STUDY OBJECTIVE: To evaluate gas exchange, pulmonary function, and lung histology during gas ventilation of the perfluorocarbon-filled lung compared with gas ventilation of the gas-filled lung in severe respiratory failure. STUDY DESIGN: Application of gas (GV) or partial liquid (PLV) ventilation in lung-injured sheep. SETTING: A research laboratory at a university medical center. SUBJECTS: Eleven sheep 17.1 +/- 1.8 kg in weight. INTERVENTIONS: Lung injury was induced by intravenous administration of 0.07 mL/kg oleic acid followed by saline pulmonary lavage. When alveolar-arterial oxygen pressure difference (P[A-a]O2) was 600 mm Hg or more and PaO2 was 50 mm Hg or less with fraction of inspired oxygen of 1.0, bijugular venovenous extracorporeal life support (ECLS) was instituted. For the first 30 min on ECLS, all animals were ventilated with gas. Over the ensuing 2.5 h, ventilation with 15 mL/kg gas was continued without intervention in the control group (GV, n = 6) or with the addition of 35 mL/kg of perflubron (PLV, n = 5). MEASUREMENTS AND RESULTS: At 3 h after initiation of ECLS, Qps/Qt was significantly reduced in the PLV animals when compared with the GV animals (PLV = 41 +/- 13%; GV = 93 +/- 4%; p < 0.005). At the same time point, pulmonary compliance was increased in the PLV when compared with the GV group (PLV = 0.61 +/- 0.14 mL/cm H2O/kg; GV = 0.41 +/- 0.02 mL/cm H2O/kg; p < 0.005). The ECLS flow rate required to maintain the PaO2 in the 50 to 80 mm Hg range was substantially and significantly lower in the PLV group when compared with that of the GV group (PLV = 25 +/- 20 mL/kg/min; GV = 87 +/- 15 mL/kg/min; p < 0.001). Light microscopy performed on lung biopsy specimens demonstrated a marked reduction in lung injury in the liquid ventilated (LV) when compared with the GV animals. CONCLUSION: In a model of severe respiratory failure, PLV improves pulmonary gas exchange and pulmonary function and is associated with a reduction in pulmonary pathology.