The uptake of F-18 deoxyglucose into dysfunction segments after myocardial infarction identifies metabolically active (FDG+) or inactive (FDG-) myocardium. Although patients with FDG+ segments have been found to be at risk for adverse events, the prognostic significance of viable myocardium in relation to other influences on postinfarction prognosis, including revascularization, remain ill defined. The purpose of this study was to investigate the relative prognostic significance of FDG+ tissue and to establish whether myocardial revascularization in patients with viable tissue attenuates the risk of adverse outcome. One hundred thirty-seven patients with left ventricular dysfunction and resting perfusion defects after myocardial infarction underwent positron emission tomography with both dipyridamole stress Rb-82 perfusion imaging and FDG imaging. After the exclusion of 4 patients proceeding to transplantation, 2 with uninterpretable scans and 2 lost to follow-up, 129 patients were followed clinically for 17 +/- 9 months. Four groups were defined: patients with FDG+ dysfunctional myocardium who were revascularized (n = 49) or treated medically (n = 21) and those with FDG- segments who were revascularized (n = 19) or treated medically (n = 40). The groups of patients with FDG+ or FDG- findings, with and without revascularization, did not differ with respect to known determinants of postinfarction prognosis: age, left ventricular ejection fraction, or the prevalence of multivessel disease. Nonfatal ischemic events occurred in 48% of medically treated FDG+ patients compared with 8% of revascularized patients with FDG+ tissue (P < .001) and 5% of patients with FDG- myocardium (P < .001). Thirteen patients died from cardiac causes; 11 (85%) had a left ventricular ejection fraction of < 30%, and these patients were evenly distributed between FDG+ and FDG- groups. Using Cox's proportional hazards model, only the presence of FDG+ myocardium (odds ratio, 12.9; P < .001) and the absence of revascularization (odds ratio, 5.8; P = .002) independently predicted ischemic events, while only age (P = .02) and ejection fraction (P < .001) but not the presence of viable myocardium were predictive of death. Residual viable myocardium after myocardial infarction may act as an unstable substrate for further events unless it is revascularized. Despite this association, age and left ventricular dysfunction remained the strongest predictors of cardiac death after myocardial infarction in these patients with a spectrum of left ventricular dysfunction.