Recently, venovenous extracorporeal life support (VVECLS) using a double lumen catheter has been clinically used to avoid neurologic complications in the treatment of respiratory failure for neonates. However, recirculation, which is a limiting factor for oxygen delivery, still exists, and thus it does not contribute to oxygenation of the patient. We developed a newly designed double lumen catheter with a double balloon (DBDL) catheter for ECLS vascular access and performed two animal preliminary experiments in normal and hypoxic dog models (normal ventilation and one lung ventilation experiments) to investigate whether the DBDL catheter could prevent recirculation and maintain oxygen delivery to systemic circulation. The DBDL catheter (JCT Co., Hiroshima, Japan) of 15 Fr was fabricated from silicone. It consists of two lumens for drainage and return of blood with two balloons (distal and proximal balloons) that prevent oxygenated blood mixing with unoxygenated blood. VVECLS using a DBDL catheter was performed in 13 mongrel dogs (8 dogs for normal ventilation experiment weighing 12.9 +/- 1.6 kg [mean +/- SD], 5 dogs for one lung ventilation experiment weighing 16.6 +/- 2.5 kg [mean +/- SD]) under anesthesia in the two experiments. The bypass flow ranged from 10-40 ml/kg per minute in the normal ventilation experiment. VVECLS in the one lung ventilation experiment was performed with maximal bypass flow for 6 hours (ranged from 25.2 +/- 8.0-28.3 +/- 8.7 ml/kg per minute at balloon inflation and deflation). Recirculation and oxygen transfer of artificial lung with or without balloon inflation during VVECLS were studied. Recirculation decreased with balloon inflation at varied bypass flows during VVECLS in the normal ventilation experiment (varied from 1.5 +/- 14.6-12.8 +/- 16.7%) and for 6 hours after VVECLS initiation in the one lung ventilation experiment (varied from 12.2 +/- 12.2-19.2 +/- 6.5%). In particular, the values at 3 and 6 hours were significantly lower than that of balloon deflation in the one lung ventilation experiment. The difference in O2 content between inlet and outlet in the artificial lung with balloon inflation was significantly higher than that of balloon deflation (varied from 3.7 +/- 1.8-4.8 +/- 1.9 ml/dl, p < 0.05) at the bypass flow of 10-30 ml/kg per minute in the normal ventilation experiment and at 5 hours after VVECLS initiation in the one lung ventilation experiment (varied from 10.6 +/- 1.6-11.7 +/- 1.8 ml/dl). The blood gas analysis of systemic circulation with balloon inflation revealed that the values of PaO2 (varied from 83.8 +/- 11.4-96.9 +/- 23.4 mm Hg) and PaCO2 (37.7 +/- 9.2-40.4 +/- 11.8 mm Hg) were higher and lower, respectively, compared with balloon deflation. In particular, PaO2 level was significantly higher than that of the preECLS value at the bypass flow of 20-40 ml/kg per minute (varied from 83.8 +/- 11.4-96.9 +/- 23.4 mm Hg, p < 0.05). In the one lung ventilation experiment, systemic PaO2 and PaCO2 levels at balloon inflation were higher and lower, respectively, compared with balloon deflation during VVECLS for 6 hours. At balloon inflation, the value of PaO2 at 6 hours after VVECLS initiation was significantly higher than that at balloon deflation. A newly designed DBDL catheter for ECLS vascular access successfully reduced recirculation and maintained oxygen delivery to systemic circulation during VVECLS. These results suggest that a high bypass flow may not be necessarily required in terms of oxygen delivery to systemic circulation when the DBDL catheter was used as an ECLS vascular access.