Patients with unstable coronary artery disease have a serious but variable prognosis. An early and specific prediction of risk is essential for stratification of treatment. Serum was obtained at a median of 9.5 hours from symptom onset in 7800 patients with unstable coronary artery disease included in the GUSTO-IV trial for analyses of creatinine, troponin-T, C-reactive protein (CRP) and N-terminal pro brain natriuretic peptide (NT-proBNP). Quartiles of troponin-T were related to an increased mortality and to an increased incidence of myocardial infarction. Increasing quartiles of C-reactive protein were also related to an increased mortality but there was no relation to the incidence of myocardial infarction. On multivariate analysis, troponin-T was the strongest marker for prediction of myocardial infarction, but reduced creatinine clearance and ST-depression at admission were also significant predictors. Prediction of subsequent mortality was possible with several risk indicators. Elevation of NT-proBNP was the strongest predictor of short and long-term mortality with a continuous increase in one-year mortality in relation to the levels. Also reduced creatinine clearance, elevation of CRP, troponin-T, ST-depression and clinical factors indicating a history of cardiovascular disease provided independent prognostic information on long-term mortality. A multimarker strategy with creatinine clearance, troponin, CRP and NT-proBNP together with ischemic ECG changes and clinical background characteristics provides the best prognostic information for choice of treatment in patients with unstable coronary artery disease.