Regional CMRglc was measured in seven healthy volunteers with positron emission tomography using [1-11C]D-glucose. Regional CBF was measured using [11C]fluoromethane. The arteriovenous differences of unlabeled glucose and oxygen together with 11C metabolites were also measured. In addition to the loss of [11C]CO2, a loss of acidic 11C metabolites was also detected. A three-compartment model was applied to the tracer data in the time interval 0-24 min. After correction for the loss of 11C metabolites, the tracer method gave an average CMRglc of 26.4 +/- 1.9 (SD) mumol/100 g/min, close to the value obtained with the Fick principle. After correction for the loss of [11C]CO2 only, the tracer method gave 23.6 +/- 2.1 mumol/100 g/min, compatible with (1/6) CMRO2, obtained with the Fick principle. These results and the time course of the loss of acidic 11C metabolites are consistent with the presence of nonoxidative metabolism of glucose that causes an early loss of mainly [11C]lactate after a bolus injection of the tracer. This implies that [1-11C]D-glucose measures the rate of glucose oxidation rather than the total CMRglc. The experiments using [1-11C]D-glucose were compared to five analogous experiments using [U-11C]D-glucose together with [15O]H2O as a flow tracer. After correction for the loss of [11C]CO2, the two glucose tracers gave similar global values of CMRglc and other parameters associated with glucose utilization, but with labeling in the carbon-1 position, the loss of [11C]CO2 was substantially delayed and the contrast between gray and white matter was improved.(ABSTRACT TRUNCATED AT 250 WORDS)