Our previous studies have suggested that there is minimal fecal excretion of carbohydrate-derived energy in premature infants fed lactose or glucose-polymer as the source of carbohydrate. In these studies, carbohydrate energy excretion was equated with the non-nitrogenous, nonfat energy. Although these studies provided consistent results, this methodological approach cannot distinguish actual excretion of dietary carbohydrate from carbohydrate in the form of large molecular compounds derived from infant or bacterial cells (which in fact comprise approximately 90% of the measured value). Therefore, in this study we compared the absorption of carbohydrate energy to the fractional absorption of 13C derived from [D-1-13C]-lactose in 10 premature infants 30-32 weeks gestational age fed either a commercial premature infant formula containing a mixture of carbohydrates (50% lactose: 50% glucose polymer) or the same formula in which lactose was the sole carbohydrate. The two methods provided comparable estimates of carbohydrate absorption (96 and 94%, respectively), although there were significant discrepancies in two infants. These studies provide evidence of external validity for the two methods.