Doubly Labelled Water analysis: preparation, memory correction, calibration and quality assurance for δ2H and δ18O measurements over four orders of magnitudes.

RATIONALE: The Doubly Labelled Water (DLW) method is an established way of determining the metabolic rate in humans and animals, with the advantage that the subjects need not be confined. The method, however, needs accurate determination of both the δ(2)H and the δ(18)O isotope values over a wide range of enrichments.
METHODS: In this paper we describe a number of crucial steps in the process of isotope determination in body fluids. These steps include micro-distillation, correction of the measurements for sample-to-sample memory and calibration of the isotope scales over many orders of magnitudes. In contrast to several published protocols and guidelines, we also take highly enriched samples into account, as they are required for studying the metabolic rate of birds and small mammals. For our isotope scale calibration, we made a set of gravimetrically prepared, double labelled waters with known isotope values. Our quality assurance includes a scheme for easy calculation of the error propagation, leading to a reliable estimate of the analytical error in the metabolic rate.
RESULTS: Our memory correction algorithm assumes the existence of three water "pools" that have different sizes and exchange rates with the injected samples. We show that the method can correct even huge memory signals, without the need for "true" values.
CONCLUSIONS: With the presented building blocks, we show how to assure a reliable and accurate isotope analysis for the DLW method, both for human and for animal applications. Although our measurements have been performed using isotope ratio mass spectrometry, most of the procedures are also useful for laser spectrometry.