BACKGROUND: QT dispersion (QTd) in 12-lead ECG, a noninvasive parameter of the degree of inhomogeneous myocardial repolarization, has been reported useful in assessing the risk of ventricular tachyarrhythmias and sudden cardiac death in patients with coronary artery disease. Restenosis after coronary stenting was not infrequent. However, there was no reliable ECG predictor for stent restenosis. This study was performed to evaluate the value of QTd as a predictor of coronary artery stent (CAS) restenosis. METHODS: One hundred eighty eight patients who underwent both successful coronary artery stenting for significant coronary artery stenosis and follow-up coronary artery angiography were included in this study. QTcd (difference of maximum and minimum QTc intervals) was measured in the 12-lead surface ECG, which was recorded 6-12 hours after the successful CAS and 12-18 hours before the follow-up coronary angiography, which was performed 1-12 months (Group I: 6+/-3 months: Group II: 7+/-2 months, p=NS) after CAS. The follow-up coronary angiography demonstrated no restenosis of the stented coronary artery in 122 patients: Group I: 97 men, 25 women: 58+/-13 years), but restenosis in 66 patients:(Group II: 58 men, 8 women: 61+/-13 years). RESULTS: There were no significant differences in the distribution of the target vessels between the two groups. The minimal luminal diameters of the target vessels were similar in the two groups before and after CAS (0.99+/-0.49 mm vs. 0.92+/-0.51 mm: 2.94+/-0.71 mm vs. 2.71+/-0.79, respectively). QTcd after CAS was 52.6+/-22.0 ms in the Group I and 51.6+/-30.5 ms in the Group II, with no significant difference between the two groups. QTcd at the follow-up examination was 50.0+/-19.8 ms in the Group I and 58.3+/-21.6 ms in the Group II, showing a significant difference between the two groups (p<0.05). The change in QTcd (DQTcd) during the follow-up period was significantly different between the two groups: QTcd significantly increased in the Group II, particularly in patients with angina pectoris, single or two-vessel disease, or left coronary artery stenosis, compared with the Group I (Group I:-2.18+/-22.11 ms, Group II: 8.77+/-28.82 ms, p<0.05). The sensitivity, specificity, negative predictive value, positive predictive value and diagnostic accuracy of DQTcd (cut-off value, 7 ms) in predicting coronary artery stent restenosis were 64%, 65%, 77%, 49%, and 64%, respectively. CONCLUSION: The change of QTc dispersion according to restenosis was significantly different, but its clinical usefulness is limited due to low positive predictive value.