Doubly Labeled Water Is a Validated and Verified Reference Standard in Nutrition Research^,

The doubly labeled water (DLW) method is currently the most relevant, albeit expensive, method for measuring free-living energy expenditure in animals and humans. The DLW method is based on the exponential disappearance from the body of the stable isotopes ²H and ¹⁸O after a bolus dose of water labeled with both isotopes. The ²H is lost as water and the ¹⁸O as both water and CO₂. After correction for isotopic fractionation, the excess disappearance rate of ¹⁸O relative to ²H is a measure of the CO₂ production rate (1). This rate can be converted to an estimate of total energy expenditure by using a known or estimated respiratory quotient and the classical principle of indirect calorimetry (2). The DLW technique, however, is not without limitations because it uses several assumptions such as a constant rate of CO₂ production and constant size of body water pool throughout the measurement period. In addition, not all researchers use the same methods to calculate the isotope pool spaces, the constant elimination rate, the fractionation factors, and the mode of CO₂ conversion to energy expenditure.

The DLW method was first proposed by Lifson et al. (3) in the middle of the last century, but its application to human studies was delayed for some 30 y (4) until improvements in analytical instrumentation made such investigations feasible. Since this pioneering work, the use of DLW has progressed, first to a research tool for the measurement of energy metabolism in free-living humans and later to the status of the “gold standard” to which most newly developed methodologies are compared. Unlike some other methods, DLW has matured relatively slowly and has been subject to frequent retrospective reviews by expert practitioners, which has led to substantial international agreement on matters of principle and practice (5). The benefit of such an approach is that it ensures compatibility between results from different laboratories so that direct comparisons can be made. Furthermore, it provides new investigators with a rigorous framework for experimental design and interpretations, thus ensuring that the results reported are of the highest quality (5).

The next significant step in establishing DLW as a reference standard in energy metabolism research is work presented in the current issue of The Journal of Nutrition by Wong et al. (6). As the authors pointed out, the long-term reproducibility of the DLW method has not been reported. Providing this clearly missing evidence could be especially important for longitudinal monitoring of changes in energy expenditure, energy intake, and body composition in human and perhaps in animal research. The experimental protocol of the study was based on the multicenter, parallel-group, randomized controlled clinical trial Comprehensive Assessment of the Long-term Effects of Reducing Intake of Energy (CALERIE), described in detail previously (7). The parent study created an ideal nest for evaluating the longitudinal reproducibility of the DLW method. The present study included 2 protocols, 1 for the study dose dilutions and 1 for test-retest reliability. The experimental procedure, which can be highly error prone (8), was conducted with a high degree of technical precision and in a double-masked fashion by using isotope ratio mass spectrometer methodology.

The first protocol showed that the theoretical fractional turnover rates for ²H and ¹⁸O and the difference between the 2 fractional turnover rates were reproducible to within 1% and 5%, respectively, over 4.4 y. The second protocol showed high reproducibility of primary DLW outcome variables such as the fractional turnover rates and isotope dilution spaces for ²H and ¹⁸O and total energy expenditure over 2.4 y. Thus, the study provided a unique set of scientific data and clearly demonstrated the feasibility of the DLW method in longitudinal studies to examine energy balance changes in humans. Although the study was conducted in adults, there is no compelling reason to believe that the results would be different in other populations, such as children and older adults. The detailed protocols presented by Wong et al. (6) could be used by other researchers in their laboratories to document the long-term reproducibility of their measurements to ensure the biologic significance of the long-term outcomes of interest. The data set could also be used to validate recently introduced new and less costly alternative methods for the high-precision water isotope abundance analyses (9, 10).

One of the unresolved issues in nutritional research over the past 100 y has been unobtrusive measurement of dynamic changes in total energy balance with high accuracy and precision and for periods longer than a few days or months. These measures are important because total energy is the only nutrient for which intake is regulated physiologically independent of weight and physical activity. It has been documented that small, albeit long-term, changes in the total energy balance could affect body weight or composition with a potential for well-known health effects such as obesity. Previous studies used labeling of total body water and measuring outflow rates to monitor changes in energy intake and expenditure and to provide estimates of energy balance–related changes in body composition (1). By using the data from the CALERIE study (7), Wong et al. (6) provided compelling evidence that the DLW method could be used in clinical trials to monitor adherence to dietary protocol, changes in energy intake, and changes in body composition. Although methodologic uncertainties are likely to persist in the near future primarily because of the described and other limitations of the DLW method, the study by Wong et al. provides us with tools to decrease these limitations.

In conclusion, the results of Wong et al. (6) and other recent advances in stable isotopes of water analysis (9, 10) solidify the validity of DLW as the reference standard for measurement of energy expenditure in the free-living and the applicability of DLW-derived applications in nutrition research.