Alterations of myocardial capillary permeability by albumin in the isolated, perfused rabbit heart.

1. Capillary permeability-surface area products for 22Na, 51Cr-EDTA (mol. wt. 357), [57Co]cyanocobalamin (mol. wt. 1353) and 125I-labelled insulin (mol. wt. approximately or equal to 6000) were measured using the single-passage, multiple-tracer dilution technique in isolated rabbit hearts perfused at constant flows between 0.2 and 4.7 ml. min-1 . g-1. 2. In hearts perfused with a Krebs-Ringer solution containing bovine albumin (10 g . l . -1), the permeability-surface area products for 51Cr-EDTA and [57Co]cyanocobalamin increased as the perfusion rate increased, but reached constant values at flows above 2 ml . min-1 . g-1. For 125I-labelled insulin a diffusion-limited value of 0.06 +/- 0.02 ml . min-1 . g-1 (mean +/- S.E., n = 10) was measured at significantly lower perfusion rates. As the value for 22Na increased continuously with increments in flow, only a flow-limited value could be estimated. 3. When hearts were initially perfused with albumin (10 g . l . -1) and then with an albumin-free Krebs-Ringer solution, a significant increase in the permeability-surface area for 22Na, 51Cr-EDTA and [57Co]cyanocobalamin was observed. 4. In hearts perfused with albumin capillary permeability coefficients calculated for 22Na, 51Cr-EDTA. [57Co]cyanocobalamin and 125I-labelled insulin were respectively: 10.5, 3.5, 2.1 and 0.21 x 10(-5) cm.sec-1. 5. These findings confirm that bovine albumin reduces the permeability of myocardial capillaries to hydrophilic solutes of varying molecular sizes and this effect may be the result of an interaction of albumin with the pathways for transcapillary exchange.