033 * RISK FACTORS ASSOCIATED WITH ADVERSE OUTCOME FOLLOWING EXTRACORPOREAL MEMBRANE OXYGENATION SUPPORT: ANALYSIS FROM 360 CONSECUTIVE PATIENTS

Abstract

Objectives: Risk factors for adverse outcome after extracorporeal membrane oxygenation applied as extracorporeal life support (ECLS) are yet to be defined. For this purpose we reviewed our institutional data from more than a decade of ECLS. Methods: Between December 2001 and January 2013, 360 consecutive highest-risk patients received ECLS for post-cardiotomy cardiogenic shock. Patient demographics, clinical characteristics, ECLS-related morbidity, as well as in-hospital and late mortality were analysed. Multivariate logistic regression analysis was performed to identify independent predictors of adverse outcome (failed ECLS weaning, in-hospital mortality). Results: Mean age was 62 ± 17 years, 76% were male and mean preoperative ejection fraction was 35% ± 16%. ECLS was established through peripheral (90%) or central thoracic cannulation. The duration of ECLS was 7 ± 4 days. Intra-aortic balloon pumps were implanted in 31%. ECLS weaning was successful in 58% and 30% could be discharged. The main cause of death was sepsis (69%). Overall, major cerebrovascular events occurred in 12% (bleeding 3%, embolic 9%), limb ischaemia in 13%, gastrointestinal complications in 16% and renal replacement therapy in 61%. Independent risk factors for adverse outcome were prior cardiorespiratory resuscitation (OR 4.1), pH <7.1 (OR 2.8), serum lactate >120 mg/dl (OR 2.6), noradrenalin dosage >0.6 μg/kg/min (OR 2.4), and age >75 years (OR 2.0). Kaplan-Meier estimates for long-term survival were 26 ± 3% at one year and 22 ± 2% at five years. Conclusions: ECLS therapy offers survival to one-third of patients with otherwise fatal prognosis. Procedural mortality is low and morbidity at the implantation site typically controllable. Thus, prolonged metabolic deterioration in combination with high-dose vasopressor support prior to ECLS therapy should be avoided, particularly in younger patients.