A population-based characterisation of the relationship between anatomical and electrical axes of the heart

Abstract

Abstract Introduction Cardiac anatomy, including heart size, position and orientation within the torso, is known to impact ECG morphology. Understanding the relationship between cardiac anatomy and ECG morphology through their main axes has the potential to lead to more accurate and reliable ECG biomarkers. Nevertheless, there is no standardised method for the definition of the anatomical and electrical axes of the heart. Aims To evaluate definitions of anatomical and electrical axes of the heart based on a derived metric in an undiseased population in the UKBioBank. Methods The anatomical and electrical axes of 2189 healthy subjects were computed separately in five different ways (Figure 1). Cardiac bi-ventricular anatomy was automatically segmented from cardiac MRI. Anatomical axes were computed from different combinations of anatomical landmarks (valves and apex) and axis of inertia. ECGs were transformed into vectorcardiograms (VCGs) using the Kors transformation. A range of combinations of dipole magnitudes, weighted averages and loop shape were used to define the electrical axes of the QRS loop. The anatomical-electrical relationship was quantified with the cosine of the angle between each of the 25 pairs of anatomical and electrical axes definitions. Spatial consistency was assessed by the standard deviation of this cosine across the cohort, where a higher level of spatial consistency corresponds to a lower standard deviation. Results Spatial consistency between anatomical and electrical axes exhibited a large range of standard deviation values from 0.61 for the worst pair to 0.14 for the best pair of definitions. The pair with the highest spatial consistency was defined by the valvular plane centre and the apex (VPA) for the anatomical axis, and the direction of maximum QRS dipole magnitude (maxQRS) in the VCG for the electrical axis (Panel A and B in Figure 2). The anatomical-electrical relationship was not significantly correlated with BMI, age or gender despite the individual anatomical and electrical axes showing significant correlation (Frontal plane: anatomical axis was most impacted by BMI, R=0.53 & p<0.0001, and electrical axis by gender R=0.37 & p<0.0001). Conclusion In this healthy population, the most spatially consistent definition of anatomical and electrical axes is defined by VPA and maxQRS. This spatial consistency is not affected by BMI, age or gender, suggesting an intrinsic relationship between the pair of measures, and represents a potential strategy to permit correction for patient anatomic factors when interpreting ECGs.Figure 1.Axes Definition MethodsFigure 2.Results