Evaluation of electrocardiogram T-wave dispersion measurement methods.

Methods for measuring the T-wave dispersion are difficult to improve, because the exact result is unknown. This study describes a flexible and cheap method for analysing the measurement problems. The analysis consisted of measuring the T-wave dispersion of electrocardiograms and vector loops with simulated dispersion. The vector loops were simulated using two primary T-waves that were identical, quasi-monophasic and located in orthogonal leads. The dispersion was the time delay between the primary T-waves. The electrocardiograms were projections of the vector loops on approximations to the orientations of leads in the standard 12-lead ECG. This simulation model was validated by matching simulated ECGs and vector loops with measured ECGs and vector loops drawn from a population of 247 patients with acute myocardial infarction. The analysis of the measurement problems was demonstrated by measuring an approximation to the classical QT-dispersion in ECGs with simulated dispersion in the interval 0-100 ms, in steps of 2 ms. This showed that the QT-dispersion was unable to measure the simulated dispersion accurately in the clinically relevant interval of 50-100 ms. The results indicated that the low prognostic accuracy of the QT-dispersion could be due to an inaccurate measurement method.