N. Frances (1), C. Meille (1), A. Caruso (1), J. Fretland (2), A. C. Harrison (2), A. Greiter-Wilke (3), M. Sanders (3), T. Lavé (1)
F. Hoffmann-La Roche Ltd., Pharma Research and Early Development, Non Clinical Safety, (1) Drug metabolism and Pharmacokinetics (DMPK), Modeling and Simulation; (2) DMPK (3) Safety Pharmacology
Objectives: QT prolongation in the electrocardiogram is a surrogate marker of ‘torsade de pointes' (fatal arrhythmias). This QT parameter is investigated during pre-clinical cardiovascular telemetry studies and it is critical, in case of QT prolongation, to define the safety margin for initiation of human trials. The objective of this study was to evaluate a threshold model to link QT with exposure and explore its translational capability.
Methods: The threshold model described below was applied to 3 compounds in pre-clinical development and for one compound for which clinical data was available. QTc (heart rate corrected QT) [1] change was linked to exposure by a threshold model using a population approach (Monolix® 3.2 [2]). This model combines a baseline parameter (QTc baseline value), a threshold parameter which corresponds to the concentration below which there is no drug induced changes in QTc expected, and a slope parameter which corresponds to the QTc increase rate according to exposure above the threshold parameter value. Inter-individual variability was estimated on baseline and slope parameter. The pre-clinically assessed threshold parameter is then used for human prediction in combination with the predicted PK profile in man by PBPK modeling (GastroPlus® [3]).
Results: The threshold model effectively described the relationship between exposure and heart rate corrected QT when data show a range of exposure with no effect on QTc and no saturation in effect. Under these circumstances, the threshold model can be used as an alternative to other existing models [4, 5]. The approach successfully described QTc change in human and showed translational properties of model parameters.
Conclusions: QTc and exposure can be linked by a threshold model and demonstrate translational properties from pre-clinical to clinical studies. In case of a QTc prolongation, this model predicts the exposure where effects are expected to occur (threshold parameter, which will likely be higher than the otherwise used No Effect Level). By using the threshold model to describe the exposure response relationship and translate the model to humans, a better prediction of where QTc prolongation is likely to occur can be estimated which provides more confidence in the dose escalation strategy.
References:
[1] Henry H. Holzgrefe, Icilio Cavero, Carol R. Gleason, William A. Warner, Lewis V. Buchanan, Michael W. Gill, Dennis E. Burkett, Stephen K. Durham; Novel probabilistic method for precisely correcting the QT interval for heart rate in telemetered dogs and cynomolgus monkeys, Journal of Pharmacological and Toxiological Methods
[2] Kuhn E., Lavielle M. "Maximum likelihood estimation in nonlinear mixed effects models" Computational Statistics and Data Analysis, vol. 49, No. 4, pp 1020-1038, 2005
[3] GastroPlus v6.1, http://www.simulations-plus.com/
[4] Kenny J. Watson, William P. Gorczyca, John Umland, Ying Zhang, Xian Chen, Sunny Z. Sun, Bernard Fermini, Mark Holbrook, Piet H. Van Der Graaf; Pharmacokinetic-pharmacodynamic modelling of the effect of Moxifloxacin on QTc prolongation in telemetered cynomolgus monkeys, Journal of Pharmacological and Toxicological methods 63 (2011) 304-313 and Poster presentation PAGE 2011, Athens.
[5] Brandquist C., Stypinski D., Teuscher N., Girard F.; PK/PD Modeling of the Effect of Intravenous Doses of Anzemet® (dolasetron mesylate) and Its Metabolite (hydrodolasetron) on the QT Interval in Healthy Subjects; Poster presentation PAGE 2011, Athens.
Reference: PAGE 21 (2012) Abstr 2577 [www.page-meeting.org/?abstract=2577]
Poster: Safety (e.g. QT prolongation)