In Part I of this series on energy expenditure, we looked at the various components of energy expenditure, as well as some of the more common field techniques for measuring total energy expenditure. In Part II, we will turn our attention to the measurement of free living energy expenditure using doubly labeled water.
Measurement of Free Living Energy Expenditure Using Doubly Labeled Water
The doubly labeled water technique was first described by Lifson et al (1955) and is based on the fact that carbon dioxide production can be estimated from observing the differences in the elimination rates between metabolic hydrogen and oxygen. The method is based on work showing that the oxygen contained in body water is in isotopic equilibrium with respiratory oxygen contained in carbon dioxide via the carbonic anhydrase reaction which liberates the oxygen from water (Lifson et al, 1949).
With this finding, and the fact that the hydrogen component of body water exits in the body almost exclusively as water (Hevesy & Hofer, 1934), they predicted they could indirectly measure carbon dioxide production in an animal model by labeling the body water pool with a stable isotope of hydrogen and of oxygen and observing the elimination rates of the isotopes from the body over time. Stable isotopes are naturally occurring forms of a compound, differing only in molecular weight, and are believed to exhibit the same chemical properties as the more commonly known form of the element.
The doubly labeled water technique uses deuterium oxide or 2H as the hydrogen tracer and H218O as the oxygen tracer to label the body water pool. This method was first carried out in the mouse model and was validated with simultaneous measurement of respiratory gas analysis (Lifson et al., 1955). Although it appeared that this technique would be successful in humans, it was not economically feasible at the time due to the lack of sensitivity in isotopic ratio mass spectrometry, the instrumentation that is used to measure the isotope concentrations in a fluid sample.
To validate the technique in humans would have required using isotope doses costing several thousand dollars at the time (Schoeller & Van Santen, 1982). Instead, the method was validated in several animal models over the next twenty years, which increased knowledge about the possible error inherent in the method (Schoeller, 1988). It was not until 1975 that the method was proposed for the possible measurement of total energy expenditure in humans, due primarily to the advancement of isotope ratio mass spectrometry sensitivity which decreased the amounts of isotopes required and, hence, the cost of the technique (Lifson et al.,1975).
As described above, the doubly labeled water method is based on the theory that hydrogen of body water exits in the body only as water while body water oxygen leaves the body as both water and as expired carbon dioxide. It has also been described by Lifson et al. (1955) that it is possible to determine energy turnover in the animal model from a given volume of carbon dioxide gas. Therefore, the difference between these two compounds, O2 and CO2, can be related to total carbon dioxide production during the metabolic turnover period of these two isotopes. By labeling both the hydrogen pool and the oxygen pool of the body water with stable isotopes it is possible to follow the isotopic decay rates by sampling the body water and measuring the isotopic concentration.
The doubly labeled water technique was first validated in humans in 1982 by Schoeller and Van Santan (1982). This protocol used a dose of 0.13 gram per kilogram body weight of 18O and 0.007 gram per kilogram body weight of deuterium in each of the four subjects. Carbon dioxide production was measured over a period of fourteen days by calculating the difference in the elimination rates of the two isotopes. The measured carbon dioxide production rate was used in conjunction with the food quotient (FQ = protein*0.81 + fat*0.71 + carbohydrate*1.0) calculated from the ingested foods of the diet to estimate oxygen consumption (O2 consumption = CO2 production / FQ). Daily oxygen consumption and carbon dioxide production rates were than used to calculate total daily energy expenditure using the equation of Weir (1949) (energy kcal/d = 0.4554(1.1*CO2 l/d + 3.9*O2 l/d).
The doubly labeled water method was validated with a controlled energy intake / energy balance measurement. Changes in the subjects’ body composition was measured on day one and on day 14 using the total body water calculated from the oxygen 18 dilution space and were used as the marker for the change in energy stores. Upon comparison of the two methods a 2.1 % difference between the two estimates of energy expenditure was reported.
Doubly Labeled Water: Methods
The basic protocol outline for the use of doubly labeled water for the measurement of total energy expenditure is as follows.
Since the stable isotopes that are being used are naturally occurring it is necessary to determine the background enrichment prior to the introduction of the loading dose.
The stable isotopes are introduced to the body as a mixed, oral dose of 2H2O and H218O at an approximate amount of 0.2 grams 2H per kilogram total body water and 0.08 grams 18O per kilogram total body water. This will produce initial enrichments of 500 – 900 parts per million for 2H and 90 – 170 parts per million for 18O which represents a level of 600 times the analytical error of biological variation of the isotope enrichment. At this level, the cost of the isotopes are kept to a minimum and the study can be carried out between 0.5 – 3 biological half-lives of the isotopes (the biological half-life for water is approximately 7 days but is dependent upon actual body water turnover rates).
Sampling of the body water pool is most commonly made through collection of urine, however it is also possible to use both saliva and plasma as alternative sources. The sampling protocol chosen to follow the monoexponential curve of the isotopes over time is the issue of some debate in the literature. The decision that has to be made is whether to implement the two-point or multi-point sample collection of the body water. It is clear that at least one sample should be taken on the first day of the study and one on the final day as well.
However, since human activity is episodic in nature there is likely variation in the flux rate of carbon dioxide production over time that would not be detected using the two-point method. In an evaluation of the two different techniques, Cole and Coward (1992) found that the multi-point method of sampling is one-third more precise than the two-point method (Cole & Coward, 1992).
Welle (1990) found that by obtaining replicate samples at point one (study day 1) and at point two (study end point) the impact of technical errors of collecting single, multi-point samples over time are greatly reduced; therefore, making the two-point method of sample collection similar to the multi-point method (Speakman, 1992). It has also been observed recently that if variation in a person’s daily expenditure is expected that the two point method is a more accurate method of body water sampling since you do not force a linear line through non-linear data (Speakman, 1992). These findings are also in agreement with the basic concept of the method which draws its allure from being able to non-invasively measure total energy expenditure in free-living individuals.
B) Isotopic Analysis of Deuterium And 18O:
The measurement of the isotopic enrichments in the body water samples from the subject are measured using isotopic ratio mass spectrometry. Isotope ratio mass spectrometry is used to determine the ratio of light (i.e. 1H or 16O) to heavy (i.e. 2H or 18O) isotopes in gaseous samples. In the analysis of the isotopes found in the water samples it is necessary to liberate either the hydrogen or the oxygen from the samples prior to analysis. This is accomplished by creating hydrogen gas for the analysis of deuterium and carbon dioxide gas for the analysis of oxygen 18.
The common method utilized for the liberation of hydrogen gas from the water samples is off-line zinc reduction (Coleman et al., 1982; Kendall & Copelan, 1985). This assay involves introducing 3 ul body water samples to specially designed reaction tubes (Wong & Schoeller, 1990). These tubes are composed of a combination of pyrex and quartz which enables the tubes to withstand the high temperature required for the reaction. The reactant used in this reduction is 100 milligrams of purified zinc which is pre-weighed into the tube. The water sample is then introduced to the tube while a positive flow of dry nitrogen gas is blown through the tubes. The samples are then frozen inside of the tubes and the tubes are than vacuumed to 1E -3 mbar to remove any nitrogen gas or atmospheric moisture that may be present in the tube. The tubes are then sealed leaving only the water sample and zinc inside the tube which is then heated to 500 degrees Celsius for thirty minutes producing zinc oxide and hydrogen gas. Hydrogen gas is generated through the following reaction:
The hydrogen gas is then introduced to the mass spectrometer for isotopic analysis
For the analysis of 18O abundance in the physiological fluids samples an equilibration procedure is preformed with standard carbon dioxide gas. First, body water samples are introduced in to sealed vaccutainers followed by purified carbon dioxide gas. The tubes are then shaken overnight at room temperature to allow the oxygen in the fluid and gas come to an isotopic equilibrium (Cohn and Urey, 1938; Epstein and Mayeda, 1953) as follows.
The carbon dioxide gas is then introduced into a mass spectrometer for analysis of isotopic enrichment against a standard reference gas.
Assumptions of The Doubly Labeled Water Method
When Lifson et al. (1955) put forth the theory of using doubly labeled water to measure total energy expenditure it was not without basic assumptions that encompass steady-state, isotopic tracer methodologies. The original assumptions have been redefined to improve the accuracy of the method as more data have become available and the sources for potential error are better understood.
– A basic assumption of all metabolic tracer methodologies is that the size of the pool in which the isotopes turnover (i.e. total body water) remain constant over the period of the study. However, under certain circumstances where the body water pool does change over the course of a study (eg newborns) (Jones et al., 1988) it has been shown that the change in water volume has to be greater than 15% to produce a significant error in the calculation of the carbon dioxide production rate. In situations where the body water pool may change, as in exercising persons (Westerterp et al., 1986), the volume of body water can be calculated at the end of the study of the study by administering a booster dose of one of the isotopes.
– The second assumption of the method is that there is a constant carbon dioxide and water flux rate, from the body, over the period of the measurement. However, human activity is episodic in nature and therefore, the flux rate of carbon dioxide production cannot be constant unless the subject is confined to very sedentary conditions (Schoeller, 1988). This assumption addresses the timing of body water sampling that is implemented during the measurement period. As mentioned above the two points sampling method will measure an average carbon dioxide production rate and the multi-point method will produce possible day to day differences in expenditure.
– The assumption put forth by Lifson that has been highly debated in the literature is the notion that the isotopes exchange with any non-aqueous materials in the body. Until recently it was believed that the dilution spaces for deuterium and 18O were 4% and 1% larger than total body water respectively (Schoeller; 1988). Therefore, correction factors of 1.04 for deuterium and 1.01 for oxygen 18 (1.04/1.03, providing a fixed ratio for the correction of isotopic exchange of 1.03) were used to account for this source of potential error in calculating the isotopic dilution spaces. These dilution spaces were calculated using very small numbers of people and were also reduced to only two decimal places for convenience. By using fixed ratios in the calculation of final carbon dioxide production individual variation in isotopic dilution space is not taken into account and therefore, is a source of potential error.
These dilution space correction factors were recently scrutinized by Speakman et al (1993) who showed that by using the fixed isotope dilution ratios it was possible to grossly over and under estimate energy expenditure (if a ratio of 1.08 actually existed and 1.03 was used, the calculated energy expenditure would be calculated 23% higher then it actually was). Speakman et al. (1993) reviewed the relationship of the 2H to 18O dilution space in 161 published values and concluded that individual dilution spaces should be calculated from the equation of (Coward et al., 1988). An updated fixed ratio of 1.0427 is also provided as an alternative which was calculated from the 161 published dilution spaces (Speakman et al., 1993). It is clear that dilution spaces of the isotopes can be a major source of error in the final calculation of carbon dioxide production by either predicting the turnover space as larger or smaller then it actually is.
– One of the most basic assumptions that is made using stable isotope methodologies is that the isotopes do not have different chemical properties (i.e. have the same specific activity) and therefore, exit the body at the same rate. However, it is known that stable isotopes of water have differing physical properties, where heavier forms of a given isotopic compound will be lost (evaporation, atmospheric moisture) disproportionately compared to the lighter form (Wong et al., 1988).
Isotopic fractionation in the doubly labeled water method is experienced whenever body surface water that is non-sweat in nature (i.e. membranes, skin) comes into contact with other exchangeable molecules (i.e. water vapor present in inspired air and body water present at the lung surface). Isotopic fractionation is most likely to occur when there are extreme environmental temperature conditions either very hot or cold and when a person may be in a state of abnormal metabolism, either hyper- or hypometabolic.
The mechanisms of isotopic fractionation can experienced via three reactions when using the doubly labeled water method: 1) evaporative loss of 2H2O 2) evaporative loss of H218O and 3) equilibration of 18O between water and atmospheric carbon dioxide. Although Lifson et al. (1975) worked extensively with the mechanics of isotopic fractionation in developing the method, the majority of his work was conducted at 25 degrees Celsius which is significantly lower then the temperature of the body (Lifson & McClintock, 1966). Recently, Schoeller (1986) and Wong (1988) revised the fractionation of deuterium and 18O and correction factors for isotopic fractionation have been further modified to adjust for this difference in temperature. Schoeller (1986) noted that fractionation of water vapor may be most affected at the surface of the nasal cavities with temperature change but, that this change would have to be at a magnitude of 5 Celsius to produce a 1% error. Wong (1988) concluded that little fractionation occurs during the secretion of saliva and urine and that when fractionation factors of the isotopes are taken into account that total body water measurements obtained from saliva, urine, respiratory water vapor and carbon dioxide were in close agreement to that measured with plasma. 18O dilution spaces using fractionation correcting factors in 20 subjects using carbon dioxide breath and plasma values were 0.37 + 0.68 kg.
– The chance of potential error also exists when there is the passage of water or carbon dioxide into the body via the lungs or skin. It has been shown that water can easily penetrate the skin and be absorbed directly into the body (Pinson and Langham, 1957). Since this type of contamination involves proportional hydrogen and oxygen, the error that is introduced is canceled. However, in a situation in which a human is exposed to higher than normal levels of environmental carbon dioxide a disproportionate increase of oxygen is observed (Schoeller, 1988). Under normal circumstances these levels of exposure are low but, in the situation of a heavy smoker or a person who is confined to a poorly ventilated area an error of approximately 4% can be experienced in the calculation of total carbon dioxide production (Schoeller, 1988).
Part II of this series on energy expenditure has covered the methods and assumptions behind the measurement of free living energy expenditure using doubly labeled water. In Part III we will turn our attention to the validation of this method’s efficacy for use in humans.