Both a long QTc and a large QTc dispersion (QTcd) can predispose infarcted patients to ventricular arrhythmias. The former simply reflects a general prolongation of ventricular recovery time, whereas QTcd is useful for revealing regional inhomogeneities of ventricular repolarization. The aim of our study was to evaluate QTc and QTcd behavior during exercise in 50 patients (all men) with previous myocardial infarction, and its possible correlation with the occurrence of exercise-induced premature ventricular complexes (EIPVC). Our patients underwent ergometric stress test with a load increase of 25 W, every 2 min, until the maximal age-related heart rate or symptoms were obtained, followed by a 10-min recovery phase. QTc and QTcd measurement was performed at rest (BS) and during exercise at two progressively increasing heart rate steps: 100-115 beats/min (T1) and 116-130 beats/ min (T2). The patients were divided into two groups according to the absence (group A; n = 22) or presence (group B; n = 28) of EIPVC. In terms of QTcd, no significant difference was found between the two groups at BS, T1, and T2. As for the mean QTc (QTcm), it was significantly longer in group B at BS (416 +/- 22 ms versus 395 +/- 19 ms; P = 0.001) and at T1 (431 +/- 24 ms versus 410 +/- 8 ms; P = 0.0001). When group B was further differentiated into two subgroups-Bx and Bz-according to the severity of EIPVC, we noted that patients with the most severe arrhythmic response (group Bz; n = 12) showed a persisting, significantly longer QTcm than group A (BS, 426 +/- 28 ms versus 395 +/- 19 ms; P < 0.05; T1, 445 +/- 24 ms versus 410 +/- 8 ms; P < 0.05; T2, 427 +/- 17 ms versus 412 +/- 14 ms; P < 0.05), and group Bx (n = 16) (BS, 426 +/- 28 ms versus 409 +/- 15 ms; P < 0.05; T1, 445 +/- 24 ms versus 420 +/- 19 ms; P < 0.05; T2, 427 +/- 17 ms versus 410 +/- 17 ms; P < 0.05). Group Bx showed a significantly longer QTcm than group A only at BS (409 +/- 15 ms versus 395 +/- 19 ms; P < 0.05). No significant difference in QTcd was found between the three groups at BS, T1, and T2. We also noted that the relationship between QTcm and QTcd was modified by the exercise, changing from a trend of direct relation at BS, towards an inverse one during effort, which reached significance at T2 (r = -0.319; P = 0.037). Based on our data, EIPVC occurrence seems to be more affected by the total duration rather than by regional inhomogeneities of the ventricular recovery time. In those patients with the most severe arrhythmic response, the autonomic modifications generated by the exercise succeed in attenuating only the regional inhomogeneities, but do not eliminate the differences in total duration of the repolarization period.