Abnormal repolarization is associated with arrhythmogenesis. Because of controversies in existing methodology, new computerized methods may provide more reliable tools for the noninvasive assessment of myocardial repolarization from the surface electrocardiogram (ECG). Measurement of the interval between the peak and the end of the T wave (TpTe interval) has been suggested for the detection of repolarization abnormalities, but its clinical value has not been fully studied. The intrasubject reproducibility and reliability of automatic measurements of QT, QT peak, and TpTe interval and dispersion were assessed in 70 normal subjects, 49 patients with acute myocardial infarction (5th day; MI), and 37 patients with hypertrophic cardiomyopathy (HC). Measurements were performed automatically in a set of 10 ECGs obtained from each subject using a commercial software package (Marquette Medical Systems, Milwaukee, WI, U.S.A.). Compared to normal subjects, all intervals were significantly longer in HC patients (P < 0.001 for QT and QTp; p < 0.05 for TpTe); in MI patients, this difference was only significant for the maximum QT and QTp intervals (P < 0.05). In both patient groups, the QT and QTp dispersion was significantly greater compared to normal subjects (P < 0.05) but no consistent difference was observed in the TpTe dispersion among all three groups. In all subjects, the reproducibility of automatic measurement of QT and QTp intervals was high (coefficient of variation, CV, 1%-2%) and slightly lower for that of TpTe interval (2%-5%; p < 0.05). The reproducibility of QT, QTp, and TpTe dispersion was lower (12%-24%, 18%-28%, 16%-23% in normal subjects, MI and HC patients, respectively). The reliability of automatic measurement of QT, QTp, and TpTe intervals is high but the reproducibility of the repeated measurements of QT, QTp and TpTe dispersion is comparatively low.