Pore models of sheep lung microvascular barrier using new data on protein tracers

Lung lymph to plasma concentration ( L P ) ratios for endogenous and tracer macromolecules were measured as a function of mean lung microvascular pressure in five chronically instrumented unanesthetized ewes. At an average lung microvascular pressure of 19 cm H 2O, L P ratios (mean ± SD) were: albumin, 0.82 ± 0.06; “globulin,” 0.55 ± 0.05; fibrinogen, 0.25 ± 0.05; and low-density lipoprotein (LDL), 0.25 ± 0.05. All L P ratios decreased proportionally as microvascular pressure increased. Fibrinogen and LDL had identical L P ratios under all conditions despite a presumed difference in their molecular conformations. The predictions of one-, two-, and three-pore models of the lung microvascular barrier were compared with L P ratio data in order to find the optimum number of pores and pore radii that could best describe our findings. No one-pore model could accurately predict all solute L P ratios. Optimum two- and three-pore models were found that could accurately predict all solute data simultaneously. These models predict that 73% of all albumin transport across the lung microvascular barrier occurs by convection. The contribution of convection to macromolecular transport increases as both vascular pressure and macromolecular size increase. We conclude that neither pore stretching, pore area recruitment, nor pressure-independent vesicular transport are necessary to explain macromolecular transport in the sheep lung.