Inadequacy of Perivascular Edema Hypothesis to Account for Distribution of Pulmonary Blood Flow in Lung Edema

Abstract

West and co-workers obtained inversion of pulmonary blood flow per unit lung volume (Q^·/V) in isolated, perfused dog lungs under special pressure-flow conditions in acute edema with high venous pressure. Repeating their experiments, we determined Q^·/V by the ¹³³xenon scanning procedure and weighed the lungs continuously to determine the rate of edema formation. Edema slowly developed spontaneously in six lungs in which pulmonary artery pressure exceeded alveolar pressure, which exceeded venous pressure (zone II). We found no Q^·/V changes. In six other zone II lungs we induced edema rapidly with alloxan. We found no changes in Q^·/V. In six lungs with high venous pressure in which both pulmonary arterial and venous pressures exceeded alveolar pressure (zone III), we obtained flow inversion when driving pressure was less than 3 mm Hg and blood flow less than 100 ml/min. This confirms West's findings but shows that flow inversion is not related to pulmonary edema in general. We rapidly froze two additional lungs in each set. All showed interstitial perivascular edema, particularly below the hilum. Alveolar walls were more congested and blood vessels more distended at all levels in lungs with high venous pressure than in the other types. We found no vascular compression, kinking, or nipping in lungs showing Q^·/V inversion. The perivascular spaces of lungs with spontaneous and alloxan-induced edema were distended with fluid, although flow distribution remained normal. We conclude that closure of dependent lung vessels by the accumulation of interstitial perivascular edema is not an adequate explanation for this phenomenon.