Quantitative assessment of myocardial glucose uptake by the glucose tracer analog 2-deoxy-2-[18F]fluoro-D-glucose (FDG) depends on a correction factor (lumped constant [LC]), which may vary. We propose that this variability is caused by different affinities of FDG and glucose for membrane transport and phosphorylation and can be predicted from the time course of FDG retention. We therefore measured the LC under steady-state metabolic conditions and compared the results with values predicted from the tracer retention alone. METHODS We measured rates of myocardial glucose uptake by tracer ([2-3H]glucose) and tracer analog methods (FDG) in isolated working Sprague-Dawley rat hearts perfused with Krebs buffer and glucose, or glucose plus insulin or beta-hydroxybutyrate. In separate experiments, we established the theoretical upper and lower limits for the LC (Rt and Rp), which are determined by the relative rates of FDG and glucose membrane transport (Rt, 1.73 +/- 0.22) and the relative rates of FDG and glucose phosphorylation (Rp, 0.15 +/- 0.04). RESULTS The LC was decreased in the presence of insulin or beta-hydroxybutyrate or both (from 1.14 +/- 0.3 to 0.58 +/- 0.16 [insulin], to 0.75 +/- 0.17 [beta-hydroxybutyrate] or to 0.53 +/- 0.17 [both], P < 0.05). The time-activity curves of FDG retention reflected these changes. Combining the upper and lower limits for the LC with the ratio between unidirectional and steady-state FDG uptake rates allowed the prediction of individual LCs, which agreed well with the actually measured values (r = 0.96, P < 0.001). CONCLUSIONS The LC is not a constant but is a predictable quotient. As a result of the fixed relation between tracer and tracee for both membrane transport and phosphorylation, the quotient can be determined from the FDG time-activity curve and true rates of myocardial glucose uptake can be measured.