Background: The relation between the electrocardiographic T-wave and the steepness of local repolarization time (RT) gradients is not completely understood.
Objective: To assess the sensitivity of 12-lead ECG T-wave markers to identify repolarization dispersion in intact hearts.
Methods: Four pig hearts were Langendorff-perfused in a human-shaped torso tank. Pinacidil (shortening RT) was infused in the left anterior descending artery, and dofetilide (prolonging RT) was infused in the rest of the heart. Maximum RT gradients were determined as the 95^th percentile of all local RT gradients (defined as the RT difference divided by interelectrode distance), as determined from 108-electrode epicardial sock recordings during atrial pacing. QT_time (time between QRS start and T-wave end), T_peak-end (time between peak-to-end of T-wave) and upslope (slope of the rising part of the T-wave) were determined from the 12-lead body surface ECG (green, blue and red lines respectively in panel A). In addition, the time difference between the end of the upslope of leads V₁ and V₆ was determined (orange lines, panel A).
Results: RT gradients ranged from 24 to 44 ms/cm (mean 34.1 ms/cm) at baseline and from 46 to 107 ms/cm (mean 76.1 ms/cm) after drug infusion. RT gradients did not show correlation with mean QT_time (Pearson’s R=0.31, panel B), but did show significant correlation with mean T_peak-end, upslope and V₁-V₆ upslope difference (Pearson’s R=0.95, 0.83 and 0.96 respectively, panel B).
Conclusion: In drug-induced pronounced RT dispersion, QT_time does not reflect the underlying RT gradient steepness, but mean T_peak-end, upslope and V₁-V₆ upslope difference are accurate markers for RT gradient steepness.