Precision of the doubly labelled water method using the two-point calculation.

The intrasubject precision of the doubly labelled water method was investigated as a function of increasing length of the metabolic period and decreasing initial isotope enrichment. This analysis was performed in seven subjects following a single dose of isotope and three subjects who received multiple doses of isotope. The intrasubject precision was 5-6 per cent for periods of up to 14 d. This is only slightly larger than that predicted from our previously published model in which we assumed that analytical error was limiting precision. Intrasubject precision worsened with increasing length of the metabolic period and with decreasing initial isotope enrichment, but the loss of precision was two to ten times greater than that predicted by the model. The use of the intercept method to determine the isotope dilution spaces did not improve precision. We speculate that one cause of the failure of the model is that it does not consider unmeasured changes in natural isotopic abundances associated with dietary intake. Results demonstrate the importance of choosing a dose that provides an initial isotope enrichment that is 500-600 times the analytical error, and a metabolic period that is between one and three biological half-lives of the isotopic tracers.