Diabetic patients manifest increased vascular permeability. To determine whether insulin per se might increase vascular permeability, five nondiabetic men were studied by the hyperinsulinemic-euglycemic clamp technique. Each subject received a 0.72-nmol/kg body wt i.v. insulin bolus, followed by a 72-pmol.kg-1.min-1 insulin infusion for 4 h. Euglycemia was maintained by the Biostator glucose controller. At 7 h of study, 10 microCi i.v. 125I-labeled albumin was injected as bolus dose. Frequent blood samples were drawn during the next 70 min for determination of the transcapillary escape rate (TER) of albumin. Subjects returned 1-2 wk later for a control study, during which 0.45% saline was infused at a rate identical to the dextrose and insulin infusion rates during the hyperinsulinemic clamp. The mean +/- SE serum insulin levels during the hyperinsulinemic clamp and saline infusion were 9786 +/- 126 and 46 +/- 4 pM, respectively, whereas serum glucose during the two sessions was similar (5.0 +/- 0.2 vs. 4.8 +/- 0.1 mM, NS). Identical fluid volumes were infused during the two sessions (1767 +/- 197 ml/7 h), and urine outputs did not differ significantly (1615 +/- 309 vs. 1035 +/- 248 ml/7 h). The TER of albumin was greater in all five men after hyperinsulinemia than after saline infusion (18.3 +/- 2.7 vs. -2.8 +/- 2.3%/h, P = 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)