Circadian variations in the QT interval (QT) and QT dispersion are decreased in patients with type 2 diabetes because of cardioneuropathy. Insulin resistance has been recently identified as an independent determinant of QT prolongation in normoglycemic women. However, the relationship between QT prolongation and the degree of insulin resistance as well as circadian variation remains unclear in diabetic patients. This study was designed to assess the relationship between insulin resistance and the circadian variation in QT measurements in patients with type 2 diabetes. In 14 patients with diabetes, QT, corrected QT (QTc), QT dispersion, QTc dispersion, and RR interval (RR) were analyzed using 12-lead Holter monitoring and commercial software. The degree of diurnal variation in each measurement was defined as the amplitude between the maximum and mean values on curves fitted using the mean cosinor method (A_QT, A_QTc, A_QT dispersion, A_QTc dispersion, and A_RR). The cosine curve was fitted to all measured values in each QT measurement and RR for 24 h. Insulin resistance (glucose infusion rate (GIR)) was measured using the euglycemic hyperinsulinemic glucose clamp method. The maximum QT, QTc, QT dispersion, and QTc dispersion were >450 ms. GIR was significantly correlated with A_QT only (r = 0.59, P < 0.05). GIR was not correlated with other variables, and was dependent only on the circadian variation in QT.