Willem de Winter, Adrian Dunne, Chyi-Hung Hsu, Xavier Woot de Trixhe, Damayanthi Devineni, David Polidori, Jose Pinheiro
Janssen Research & Development LLC
Objectives: Canagliflozin is an orally active inhibitor of SGLT2 for the treatment of patients with T2DM. A dynamic pop PK/PD model was developed to assess the effect of once daily (QD) versus twice daily (BID) dosing regimens of canagliflozin as add-on to metformin on the relationship between canagliflozin plasma exposure and HbA1c response in diabetic patients on metformin background medication.
Methods: A dynamic pop PK/PD model based on a turnover model for HbA1c was implemented in NONMEM 7.2 using observed HbA1c responses and predicted 24 hour individual canagliflozin plasma exposures from 1,347 type 2 diabetic patients on metformin background medication at screening from two placebo-controlled canagliflozin studies: a 12 week Ph 2b study with 50, 100, 200 and 300 mg QD and 300 mg BID dosing arms, a 26 week Ph 3 study with 100 and 300 mg QD arms, and the baseline and placebo-arm data from an 18 week Ph 2 study in patients receiving placebo, 50 and 150 mg BID. 24 hr exposure profiles were predicted using a population PK model for canagliflozin developed on 9,061 PK samples from 1,616 subjects [1]. An efficient method of averaging was developed for numerically solving the ordinary differential equations of the dynamic model [2]. The final model was validated internally as well as externally on its ability to predict the post-baseline BID dosing observations from the 18 week Ph 2 study.
Results: The pop PK/PD model provided a satisfactory fit to the observed HbA1c data. The combined HbA1c-lowering effects of placebo (study specific) and canagliflozin were found to be baseline dependent. No other significant covariate effects could be identified using this patient population which consisted mostly of subjects with normal to mildly impaired renal function. The model could adequately predict the external data from the 50 and 150 mg BID dosing arms from the 18 week phase 2 study, and was used to demonstrate that the differences in HbA1c reduction between the BID and QD regimens in patients with similar baseline conditions and total daily doses were small and not clinically meaningful [3].
Conclusions: A dynamic pop PK/PD model for the relationship between canagliflozin exposure and HbA1c response was developed and used to support the regulatory approval of the canagliflozin/metformin IR fixed-dose combination for BID administration by bridging efficacy between the QD and BID regimens, thus avoiding the need for an additional clinical trial.
References:
[1] Hoeben E, Vermeulen A, de Winter W, Neyens M, Dunne A. (under review) Population Pharmacokinetic Modeling of Canagliflozin in Healthy Participants and Patients with Type 2 Diabetes Mellitus. Clin. Pharmacokinet.
[2] Dunne A, de Winter W, Hsu C-H, Mariam S, Neyens M, Pinheiro J and Woot de Trixhe X (under review) The Method of Averaging applied to Pharmacokinetic/Pharmacodynamic Modeling. J. Pharmacokinet. Pharmacodyn.
[3] Marathe A, Jain R, Jain L, Hsu C, Pinheiro J, de Winter W, Sahajwalla CG, Sinha V, Mehrotra N (2015). Exposure response analysis as evidence for approval of canagliflozin-metformin immediate release fixed-dose combination product: A regulatory perspective. Oral Presentation at the 2015 ASCPT Annual Meeting, New Orleans, LA, USA.
Reference: PAGE 24 (2015) Abstr 3626 [www.page-meeting.org/?abstract=3626]
Poster: Drug/Disease modeling - Other topics