Exercise testing helped in diagnosing postinfarction patients in the prethrombolytic era. Over the past decade acute myocardial infarction treatment has changed because of new thrombolytic therapies and consequently, the value of exercise testing is under debate. The GISSI-2 database allowed us to reevaluate the prognostic role of exercise testing in thrombolysed patients. The exercise test was performed in 6296 patients, on average 28 days after randomisation. The test was not performed in 3923 patients because of contraindications. The test was judged positive for residual ischaemia in 26% of the patients, negative in 38%, and non-diagnostic in 36%. Among the patients with a positive stress test result, 33% had symptoms, whereas 67% had silent myocardial ischaemia. The mortality rate was 7.1% among patients who did not have an exercise test and 1.7% [correction of 7.1%] for those with a positive test, 0.9% for those who had a negative test, and 1.3% for those who did not have a diagnostic test. In the adjusted analysis, symptomatic induced ischaemia, submaximal positive result, low work capacity, and abnormal systolic blood pressure were independent predictors of 6-month mortality (relative risks [RR] 2.54, 95% CI 1.27-5.08, 2.28, 1.17-4.45, 2.05, 1.23-3.42, and 1.86, 1.05-3.31, respectively). However, when these factors were tested simultaneously, only symptomatic induced ischaemia and low work capacity were confirmed as independent predictors of mortality (RR Cox 2.07, 95% CI 1.02-4.23 and 1.78, 1.06-2.99, respectively). Patients with a normal exercise response have an excellent medium-term prognosis and do not need further investigation. However, more evaluation should be devoted to the patients who cannot undergo exercise testing, because the potential to influence outcome appears to be much greater.