RESUMO
OBJECTIVE: To determine the test performance of 24-lead variance cardiography (VC), an ECG technique that measures QRS morphologic variability, for ED evaluation of chest pain associated with coronary artery disease (CAD). METHODS: A prospective, single-blind study of VC was performed in a community teaching hospital ED. All chest pain patients (> 30 years of age) who, after initial emergency physician evaluation, were believed to have pain of potential cardiac etiology and were admitted to the hospital were eligible. Exclusion criteria included obvious noncardiac etiology for discomfort, bundle-branch block, atrial fibrillation, and incomplete subsequent cardiac evaluation. After initial evaluation and stabilization, VC was obtained. The numerical output of VC was a CAD index (CADI). Serum myoglobin and creatine kinase (CK)-MB levels were obtained at the time of presentation and after one, two, and six hours. Hospital records were reviewed to determine final diagnosis and in-hospital evaluation results. RESULTS: Fifty-two of 75 eligible patients had complete data. Final diagnoses were as follows: 27/52 (52%), noncardiac; 13/52 (25%), acute myocardial infarction (AMI); and 12/52 (23%), unstable angina due to CAD. Twenty-three percent (12/52) of the patients had CADIs < 75. Eleven of these were found to have noncardiac origins for their chest pain. The twelfth patient had a 12-lead ECG revealing AMI and had been given thrombolytic therapy with subsequent reperfusion prior to VC. Using a CADI < 75 as the cutoff for a negative study, VC alone had a negative predictive value of 92%, a sensitivity of 96%, a positive predictive value of 60%, and a specificity of 41%. CONCLUSION: A CADI < 75, in addition to clinical impression and initial ECG, may identify chest pain patients who do not have significant CAD. Further prospective assessment of VC is warranted.