Model‐Based Evaluation of QTc Interval Risk: An Increasing Emphasis on Early Decision Making

In this issue of the Journal of Clinical Pharmacology, Shashank Rohatagi and colleagues1 report on the use of pharmacokinetic/QTc relationships from rising-dose studies to better understand the risk of QTc interval prolongation. Although the methods proposed are consistent with current standards on the development and application of pharmacokinetic/pharmacodynamic (PK/PD) modeling and simulation, what is unique about this work is that it questions the need to have a separate thorough QT (TQT) study. Such an approach is not inconsistent with the presumed intent of the ICH E14 guidance,2 where there is a provision on Alternative Strategies to Assess QT/QTc Interval Effects (2.2.5), specifically stating that “alternatives to the use of the ‘thorough QT/QTc study’ are under active investigation. Examples include evaluating the relationship between concentration and QT/QTc effects or more intensively evaluating ECGs, based on data collected during early phase clinical studies.” A reduced sample size approach to study the effect of drug development candidates on QTc interval was proposed earlier by Malik and coworkers3 in the same journal. Malik et al3 argued that careful electrocardiogram (ECG) acquisition was a central tenet to the early assessment of risk for QTc interval prolongation, and this aspect may be singularly rate limiting given that the clinical trial sites may need to have an additional level of expertise related to the ECG data capture issues. Often times, there have been limitations at the clinical site on the extent of continuous ECG capture. Whereas enhanced ECG monitoring is clearly possible with 12-lead ECG, the diagnostic reference standard, at baseline (predose) and at several time points postdose in triplicates, most first-in-man (FIM) study sites are limited to a 3- or 5-lead cardiac telemetry monitoring. Furthermore, aspects related to systematic checks of heart rate stability before QT-interval measurement, ECG morphological assessments, and prespecification of heart rate correction are additional areas of improvement within initial rising-dose studies.3 With the improvements in continuous ECG monitoring, access to a cardiologist at the site, and building sufficient infrastructure capabilities ensuring data capture, integrity, and quality, leveraging PK/QTc analysis may be a quick and nimble approach to better understand the probability of observing a QTc signal and potentially aiding early decision making on compound candidates.1,3,4 That said, whether TQT studies can be completely replaced by modeling and simulation (M&S) based on early clinical experiments still remains at large. These M&S approaches remain hypothesis-generating experiments but contribute to an early assessment of model-based QTc prolongation risk. Their use in hypothesis testing still remains an ambitious opportunity given limitations inherent in small sample size in phase I studies. More specifically, the predictive value of M&S to delineate risk when the effect size is small needs to be more carefully investigated. It is becoming increasingly clear, however, that M&S can inform the probability of risk based on a preponderance of cumulative information from earlier stages of drug discovery and preclinical development, such as hERG binding experiments and telemetry studies in conscious dogs, before formally assessing clinical risk.5 It is certainly possible that model-based risk assessments can aid in early decision making, such as compound prioritization, and potentially in decisions around continuance on the development path. Financial disclosure: The author is an employee of Merck & Co, Inc, and potentially owns stock and/or holds stock options in the company.