The estimation of the hepatic protein synthesis precursor pool was investigated through the measurement of very low-density lipoprotein (VLDL) apolipoprotein (apo)B100 labeling in healthy volunteers. L-[1-13C]leucine and L-[5,5,5-2H3]leucine were administered intravenously and intragastrically, respectively. Subjects were continuously fed with isoenergetic meals providing either 16% protein or no protein. The labeling of leucine incorporated into VLDL apoB100 (leucine-apoB) was lower than plasma leucine or alpha-ketoisocaproate (KIC) enrichments with the intravenous tracer. By contrast, with the oral tracer, leucine-apoB enrichment was higher than either plasma free leucine or KIC labeling. The KIC and leucine-apoB enrichments relative to plasma leucine enrichment were not affected by protein intake. Albumin or fibrinogen synthesis rates were similar whatever the administration route of the tracer when leucine-apoB was used to indicate the precursor, which was not the case for plasma leucine or KIC. The present data suggest that leucine-apoB enrichment represents a reliable indicator of the hepatic precursor pool for protein synthesis. The effect of dietary protein on the calculated rates of albumin and fibrinogen synthesis is also reported in relation to the choice of the precursor.