The clinical and haemodynamic effects of adrenaline infusion (30 ng kg-1 min-1) producing plasma adrenaline concentrations in the range seen during acute myocardial infarction and of placebo were investigated in a crossover design in 14 patients with stable coronary heart disease. Adrenaline infusion resulted in electrocardiographic evidence of myocardial ischaemia (greater than or equal to 1 mm (0.1 mV) horizontal or downsloping ST segment depression) in 10 patients and angina in four, although the mean (SEM) increase in heart rate was modest (14 (2) beats/min) and mean coronary vascular resistance fell from 1.56 (0.21) to 1.16 (0.14) mm Hg min ml-1 (p less than 0.005). New or increasingly frequent or complex ventricular arrhythmias occurred in five patients. Placebo infusion had no effect on the variables measured. Supine bicycle exercise during infusion of the saline placebo was associated with a similar degree of ST segment depression (0.9 (0.2) mm) as adrenaline infusion at rest (0.9 (0.1) mm) but exercise performed during adrenaline infusion (10 patients) resulted in more pronounced ST segment depression (1.9 (0.3) mm) (p less than 0.005) than either intervention alone. Angina occurred in three of 11 patients during control exercise and in six of 10 during the combination of adrenaline infusion and exercise. Such potentially adverse consequences of low dose adrenaline infusion in patients with stable coronary heart disease are consistent with the suggestion that adrenal activation is detrimental during acute myocardial infarction, being both arrhythmogenic and proischaemic.