The relationship between 31 variables and survival after acute myocardial infarction was evaluated in 432 patients discharged from our Coronary Care Unit from 1975 to 1984. The patients were followed for 1 to 105 months and either univariate and multivariate analysis were performed. For end-point death the significant variables (p less than 0.05) selected by the univariate analysis were: age, diabetes, smoke, heart rate at recovery, supraventricular arrhythmias, cardiac failure and complex ventricular arrhythmias either during recovery, either after discharge and finally spontaneous angina after hospital discharge. Meanwhile, for the end-point cardiac death age, smoke and supraventricular arrhythmias were not yet significant while arterial pressure at recovery and effort angina after hospital discharge were. Multivariate analysis identified cardiac failure during recovery, diabetes, complex ventricular arrhythmias before and spontaneous angina after discharge as independent variables contributing to total mortality: effort angina was a further significant one relatively to cardiac death. Thus, our study points out the importance of multivariate survival analysis when evaluating the relationship between survival after discharge for the effect of other prognostic factors. Moreover, providing identification of high risk cohorts permits appropriate interventions designed to lessen risk.