From a symptom-limited bicycle exercise test 3 wk after acute myocardial infarction is 205 patients the prognostic significance for the 1-yr prognosis of the following variables was examined by a multivariate analysis (Cox's model): reasons for stopping, duration of work, maximal heart rate, maximal product heart rate x systolic blood pressure, maximal ST-deviation, time to maximal ST-deviation and ventricular ectopic beats (type and frequency). The reason for stopping was fatigue in 59% and angina pectoris in 15%. 48% exercised up to 6 min and 14% beyond 12 min with median duration of 7 min. The maximal heart rate was over 140/min in 30%. ST-deviations were found in 77%; in most patients below 3 mm. 44% had ventricular ectopic beats during the exercise. The only significant variable was the duration of of work with prognostic index for the 1-yr prognosis SE: SE = 1.21-0.16 x (duration of work). A probability of survival of over 0.95 after 1 yr required a duration of work over 13 min. The observed deaths were in good accordance with the expected with a little overestimation. 75% died from a definite cardial cause. There is a good predictive value for the 1-yr prognosis of the duration of work from an exercise test 3 wk after AMI.