QT dispersion has been adopted as a new index for the noninvasive assessment of the inhomogeneity of repolarization and has been evaluated in several clinical studies as an index of arrhythmia propensity. In most of these studies, indices of dispersion of repolarization were rate corrected by the Bazett formula calculating QT dispersion as QT cmax-QT cmin or JT dispersion as JT cmax-JT cmin, implying that dispersion of repolarization also changes with heart rate. This study aimed to determine in the electrically paced isolated heart whether dispersion of ventricular repolarization is rate dependent. Multiple (5-7) monophasic action potentials (MAPs) were recorded simultaneously from the epicardium and endocardium of both ventricles in 18 isolated Langendorff-perfused rabbit hearts. Hearts were paced from a right ventricular site at basic cycle lengths (CL) between 1,200 and 300 ms in 100-ms decrements. Action potential duration was measured at 90% repolarization (APD90), and recovery time (RT) was defined as the sum of APD90 and activation time in each of the simultaneous MAP recordings. The dispersion of APD90 and RT, respectively, were calculated as the maximal difference among all recordings. APD90 and RT shortened continuously throughout the range of paced steady-state CLs from 1,200 to 300 ms. APD90 was 197.6 +/- 6.1 ms at a CL of 1,200 ms and decreased to 148.5 +/- 2.5 ms at a CL of 300 ms (P < 0.0001). RT was 228.2 +/- 6.2 ms at a CL of 1,000 ms and decreased to 175.9 +/- 2.9 at a CL of 300 ms (P < 0.0001). In contrast, dispersion of APD90 and RT did not change significantly. Dispersion of APD90 was 24.8 +/- 2.3 ms at a CL of 1,200 ms, 26.1 +/- 1.9 msec at a CL of 1,000 ms, and 21.6 +/- 2.1 at a CL of 300 ms (NS). Dispersion of RT was 29.7 +/- 3.4 ms at a CL of 1,200 ms, 29.0 +/- 3.0 ms at a CL of 1,000 ms, and 32.7 +/- 3.2 ms at a CL of 300 ms (NS). In contrast to the duration of the QT interval, dispersion of ventricular repolarization does not change significantly with pacing induced changes in CL. Assuming that the rate-dependent behavior of action potential duration is similar between the rabbit and human heart, a rate correction of parameters of dispersion of repolarization is probably unnecessary.