Torsade de pointes is a polymorphic ventricular tachycardia that has been linked to sudden cardiac death. While typically rare and fleeting, increased risk for occurrence of and problematic outcomes from this arrhythmia has been attributed to prolonged QT intervals. These interval measurements of electrocardiogram waveforms represent the time between depolarization and repolarization of the ventricles. Prolongation of this period increases the likelihood of disruptions in ventricular cardiomyocyte conduction, which can lead to torsade de pointes events. Due to the potentially fatal outcomes associated with QT prolongation, it was not long until drugs found to induce this prolongation began to be removed from the market. These concerns led to the creation of safety pharmacology guidelines S7B and E14 that outlined suggested QT prolongation risk assessments to be performed during the preclinical and clinical stages of drug development, respectively. Studies based on these guidelines are intended to identify drug-induced changes to the QT interval, but must first control for the effect heart rate has on QT. To isolate drug effect QT correction methods are used to estimate corrected QT values as if the data was collected at a fixed heart rate. Using these values as a biomarker has led to highly sensitive assessments that have prevented any new drugs from reaching the market with unacceptable QT prolongation risk. However, these assessments are still in need of improvement as the pharmaceutical industry must deal with costly and time-consuming clinical safety studies despite the high sensitivity of preclinical assessments, leading to calls for the integration of preclinical and clinical guidelines. To achieve this, the translatability of preclinical results must be improved. This dissertation aims to increase the reliability of preclinical results and improve their translatability by optimizing the QT correction methods they rely on. To do this, I analyzed large ECG datasets from preclinical safety pharmacology studies obtained through Eli Lilly and Company (Indianapolis, IN). The central hypothesis of this dissertation is that through statistical analysis of these data, the collective understanding of QT correction methods will be expanded, and an improved method can be developed. In pursuit of this goal, the effectiveness of various preclinical QT correction methods was evaluated in simulated drug treatment scenarios, against 130,000+ bootstrap resampled ECG recordings of vehicle control treatments, and with study data from non-human primates treated with known QT prolonging drugs. The results of these evaluations provided evidence of how assumptions inherent to these methods affect the result of correction. Examples of such assumptions include what heart rate is relevant for the species, that a predetermined population-based estimate of the QT-rate relationship is appropriate for individuals, that this relationship will not change over time or between treatments, and that assuming this relationship before correction is appropriate in the first place. All of this led to the development of the novel Ratio QT correction method designed to be applicable to any scenario by dynamically responding to moment-to-moment changes in the relationship between timepoints. This novel method combines the simplicity and ease-of-use of population-based methods with the effectiveness of individual methods. Taken together, this research has increased the collective understanding of QT correction methods and resulted in a novel method that is as effective as it is simple to use. The investigations presented in this dissertation have explored aspects of QT correction methods that have been in question for years. Optimizing these methods is now easier thanks to the information gained through these analyses of large preclinical ECG datasets. An integral step has been made towards the creation of a new industry standard of QT correction capable of bridging the gap between preclinical and clinical safety pharmacology studies. Such improvements can be used to help reduce the number of research subjects necessary for preclinical and clinical QT prolongation assessments.
Read
