We designed a series of experiments to compare the pulmonary dysfunction observed in models of cardiogenic and noncardiogenic pulmonary edema in chronically instrumented awake sheep. Cardiogenic pulmonary edema was induced by inflating the balloon of a Foley catheter surgically positioned in the mitral valve orifice causing increased left atrial pressure (increases PLA). Noncardiogenic pulmonary edema was induced by intravenous infusion of Perilla ketone (PK). Calculated microvascular pressure remained constant during PK infusion but increased from 9.4 +/- 0.7 to 42.8 +/- 2.4 cm H2O during increases PLA. Comparable increases in lung lymph flow (QL) were observed in the two protocols (five to seven times baseline). Pulmonary edema as quantified by chest radiograph scores increased from 0 (normal) to 2.9 +/- 0.5 and 3.4 +/- 0.1 in the PK and increases PLA groups, respectively. Room air alveolar to arterial oxygen pressure difference (P[A-a]O2) increased from 24 +/- 3 to 46 +/- 7 mm Hg in the PK group and from 23 +/- 4 to 56 +/- 6 mm Hg in the increases PLA group. Dynamic compliance of the lungs (Cdyn) expressed as the percentage of the baseline value decreased to 53 +/- 7 and 50 +/- 7% in the PK and increases PLA groups, respectively. Resistance to airflow across the lungs (RL) increased from 2.5 +/- 0.6 to 3.3 +/- 0.8 cm H2O.L-1.sec-1 in the PK group and from 1.4 +/- 0.3 to 4.2 +/- 1.1 in the increases PLA group. Significant correlations were observed between changes in the severity of pulmonary edema observed on chest radiographs, Cdyn, delta P(A-a)O2, and QL in both the increases PLA groups. We conclude that similar degrees of pulmonary edema, regardless of the mechanism, are associated with similar changes in QL, Cdyn, and delta P(A-a)O2. Hydrostatic pulmonary edema appeared to cause greater changes in RL than that resulting from increased microvascular permeability.