An extra corporeal venovenous bypass circuit (right atrium to pulmonary artery), including an intravascular gas exchanger in a blood chamber with a variable inner diameter, was developed for ex vivo evaluation of the host vessel diameter/intravascular oxygen transfer relationship. Three host vessel diameters mimicking different configurations of the caval axis were studied in three bovine experiments (body weight 82 +/- 3 kg). Blood flow was 3,000 ml/min and device oxygen inflow was 2,300 ml/min. Serial blood samples were taken for 26 mm, 23 mm, and 20 mm inner blood chamber diameters after hemodynamic stabilization before and after exposure of the circulating blood to the intravascular gas exchanger (sampling ports at blood chamber inlet and outlet). Measured oxygen saturation at the blood chamber inlet was 25.0 +/- 11.7% for the 26 mm diameter as compared to 31.7 +/- 12.6% for 23 mm, and 28.7 +/- 9.2% for 20 mm. At the outlet, the corresponding O2 saturations were 34.5 +/- 11.5% for 26 mm, 42.9 +/- 8.8% for 23 mm, and 43.2 +/- 6.2 for 20 mm. Total O2 transfer was 24.9 +/- 11.5 ml/min for 26 mm, 31.9 +/- 7.4 ml/min for 23 mm, and 35.9 +/- 12.2 ml/min for 20 mm (p < 0.05). Likewise, O2 transfer rate was 8.3 +/- 3.8 ml/L, 10.6 +/- 2.4 ml/L, and 12.0 +/- 4.0 ml/L (p < 0.05). Parallel analyses of total CO2 transfer and CO2 transfer rates provided less consistent findings. At 3 L/min, the pressure drop between the inlet and outlet of the blood chamber was 12 +/- 3 mmHg for 26 mm, 26 +/- 1 mmHg for 23 mm, and 38 +/- 2 mmHg for 20 mm diameters (p < 0.001). The authors conclude that oxygen transfer of a given intravascular gas exchanger appears to be indirectly proportional to the host vessel diameter. Increasing blood pressure gradient as a function of decreasing diameter has to be considered in clinical application.