Exercise echocardiography or exercise SPECT imaging? A meta-analysis of diagnostic test performance.

Abstract

Kirsten E. Fleischmann, MD, MPH; Maria G. M. Hunink, MD, PhD;Karen M. Kuntz, ScD; Pamela S. Douglas, MDContext.— Cardiac imaging has advanced rapidly, providing clinicians with sev-eralchoicesforevaluatingpatientswithsuspectedcoronaryarterydisease,butfewstudies compare modalities directly.Objectives.— To review the contemporary literature and to compare the diag-nostic performance of exercise echocardiography (ECHO) and exercise single-photon emission computed tomography (SPECT) imaging in the diagnosis ofcoronary artery disease.Data Sources.— Studies published between January 1990 and October 1997identifiedfromMEDLINEsearch;bibliographiesofreviewsandoriginalarticles;andsuggestions from experts in each area.Study Selection.— Articles were included if they discussed exercise ECHOand/or exercise SPECT imaging with thallous chloride TI 201 (thallium) or techne-tium Tc 99m sestamibi for detection and/or evaluation of coronary artery disease,if data on coronary angiography were presented as the reference test, and if theabsolute numbers of true-positive, false-negative, true-negative, and false-positiveobservations were available or derivable from the data presented. Studiesperformed exclusively in patients after myocardial infarction, after percutaneoustransluminal coronary angioplasty, after coronary artery bypass grafting, or with re-cent unstable coronary syndromes were excluded.Data Extraction.— Clinical variables, technical factors, and test performancewere independently extracted by 2 reviewers on a standardized spreadsheet. Dis-crepancies were resolved by consensus.Results.— Forty-four articles met inclusion criteria. In pooled data weighted bythe sample size of each study, exercise ECHO had a sensitivity of 85% (95% con-fidence interval [CI], 83%-87%) with a specificity of 77% (95% CI, 74%-80%). Ex-ercise SPECT yielded a similar sensitivity of 87% (95% CI, 86%-88%) but a lowerspecificity of 64% (95% CI, 60%-68%). In a summary receiver operating charac-teristic model comparing exercise ECHO performance to exercise SPECT, exer-cise ECHO was associated with significantly better discriminatory power (param-eter estimate, 1.18; 95% CI, 0.71-1.65), when adjusted for age, publication year,andasettingincludingknowncoronaryarterydiseaseforSPECTstudies.Inmodelscomparing the discriminatory abilities of exercise ECHO and exercise SPECT vsexercise testing without imaging, both ECHO and SPECT performed significantlybetter than exercise testing. The incremental improvement in performancewas greater for ECHO (3.43; 95% CI, 2.74-4.11) than for SPECT (1.49; 95% CI,0.91-2.08).Conclusions.— Exercise ECHO and exercise SPECT have similar sensitivitiesfor the detection of coronary artery disease, but exercise ECHO has better speci-ficityand,therefore,higheroveralldiscriminatorycapabilitiesasusedincontempo-rary practice.