The aim of this study was to address whether albuminuria could predict myocardial dysfunction in diabetic patients without overt heart disease. We studied 67 patients with normal left ventricular (LV) ejection fraction and no evidence of LV hypertrophy or coronary artery disease (47 patients with type 2 diabetes mellitus and hypertension and 20 patients with hypertension only). Diabetes patients were divided into 3 groups based on albuminuria status: group II = no albuminuria (n = 20, <30 mg/d), group III = microalbuminuria (n = 13, 30-300 mg/d), and group IV = macroalbuminuria (n = 14, >300 mg/d). Twenty patients with hypertension only served as a control group (group I). Conventional 2-dimensional and Doppler echocardiography was done. Peak strain, peak systolic strain rate (SR), and peak diastolic SR of 6 LV segments in the apical views were measured and averaged in each patient. Conventional 2-dimensional parameters such as LV ejection fraction; left atrium volume index; LV mass; deceleration time; and mitral early peak, mitral late peak, myocardial early peak diastolic, and myocardial peak systolic velocities were not different among the 4 groups. However, peak strains were significantly lower in group III (P = .002) and group IV (P < .001) than in group I; and the absolute value of peak systolic SR was lower in group III (P = .033) and group IV (P < .001) than in group I. Furthermore, the value of peak diastolic SR was lower in group IV than in group I (P = .014). In diabetic patients with albuminuria, Doppler strain and SR imaging detected subclinical LV systolic and diastolic dysfunction; and albuminuria was associated with myocardial dysfunction in diabetic patients without overt heart disease.