The prognostic value of information derived from clinical characteristics and exercise treadmill tests performed before discharge and repeated at three months was evaluated in 205 consecutive patients followed for five years. Recurrent myocardial infarction, unstable angina and mortality were tabulated. Survival was analyzed by the Kaplan-Meier life-table method and the Cox regression model. The major difference between the predischarge and three month intervals was the failure of exercise-induced ST depression to predict mortality from the predischarge test. However, it was predictive of mortality at three months when 76% survived five years with a positive ischemic response compared to 94% with a negative response (P less than 0.0005). In contrast, resting ST depression of at least 1 mm was associated with a very poor five year survival rate of 58.3% and 50% when assessed at both predischarge and three months (P less than 0.0005 and P less than 0.004, respectively). Selected univariately at the predischarge interval, the following characteristics were ranked in descending order of predictive power for five year mortality by discriminant analysis: history of previous infarction; exercise capacity; and ST depression on resting ECG greater than 1 mm. At three months, the same characteristics were selected. However, recurrent infarction and unstable angina were not predictable at either interval by any clinical or treadmill variable. Characteristics tending to reflect poor exercise capacity are stronger predictors of five year outcome than exercise-induced ischemia. While predischarge exercise testing for ST segment response failed to predict survival, this variable showed improved predictive power with repeat testing at three months.(ABSTRACT TRUNCATED AT 250 WORDS)