BACKGROUND--Many studies have shown that coronary flow reserve is reduced in patients with chest pain and angiographically normal coronary arteries. The methods used to assess coronary blood flow have varied, but in nearly all reports dipyridamole has been used to bring about vasodilatation. This study was designed to assess whether the apparent impairment of coronary flow reserve seen with dipyridamole could be reproduced with either papaverine or adenosine, which induce maximum coronary blood flow by different mechanisms. METHODS--25 patients with chest pain and angiographically normal coronary arteries were studied with an intracoronary Doppler flow probe and quantitative angiography to determine epicardial coronary artery area, coronary blood flow velocity, coronary flow reserve, and coronary vascular resistance index (CVRI, the ratio of resistance after intervention to basal resistance). All patients received papaverine 8 mg. Eight patients with positive exercise tests received intracoronary papaverine (8 and 10 mg), intracoronary adenosine (6, 20, 60 micrograms), and high-dose intravenous dipyridamole (0.84 mg/kg). RESULTS--The velocity ratio (peak after intervention: baseline) (mean (SEM)) after 8 mg papaverine was 3.3 (0.2) (n = 25) and the coronary flow reserve was 4.1 (0.3) (n = 25). There were no differences between patients with a positive (n = 16) or negative (n = 9) exercise test. In eight patients coronary flow reserve was measured after increasing doses of papaverine, adenosine, and dipyridamole. Coronary flow reserve was 4.5 (0.3) with papaverine, 4.8 (0.3) with adenosine, and 3.5 (0.4) with dipyridamole (p = 0.08 v papaverine and adenosine). CVRI was 0.22 (0.01) with papaverine, 0.21 (0.02) with adenosine, and 0.29 (0.03) with dipyridamole (p < 0.05 v papaverine, p = 0.09 v adenosine). CONCLUSIONS--These results indicate that measurement of coronary flow reserve and CVRI in patients with chest pain and normal coronary arteries depends on the pharmacological stimulus. Normal values were obtained with papaverine in all patients, irrespective of the exercise test response. In patients with a positive exercise test significantly lower values were obtained with dipyridamole than with papaverine, or adenosine. The reported impairment of coronary flow reserve in patients with angina and normal coronary arteries may reflect the variability in response to different pharmacological agents. The mechanism underlying this variability is unknown, but may involve an abnormality of adenosine metabolism in the myocardium.