The optimum ST elevation model correctly classified 83.1% of subjects (55.8% sensitivity, 94.0% specificity) for acute myocardial infarction; adding QRST features only marginally improved this.
Observational (n=1,190)
Do different ST elevation criteria and additional QRST features improve the diagnosis of acute myocardial infarction in patients presenting with chest pain?
Different definitions of ST elevation significantly impact the sensitivity and specificity for diagnosing acute myocardial infarction, while adding other QRST features provides only marginal improvement.
AIMS: The optimum definition of ST elevation for diagnosis of acute myocardial infarction, with respect to both the minimum height and the minimum numbers of leads, is unknown. Furthermore, only 50% of patients with acute myocardial infarction present with ST elevation. We thus quantified the sensitivity and specificity of different ST elevation criteria for diagnosis of acute myocardial infarction, and determined whether models incorporating multiple QRST features in addition to ST elevation, could improve detection of acute myocardial infarction. METHODS AND RESULTS: The study population comprised 1190 subjects: 1041 consecutive patients presenting with chest pain (335 with acute myocardial infarction) and 149 controls without chest pain. Subjects were randomly divided into a training set (587) and a validation set (603). ECG prediction models for acute myocardial infarction incorporating different ST elevation criteria and/or additional QRST features (Q waves, ST depression, T wave inversion, bundle branch block, axes deviations, and left ventricular hypertrophy) were developed in training set patients using forward stepwise multiple logistic regression. Models were then prospectively tested in the validation set patients. The optimum ST elevation model (based on > or =1 mm ST elevation in > or = 1 inferior/lateral leads, or > or =2 mm ST elevation in > or =1 anteroseptal leads) correctly classified 83.1% of subjects (55.8% sensitivity, 94. 0% specificity). The choice of ST elevation definition had marked influence on the sensitivity (45.4-68.6%) and specificity (81.2-98. 1%) for diagnosis of acute myocardial infarction. The addition of multiple QRST variables only marginally improved overall classification but did result in high specificity (92.6-96.1%). CONCLUSION: Different definitions of 'significant' ST elevation led to marked variations in sensitivity and specificity for diagnosis of acute myocardial infarction. Multiple QRST features in addition to ST elevation only marginally improved overall classification.
Ian Menown (Tue,) conducted a observational in Acute myocardial infarction (n=1,190). ECG prediction models incorporating different ST elevation criteria and additional QRST features was evaluated on Correct classification of acute myocardial infarction. The optimum ST elevation model correctly classified 83.1% of subjects (55.8% sensitivity, 94.0% specificity) for acute myocardial infarction; adding QRST features only marginally improved this.