The aim of this study was to assess the social fate (work resumption) and long-term (four years) survival in 141 patients who completed a cardiac rehabilitation programme after a recent myocardial infarction. Out of the 100 patients who had been working up to recently before the myocardial infarction, 58 resumed their work. Of the predischarge evaluation (clinical data, resting radionuclide ventriculography, bicycle ergometry and 24-h ambulatory ECG monitoring) and bicycle ergometry after the rehabilitation, the only significant predictor of work resumption was a better exercise tolerance at discharge (P less than 0.02). Work was resumed by 68% of white-collar workers and by 52% of blue-collar workers. The four-year cardiac mortality in patients who completed the rehabilitation was 8.5% (N = 12). Four patients died during the first year. Clinical, ventriculographic and ergometric variables collected at hospital discharge, which were related to left ventricular dysfunction, were predictive of survival, while ventricular arrhythmias and markers of myocardial ischaemia were less predictive. The exercise testing performed after the rehabilitation programme was not useful for risk assessment. It is concluded that markers of left-ventricular dysfunction are predictive of a poor outcome; however, due to the low risk of patients who were referred to our rehabilitation unit and completed the rehabilitation programme, it seems reasonable for return to work to be based primarily on clinical information, exercise tolerance, and on psychological and social grounds. An additional extensive cardiological evaluation should be individually tailored for patients with specific symptoms.