Joanna Parkinson (1), Ahmad Ebrahimi (2), Christer Gottfridsson (2), Jacob Leander (1), Dheeraj Rupani (3), Corina Dota (2), Dinko Rekić (4)
(1) Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Gothenburg, Sweden (2) Cardiovascular Safety Center of Excellence and SKGs, Global Patient Safety, Oncology R&D, AstraZeneca, Gothenburg, Sweden (3) Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Wilmington, US (4) BioPharmaceuticals Business Unit, AstraZeneca, Gothenburg, Sweden
Introduction
All drugs with systemic bioavailability need assessment of the QT interval which represents ventricular repolarization activity of the heart by electrocardiogram (ECG) in clinical trials [1]. This is done to rule out risk of QT prolongation and the potentially fatal but rare ventricular arrhythmia (VT/VF), Torsade de Pointes. Historically, QT assessment was done in expensive and burdensome dedicated clinical thorough QT/QTc (TQT) studies. Currently, this assessment includes concentration-response modelling of heart-rate-corrected QT (C-QTc) using digitally recorded ECG data from early phase studies, most often in single and multiple ascending dose studies. Using C-QTc analysis as primary analysis also offers the opportunity to conduct TQT studies with fewer participants and therefore more efficient and less costly trials.
Objectives
Technical requirements for the C-QTc modelling assessment are clearly defined in the Scientific White Paper [2], published by FDA and endorsed by the ICH E14 guidance [3]. Our goal was to standardize and automate this model-based analysis within AstraZeneca, to deliver C-QTc modelling reports that comply with regulatory standards with great speed and efficiency.
Methods
The Clinical Pharmacology department at AstraZeneca, together with AstraZeneca Cardiac Safety Knowledge Group, took advantage of the clear technical requirements for the assessment and created a semi-automated workflow, an efficient and fully reproducible C-QTc analysis process. This workflow involves templates for the modeling analysis plan and the modeling report, dataset standards as well as a series of analysis R scripts within a GxP compliant pharmacometrics environment (Metworx and Github Enterprise). As a result, the C-QTc assessment can be done semi-automatically using templated R scripts starting from data formatting, and including data exploration, modelling, model diagnostics, model prediction, producing tables and figures, as well as the analysis report.
Results
The semi-automated workflow was successfully developed and validated. This process allows us to reduce cost and save time of the QT assessment and also guarantees high quality output, that meets Regulatory Agencies requirements. A typical C-QTc modeling assessment using existing data can reduce cost of up to 98% compared to a dedicated TQT study (cost of 1.5-4 million USD, including up to 40-60 healthy volunteers). Recent C-QTc assessments that were conducted using the developed QT framework and were successfully approved by the FDA include AZD8233 [4], capivasertib [5], mitiperstat [6] and others. The successful submission of compliant C-QTc assessments leads to removing the need for a dedicated TQT study and more extensive ECG monitoring in phase 3.
Conclusions
An efficient semi-automated C-QTc analysis workflow was successfully developed in AstraZeneca, which is fully reproducible and entirely aligned with the requirements of Regulatory Agencies. Several C-QTc reports generated using this process were successfully approved by Regulatory Agencies and the workflow is now used routinely for all C-QTc assessments across AstraZeneca portfolio. Employing our standard templates ensures that each generated report is submission-ready. The development and implementation of this workflow have led to significant cost, time, and resource savings.
References:
[1] E14 Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs Guidance for Industry. 2005
[2] Garnett C, Bonate PL, Dang Q, et al. Scientific white paper on concentration-QTc modeling. J Pharmacokinet Pharmacodyn. 2018;45(3):383-397
[3] E14 Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs. Questions and Answers (R3) Guidance for Industry. 2017
[4] Rekić D, Azarov I, Knöchel J, Sokolov V, Nilsson C, Wernevik L, et al. AZD8233 antisense oligonucleotide targeting PCSK9 does not prolong QT interval. Br J Clin Pharmacol. 2022;88(11):4839–44
[5] Voronova V, Cullberg M, Delff P, Parkinson J, Dota C, Schiavon G, et al. Concentration-QT modelling shows no evidence of clinically significant QT interval prolongation with capivasertib at expected therapeutic concentrations. Br J Clin Pharmacol. 2022;88(2):858–64
[6] Parkinson J, Sundell J., Rekić D, Nelander K, Ericsson H, Ebrahimi A, Dota C, Sunnåker M. The myeloperoxidase inhibitor mitiperstat (AZD4831) does not prolong QT interval at expected therapeutic doses. Pharmacology Research & Perspectives 2024;12:e1184
Reference: PAGE 32 (2024) Abstr 10873 [www.page-meeting.org/?abstract=10873]
Poster: Drug/Disease Modelling - Safety