Exercise stress testing is often used as the initial non-invasive diagnostic test in symptomatic patients with suspected obstructive coronary artery disease. Positive standard ECG criteria are quite specific for obstructive coronary artery disease, but there may be a substantial number of false negative tests, including patients with severe coronary artery disease. Also, exercise stress tests frequently yield equivocal results. Instead of detecting the functional consequences, electron-beam computed tomography visualized atherosclerotic plaque disease directly, but its relationship to functional testing has not been clearly delineated. It was the aim of the current study to examine electron-beam computed tomography for the identification of obstructive coronary artery disease in patients with a normal, abnormal, or equivocal exercise stress test. Symptomatic patients referred for coronary angiography were prospectively included in a consecutive manner if they had no prior diagnosis of coronary artery disease and an unremarkable resting ECG. All patients underwent both exercise stress test and electron-beam computed tomography on the day before coronary angiography. Standard protocols and ECG criteria to diagnose inducible ischaemia were used for the exercise stress test. The electron-beam computed tomography-derived total calcium score was computed according to standard Agatston criteria. Of the 323 patients (mean age, 55+/-11 years; 77% male), 179 (55%) had obstructive coronary artery disease, defined angiographically as luminal diameter narrowing >/=50%. A normal exercise stress test was documented in 105 patients (32.5%), an abnormal exercise stress test ('diagnostic for ischaemia') in 113 (35%), and an equivocal exercise stress test ('inadequate exercise or non-diagnostic ECG-changes') in 105 (32.5%). Multivariate analysis indicated that exercise stress test and electron-beam computed tomography yielded independent information for predicting obstructive coronary artery disease. Sensitivity, specificity and overall accuracy of the exercise stress test were 71%, 75% and 73%, respectively, if equivocal tests were not included, and 50%, 84% and 65% if they were included. Irrespective of the cutpoint regarded as 'positive', the overall accuracy of the electron-beam computed tomography-derived calcium score remained approximately 80% in patients with a normal, abnormal or equivocal exercise stress test. In patients with an equivocal and - to a lesser degree - with a normal exercise stress test, electron-beam computed tomography was able to significantly improve classification regarding obstructive coronary artery disease. Electron-beam computed tomography added no incremental predictive value in patients with an abnormal exercise stress test. In patients who are judged to have an intermediate post-test probability of disease after exercise stress test, electron-beam computed tomography scanning may be a meaningful strategy for further stratification regarding the likelihood of obstructive coronary artery disease. Copyright 2000 The European Society of Cardiology.
