Prediction of late mortality after myocardial infarction from variables measured at different times during hospitalization

Abstract

The long-term prognostic importance of sets of variables from different times in the hospital course after acute myocardial infarction was examined in 818 patients discharged from the hospital. Cardiac mortality during the first year after discharge was 11.1 %. For the end point death within 1 year after admission, discriminant function analysis identified 5 important factors from the history and the first 24 hours of hospitalization: maximal level of blood urea nitrogen, previous myocardial infarction, age, displaced left ventricular apex (abnormal apex) on physical examination, and sinus bradycardia (negative correlation). When data from the entire hospitalization were included, extension of infarction and maximal heart rate were also selected. When variables obtained at discharge were included, only the presence of S 3 gallop and abnormal apex were selected. In subgroups of patients, neither the left ventricular ejection fraction nor the presence of complex ventricular arrhythmias during a 24-hour ambulatory monitoring were independent predictors. Correct prediction was similar for each analysis, with 55 to 60% of the deaths and 79 to 81 % of survivors correctly identified. The high-risk group consisted of 25 % of the patients with 28 to 30 % predictive value for death in the first year. In conclusion, outcome up to 1 year after acute myocardial infarction can be predicted early after admission. Addition of more information later during the hospitalization and at discharge did not improve correct prediction and may be redundant for prognostic evaluation.