Predictors of Poor Outcomes in Pediatric Venoarterial Extracorporeal Membrane Oxygenation

Abstract

Background: Venoarterial extracorporeal membrane oxygenation (VA-ECMO) provides respiratory and hemodynamic support to pediatric patients in severe cardiac failure. We aim to identify risk factors associated with poorer outcomes in this population. Methods: A retrospective chart review was conducted of pediatric patients requiring VA-ECMO support for cardiac indications at our institution from 2004 to 2015. Data were collected on demographics, indication, markers of cardiac output, ventricular assist device (VAD) insertion, heart transplantation, or left atrial (LA) decompression. Univariate Cox proportional hazards models were used to calculate hazard ratios (HRs) for variables associated with the composite primary outcome of transplant-free survival (TFS). Results: Of the 68 reviewed patients, 65% were male, 84% were white, 38% had a prior surgery, 13% had a prior transplant, 10% had a prior ECMO support, and 87.5% required vasoactive support within six hours of cannulation. The ECMO indications included congenital heart disease repaired >30 days prior (12%), cardiomyopathy (41%), posttransplant rejection (7%), and cardiorespiratory failure (40%). The TFS was 54.5% at discharge and 47.7% at one year. Predictors of transplant and/or death include epinephrine use (hazard ratio [HR] = 2.269, P = .041), elevated lactate (HR = 1.081, P = 0005), and elevated creatinine (HR = 1.081, P = .005) within six hours prior to cannulation. Sixteen (23.6%) patients underwent LA decompression. Placement of VAD occurred in 16 (23.5%) patients, for which nonwhite race (HR = 2.94, P = .034) and prior ECMO (HR = 3.42, P = .053) were the only identified risk factors. Conclusions: Need for VA-ECMO for cardiac support carries high inpatient morbidity and mortality. Epinephrine use and elevated lactate and creatinine were associated with especially poor outcomes. Patients who survived to discharge had good short-term follow-up results.