An integrated model for glucose-insulin regulation to describe oral glucose tolerance test data in healthy volunteers
Hanna E Silber (1,3), Nicolas Frey (2), Mats O Karlsson (1)
(1) Department of biosciences, Uppsala University, Sweden, (2) Hoffmann-La Roche, Basel Switzerland, (3) Current adress Novartis M&S, Basel
Introduction: Oral glucose tolerance tests (OGTT) are commonly used in drug development. The study design is less invasive than the intravenous provocations and often sparse in sampling which makes it attractive both for patients and researchers. An integrated model for glucose and insulin regulation has previously been developed based on different intravenous glucose provocations in healthy volunteers and type 2 diabetes mellitus (T2DM) patients and extended to also include the OGTT in T2DM patients (1, 2). With the extension to also include healthy volunteer OGTT valuable information can be gained regarding differences between healthy and diabetic individuals during oral provocations.
Methods: Data from 23 healthy volunteers receiving an OGTT was included for the analysis. The subjects were given 75 g of glucose solution to drink and blood samples were drawn every 15 minutes during 4 hours following the glucose dose and analyzed for glucose and insulin. Non-linear mixed effects analysis was performed with NONMEM VI (FOCE) and was based on the previous models. The disposition parameters were initially fixed to the final parameters of the intravenous model (1). The validity of this assumption was later tested by estimation of separate disposition parameters for the OGTT. The validity of the included control mechanisms and the need for incorporation of additional control mechanisms were also evaluated. Model extensions for glucose absorption and for the incretin effect on insulin secretion were added and the starting point was the final model for OGTT in patients (2). Complex glucose and insulin profiles, with multiple glucose and insulin concentration peaks, were observed in most individuals. Both semi-mechanistic and empirical models were evaluated for the description of this behavior. The incretin effect was included as a direct effect on insulin secretion, which was previously done also in the patient population (2). The predictive properties of the final model were evaluated by simulation through the visual predictive check (VPC) and the uncertainty in parameter estimates was quantified by a bootstrap.
Results: The OGTT data was successfully described by the basic intravenous model with extensions for the absorption, the incretin effect on insulin secretion and an additional control effect on glucose production. The final model could successfully capture the multiple peaks in glucose and insulin plasma concentrations seen in most individuals. An empirical flexible input model (3) with 12 steps was used to describe the absorption phase and revealed a complex absorption pattern in line with previous literature indicating inhibition of gut emptying of glucose. Most patients displayed at least a biphasic absorption pattern. The bioavailability was estimated to 72% which is in line with previous results in patients where it was estimated to be 81%. The incretin effect was described using a linear function of glucose absorption rate and was incorporated as a direct effect on insulin secretion. An additional control mechanism of insulin on glucose production was included and was found to be important for the description of the hypoglycemic episode seen in most individuals during the second half of the experiment. The inhibitory effect of insulin on glucose production is a well known effect of insulin. In previously publications with this model it has, however, not been possible to separate it from the effect of glucose on its own production. The effect of insulin on glucose production was found to be stronger than the previously included effect of glucose and to affect the glucose production mainly during the first half of the experiment. Insulin dependent clearance in patients has previously been found to be higher during oral provocations compared to intravenous (2). In this study it was however found to be similar. A possible explanation for the difference could be that in the patient population it was not possible to identify an inhibitory effect on glucose production and that the higher insulin dependent clearance was needed to compensate for the constant glucose production. In contrast to the previously analyzed intravenous data in healthy volunteers it was not possible to identify the first-phase secretion of insulin during the oral provocation experiment. As the first phase insulin response has been shown to be important for glucose regulation it is likely that it is present also in this population but that it is described by the incretin effect on insulin secretion which sets in quickly after the administration of the glucose dose. The VPC showed that the model was able to predict real-life like data although the variability in the simulated data of glucose was somewhat higher compared to the observed data. The bootstrap showed that most parameters were estimated with good precision. Population parameters were in general estimated with a standard error of less than 35 %.
Conclusions: In conclusion, a previously developed model has been extended to also include the OGTT in healthy volunteers. New structures include description of the biphasic absorption profiles and identification of the inhibitory insulin effect on glucose production. As the previously developed model in T2DM patients this model could become important in drug development.
References:
[1] H. E. Silber, P. M. Jauslin, N. Frey, R. Gieschke, U. S. H. Simonsson and M. O. Karlsson. An integrated model for glucose and insulin regulation in healthy volunteers and type 2 diabetic patients following intravenous glucose provocations. J Clin Pharmacol 47: 1159-71, 2007.
[2] P. M. Jauslin, H. E. Silber, N. Frey, R. Gieschke, U. S. Simonsson, K. Jorga and M. O. Karlsson. An integrated glucose-insulin model to describe oral glucose tolerance test data in type 2 diabetics. J Clin Pharmacol 47: 1244-55, 2007.
[3] A. Lindberg-Freijs and M. O. Karlsson. Dose dependent absorption and linear disposition of cyclosporin A in rat. Biopharm Drug Dispos 15: 75-86, 1994.