- 期刊
- OpenAccess
A Box-Cox Approach to QT Interval Correction for Heart Rate
並列摘要
QT interval on an electrocardiogram (ECG) trace is a widely used surrogate parameter in drug development to detect the potential of drugs to cause life threatening cardiac arrhythmias. It changes inversely with the magnitude of heart rate (Heart rate=60/RR, RR stands for RR interval on an ECG reading) and is often corrected to a heart rate independent value known as the corrected QT interval (QTc). Various parametric QT correction formulae have been proposed by specifying the QT and heart rate relationships explicitly by a fixed number of parameters. This paper proposes to use the Box-Cox transformation to fit the QTRR relationship, and develops a new method for correcting the QT interval for heart rate. The six commonly used parametric models are the special cases of Box-Cox transformed model. We discuss the model parameter estimation and assess the performance of QT correction formulae derived from four types of Box-Cox transformations using four off-drug ECG datasets and one on-drug ECG dataset, as well as a simulated dataset. The results show that all four derived QT correction formulae from Box-Cox transformation generate a heart rate independent QTc and that the QT correction formula derived from transformation of both QT and RR generates QTc with smaller variations. The Box-Cox transformation represents a very flexible family for modelling QT-RR relationships including the six commonly used parameter models, thus providing a potentially better QT correction method than the existing parametric models.