Linnéa Bergenholm (1,2), Neil Evans (2), Michael Chappell (2), Teresa Collins (3), Joanna Parkinson (4)
(1) Discovery Safety, Drug Safety and Metabolism, iMED, AstraZeneca, Mölndal, Sweden, (2) Systems Modelling and Simulation, School of Engineering, University of Warwick, Coventry, UK, (3) Translational Safety, Drug Safety and Metabolism, iMED, AstraZeneca, Alderley Park, UK, (4) Early Clinical Development, Quantitative Clinical Pharmacology, iMED, AstraZeneca, Mölndal, Sweden.
Objectives: To develop population PKPD models for describing baseline and drug-induced changes in QRS and PR intervals in dogs using standard telemetry data accessible in drug development. Compare modelling results for a selection of compounds with different mechanism of action.
Methods: Cardiac effects of three antiarrhythmics (AZD1305, flecainide and quinidine) and two anticholinergics (AZD8683 and AZD9164) were investigated in conscious dogs. Plasma concentrations were quantified. Cardiac effects were monitored using surgically implanted telemetry devices. Mixed effects PKPD models were developed sequentially. Individual PK models were used to drive PD modelling of vehicle and treatment QRS and PR interval data. RR interval correction and circadian rhythm models were investigated for identification of appropriate baseline models. Linear and non-linear direct and link models were tested to describe the drug-induced effects on QRS and PR intervals.
Results: (preliminary) Prolongation of QRS ranged from 9-16% in AZD1305, flecainide and quinidine and prolongation of PR from 21-34% in for AZD1305, AZD8683, AZD9164 and flecainide. Population PK models were able to characterise the drug exposures well. Preliminary results suggest that RR correction and circadian rhythm improve fit to PR but not QRS interval data. Also, drug effects on PR and QRS are best explained by linear models rather than power and Emax models. The slopes of the PK-QRS relationships varied from 2-7 ms/uM unbound drug, while the slopes of the PK-PR relationships were larger.
Conclusions: Baseline and drug-induced effects on QRS and PR intervals were successfully described for five compounds. These models improve our understanding of the concentration-effect relationships and as such add value to the project teams and decision-makers. Also, these analyses could be used to investigate the translational relationship between dogs and humans, potentially improving the prediction of QRS and PR effects in humans before clinical trials.
Reference: PAGE 24 (2015) Abstr 3650 [www.page-meeting.org/?abstract=3650]
Poster: Drug/Disease modeling - Safety