SUMMARYMicrovolt T-wave alternans (TWA) and QT interval dispersion (QTD), which reflect temporal and spatial repolarization abnormalities, respectively, have been proposed as useful indices to identify patients at risk for ventricular tachyarrhythmias (VTs). The purpose of this study was to clarify which repolarization abnormality marker is more useful in predicting arrhythmic events in patients with dilated cardiomyopathy (DCM).Forty-two consecutive nonischemic DCM patients underwent the assessment of TWA and QTD. Patients undergoing antiarrhythmic pharmacotherapy, except β-blockers and those with irregular basic rhythms, were excluded from entry. Eight patients were also excluded because of indeterminate test results. Therefore, 34 DCM patients were prospectively assessed. The end point of the study was the documentation of VT defined as ≥ 5 consecutive ectopic beats during the follow-up period.TWA and QTD (≥65 msec) were positive in 24 (80%) and 11 (37%) of 30 patients with available follow-up data, respectively. There was no relationship between TWA and QTD. During a follow-up of 13±11 months, VTs occurred in 13 patients (43%). In Cox regression analysis, TWA was a significant risk stratifier (p=0.02), whereas QTD was not. The sensitivity, specificity, and positive and negative predictive values of TWA in predicting VTs were 100%, 35%, 54%, and 100%, respectively.TWA could be a useful noninvasive index to identify patients at risk for VTs in the setting of DCM. This study may suggest that temporal repolarization abnormality is associated more with arrhythmogenesis than with spatial repolarization abnormality in DCM patients. (Jpn Heart J 2001; 42: 451-457)