Posthepatic rate of appearance of insulin: measurement and validation in the nonsteady state

To assess the accuracy with which insulin appearance rates in the peripheral circulation can be measured out of steady state, seven conscious dogs were simultaneously infused with somatostatin and insulin at known variable rates. Tritiated insulin was infused concurrently at a constant rate. Insulin rates of appearance were estimated continuously on the basis of a two-compartment model for systemic insulin kinetics. The calculations were performed assuming that insulin kinetics were linear (tracer data not used) and nonlinear or time varying (tracer data used to assess the variation). The average error in areas under the curve was -3.5 +/- 2.5 and 27.0 +/- 14.2% when nonlinear or linear kinetics were assumed. The maximal errors when linearity was assumed was 39.9 +/- 11.3% and decreased to 16.3 +/- 2.6% when the tracer data was used to account for changes in the fractional removal rate of insulin. The accuracy of the linear estimates improved as the fractional removal rate remained closer to constant. These data suggest that a priori assumptions should not be made on the linearity of the insulin system in a given experimental situation.