The [14C]deoxyglucose ([14C]DG) method depends upon quantitative trapping of metabolites in brain at the site of phosphorylation, and in the usual procedure it is assumed that all the label in plasma is in free DG. Our previous finding of labeled nonacidic derivatives of DG in plasma raised the possibility that some metabolites of DG might not be fully retained in body tissues and therefore cause overestimation of the integrated specific activity of the precursor pool determined from assay of label in plasma and/or underestimation of the true size of the metabolite fraction in brain. In the present study, metabolism of DG in rat tissues by secondary pathways was examined and found to be more extensive than previously recognized. When 14C-labeled compounds in ethanol extracts of either plasma or brain were separated by anion exchange HPLC, eight fractions were obtained. 14C-labeled metabolites in plasma were detected after a 35-min lag and gradually increased in amount with time after an intravenous pulse. In brain, deoxyglucose-6-phosphate was further metabolized, mainly to deoxyglucose-1-phosphate and deoxyglucose-1,6-phosphate. These are acid-labile compounds and accounted for approximately 20% of the 14C in the metabolite pool in brain. The rate constants for net loss of 14C from the metabolite pool between 45 and 180 min after a pulse were similar (0.4-0.5%/min) in vivo and in intact postmortem brain. The rate constant for loss of deoxyglucose-6-phosphate (DG-6-P) in vivo (approximately 0.7%/min) was, however, about twice that for postmortem brain, suggesting that a significant fraction of the DG-6-P lost in vivo is due to its further metabolism by energy-dependent reactions. 14C-labeled metabolites of [14C]DG in plasma and brain do not interfere with determination of local rates of glucose utilization in brain in normal, conscious rats by the autoradiographic method if the prescribed procedures and a 45-min experimental period are used.