Clinical, electrocardiographic, and thallium-201 single-photon emission computed tomography data were evaluated in 397 consecutive patients divided into 3 groups according to coronary hyperemic stimulation: 186 patients (group I; Ex) had maximal symptom-limited exercise ergometric stress testing, 93 patients (group II; Dip) had intravenous dipyridamole (0.7 to 0.8 mg/kg) stress testing, and 118 patients (group III; Dip+Ex) had dipyridamole (0.7 to 0.8 mg/kg) plus nonlimited (i.e., symptom-limited) exercise stress testing, achieving a maximal workload (mean +/- SD) of 102 +/- 37 W. Clinical tolerance was higher in Ex than in Dip groups (p < 0.01), and tended to be higher in Dip+Ex than in Dip groups (p = NS). Image quality--as judged by signal-to-noise ratios--was superior in Ex and Dip+Ex groups when compared with the Dip group (p < 0.01). Chest pain and electrocardiographic positivity were more frequent in the Dip+Ex group than in the Dip group (p < 0.05), despite more extensive coronary artery disease (CAD) in the Dip group; and reversible scintigraphic defects were more frequent in Dip+Ex versus Dip (p < 0.01) and in Ex versus Dip groups (p < 0.05) in patients with established CAD, as well as for the whole group. We conclude that, in patients unable to achieve 85% of their maximal predicted heart rate, the combination of high-dose dipyridamole plus nonlimited exercise stress testing is superior to dipyridamole stress testing alone, and comparable to maximal exercise testing.