We evaluated the utility of positron emission tomography in differentiating patients with idiopathic dilated cardiomyopathy from those with ischemic cardiomyopathy. Twenty consecutive non-diabetic patients with dilatation (end-diastolic volume > or = 120 cc/m2) and reduced systolic function (ejection fraction < or = 40%) of the left ventricle on cineangiography, underwent coronary angiography, F18 fluorodeoxyglucose (F18-FDG) (glucose load technique) and N13-ammonia (N13-NH3) positron emission tomography. A semiquantitative score based on the extension and the severity of the uptake defects was calculated. Endomyocardial biopsy was performed in patients with normal coronary arteries. Ten patients (group A) had normal coronary arteries and histologic features of the endomyocardium fitting with the diagnosis of idiopathic dilated cardiomyopathy. Cineangiography showed critical stenosis of at least one major coronary artery in the other 10 patients (group B). The two groups were similar in age. left ventricular end-diastolic volume and ejection fraction. Both N13-NH3, positron emission tomography and F18-FDG positron emission tomography scores were lower in group A than in group B: 0.1 +/- 0.3 vs. 10.6 +/- 5.1 (P<0.0001) and 2.4 +/- 4.4 vs. 9.9 +/- 4.1 (P<0.0001) respectively. but only N13-NH3 positron emission tomography allowed a complete separation of the two groups (score range 0-1 group A vs. 4-12 group B). The F18-FDG score value showed some overlapping between the two groups (score range 0-12 in the group A vs. 2-17 in the group B). All three idiopathic dilated cardiomyopathy patients with a F18-FDG score value >2 had left bundle branch block on standard ECG. Positron emission tomography imaging with N13-NH3 and F18-FDG provided a complete differentiation between idiopathic dilated cardiomyopathy and ischemic cardiomyopathy patients. However patients with left bundle branch block on ECG could present defects in FDG uptake even if affected by idiopathic dilated cardiomyopathy.