A comparative assessment of regional coronary flow reserve, quantitative percent diameter coronary stenosis and exercise-induced perfusion and wall motion abnormalities was performed in 39 patients with coronary artery disease. Coronary flow reserve was determined by a digital angiographic technique utilizing contrast medium as the hyperemic agent. Percent diameter stenosis was calculated by an automated quantification program applied to orthogonal cineangiograms. Thallium-201 scintigraphy and radionuclide ventriculography were used to assess regional perfusion and wall motion abnormalities, respectively, at rest and during exercise. In Group A, 19 patients without transmural infarction or collateral vessels, coronary flow reserve was inversely related to percent diameter stenosis (r = -0.61, p less than 0.0001), and scintigraphic abnormalities occurred only in vascular distributions with a coronary flow reserve of less than 2.00. There was a strong relation among abnormal regional exercise results, stenoses greater than 50% and reactive hyperemia of less than 2.00. Patients with multivessel disease, however, often had normal exercise scintigrams in regions associated with greater than 50% stenosis and low coronary flow reserve when other regions had a lower coronary flow reserve or higher grade stenosis, or both. In Group B, 20 patients with angiographically visible collateral vessels, 12 of whom had prior myocardial infarction, coronary flow reserve correlated less well with percent diameter stenosis than in Group A (r = -0.47, p less than 0.004). As in Group A patients, there was a significant relation between abnormal exercise test results and stenoses greater than 50%. However, reactive hyperemia values were generally lower than in Group A, and positive exercise stress results were strongly correlated only with highly impaired flow reserves of 1.3 or less. In Group B patients, the coronary flow reserve of vessels with less than 50% stenosis was significantly lower than that of similar vessels in Group A patients (2.40 +/- 0.79 versus 1.56 +/- 0.43; p less than 0.0002). It is concluded that: there is a general relation between quantitative percent diameter stenosis and reactive hyperemia that is not of sufficient precision to allow accurate prediction of coronary flow reserve in individual cases; exercise scintigraphic abnormalities are usually associated with low coronary flow reserve, and the relation between these two functional tests is stronger than the relation between exercise test results and quantitative percent diameter stenosis.(ABSTRACT TRUNCATED AT 400 WORDS)