P1714Short-term survival prognostic factors in patients supported with veno-arterial extracorporeal membrane oxygenator

Abstract

Abstract Introduction Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) provides cardiopulmonary support in different conditions with circulatory compromise, assuming high complication and mortality rates. Purpose Identify predictive factors of short-term survival. Methods Retrospective study including all VA-ECMO implants in a referral hospital. Univariate and multivariate analysis of factors related to discharge survival. Results 85 VA-ECMO were implanted from 2013 to Jan-2019 (table). Survival at discharge was 42.4%. Deaths were due to multiorgan dysfunction syndrome (44.9%), anoxic encephalopathy (18.4%) and bleeding (10.2%). A lower post-implantation LVEF (at weaning or later) was associated with a decrease in survival (p0.002), as well as higher lactate (p<0.001) and LDH (p0.019) or presence of preimplantation cardiac arrest (p<0.001) and its duration (p0.049). The appearance of peripheral ischemia, ischemic/hemorrhagic stroke, bleeding, infections and need for renal replacement therapy were associated with shorter survival (p<0.001 all variables). At the multivariate analysis, a better post-implant LVEF was a protective factor (HR 0.9, IC95% 0.83–0.97, p0,002) and higher pre-implant lactate level was independent predictor of mortality (HR 1.3, IC95% 1.1–1.98, p=0,019). ROC analysis revealed the best cut-off for predicting survival at hospital discharge of postimplant LVEF (33,5%, Sen 86%, Spe 70%, AUC=0.76) and lactate (8.5 mmol/L, Sen 88%, Spe 45%, AUC=0.706) (Figure). Patients' characteristics Baseline characteristics (n=85) Periimplantation characteristics Complications (n,%) Age (years) (mean ± SD) 61.2±10 Peripheral cannulation (n, %) 79 (92.9%) Vascular 16 (18.8%) Male (n, %) 64 (75.3%) Percutaneous implant (n, %) 58 (68.2%) Peripheral ischemia 12 (14.%) Situation at the admission Intraaortic balloon pump (n, %) 40 (47.1%) Critical care patient infections 38 (47.7%) Bridge to recovery (n, %) 73 (85.9%) Noradrenalin (n, %) 71 (83.5%) Minor or mayor bleeding 39 (45.9%) Indication (n, %) Dobutamine (n, %) 71 (83.5%) Need for transfusion 60 (70.6%) Cardiogenic shock 38 (44.7%) Adrenaline (n, %) 26 (20.6%) Ischemic stroke 4 (4.7%) Cardiac arrest 10 (11.8%) Blood test (mean ± SD) Hemorrhagic stroke 3 (3,5%) Electrical storm 7 (8.2%) pH 7.26±0,2 Renal replacement therapy 16 (18.8%) High-risk percutaneous intervention 8 (8.2%) Lactate (mmol/L) 6.56±4,4 Critical patient polyneuropathy 16 (18.8%) Postcardiotomy shock 21 (24.7%) Creatinine (mg/dl) 1.58±1,5 Tracheostomy 14 (16.5%) Others 2 (2.4%) Bilirrubin (mg/dl) 1.1±1 ECMO circuit 7 (8.2%) Preimplant cardiac arrest (n, %) 45 (52.9%) LDH (U/L) 959±67 Cardiac arrest duration (min) (n, %) 29.6±23 Time ECMO support (days) (mean ± SD) 4.84±4 LVEF (%) (mean ± SD) 29.9±17.5 ECMO-CPR (n, %) 19 (22.4%) LVEF weaning (%) (mean+SD) 40.2±15.9 RV dysfunction (n, %) 41 (48.2%) ROC curves for survival at discharge Conclusion VA-ECMO is an effective tool for hemodynamic support in circulatory compromise cases but it has high morbi-mortality. The higher lactate and LDH level, cardiac arrest prior to implantation and the appearance of complications decrease survival significantly. Lactate at implantation and post-implantation LVEF (at weaning or later) were independent predictors of survival.