The relation between QT interval and heart rate during ramp exercise tests on a bicycle was investigated in 37 healthy individuals (21 women) without regular medication and with a normal thallium-201 exercise scintigram (mean age 52.9 +/- 8.3, range 38-68). The test started at 20 W and the load increased by 10 W min-1. A 12-lead ECG was recorded twice every min and mean complexes (during a 15 s period) were calculated by computer. At rest the QT interval (in s) corrected for heart rate (QTc) for women and men was 0.408 +/- 0.004 and 0.399 +/- 0.005, respectively, P > 0.05). During exercise there was no difference in QT interval between women and men or between younger (< 50 years) and older (> 50 years) individuals. A straight line was used to describe the relation between QT interval and heart rate (beats min-1; QT = 0.459-12.3 x 10(-4)*HR). A 95% prediction interval around the regression line was determined using a non-parametric statistical method. When QTc was calculated using Bazett's formula with a cut-off value of QTc = 0.46, 19 individuals (11 women) had a prolonged QT interval during exercise. It is concluded that the relation between QT interval and heart rate can during exercise be described by a straight line for normal individuals. It is not valid to use Bazett's formula for correction of QT intervals during ramp exercise tests.