We would like to thank Dr. Madias for his interest in and comments on our report. We take this opportunity to attempt to address some of the points raised.

First, we did not find a significant correlation between the percentage change in right ventricular end-diastolic volume and QRS duration after percutaneous pulmonary valve implantation at 1 year (p = 0.16). This is in keeping with our published findings of a significant decrease in right ventricular end-diastolic volume at 1 year, with no associated significant shortening of QRS duration.

With regard to Dr. Madias’s second point, the electrocardiographic (ECG) machines in use at the 3 study locations are unable to provide automatic measurements of QRS, QT, or JT dispersion, thereby necessitating manual measurement of all leads. Furthermore, given that different ECG machines would necessarily be used at each site over the study span of 7 years, the assumption that all automatic readings for QRS duration and corrected QT interval are comparable and accurate may render statistical analysis invalid, particularly in these patients, in whom inaccurate measurements, such as including a prominent but definite U wave in the QT measurement, are possible. Last, most published research in this field has determined ECG parameters by manual measurement or confirmed automatic measurement with subsequent manual assessment.

We appreciate that further information on the characterization of the electrocardiograms may be of interest to readers of The American Journal of Cardiology . However, we chose to focus our study primarily on the right ventricle, because this was the cardiac chamber subject to the main hemodynamic changes with intervention. We elected not to study other ECG parameters on this occasion, because the varied anatomic substrates presented would themselves influence P-wave amplitude and duration, possibly preventing reliable statistical analysis. Future studies of larger homogenous cohorts would enable such an analysis.

Dr. Madias further proposes that the presence of peripheral edema in our study cohort may have led to a blunting of the ECG changes with corrective intervention. We would like to point out that at the time of percutaneous pulmonary valve implantation, none of the included patients were in acute right ventricular failure. In general, the finding of peripheral edema is extremely rare in patients with right ventricular outflow tract dysfunction and congenital heart disease. Furthermore, we did not observe any significant changes in QRS duration at 24 hours after percutaneous pulmonary valve implantation.

Finally, Dr. Madias suggests that any reduction in QRS dispersion (48 ms rather than 32 ms in the example he gives) is due purely to the reduction of maximum and minimum QRS width alone and that as such, a proportional reduction in minimal QRS width may not be physiologically important. A reduction in QRS dispersion, however, implies global reduction in activation delays between myocardial segments, even if the overall activation time of the ventricles (represented by QRS width) is reduced as well. This homogenization in activation time will thus reduce the opportunity for the dispersions in conduction and repolarization times to promote conduction block and reentry and precipitate ventricular arrhythmias. Therefore, although these changes may appear small in absolute terms, they have important physiologic consequences: in the general heart failure population, as in those with congenital heart lesions, there is a recognized correlation between QRS dispersion and risk for sudden death, implying that this parameter has important clinical implications. Furthermore, patients with repaired tetralogy of Fallot have been found to have prolongation of QRS dispersion in association with right ventricular regional wall motion abnormalities, implying a role for a mechanoelectric feedback mechanisms in future arrhythmic risk. Therefore, reductions in dispersion and absolute QRS width are manifestations of positive electrical remodeling.