OBJECTIVE To evaluate a two-isotope technique to detect graded changes in pulmonary microvascular permeability. METHODS Open experimental study. METHODS University hospital, Sweden. METHODS Fifty-seven sheep. METHODS Catherisation of one carotid artery, pulmonary artery and central vein. Control group 1 (n = 10), control group 2 (n = 12) had an additional pulmonary artery balloon catheter inserted, experimental group 3 (n = 9) was given oleic acid 0.005 mg kg-1 BW, experimental group 4 (n = 12) received oleic acid 0.02 mg kg-1 BW and experimental group 5 (n = 11) 0.05 mg kg-1 BW. Groups 3-5 had all PA catheters. All animals were intubated and ventilated artificially. Duration of experiments was 6 hours. METHODS Transferrin was labelled in vivo with 113mIn chloride and erythrocytes with 99mTc following injection of stannous chloride. External gamma counting was corrected for background, decay and scatter. Blood activity was used as reference. Normalised slope index (NSI) and transferrin leak index (TLI) were calculated as measures of pulmonary microvascular permeability. RESULTS A graded response in both NSI and TLI was found. Insertion of the PA catheter (group 2) significantly increased NSI from (group 1) (1.4 (0.1)) 10(-4) min-1 to (11 (2)) 10(-4) min-1 (p < 0.05). TLI increased significantly from (9 (2)) 10(-4) min-1 to (72 (13)) 10(-4) min-1. Oleic acid increased NSI significantly to (13 (1)) 10(-4) min-1, (32 (2)) 10(-4) min-1 and (61 (5)) 10(-4) min-1 in groups 3-5, respectively. Corresponding values for TLI were (95 (13)) 10(-4) min-1, (162 (6)) 10(-4) min-1 and (228 (26)) 10-4 min-1, respectively. CONCLUSIONS The double-isotope technique of external monitoring of permeability changes to protein in the lungs was sensitive to pick up graded increments in leakage, related in a dose-dependent way to lung injury.