A meta-analytic comparison of echocardiographic stressors

Background: The relative performance of alternative stressors for stress echocardiography for the diagnosis of coronary artery disease (CAD) is not well established. Methods: All studies published between 1981 to December 2001 who met inclusion criteria were included in this analysis. We performed a summary receiver operator characteristic (SROC) analysis and calculated weighted mean of the likelihood ratio and sensitivity/specificity. A covariate analysis using meta-regression methods was also performed. Results: Forty-four studies presented data on Exercise, 11 on Adenosine, 80 on Dobutamine, 40 on Dipyridamole, 16 on transatrial pacing transesophageal echocardiography (Tap-TEE), and 7 on transatrial pacing transthorasic echocardiography (Tap-TTE). SROC analysis showed that the following order of most discriminatory to least: Tap-TEE, Exercise, Dipyridamole, Dobutamine and Adenosine. Weighted means sensitivity/specificity were Exercise: 82.6/84.4%, Adenosine: 68.4/80.9%, Dobutamine: 79.6/85.1%, Dipyridamole: 71.0/92.2%, Tap-TTE: 90.7/86.1%, and Tap-TEE: 86.2/91.3%. Covariate analysis showed that the discriminatory power of Exercise decreased with increasing mean age. Conclusions: Tap-TEE is a very accurate test for both ruling in and ruling out CAD although its invasiveness may limit its clinical acceptability. Exercise is a well-balanced satisfactory test for both ruling in and ruling out but performance might be lower for the elderly. Dobutamine offers a reasonable compromise for Exercise. Dipyridamole might be good for ruling in but not for ruling out CAD. The incapability in ruling-out CAD was a major problem in clinical application of the stress. Adenosine was the least useful stressor in diagnosing CAD.