Reconstruction of insulin secretion under the effects of hepatic extraction during OGTT: a modelling and convolution approach

The reconstruction of insulin secretion rates deriving from insulin and C-peptide concentrations is studied. Two sets of data simulating profiles of insulin and C-peptide concentrations during glucose administration are generated, one of which is generated using an extended combined model (Watanabe et al., 1998) under the assumption of constant fractional hepatic extraction, while the second data set is from assuming non-constant fractional hepatic extraction. A deconvolution approach based on the extended combined model with the constant extraction is then applied to estimate the rate of insulin secretion. The result indicates that variation in the fractional hepatic extraction significantly effects the reconstruction of insulin secretion. When the convolution approach is modified to accommodate the assumption of non-constant fractional hepatic extraction, it can provide more accurate estimates for the rates of insulin production. The modified approach discussed in this paper may thus offer a better option in the attempt to determine an estimation of insulin secretion crucial for the control of diabetes mellitus.