SUMMARY Although the exchange of labeled water between blood and tissue in the heart has usually been assumed to be flow-limited, the outflow patterns of labeled water, relative to intravascular references, in a multiple indicator dilution experiment, have appeared to be anomalous in terms of the models used to explain the transport of less permeable substances. Data showing a change in the shape of the labeled water outflow curve after vasodilation and after the infusion of toxic doses of 2,4-dinitrophenol led us to propose a new model for labeled water permeation which includes barriers at both the capillary wall and the sarcolemmal membrane. This model explains adequately the form of the outflow curve, provides parameters related to the permeability at the two barriers, and gives an estimate of the ratio of the intracellular to interstitial space. Dinitrophenol infused intra-arterially in a dose sufficient to cause S-T elevation in the electrocardiogram is found to reduce the sarcolemmal water permeability by an order of magnitude, but to have no effect on capillary water permeability. We conclude that water transport in the heart is barrier-limited at both the capillary and sarcolemmal membranes and that sarcolemmal water permeability is probably mediated at least in part by a structure sensitive to the effects of dinitrophenol, presumably a protein channel. Since the outflow patterns of inert gases resemble that of labeled water, it is possible that oxygen distribution is also barrierlimited.ALTHOUGH cell membranes are known to be highly permeable to water, they do reduce the rate of diffusion of water molecules to about a hundred thousandth that of free diffusion.1 Despite this fact it has been hypothesized that the membrane permeability is high enough and blood flow is slow enough that the distribution of labeled water in an organ such as the heart, where intercapillary distances are small, is flow-limited at physiologic rates of perfusion.213 At this point in time, however, it has been possible to quantitatively corroborate this assumption only for the liver . 4 This organ is very specialized, in terms of the structure of its basic microcirculatory unit, the hepatic sinusoid. There is no significant barrier to small molecules corresponding to the capillary membrane, and the plasma membranes of the hepatocytes are massively expanded in area by virtue of their innumerable microvillous processes. In addition, the mean sinusoidal transit time is much longer than that in the capillaries of a visceral organ perfused at arterial pressure.On the other hand, there is evidence that the distribution of labeled water in the brain is barrier-limited at high but physiologic perfusion rates.
"7 In the heart, the organ we wish to consider here, Ziegler and Goresky 8 have shown that the shape of the outflow concentration-time curve for labeled water in a multiple indicator dilution experiment could not be explained by the assumption of flow-limited exchange unless the additional assumption of random diffusional capillary...