Abstract

OBJECTIVES--To determine the incidence of the earliest electrocardiographic changes in patients with suspected myocardial infarction and their sensitivity and specificity for predicting the final diagnosis of acute myocardial infarction. DESIGN--Retrospective study of paired electrocardiograms recorded at home and on admission to hospital. SETTING--29 rural practices in Grampian and teaching hospitals in Aberdeen. PATIENTS--137 patients participating in the early anistreplase trial in the Grampian region, who received placebo at home and for whom paired electrocardiograms were available. MAIN OUTCOME MEASURES--Classified electrocardiographic abnormalities and diagnosis at discharge. RESULTS--Electrocardiograms were recorded immediately before injection of placebo at home and anistreplase in hospital at median times of 110 and 240 minutes after the onset of symptoms. Definite or probable myocardial infarction was later confirmed in 93 (68%) patients. Of these, 66 (71%) had the same findings on both electrocardiograms of either ST elevation, bundle branch block, or a non-specific abnormality, while 27 (29%) showed a major change of classification between home and hospital recordings; 21 (23%) had ST elevation or bundle branch block on only one of the paired recordings. Although ST elevation was the commonest abnormality in the 93 patients with myocardial infarction, in only 51 was it recorded at home (sensitivity 55%) and in 49 on admission (sensitivity 53%). Of 57 patients with ST elevation at home, six did not have infarction (specificity 86%), while of 51 with ST elevation on admission, two did not have infarction (specificity 95%). CONCLUSIONS--Elevation of the ST segment is a transient electrocardiographic abnormality that has high specificity but low sensitivity for predicting the diagnosis of acute myocardial infarction; it is an unsatisfactory precondition for giving thrombolytic treatment to patients with suspected acute myocardial infarction.