Quantification of myocardial glucose uptake (MGU) by 18F-fluoro-2-deoxyglucose (FDG) using PET may be inaccurate, because the correction factor that relates myocardial FDG uptake to MGU, the lumped constant (LC), is not a true constant. Recent studies have shown that analysis of FDG time-activity curves allows determination of individual LCs and that variable LCs yield accurate determination of MGU. We compared the magnitude of the LC in different regions of the heart in patients with ischemic cardiomyopathy. METHODS: Twenty patients with ischemic cardiomyopathy and an average ejection fraction of 33% underwent dynamic 13N-ammonia and FDG PET. We determined myocardial perfusion and MGU in 177 regions classified as control (71 regions), mismatch (50 regions) and match (56 regions), according to findings on PET and echocardiography. Regional MGU was calculated with both regional LCs and a fixed LC of 0.67. RESULTS: All results were expressed as mean +/- SD. Myocardial perfusion was highest in control regions (0.52+/-0.18 mL/g/min), reduced in mismatch regions (0.43+/-0.19 mL/g/min; P < 0.05 versus control) and severely reduced in match regions (0.28+/-0.17 mL/g/min; P < 0.001 versus control and mismatch). Regional LCs ranged from 0.45 to 1.30 and differed between patients (P < 0.001). Regional LCs were similar in regions diagnosed as control (0.78+/-0.23), mismatch (0.80+/-0.24) and match (0.72+/-0.21). MGU (micromol/g/min) calculated by regional LCs was similar in control (0.52+/-0.16) and mismatch (0.49+/-0.19) regions and decreased in match regions (0.31+/-0.12, P < 0.001). The agreement between MGU calculated with variable and fixed LCs was poor. CONCLUSION: The LC used in the calculation of MGU was not affected by regional differences in the metabolic state of the myocardium. However, the LC varied substantially between patients in control, mismatch and match regions. These findings indicate that quantitative measurements of MGU using a fixed LC must be interpreted with caution.