To study whether absolute quantitation of regional myocardial glucose utilization (rMGU) enhances detection of myocardial viability, 70 nondiabetic patients with prior myocardial infarction and angiographically confirmed coronary artery disease were studied with [18F]FDG PET after oral glucose loading. Forty-eight patients were also revascularized and underwent echocardiography after revascularization to detect wall motion recovery. The rMGU was calculated in eight myocardial segments in each patient and the results were compared to normalized (relative) [18F]FDG uptake values. In normal segments (n = 225), rMGU was 56 +/- 18 mumole/min/100 g (mean +/- s.d.) and relative [18F]FDG uptake 97% +/- 12%. The interindividual variation of rMGU in normal myocardium was greater than the intraindividual variation (s.d. 31% versus 11%). The respective values for relative [18F]FDG uptake were 9% and 10%. Both rMGU and [18F]FDG uptake were significantly reduced in segments with scarring observed visually during bypass surgery (29 +/- 19 mumole/min/100 g and 45% +/- 22%, n = 26). The rMGU and [18F]FDG uptake were higher in segments that recovered after revascularization (53 +/- 17 mumole/min/100/g and 110% +/- 21%, n = 27) than in those that did not (37 +/- 20 mumole/min/100 g and 65% +/- 24%, n = 63). However, due to larger variability of rMGU values, normalized [18F]FDG uptake was superior to rMGU in separating normal and scar segments as well as in predicting wall motion recovery. We conclude that rMGU variability is notable and is caused mainly by variations between patients. Interindividual variation is reduced by normalization, which results in more accurate assessment of myocardial viability. Thus, static imaging and semiquantitative analysis are sufficient for the clinical assessment of myocardial viability.