Model-Based Evaluations to Select and Confirm Doses in the Clinical Development of Exenatide

Conference: ASCPT
Division: Cognigen


Background: Population pharmacokinetic (PK) and pharmacodynamic (PK/PD) models based on early Phases 1 and 2 data for the new antidiabetic medication exenatide (ex) supported transition from weight-based to fixed 5 and 10 μg subcutaneous BID dosing for Phase 3 trials (CPT 2002;71:P29). The model-based dose selection process was designed to target exenatide exposure that would maximize glucose reduction and mitigate gastric AEs. The models were iteratively refined with the inclusion of additional Phases 2 and 3 data to further describe the PK and exposure-response (ER) relationships. These models confirmed the appropriateness of the selected doses and allowed for the quantitative assessment of the influence of intrinsic variables such as age, weight, and gender.

Methods: 4870 ex concentrations from 242 patients with type 2 diabetes (subj) across 10 clinical trials were applied to the existing PK model (1-CMT with nonlinear plus linear absorption and linear elimination). The relationship between ex AUC0-3h and glucose (gluc) AUC0-3h (540 pairs from 183 subj) was assessed in an inhibitory sigmoid-Emax model. Covariate evaluation included the influence of selected subject descriptors, clinical chemistries, concomitant oral antidiabetic agent use, and anti-ex antibody (Ab) status on PK and PK/PD. Model validation was performed by bootstrapping.

Results: For the PK model, weight and anti-ex Ab status were significant covariates for CL/F and V/F, and gender for absorption rate. In the PK/PD model, significant factors for Emax were baseline gluc AUC and anti-ex Ab status. The magnitude of these effects did not support dose adjustment. The population mean estimated decrease in gluc was ~27% and 33% for the 5 and 10 μg BID doses, respectively. Both doses approach the asymptotic portion of the exposure-response curve. With ex doses of 5 μg, 55% of subj achieved 75% of their predicted maximum glucose reduction. With 10 μg doses, 90% achieved that same target.

Conclusion: This model-based drug development approach demonstrated the value of the learn-and-confirm paradigm. Dose selection was supported by prelim PK and PK/PD models in a small number of subj, and confirmed with larger numbers of subj from Phases 2 and 3 trials. These results verified that ex doses of 10 μg BID optimize clinical benefit in terms of glycemic control.

American Society for Clinical Pharmacology and Therapeutics (ASCPT), Anaheim, California, March 2007

By Mark Fineman, Luann Phillips, David J Jaworowicz, Brenda Cirincione, Elizabeth Ludwig, Kristin Taylor, Prajakti A Kothare, Alain D Baron, Thaddeus Grasela