Twelve patients with documented coronary arteriosclerosis and severe stable angina pectoris were treated with the beta blocker pindolol, 5 mg four times a day, utilizing a double-blind crossover protocol. Following 4 weeks of baseline observation with no active treatment, pindolol and placebo were given for 4 weeks each. End points evaluated were episodes of angina pectoris/week, number of nitroglycerin tablets used per week, time on treadmill test until onset of angina pectoris, double product of heart rate and blood pressure at onset of angina pectoris, and amount of ST depression during the treadmill exercise test. Episodes of angina pectoris and nitroglycerin consumption were reduced by 18% on placebo and 32% on pindolol (not significant). All of the improvement occurred in the third and fourth weeks of pindolol treatment. Less difference between pindolol and placebo was noted when the placebo period came after the pindolol period, suggesting a carry-over effect of pindolol. With pindolol, treatment exercise tolerance was increased 13% (33 seconds) over baseline levels but only 2% over the levels achieved with placebo treatment (not significant). ST depression with exercise was 6% less when patients were on pindolol than when they were on placebo (not significant). There was a marked decrease in myocardial oxygen demand as measured by the double product of blood pressure and pulse during exercise (23% reduction when on pindolol and no change when on placebo, p less than 0.01). This study shows that there was an important placebo effect when treatment of angina pectoris was evaluated and that pindolol significantly reduced myocardial oxygen demand but evidence of ischemia was not significantly reduced. Possible mechanisms to explain the disparity between reduction in estimated myocardial oxygen demand (double product) and objective improvement in ischemia include coronary spasm and altered regional flow resulting from beta blockade. Alternative explanations may be the relatively small fixed dose of pindolol and the small number of patients studied.