The general use of first order kinetics to describe the disappearance of insulin from plasma in man contrasts the available evidence of saturation kinetics for insulin. In order to bridge this gap, we have put forward three alternative models of insulin kinetics. Model 1 considers the combined existence of first order and saturation (Michaëlis-Menten) kinetics. Model 2 considers saturation kinetics alone. Model 3 considers first order kinetics alone. The validity of the models was studied in normal and type I (insulin-dependent) diabetic subjects. Sequential constant intravenous infusions of insulin at different rates were used to achieve different levels of steady state plasma insulin concentrations, while the glycaemic level (usually normoglycaemia) was maintained by a glucose clamp. Appropriate validation procedures demonstrated that the model of saturation kinetics alone (model 2) was superior to the other models in normal subjects at physiological and supraphysiological plasma insulin concentrations, and in diabetic patients at supraphysiological concentrations. The minimal model at physiological insulin concentrations in type I diabetic patients was that of first order kinetics (model 3). The kinetics of insulin was independent of the species of insulin (human or porcine) in both study groups. The actual glycaemic clamp level (normoglycaemia and moderate hyperglycaemia) did not influence the insulin disappearance rate. Binding of insulin to its receptor is considered to be the initial step in insulin degradation. Saturation kinetics of insulin may therefore be influenced by the saturation of binding of insulin molecules to their receptors. We found values of Km (i.e. the concentration of plasma insulin at which the insulin disappearance rate is half maximal) in normal subjects comparable to values of Kd (i.e. the dissociation constant for insulin-receptor binding) from receptor studies in isolated cells. Changes in regional (hepatic and/or renal) blood flow at hyperinsulinaemia represent an alternative explanation to a model of saturation kinetics. Increases in Vmax (i.e. the maximal insulin disappearance rate) and Km with increasing insulin dose were observed in normal subjects. This finding suggests that insulin may disappear from plasma by more than one saturable pathway. Additional studies are needed to confirm this observation. The clearance rate of insulin at infinitesimal plasma insulin concentrations (Vmax/Km) was 28 +/- 8 ml.kg-1.min-1 in normal subjects. This value is higher than most clearance rates previously reported in studies using first order kinetics. The clearance rate of insulin in type I diabetic patients was 20 +/- 4 ml.kg-1.min-1, corresponding to a reduction in clearance of 30% compared to normal subjects.(ABSTRACT TRUNCATED AT 400 WORDS)