To enable assessment of myocardial viability, myocardial glucose utilization has commonly been stimulated by oral glucose loading. To compare the effects of glucose loading and insulin and glucose infusion (insulin clamp) on PET fluorodeoxyglucose ([18F]FDG) myocardial scan image quality and regional myocardial glucose utilization rate (rMGU), eight patients with angiographically documented coronary artery disease and previous myocardial Q-wave infarction were studied twice, once during insulin clamp and once 1 hr after oral glucose loading. The rMGU rates were derived by graphic Patlak analysis in 33 normal, 10 scar and 6 "hot spot" myocardial segments. Infusion of insulin and glucose gave stable plasma-glucose and serum-insulin levels during imaging. In contrast, glucose loading caused marked changes in plasma-glucose and insulin concentrations. The image quality was clearly superior and the fractional utilization rates of [18F]FDG were twice as high during insulin clamp than after glucose loading (p less than 0.0001). Due to the higher plasma-glucose levels after glucose loading, the calculated rMGU in normal, scar and hot spot myocardial segments was comparable between the two protocols. The insulin clamp technique makes it possible to adjust and maintain a metabolic steady state during the PET study. It does not alter [18F]FDG uptake patterns in different myocardial areas when compared to the standard glucose loading protocol, but this technique results in superior image quality and permits the use of smaller [18F] FDG patient doses.