Effects on anginal symptoms of sudden withdrawal of large doses of propranolol or placebo were evaluated in 20 patients in a double-blind crossover efficacy trial requiring sudden cessation of the agent. With propranolol, 160 to 320 mg per day for six and 12 weeks, no patients had increased angina or nitroglycerin use, and there were no hospitalizations or deaths. However, within two weeks of discontinuance of propranolol, untoward ischemic events developed in 10 patients. Six had serious withdrawal complications: intermediate coronary syndrome in three, and ventricular tachycardia, fatal myocardial infarction, and sudden death in one each. In four patients discontinuance of placebo increased anginal symptoms; in the remaining 10, ischemic symptoms were not provoked. The rebound phenomenon was related to degree of pre-propranolol angina and relief of pain by the agent. Thus, chronically administered propranolol should be gradually reduced, and activity restricted during its withdrawal.