OBJECTIVE This clinical study was designed to compare rate-dependent effects of class III agents on QT prolongation. BACKGROUND Clinical data that compare the electrophysiologic differences among class III agents with different selectivity for potassium channels are still lacking. METHODS QT intervals were measured over a wide range of preceding RR intervals during sinus rhythm by 24-h Holter electrocardiography before and after oral administration of four class III agents: E4031, dofetilide, MS551 and d-sotalol. Rate-dependent changes in the QT interval were assessed by the slope of the linear regression line estimating the QT-square root of RR relation. RESULTS All agents significantly increased the mean slope: E4031 increased the mean [+/- SD] value from 0.32 +/- 0.05 to 0.42 +/- 0.13 (p < 0.01), dofetilide from 0.32 +/- 0.03 to 0.50 +/- 0.12 (p < 0.03), MS551 from 0.35 +/- 0.06 to 0.45 +/- 0.10 (p < 0.02) and d-sotalol from 0.31 +/- 0.05 to 0.33 +/- 0.04 (p < 0.05). However, in those patients given either E4031, dofetilide or MS551, the degree of QT prolongation was smaller at shorter square root of RR intervals and was better preserved at shorter square root of RR intervals by d-sotalol, with a smaller increase in slope (p < 0.02 vs. dofetilide and MS551). CONCLUSIONS On ambulatory electrocardiography, reverse use dependence in QT prolongation was least prominent with d-sotalol among the four study drugs. In the range of physiologic heart rates, class III agents could manifest different profiles of rate dependence in their QT-prolonging effect.