This review of the literature concerns the prognosis of patients suspected of having myocardial infarction subsequent to its exclusion as the diagnosis. Several investigations show a surprisingly bad prognosis for patients in this category, almost comparable to that of patients with a confirmed infarction. When the results of the different studies are pooled, however, there is a significant difference between those patients with true infarction, and those in whom infarction was excluded, in terms of overall mortality (12% and 7%; P less than 0.0001) and the development of subsequent non-fatal infarction (11% and 6%; P less than 0.05) when the results are analysed for a period of follow-up of one year. The difference was significant even when both fatal and non-fatal infarctions were taken into account over the one-year period of follow-up (13% and 8%; P less than 0.0001). The analysis shows that electrocardiographic ST-T changes are a risk factor for coronary events, but the results are conflicting for other possible risk factors. The selection of patients varies between the different studies, which probably contributes to the different results reported. Prospective studies with well defined groups of patients large enough to permit analysis of subgroupings will be needed to resolve the outstanding questions.