Background Sex differences have been described in resting cardiac repolarization and susceptibility to torsade de pointes in humans. This study compares the QT-interval and QT-dispersion dynamics during exercise and recovery between healthy men and women.Methods Twenty healthy subjects (10 males aged 30 ± 4 years, 10 females aged 31 ± 11 years) underwent symptom-limited bicycle ergometry followed by a 10-minute recovery period. Digital 12-lead electrocardiograms (ECG) were recorded every 10 seconds during exercise and recovery. For each lead, the QTp interval (Q onset to T peak) was automatically measured by use of QT Guard (GE Marquette, Milwaukee, Wis). QTp dispersion was defined as the difference between the maximum and minimum QTp for each ECG. To quantify QT dynamics, we fit the QTp in lead V3 (QTpV3) versus cycle length (CL) relationship to a quadratic function during exercise and recovery with nonlinear regression analy- sis. Similar regression analysis was performed for the QTp dispersion versus CL relationship.Results At baseline, QTpcVs was longer in women than in men (338 ± 25 vs 278 ± 15 ms, P < .0001), but QTp dispersion was similar (35 ± 18 vs 41 ± 19 ms). At peak exercise, QT dispersion decreased compared with baseline in both men and women. During exercise and recovery, women had a steeper QTpVs-CL relationship. QTpVs hysteresis, a measure of the exercise and recovery QTpVs-CL curve separation, was greater in women than in men when measured 1 minute into recovery (33 ± 20 vs 6 ± 8 ms, P < .001). No sex difference in QTp-dispersion-rate adaptation was observed during exercise or recovery.Conclusions Healthy women exhibit greater QT-interval-rate adaptation during both exercise and recovery than men, resulting in more QT-interval hysteresis. Greater QT prolongation during decelerating heart rates in recovery may play a role in increasing proarrhythmia risk in women.