Extracorporeal membrane oxygenation (ECMO) for critically ill adults in the emergency department: history, current applications, and future directions

Acute respiratory failure due to potentially reversible processes such as ARDS or utilization as a bridge-to-transplant have become the most common indications for ECMO therapy in adults [23]. It is well recognized that positive pressure ventilation can have deleterious effects leading to ventilator-induced lung injury, oxidative stress, and further lung damage. Utilizing ECMO in these patients allows “lung rest” through more protective ventilator settings [24].

Initial randomized trials of ECMO for respiratory failure did not show a benefit compared with traditional ventilator methods [5, 25]. In the 1990s and 2000s, survival rates reported in case series with ECMO improved to 52–75 % [26, 27]. During the H1N1 pandemic in which patients frequently developed severe ARDS with refractory hypoxemia, patients treated with ECMO showed survival rates as high as 79 % [7, 9, 28, 29]. The first modern randomized controlled trial of ECMO for adult patients with ARDS, the CESAR trial, was published in 2009. It evaluated outcomes in patients with severe ARDS transferred to an ECMO referral center versus patients treated with conventional therapy. Although mortality at any point was not significantly different, the study identified significantly higher 6-month survival rates in the group transferred to the ECMO referral center—of which only 75 % of the patients received ECMO—versus the control group that were not transferred [30]. Thus, it may have been other aspects of care at the ECMO center, not necessarily the ECMO itself, that led to improved outcomes. The Extracorporeal Membrane Oxygenation for Severe Acute Respiratory Distress Syndrome (EOLIA) trial is currently underway, evaluating early ECMO within 3 hours of initiation of mechanical ventilation for patients with refractory hypoxemia and severe ARDS [31]. If this trial finds a benefit of early ECMO for ARDS, coordinating transfer to ECMO-capable referral centers directly from the ED for early initiation of ECMO support may become important for patient outcomes [32]. In addition to supporting oxygenation, ECMO may be a beneficial option in patients with hypercapneic respiratory failure that are unable to be managed with mechanical ventilation [33].

Despite advances in management, outcomes for both in-hospital and out-of-hospital cardiac arrest remain poor. In-hospital cardiac arrest treated with conventional cardiopulmonary resuscitation (CPR) typically has a survival rate of 15–17 % and out-of-hospital cardiac arrest (OHCA) survival is even lower at only 8–10 % [34, 35]. The worst outcomes are in patients with prolonged time to return of spontaneous circulation. Prolonged cerebral hypoperfusion leads to significantly worse neurologic sequelae and early initiation of ECPR with veno-arterial (VA) ECMO may be a useful adjunct to reducing the interval time from arrest to restoration of cerebral perfusion.

Data for in-hospital arrest are the most promising, likely due to the shorter interval from the onset of arrest to initiation of ECMO flow. While there are no randomized trials to date, observational studies have reported an association between ECPR and improved survival. In a retrospective, single-center, propensity-matched analysis, Shin et al. [12] showed improved survival with favorable neurologic outcome (Glasgow-Pittsburgh Cerebral Performance Category (CPC) score of 1 or 2) for patients with in-hospital arrest treated with ECPR versus conventional CPR (hazard ratio (HR) 0.17, 95 % confidence interval (CI) 0.04–0.68). Chen et al. [11] found similar 30-day (HR 0.47, 95 % CI 0.28–0.77) and 1-year (HR 0.53, 95 % CI 0.33–0.83) survival for ECPR when compared with conventional CPR. Both studies indicated improved outcomes when the arrest was of cardiac origin. Another retrospective review of prolonged in-hospital arrests (>15 minutes) or refractory shock after return of spontaneous circulation showed nearly half of the patients who survived had a CPC score of 1 or 2 with the use of ECPR [10]. Other observational studies have found variable improvements in mortality with the use of ECPR [36, 37]. A recent meta-analysis performed by Cardarelli et al. [38] in 135 patients from 1990 to 2007 showed a hospital survival rate to discharge with ECPR of 40 %.

Reports from OHCA studies are not as robust, although there are dramatic reports of otherwise hopeless cases rescued by ECMO [39]. Haneya et al. [40] compared ECPR initiated in the ED for OHCA with ECPR initiated for in-hospital cardiac arrest, and found a survival rate of 42 % for in-hospital arrest patients versus only 15 % for OHCA patients. ECPR combined with therapeutic hypothermia and intra-aortic balloon pump placement was recently shown to improve neurologic outcomes for OHCA patients with ventricular fibrillation or pulseless ventricular tachycardia. This study demonstrated survival with a favorable CPC score in 11.2 % of those with ECPR versus 2.6 % with conventional CPR at 6 months [14]. Similar results were reported in a small observational pilot study in Australia combining mechanical compression devices, ECPR, and hypothermia for patients with refractory cardiac arrest. In this small study, 5 of 11 OHCA patients and 9 of 15 in-hospital cardiac arrest patients survived, and favorable neurological outcome was achieved in half of the survivors [15].

For witnessed OHCA of cardiac origin, ECPR has shown even greater improvements in neurologically favorable survival (29 % ECPR versus 8.9 % CPR) [13]. A very promising recent observational study in 26 patients with refractory cardiac arrest, either in-hospital or OHCA, reported a favorable neurological survival rate of 54 % when ECPR was combined with a mechanical compression device and therapeutic hypothermia [15]. However, these data have not been replicated and similar studies have reported survival rates ranging from 4–15 % [41‐43]. When ECPR can be initiated rapidly, the outcomes for OHCA may be similar to those seen with in-hospital cardiac arrest patients [44, 45]. Although there are no large prospective randomized trials and survival from refractory cardiac arrest is poor, ECPR may provide a tool to improve survival with good neurologic outcomes when initiated early in selected patients.

While an optimistic estimate of survival from OHCA with the use of ECPR may be in the 15–20 % range, the critical factor that determines success appears to be the duration from the onset of arrest to achieving ECMO flow. This may be why in-hospital cardiac arrest studies have generally reported better outcomes [45‐47]. Furthermore, the volume of patients meeting optimal criteria for ECPR, including witnessed ventricular fibrillation/tachycardia with a short interval to initiation of CPR, refractory arrest despite optimal resuscitative efforts, and <75 years of age comprises a very small (<10 %) subset of all patients with OHCA [48, 49]. Additionally, complication rates of ECPR remain high [43, 50, 51]. Currently the American Heart Association’s position is that evidence does not support a recommendation for ECPR, although it may be considered in highly specialized centers in patients who have a potentially reversible disease and a short duration of cardiac arrest [52]. The recent Institute of Medicine report on cardiac arrest care states ECMO is an emerging technology that has promise in improving cardiac arrest care and should be developed and researched [53].

In addition to ECPR, ECMO may have a role in select patients with cardiogenic or septic shock [54‐56], toxic ingestions [57], thyrotoxicosis [58], or trauma [59]. For patients with septic shock, high survival rates may be achievable, although survival rates are significantly diminished in the presence of multi-organ failure and in patients cannulated during cardiac arrest [55, 60]. ECMO has shown promising results as a method of supporting hemodynamics in patients with cardiogenic shock due to acute myocardial infarction [61], massive pulmonary embolism [62] or myocarditis [63]. ECMO may be preferable to aortic balloon pumps as it can provide more robust biventricular support by increasing right ventricular drainage in addition to supporting gas exchange, and it can be employed quickly at the bedside. However, the downside of ECMO for cardiac support is that because of the retrograde aortic flow, left ventricular afterload and oxygen demand increase without the placement of a left ventricular drain. ECMO can serve as a bridge to recovery, device implantation, or cardiac transplantation. Small trials have shown improved survival rates in patients placed on ECMO as a bridge to a left ventricular assist device and subsequently to transplantation [64]. Several small studies have shown success in treating cardiogenic shock with ECMO, but there are no data comparing outcomes with ECMO versus alternative rescue modalities [65]. Arranging the initiation of ECMO may be a viable option in patients with known severe cardiac dysfunction in shock refractory to conventional therapy in the ED.

A successful ECLS program requires a significant multidisciplinary and organizational commitment to ensure necessary resources and personnel [66]. Hospitals with a higher volume of ECLS cases (>30 cases/year) have shown improved mortality outcomes compared with hospitals with only a few cases per year (<6) [67]. Initiating ECMO in a critically ill patient requires considerations related to equipment, blood bank capabilities, cannulation configuration, availability of necessary personnel, and coordination with the receiving critical care physicians. In a recent trial evaluating ECMO for OHCA, initiation of ECMO required two critical care physicians for cannulation, a third physician providing ultrasound guidance of cannula placement in the inferior vena cava, a fourth physician dedicated to leading the resuscitation, a nurse coordinator to initiate the circuit, and a sixth person to infuse cold saline for intra-arrest hypothermia [15]. Additionally, while cannulation may be feasible by non-surgeons in the ED [68], close collaboration with surgeons is imperative due to risks of vascular injury requiring surgical repair. A successful ECLS program requires physicians, nurses, perfusionists, and respiratory therapists trained and competent in cannulation and management of the ECLS circuit to be available in sufficient numbers to provide 24/7 coverage. A streamlined exit strategy from the ED should be established, whether that includes transfer to the cardiac catheterization lab, ICU, or transfer to an ECMO receiving hospital. Predetermined inclusion and exclusion criteria, as well as guidelines on how and when to separate from ECMO in a patient who fails to improve are important to optimize outcomes and resource utilization. Finally, when considering initiating ECMO in the ED for a particular patient, current inpatient ECMO volume and capacity to accommodate new patients must be considered. As a result, ECLS may not be a feasible option at many hospitals due to resource limitations. Local, regional, and inter-regional networks of hospitals with direct communication with the ECMO referral center that provides a mobile rescue team for cannulation and transport of patients with critical cardiopulmonary failure may be the optimal solution to alleviate many of these costly programmatic considerations.

Cannulation strategies must be carefully considered based on the underlying pathophysiology. Configurations include veno-venous (VV) and VA approaches, with hybrid modes possible for specialized uses (e.g., veno-arterial-venous for patients with cardiogenic shock and differential hypoxemia). In VV ECMO, patients are typically cannulated through the femoral vein and internal jugular/superior vena cava with two cannulas, or a single dual-lumen cannula placed into the right atrium and inferior vena cava via the right internal jugular vein (Fig. 1). Venous blood is drained from the vena cava regardless of the cannulation configuration. After passing through the oxygenator, blood is returned to the right atrium via the return cannula (Fig. 2). This configuration requires blood to pass through the pulmonary circulation and is systemically circulated by the native cardiac output. Thus, VV ECMO provides no cardiac support.

Percutaneous VA ECMO in adults most commonly involves femoral artery and vein cannulation where oxygenated blood is returned to the femoral artery and flows retrograde into the aorta. Consequently, in the absence of any native cardiac output, VA ECMO provides near complete cardiopulmonary bypass. In the presence of native cardiac output, blood flow out of the left ventricle mixes with retrograde ECMO return flow in the aorta. If there is no intrapulmonary shunt, the native cardiac output is well oxygenated and this mixing is not detrimental. However, if the patient’s lungs are failing (e.g., because of ARDS), blood flow from the native cardiac output is deoxygenated and competes with retrograde flowing oxygenated blood from the ECMO circuit. These competing flows mix somewhere in the aorta and can create differential hypoxemia (“North–south” syndrome) whereby poorly oxygenated blood preferentially perfuses the upper body, including the brain and heart, potentially leading to coronary and cerebral ischemia.