperative cardiac arrest, have been suggested as inclusion criteria for extracorporeal CPR. 

Although there are no universally accepted criteria for extracorporeal CPR, the Extracorporeal Life Support Organization lists some accepted indications for extracorporeal CPR in the 2021 Interim Guideline Consensus Statement.  These indications include (1) age less than 70 yr; (2) witnessed arrest; (3) arrest to first CPR (no-flow time) less than 5 min; (4) arrest to venoarterial ECMO flow less than 60 min; (5) end-tidal CO₂ greater than 10 mmHg during conventional CPR; (6) intermittent return of spontaneous circulation or recurrent ventricular fibrillation; (7) initial cardiac rhythm of ventricular fibrillation, pulseless ventricular tachycardia, or pulseless electrical activity; (8) absence of life-limiting comorbidities (e.g., end-stage heart failure); and (9) absence of significant aortic insufficiency (i.e., moderate or severe). The guidelines acknowledge that the decision to initiate extracorporeal CPR is frequently made with incomplete clinical information and that it is reasonable to institute venoarterial ECMO support as a “bridge to seeking” additional clinical information in patients who may benefit from extracorporeal CPR. 

Several studies have evaluated the prognostic factors for extracorporeal CPR. Younger age is associated with higher survival rates and more favorable functional outcomes after extracorporeal CPR.  In children and young adults, good outcomes have been reported with extracorporeal CPR even after hours of conventional CPR. Given the reduced low-flow times and feasibility of rapidly establishing venoarterial ECMO support in perioperative cardiac arrests, it may be reasonable to accept a higher age threshold for initiating perioperative extracorporeal CPR. An initial shockable cardiac rhythm (ventricular fibrillation, pulseless ventricular tachycardia) at the time of arrest is associated with higher survival rates after extracorporeal CPR. In a recent randomized trial evaluating out-of-hospital cardiac arrest, the reported survival rate after extracorporeal CPR in patients with an initial shockable rhythm was 48.6% compared to 23.9% for all patients who received extracorporeal CPR.  Asystole as the initial documented rhythm is associated with the lowest probability of survival. In a multicenter, retrospective study of 1,644 patients managed with extracorporeal CPR for out-of-hospital cardiac arrest, favorable neurologic outcome at hospital discharge was found in 16.7%, 9.2%, and 3.9% of patients whose initial rhythm was ventricular fibrillation/tachycardia, pulseless electrical activity, and asystole, respectively.  Elevated lactate levels and pre-existing renal disease are associated with less favorable outcomes after extracorporeal CPR. The Refractory End-Stage Shock Cured with ECLS (RESCUE) in-hospital cardiac arrest score was proposed as a tool that can be used to predict the likelihood of mortality among patients with in-hospital cardiac arrest managed with extracorporeal CPR.  Clinical factors associated with higher mortality included older age, nonshockable arresting rhythm, pre-existing renal disease, increasing duration of arrest, and nonsurgical illness. The authors found that survivors were more likely to be in procedural settings, including the operating room.

The factors that affect suitability for extracorporeal CPR are summarized in table 1. Contraindications include indicators of medical futility, such as metastatic cancer, and contraindications to venoarterial ECMO therapy, such as aortic insufficiency that is greater than mild in severity and severe peripheral vascular disease prohibiting rapid vascular access.  Because venoarterial ECMO results in retrograde flow up the descending aorta and increases left ventricular afterload, significant aortic valve insufficiency will cause left ventricular distension and is a contraindication to extracorporeal CPR.

There is a lack of evidence to guide the use of extracorporeal CPR for specific etiologies of perioperative cardiac arrest. Case reports have described its use in perioperative arrest due to local anesthetic systemic toxicity, malignant hyperthermia, anaphylaxis, and amniotic fluid, fat, and pulmonary embolism.  A recent expert review on perioperative arrest indicates that extracorporeal CPR should be considered for cardiac arrest in the operating room and more specifically for arrest due to local anesthetic systemic toxicity, severe hyperkalemia, anaphylaxis, and pulmonary, gas, or fat embolism.  Extracorporeal CPR should also be considered for perioperative arrest due to cardiac arrhythmias, myocardial infarction, pulmonary hemorrhage, or loss of airway.  On the other hand, arrest due to uncontrolled hemorrhagic shock may not be amenable to venoarterial ECMO therapy. The inability to control blood loss or effectively maintain intravascular volume leads to inadequate venoarterial ECMO flows.

The optimal timing for transitioning patients from conventional to extracorporeal CPR is unknown. If extracorporeal CPR is initiated too early, patients who may have responded to conventional CPR are exposed to unnecessary risk, such as cannulation-related complications, bleeding, and acute kidney, neurologic, and liver injury (table 2). On the other hand, delaying the initiation of extracorporeal CPR in patients with refractory cardiac arrest reduces the likelihood of survival. The probability of survival with a good functional outcome diminishes with every minute of conventional CPR without return of spontaneous circulation.  Although the threshold duration at which CPR becomes medically futile is unclear and varies between patients, data from out-of-hospital cardiac arrest suggest that conventional CPR may be most effective within the first 10 to 15 min, with a decreased likelihood of favorable outcomes if return of spontaneous circulation is not achieved within that time frame.  Use of extracorporeal CPR may result in higher survival rates for a given low-flow time compared to conventional CPR, thus allowing for a longer duration of conventional CPR before the presumption of medical futility, particularly if an automated chest compression device is used.

Although there is limited evidence to support specific recommendations, the Extracorporeal Life Support Organization suggests that initiation of cannulation for extracorporeal CPR be considered after 10 to 20 min of unsuccessful conventional CPR, with the goal of establishing venoarterial ECMO support within 60 min of arrest.  Due to the time required to mobilize ECMO resources and perform cannulation for venoarterial ECMO support, along with the clear association between reduced low-flow times and improved survival, it may be reasonable to consider extracorporeal CPR as soon as possible after initiation of conventional CPR. While the time required to initiate venoarterial ECMO for in-hospital cardiac arrest varies between institutions, one center reported an average time of conventional CPR until initiation of venoarterial ECMO of 49.6 ± 5.9 min.  Other experienced centers have also reported deployment of venoarterial ECMO within 30 to 50 min.

Based on the available evidence and the society recommendations presented above, we provide a potential algorithm for the perioperative implementation of extracorporeal CPR (fig. 2). Definitive data are absent, and determinations regarding favorable patient characteristics and suggested timing for extracorporeal CPR are extrapolated from existing evidence to the perioperative period.

Standard perioperative ACLS should be instituted immediately after cardiac arrest is recognized. Given the association between prompt bystander CPR and improved survival in out-of-hospital cardiac arrest, maintaining high-quality conventional CPR and ACLS until venoarterial ECMO support is established is critical to optimizing outcomes. Patient suitability for extracorporeal CPR should be assessed as soon as possible after ACLS is initiated (table 1).

The extracorporeal CPR team should be immediately consulted if the patient is deemed an appropriate candidate or if eligibility requires further evaluation by the team. Timely activation of the extracorporeal CPR team minimizes low-flow time by allowing initial management with conventional CPR to occur simultaneously with the extracorporeal CPR team’s evaluation and preparation for venoarterial ECMO. Preparation for venoarterial ECMO support should begin immediately after the extracorporeal CPR team confirms that the patient is an appropriate candidate, and venoarterial ECMO support should be initiated within 60 min of arrest onset if sustained return of spontaneous circulation is not achieved after at least 10 to 20 min of conventional CPR and ACLS.

While a detailed discussion of cannulation technique is beyond the scope of this review, a brief overview of the technical considerations is presented. Cannulation during concomitant CPR is technically challenging and may require interruption of mechanical compressions; the duration of interruption should be minimized to the extent possible. Once the decision to proceed with extracorporeal CPR has been made, the code leader, who should not be performing the cannulation procedure, should ensure that conventional CPR and deployment of venoarterial ECMO for extracorporeal CPR occur concurrently. Cannulation should be performed in a location that allows for safe cannulation and minimizes the time to initiation of venoarterial ECMO support. Although this location will vary between institutions, transporting a patient in refractory arrest is challenging and cannulation at the site of perioperative arrest is often the safest and most expeditious option.

Vascular access for cannulation is most commonly obtained percutaneously or by surgical cutdown. Percutaneous access is more frequently utilized and may be associated with lower rates of bleeding and infection. Surface ultrasound guidance can reduce time to cannulation and is recommended during percutaneous placement.  Transesophageal echocardiography (TEE) may be used to visualize the venous and arterial guidewires and to position the cannulas. Fluoroscopic imaging, shown to reduce major vascular complications, may also be used when available. After guidewire placement, the Seldinger technique is utilized for cannulation. The tip of the venous cannula should reside within the right atrium and can be visualized with TEE in the midesophageal bicaval view. The arterial cannula is inserted up to the iliac artery or abdominal aorta and cannot be visualized with TEE. However, the guidewire can be visualized in the descending aorta before vessel dilation and cannula insertion.

Cannula size is determined by the patient’s body surface area and anticipated need for circulatory support but may be limited by vessel size. Due to the emergent nature of extracorporeal CPR and the need for timely initiation of venoarterial ECMO support, smaller cannula sizes (e.g., 15 F) may be acceptable to reduce the risk for vascular injury, particularly if the presence of peripheral vascular disease is unknown.  Arterial cannulation often results in partial to complete obstruction of flow distal to the cannulation site. A distal perfusion cannula (6 to 8 F), which provides antegrade blood flow to the lower extremity, is commonly placed in the ipsilateral superficial femoral artery to reduce the risk of limb ischemia (fig. 1).  Because cannulation is performed under emergent circumstances where it may not always be possible to adhere to all elements of sterile technique (i.e., use of maximal sterile barrier precautions), antibiotic prophylaxis before and after insertion of ECMO cannulas should be considered.

The optimal anticoagulation regimen in patients receiving extracorporeal CPR is unknown; it may be reasonable to utilize institutional ECMO anticoagulation protocols for patients receiving extracorporeal CPR. Unless there is significant bleeding or concern for cerebral hemorrhage, systemic anticoagulation is typically administered, most commonly with unfractionated heparin. A bolus of intravenous unfractionated heparin (~50 to 100 units/kg) is frequently given just before cannulation.  Anticoagulation can be minimized or omitted in patients with bleeding complications during cannulation, and in some cases, venoarterial ECMO support can be safely maintained without any anticoagulation. However, evidence for this practice is based on small cohort studies. 

Weaning of epinephrine and other vasoactive medications should be considered when cannulation and initiation of ECMO flow are imminent so that severe hypertension does not occur with the onset of mechanical support. Chest compressions should be continued until ECMO flow is 3 l/min. Extracorporeal CPR provides additional time for treatment of the underlying etiology of arrest and recovery of native circulation after treatment. In some cases, interventions such as percutaneous or surgical coronary revascularization or pulmonary embolectomy will be required. These may be performed after stabilization with venoarterial ECMO therapy. In all cases, treatment should be carried out in a timely manner to optimize patient outcomes. After initiation of venoarterial ECMO support, several aspects of patient care should be immediately considered: gas exchange, temperature management, volume status, maintenance of adequate circulatory support, and the monitoring and management of complications such as left ventricular distension and differential hypoxemia (table 3).

Initiation of venoarterial ECMO during extracorporeal CPR significantly increases arterial oxygen content and reduces Paco₂ depending on the initial sweep gas flow rate. Although there are limited data on optimal blood oxygen levels in patients receiving extracorporeal CPR, hyperoxia should be avoided. Hyperoxia is associated with worse neurologic outcomes and increased mortality rates in patients receiving venoarterial ECMO support. Hyperoxia can be avoided by titrating the Fio₂ in mechanically ventilated patients and the fraction of oxygen in the fresh gas delivered to the ECMO oxygenator (sweep gas). Target ranges for patient Pao₂ (60 to 100 mmHg) and arterial oxygen saturation (92 to 97%) have been suggested.  Results from an ongoing registry-based multicenter randomized trial comparing a conservative oxygen management strategy (ECMO oxygen fraction titrated to achieve a postoxygenator saturation of 92 to 96%) versus a liberal strategy (ECMO oxygen fraction set at 1.0 at all times) may help guide oxygen management in the future (NCT03841084). Due to the possibility of differential hypoxemia (see section “Differential Hypoxemia”), reductions of the sweep gas oxygen fraction and ventilator Fio₂ should be based on arterial partial pressure of oxygen values sampled from a right upper extremity arterial line.

There is some evidence to suggest that rapid decreases in Paco₂ after initiation of ECMO may be associated with decreased survival.  After cardiac arrest, hypocapnia may be associated with higher in-hospital mortality and lower rates of discharge to home. However, the optimal Paco₂ levels and rates of Paco₂ correction are unknown. It is generally accepted that respiratory acidosis should be avoided.  CO₂ levels may be controlled by adjusting sweep gas flows and ventilation settings in mechanically ventilated patients. Increasing sweep gas flows relative to blood flow through the ECMO oxygenator and increasing minute ventilation will increase CO₂ clearance.

There are no definitive recommendations for the management of mechanical ventilation during the postresuscitative phase. Lung protective ventilation with a tidal volume of 6 ml/kg of predicted body weight and a plateau pressure less than 25 cm H₂O is reasonable and may reduce the risk of lung injury. In cases of severe accompanying lung injury, it may be reasonable to use an ultra-protective lung strategy with even greater reductions in tidal volumes (4 ml/kg of predicted body weight). However, the potential benefits of lung rest must be weighed against the risk of diaphragmatic atrophy.

Notably, the impact of changes in ventilator settings on overall gas exchange will depend not only on native lung function but also on the relative amount of blood flow through the heart and lungs compared to the extracorporeal circuit. Changes to ventilator settings will have more of an effect as transpulmonary blood flow increases relative to flow through the ECMO circuit, and alterations to the sweep gas will have more of an effect as flow through the ECMO circuit increases relative to transpulmonary blood flow.

Targeted temperature management has previously been shown to improve neurologic outcomes and survival after cardiac arrest. However, two recent large, randomized studies evaluating its benefit in patients after cardiac arrest are conflicting.  Patients (n = 584) with nonshockable rhythms were randomized to either hypothermia (33°C) or normothermia in the Therapeutic Cardiac Arrest in Nonshockable Rhythm (HYPERION) trial. Patients in the hypothermia group were more likely to survive (10.2% vs. 5.7%) with a favorable neurologic outcome at 90 days compared to patients in the normothermia group. The Targeted Hypothermia versus Targeted Normothermia after Out-of-Hospital Cardiac Arrest (TTM2) trial randomized patients with out-of-hospital cardiac arrest with cardiac or unknown cause to either targeted hypothermia (33°C) or normothermia. The authors found no difference in 6-month survival or functional outcomes. Arrhythmia with hemodynamic compromise was more common in the hypothermia group (24% vs. 17%; P < 0.001).

The 2021 Extracorporeal Life Support Organization interim guideline on extracorporeal CPR in adults recommends therapeutic hypothermia (33° to 36°C) for 24 h, followed by gradual rewarming to 37°C. However, these guidelines were released before publication of the TTM2 randomized trial. The International Liaison Committee on Resuscitation recommends preventing fever (37.5°C or lower) in patients who remain comatose after cardiac arrest, while acknowledging that the impact of temperature control on outcomes after extracorporeal CPR is unknown.  The heat exchanger on the venoarterial ECMO oxygenator can be used to maintain target temperatures.

Volume status can be monitored with echocardiography, a pulmonary artery catheter, and physical examination of both the patient and venoarterial ECMO circuit. Hypovolemia may cause insufficient venous drainage and inadequate venoarterial ECMO flow. Hypovolemia often manifests in venoarterial ECMO patients as “chugging” of the venoarterial ECMO circuit, during which the tubing can be observed shaking when negative pressure is applied to a collapsed vein in the patient. Additional volume should be administered as needed, and potential sources of bleeding, including vascular injury from ECMO cannulation, should be evaluated. Venoarterial ECMO flow may be transiently reduced until additional volume is administered. Fluid overload can be managed with diuretics when renal function is intact. If needed, renal replacement therapy can be integrated directly into the venoarterial ECMO circuit, or a hemoconcentrator can be integrated into the circuit and used to remove volume from the patient. 

The adequacy of circulatory support should be continuously evaluated. Initial venoarterial ECMO flows of 3 to 4 l/min are recommended by the Extracorporeal Life Support Organization interim guidelines and should be titrated to achieve and maintain adequate organ perfusion. The central venous oxygen saturation can be observed continuously on many venoarterial ECMO circuits and provides information about the adequacy of venoarterial ECMO flow and perfusion. Blood pressure, urine output, and lactate concentration can also be used to assess tissue perfusion in venoarterial ECMO patients. The optimal mean arterial pressure for patients receiving extracorporeal CPR has not been determined. However, it is reasonable to target a mean arterial pressure between 60 and 80 mmHg to maintain adequate organ perfusion while minimizing the risk of left ventricular distension (discussed below). 

Understanding the physiologic principles of peripheral venoarterial ECMO is critical to managing patients receiving extracorporeal CPR. As described previously, venous blood is drained into the extracorporeal circuit and pumped through an oxygenator, where gas exchange occurs, before it is returned into the arterial system with retrograde flow up the descending aorta. Venous drainage reduces venous congestion, right ventricular preload, and wall stress, while blood return into the arterial circulation increases arterial blood pressure and left ventricular afterload. To avoid left ventricular distension, the left ventricle must overcome the increase in afterload imposed by venoarterial ECMO flow. An inability to do so will result in reduced left ventricular stroke volume and increased left ventricular volume and pressure. In cases of severe ventricular dysfunction or very high afterload, the left ventricle may stop ejecting blood completely, increasing the risk of thromboembolic complications, particularly intracardiac thrombosis.

Left ventricular distension increases left atrial and pulmonary artery pressures, which can cause acute pulmonary edema, leading to respiratory failure. Furthermore, elevated left ventricular diastolic pressures may reduce coronary perfusion pressure. A reduction in coronary perfusion, coupled with an increase in myocardial oxygen consumption from an increase in ventricular afterload, can significantly impair myocardial recovery and worsen myocardial injury. Thus, venoarterial ECMO support, while augmenting oxygenation and perfusion to vital organs, may simultaneously induce a damaging cycle of cardiopulmonary injury. This highlights the importance of prompt recognition and treatment of left ventricular distension.

The Extracorporeal Life Support Organization Interim 2021 venoarterial ECMO guidelines state that a pulmonary artery catheter should be considered to identify elevated left ventricular pressures. If not already present, venoarterial ECMO support should be temporarily reduced during insertion to increase blood flow through the right heart and allow for catheterization. This may need to be done under fluoroscopic guidance in the cardiac catheterization laboratory because of the low-flow state. Pulmonary artery diastolic or pulmonary capillary wedge pressures are often used as surrogate measures of left ventricular end-diastolic pressure and can be used to identify increasing left-sided pressures (fig. 3). Additionally, the arterial pressure waveform may be used to monitor left ventricular ejection, with a reduction in pulse pressure signifying a decrease in stroke volume and aortic valve opening and an increased likelihood of left ventricular distension. Finally, echocardiography can be used to monitor ventricular size and the degree and frequency of aortic valve opening, as well as to identify spontaneous echo contrast or “smoke” in areas at higher risk for low-flow and thrombus formation, such as the left ventricle and aortic root, which could lead to stroke, valve thrombosis, or systemic embolism.

If ventricular distension is identified, the left ventricle should be unloaded to prevent further deterioration of cardiopulmonary function. Multiple strategies have been used to unload the left ventricle.  Because there is little evidence to support an optimal approach to unloading, clinical practice is typically based on institutional experience. Noninvasive options include the use of inotropes, vasodilation, and modest reductions in venoarterial ECMO flow to enhance left ventricular ejection. Reducing venoarterial ECMO flow decreases left ventricular afterload but may also reduce overall circulatory support and should only be implemented when the patient’s circulatory needs can still be met at lower flows or as a temporary measure while interventions to reduce left ventricular distension are instituted.

Administration of inotropes or vasodilators will enhance ventricular unloading by promoting left ventricular ejection, thus reducing left ventricular volume. Maintaining a pulse pressure of 10 to 20 mmHg helps to ensure adequate left ventricular emptying and reduces the risk of thrombus formation. However, increasing contractility with inotropic agents increases myocardial oxygen consumption and may impair myocardial recovery. The ability to reduce systemic vascular resistance with vasodilators may be limited by insufficient systemic blood pressure.

More invasive interventions for left ventricular decompression include trans-septal left atrial venting creation of an atrial septostomy, and placement of a surgical left ventricular vent, intra-aortic balloon pump, or percutaneous transvalvular heart pump (i.e., Impella; Abiomed, USA).  The Extracorporeal Life Support Organization interim 2021 venoarterial ECMO guidelines recommend the use of invasive catheter-based left ventricular unloading maneuvers in the absence of arterial pulsatility, central venous pressure greater than 20 mmHg, severe left ventricular distension on echocardiography, and a pulmonary capillary wedge pressure greater than 25 mmHg.  The EARLY-UNLOAD randomized controlled trial did not show a reduction in mortality with early routine left ventricular unloading with trans-septal left atrial cannulation compared to conventional management. Intra-aortic balloon pump counterpulsation reduces left ventricular afterload and augments diastolic pressure, which can improve ventricular ejection and coronary blood flow. Study results have been conflicting, but there are observational data to show that concomitant intra-aortic balloon pump placement in patients supported with venoarterial ECMO may improve survival in certain populations.  A percutaneous transvalvular heart pump, such as the Impella, directly unloads the left ventricle by pumping blood from the left ventricle into the aortic root and reduces aortic root blood stasis while also contributing to overall systemic flow (the combination of venoarterial ECMO and Impella support is sometimes referred to as ECPella).  Its use as a left ventricular vent was associated with improved 30-day survival, despite a higher rate of vascular complications and increased need for renal replacement therapy.

It is also critical to monitor for and manage upper body hypoxia. Antegrade flow from the native circulation and retrograde ECMO flow create a variable mixing zone, or watershed area, within the thoracic aorta in patients on peripheral venoarterial ECMO (fig. 4). The location of this mixing zone depends on the relative contributions of the heart and ECMO flow to total output. The larger the relative contribution of ECMO flow, the more proximally (toward the aortic root) the mixing zone will be located. The larger the relative contribution of the heart to systemic blood flow, the more distally the location of the mixing zone will be. While blood from the venoarterial ECMO circuit will be fully oxygenated, blood leaving the left ventricle will have variable oxygen content, depending on the patient’s pulmonary function. Ineffective pulmonary gas exchange, combined with a mixing zone located distal to the brachiocephalic artery may result in the delivery of poorly oxygenated blood to the heart and brain. This condition is known as differential hypoxemia (i.e., Harlequin syndrome) and should be suspected in patients with persistent pulmonary failure.

Differential hypoxemia is detected by measuring Pao₂ values from a right upper extremity arterial blood sample, which is more representative of cerebral oxygenation than blood from the left upper extremity. When feasible, a right upper extremity arterial catheter should be placed in patients receiving extracorporeal CPR because of the risk of differential hypoxemia from peripheral venoarterial ECMO. Arterial line placement also allows for continuous assessment of pulsatility as a rough estimate of cardiac function. A pulse oximeter should also be placed on the right hand to continuously monitor oxygen saturation. If differential hypoxemia is identified, ventilator settings should be optimized to improve oxygenation, and underlying pulmonary disease should be treated. Left ventricular venting may be required if pulmonary edema or lung injury is due to left ventricular distension, as described above. Increasing the proportion of total output supplied by the ECMO circuit may also improve cerebral oxygenation by moving the mixing zone more proximally. This can be achieved by reducing native cardiac output (reducing the dose of inotropic agents) or increasing ECMO flows. However, increasing ECMO flow may induce left ventricular distension and pulmonary edema. An additional venous return cannula may be placed in the right internal jugular vein (conversion to venoarteriovenous ECMO) if persistent differential hypoxemia occurs. In this configuration, blood is drained from the femoral venous cannula as usual, oxygenated, and then delivered back into the femoral arterial cannula and internal jugular venous cannula. The venous cannula in the internal jugular vein supplies oxygenated blood to the pulmonary circulation and improves the oxygenation of blood ejected by the left ventricle. It is essential to recognize that if the myocardium recovers in the setting of persistent pulmonary dysfunction, patients may develop differential hypoxemia, leading to cerebral hypoxia and injury, if this is not promptly recognized and treated.

Venoarterial ECMO support may be weaned after adequate recovery of myocardial and pulmonary function (pulsatile pulmonary and systemic arterial waveforms, adequate cardiac output with improvement in indicators of tissue perfusion and resolving end-organ dysfunction, mean arterial pressure of 65 mmHg or greater on less than moderate doses of pharmacologic support). The weaning process typically involves a stepwise reduction in ECMO flow with serial assessment of hemodynamic parameters and biventricular function with echocardiography.  Infusions of inotropic and vasopressor medications may need to be continued or restarted to facilitate successful weaning. Before weaning, the patient should be adequately anticoagulated (within target activated clotting time, activated partial thromboplastin time, or anti-Xa range) to prevent thromboembolic complications as ECMO flows are decreased. If irreversible cardiac injury has occurred, durable ventricular assist device implantation or heart transplantation may be indicated. Patients who are not candidates for either therapy may need to be considered for terminal decannulation with palliation.

Extracorporeal CPR is resource-intensive and should be performed by highly specialized teams in major centers with significant venoarterial ECMO experience.  As of 2022, 583 centers worldwide have reported venoarterial ECMO cases to the Extracorporeal Life Support Organization registry, including 14,836 adult extracorporeal CPR runs; the number of centers performing extracorporeal CPR has not been reported. The feasibility of extracorporeal CPR outside of large tertiary care hospitals with established ECMO programs is unknown. The Extracorporeal Life Support Organization interim guidelines for adult extracorporeal CPR recommend that centers conduct team-based simulations of extracorporeal CPR to practice cannulation during resuscitative efforts and monitor process and outcome metrics as part of quality improvement efforts. A hub–spoke model in which extracorporeal CPR is initiated at hospitals with cardiac surgical programs (there are 1,065 centers that perform cardiac surgery in the United States) and then transferred to large centers with venoarterial ECMO programs has been proposed.  However, this model is untested.

The cost–utility of extracorporeal CPR is approximately $56,156 per quality-adjusted life year gained.  This cost–utility is comparable to other well established interventions such as hemodialysis ($72,476 per quality-adjusted life year), kidney transplantation ($39,939 to $80,486 per quality-adjusted life year), and heart transplantation (less than $100,000 per quality-adjusted life year).  In the United States, a cost–utility of $50,000 to $150,000 per quality-adjusted life year has generally been considered a cost-effective healthcare intervention.  However, this estimate does not account for the cost of establishing an ECMO or extracorporeal CPR program.  Most patients who receive extracorporeal CPR can be triaged within a few days based on neurologic recovery, resulting in relatively short durations of venoarterial ECMO support for many patients who do not survive. Moreover, some patients with devastating brain injury may become organ donors and would not have been able to donate their organs without the use of venoarterial ECMO. The decision to pursue organ donation in a patient with a devastating brain injury should ensure that the preferences of the individuals and their family members are respected and may lead to new ethical challenges given the lack of familiarity of family members with extracorporeal CPR and venoarterial ECMO.

Extracorporeal CPR is always deployed in emergent situations, thereby negating the opportunity for detailed discussions with patients and their surrogates. As a result, decisions regarding the implementation of extracorporeal CPR are often made with limited information about patient preferences concerning venoarterial ECMO support. In extracorporeal CPR-capable centers, it may be reasonable to include extracorporeal CPR as part of the preoperative informed consent process when caring for high-risk patients undergoing procedures with significant risk for morbidity and mortality. Even before this perioperative patient encounter, institutions performing extracorporeal CPR could consider discussing its use when determining the code status of high-risk patients upon hospital admission. 

As the use of extracorporeal CPR continues to grow, it will be critically important for the medical community to engage in further discussion about whether extracorporeal CPR is an extension of conventional CPR (where clinicians often unilaterally cease the intervention in the setting of futility) or is instead a life-sustaining therapy like mechanical ventilation.  Rapid clinical uptake of extracorporeal CPR, without a clear understanding of which patients are most likely to benefit, increases the potential for “bridge to nowhere” scenarios. This occurs when patients whose cardiopulmonary function is maintained on venoarterial ECMO support do not have a meaningful chance for recovery and are not candidates for definitive therapies, such as durable ventricular assist devices or heart and lung transplantation. This creates ethically challenging situations for patients, who may be neurologically intact, as well as surrogates and healthcare teams. Cessation of venoarterial ECMO support against a patient’s or surrogate’s wishes is a violation of the ethical principle of autonomy. On the other hand, extending life with temporary venoarterial ECMO support despite no foreseeable chance of recovery may be considered an inappropriate use of limited resources and medically futile. A multidisciplinary team discussion with early involvement of the ethics service, palliative care, and spiritual care may help patients, surrogates, and healthcare providers navigate these challenging situations. 

Extracorporeal CPR is increasingly used to manage patients with cardiac arrest refractory to conventional CPR. When implemented expeditiously in select patients, extracorporeal CPR may improve patient outcomes in the perioperative setting. Anesthesiologists and other perioperative physicians should be familiar with its implementation and management and consider how it may be best utilized in their clinical practice. In this review, we have provided a potential algorithm for deploying perioperative extracorporeal CPR. However, there is an urgent need for high-quality randomized controlled trials to further guide the appropriate implementation of extracorporeal CPR in the perioperative setting and to help identify which patients are most likely to benefit from this resource-intensive therapy.