Eleonora Marostica (1), Karel Van Ammel (2), Ard Teisman (2), Jan Van Bocxlaer (1), Koen Boussery (1), Filip De Ridder (3), An Vermeulen (1,3) and David J. Gallacher (2)
(1) Laboratory of Medical Biochemistry and Clinical Analysis, Ghent University, Belgium; (2) Global Safety Pharmacology and (3) Model Based Drug Development, Janssen R&D, Beerse, Belgium
Objectives: Drug-induced Torsades de Pointes (TdP) and consequent sudden cardiac death are positively correlated with QT-interval prolongation. Conversely, not all “QT-prolonging drugs” induce TdP. Irrespective of this inconsistency, the early risk biomarker QT-interval prolongation is frequently used to deselect drugs. Hence, the need for an improved understanding of it is obvious. The work’s aims were: to develop a pharmacokinetic/pharmacodynamic (PK/PD) model to describe “QT data” in awake dogs and humans receiving drugs that are known to affect the QT interval (i.e. moxifloxacin, C1, C2, and C3); to determine the unbound concentration needed to reach 50% probability (CP50) of QT-interval prolongation/shortening; to assess the approach’s translational opportunities between species.
Methods: A population PK model was developed for each compound and species using NONMEM 7.1. The QT interval was then modelled as a sum of individual heart-rate correction, circadian rhythm, and drug effect [1]. Different model structures were assessed. The final PK/PD models were then implemented in WinBUGS 1.4.3 according to a fully Bayesian approach. The probability of QT-interval prolongation/shortening greater than 10 ms was assessed for all the compounds. Based on the posterior distributions of the parameter estimates, the probability curves of the typical and the new subject/dog were calculated.
Results: For each compound and species, the QT profiles were well described by the PK/PD models. Similar CP50 of moxifloxacin, C1, and C2 are needed in dogs and humans to reach the same effect. Dogs were less sensitive than humans to QT-interval prolongation when receiving C3 (CP50 dogs > CP50 humans). However, in all cases, when looking at the typical subject/dog, the probability curve for humans was steeper than the one for dogs. This finding is in agreement with [1], where moxifloxacin and two other drugs were analyzed. The drug-effect parameter estimates provided a translational scaling factor of 2.3.
Conclusions: The proposed model was able to describe both QT-interval prolongation and shortening. The knowledge of the translational scaling factor can provide an insight into the possible effect that a new compound may have on the QT interval in humans, based on the drug effect observed in dogs. The inclusion of more compounds could improve the estimate of the translational scaling factor as well as our understanding of the relationship between dogs and humans.
References:
[1] Anne S.Y. Chain, Vincent F.S. Dubois et al, Identifying the translational gap in the evaluation of drug-induced QTc interval prolongation, Br J of Clin Pharmacol 76, pp. 708-724, 2013.
Reference: PAGE 24 () Abstr 3465 [www.page-meeting.org/?abstract=3465]
Poster: Drug/Disease modeling - Safety