Probabilistic risk assessment for QT prolongation and heart rate increase
Helene Karcher, Tobias Sing, Mick Looby, Olivier Luttringer
Novartis Pharmaceuticals AG
Objectives: Oral drug A is in clinical development for several indications between Proof-of-Concept and Phase II. The objective of the modeling work was to quantify the risk of QT prolongation or heart rate increase in a population of patients taking drug A, and thereby characterizing the safety profile of drug A at an early stage.
Methods: Assessment of the probability of a patient experiencing QT prolongation or heart rate increase at a certain dose and regimen was performed in including inter-individual variability in pharmacokinetics and inter-individual variability in heart rate or QT prolongation at given plasma concentrations. Namely, the following steps were taken:
- 1. A mixed-effects PK analysis was performed, and the model used to simulate a population of patients treated at different doses and regimens. To obtain the best possible prediction on PK profiles for patients, all available data from disease and healthy volunteer Phase I trials that did not study drug-drug interaction were included to build the mixed-effects PK model.
- A mixed-effects analysis of the QT (or heart rate) – concentration relationship was conducted on controlled data from a thorough QT study with time-matched electrocardiograms and PK samples.
- At each considered dose and regimen, the two mixed-effects models above (pharmacokinetic and concentration-QT (or heart rate)) were used sequentially to simulate 1000 patients’ steady-state Cmax and corresponding QT prolongation or heart rate increase. The distribution of QT prolongation or heart rate increase in a 1000-patient population was obtained, and used to define the percentage of patients likely to experience a QT prolongation or heart rate increase at a range of threshold values (defined by the clinicians, e.g., 5 ms, 10ms, 20 ms, 30 ms for QT, 2 bpm, 5bpm, 10bpm for heart rate).
Results: Distribution of predicted Cmax at steady-state for several doses and regimens highlighted the large variability in patient’s pharmacokinetics for drug A. Prediction of the corresponding percentage of patients likely to experience a QT prolongation or heart rate increase at a range of threshold values could be interpreted as either quantification of the risk associated with each dose in a large disease population, or the individual risk for a given patient to experience a QT prolongation or a heart rate increase at that dose. The way the safety profile changed when using once-daily regimen vs. twice-daily regimen was reconciled with other factors (efficacy, market access) to inform the decision on a final regimen.
Conclusions: The probability of a patient experiencing QT prolongation or heart rate increase over certain defined thresholds enabled the clinical team to assess and communicate the risk of administering compound A at several doses and regimens to a larger patient population.