Extracorporeal Membrane Oxygenation for Treatment of Perioperative Cardiogenic Shock - Short and Long-Term Outcomes in 576 Patients

Purpose The aim of this study was to analyze short and long-term outcomes in patients who were treated with veno-arterial extracorporeal membrane oxygenation (vaECMO) for treatment of perioperative low-output-syndrome (LOS) and identify risk factors for in-hospital mortality. Methods All consecutive patients who received vaECMO during or after cardiac surgery at a high volume center between 01/2008 and 12/2017 were identified. This patient cohort was characterized and long-term survival (9 years) was analyzed. Furthermore, a regression analysis was performed to identify risk factors for in-hospital mortality. Results n=576 patients were treated with vaECMO. The mean age was 65±12 years and 37.2% were female. 46.9% underwent elective surgery. Patients underwent isolated coronary bypass in 21.5%, aortic valve replacement in 7.3%, mitral valve replacement or repair in 5%, tricuspid replacement or repair in 2.2% and heart transplantation in 13%. The median logistic EuroSCORE was 23.9% (IQR 10.3-46.2). vaECMO was implanted peripherally through femoral vessels in 56.4% and centrally in 34.9%. 43% could be weaned off vaECMO, while 4.6% were switched to a durable ventricular assist device and 0.4% underwent heart transplantation. In 42%, treatment was terminated at the time of vaECMO explantation. In-hospital and 1-year mortality were 66% and 79.9%, respectively. In univariable analysis, age, severe aortic valve stenosis and EuroSCORE II emerged as risk factors for in-hospital mortality. After multivariable adjustment, age and EuroSCORE II persisted as independent risk factors for in-hospital mortality. Estimated 9-year survival was 8%. For patients who were discharged alive, estimated 9-year survival was 24.8%. Conclusion After treatment of perioperative LOS using vaECMO, 34% of patients could be discharged alive. Overall, in-hospital and long-term mortality were high but acceptable for patients who survived the in-hospital period. In multivariable regression, we found age and EuroSCORE II to be associated with in-hospital death but could not identify more specific risk factors.