A three-compartment model of osmotic water exchange in the lung microvasculature

A bolus injection of hypertonic NaCl into the pulmonary arterial circulatio n of an isolated perfused dog lung causes the osmotic movement of water fir st into, and then out of the capillary. The associated changes in blood con stituent concentrations and density are referred to as the osmotic transien t (OT). Measurement of the sound conduction velocity of effluent blood duri ng an OT is a highly sensitive way to monitor water movement between the va scular and extravascular spaces. It was our objective to develop a mathemat ical model that adequately describes this transient change in the sound con duction velocity and evaluate its application under conditions of homogeneo us and heterogeneous capillary flow distributions. The model accounts for o smotic water exchange between the capillary and two parallel extravascular compartments, and includes as parameters the osmotic conductances (sigmaK(1 ),sigmaK(2)) of the two compartments. The osmotic conductance parameters in corporate the filtration coefficient for water and reflection coefficient f or salt for the two pathways of water exchange. The partition of total extr avascular lung water (EVLW) between the two extravascular compartments is a third parameter of the model. The homogeneous model parameter estimates (p er gram wet lung weight +/-95% confidence limits) from the best-fit analysi s of a typical curve were sigmaK(1) = 2.15+/-0.07, sigmaK(2) = 0.03+/-0.00 [ml h(-1) (mosmol/liter)(-1) g(-1)] and V-1 = 23.83+/-0.12 ml, with a coeff icient of variation (CV) of 0.08. The heterogeneous parameter estimates for a capillary transit time distribution with mean transit time (MTTC) = 1.72 s, and relative dispersion (RDC) = 0.35 were sigmaK(1) = 2.38+/-0.05, sigm aK(2) = 0.03+/-0.00 [ml h(-1) (mosmol/liter)(-1) g(-1)], V-1 = 23.91+/-0.08 ml, and CV = 0.05. EVLW was 42.1 ml for both models. We conclude that the three-compartment mathematical model adequately describes a typical OT unde r both homogeneous and heterogeneous blood how assumptions. (C) 2000 Biomed ical Engineering Society. [S0090-6964(00)01108-5].