The aim of this study was to assess whether data related to predischarge clinical examinations, resting radionuclide ventriculography and symptom-limited bicycle ergometry can predict the achievement of a normal exercise capacity after a rehabilitation program in patients with a recent myocardial infarction. The study population consists of 141 consecutive patients who completed a 3-month training program. Patients with heart failure and/or severe angina were excluded. The rehabilitation program included two training sessions weekly during the 3 months. Working capacity (WC) increased from 79 +/- 17% at hospital discharge to 105 +/- 21% of normal values after rehabilitation (P less than 0.001), by 33% on average. Ninety-five patients achieved a normal WC. Conventional predischarge clinical evaluation, resting left ventricular ejection fraction, exercise induced angina, or ST segment depression were not predictive of normal WC after rehabilitation. Predischarge WC was the single best predictor of a normal WC after rehabilitation compared to those with a persistently low WC (84 +/- 15% in patients with normal WC vs 69 +/- 14% in those with persistently low WC, P less than 0.001). Nevertheless, 49% of patients with a baseline WC of less than 80% achieved a normal WC after rehabilitation. No correlation was found between the change of WC after rehabilitation and predischarge WC or ejection fraction. Therefore, the selection of patients for cardiac rehabilitation after a myocardial infarction should be based primarily on clinical grounds. Exclusion based on exercise induced angina, ST segment depression or low resting ejection fraction at hospital discharge or at entry in the rehabilitation program is not justified.