Clémence Rigaux, Bernard Sébastien, Michel Dubar
Sanofi R&D, Chilly-Mazarin, France
Objectives: To extend a recently published disease progression [1] model for lixisenatide, a prandial GLP-1 receptor agonist [2], in order to explicitly account for 2H post prandial glucose (PPG) data for describing and predicting HbA1c over long term.
Methods: Fasting plasma glucose (FPG), 2H glucose excursion (GLUEX) data, defined as the difference between FPG and PPG, fasting serum insulin (FSI) and HbA1c data were collected in 5 phase II and III clinical trials. Data from another phase III trial were used for external validation. Observations from 2.470 patients were used. Data were analyzed with a non-linear mixed modelling approach using Monolix 4.1.3. and SAEM algorithm. FPG and FSI were described by the former semi-mechanistic model, while GLUEX was empirically modelled. HbA1c was then described with a turn-over model linked to FPG and PPG.
Results: An immediate drug effect was identified on GLUEX, with dose response modelled with an Emax function, leading to a predicted mean decrease of 2.8 mmol/L for a 20 μg QD lixisenatide dose. An additional reduction of 1 mmol/L accounted for a placebo effect. The model predicted a disease progression corresponding to a mean increase of 0.5 mmol/L after 24 weeks. Inter-individual variability was identified on baseline GLUEX value and on ED50, with baseline HbA1c and Asian race as covariates, and on the disease progression coefficient.
HbA1c synthesis was described as a linear function of FPG and GLUEX, combined with an elimination term. Inter-individual variability was identified on the FPG and GLUEX contributions coefficients, with baseline HbA1c, baseline FPG, and Asian race as covariates. GLUEX reduction (-2.8 mmol/L on average) accounted for about 3 times more than FPG reduction (-1.1 mmol/L on average) in the decrease of the HbA1c synthesis rate induced by lixisenatide.
Standard diagnostic tools, including external qualification procedure, showed an acceptable quality of fit of the data.
Conclusions: For up to 24 weeks, the model properly describes the therapeutic reduction of HbA1c levels and discriminates the relative contribution of lixisenatide effects on FPG and on PPG. This extended model of disease progression in type II diabetes adequately predicts the long term HbA1c therapeutic changes of a prandial GLP-1 receptor agonist, lixisenatide.
References:
[1] Wilkins J, Dubar M, Sébastien B, Laveille C. A drug and disease model for lixisenatide, a GLP-1 receptor agonist in type 2 diabetes, J Clin Pharmacol. 2014 Mar;54(3):267-78.
[2] Bolli GB, Owens DR. Lixisenatide, a novel GLP-1 receptor agonist: efficacy, safety and clinical implications for type 2 diabetes mellitus. Diabetes Obes. Metab. 2013 [Epub ahead of print]
Reference: PAGE 23 () Abstr 3153 [www.page-meeting.org/?abstract=3153]
Poster: Drug/Disease modeling - Endocrine