To evaluate the influence of acute beta-blockade on the ability of predischarge exercise test data to predict long-term prognosis in patients admitted for suspected acute myocardial infarction, patients randomized at hospital admission to intravenous metoprolol or placebo were studied. Among 190 patients discharged alive, total 4-year mortality was 20.5% (n = 39); (33 cardiac deaths, 6 non-cardiac deaths). Non-fatal infarction rate was 6.8% (n = 13). Multiple logistic regression analysis revealed that total mortality and non-fatal infarctions were independently predicted by (a) inability to perform predischarge stress testing (event-free survival for patients exercise tested 79.5% vs 56.9% for patients not eligible for testing; relative risk (RR) 1.40, 95% confidence interval (CI) 1.10-1.78; P = 0.01), and (b) low predischarge exercise capacity (RR 1.44, CI 1.08-1.93; P = 0.034). ST segment shift > or = 1 mm did not predict mortality or reinfarction. Administration of metoprolol in the acute phase did not influence the predictive value of these parametres. It is concluded that assessment of exercise capacity at early exercise testing yields independent information for later death and myocardial infarctions, and that beta-blockade with metoprolol does not influence the predictive value of early exercise testing.