Correlation of preoperative myocardial function, perfusion, and metabolism with postoperative function at rest and stress after bypass surgery in severe left ventricular dysfunction

Previous studies of dobutamine echocardiography (DE) and positron emission tomography (PET) showed similar accuracy for predicting improvement in resting wall motion after revascularization, although limited direct comparative data are available. We sought to compare the relative accuracy of detecting contractile reserve, ischemia, perfusion, and myocardial metabolism for predicting functional recovery after coronary bypass surgery in 94 consecutive patients (aged 63 ± 11 years) with chronic coronary disease and depressed left ventricular function (ejection fraction 28 ± 5%). PET imaging comprised rest and dipyridamole stress myocardial perfusion images, with fluorodeoxyglucose to define metabolism-perfusion mismatch. A standard dobutamine-atropine stress was used, with evaluation of low- and peak-dose echocardiographic responses. Regional function was assessed after 13 ± 16 weeks at rest in 68 patients who underwent isolated coronary bypass operation without evidence of perioperative infarction, and at rest and stress in a subgroup of 29 patients. Concordance between methods for evaluating abnormal segments (ischemic, viable, and scar) and accuracy of both tests for predicting improvement in regional function were identified. Concordance between PET and DE for identifiying viable or nonviable myocardium was 63% using a 16-segment model. For predicting improved resting function after surgery, the sensitivity of PET (84%) was superior to DE (69%, p <0.001), but DE was more specific (78% vs 37%, p <0.0001) and more accurate (75% vs 53%, p <0.001) in predicting recovery at rest. Analysis of postoperative recovery of segmental function during stress also showed the specificity of DE to exceed that of PET (89% vs 32%, p <0.001). The accuracy of DE was enhanced by evaluation of function during stress (86%, p <0.001), but this was not altered with PET (52%, p = NS). Thus, PET is more sensitive than DE in predicting functional recovery, but DE is more specific than PET. Evaluation of left ventricular functional recovery during stress may be preferable to assessment at rest.