Although unstable angina is an extremely common and often initial manifestation of coronary disease, few controlled studies of its treatment have been carried out. This relative dearth of information is due to the methodological problems raised by the evaluation of unstable angina. Unlike the definition of myocardial infarction, that of unstable angina--i.e. of a population of coronary patients who from time to time are at a high risk of myocardial infarction or death--is neither unequivocal nor easy to standardize. It follows that the patient population ultimately selected for controlled trials is but a small part of all unstable angina patients. The representativeness of patients involved in therapeutic trials is probably approximate. Moreover, the current criteria for assessment of effectiveness are either the clinical signs of angina in the short term or the incidence of myocardial infarction and changes in survival curves in the mid- and long terms. A more precise definition of criteria of inclusion, leading to an homogeneous population, and the development of a simple and reliable method for detecting and quantifying myocardial ischaemia, both being used as intermediate criteria of assessment, would undoubtedly improve the quality of therapeutic trials in unstable angina and, mostly, their applicability to daily therapeutic practice.