Klaus Lindauer 1, Anja Tündermann 1, Birger Wenge 2, Feras Khalil 1, Heiko Hagedorn 1
1 Grünenthal GmbH (Aachen, Germany), 2 at the time of contribution Grünenthal GmbH (Aachen, Germany)
Introduction
Drug-induced QT prolongation is a serious side effect where certain medications delay heart muscle recharging, risking a fatal, rapid, irregular heart rhythm called torsades de pointes. The assessment of drug-induced prolongation of the QT interval has increasingly shifted from classical by time point evaluations to C–QTc modeling, following updates to the ICH E14 Q&A[1] and broader regulatory acceptance of model based methodologies. These regulatory advancements have led to reduced reliance on Thorough QT studies and widespread adoption of exposure–response approaches as the primary framework for quantifying QT liability. However, QTcF, the standard heartrate correction method, may inadequately separate drug and time related physiological effects, particularly in disease settings characterized by dynamic heart rate changes. As demonstrated in recent analyses of antituberculosis therapy, exposure–QTcF models required incorporation of secular (time related) trends to disentangle drug effects[2,3]. Together, these developments highlight the need for modeling frameworks that simultaneously address exposure and time dependent components of QT dynamics to ensure accurate characterization of the proarrhythmic risk of drugs in clinical development.
Data obtained from a randomized, single-center, double-blind, placebo-controlled first-in-human (FIH) trial with single (SAD) and multiple ascending (MAD) doses of an undisclosable drug was used for this analysis.
Objective
C-QTc analysis using pharmacokinetic/pharmacodynamic (PK/PD) modeling to evaluate time- and exposure-dependent drug-induced QT interval prolongation
Method
The SAD dataset of the FIH trial consists of 8 cohorts with different doses, each including 6 subjects on active test drug and 2 on placebo. Pharmacokinetic (PK) samples were taken at 0,0.5,1,2,3,4,6,8,10,12,15,24h post-dose. Continuous 12-lead-Holter ECGs and triplicate-ECG extractions at Day -1 and 1 were performed at matching time points of the PK samples taken.
The MAD dataset consists of 3 cohorts with ascending doses, each including 8 subjects on IMP and 2 on placebo (14 days bid treatment). Continuous 12-lead-Holter ECGs and triplicate-ECG extractions at day –1 (baseline), day1 and day14 (0,0.5,1,2,3,4,6,8,10,12,15,24h post-dose) were performed and PK samples were taken at matching time points. 12-lead-safety ECGs were extracted at Day -1 to 14 and 16 at pre-dose and 1h post-dose.
QTcF was the primary variable and time matched QTcF measurements were compared to baseline (ΔQTcF). First, the individual ΔQTcF values were assessed graphically and corrected to placebo (ΔΔQTcF) The following functions of the exposure-response relationship were tested and evaluated:
ΔΔQTcF=ΔΔQTcFt=0 +TE +CIRC +DE
• ΔΔQTcFt=0 is baseline
• TE – indirect drug induced time dependent effect
o ΔΔQTcFmax×(1.0 − e−r×time) with ΔΔQTcFmax maximal effect, r – scaling factor
• CIRC – circadian rhythm like term
o A0×cos(2π×(time−ϕ)/(period)) with A0 amplitude, ϕ phase shift
• DE – drug effect
o Emax×c/(EC50+c) with Emax maximal effect, c concentration
o Escale×cfactor with Escale scaling factor
o Emax/(1.0+e−SLOP×(c−c05)) with Emax maximal effect
The results were reported as described in[4,5].
Results
The analysis of the SAD data alone did not reveal an statistically significant ΔΔQTcF increase (p-value=0.46).
The ΔQTcF data obtained from safety and 12-lead Holter ECGs from the MAD part were evaluated graphically. The two different data types were comparable and were merged into a dataset, which represents the ΔQTcF dynamics of the study duration. The graphical evaluation of the ΔQTcF data demonstrated a drug induced effect, which increased over time. The lowest dose groups (cohorts 1 and 2) did not clearly differentiate from placebo, whereas the highest dose group showed an increase in ΔQTcF after multiple dosing within the first 10 treatment days, reaching a plateau thereafter.
Therefore, time (TE) and drug (DE) effect terms were considered for the placebo corrected ΔΔQTcF analysis. Estimates of ΔΔQTcF exhibited considerable variability, which could be partly explained by introducing the CIRC term. The period was estimated to be approximately 2.7 days. The upper bound of the 90% confidence interval for ΔΔQTcF approached the 10ms threshold[1] at exposure levels of ~300ng/mL.
Conclusions
Analysis of ECG data indicated that the two data sources can be merged. Although, the SAD data alone did not indicate any significant ΔΔQTcF increase, the ΔΔQTcF analysis of the MAD dataset was estimated to be both time- and exposure-dependent, with a plateau reached between 10 to 14days . The underlying mechanism for this time-dependency of the observed QTc prolongation remains unclear and cannot be explained by the block of the hERG channel as a drug-mediated off-target interaction.
References:
[1] ICH guideline E14/S7B: clinical and Nonclinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential – questions and answers. https://www.ema.europa.eu/en/documents/scientific-guideline/ich-guideline-e14s7b-clinical-and-nonclinical-evaluation-qtqtc-interval-prolongation-and-proarrhythmic-potential-questions-and-answers-step-5_en.pdf last accessed 17. Feb. 2026
[2] Tanneau et. al. Exposure–¬safety analysis of QTc interval and transaminase levels following bedaquiline administration in patients with drug-¬resistant tuberculosis CPT Pharmacometrics Syst Pharmacol. 2021
[3] Vongjarudech et. al. Establishing the Exposure-QT Relationship During Bedaquiline Treatment Using a Time-Varying Tuberculosis-Specific Correction Factor (QTcTBT) CPT: Pharmacometrics & Systems Pharmacology, 2025
[4] Garnett et al. Scientific white paper on concentration-QTc modeling Journal of Pharmacokinetics and Pharmacodynamics (2018) 45:383–397
[5] Parkinson et al Practical guide to concentration QTc modeling: a hands on tutorial Journal of Pharmacokinetics and Pharmacodynamics (2025) 52:43
Reference: PAGE 34 (2026) Abstr 11894 [www.page-meeting.org/?abstract=11894]
Poster: Drug/Disease Modelling - Safety