BACKGROUND: The majority of patients hospitalized due to suspected acute myocardial infarction (AMI) will eventually not develop infarction. Information about the long-term prognosis in this patient population is limited. AIM: To describe the mortality during 10 years of follow-up in patients hospitalized due to an initially strong suspicion of AMI, but in whom the diagnosis of AMI could not be confirmed. PATIENTS: All patients participating in an early intervention trial with metoprolol in suspected AMI, but in whom the diagnosis was not confirmed. Patients were included during 1976-1981. RESULTS: In all 1395 patients were included in the study, of whom 586 did not fulfil the criteria for confirmed AMI. The overall mortality during 10 years of follow-up in this population was 26%. In a multivariate analysis considering age, sex, history of cardiovascular diseases, initial heart rate and various complications during the hospital stay, including congestive heart failure, severe ventricular arrhythmias, tachycardia, hypotension, high degree AV-block and severe chest pain, the following appeared as independent predictors of death: previous infarction (P < 0.001), age (P < 0.001), history of diabetes mellitus (P < 0.001) history of smoking (P < 0.05), history of hypertension (P < 0.05), male sex (P < 0.05), and the initial heart rate (P < 0.05). CONCLUSION: Among patients in whom AMI was not confirmed the major risk indicators for death during 10 years of follow-up were: a history of cardiovascular diseases and smoking, age, male sex and high heart rate on admission to hospital.