Evaluation of a Glomerular Filtration Term in the DISST Model to Capture the Glucose Pharmacodynamics of an Insulin-Resistant Cohort

Abstract

Glomerular filtration (kidney clearance) of glucose occurs at high glucose concentrations. Thus, the applicability of glomerular filtration terms in models of insulin and glucose pharmacodynamics (PD) should be investigated. To evaluate such a term, data from 36 insulin sensitivity tests on 12 participants with type 2 diabetes in an Atkins diet intervention study was analysed using three PD models. The models include the dynamic insulin sensitivity and secretion test (DISST) model the DISST model with an added glomerular filtration rate (GFR) term and the well-known Minimal Model (MM). The identified insulin sensitivity values and simulation fit-to-data residuals are analysed to test performance and differences. The Minimal Model produced the best fit-to-data with a median residual of 0mmol/L and IQR of -0.13, 0.18mmol/L. Both DISST models also produced residuals with a median of 0mmol/L but an IQR of approximately -0.41, 0.37mmol/L. However, the DISST derived sensitivity values were considerably more in accordance with expected trends, showing the expected 20-50% increase in sensitivity for most subjects due to the intervention. In contrast, the Minimal Model repeated the variable trade-off issue previously recorded for this model with insulin resistant participants. The Minimal Model sensitivity values were effectively random, and did not capture observable changes in insulin sensitivity of glucose clearance. The addition of a GFR term had a positive impact on the identified insulin sensitivity by shifting some values more toward expected and observable behaviour. However, more data must be made available for an exhaustive investigation of the applicability of this term for this type of usage.