Role of extracorporeal membrane oxygenation in COVID‐19: A systematic review

We aimed to examine the literature evidence behind using extracorporeal membrane oxygenation in COVID‐19 patients in a systematic review manner.

Organisation (WHO) released interim guidelines that advocate the use of extracorporeal membrane oxygenation (ECMO) to support the cardiorespiratory system in patients who fail maximal conventional therapies with ARDS. 3 ECMO principally functions as a rescue cardiopulmonary bypass, exchanging oxygen with carbon dioxide over an artificial membrane to deoxygenated venous blood which is then returned to the patient via the venous or arterial system. Previous pandemics have proven the role of ECMO to support the recovery from severe respiratory and cardiovascular compromise resulting from ARDS. 4,5 However, the role of ECMO in COVID-19 and its implications are yet to be understood with data being collected and reported from large centers internationally.
This study aims to investigate the current literature and explore the effectiveness of ECMO on patients with COVID-19. Our secondary objective was to identify patterns between types of ECMO Keywords and MeSH terms relating to these categories were used to optimize the output from the database search including "Coronavirus" OR "nCoV*" OR "2019-nCoV" OR "COVID*" OR "SARS-CoV*" AND "ECMO" OR "VV-ECMO" OR "VA-ECMO" OR "Extracorporeal membrane oxygenation." All the relevant articles were screened and selected for inclusion by two authors and any disagreements were resolved through consensus and vote.

| Inclusion and exclusion criteria
The main exclusion criteria were narrative reviews, consensus document, editorials and commentaries without reporting on patient data or outcomes. Studies were included if they contained primary data on patients who were diagnosed with COVID-19 and were subsequently put on ECMO.

| Quality assessment
A quality assessment for all the included articles was undertaken, using the NIH quality assessment tool for the appropriate studies.
No articles were excluded based on their quality score.

| Data extraction
Data extracted from the included articles were tabulated, and then, a narrative synthesis was undertaken to identify key themes in the literature.

| RESULTS
A total of 102 articles were retrieved from the database search and snowballing. Following the exclusion of duplicates and screening, a total of 25 articles were selected for inclusion in this systematic review   (Figure 1). The characteristics of these studies are summarized in Table 1.
Three main themes were identified following the data extraction:

| DISCUSSION
ECMO was often adopted as salvage therapy for patients commonly experiencing COVID-19-induced ARDS and/or other COVID-19 complications. 6,12,22,27 The overall mortality rate following the collation of the data from the 25 articles selected in this review was 19.83%. This value can, however, only be used as an estimate as some articles did not report mortality outcomes for their patients put on ECMO, making the mortality rate subject to increase. Despite this, this figure shows a promise that ECMO is not detrimental for critically ill patients with COVID-19.

| Evidence against/inconclusive regarding ECMO for COVID-19
A small number of studies presented high rates of mortality for patients with COVID-19. Three studies in this review reported 100% mortality for patients with ARDS put on ECMO, whereas Yang et al reported a similarly high mortality rate of 83.33% (15 deaths in total). 18,25,26,30 In addition, Guan et al reported that all five patients that were put on ECMO experienced the composite primary endpoint that consisted of admission to the ICU, use of mechanical ventilation or death. 10 Other studies reporting poor outcomes for ECMO include Zeng et al. 28 Although three patients did recover following ECMO, four patients (two of which were comatose) remained on ECMO and five patients died. Despite this, however, the study attributed half the deaths to septic shock and multiple organ failure. We are not sure whether the patients in question experienced multiple organ 2680 | failure whilst on ECMO; if this was the case, it would explain the negative outcome due to the absolute contraindication between ECMO and multiple organ failure (as depicted in Extracorporeal Life Support Organisation (ELSO) guidelines). 31 Li et al 14  Although these studies have reported either negative or equivocal results, several considerations should be noted. Firstly, many of these articles consist of a small sample, and thus, no reliable conclusions can be made. Secondly, some of the articles did not provide information with regard to the patient's disease severity at the time of ECMO initiation, so we cannot know whether ECMO was perhaps administered too late to have a significant effect in a severely deteriorating patient.
Main risk factors associated with a high mortality rate include age ≥60 years, various comorbidities (such as cardiovascular disease and diabetes) and low lymphocyte count <0.8 (×10*9/L) and D-dimer levels >1 µg/L on admission. 33 We find that many of the patient deaths reported in this review possessed some of the above characteristics, allowing us to, therefore, assume that such predispositions were potential determinants that had a strong influence on the prognosis of patients, despite the initiation of ECMO.

| Evidence supporting ECMO for COVID-19
In our study, six case reports and two case series reported positive endpoints (weaned off ECMO/discharged from hospital) for patients on ECMO. [7][8][9]11,17,19,23,29 These outcomes supporting ECMO for COVID-19 are likely to have occurred for the following reasons.  Third, in cases where the association between COVID-19, cytokine storm and mortality has been established, 34 the role of ECMO in reducing inflammatory substances has also been attributed to patient survival. 23    Although beliefs and traditions regarding death vary across cultures around the world, it seems that this final contact is paramount to optimal palliative care.

| EVIDENCE-BASED RECOMMENDATIONS
The use of VV-ECMO in COVID-19 ARDS is supported by multiple guidelines. 3 Drawing from data published so far, we propose the following algorithm outlined in Figure 2. Patients presumed suitable for ECMO should be identified early, to minimize the risk of complications associated with prolonged ventilator use. We also recommend the absolute contraindications outlined in the guidelines be followed (see Table 2).
In addition to existing guidelines, several prediction tool scores including the PRESERVE ((AUC 0.75 (95% CI, 0.57-0.92; P = .01) and RESP scores (AUC 0.81 (95% CI, 0.67-0.95; P = .035) have been developed to aid decision-making regarding the use of VV-ECMO based on best predicted outcomes. 46 The survival after veno-arterial-ECMO (SAVE) score can be used to identify patients that would benefit more from VA-ECMO and balance use with the availability of resources. 47  F I G U R E 2 Algorithm for decision-making regarding ECMO provision in COVID-19 patients. PaO 2 :FiO 2 = ratio of partial pressure of oxygen in arterial blood to the fractional concentration of oxygen in inspired air. PaCO 2 = partial pressure of carbon dioxide in arterial blood. EMCO = extracorporeal membrane oxygenation. PEEP = positive end-expiratory pressure. *L-phenotype has been associated with preserved lung compliance and shown to have favorable outcomes with ECMO. 21 Adapted from the Extracorporeal Life Support Organisation (ELSO) 31 HAIDUC ET AL.

| FUTURE RESEARCH
The use of ECMO in COVID-19 patients is a topic that still requires extensive research before significant recommendations can be made.
More data are required in terms of prognosis for patients to be put on both VV and VA-ECMO so that we can make reliable conclusions on when and on whom this treatment should be used. Future prospective multi-center studies should be done to validate findings in a larger cohort of patients. These should try to quantify the true isolated effect on COVID-19 outcomes, whilst adjusting for significant covariates such as comorbidities and age, which also have a large influence over patients' outcomes.
More research should also be conducted on the effectiveness of blood filters such as Cytosorb, which can be used during ECMO Therefore, we must ensure that a risk-benefit analysis for each candidate is conducted thoroughly so that patients that have increased probability of survival can benefit from this scarce resource.