Myocardial blood flow (MBF) was measured using continuous inhalation of oxygen-15-labeled carbon dioxide, and positron emission tomography before and after intravenous dipyridamole in 13 patients with syndrome X (angina pectoris, angiographically normal coronary arteries, positive exercise test and negative ergonovine test), 7 healthy subjects and 8 patients with 1-vessel coronary artery disease (CAD). In patients with syndrome X, baseline MBF was greater than in healthy subjects and patients with CAD (1.24 +/- 0.27 vs 0.87 +/- 0.07 and 1.03 +/- 0.23 ml/g/min, respectively; p < 0.05), and more heterogeneous (34 +/- 7 vs 26 +/- 5 and 25 +/- 6, respectively; p < 0.05) as assessed by the coefficient of variation among myocardial regions < or = 2.3 cm3. After dipyridamole, MBF in patients with syndrome X was similar to that in healthy subjects (2.95 +/- 0.75 vs 3.40 +/- 0.82 ml/g/min; p = NS) and greater than in patients with CAD (1.78 +/- 0.76 ml/g/min; p < 0.05). However in patients with both syndrome X and CAD, MBF was more heterogeneous than in healthy subjects (48 +/- 12 and 48 +/- 11, respectively, vs 30 +/- 7; p < 0.01). Thus, in patients with syndrome X, MBF is abnormally heterogeneous both at baseline and after dipyridamole. These findings are compatible with the presence of dynamic alterations of small coronary arteries. Because these alterations appear to be very sparse within the myocardium, they can be undetected when myocardial perfusion, function and metabolism are assessed using conventional methods that are unable to detect small myocardial regions.