Measurements of whole-body protein turnover in preterm infants have been made using different stable isotope methods. Large variation in results has been found, which could be due to different clinical conditions and/or the use of different tracers. We studied 14 appropriate for gestational age and nine small for gestational age orally fed preterm infants using [15N]glycine and [1-(13)C]leucine simultaneously, which allowed us to make a comparison of commonly used methods to calculate whole-body protein turnover. Whole-body protein turnover was calculated from 15N enrichment in urinary ammonia and urea after [15N]-glycine administration and from the 13C enrichment in expired CO2 after administration of [1-(13)C]leucine. Enrichment of alpha-ketoisocaproic acid after [1-(13)C]leucine constant infusion was measured as a direct parameter of whole-body protein turnover. Group means for whole-body protein turnover using [15N]glycine or [1-(13)C]leucine ranged from 10 to 14 g.kg-1.d-1, except when using the end product method that assumes a correlation between leucine oxidation and total nitrogen excretion. We found very low 15N enrichment of urinary urea in the majority of small for gestational age infants. These infants also had a lower nitrogen excretion in urine and oxidized less leucine. Nitrogen balance was higher in small for gestational age infants (416 +/- 25 mg.kg-1.d-1) compared with appropriate for gestational age infants (374 +/- 41 mg.kg-1.d-1, p = 0.003). [15N]Glycine does not seem to exchange its label with the body nitrogen pool to a significant degree and is therefore not always suitable as a carrier for 15N in protein turnover studies in premature infants.