Hanna E. Silber Baumann, Christophe Schmitt and Valerie Cosson
Clinical Pharmacology, Roche Pharma Research and Early Development, Roche Innovation Center Basel, F.Hoffmann-La Roche Ltd
Objectives: The objective of the analyses was to support the Phase II dose selection through the characterization of the non-linear properties of RG7696 PK, the identification of the model describing the effects on mean plasma glucose (MPG) as a function of exposure, and the extrapolation of MPG decrease to 12-week HbA1c reduction.
Methods: Data from 92 individuals which were treated with at least one dose of RG7697 were included in the PK analysis. Doses ranged between 0.03-5 mg (s.c. injection) once daily. Rich sampling was performed on day 1 and 14 with additional sparse sampling. In 54 type 2 diabetics patients receiving 2-week treatment, the pharmacological effect of RG7697 was assessed with repeated meal tolerance tests (MTT). 24-hour glucose profiles were collected on days -1, 1, 7 and 14. MPG was calculated as the average of the collected samples over 24 hours. PK and PKPD models were developed based on 2-week data. Finally, 12-week HbA1c extrapolations were generated based on the 2-week PKPD model, coupled with a life-span model of HbA1c [1]. The results were compared to literature data of GLP-1 compounds through a model-based meta analysis. Analysis and simulations were performed using NONMEM 7.2.
Results: The PK of RG7697 was described using a 2-compartment model. Non-linear functions were incorporated to account for dose and time-dependent kinetics with respect to bioavailability and duration of absorption. Steady-state was reached after approximately 1 week of treatment. The time course of MPG over 2 weeks was captured using an indirect response model (IDR). Individual predicted RG7697 steady state exposure was included as an Emax function on Kout. Fasting plasma glucose was included as a covariate on the baseline MPG. Predicted 12-week HbA1c reduction across a range of RG7697 doses was compared to prediction from a meta-analysis model of GLP-1 compounds. The results predict that the reduction following RG7697, compared to Liraglutide at the recommended dose, would follow a faster onset and that the HbA1c reduction would be larger with RG7697.
Conclusions: RG7697 non-linear PK could be adequately accounted for by including functions of dose and time on the absorption phase. The reduction in MPG over 2 weeks was adequately described through an IDR model with the drug effect on Kout. The predicted 12-week reduction in HbA1c indicated that well tolerated doses of RG7697 should compare favorably to marketed GLP-1 compounds in Phase II.
References:
[1] Lledo-Garcia R et al. A semi-mechanistic model of the relationship between average glucose and HbA1c in healthy and diabetic subjects. J Pharmacokin Pharmacodyn, 2013 Apr;40(2) :129-42
Reference: PAGE 24 () Abstr 3369 [www.page-meeting.org/?abstract=3369]
Poster: Drug/Disease modeling - Endocrine