Extracorporeal Cardiopulmonary Resuscitation—Concise Definitive Review

Background and theoretical benefit

Cardiac arrest occurs in >600,000 people annually in the US; 350,000 arrests occur outside of the hospital (out-of-hospital cardiac arrest [OHCA]), with a 12% survival overall (1). Among the 290,000 in-hospital cardiac arrests (IHCA), survival is better at 20% (1, 2). While the causes of cardiac arrest are heterogeneous, the leading cause is by far acute fatal arrhythmia due to coronary artery disease (CAD) and acute myocardial ischemia. The American Heart Association Advanced Cardiac Life Support (ACLS) guidelines offer a limited number of diagnostics and empiric interventions to treat reversible causes of the arrest (3–5), including defibrillation. During cardiac arrest, vital organ perfusion must be maintained to enable survival with favorable neurological outcomes. Perfusion is accomplished via closed chest compressions (cardiopulmonary resuscitation [CPR]), which for the last 50 years has been the standard for cardiac arrest resuscitation (5).

Unfortunately, CPR is often ineffectual at achieving normal perfusion/blood pressure, is time-limited, and precludes many diagnostics and medical therapies. Pauses in CPR and poor-quality CPR are associated with the unsuccessful return of spontaneous circulation and decreased survival (6–14). Each minute of CPR dramatically decreases the probability of survival, which reaches 5% after 20 minutes of CPR (15, 16). CPR is also inherently traumatic and impedes many diagnostic tests and procedures during the resuscitation (17–21). For example, coronary revascularization, radiographic studies, and medication distribution throughout the body are all difficult or impossible during CPR. If clinicians do not diagnose the arrest etiology, they cannot effectively treat it.

Extracorporeal cardiopulmonary resuscitation (ECPR) is the application of extracorporeal membrane oxygenation (ECMO) to a patient in refractory cardiac arrest. ECMO provides immediate life support and can be continued for days. ECPR can be performed in myriad settings, including in the emergency department, coronary catheterization laboratory, intensive care unit, or even in the street, in as little as 6–15 minutes (22, 23) even after prolonged CPR (60–75 min)(24–26). The establishment of ECMO flow enables cessation of conventional CPR, which both ceases the trauma of manual compressions and facilitates medical diagnostic tests and therapies typically precluded by conventional CPR. Despite this potential, establishing vascular access (atraumatically) and initiating sufficient ECMO flow is, at best, challenging in patients without a pulse who are receiving chest compressions, and requires providers who are both experienced in cannulating and in managing ECMO patients. Herein, we discuss the use of ECPR for adult patients in cardiac arrest, addressing the relevant aspects of candidacy, implementation, and care. While many aspects of ECPR in adults are similar to ECPR in children, many are quite different and a proper discussion of pediatric ECPR is outside the scope of this review.

Patient selection

Optimal patient selection for ECPR depends upon the absolute added benefit of ECPR over conventional CPR; i.e., it depends upon both the probability of survival of a given patient at a given time during their resuscitation, and also upon the risk/benefit of cannulation. As an example, a young healthy patient with a witnessed cardiac arrest, immediate high-quality CPR, and initial ventricular fibrillation (VF) / pulseless ventricular tachycardia (pVT) rhythm has a good probability of converting to normal sinus rhythm with the first defibrillation. In this patient, even rapid initiation of ECMO flow may still confer more risk (e.g., anticoagulation and the risks of the cannulation procedure) than conventional ACLS management. In contrast, in the same patient, after 2–3 unsuccessful defibrillation attempts and continued high-quality CPR, the probability of return of spontaneous circulation (ROSC) falls dramatically (15, 16) and the absolute potential benefit of ECPR could be in excess of 30–40% compared to continued advanced cardiac life support (ACLS) (27). By 20 minutes of refractory arrest, the probability of favorable neurological outcome is ~3%, meaning the absolute benefit of ECPR may be 30–50% (26, 28).

The probability of benefit from ECPR often depends upon the initial rhythm. Initial rhythm provides two critical pieces of information. Initial rhythm is both 1) suggestive of the cardiac arrest etiology, and also 2) often correlates with the duration of no-flow (no CPR) time prior to CPR. Survival is best with initial shockable rhythms (29–31). Explaining these statements, shockable rhythms strongly correlate with acute myocardial ischemia (32)—which can be treated with percutaneous coronary intervention (PCI). Nonshockable rhythms are more likely to be caused by etiologies other than ischemia (30, 33). Further, VF/pVT will only persist untreated for limited periods of time until they devolve into non-shockable rhythms (34, 35). Thus, initial shockable rhythms indicate that the arrest etiology is treatable and suggest shorter durations of no-flow time.

In comparison, an initial rhythm of pulseless electrical activity (PEA) could indicate a massive saddle pulmonary embolism, in whom ECPR may enable superior perfusion by bypassing the native circulation. Myocarditis, which can lead to cardiogenic shock, and in its fulminant form can present as sudden cardiac arrest, is a typically underdiagnosed condition that can effectively be supported with ECMO with favorable outcomes; these patients may present with non-shockable initial rhythms (36). Alternatively, PEA can indicate prolonged respiratory insufficiency and hypoxemia that eventually manifested in cardiac arrest. Restoration of circulation with ECMO after prolonged hypoxemia would not reverse established end organ damage preceding the actual loss of pulses. It was recently shown that non-cardiac causes of arrest are less likely to present with arrhythmia and less likely to be successfully resuscitated (30). Likewise, asystole has the worse outcome (37); this could be due to the possibility that the asystole reflects a less-treatable/untreatable etiology, and/or that the duration of no-flow or low-flow (CPR) time may have been sufficient to cause a devolution of the rhythm into asystole (34). Both possibilities are associated with worse outcomes.

The real-world evidence of this spectrum can be partially explained by the difference in survival between patients in the ARREST trial (~43%), which was limited to ventricular tachycardia (VT)/VF (26), and patients in the Prague OHCA trial (~32%), which included PEA and asystole (25).

Finally, it is important to consider the available options for treatment and reversal of the presumed etiology of cardiac arrest when electing to use ECPR. Treatment of acute myocardial infarction (MI) resulting in cardiac arrest should lead to prompt PCI (26, 38). Treatment of cardiac arrest in the context of decompensated chronic heart failure may require a transition to a durable left ventricular assist device or transplant (39, 40), similar to what has been observed for a subset of patients with cardiogenic shock treated with ECMO (41).

The optimal patient for ECPR is one in whom there has been minimal no-flow time, in whom the cardiac arrest is refractory to conventional treatment, who has an etiology that can be treated after initiation of ECMO flow with a rapid, and atraumatic cannulation.

Controversy:

While there is little controversy over the use of ECPR for this idealized patient in an idealized system, there is controversy over whether broad implementation of ECPR, even with this idealized patient, improves patient outcomes. The INCEPTION trial, discussed below, did not show a difference in survival outcomes, and was a more pragmatically implemented ECPR trial, suggesting that outside of tightly controlled, highly practiced systems, ECPR may not confer as much benefit. Further, there is also controversy over whether use of ECPR for patients with non-shockable rhythms, older age, extensive comorbidities, and no bystander CPR is of benefit. The above observational data are mixed, and RCTs have not yet enrolled these patients (the Prague trial did enroll non-shockable rhythms).

Transport

For patients treated with ECPR, there is a strong inverse relationship between the duration of CPR prior to cannulation and survival (37, 42, 43). In fact, at one high-volume ECPR center, survival of patients cannulated for refractory VT/VF arrest within 30 minutes have had nearly 100% survival (42). Cannulation within 60 minutes of cardiac arrest is a reasonable target, which was achieved in both the Prague and ARREST trials (25, 26). Data suggest that arrest durations >60 minutes are associated with worse survival (24), which may be a contributing reason for the lack of benefit of ECPR in the INCEPTION trial, among other important protocol differences (44). For patients treated with ECPR, rapid cannulation should be the goal, at least cannulating within 60 minutes of arrest (38, 45). Finally, because most resuscitations do not utilize dynamic markers of perfusion to guide interventions (many OHCA do not even utilize continuous arterial pressure monitoring), the decision to initiate ECPR has been historically based on the duration of refractory arrest. In the future, adoption of continuous perfusion monitoring may meaningfully influence the optimal timing of ECPR initiation.

For patients with in-hospital cardiac arrest (IHCA), rapid transport to the cannulation location (if necessary) should occur in order to expedite cannulation. Cannulation is both more likely in procedural areas (46) and also predictive of better survival (37). For patients with OHCA, the majority of patients are transported to the hospital for cannulation in the Emergency Department or coronary catheterization laboratory (45, 47)—with notable exceptions in select cities utilizing pre-hospital cannulation (48–51). For patients needing to be transported, modification of pre-hospital EMS protocols may be required in order to minimize pre-hospital durations, given recent data suggesting a benefit to prioritizing continued pre-hospital resuscitation over rapid transport for OHCA patients not treated with ECPR (52).

While there are no randomized data suggesting superiority of mechanical compressions for patients with OHCA (53, 54), it is known that avoiding interruptions of chest compressions confers superior outcomes for patients with cardiac arrest (55). Given the difficulties with performing high quality continuous CPR during patient transport, many high-quality programs and guidelines favor automated mechanical compressions prior to ECPR (25, 38, 45, 56).

Controversy:

While there is agreement that cannulation within 60 minutes is a reasonable target, there are different beliefs around whether this should result in a change in EMS transport to achieve faster time to hospital arrival. Further, some centers argue for bringing the cannulators to the patient, rather than the patient to the hospital, to expedite cannulation (57). There is not consensus on these points.

Cannulation

Atraumatic cannulation of the femoral vessels is highly challenging in patients receiving CPR, due to decreased vascular volume and physical motion, both of which increase the probability of vascular injury and need for repeated attempts. These injuries, if unrecognized and untreated, can result in vasospasm or vascular thrombus, leading to distal ischemia. Limb ischemia is a known complication from cannulation during VA ECMO (58). While the injury can be mitigated by the placement of perfusion catheters in the femoral artery distal to the ECMO return cannula, the use of these cannulas for adult ECPR patients may be as low as 5% (59). Further, unrecognized vascular injury from multiple needle attempts precludes repair, which can lead to complications despite routine placement of a distal perfusion catheter (DPC).

The probability of atraumatic and rapid cannulation vs traumatic and prolonged cannulation is influenced by patient body habitus, cannulation resources, environment, and proceduralist skill. As an example of these, in the ARREST trial cannulation was carried out in a fluoroscopically capable environment by one of 4 proceduralists, most of whom had been doing high volume ECPR for years (22); the cannulation only took an average of 6 minutes, and no patient suffered limb ischemia. The ARREST trial showed a ~30% absolute survival benefit at discharge. In contrast, the INCEPTION trial took place across 10 centers, multiple cannulation locations and with providers from three specialties; cannulation lasted ~15 minutes, with 6 patients (9%) having unsuccessful cannulation and 6% having limb ischemia. There was no significant difference in primary outcomes (neurological outcome at 6 months) in this trial. It is important to highlight that a strong center-level volume outcome relationship exists for adult ECPR for centers doing more than 12 patients per year (59); extrapolating from the surgical literature (60), it is likely that on an individual level, higher procedural volumes lead to improved ECPR outcomes, for both the patient and at a system level. With or without fluoroscopy, the use of surface ultrasound has become the standard for central venous and arterial catheterization and has been associated with shorter cannulation times for adult ECPR (61).

In patients who are obese, cannulation difficulty is significantly increased, prolonged and comparative outcomes are worse (62), but reports show that if ECPR cannulation is successful, these patients still have acceptable outcomes and obesity should not in and of itself be considered a contraindication (62, 63). Further, many patients with coronary vascular disease also have significant peripheral vascular disease, including narrowed femoral arteries and arterial calcium. These features can make cannulation difficult, necessitate smaller cannulas, or result in distal embolization and bleeding.

Consideration of distal limb perfusion is likewise paramount during cannulation. The return cannula (which is located in the femoral artery) is typically 15–17 French. In patients with smaller femoral arteries, concurrent or preceding use of vasoconstrictors (including high dose resuscitative epinephrine), or multiple arterial access during cannulation (leading to thrombus formation), arterial flow distal to the cannula can be minimal or absent (58). Distal perfusion is variable over the first few days of ECPR as the patient condition changes, and insufficient distal perfusion is often occult. Additional arterial distal perfusion catheters (DPC) placed distal to the return cannula can improve peripheral perfusion and have been associated with improved outcomes in adult ECPR patients (59).

During the cannulation process, it is important to remember that the resuscitation must continue until flow is established. As part of this, there needs to be a series of people to perform chest compressions, or a mechanical compressor device, along with individuals who can secure or manage the airway, those who can administer ACLS medications, and those who can “run” the resuscitation (64, 65). Due to the complexity of the cannulation process, simulation may play an important role for centers in maintaining competency and practice (66).

Controversy:

There is active controversy over whether cannulation should occur in a fluoroscopically capable setting, or in the ED. As above, many programs are ED based, but the two best outcome RCTs were catheterization laboratory based (25, 26). Along with this, there is controversy over optimal specialty for the cannulator. There is controversy over whether there should be a few proceduralists (such as interventional cardiology) with high procedural volumes, but possibly not 24/7 availability, or a “wider bench” with more proceduralists (such as emergency physicians), but lower individual volumes. The former may not always be available, or widely available, and the later may not have sufficient capabilities (fluoroscopic ability or procedural volume).

Treatment of cardiac arrest etiology

When considering treatment of the arrest etiology, it important to remember that the establishment of ECMO flow with ECPR only temporizes hemodynamics and provides cardiopulmonary support, and does not treat the cause of the arrest. Both the definitive treatment and its temporal urgency will vary for different etiologies. For example, acute myocardial ischemia warrants emergent revascularization after establishing ECMO flow, as the ischemia will be ongoing despite ECMO support. In the case of a massive pulmonary embolism, the establishment of ECMO will decompress right ventricular outflow, preload and thus RV overdistension/ischemia as well as provide cardiopulmonary support, but will not remove the clot, which may still necessitate embolectomy/lysis. In the case of a respiratory arrest from drugs or medications, the establishment of ECMO flow will normalize hypoxemia/hypercarbia, and time and metabolism alone will often resolve the underlying problems.

Practically, despite myriad causes of arrest, initial treatment can be thought of in a hierarchical way. The potential treatments after ECPR cannulation are ordered both by 1) urgency and by 2) prevalence; prevalence is informed by the initial rhythm. For patients with initial shockable rhythms, prioritization should be given to emergent coronary angiography and PCI. For patients with ST-elevation MI (STEMI) on their initial electrocardiogram, the prevalence of CAD is 70–85%, and up to 90% of these ECPR patients have successful PCI (67). For those with initial shockable rhythm without ST-elevation, the prevalence of CAD is still 25–50%. In patients with refractory arrest requiring ECPR, this prevalence is likely higher. Patients with ECPR, especially with initial shockable rhythms, should receive coronary angiography—with the same urgency as patients with STEMI—immediately following cannulation (40, 45, 67, 68). The Extracorporeal Life Support Organization (ELSO) Guidelines recommend that all patients with ECPR and without an obvious non-cardiac cause undergo emergent coronary angiography after cannulation (38).

Patients without initial shockable rhythms may still have had an arrest from acute myocardial ischemia, and it is reasonable to perform urgent/emergent coronary angiography in select patients. Other causes of cardiac arrest, including massive pulmonary embolism can also be diagnosed in the catheterization laboratory using pulmonary angiography.

While coronary or pulmonary angiography will not help diagnose other non-MI/non-PE causes of sudden cardiac arrest, it will eliminate those as an etiology, and by stabilizing patients on ECMO, other causes of sudden cardiac arrest can also potentially be diagnosed and treated after the coronary catheterization laboratory, including respiratory arrest and cardiotoxic medication overdose.

Controversy:

As evidenced by the INCEPTION trial, there is not widespread agreement on the necessity of mandated coronary angiography after ECPR.

Imaging

Radiographic imaging in patients treated with ECPR is both diagnostic and enables therapeutic treatment. The two primary modalities to consider are coronary angiography, discussed earlier, and (whole body) computed tomography (CT). Unfortunately, data are limited on the value of diagnostic computed tomography in ECPR patients. One study showed that CT can identify injuries that may have precipitated the arrest, (e.g. aortic dissection), as well as traumatic complications from the resuscitation (e.g. pneumothorax, fractures) and from cannulation (e.g. cannulation site bleeding, vascular injury) (69). There are also good data from patients with OHCA who achieve ROSC without ECPR that describe the prevalence and distribution of conditions and injuries among resuscitated OHCA patients when imaged with “full body” CT (21, 70), reinforcing the high prevalence of injuries and the value of comprehensive post-arrest diagnostic imaging.

Different from patients without ECMO, during VA ECMO, imaging is more difficult due to the extracorporeal flow and its effect on contrast distribution and timing. Broadly, scans either need to be non-contrasted (which can show hemorrhage in the head and large volume blood in other locations), or if contrast is desired, there must be temporary clamping of the ECMO circuit or protocol modification (71). Similar to trauma patients, local protocols may allow for contrast administration despite evidence or lack of knowledge of kidney injury, given the criticality the situation.

Controversy:

There is not agreement on mandated imaging after cannulation.

The randomized trials of ECPR

There have been four randomized controlled trials of ECPR published to date. We briefly discuss each.

The Advanced reperfusion strategies for patients with out-of-hospital cardiac arrest and refractory ventricular fibrillation (ARREST) trial, was an open label trial comparing ECPR for refractory OHCA with an initial rhythm of ventricular fibrillation, to continued ACLS, conducted at the University of Minnesota (26). Patients were enrolled if they had not achieved ROSC after 3 shocks. All patients were rapidly transported from the field to the hospital using an automated compression device and were randomized (1:1) upon hospital arrival. Patients randomized to ECPR were rapidly transported to the coronary catheterization laboratory and cannulated for ECPR, and then received coronary angiography, with revascularization as needed. Patients randomized to ECPR had additional bundled care, including transfer to the cardiac intensive care unit, where they received temperature management at 34 °C , no neuroprognostication for at least 72 hours, immediate CT head and continuous electroencephalogram monitoring. Patients randomized to continued ACLS remained in the Emergency Department under the care of the emergency physicians. The primary outcome was survival to hospital discharge. The trial was stopped early at an interim analysis by the data and safety monitoring board (DSMB) after enrollment of 30 patients for superiority of ECPR arm for the outcome of survival to hospital discharge (43% vs 7%; risk difference 36% 3.7–59.2; 0.9861 posterior probability of ECMO superiority).

The Prague OHCA trial enrolled 256 adult patients with witnessed OHCA of presumed cardiac origin and without ROSC (25). The trial included patients with shockable (VF/VT) and non-shockable (PEA and asystole) initial rhythms. Randomization occurred in the field, with patients randomized (1:1) to be transported to the coronary catheterization laboratory, where they underwent ECMO cannulation and coronary angiography with revascularization as needed vs continuing standard ACLS on scene, unless they achieved ROSC, in which case they were transported to the Emergency Department for continued care. Post resuscitation care was standardized for both groups who were admitted to the hospital alive, including temperature maintained at 33 °C , and whole-body CT if feasible. The primary outcome was survival with good neurologic outcome (cerebral performance category [CPC] 1 or 2) at 180 days, by an independent assessor. The trial was stopped by the DSMB for failing to achieve the prespecified effect difference for sufficient improvement in survival. At 180 days, neurologically good survival for patients randomized to ECPR vs standard care was 31.5% vs 22.0% (p=0.09). Survival with favorable neurological outcomes at 30 days was 30.6% vs 18.2% (p=0.02). Ten patients randomized to the standard care arm crossed over to the invasive arm and received ECPR (with 40% [4/10] survival).

The INCEPTION trial was a multi-center, randomized (1:1) controlled trial in the Netherlands that enrolled patients with witnessed OHCA, initial VF, bystander CPR and at least 15 minutes of refractory cardiac arrest, and randomized them to ECPR vs standard ACLS (24). The trial included 134 patients in the intention to treat analysis. Patients were randomized pre-hospital, and then entered each treatment arm upon hospital arrival after re-reviewing inclusion/exclusion criteria. The trial occurred across 10 centers, with 12 EMS systems. The trial protocol allowed for cannulation in a variety of locations, including in the emergency department and in the coronary catheterization laboratory, and allowed for a diversity of cannulator specialties including cardiothoracic surgery, intensive care, and interventional cardiology. Postresuscitation management, such as targeted temperature management, was carried out according to local center practice. The hypothesized effect benefit from ECPR was 22%. Survival to 30 days for patients treated with ECPR vs standard ACLS was 20% vs 16% (p=0.52).

The EROCA trial was a randomized controlled trial of the feasibility of ECPR vs standard care, conducted at the University of Michigan (72). The trial screened patients with OHCA, initial shockable rhythm, or witnessed arrest, in whom the predicted time from emergency medical services (911) call to ED arrival was ≤ 30 minutes. The trial randomized patients in a 4:1 ratio to expedited transport for planned ECPR vs continued care in the field. Of 15 patients enrolled, 12 received expedited transport, and 3 received standard care. The primary endpoint was the proportion of patients with a 911 call to ED arrival time of ≤30 minutes. Of the 12 patients randomized to expedited transport, 5 achieved ED arrival within 30 minutes and were treated with ECPR. Of these 5, 3 had ECMO flow initiated within 30 minutes of arrival. No subjects in either arm survived with good neurologic outcome.

Notable differences in the trials include the restriction to VF for patients in the ARREST, INCEPTION and EROCA trials, the single center design of all but the INCEPTION trial, and the protocolized post arrest care, including coronary catheterization and temperature management, of the ARREST and Prague trials. Both the Prague and ARREST investigators had performed many tens of ECPR cannulations each, whereas the previous cannulation volume in the INCEPTION and EROCA trials was more limited. Finally, Prague and ARREST had cannulations performed in the coronary catheterization lab by interventional cardiologists, INCEPTION was not protocoled, and EROCA was ED based. We and many other experts suggest that the trials, taken together, suggest that survival is much better within highly protocoled, high-volume systems. This contrasts with a more pragmatic, multi-center, less protocolized approach. The catheterization lab and interventional cardiology were considered as the optimal location and specialty for ECPR for some centers, which may simply reflect the high volume percutaneous vascular access experience of these providers and fluoroscopic capabilities of the coronary catheterization laboratory. We suggest that the 10% survival difference (p=0.09) in the Prague trial, and 36% difference in ARREST (3.7–59.2; posterior probability of ECMO superiority 0.9861) confirms at least a 10% neurologically intact survival mortality benefit to ECPR vs standard of care within high volume/high experience highly protocolized systems, and that VF is a strong indicator of favorable survival compared to non-shockable rhythms. The differences between the 36% survival improvement in the ARREST and 10% in the Prague trial may be attributable to the inclusion of non-shockable rhythms, a large increase in survival in the control arm and the crossover in the Prague trial compared to historically observed outcomes. Indeed, it was demonstrated that a per protocol analysis pooling the ARREST and Prague patients but excluding crossover resulted in a statistically improved 180 neurologically intact survival with ECPR, even including all rhythms (32.4% vs 19.7%; p=0.015) (44). Additionally, because withdrawal of care for expected poor outcomes is a pervasive practice (73, 74), and it has been demonstrated that a significant portion of patients treated with ECPR may take weeks to regain consciousness (75), it is not known how early withdrawal of care may have influenced outcomes in the trials. In summary, an important recognition from these trials is that the procedure of ECPR alone, without an otherwise high volume/highly practiced system, and possibly implemented more pragmatically, does not confer a survival benefit.”

Neurologic outcomes

It is important to remember that the eventual neurologic outcome of the patient will be dependent upon multiple required steps in the chain of resuscitation. Conceptually, survival is highly dependent upon patient characteristics (37, 76, 77), and then three phases of care: 1) arrest phase, 2) cannulation and immediate treatment phase, 3) postresuscitation care phase. Factors during cannulation that predict outcome are discussed under “Cannulation.” Factors during postresuscitation care that predict outcome are discussed in an adjoining manuscript in CCM (78).

Within the arrest phase, there are three sub-components that impact outcome:

The first is whether the patient experiences an immediate arrest (most prevalent with arrhythmic arrests or massive pulmonary emboli), or a preceding period of hypoventilation or hypoperfusion prior to actual cardiac arrest (most commonly seen with progressive, untreated cardiac or respiratory insufficiency). This is important as it determines the “burden” of cerebral hypoxemia/hypoperfusion prior to the recognition of cardiac arrest.

The second is duration of “no-flow” time prior to initiation of CPR. Witnessed arrests and immediate CPR (minimal/no “no-flow” time) are most strongly associated with favorable neurological outcomes (76, 77).

The third is the quality of CPR and inverse duration of CPR prior to cannulation, which have been associated with survival for patients with cardiac arrest (79, 80) and ECPR (37, 76, 77, 81, 82).

A number of studies have reported on factors associated with survival and neurologic outcome in patients treated with ECPR (76, 81, 83–87), with two noteworthy survival prediction models. One (TiPS65) was derived primarily among patients with OHCA and was internally validated (77); another (RESCUE-IHCA) was derived among patients with IHCA and externally validated (37).

Context within an ECMO Program

ECPR is the combination of ECMO plus cardiac arrest care. Accordingly, ECPR is more complex than in isolated cardiac arrest. It is for this reason the ELSO Guideline advocates that ECPR programs grow out of existing high performing ECMO programs, and not in isolation (45). This ensures that the components necessary for ECMO in general are also present for ECPR, (e.g. perfusion and/or ECMO specialists to prime and manage circuits 24/7, blood bank availability, surgeons for vascular repair). Given the dependency of clinical outcomes on both postresuscitation care after cardiac arrest (88) and on post-cannulation clinical management during ECMO (59, 89), patients receiving ECPR should do so within existing ECMO programs, ideally with protocoled post-cardiac arrest care. It is known that longer durations of participation in a national cardiac arrest registry are associated with improved patient outcomes (90); it is recommended that centers performing ECPR participate in a national/international registry, such as ELSO, for quality benchmarking and research (45). It is also known that centers performing higher annual volumes of adult ECPR have better patient survival (59). When initiating an ECPR program, cannulators must be available 24/7—or at least when ECPR is offered. Most adult programs utilize pre-primed waiting circuits, which then can be available for ECPR 24/7. Further, because ECPR patients often require specialized post arrest care, including coronary angiography, temperature management, and neurologic consultation, ECPR programs should have these specialties/resources available.

Conclusions

ECPR has the potential to provide full cardiopulmonary support, negating the need for traumatic chest compressions, and enabling subsequent interventions and procedures to reverse the cause of the arrest. The benefit of ECPR is predicated on a) optimal patients (such as young patients with initial shockable rhythms, minimal no-flow time and high-quality CPR—such that they more likely have treatable etiologies and have sustained minimal neurologic injury at the time of cannulation), b) atraumatic cannulation, and c) high quality post-cannulation care. Femoral cannulation for ECMO during active chest compressions is, at best, difficult, and in select patients with peripheral vascular disease or obesity, may be more challenging or impossible. ECPR within a high performing system provides superior survival for select patients. Beyond these optimal conditions, ECPR may still provide benefit, but may also be insufficient or can even cause harm. Further studies should define the value of individual aspects of care, the limits of benefit within different patients and over different durations of CPR, and the optimal way to implement ECPR within new or lower-volume systems.