In a double-blind study, eight patients with symptomatic myocardial ischaemia and nine with asymptomatic myocardial ischaemia were compared during physical exercise under naloxone (6 mg i.v.) or placebo. Plasma beta-endorphin, cortisol and catecholamines were measured before exercise, during maximal exercise, and 10, 20 and 60 min after exercise. A tourniquet pain test (on the forearm, under control of transcutaneous PO2), and an electrical pain test (intracutaneous electrode placed in the finger with the electrical stimulus under computer control and two-interval forced-choice psychophysical technique) were performed before exercise as well as immediately after, and 60 min after exercise. Plasma beta-endorphin levels increased significantly (P < 0.01) during exercise in symptomatic and asymptomatic patient groups; every patient showed an increase on betaendorphins during and after exercise. However, the increase found in beta-endorphins during and after exercise was significantly larger (P < 0.01) in asymptomatic than in symptomatic patients. After naloxone, this difference was no longer evident. Angina pectoris during exercise was reported with less latency in symptomatic patients (P < 0.05) and occurred in two of nine asymptomatic patients following naloxone. The time course of plasma cortisol levels exhibited the same pattern as beta-endorphins with the same significant differences between symptomatic and asymptomatic groups. Electrical pain thresholds, though on average higher in asymptomatic patients (2.21 mA vs. 0.79 mA), were not affected by exercise or naloxone. Asymptomatic patients required more time to reach pain thresholds in the tourniquet pain test (P < 0.02). After exercise, tourniquet pain thresholds were significantly lower (P < 0.01) under naloxone compared with placebo. The results suggest that there arequantitative differences in the endorphinergic regulation of pain in patients with symptomaticand asymptomatic myocardial ischaemia.