Intracellular glucose concentration in skeletal muscle of awake rats was determined under conditions of hyperglycemic (10.2 +/- 0.6 mM) hyperinsulinemia (approximately 1,200 pM) and hyperglycemic (20.8 +/- 1.5 mM) hypoinsulinemia (< 12 pM) by use of 13C nuclear magnetic resonance (NMR) spectroscopy during a prime-constant infusion of [1-13C]glucose and [1-13C]mannitol with either insulin (10 mU.kg-1.min-1) or somatostatin (1.0 microgram.kg-1.min-1). Intracellular glucose was calculated as the difference between the concentrations of total tissue glucose (calculated from the in vivo 13C NMR spectrum with mannitol as an internal concentration standard) and extracellular glucose, corrected by the ratio of intra- and extracellular water space. Extracellular concentration was corrected for an interstitial fluid-to-plasma glucose concentration gradient of 0.83 +/- 0.07, determined by open-flow microperfusion. The mean ratio of intra- to extracellular glucose space, determined from the relative NMR signal intensities and concentrations of mannitol and total creatine, was 9.2 +/- 1.1 (hyperglycemic hyperinsulinemia, n = 10), and 9.0 +/- 1.7 (hyperglycemic hypoinsulinemia, n = 7). Mean muscle intracellular glucose concentration was < 0.07 mM under hyperglycemic-hyperinsulinemic conditions (n = 10) and 0.32 +/- 0.06 mM under hyperglycemic-hypoinsulinemic conditions (n = 7). This method is noninvasive and should prove useful for resolving the question of whether glucose transport or phosphorylation is responsible for the reduced rate of muscle glycogen synthesis observed in diabetic subjects.