Predictors of In-Hospital Mortality and Mid-term Outcomes of Patients Successfully Weaned from Veno-arterial Extracorporeal Membrane Oxygenation

Abstract Objectives There is limited evidence to guide the decision to proceed with weaning from venoarterial extracorporeal membrane oxygenation (VA-ECMO) and ∼30% of patients weaned “successfully” do not survive to hospital discharge. We evaluated predictors of in-hospital mortality and mid-term outcomes of patients successfully weaned from VA-ECMO after support for cardiogenic shock, surviving more than 24 hours after weaning, with the aim of improving patient selection for durable weaning. Methods We performed a retrospective analysis of 92 patients supported on VA-ECMO and successfully weaned between January 2013 and February 2018. Survival was estimated by the Kaplan-Meier method. Predictors of in-hospital mortality were identified using a Cox proportional hazards model and an Akaike Information Criterion-selected multivariate model. Results Overall survival at hospital discharge was 64.2%; survival was 54.6% 1-year post-support and 51.4% 3-years post-support. History of diabetes, previous myocardial infarction, prolonged ECMO support, and hypoxemia at ECMO weaning were independent predictors of in-hospital mortality. At mid-term follow-up, New York Heart Association class I heart function was observed in 53% of patients, class II in 19%, class III in 16% and class IV in 12%. Average left ventricular ejection fraction was 46.5% ± 18.2%, and 50% of the patients had been readmitted to the hospital due to heart failure. Conclusions Durable ECMO weaning with acceptable mid-term functional status is obtainable in well selected patients. Previous myocardial infarction, diabetes, prolonged ECMO support and pulmonary dysfunction strongly predicted in-hospital mortality after VA-ECMO weaning. In this high-risk situation other heart replacement therapies should be considered.