Background: To evaluate the accuracy of myocardial perfusion imaging (MPI) on a novel cadmium-zinc-thelluride camera in detecting significant coronary artery disease (CAD) in patients with atrial fibrillation (AF).
Methods and results: Seventy-four subjects with AF submitted to stress-rest MPI and coronary angiography were consecutively enrolled. One hundred and forty-eight patients in sinus rhythm, matched for age, sex, and type of stress-test protocol and with known coronary anatomy served as controls. The summed difference score, as measure of reversible myocardial ischaemia, was calculated. A coronary stenosis ≥70% was considered significant. The prevalence of significant CAD did not differ between AF patients and controls. At receiving operating characteristic analysis MPI showed relevant accuracy in unmasking the presence of significant CAD both in AF and in control patients (areas under the curve 0.71 vs. 0.80, P for difference: 0.212). However, after stratifying patients according to the stress protocol, a significant interaction between the presence of AF and MPI diagnostic power was evident. While in the case of a vasodilator stress-test MPI diagnostic accuracy remained high in both groups of patients (P for difference: 0.664), in those submitted to an exercise stress-test the diagnostic power of MPI was significantly lower in the presence of AF (P for difference: 0.039), because of a lower specificity. Interestingly, at multivariate analysis, a lower exercise duration (P = 0.017) was the major predictor of reduced MPI specificity.
Conclusion: The presence of AF impairs MPI accuracy on the detection significant CAD. This effect was only apparent in the case of an exercise stress-test, while disappeared in patients submitted to vasodilator stress.
Keywords: Atrial fibrillation; Cadmium-zinc-telluride; Coronary artery disease; Diagnostic accuracy; Myocardial perfusion imaging.
Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2015. For permissions please email: journals.permissions@oup.com.