Phase I trials: Model-based assessment to identify a clinical relevant change in heart rate
Martin Fink (1), Olivier Luttringer (1), Peter Gergely (2), Erik Wallstroem (2), Michael Looby (1), Goonaseelan Pillai (1)
(1) Modeling & Simulation, Novartis Pharma AG, Basel, Switzerland; (2) Translational Medicine, Novartis Pharma AG, Basel, Switzerland
Objectives: The primary objective of this work was to ascertain, through an integrated PK/PD model-based approach, what measure of change in heart-rate (HR) should be considered as clinically relevant in phase I trials. Interindividual and between-study variability of circadian variations of HR demand a more sophisticated approach than simple baseline-correction, when attempting to distinguish drug effects from usual changes in HR.
Methods: Placebo and pre-dose heart rate data from 24-h holter monitoring from 7 phase I clinical studies were pooled (n=405, >700 full days of recordings). The basic mathematical model consisted of a sum of five cosine functions to replicate the circadian variations (with periods of 24, 12, 8, 6, and 4.8 hours, respectively). Study, sex, and weight were tested during the covariate building of a non-linear mixed effects model as well as the placebo effect compared to pre-dose.
Results: HR over the day ranged from 60 to 80 bpm (in a typical male). Gender differences could be found for the mesor (>5 bpm), but no study dependence was noted. The placebo effect (observed mainly) on the mesor was smaller than the gender difference. Study dependence was much more often found on the phase shifts than on the amplitude of the cosine functions - and the latter in general denoted for changes of less than 2 bpm for the typical subjects. Nonetheless, due to the phase shifts the maximum time-wise differences between studies were larger than 10 bpm.
Discussion: Not all the effects found to be statistically significantly different during model building could also be considered clinically relevant, especially the study dependent effects on the amplitude parameters. However, overall gender and study dependent effects influenced the circadian changes of HR of the typical subject more than what is considered to be a clinically relevant drug effect.
Conclusions: Due to the high variability in HR over the day and the large study and gender dependencies, it is recommended to consider a model based approach when estimating any potential drug effect compared to baseline and placebo during clinical trials.
