OBJECTIVE To determine the sensitivity and specificity of a new myoglobin assay for acute myocardial infarction (AMI), considering both the total amount of serum myoglobin and its percentage change over 2 hours. METHODS A prospective, observational test performance study for the recognition of AMI was done using serial myoglobin assays of 42 admitted chest pain patients at a large, urban teaching hospital ED. Myoglobin testing was performed at presentation (time 0) and at 1 and 2 hours after arrival. A myoglobin level > 100 micrograms/L (ng/mL) or a change > or = 50% from baseline (increase or decrease) any time during the 2-hour period was considered positive. Patients and their physicians were blinded to the myoglobin results. The managing clinician's final diagnosis of the presenting event was used as the diagnostic criterion standard. RESULTS The sensitivity of the myoglobin technique for detection of AMI in the first hours in the ED was 13/14 (93%; 95% CI: 66-100%). The 1 patient who had a false-negative test had evidence of AMI on the ECG and an initially abnormal creatine kinase-MB (CK-MB) assay. The specificity was 22/28 (79%; 59-92%). However, of the 6 patients who had "false-positive" myoglobin tests, all had serious illness: significant cardiac disease (n = 4), in-hospital death (n = 1), or deep venous thrombosis (n = 1). CONCLUSIONS Myoglobin level determinations are sensitive tests to detect AMI during the first 2 hours of a patient's stay in the ED and may complement current clinical tools.