O nset of lung edema is usually associated with increase in the pulmonary transvascular flux of water and proteins. Clinical measurement of these parameters may aid in early diagnosis of pulmonary edema, and allow differentiation between "cardiogenic" and "noncardiogenic" types base on the magnitude of the detected changes. We have previously described a noninvasive method for estimating transvascular protein flux in lung (Gorin, A. B., W. J. Weidner, R. H. Demling, and N. C. Staub, 1978. Noninvasive measurement of pulmonary transvascular protein flux in sheep. J. Appl. Physiol. 45: 225-233). Using this method we measured the net transvascular flux of [113mIn]transferrin (mol wt, 76,000 in lungs of nine normal human volunteers. Plasma clearance of [113In]transferrin occurred with a T1/2 = 7.0 +/- 2.6 h (mean +/- SD). The pulmonary transvascular flux coefficient, alpha, was 2.9 +/- 1.4 X 10(-3) ml/s (mean +/- SD) in man, slightly greater than that previously measured in sheep (2.7 +/- 0.7 X 10(-3) ml/s; mean +/- SD). The pulmonary transcapillary escape rate is twofold greater than the transcapillary escape rate for the vascular bed as a whole, indicating a greater "porosity" of exchanging vessels in the lung than exists for the "average" microvessel in the body. Time taken to reach half-equilibrium concentration of tracer protein in the lung interstitium was quite short, 52 +/- 13 min (mean +/- SD). We have shown that measurement of pulmonary transvascular protein flux in man is practical. The coefficient of variation of measurements of alpha (between subjects) was 0.48, and of measurements of pulmonary transcapillary escape rates was 0.39. In animals, endothelial injury commonly results in a two- to threefold increase in transvascular protein flux. Thus, external radioflux detection should be a suitable means of quantitating lung vascular injury in human disease states.