December 12, 2022

Article Review - Journal Rewiew Club

A multi-centre-matched cohort study


Current data are conflicting in terms of outcomes for the benefit of treatment with ECMO in patients with COVID-19. Consensus guidelines have given guarded recommendations for ECMO in selected COVID-19 patients. At the start of the pandemic, a centralised national referral system was set up for the referral of patients to be considered for ECMO in the United Kingdom.

A multi-centre retrospective study was conducted on COVID-19 patients from 111 hospitals and referred to two specialist ECMO centres from March 2020 to February 2021. All referred patients were eligible for inclusion. Patients who did not have COVID-19 or those referred for non-respiratory ECMO were excluded.

The study group were patients who were retrieved for ECMO to a specialist centre. The control group were patients who were not accepted and remained in the referring centres receiving conventional care.

The primary outcome was in-hospital mortality. There were no secondary outcomes.

After online referral, the data were reviewed by the ECMO centre, and specialists in referring centre and ECMO teams decided in the absence of validated scoring systems. The clinical outcome was either retrieved, managed at referring hospital, or perceived futility.

Propensity score matching was used to analyse the added benefit of ECMO treatment. Patients with an outcome of perceived futility were excluded from matching to prevent confounding of too severe disease.

A total of 1363 patients, of which 243 received ECMO, were analysed. 430 patients were excluded as the outcome was perceived as futility. In the paper, there is a comparison made between the 2 waves. Mortality for conventional treatment was increased in wave 2 (51.9% vs 62.4%, p = 0.001) but was not different in the study group (22.9% vs 26.1%, p=0.672).

Three different matching techniques were used. GenMatch resulted in the best combination and resulted in a minor odds ratio of 0.44 (95% CI 0.29-0.68, P<0.001) and an absolute mortality reduction of 18.2%% (44% vs 25.8%, p<0.001) for ECMO treatment.

In this retrospective analysis by Whebbel et al., the use of ECMO at a specialist centre in patients with COVID-19 compared to those without, using propensity score matching, had a significant survival benefit. The results of this study suggest that critically ill COVID-19 patients should be referred to specialised centres for treatment with ECMO.



  • Extensive independent data available
  • Well-defined treatment group (COVID-19)
  • Robust statistical analyses with multiple sensitivity analyses


  • The analysis was limited to in-hospital mortality, 1-year mortality nor functional status, nor morbidity was analysed
  • Residual confounding cannot be excluded and is discussed by the authors
  • Assessment of the pure effect of ECMO cannot
  • Separation between the benefit of pure ECMO treatment and referral to highly experienced ARDS centre unit cannot be made as all patients were retrieved on mobile ECMO
  • Low mortality rates in the treatment group suggest a ‘better’ population reflected in younger age, shorter ventilation before referral, lower SOAFA scores (5 (4-7)) and little to no comorbidities favouring outcome for ECMO
  • The potential positive effect of advice given by the ECMO specialists to the referring clinician may have had a positive effect on mortality for conventional treatment


The benefit of ECMO in COVID-19 has been conflicting. In this retrospective trial, a well-matched cohort of ECMO-referred patients, retrieval on ECMO to a specialist centre had an impressive survival benefit.

This article review was prepared and submitted by Anna Hall, Guy’s & St. Thomas’ NHS Foundation Trust, on behalf of the ESICM Journal Review Club.


Whebell S. et al. Survival benefit of extracorporeal membrane oxygenation in severe COVID-19: a multi-centre-matched cohort study. Intensive Care Med. 2022 Apr;48(4):467-478. doi: 10.1007/s00134-022-06645-w. Epub 2022 Mar 3. Erratum in: Intensive Care Med. 2022 Mar 18;: PMID: 35238946; PMCID: PMC8892395.

Comment on this news