AKA-TPG: a program for kinetic and epidemiological analysis of data from labeled glucose investigations using the two-pool model and database technology.

BACKGROUND The Two-Pool Glucose (TPG) model has an important role to play in diabetes research since it enables analysis of data obtained from the frequently sampled labeled (hot) glucose tolerance test (FSHGT). TPG modeling allows determination of the separate effects of insulin on the disposal of glucose and on the hepatic production of glucose. It therefore provides a basis for the accurate estimation of glucose effectiveness, insulin sensitivity, and the profile of the rate of endogenous glucose production. Until now, there has been no program available dedicated to the TPG model, and a number of technical reasons have deterred researchers from performing TPG analysis. METHODS AND RESULTS In this paper, we describe AKA-TPG, a new program that combines automatic kinetic analysis of the TPG model data with database technologies. AKA-TPG enables researchers who have no expertise in modeling to quickly fit the TPG model to individual FSHGT data sets consisting of plasma concentrations of unlabeled glucose, labeled glucose, and insulin. Most importantly, because the entire process is automated, parameters are almost always identified, and parameter estimates are accurate and reproducible. AKA-TPG enables the demographic data of hundreds of individual subjects, their individual unlabeled and labeled glucose and insulin data, and each subject's parameters and indices derived from AKA-TPG to be securely stored in, and retrieved from, a database. We describe how the stratification and population analysis tools in AKA-TPG are used and present population estimates of TPG model parameters for young, healthy (without diabetes) Nordic men. CONCLUSION Researchers now have a practical tool to enable kinetic and epidemiological analysis of TPG data sets.