Interpreting QT in a Patient Population After Surgical Intervention
Helene Karcher (1), Ken Abrams (2), Alan Slade (2), Olivier Luttringer (1)
(1) Modeling & Simulation, Novartis Pharma AG, Basel, Switzerland; (2) Novartis Pharmaceuticals Corporation, East Hanover, NJ USA
Objectives: A drug D was tested patients after a surgical intervention in two Phase II trials. The standard of care includes drugs such as M that can have deleterious side effects. The goal of clinical program is to assess the minimization or replacement of M with the potential new agent D. To that purpose the two trials included treatment arms with M as a comedication or alone as control. A cardiac safety concern arose in one study as a higher proportion patients with at least one ECG (electrocardiogram) sample for which deltaQTcF>30ms in the D + M treatment arm in comparison to the control arm. The aim of this analysis was to assess the robustness of this finding by looking at the raw QTcF data and furthermore determine whether any apparent change in QTcF is correlated with D concentration.
Methods: Pharmacokinetic data and ECG (electrocardiogram) samples were collected in two phase II trials with drug D and comedication/standard of care M. Clinical data exploration and analysis were carried out with the software R. Assessment of cardiac safety was based on QTcF data, the time between the start of the Q wave and the end of the T wave in an ECG signal, Fridericia-corrected by the heart rate [1].
Results: (i) Neither study has a constant QTcF signal over time. Instead, large (~40ms), systematic change in QTcF over time are seen in control arms in both studies after the surgical intervention. Hence, evaluating cardiac safety with QTcF does not make sense in these studies. In all 5 arms of the 2 studies, the whole distribution of QTcF values changes systematically with time, independent of drug treatment. The magnitude of systematic changes in QTcF across trial periods is much larger than any apparent between-treatment differences. (ii) Observed maximum QTcF values in a patient depend on the number of ECG samples taken. The number of ECG samples per patient, per treatment and per treatment period is not constant in both studies. In particular, more samples were taken in the 0-12 week period in all treatment arms. This is consistent with a higher maximum QTcF in this period. In contrast, the median QTcF decreases consistently with time in all treatment arms. As a consequence, the higher incidence of elevated QTcF values in this period is most likely an artifact of sampling frequency. (iii) No apparent correlation between QTcF and drug concentration was observed for any treatment (D or M).
Conclusions: Text The analysis shows that random & systematic variability in QTcF confounds assessment of any potential drug-induced effect in both studies. A large, systematic treatment-independent change in QTcF of approximately 40 ms in the 12 weeks following the surgical intervention impedes the cardiac safety evaluation using deltaQTcF.
Reference:
[1] ICH Expert Working Group (2005) The Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs: E14, May 12, 2005.