Insulin action to augment glucose utilization (Rd) and suppress endogenous glucose production is not directly determined by changing plasma insulin, but rather by that insulin which traverses the capillary endothelial boundary to enter the interstitial space bathing insulin-sensitive cells. We have examined the importance of transcapillary insulin transport to the efficiency of insulin action by sampling insulin in thoracic duct lymph, believed to represent interstitial fluid, during euglycemic glucose clamps or intravenous glucose tolerance tests (IVGTTs) in conscious dogs. During clamps (insulin infusion: 0.6 mU/min per kg), we observed a 3:2 gradient between plasma and lymph insulin both at basal and at hyperinsulinemic steady state, which was reestablished after termination of infusion. No such gradient was observed for inulin, a non-metabolizable diffusionary marker infused along with insulin. Furthermore, lymph insulin was proportional to Rd during clamps, and these two independently measured variables were strongly correlated (r = 0.96). These results indicate that transcapillary insulin transport is rate-limiting for insulin action during clamps. Compartmental analysis of insulin and inulin data was consistent with receptor-mediated transport of insulin across capillary endothelium from plasma to interstitium. Glucose tolerance tests were subjected to minimal model analysis, which yields the insulin sensitivity index (SI) and X(t), the rate of net glucose disposition during the intravenous test. Preliminary data reveal a striking similarity between the time courses of X and measured lymph insulin during IVGTTs, evidence that interstitial insulin is virtually at equilibrium with the rate of glucose uptake. Thus, even in the non-steady state, transendothelial insulin transport is the rate determining step for insulin action.(ABSTRACT TRUNCATED AT 250 WORDS)