Background QT dispersion has been considered a surrogate for heterogeneity of repolarization, leading to ventricular arrhythmias. Methods High-resolution 12-lead electrocardiograms were obtained in 15 patients with a history of ventricular tachycardia or ventricular fibrillation, 15 patients with congestive heart failure, 17 patients with a history of previous Q-wave myocardial infarction without heart failure, and 23 healthy control subjects. Results QTc dispersion was prolonged in all 3 patient groups compared with controls (71 ± 22, 68 ± 31, 61 ± 27 vs 44 ± 17 msec, P =.003), but no difference was seen between heart disease groups. QTc dispersion was strongly correlated with the QTc max (r = 0.73, P <.0001) but did not correlate with the QTc min (r = 0.04, P =.76). QTc dispersion also strongly correlated with the JTc max (r = 0.54, P <.0001) but did not correlate with JTc min (r = –0.007, P =.95). QTc dispersion correlated inversely with T-wave amplitude (r = –0.35, P =.003). When all 876 electrocardiographic signals were considered, a significant negative correlation was present between QTc duration and T-wave amplitude (r = –0.133, P =.0002). Logistic regression analysis failed to demonstrate any independent risk factors that predicted ventricular arrhythmias, including all measures of dispersion. Conclusions The measurement of QT dispersion is strongly influenced by the maximum QT interval, as well as by changes in T-wave amplitude. QT “dispersion” may represent a summary of these changes that reflect the underlying myocardial process but does not represent an accurate quantitative measure of heterogeneity of refractoriness. (Am Heart J 2001;141:111-6.)