Contents - Previous - Next


Modern scientists are very quick to identify, assimilate and apply new techniques. This trait often leads to astonishing rates of progress in which new methods can become routinely used throughout the world in just a few years. Monoclonal antibodies are a good example of one such success story. However, in other instances the over-enthusiastic adoption of a new method, together with the pressure to publish new results as soon as possible, has led to the premature application of techniques which have not undergone the usual rigours of a full methodological work-up. This document, and the work leading up to it, are an attempt to ensure that the Doubly-Labelled Water (DLW) method does not fall into the latter category.

The method has undergone a period of arrested development with respect to its applications in man. It was over 30 years between Professor Nathan Lifson's initial idea and Dr Dale Schoeller's realisation in the early 1980's that improvements in the precision of isotope ratio mass spectrometers had reduced the cost of applying the technique to such an extent that it had become financially viable for studies in humans. There followed a period in which several laboratories built on the experience gained from applying DLW in small animals, and concentrated on refining the technique for use in man. The extrapolation from one to the other was by no means straight-forward. As indicated by Kleiber's Law, humans have a relatively low energy expenditure per unit body mass compared to small animals. Together with their relatively profligate use of water, and hence high water turnovers, this tends to increase the potential errors in the method. On the other hand humans have some distinct advantages over other animals. They are easier to dose and can supply regular serial samples for analysis thus removing the need to use the traditional capture-release-recapture procedure.

The early papers on human applications of DLW all concentrated on theoretical and technical aspects, and on the results of new cross-validation studies. Many of the leading groups in the field participated in a methodological workshop during the XIII International Congress of Nutrition at Brighton in 1985. The first biological results were published in the same year, and it was announced that Nathan Lifson was to be awarded the Rank Prize in Nutrition for his pioneering work in developing DLW. In 1986 several groups from laboratories throughout the world participated in a symposium in Cambridge on 'Stable Isotopic Methods for Measuring Energy Expenditure'. This was an exciting time during which the results from many new studies were presented and when there was a general acceptance by most nutritionists that the method was probably working well. However, there remained a healthy scepticism which was encapsulated by Dr Elsie Widdowson in her now famous description of the method as "doubly-indirect calorimetry".

In addition to this residual scepticism the main proponents of the method had two other concerns. The first was that it might be difficult to make inter-laboratory comparisons of results if each laboratory used a slightly different variant of the technique. This problem has been particularly acute in another field where stable isotopes are used in nutritional studies, namely protein turnover. The diversity of tracers, end-products, dosing protocols and kinetic models employed, together with the sparsity of cross-validation studies, make it difficult to compare protein turnover results from different laboratories. The second concern was that new workers in the field may underestimate the complexities of DLW and publish invalid data which could potentially discredit the method.

In order to circumvent these problems it was decided to convene a workshop in which to seek a consensus view on the various technical aspects of applying the method. The International Dietary Energy Consultancy Group (IDECG) endorsed this proposal and it was financed by the Nestlé Foundation. The International Atomic Energy Agency (IAEA) supported the publication of this document as part of their Co-ordinated Research Programme on Applications of Stable Isotope Tracers in Human Nutrition Research.

The workshop was held in Clare College, Cambridge in September 1988. It was preceeded by the exchange of 32 DLW data sets from 6 of the participating laboratories. These were recalculated by a number of the participants using a total of 17 variants of the initial Lifson equation, and using different fractionation assumptions in order to quantify the maximum possible methodological discrepancies, and to identify the causes of such variance. The central participants prepared position documents which formed the basis of the subsequent discussions held over 4 days.

There was a remarkable concurrence of views concerning the causes, consequences and solutions to all of the major problems associated with DLW. Following the meeting these views were summarised by a number of the participants whose chapters were re-circulated for comments in order to ensure that the consensus had been fairly represented. The resultant recommendations contained in this report do not stipulate exact rules as to how the method should be applied, but instead provide a framework of guidelines which will ensure that published data is of a high quality. All current and potential users of the doubly-labelled water method are therefore strongly encouraged to make full use of these guidelines.

Readers who are completely unfamiliar with the doubly-labelled water method may find certain sections of this report rather impenetrable. They may find it easier to read Chapter 1 followed by Chapter 11, which contains some worked examples, before returning to Chapters 2 to 10 which examine the detailed arguments in support of the final recommendations.

Andrew Prentice, April 1990


I am most grateful to Gail Goldberg for the enormous amount of work which she contributed to the organisation of the IDECG Workshop and to the preparation of this document.

Contents - Previous - Next