Estimating left ventricular ejection fraction after myocardial infarction by various clinical parameters

Abstract

The clinical predictors of left ventricular (LV) ejection fraction (EF) were determined in 760 survivors of an acute myocardial infarction (AMI). LVEF was dichotomized at ≤0.40 (n = 269) and >0.40 (n = 491). Logistic regression showed that 4 of 20 preselected, clinically meaningful variables were strong and independent (p < 0.001) predictors of LVEF ≤0.40. Independent variables together with their odds ratio (odds of having a LVEF ≤0.40 with factor present to odds with factor absent) in order of decreasing importance were: anterior AMI (4.7), congestion on chest x-ray (2.9), previous AMI (2.3) and creatine kinase >1,000 U (2.1). There was a stepwise decrease in LVEF and an increase in the proportion of patients with a low LVEF for each additional clinical variable. A general estimate of LVEF was made by simply considering the total number of clinical factors present. The presence of 0 or 1 clinical variable (n = 466) predicted a high LVEF (>0.40) with an accuracy of 80%. Two or more variables (n = 294) predicted a low LVEF (≤0.40) with an accuracy of 60%. The overall predictive accuracy was 72%. Clinically significant major misclassifications were rare (<10%). Readily obtainable clinical variables provide the clinician with a useful bedside method of estimating LVEF after AMI.