Assessment of minimal model applicability to longitudinal studies.

摘要

The performance of the minimal model of glucose kinetics on longitudinal studies is being evaluated by use of Monte Carlo simulations. A new method for estimating SI is being proposed, based on the fact that glucose effectiveness does not change considerably during the time between clinical tests. Results show that the minimal model has the potential to represent the insulin sensitivity index (SI) accurately (r = 0.90--0.96) even in the presence of large levels of measurement error (2% or 4%). However, high variability on the insulin profile induces a slight underestimation of the insulin sensitivity index (SIMINMOD = 0.8 SIth + 0.19). The method has been applied to simulated data for both the insulin-modified and regular IVGTT protocols and showed that the minimal model reproduces the insulin sensitivity indices accurately for both tests.