To determine thromboxane A 2 release in coronary artery disease, we measured its stable metabolite thromboxane B 2 by radioimmunoassay in 20 patients. In 15 patients with stable disease (last angina episode >96 hours before study), coronary venous thromboxane B 2 concentrations were lower than in aortic blood (mean 109 ± 36 vs 194 ± 40 pg/ml, p < 0.001). In contrast, in five other patients with spontaneous angina, coronary venous thromboxane B 2 concentrations were higher than aortic thromboxane B 2 concentrations during the angina episode (mean 1716 ± 316 vs 875 ± 388 pg/ml, p < 0.02). Plasma thromboxane B 2 levels were in the normal range (mean 175 ± 35 pg/ml) in patients with stable angina but significantly ( p < 0.02) higher in patients with spontaneous angina. With atrial pacing to the point of chest pain and/or ECG changes in patients with stable coronary artery disease, aortic thromboxane B 2 concentrations increased in 10 of 13 patients (mean 283 ± 70 pg/ml, p < 0.02). Coronary venous thromboxane B 2 concentrations increased in seven patients at peak pacing rates (mean 223 ± 76 pg/ml) and in three other patients after termination of pacing. These data indicate that release of thromboxane A 2 is much greater during spontaneous angina than with pacing stress in patients with coronary artery disease. Thromboxane A 2 released during spontaneous or pacing-induced angina may modulate coronary and systemic vascular tone. Enhanced thromboxane A 2 activity may either precede or follow myocardial ischemia and could be a factor in the initiation and propagation of the ischemic episode.