Patients with non-insulin-dependent diabetes (NIDDM) are at independent risk of cardiovascular death. The reason is only partially understood. The aim of our study was therefore to evaluate the impact of corrected QT interval length (QTc) and QT dispersion (QT-disp) on mortality in a cohort of 324 Caucasian NIDDM patients. A resting 12-lead ECG was recorded at baseline. Maximum (QT-max) and minimum QT (QT-min) intervals were measured, and QT-max was corrected for heart rate (QTc-max). QT-disp was defined as the difference between QT-max and QT-min. QTc-max was 454 (376-671) ms(1/2) (median (range)) and QT-disp 61 (0-240) ms. Prolonged QTc interval (PQTc), defined as QTc-max > 440 ms(1/2), was present in 67% of the patients and prolonged QT-disp (PQT-disp), defined as QT-disp > 50 ms, was present in 51%. During the 9-year follow-up period, 100 patients died (52 from cardiovascular diseases). Thirty-seven percent of the patients with PQTc died compared with 17% with normal QTc interval (p<0.001). The Cox proportional hazard model, including putative risk factors at baseline, revealed the following independent predictors of all cause mortality; QTc-max (p<0.05), age (p<0.0001), albuminuria (p<0.01), retinopathy (p<0.01), HbA1c (p<0.05), insulin treatment (p<0.01), total cholesterol (p<0.01), serum creatinine (p<0.05) and presence of cardiac heart disease based on Minnesota coded ECG (p<0.001). Whereas QT-disp was not a predictor, QTc-max interval was an independent predictor of cardiovascular mortality. Our study showed a high prevalence of QTc and QT-disp abnormalities and indicated that QTc-max but not QT-disp is an independent predictor of all cause and cardiovascular mortality in NIDDM patients.