Outcome of Emergency Pulmonary Lobectomy under ECMO Support in Patients with COVID-19

Background Not much is known about the results of nonelective anatomical lung resections in coronavirus disease 2019 (COVID-19) patients put on extracorporeal membrane oxygenation (ECMO). The aim of this study was to analyze the outcome of lobectomy under ECMO support in patients with acute respiratory failure due to severe COVID-19. Methods All COVID-19 patients undergoing anatomical lung resection with ECMO support at a German university hospital were included into a prospective database. Study period was April 1, 2020, to April 30, 2021 (first, second, and third waves in Germany). Results A total of nine patients (median age 61 years, interquartile range 10 years) were included. There was virtually no preexisting comorbidity (median Charlson score of comorbidity 0.2). The mean interval between first positive COVID-19 test and surgery was 21.9 days. Clinical symptoms at the time of surgery were sepsis (nine of nine), respiratory failure (nine of nine), acute renal failure (five of nine), pleural empyema (five of nine), lungartery embolism (four of nine), and pneumothorax (two of nine). Mean intensive care unit (ICU) and ECMO days before surgery were 15.4 and 6, respectively. Indications for surgery were bacterial superinfection with lung abscess formation and progressive septic shock (seven of nine) and abscess formation with massive pulmonary hemorrhage into the abscess cavity (two of nine). All patients were under venovenous ECMO with femoral-jugular configuration. Operative procedures were lobectomy (eight) and pneumonectomy (one). Weaning from ECMO was successful in four of nine. In-hospital mortality was five of nine. Mean total ECMO days were 10.3 +/- 6.2 and mean total ICU days were 27.7 +/- 9.9. Mean length of stay was 28.7 +/- 8.8 days. Conclusion Emergency surgery under ECMO support seems to open up a perspective for surgical source control in COVID-19 patients with bacterial superinfection and localized pulmonary abscess.