BACKGROUND: The prevalence of left ventricular hypertrophy (LVH), coronary artery disease, and subclinical cardiomyopathy in diabetic patients without known cardiac disease is unclear. We sought the frequency of these findings to determine whether plasma brain natriuretic peptide (BNP) could be used as an alternative screening tool to identify subclinical LV dysfunction. METHODS: Asymptomatic patients with diabetes mellitus without known cardiac disease (n = 101) underwent clinical evaluation, measurement of BNP, exercise stress testing, and detailed echocardiographic assessment. After exclusion of overt dysfunction or ischemia, subclinical myocardial function was sought on the basis of myocardial systolic (Sm) and diastolic velocity (Em). Association was sought between subclinical dysfunction and clinical, biochemical, exercise, and echocardiographic variables. RESULTS; Of 101 patients, 22 had LVH and 16 had ischemia evidenced by exercise-induced wall motion abnormalities. Only 4 patients had abnormal BNP levels; BNP was significantly increased in patients with LVH. After exclusion of LVH and coronary artery disease, subclinical cardiomyopathy was identified in 24 of 66 patients. Subclinical disease could not be predicted by BNP. CONCLUSIONS: Even after exclusion of asymptomatic ischemia and hypertrophy, subclinical systolic and diastolic dysfunction occurs in a significant number of patients with type 2 diabetes. However, screening approaches, including BNP, do not appear to be sufficiently sensitive to identify subclinical dysfunction, which requires sophisticated echocardiographic analysis.