Exercise Testing in Patients with Chest Pain and Normal Coronary Arteries: Improving Test Specificity by Use of a Simple Logistic Model

Abstract

Despite an exercise electrocardiogram (ECG) positive for ischemia by established criteria, many patients referred for coronary angiography to evaluate chest pain are found to have angiographically normal coronary arteries (NCA). Exercise ECG were analyzed from 27 patients with chest pain and angiographically NCA and 28 patients with chest pain and coronary artery disease (CAD) using univariate and multivariate logistic regression analysis. We derived the following logistic model for the logit probability of CAD: 3 + SEX × 4 – METs × 0.7 + STDV5 × 0.8, where SEX = 0 for female and SEX = 1 for male, METs = maximal estimated work load (metabolic equivalents) and STDV5 = horizontal or downsloping ST depression (mm) in V5. A logit probability ≧O identified CAD with a sensitivity of 79% and a specificity of 89%. The model correctly identified 28/36 (78%) patients with CAD, and 7/10 (70%) patients with NCA (correct diagnosis 76%; p < 0.02) in a separate random group of 46 unselected patients with positive exercise tests undergoing diagnostic coronary angiography.