Use of venovenous extracorporeal life support in pediatric patients for cardiac indications

OBJECTIVE: To describe survival outcomes for pediatric patients supported on venovenous extracorporeal life support with cardiac indications and identify predictors of successful application of venovenous extracorporeal life support. DESIGN: Retrospective review of Extracorporeal Life Support Organization registry database. SETTING: Data reported from extracorporeal membrane oxygenation centers to the Extracorporeal Life Support Organization. PATIENTS: Patients ≤18 yrs of age with cardiac diagnoses initiated on venovenous extracorporeal life support during 1985 to 2007. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Of 8,551 reported pediatric extracorporeal life support cases for cardiac indications during 1985 to 2007, 133 patients received venovenous extracorporeal life support (1.6%); 56 (42%) survived to hospital discharge, comprising the venovenous success group. Of 77 (58%) in the venovenous failure group, 45 (34%) died on venovenous extracorporeal life support and 32 (24%) were converted to venoarterial extracorporeal life support. Median duration of extracorporeal life support course was shorter in the venovenous success group (76 vs. 133 hrs, odds ratio 1.01, 95% confidence interval 1.00-1.01). In the univariate analysis, patients in the venovenous failure group had lower median arterial pH (odds ratio 0.06, 95% confidence intervals 0.01-0.61) and higher PaO(2) (odds ratio 1.02, 95% confidence interval 1.00-1.04). Complications from extracorporeal life support, including receipt of renal replacement therapy (odds ratio 4.35, 95% confidence interval 1.87-10.11), surgical hemorrhage (odds ratio 2.56, 95% confidence interval 1.05-6.25), use of inotropic infusions (odds ratio 2.53, 95% confidence interval 1.24-5.15), and infections (odds ratio 4.99, 95% confidence interval 1.07-23.25), were associated with increased odds for venovenous failure. In a multivariable model, the highest PaO(2) (PaO(2) ≥52 torr) compared to the lowest (PaO(2) ≤ 22 torr) (odds ratio 3.75, 95% confidence interval 1.11-12.57), and use of renal replacement therapy (odds ratio 4.35, 95% confidence interval 1.8710.11) were associated with increased odds of venovenous failure. CONCLUSION: Venovenous extracorporeal life support appears to be an appropriate choice in some children with cardiac failure but better definition of this population is needed.