Silver and colleagues [1] have derived a clinical rule that identifies patients who have preserved left ventricular systolic function after myocardial infarction. The validity of this rule must be tested in other patient populations. We have now validated this prediction rule in an unselected group of patients in the United Kingdom.

Our study entailed a retrospective analysis of a prospective cohort study of 75 randomly selected patients admitted to our coronary care unit [2]. Study patients who survived the first 2 days after an acute myocardial infarction had transthoracic echocardiography, radionuclide ventriculography, or both to measure the left ventricular ejection fraction (LVEF). Only two patients' electrocardiograms were uninterpretable because of the presence of a left bundle-branch block pattern. Twenty-three patients (36%) had index Q-wave anterior infarctions, 18 (28%) had a history of infarction, and 12 (19%) had a history of congestive heart failure. On the basis of the prediction rule, 13 patients (24%) were expected to have an LVEF of 40% or more, but only 9 (16%) were found to have preserved left ventricular function. Therefore, the positive predictive value of the rule was only 69% in our patients.

We observed a much lower positive predictive value (69%) than the 90% reported by Silver and colleagues [1]. One possible explanation for the contrasting findings may be that our sample consisted of an unselected and heterogeneous group of patients, whereas the study of Silver and colleagues included some selective screening of the patient population by attending physicians. In the latter study, 65 patients (17%) in the database had no LVEF assessment, and this group differed somewhat from the 314 patients included in the study. Other evidence for differences between the two populations comes from the fact that the rule allowed 40% of patients to be excluded from unnecessary screening for left ventricular dysfunction [1] compared with only 25% of the unselected patients in our study.

Our findings question the universal applicability of the preserved LVEF prediction rule to all patients after myocardial infarction. Given the low positive predictive value observed in our patients, this rule would not identify one third of patients with left ventricular dysfunction. Because of this, patients with left ventricular dysfunction after myocardial infarction would not receive the mortality benefit associated with angiotensin-converting enzyme inhibitor therapy.