Comparison of single-lead and multi-lead ECG for QT variability assessment using autoregressive modelling

Abstract

Abstract Objective. Beat-to-beat fluctuations in the QT interval—QT variability (QTV)—have been shown to vary amongst the different ECG leads. This study aims to compare the utility of single and multi-lead ECG to disentangle the mechanisms contributing to QTV. Approach. Twelve-lead ECG was analysed in 57 coronary artery disease patients before and after an elective percutaneous transluminal coronary angiography (PTCA) procedure. QT, RR and respiration time series were extracted. QTV was decomposed into contributions by heart rate, respiration and QTV independent of heart rate and respiration using parametric autoregressive modelling. Signal-to-noise ratio, model goodness-of-fit, mean QT, corrected QT, QT variability and RR variability were also computed. Results from two single leads (Lead II and V5) and three one-dimensional representations of 12-lead ECG (principal component analysis (PCA), vector magnitude (VM), and root mean square of the 8 independent leads of the standard 12 leads (RMS8)) were compared during resting conditions, before and after PTCA, and between patients with myocardial infarction and those without. Main results. At baseline, mean QT and corrected QT were significantly lower in VM and RMS8 compared to single leads. While overall QT variability was not different between the leads, QT independent of heart rate and respiration was significantly lower in VM and RMS8. Following PTCA, changes in these variables were similar in all leads. Differences between patients with MI and those without MI were consistent in all leads. Significance. Despite the differences in some QTV components amongst various leads, single-lead ECG could be sufficient for analyzing QTV in populations with pathological cardiovascular conditions compared to those without, or for quantification of intervention effects.