Extracorporeal Membrane Oxygenation for COVID-19

Introduction

The current coronavirus disease 2019 (COVID-19) pandemic represents an international public health crisis unprecedented in modern times with death rates in highly impacted countries approaching in excess of 400 deaths per one million persons. ¹ Acute respiratory failure is a prominent clinical feature of patients with severe COVID-19, in part due to similarities between viral surface proteins of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) and those involved in human pulmonary surfactant synthesis (Fig. 1). ² While underlying molecular pathways that mediate the pathophysiology of COVID-19 have yet to be completely defined, it is clear that the disease can affect multiple organ systems resulting in devastating clinical sequelae.

Fig. 1

Chest x-ray of a patient with COVID-19 on ECMO. The lung fields are opacified and ECMO is necessary for life-sustaining gas exchange. COVID-19, coronavirus disease 2019; ECMO, extracorporeal membrane oxygenation.

Patients with acute respiratory distress syndrome (ARDS) and persistent hypoxemia despite recruitment maneuvers, escalation of positive end-expiratory pressure, and proning may benefit from veno-venous (VV) extracorporeal membrane oxygenation (ECMO). ECMO has been used increasingly to support patients with severe ARDS and also has demonstrated utility for treatment of ARDS related to H1N1 influenza. ³ Recent data from the Extracorporeal Life Support Organization (ELSO) Registry demonstrate a 65% survival rate for patients supported with VV ECMO. ⁴

First pioneered in the late 1960s, ECMO was initially utilized for treating neonates and infants with cardiopulmonary failure but was successfully translated to adult patients in 1972. ^5,6 Since then, a number of improvements in oxygenators, cannulas, pumps, and clinical strategies have contributed to increasingly broadened use of ECMO. A 2009 landmark study of intensive care unit patients with confirmed H1N1-associated ARDS in Australia and New Zealand found ECMO survivorship to be 79% with a median of 25 days of mechanical ventilation. ³ This observation contributed to an exponential increase in the use of ECMO for acute respiratory failure, specifically with the use of VV ECMO in the 2013 H7N9 avian influenza A pandemic several years later. ⁷ A subsequent outbreak in the next year of Middle East respiratory syndrome coronavirus (MERS-CoV) suggested superior clinical outcomes in patients receiving ECMO compared to those who did not, with 65% mortality compared to 100% mortality for those receiving conventional therapy. ^8,9

While the precise role has yet to be defined, VV ECMO is currently being utilized to treat many critically ill patients with COVID-19-related respiratory failure. According to the ELSO COVID-19 registry at time of this writing, 1,042 patients confirmed to have COVID-19 have been cannulated for ECMO and 237 out of 445 (53%) whose outcome have been determined have been discharged alive. ¹⁰ At our institution we have seen an increase in the use of ECMO in the setting of the COVID-19 pandemic. In the first 5 months of 2020 we have supported 100 patients on ECMO, 33 of which have been for respiratory disease due to COVID-19.

Cannulation Strategies

Veno-venous ECMO

In patients with predominantly impaired gas exchange as often observed in COVID-19, VV ECMO is appropriate. VV ECMO extracts blood from the venous system, generally via the vena cava or right atrium, performs gas exchange, then returns blood to the venous system while simultaneously allowing mechanical ventilatory pressures to be safely reduced. In turn, this decreases risk of ventilator-induced lung injury (VILI) from barotrauma, hyperinflation, and reduces sedation requirements while potentially improving hemodynamics by reduction in intrathoracic pressure. Increased blood oxygen content improves myocardial perfusion, reduces pulmonary vascular resistance, and right ventricular strain, thereby resulting in improved hemodynamics.

Our institution preferentially performs femoral vein–right internal jugular vein cannulation. A percutaneous approach that employs ultrasound guidance is often used at the bedside without need for fluoroscopy. Cannulas are secured at the skin. A 25 French (Fr) femoral venous drainage cannula is commonly employed. Return cannulas ranging from 19 to 23 Fr are employed in the right internal jugular position. The use of large cannulas may facilitate higher flows, thus potentially achieving improved oxygenation, however at an increased rate of intracranial hemorrhage. ¹¹ In smaller patients, should partial or total venous obstruction occur from the cannula, extremity edema and venous thrombosis may result. Alternative cannulation strategies may similarly be efficacious, particularly in patients who require mechanical support for the right ventricle. Dual lumen cannulas such as the Protek Duo can provide both VV ECMO and right ventricular mechanical support. These strategies typically require fluoroscopy or transesophageal echocardiographic guidance and may be less desirable in the setting of a viral pandemic. We have not utilized these cannulas and do not recommend their use unless there is a compelling reason.

ELSO has established guidelines for the implementation of prolonged extracorporeal life support for adult patients with respiratory failure. ¹² Per their recommendations, optimal ECMO outcomes are obtained when initiated early after respiratory failure according to the criteria listed in Table 1. ^{13 -15}

Veno-arterial ECMO

Veno-arterial (VA) ECMO offers hemodynamic support as well as gas exchange, conferring greatest benefit in patients who develop significant impairment in cardiac output. This has been used only sparingly in COVID-19 but has potential utility for patients with myocardial infarction or myocarditis causing cardiogenic shock. ¹⁶ There have been a significant number of ST elevation myocardial infarctions reported in patients with COVID-19 possibly due to the virus’ potential to replicate within cardiomyocytes and pericytes resulting in viral myocarditis, although this is an ongoing area of study. ^17,18 To our knowledge, only one case of cardiogenic shock secondary to COVID-19 myocarditis has been successfully managed with VA ECMO. ¹⁹

Cannulation can be performed percutaneously at the bedside. Our practice generally utilizes a 25 Fr venous drainage cannula, with arterial cannula sizes ranging from 17 to 21 Fr for adult patients. To avoid leg ischemia, we routinely place a 6 Fr reinforced distal perfusion catheter percutaneously under ultrasound guidance. Of note, if cardiac output is initially robust or improves substantially and pulmonary gas exchange remains poor, ejection of deoxygenated blood from the heart can cause selective upper body hypoxemia (North–South or Harlequin syndrome), which can lead to coronary or cerebral ischemia. ²⁰ Development of pulmonary edema or echocardiographic evidence of left ventricular distension should prompt consideration left ventricular venting. A number of strategies, including direct surgical venting, the use of temporary left ventricular assist devices, and atrial septostomy, can be employed to vent the left ventricle.

Hybrid ECMO Strategies

Hybrid strategies may be useful for patients with significant hypoxemia and some degree of cardiac dysfunction resulting in hemodynamic compromise that is not resolved by correcting hypoxemia and hypercarbia. One case report describes a patient rescued from cardiogenic shock following poor response to cannulation for VV ECMO with left femoral 25 Fr drainage and left internal jugular 20 Fr return cannulas. Subsequent addition of a 15 Fr arterial limb to the right femoral artery improved arterial flow and significantly diminished pressor requirements with normalized lactic acid levels. ²¹ Such a strategy would be useful for patients initially cannulated for VV ECMO that develop myocarditis or thrombotic complications resulting in coronary ischemia. Conversely, hypoxemia may be alleviated by adding a right internal jugular inflow cannula should hypoxemia develop in a patient initially supported on VA ECMO.

ECMO-Assisted Cardiopulmonary Resuscitation for Cardiac Arrest

ECMO may be a useful adjunct to conventional cardiopulmonary resuscitation in the setting of refractory cardiac arrests. ²² Unfortunately, ECMO-assisted cardiopulmonary resuscitation (ECPR) may not be practical for COVID-19 patients because of the extra time required to don protective equipment for all members of the team and transport equipment into the room. ²³ Further, patients with COVID-19 who have refractory cardiac arrest tend to have a grave prognosis and the risk–benefit profile does not favor performing ECPR. Ethical considerations and expectations should be discussed early with patients and their surrogate(s) during the hospital course. Our institution requires patients or family members to discuss expectations for ECMO therapy during the initial few days on ECMO.

Complications of ECMO Therapy

ECMO is associated with serious complications that include bleeding, stroke, ventilator-associated pneumonia, catheter-related blood stream infections, metabolic derangements, limb ischemia, and brain death in cases of severe brain injury. ²⁴ Bleeding is the most common severe complication during ECMO occurring in up to 70% of VA ECMO patients. ^{25 -27} Procedures such as chest tube insertion and nasal gastric tube insertion also lead to much higher complication rates in patients on ECMO for COVID-19 because of increased bleeding risk. Severe inflammation and dysregulation of the coagulation cascade are factors that may lead to increased complication rates for ECMO when used for patients with COVID-19 compared to other disease processes.

Special Considerations

Personal Protective Equipment

Guidelines from the World Health Organization (WHO) recommend combined droplet and contact precautions for patients with COVID-19. The US Centers for Disease Control and Prevention (CDC) have suggested that airborne precautions are necessary. ²⁸ Because our understanding of the infectivity of SARS-CoV-2 is incomplete, there remains controversy about what level of personal protective equipment (PPE) is necessary. ²⁹ Regardless, all suspected patients should be treated as positive until tests have confirmed negative viral status. Routine hand hygiene prior to and following contact with suspected or confirmed COVID-19-positive patients, materials, or surfaces with either soap and water or alcohol-based disinfectants remains universally endorsed for all levels of patient care. ³⁰

As ECMO is often initiated as a bedside sterile procedure, PPE should be collected and then donned in the correct order. Following an initial round of hand hygiene, sterile gown should be donned, then N95 respirator, powered air-purifying respirator (PAPR) with high-efficiency particulate arrestance filter, sterile gloves, then a second sterile gown and gloving, taking care to overlap with the cuff of the gown at each step. When doffing, the order is to remove external and internal gown into both layers of sterile gloves, then to remove PAPR, keeping the respirator worn until exited from the contaminated area. Appropriate training and fit testing should be provided to all personnel prior to patient care, with additional staff present outside of designated hot zones to ensure correct donning and doffing procedures. The additional time necessary to comply with these procedures should also be considered, as time-sensitive encounters may be complicated by correct PPE application. Similarly, when procedures such as bronchoscopy or tracheostomy are being performed, appropriate PPE and practices must be utilized (Fig. 2, Fig. 3).

Fig. 2

Performing bedside tracheostomy on COVID patient. Appropriate PPE is critical for any contact with patients with confirmed or suspected COVID-19. During the tracheostomy, ECMO flow and sweep are increased and the ventilator is intermittently paused when the airway is open to help prevent aerosolization. COVID, coronavirus disease; ECMO, extracorporeal membrane oxygenation; PPE, personal protective equipment.

Fig. 3

Cannulation for ECMO in a patient with COVID-19. VV ECMO, used in this case, is appropriate for most patients with COVID-19, where impaired gas exchange is the predominant problem. A femoral venous cannula and right internal jugular cannula were used in this case. Appropriate PPE is also an important consideration. COVID-19, coronavirus disease 2019; ECMO, extracorporeal membrane oxygenation; PPE, personal protective equipment; VV, veno-venous.

Personnel

Implementation of multidisciplinary ECMO teams has been documented to improve clinical outcomes for patients with ARDS compared to decentralized and informal chains of command. ³¹ When initiating ECMO for COVID-19-positive patients, personnel should be limited to reduce personal risk of acquiring the virus and to minimize rates of transmission and with clearly designated roles. Our general strategy utilizes 2 physicians in sterile garb for cannulation, 1 intensivist, 1 nurse, and 1 ECMO specialist or perfusionist for cannulation and initiation. Centers may adjust to their needs accordingly. ³²

Patient Selection and Exclusion Criteria

Appropriate patient selection and timing of ECMO initiation is crucial to achieving favorable outcomes in ECMO; however, this process is in evolution for COVID-19. Terminal conditions, severe central nervous system compromise, and do-not-resuscitate orders preclude treatment with ECMO. According to ELSO guidelines, patients with extensive comorbid disease and advanced age should be scrutinized, while patients requiring mechanical ventilation in excess of 7 days should be excluded. ²³ Once initiated, if cardiopulmonary function does not improve over a reasonable period of time, care may be deemed futile. ELSO has issued guidelines for the institution of ECMO in respiratory failure (Table 1). At our institution we employ similar criteria.

Table 1

ELSO Guidelines for the Implementation of Prolonged Extracorporeal Life Support for Adult Patients with Respiratory Failure.

Initiation of ECMO should be considered according to the following criteria: Consideration should be given in hypoxic respiratory failure with associated mortality equal to or greater than 50% and indicated in those whose mortality is 80% or greater. Fifty percent mortality risk is associated with PaO2/FiO2 <150 on FiO2 >90% and/or equivalent risk stratification scores (Murray score 2 to 3, AOI score 60, or APSS score 3) while 80% mortality risk is associated with PaO2/FiO2 <100 on FiO2 >90% (and/or Murray score 3 to 4, AOI >80, or APSS 8) despite optimal care for 6 hours or less. CO2 retention on mechanical ventilation despite high plateau pressure in excess of 30 cm H2O. Severe air leak syndromes. Intubation requirement in pre-lung transplant patient. Immediate cardiac or respiratory collapse (including PE, blocked airway, or nonresponsive to optimal care).

Relative contraindications resulting in suboptimal outcomes following ECMO therapy include: High mechanical ventilation settings for 7 days or more, including FiO2 >90% or plateau pressure >30 cm H2O. Major pharmacologic immunosuppression (absolute neutrophil count <400/ mm3). Recent or expanding CNS hemorrhage. Nonrecoverable or terminal injury (CNS injury, terminal malignancy). Increasing age associates with increasing risk of poor outcome.

The above guidelines have been set forth according to the ELSO respiratory support guidelines v1.4 dated August 2017 which are available at https://www.elso.org/Resources/Guidelines.aspx. ¹² Per ELSO guidelines, optimal ECMO outcomes are obtained when initiated early after respiratory failure within 1 to 2 days according to these criteria. Mortality scores include the Murray Score for Acute Lung Injury, AOI, APSS.

Abbreviations: AOI, Age-Adjusted Oxygenation Index; APSS, Acute Physiology of Stroke Score; CNS, central nervous system; ECMO, extra corporeal membrane oxygenation; ELSO, Extracorporeal Life Support Organization; PE, pulmonary embolism; PPE, personal protective equipment; VV, veno-venous.

Postcannulation Care

Ventilation Strategies

COVID-induced acute respiratory failure represents one of the chief insults secondary to infection. According to the Surviving Sepsis Campaign (SSC) COVID-19 panel in conjunction with the WHO and CDC, clinical care guidelines have been offered to address the oxygenation and ventilation of COVID-19 patients. ³³ For those who fail to respond to noninvasive ventilation including high-flow nasal cannula or noninvasive positive pressure ventilation, invasive mechanical ventilation is necessary. The SSC advocates low tidal volume (V_T) ventilation V_T 4 to 8 mL/kg of predicted body weight over higher tidal volumes defined as V_T >8 mL/kg and targeting a plateau pressure (P_PLAT) of <30 cm H₂O, as more aggressive ventilation pressures have been documented as causing VILI and contributing to multiorgan failure in patients with ARDS. ³⁴ Interventions should focus on higher positive end-expiratory pressure (PEEP) as tolerated, with patients observed for signs of barotrauma for PEEP >10 cm H₂O. Limiting driving pressure is key to reducing VILI, as this is the key determinant of VILI and highly predictive of mortality in ARDS. ³⁵ PEEP may be titrated at the bedside and, when optimized, driving pressure should be at its lowest level to achieve a particular tidal volume. Fluid support should be used conservatively as COVID-associated cardiac failure either alone or in combination with ARDS has been documented to cause as many as 40% of COVID-19 deaths. ³⁶ Judicious fluid administration may be required to help prevent renal failure. Prone positioning for ventilation has also been demonstrated to be useful in ARDS and is recommended for 12 to 16 hours but should be used in conjunction with diligent observation for the development of pressure ulceration. Paralytic agents should be used as necessary and partial paralysis along with deep sedation is often helpful to improve ventilator synchrony and reduce work of breathing. Intermittent dosing should be used when possible, reserving continuous infusion for those with persistent ventilator dyssynchrony. These principles are also employed after a patient has been cannulated for ECMO.

Anticoagulation

Early observations suggest that COVID-19 patients are hypercoagulable and require increased anticoagulation, with some data suggesting that heparin administration may be beneficial. ^{37 -40} Patients have elevated fibrinogen and factor VIII levels that likely contribute to hypercoagulability. These changes can be further exacerbated by ECMO, which also leads to increased factor VIII. Before initiating a loading dose of heparin, baseline laboratory data should be obtained if possible including complete blood count, prothrombin time (PT), activated partial thromboplastin time (aPTT), d-dimer and fibrinogen levels, activated clotting time (ACT), and antithrombin activity. Visoelastic testing may be utilized in some centers and can be helpful in identifying hypercoagulable states in patients with COVID-19. Anticoagulation protocols vary, and some centers preferentially use direct thrombin inhibitors. ^{41 -43} There is no compelling data to support a single approach and heparin remains the gold standard anticoagulant. At our center, we have increased our aPTT goal to 60 to 80 for VV ECMO due to concerns for hypercoagulability. Values are checked every 6 hours until a consistent, stable level is reached, and then every 6 hours thereafter.

Routine monitoring for heparin-induced thrombocytopenia and thrombosis should be performed with platelet counts and skin assessments. Acquired hypofibrinogenemia may be observed in high consumption states as in disseminated intravascular coagulation and should prompt investigation for large volume thrombosis. Typically such low fibrinogen states are associated with increased bleeding risk and warrant necessary precaution. Patients with pronounced coagulopathy prior to initiation of ECMO or known heparin resistance should receive appropriate blood products including fresh frozen plasma, cryoprecipitate, or platelets as indicated. Laboratory data suggesting increased risk of thrombosis, as measured by elevated d-dimer levels, increased fibrinogen, or as observed changes in thromboelastogram should prompt careful monitoring of the ECMO circuit, particularly the oxygenator. ⁴⁴

Medical Therapies

Data on medical therapy for COVID-19 are currently in evolution. Initial therapy with corticosteroids has been discouraged for generalized use in a recent meta-analysis by the SSC and previous studies have demonstrated a possible role in promoting increased viral shedding in coronavirus strains other than SARS-CoV-19. ^33,45 Still, corticosteroids may benefit patients with septic shock or who are worsening without other suitable options. Use of acetaminophen for antipyresis and broad-spectrum antimicrobial coverage for opportunistic secondary infections are also strongly encouraged by most professional societies. Intravenous immunoglobulin and convalescent plasma may be useful but have previously proven ineffective in modulating morbidity or mortality in other viral respiratory infections or have proven too difficult to pool in sufficient quantities to provide on clinically meaningful scales. ^46,47

Antiviral Drugs

Limited data exist to support efficacy of antiviral therapy in COVID-19; however, this represents an area of ongoing investigation. Chloroquine and its metabolite hydroxychloroquine have previously been studied for their effects on the related SARS-CoV virus responsible for the SARS pandemic of 2003. Some Chinese centers have initiated recommendation guidelines for the implementation of chloroquine and hydroxychloroquine, although care should be taken in monitoring for cardiotoxic prolonged QTc changes that have been well documented with its use. ^48,49 Preliminary data from the Brazilian CloroCovid-19 trial have identified increased adverse events and cardiotoxicity with increased dosing of chloroquine in severe COVID-induced ARDS including QTc prolongation and myocarditis. ⁵⁰ To date there remains insufficient evidence to recommend use at this time, with the National Institutes of Health (NIH) currently recommending against use of high-dose chloroquine or hydroxychloroquine as of this writing. ^{51 -54}

Lopinavir either alone or in combination with ritonavir as an adjunctive agent to inhibit metabolism of lopinavir has not been proven effective either in symptom abatement or mortality reduction. ⁵⁵ Only nominal benefits including modest reductions in fever were observed in nonrandomized data on 47 patients receiving lopinavir/ritonavir combination therapy. ⁵⁶

Remdesivir is a similarly structured nucleoside analog to lopinavir previously studied in the treatment of Ebola and Nipha viruses. Therapeutic administration has demonstrated reductions in viral titers as early as 12 hours after administration of therapy with improved respiratory status and reduced pulmonary infiltrates on radiography in rhesus macaque models. ⁵⁷ In a statement by the NIH, it was reported that the Adaptive COVID-19 Treatment Trial (ACTT) sponsored by the National Institute of Allergy and Infectious Diseases was able to demonstrate 31% faster time to recovery in patients receiving remdesivir than those who received placebo, with a median time to recovery of 11 days. ⁵⁸ Similarly, in a double-blind randomized controlled trial of 1,063 hospitalized adults, a 10-day course was superior to placebo in shortening time to recovery with estimated mortality at 14 days 7.1% in the remdesivir group compared to 11.9% in the placebo group. ⁵⁹

Tocilizumab is an anti-IL-6-receptor human monoclonal antibody currently used in the treatment of rheumatoid and systemic juvenile idiopathic arthritis, which modulates release of proinflammatory cytokines and has been proposed as a possible modulator of COVID-19 symptomatology. ⁶⁰ SARS-CoV-2 has demonstrated similar cytokine profiles to SARS-CoV and MERS-CoV, with the constellation of these symptoms now referred to as cytokine release syndromes. ⁶¹ The drug is currently under investigation for its potential role in the treatment of SARS-CoV-2. ⁶² Single-center experiences and early reports from multicenter trials are promising, suggesting increased likelihood of survival but are not as yet definitive. ^63,64

Conclusions

The emergence of SARS-CoV-2 has led to a global pandemic resulting in large numbers of cases of respiratory failure. ECMO can be life saving for select patients with refractory cardiopulmonary failure. VV ECMO is the appropriate configuration for most patients with predominantly respiratory failure, although VA ECMO or hybrid strategies may be appropriate for patients with impaired cardiac output. Regionalization of care to high-volume ECMO centers is most likely to lead to improved outcome as management of COVID-19 patients on ECMO is complex and requires a team-based approach. As experience with COVID-19 grows, ECMO is likely to have an important role in caring for severely ill younger patients with few comorbidities.