Comparison of three PK/PD models for glycated haemoglobin in diabetes type 2 patients treated with lixisenatide
Valerie Nock (1,2), Diether Rüppel (3), Charlotte Kloft (1)
(1) Dept. Clinical Pharmacy & Biochemistry, Freie Universitaet Berlin and (2) Graduate Research Training program PharMetrX, Germany, (3) Sanofi-Aventis Deutschland GmbH Frankfurt, Germany
Objectives: Lixisenatide, a new glucagon-like-peptide receptor agonist, is known to act on the fasting plasma glucose (FPG) and on postprandial glucose concentrations (PPG) [1]. The aim of this work was to investigate different models to mechanistically describe the effect of lixisenatide on glycated haemoglobin (HbA1c) in type 2 diabetes (T2D) patients..
Methods: Information on HbA1c values from 162 patients from 2 randomised, double-blind, placebo-controlled, parallel-group studies in T2D patients inadequately controlled on metformin and treated with doses of 5, 10, 20 or 30 µg once or twice daily was analysed using NONMEM 7.1.. Studies were sponsored by Sanofi-Aventis. Three models were investigated: (i) a turnover model (TO) with a sigmoidal inhibitory drug effect on the glycation rate (KGL), (ii) an FPG dependent lifespan model (LS) [2] and (iii) an extended version (ELS) of this model with an additional, FPG independent glycation rate (KGL2) which was linked to lixisenatide concentrations via a sigmoidal Imax model. Information on PK and PD (model for FPG) parameters from previous analysis [3,4] was used as input following a sequential modelling approach. Model comparison was guided by AIC, GOF plots, acceptable precision of parameter estimates and shrinkage..
Results: Due to stability reasons and based on estimates the lifespan of the erythrocytes was fixed to 101 d (LS, ELS), the rate of erythrocytes entering the circulation to 1.16 g/L/d (LS) and the exponent (γ) linking FPG to KGL was fixed to 0.75 (ELS). The TO model described the data sufficiently well resulting in the lowest AIC. HbA1c values were not sufficiently described when taking only FPG as a predictor into account (LS) whereas the introduction of KGL2 to the model (ELS) improved the fit considerably. The inter-individual variability on IC50 (62.8 ng/L) of KGL2 was estimated to be high (108%) possibly representing the remaining ability of insulin secretion of the patients. Additionally the ELS model enabled the separation and quantification of 2 glycation pathways, an FPG dependent and independent one. The latter, possibly being attributed to PPG, explained about half of the reduction in HbA1c..
Conclusions: For lixisenatide the TO model and the ELS model were superior compared to the LS model which originally was developed for tesaglitazar, a PPAR α/γ agonist with probably no action on PPG. The ELS model was most mechanistically motivated and enabled the quantification of an FPG dependent and FPG independent glycation pathway..
References:
[1] Christensen M, Knop FK, Vilsboll T, Holst JJ. Lixisenatide for type 2 diabetes mellitus. Expert Opin. Investig. Drugs 20 (49): 549-557 (2011).
[2] Hamrén B, Björk E, Sunzel M, Karlsson MO. Models for Plasma Glucose, HbA1c, and Hemoglobin Interrelationship in Patients with Type 2 Diabetes Following Tesaglitazar Treatment. Clin Pharmacol Ther. 84: 228-35 (2008).
[3] Frank T, Rüppel D, Weber W. Population Pharmacokinetics of Lixisenatide, a once-daily Human Glucagon-Like Peptide-1 Analoge, in Patients with Type 2 Diabetes. In preparation.
[4] Rüppel D, Weber W, Liu YH, Steinstraesser A. PK/PD Modelling of Lixisenatide in Diabetes Type 2. 6th Interntional Symposium on Measurement and Kinetics of in Vivo Drug Effects. 21-24 April 2010, Noordwijkerhout, Netherlands. Advances in Simultaneous Pharmacokinetic/Pharmacodynamic Modeling, edited by Danhof M, van der Graaf PH, Holford NHG. Part 2, C43, 227-228.
