Circuit change during extracorporeal membrane oxygenation: single-center retrospective study of 48 changes

BACKGROUND: Bleeding and thrombosis induce major morbidity and mortality in patients under extracorporeal membrane oxygenator (ECMO). Circuit changes can be performed for oxygenation membrane thrombosis but are not recommended for bleeding under ECMO. The objective of this study was to evaluate the course of clinical, laboratory, and transfusion parameters before and after ECMO circuit changes warranted by bleeding or thrombosis. METHODS: In this single-center, retrospective, cohort study, clinical parameters (bleeding syndrome, hemostatic procedures, oxygenation parameters, transfusion) and laboratory parameters (platelet count, hemoglobin, fibrinogen, PaO(2)) were collected over the seven days surrounding the circuit change. RESULTS: In the 274 patients on ECMO from January 2017 to August 2020, 48 circuit changes were performed in 44 patients, including 32 for bleeding and 16 for thrombosis. Mortality was similar in the patients with vs. without changes (21/44, 48% vs. 100/230, 43%) and in those with bleeding vs. thrombosis (12/28, 43% vs. 9/16, 56%, P = 0.39). In patients with bleeding, numbers of bleeding events, hemostatic procedures, and red blood cell transfusions were significantly higher before vs. after the change (P < 0.001); the platelet counts and fibrinogen levels decreased progressively before and increased significantly after the change. In patients with thrombosis, numbers of bleeding events and red blood cell transfusions did not change after membrane change. No significant differences were demonstrated between oxygenation parameters (ventilator FiO(2), ECMO FiO(2), and PaO(2)) and ECMO flow before vs. after the change. CONCLUSIONS: In patients with severe and persistent bleeding, changing the ECMO circuit decreased clinical bleeding and red blood cell transfusion needs and increased platelets and fibrinogen levels. Oxygenation parameters did not change significantly in the group with thrombosis.