The goal of this study was to determine whether dipyridamole-induced ST-segment depression reflects more severe or extensive myocardial hypoperfusion than the absence of this electrocardiographic finding. The clinical, hemodynamic and scintigraphic correlates of ST-segment depression during intravenous dipyridamole infusion were studied in 204 consecutive patients undergoing dipyridamole stress thallium-201 (Tl-201) imaging for evaluation of coronary artery disease. Of 182 patients with a diagnostic baseline electrocardiogram, 28 (15%) developed ST depression after dipyridamole. Patients with ST depression, compared with those without, were older (64 +/- 1 vs 60 +/- 1 years; p less than 0.03) and had a higher frequency of chest pain (57 vs 23%; p less than 0.001) and a higher heart rate-blood pressure product (12.7 +/- 0.6 vs 11.2 +/- 0.2 x 10(3); p less than 0.008) after dipyridamole. Patients with ST depression were more likely to have Tl-201 redistribution (64 vs 38%; p less than 0.02) and a greater number of redistribution defects (2.3 +/- 0.04 vs 0.9 +/- 0.1, p less than 0.001) than were those without ST depression. By multivariate logistic regression analysis, the most powerful correlate of ST depression was the number of segments having Tl-201 redistribution (p less than 0.001). Other independent correlates were presence of chest pain, heart rate at Tl-201 injection, and age. Thus, the determinants of dipyridamole-induced ST-segment depression include the scintigraphic extent of reversible hypoperfusion, as well as indexes of myocardial oxygen demand.