Cardiac PET imaging using 13N-ammonia and 18FDG was performed on 17 normal volunteers in the glucose-loaded and fasted states. PET images (glucose-loaded and fasted) were subjected to qualitative visual analysis using the descriptors: image quality, FDG blood-pool activity, ammonia lung field activity, and clinical interpretability. Glucose-loaded studies had better image quality, diminished FDG blood-pool activity, and were more often clinically interpretable compared to fasted studies. Dietary state did not affect 13N-ammonia study parameters. Using a semi-quantitative region of interest method, 18FDG myocardial activity was homogeneous and significantly affected by dietary state (2.48 +/- 1.0 fold increase in myocardial-to-blood pool ratio in glucose-loaded versus fasted states). Serum glucose, insulin, and free-fatty acid concentrations responded physiologically to glucose loading and fasting but showed no correlation with myocardial 18FDG uptake. PET imaging using 13N-ammonia and 18FDG appears optimal in the glucose-loaded state.