To evaluate whether or not coronary microvascular dysfunction is associated with exercise-induced myocardial ischemia in left ventricular dysfunction of unknown cause, both the treadmill exercise test (TET) and coronary hemodynamics were studied in 20 patients with impaired left ventricular ejection fraction (<50% by radionuclide ventriculogram), normal cardiac size, normal coronary angiogram and no evidence of clinical heart failure. Ten subjects with atypical chest pain were studied as the control. Coronary hemodynamics were studied both at baseline and after dipyridamole infusion (0.56mg/kg, i.v. for 4'). There was no difference in age, gender, blood pressure, baseline great cardiac venous flow (GCVF) and coronary vascular resistance between ten patients with a positive TET and the other ten with a negative TET. At baseline, coronary sinus oxygen concentration was increased and myocardial oxygen consumption reduced in patients with a positive TET compared with those with negative a TET. After dipyridamole infusion, maximum GCVF (102+/-47 vs. 144+/-31 ml/min, P=0.027) and coronary flow reserve (2.31+/-0.49 vs. 3.00+/-0.61, P=0.012) were significantly reduced and minimum coronary vascular resistance was higher (1.00+/-0.42 vs. 0.63+/-0.12 mmHg/ml/min, P=0.016) in patients with a positive TET than in those with a negative TET. At follow-up, 40% of patients with a positive TET and 10% of those with a negative TET developed clinical heart failure with a dilated left ventricle during a period of 45 months. Thus, coronary microvascular function is heterogeneous in patients with left ventricular dysfunction of unknown cause. In some of them, coronary microvascular dysfunction could be related to the presence of exercise-induced myocardial ischemia, suggesting that similar pathophysiology underlies the early stage of dilated cardiomyopathy and syndrome X.