Spatial Approach to Cardiac Electrical Dynamics: Evaluation of Surrogate Biomarkers of Differences in Drug-Induced Multichannel Block

Abstract

Background and Aims

Torsade de Pointes (TdP), a side effect of many marketed drugs, can lead to sudden cardiac death. Regulatory guidelines require quantification of hERG channel block by QT interval prolongation on ECG, although its predictive value remains low. To propose a novel normalization technique for vectorcardiographic loops, enabling improved derivation of conventional and new indices for the robust identification of multiple cardiac ion channel blockades associated with TdP risk.

Methods

A robust method was developed to obtain angular parameters from ECG loops by normalizing for baseline drift using principal component analysis. Linear (via differentiation) and angular (via quaternion algebra) velocities were assessed to identify differential features between multichannel and selective hERG-blocking drugs. Furthermore, bidirectional baseline correction allowed more accurate extraction of ECG wave extrema and peaks, improving the robustness of the QT interval and other temporal measurements. The proposed dynamic biomarkers were evaluated in 22 subjects enrolled in a clinical trial of four known QT-prolonging drugs.

Results

The two high-risk drugs exhibited drug-induced changes (p < 0.0005) in velocity during ventricular repolarization. Strong calcium or sodium blockers reduced the effect on velocity caused by hERG potassium channel block. A tendency to symmetry of angular values was observed with high-risk drugs. The alternative temporal indices showed a high correlation (r > 0.9) with standard indices. Differences emerged between the T-wave end and the angular velocity marker of ventricular repolarization end.

Conclusion

Spatial analysis of cardiac signals and the new dynamic measures could complement current standards and support safer drug development.