ObjectiveTo compare the efficacy of venovenous to venoarterial bypass in an unselected cohort of infants with refractory cardiorespiratory failure.DesignRetrospective cohort analysis.SettingTwo tertiary hospitals capable of providing extracorporeal life support for neonates with acute respiratory failure.PatientsAll San Diego Regional Extracorporeal Membrane Oxygenation (ECMO) Program patients treated after the adoption of a policy which eliminated traditional restrictions to venovenous support.InterventionsVenoarterial or venovenous extracorporeal life support.Measurements and Main ResultsFifty-four infants were treated with venovenous bypass; 30 were treated with venoarterial bypass due to unsuccessful placement of the double lumen venovenous catheter or inability to exclude congenital heart disease before cannulation. No patient required conversion from venovenous to venoarterial ECMO. There were no differences in birth weight, gestational age, diagnosis, or pre-ECMO condition in the two groups. Patients who met ECMO criteria early were more likely to be successfully cannulated with a double-lumen venovenous catheter. Severe hemodynamic compromise was present before cannulation in a comparable percentage of venovenous and venoarterial patients. During venovenous bypass, mean PaO2values were lower but remained in the normoxic range; PaCO2values, ventilatory settings, intravascular volume requirements, inotropic support, and mean duration of ECMO support were not different. The frequency rate of patient and mechanical complications were also comparable, except that the frequency of intravascular thrombosis was significantly lower in patients receiving venovenous ECMO. Survival, the frequency rate of chronic lung disease, and neurodevelopmental outcome were similar in the two groups.ConclusionsWe conclude that venovenous ECMO using a double-lumen venovenous catheter can provide results comparable with venoarterial bypass without the need for carotid artery ligation in an unselected population of neonatal ECMO candidates. In our experience, reported contraindications to venovenous ECMO did not prove to be valid.(Crit Care Med 1996; 24:1678-1683)