Evolving outcomes of extracorporeal membrane oxygenation support for severe COVID-19 ARDS in Sorbonne hospitals, Paris

Schmidt, Matthieu; Langouet, Elise; Hajage, David; James, Sarah Aissi; Chommeloux, Juliette; Bréchot, Nicolas; Barhoum, Petra; Lefèvre, Lucie; Troger, Antoine; de Chambrun, Marc Pineton; Hékimian, Guillaume; Luyt, Charles-Edouard; Dres, Martin; Constantin, Jean-Michel; Fartoukh, Muriel; Leprince, Pascal; Lebreton, Guillaume; Combes, Alain

BACKGROUND: Extracorporeal membrane oxygenation (ECMO) was frequently used to treat patients with severe coronavirus disease-2019 (COVID-19)-associated acute respiratory distress (ARDS) during the initial outbreak. Care of COVID-19 patients evolved markedly during the second part of 2020. Our objective was to compare the characteristics and outcomes of patients who received ECMO for severe COVID-19 ARDS before or after July 1, 2020. METHODS: We included consecutive adults diagnosed with COVID-19 in Paris–Sorbonne University Hospital Network ICUs, who received ECMO for severe ARDS until January 28, 2021. Characteristics and survival probabilities over time were estimated during the first and second waves. Pre-ECMO risk factors predicting 90-day mortality were assessed using multivariate Cox regression. RESULTS: Characteristics of the 88 and 71 patients admitted, respectively, before and after July 1, 2020, were comparable except for older age, more frequent use of dexamethasone (18% vs. 82%), high-flow nasal oxygenation (19% vs. 82%) and/or non-invasive ventilation (7% vs. 37%) after July 1. Respective estimated probabilities (95% confidence intervals) of 90-day mortality were 36% (27–47%) and 48% (37–60%) during the first and the second periods. After adjusting for confounders, probability of 90-day mortality was significantly higher for patients treated after July 1 (HR 2.27, 95% CI 1.02–5.07). ECMO-related complications did not differ between study periods. CONCLUSIONS: 90-day mortality of ECMO-supported COVID-19–ARDS patients increased significantly after July 1, 2020, and was no longer comparable to that of non-COVID ECMO-treated patients. Failure of prolonged non-invasive oxygenation strategies before intubation and increased lung damage may partly explain this outcome. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13054-021-03780-6.