Background  

It has been suggested that in patients with a normal resting electrocardiogram (ECG), exercise radionuclide myocardial perfusion imaging (MPI) does not add significant incremental diagnostic information to exercise ECG.

Methods and Results  

Of 840 consecutive patients referred for physical exercise MPI, 313 (37%) had normal resting ECGs. There were 189 men and 124 women with a mean age of 54±11.9 years. Exercise MPI was performed with either Tl-201 or ^99mTc-labeled sestamibi. Overall concordance between exercise ECG result and MPI result was 79% (kappa agreement=.54). One hundred eighty-four patients had normal exercise ECG; 181 (98.4%) of these also had normal exercise MPI. In 271 patients with low (≤20%) to intermediate (21% to 70%) pre-exercise likelihood of coronary artery disease (CAD), concordance between normal exercise ECG and normal MPI was 100%. In the high likelihood (≥71%) group 3 (15%) patients with normal exercise ECG had abnormal exercise MPI. Of 129 patients with abnormal exercise ECG, 67 (52%) patients also had abnormal MPI, but 62 (48%) patients had normal MPI. Complete follow-up was obtained in 89% of patients at 9 months. Only 1 hard cardiac event occurred: nonfatal myocardial infarction. Twenty-one (8%) patients had subsequent coronary revascularization or admission with unstable angina. Although both abnormal stress ECG and abnormal exercise MPI were significantly (P<.0001) associated with hard and “soft” events, the association of abnormal exercise MPI was significantly stronger. Because all patients with a low and intermediate likelihood of CAD who had normal exercise ECG also had normal exercise MPI, we propose a stepwise diagnostic testing strategy whereby exercise MPI imaging is performed only in patients with a low to intermediate likelihood of CAD when the exercise ECG is abnormal. When the exercise ECG is performed first, and exercise MPI is performed only when the exercise ECG is abnormal, substantial (38%) cost savings can be achieved. In patients with a high likelihood of CAD, the exercise ECG may be falsely negative, and exercise MPI is preferred.

Conclusion  

In patients with normal resting ECGs a stepwise diagnostic strategy can reduce costs of exercise testing without compromising diagnostic yield when pretest likelihood of coronary artery disease is taken into consideration.