Cause of In-Hospital Death After Weaning from Venoarterial-Extracorporeal Membrane Oxygenation

Abstract

Purpose A survival gap between weaning from venoarterial-extracorporeal membrane oxygenation (VA-ECMO) and the hospital discharge has been consistently reported. The aim of this study is to investigate the clinical features of patients who underwent successful VA-ECMO decannulation at our institution and to identify the major contributors responsible for adverse outcomes. Methods We retrospectively reviewed all patients supported with VA-ECMO in our institution between January 2013 and June 2020. Only patients that survived VA-ECMO and underwent successful decannulation were included and dichotomized based on survival to hospital discharge: non-survivors versus survivors. The primary study outcome was the cause of death after successful VA-ECMO decannulation. Results Of the 262 adult patients who underwent VA-ECMO decannulation, 72 (27.5%) patients did not survive to hospital discharge. Non-survivors were older (62 vs. 54 years, p < 0.001) and suffering from many pre-existing comorbidities. Pneumonia and sepsis were the most frequent infectious complication and almost twice as likely in non-survivors. Major causes of death were: cardiovascular (31.9%), infections (25.0%) and neurological (20.8%). The survival curve demonstrated that 51.4% of our patients died within 8 days after decannulation. Multivariate analysis identified older age, central venous cannulation, pulmonary bleeding and infection, dialysis after VA-ECMO, sepsis, and ischemic stroke (OR = 7.86, 95% CI: 2.76-2.43, p < 0.001) as factors significantly predisposing to patients’ death. Conclusion In our study, one-third of patients decannulated off VA-ECMO did not survive to hospital discharge due to end-stage heart failure, infections or neurological injury. The first 8 post-decannulation days were recognized as a critical period where thorough strategies to prevent acquired infections and cautious support of end-organ function should be warranted. Future large-scale trials are needed to confirm our results.