COOPER, JAMIE A.1; NGUYEN, DAVID D.2; RUBY, BRENT C.3; SCHOELLER, DALE A.2
With the growing popularity of long-duration, physically demanding events, as well as extreme energy expenditures (EE) in some occupational or military professions, a need arises to determine the precise energy requirements of training and performing in individuals involved in these events and professions. The World Health Organization (WHO) has recommended that energy requirements be based on EE to maintain body weight with neither excessive gain nor loss. That, however, is a difficult task under conditions of high EE because few portable methods of measuring EE have been developed and validated. During the last few decades, however, the doubly labeled water method (DLW) for measuring EE in free-living subjects has been optimized for human use. This method is based on the differential fluxes of oxygen and deuterium in body water. The oxygen flux is a measure of both water and carbon dioxide flux, whereas the deuterium flux is a measure of water flux alone. The difference between the two fluxes is therefore a measure of carbon dioxide production (24). In addition, the relative ease for sample collection (via urine, blood, or saliva) allows this method to be used in a variety of locations without hindering or obstructing the performance of the athlete. The accuracy and simple procedures involved with DLW provide the versatility to be used in extreme conditions and thus has been applied to an increasing number of situations involving extreme EE.
Total EE (TEE) is influenced by body size, and therefore, it is difficult to compare EEs between species of different body sizes. One means of adjusting for body size is to use physical activity level (PAL) measures. The PAL is defined as TEE divided by the basal metabolic rate (BMR) (28). Westerterp and Bryant (27) took initial steps in determining upper limits of PALs that could be achieved from their studies on migrating birds. They demonstrated that migrating birds were able to reach and sustain an EE that was five times their BMR (27). Other studies in different animal species have also reported sustained energy budgets greater than or equal to five times their BMR but not exceeding seven times their BMR (9,10). These high expenditures were found in three species of marsupials (10,19) and one more species of birds (2). Whereas most animal species sustain PALs below 5.0, the question has been raised whether or not humans can meet or exceed sustained EEs that are equivalent to what these few species can attain. It is well known that humans can achieve PALs much greater than 5.0 during short, intense exercise; however, whether or not high PALs can be sustained for a minimum of five consecutive days remains to be determined.
Using DLW as an accurate measure of EE, our aim was to determine human EE endurance limits by compiling the growing number of DLW studies performed during extreme EE. Using this growing database, coaches, athletes, and researchers may be able to more accurately estimate caloric intake requirements needed to sustain elite athletes and military and occupational personnel. We hypothesized that trained humans would be able to sustain a PAL of 5.0 but that higher levels would be associated with weight loss, suggesting a catabolic state. We also hypothesized that athletes exhibiting a lower body weight would be able to sustain higher PALs than “heavier” athletes because simply carrying less body mass may allow these lighter athletes to sustain higher levels of activity for longer periods of time.
To identify studies measuring EE of athletes and military personnel, we used electronic databases PubMed, Medscape, and Web of Science (renamed Web of Knowledge). Text words used for initial searches were “Doubly Labeled Water,” “Doubly Labelled Water,” “Extreme Physical Activity,” “Marathon + Expenditure,” and “Environment + Expenditure.” In addition to these searches, we searched the referenced material within the articles identified from the database searches. Full articles for each citation were obtained and reviewed.
Doubly labeled water methods
To compare or contrast EE from different studies, it is important to ensure that standard DLW methodology was used. DLW can be administered via intravenous, intramuscular, intraperitoneal, subcutaneous, or ingestion methods (24). Accurate animal measurements have been possible through the first four aforementioned methods, but simple ingestion has been used for human testing because it is less invasive and safer. An initial urine sample must be taken to determine natural levels of isotopes found within the body and to provide a baseline for comparison (30). Here, we report the “typical” DLW protocol. For reported subjects, in every study, DLW was given via ingestion where 2H218O (deuterium oxide) was “chased” with 100 mL of regular water. The amount of DLW administered is determined by the subject’s body mass. Dosages were 2 g of 10 AP H218O per kilogram and 0.1 g of 99.9 AP (atom percent) 2H2O per kilogram. Once ingested, all subjects were required to fast for 4 h before training, competing, or eating. Saliva or urine samples were taken immediately after this 4-h period. Urine samples were used to provide a reading for the removal of both isotopes. Once samples were taken, they would be transported to a designated laboratory for mass spectrometry analysis (30). For detailed information on the DLW methodology for each study, please see Table 1.
To determine the PAL or physical activity index, we used the following equation: (PAL or physical activity index) = TEE/BMR (28). If BMR was not provided, we calculated it by using methods established through the 1985 Food and Agriculture Organization (FAO)/WHO/United Nations University (UNU) collaboration (21). Where BMR was calculated, we used the subjects’ beginning weight to determine the PAL from reported TEE for each athlete. In addition, the established PAL was determined by comparing the highest level of EE achieved over a significant period (≥5 d). The PAL was then used to determine whether or not the athletes exceeded five times his or her BMR.
Study demographics and PALs
Twenty studies were included in our review, and a description of the subjects and physical activities can be found in Table 2. We recognized three categories for these studies: athletic training and job professionals, athletic events, and military training. Each study was placed into one of these categories and is shown this way in subsequent tables. The PALs from all studies ranged from 1.8 (3) to 6.9 (25) (Table 3). Studies categorized as “athletic training and job professionals” had PAL ranges of 1.9 (16) to 4.0 (23). The PAL of 1.9 was found in eight collegiate swimmers during a 10-d tapering period (training of low intensities to remove high levels of physical stress on the body to prepare for a competition). The highest PAL (4.0 ± 0.5) considered in the athletic training category was established by the Swedish national cross-country team during a 7-d period of regular seasonal training (23). Studies falling under the military category showed a PAL range of 1.8 (3) to 4.0 (14). PALs recorded at 1.8 were supplied during a 28-d field exercise by special operations soldiers who were given rations of only 8.3 MJ·d−1 (1980 kcal·d−1) of food. The uppermost PAL achieved by military personnel was during high-altitude cold-weather training during a span of 11 d. However, we were unable to measure the catabolic response to higher PAL levels found in military training because of insufficient data. Finally, PAL levels found in athletic events showed a range between 2.7 (18) and 6.9 (25). The lowest PAL of 2.7 was in a transatlantic yacht race that spanned a period of 13 d. The highest PAL was reported by a man trekking 2300 km across the Antarctic. His highest PAL of 6.9 was obtained during a 10-d period that was dominated by ascending to higher elevations.
Body mass and body composition
Most of the studies included in this analysis reported before and after body weights; however, only 8 of the 19 studies reported before and after body composition measures (Tables 2 and 3). Those studies reporting PALs around 2.5 and lower showed little to no change in body composition (8,15,20). Similarly, those studies reporting before and after body weight measures found little to no change in body weight at lower PALs. Even for PALs of 2.8 ± 0.9, Hill and Davies (12) reported a change in body mass from only 60.9 ± 2.3 to 59.7 ± 1.1 (Table 3). This was similar to Sjödin et al. (23), who showed no significant changes in body mass at PALs of 3.4 ± 0.3 and 4.0 ± 0.5. Conversely, PALs of 5.0 and higher were accompanied by larger changes in body mass and composition. Large changes in body fat percentage were recorded (from 17.4% to 1.9% and from 16.3% to 2.5%) (Table 2) (25) for peak PALs of 5.56 and 6.94 in two individuals completing a 95-d Antarctic Trek. Not surprisingly, this was accompanied by decreases in body mass of 18.9 and 16 kg, respectively. In another study of cyclists during the Tour de France, a calculated PAL of 5.2 led to a change in body fat percentage from 11.6% to 9.3% (Table 2) (29). Some of the other studies did report some losses or gains of body mass, but the amount was no more than 2% body fat or a 5-kg change in body composition weight. We also calculated changes in body weight per week of activity to normalize the results for studies of different length (Table 3). Changes in body weight ranged from −2.3 to +0.60 kg·wk−1.
We tested for correlations between both PAL and TEE versus initial body weight and body weight change. The results of these correlations are shown in Figures 1 and 2. There was a significant positive correlation between TEE and body weight (r = 0.43, P < 0.05). This positive correlation remained significant when we looked at TEE versus body mass index rather than body weight (r = 0.35, P < 0.05). Conversely, there was no significant correlation between PAL and body weight (r = 0.23, not statistically significant (NS)) (Fig. 1). When we correlated PAL and TEE levels with body weight change (expressed as change in body weight per week), there was a nonsignificant negative correlation between PAL and weight change (r = −0.28, NS) as well as TEE and weight change (r = −0.31, NS) (Fig. 2). Finally, we correlated the activity EE or non-BMR EE (TEE − BMR) versus body weight and found a trend for a positive correlation between the two variables (r = 0.38, P = 0.07) (Fig. 3).
Two studies of humans sustaining PALs in excess of 5.0 were found (25,29). These PALs were sustained for 3 wk in one study and 3 months in the other study. Furthermore, these high PALs were achieved during athletic events (Antarctic Trek and Tour de France race). Military operations showed sustained PALs ranging from 1.8 to 4.0, athletic events showed sustained PALs ranging from 4.0 to 6.9, and athletic training showed sustained PALs ranging from 1.9 to 4.0.
The study by Stroud et al. (25) provided us with two males maintaining extraordinarily high peak TEEs of 48.7 MJ·d−1 (11,631 kcal·d−1) and 44.6 MJ·d−1 (10,652 kcal·d−1) during 10 d of a 95-d Arctic Trek (days 20–30). After calculating BMR from their age and height using the FAO/WHO/UNU equation, we found that their peak PALs reached 6.9 and 5.6, respectively (25). These high TEEs, however, were not found for the entire 95 d. Average TEE for the first 50 d of the trek for each subject were 35.5 and 29.1 MJ·d−1, whereas days 51–96 showed average TEEs of 23.1 and 18.8 MJ·d−1. There is some question as to the precision of the TEE measurements during days 20–30. Therefore, we also calculated days 1–10 for that study. We found TEEs of 39.8 and 26.8 MJ·d−1, which correspond to PALs of 5.0 and 3.8. Therefore, at least one of the subjects did achieve a PAL of 5 during another period of the 95-d race.
Falling just below these extraordinarily high PALs were elite cyclists competing in the Tour de France studied by Westerterp et al. (29). These four cyclists had a mean TEE of 35.7 MJ·d−1 (8527 kcal·d−1) for the third period (last 7 d) of the race. This translated into a peak PAL of 5.2 (29). Interestingly, the average daily PALs for the cyclists ranged from 3.4 to 5.2 times their BMR, which suggests that their daily PALs were similar but slightly below the peak PAL of 5.2 found in one cyclist. On the basis of these aforementioned studies, it is possible for humans to reach and even exceed a PAL of 5.0 during several consecutive days of exercise. Additionally, of the 22 other studies included in our analysis, 6 reported PALs between 3.0 and 4.0, which is still a very high level of activity. Of those 6 studies, 3 were military studies (13,14,17), 2 were training (runners and swimmers) (23,26), and 1 was a study on an ultra-endurance run around Australia (11). Finally, 13 other studies reported modestly high PALs between 2.0 and 3.0 and were from military, training, and athletic events.
We hypothesized that catabolism of body tissue would occur in studies where the PAL approached or exceeded 5.0. Although two studies showed that athletes could achieve high PALs (above 5.0) for extended periods of time (longer than 5 d), they were unable to meet their energy requirements through consumption of foods and thus catabolized their body tissues (likely fat tissue as well as muscle tissue). This indicates that their bodies could not sustain the high levels of EE indefinitely. Athletes who surpassed a PAL of 5.0 showed some of the most extreme levels of catabolism, going from 69 to 53 kg of body weight (loss of 1.2 kg·wk−1), from 89.9 to 71 kg of body weight (loss of 1.4 kg·wk−1) (25) during the Arctic Trek, and from 69.2 to 67.8 kg of body weight (loss of 0.5 kg·wk−1) during the Tour de France (29). As expected, these extreme changes in absolute measures of body weight were from the studies of the longest duration (events lasting 22 d all the way up to 95 d). Not surprisingly, body composition analysis 7 d after the conclusion of the 95-d Arctic Trek showed that both men had drastically reduced their body fat percentage to minimal levels (1.9% and 2.5%) (25). Although TEEs were not consistent during the 95-d trek, the largest weight loss occurred during the first 50 d (22.8 and 18.2 kg), which, not surprisingly, coincided with the higher TEEs during the first 50 d (35.5 and 29.1 MJ·d−1). Evidence of catabolism, however, extended beyond data from those two studies raising concern about how long humans can sustain high PAL levels. Although participants in many of these studies were able to complete their athletic endeavor, there was a metabolic cost. To further examine this issue, we analyzed the weight loss data by calculating weight loss per week to account for differences in study durations. When calculated this way, some of the largest decreases in body weight were found in studies of shorter duration. Forbes-Ewan et al. (7) reported a 2.3-kg·wk−1 weight loss (study period was 7 d), and Myers et al. (18) found a 1.8-kg·wk−1 weight loss (study duration was 13 d). Interestingly, the PALs in those studies, 2.7 for both, were some of the lowest reported in our analysis. Similarly, Mudambo et al. (17) reported a 1.75-kg·wk−1 weight loss (study duration was 12 d) with a slightly higher PAL of 3.3.
A correlation analysis of body weight lost per week versus PAL revealed a slight, nonsignificant negative correlation. This indicates that the highest PALs do not seem to be associated with the greatest changes in body weight. This was not necessarily surprising because some of the high PALs are associated with situations in which dietary provisions are scarce because they have to carry all food with them and/or food is just not readily accessible. Therefore, it may be the lack of energy intake (EI) rather than PAL achieved that would be affecting weight loss. We decided to explore this a little more and split the studies up into those that seemed to have some energy restriction (3,7,11,13–15,17,18,20,25,29) versus those that did not have energy restriction (1,3–6,8,12,16,22,23,26). There was no correlation between weight change per week and PAL for energy-restricted studies (r = −0.03, NS); however, there was a fairly strong positive correlation, albeit not statistically significant, between weight change per week and PAL for the non–energy-restricted studies (r = 0.52, P = 0.15). In those studies where EI was not restricted, it seems that the higher PALs may result in the greatest changes in body weight, which would mean that PAL, rather than EI, was the driving force behind changes in body weight. These data should be regarded with caution, though, because nearly all of the PALs in those studies were moderate (average PAL of 2.63) and the weight change per week was much less than that in the energy-restricted studies (−0.20 vs −1.1 kg·wk−1 for energy restricted and non–energy restricted, respectively). Conversely, in studies where there is likely a component of energy restriction, it is the dietary restriction, rather than PALs achieved, that seems to be the driving force behind weight loss because in those studies, there was no relationship between PAL and weight change per week. Regardless of whether or not body weight is calculated in absolute terms or per unit of time, humans are unlikely to sustain even modest to high PALs without suffering significant catabolism of body tissues if the events are long in duration and especially if food is not widely available for subjects to self-select. The mild to severe losses of body fat shown in some of these studies would suggest that humans may only be able to sustain high levels of activity (four to five times greater than their BMR) as long as excess body fat is available to provide additional energy.
In contrast to extreme catabolism that was observed in athletic events and military training, a significantly lower level of catabolism was observed in studies reporting on athletic training regimens. This is shown in the study by Fudge et al. (8) on professional Kenyan endurance runners 1 wk before their national event where body fat did not significantly change (decreased from 7.1% ± 2.5% to 6.9% ± 2.1%). Another study on regular training in four female and four male Swedish national cross-country skier athletes showed no changes in body weight while sustaining a PAL of 3.4 and 4.0, respectively (23). One study on male and female firefighters during wildfire suppression showed no significant decrease in body weight or body fat percentage with PALs of 2.8 and 2.5, respectively (20). Finally, seven male Japanese professional soccer players competing in two games and training sessions for five other days showed a nonsignificant increase in body mass (69.8 ± 4.7 to 69.9 ± 4.7 kg) (5). Together, these studies indicate that catabolism of body tissue is preventable in nonextreme events or competition with proper EI. Further, small changes in body weight that do occur in some of the studies are more likely due to changes in whole-body hydration status rather than catabolism of body tissue.
It is known that EI is often limited during exercise, but on the basis of the results of this analysis, we speculate that the amount of EI does not have a direct effect on TEE or PAL. It does, however, likely affect the degree or magnitude of weight loss that occurs at several different PALs. A lack of effect of EI on TEE or PAL is supported by DeLany et al. (3), who studied two special forces groups undergoing training maneuvers, each given a different food ration. One group was given meals containing 16.8 MJ·d−1 (4022 kcal·d−1) of energy, whereas the second group was only given 8.3 MJ·d−1 (1987 kcal·d−1). TEEs, which did not differ between groups, were 14.6 ± 0.9 and 13.9 ± 1.2, respectively. Further, the calculated PALs for each group were 2.0 and 1.8, respectively (3). Because the TEE and PAL for both groups were very similar, we may extrapolate that the difference in EI did not significantly affect the TEE or PAL. In a different study, Ruby et al. (20) reported a TEE of 20.4 MJ·d−1 and an EI of 17 MJ·d−1 in male firefighters as well as a TEE of 14.8 and an EI of 13.5 in female firefighters. Very small changes in body composition and body weight were found (average weight changes of 0.7 and 0.14 kg·wk−1 for men and women, respectively). However, the energy deficit was greater in male firefighters, and they also showed slightly more weight loss than female subjects who had a smaller energy deficit. Importantly, these results may be due to differences in hydration status rather than catabolism of body tissue. The study duration was only 5 d, so it would likely take more days of these firefighters in an energy deficit to find significant changes in body weight or composition.
We determine that limited EI would have greatly influenced the weight loss that occurred at several different PALs. We further suggest that EI rather than PALs achieved must be considered when predicting catabolism of body tissue because we did not find a significant correlation between body weight lost per week and PAL. We found that the greatest changes in body weight (when calculated per week) were in studies with modest PALs. Two of those studies were military training in the jungle or African bush. This indicates that EI and food availability may be the primary driving force in determining whether catabolism occurs. Forbes-Ewan et al. (7) had the greatest weight loss per week (−2.3 kg) during jungle training in soldiers but had a modest PAL of 2.7. Mudambo et al. (17) reported a weight loss of 1.8 kg·wk−1 and a PAL of 3.3 during military training in the African bush. This shows that military rations alone and in restricted availability will cause eventual problems even when TEE and PAL are conservative. Finally, a study in one racer competing in a transatlantic yacht race had a PAL of 2.7 but weight loss of 1.8 kg·wk−1 (18). To summarize, it is quite likely that EI was limited in each of these three studies and that limitation, rather than extremely high EEs or PALs, led to more extreme catabolic losses in the body.
A final area of exploration was the relationship between initial body weight of the athletes and PALs achieved. We hypothesized that the highest PALs would be observed in those with low or modest body weight. This was not the case, however. There was, in fact, a slight but nonsignificant positive correlation between PAL and body weight. We also looked at the relationship between TEE and body weight, which showed an even stronger (and significant) positive correlation. Therefore, people exhibiting higher initial body weights tended to expend slightly more energy than their lower body weight counterparts. Larger individuals, however, have a larger TEE when expressed as megajoules per day because of higher BMRs. Therefore, we correlated activity EE (TEE − BMR) with initial body weight (Fig. 3). Accounting for body weight still showed a trend for a positive correlation between activity EE and initial body weight (r = 0.42, P = 0.06). Therefore, it seems that even when BMR is removed from the equation, individuals with a higher initial body weight may be able to expend more total energy and possibly achieve higher PALs. Because catabolism of body tissues was also observed in many studies, it may be beneficial for athletes with slightly higher initial body weights and body fat percentages to achieve and sustain high PALs as they may be able to afford more catabolism of body tissue (fat mass) without hurting performance.
While performing this analysis, we found that humans could reach a level higher than 5.0 times their BMR for periods of longer than a day. There were weaknesses, however, to this observation. In the event with the highest sustained PAL as well as several others, we used a calculated BMR rather than one that was measured. However, we calculated BMR using the WHO equation in those studies with a measured BMR and found very similar results (6.24 ± 1.11 vs 6.19 ± 1.08 MJ·d−1 for measured vs calculated BMR, respectively). Therefore, we feel confident in using calculated BMR in those studies where it was not measured. Another weakness was that throughout many of the events, athletes’ bodies underwent weight loss and thus were catabolic. It is therefore possible that these changes in body weight or composition created slight inaccuracies regarding calculated BMRs, which were based on initial body weight. This becomes more of a concern for those studies that were the longest in duration or those that reported the largest changes in body weight. As mentioned above briefly, one final limitation to these studies is EI. We believe that in most cases, EI was extremely limited because of participation in each athletic event or military training. It is likely that in a military training setting or training for an athletic event, EI is not a limiting factor. However, when allowed to free range in hostile settings, limitations of provisions are a function of logistics and availability, including EI. A few additional limitations to our review of these studies was that reported TEE, PAL, and other athlete variables were presented as a collective average or mean of the data rather than on an individual basis. Extremely high or low values could be hidden from these averages, thus masking accurate PAL and TEE measurements. This could greatly dampen or enhance peak values, EI, or body weight and body compositions of each report. Finally, the individuals that were included in this analysis are probably quite physically fit and are not likely to be representative of the general population. Consequently, most individuals in the general population are unlikely to be able to sustain such high TEEs for extended periods of time. However, if less fit individuals were to complete some of these extreme events, we would expect to see similar results in terms of catabolism of body tissue with high sustained PALs. Therefore, the greatest difference between the individuals studied here and the general population may be the lack of ability for people in the general public to reach and sustain PALs over 5.0.
In summary, the findings from this analysis show that some elite humans can achieve sustained PALs in excess of 5.0. The ability to sustain PALs above 5.0, however, may be limited to the amount of energy they can consume because large decreases in body weight and fat were found in those studies with PALs above 5.0. Further, significant losses of body mass or fat mass also occurred in some but not all of the studies with modest PALs (ranging from 2.5 up to 5.0). The energy balance equation dictates that both TEE and EI must be considered when predicting the catabolic consequences of extreme EE events. EI may also be a more important determinant of sustainability of high EEs rather than the magnitude of PAL achieved. Finally, we found that initial body weight was not related to the PAL that could be achieved or sustained. Future studies that measure body mass and body composition at the same period that DLW analysis of TEE is done are needed. In addition, studies that provide controlled rations of varying energy contents are needed to truly measure the effect of EI on achievable and sustainable PALs.
This study was supported in part by the National Institutes of Health grants RR025011 and DK007665.
The authors report no conflict of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Branth S, Hambraeus L, Westerterp K, et al.. Energy turnover in a sailing crew during offshore racing around the world. Med Sci Sports Exerc. 1996; 28 (10): 1272–6.
2. Brit-Friesen VL, Montevecchi WAK, Cairns DK, Macko SA. Activity-specific metabolic rates of free-living northern gannets and other seabirds. Ecology. 1989; 70: 357–67.
3. DeLany JP, Schoeller DA, Hoyt RW, Askew EW, Sharp MA. Field use of D2 18
O to measure energy expenditure of soldiers at different energy intakes. J Appl Physiol. 1989; 67 (5): 1922–9.
4. Ebine N, Feng JY, Homma M, Saitoh S, Jones PJ. Total energy expenditure of elite synchronized swimmers measured by the doubly labeled water method. Eur J Appl Physiol. 2000; 83 (1): 1–6.
5. Ebine N, Rafamantanantsoa HH, Nayuki Y, et al.. Measurement of total energy expenditure by the doubly labelled water method in professional soccer players. J Sports Sci. 2002; 20 (5): 391–7.
6. Edwards JE, Lindeman AK, Mikesky AE, Stager JM. Energy balance in highly trained female endurance runners. Med Sci Sports Exerc. 1993; 25 (12): 1398–404.
7. Forbes-Ewan CH, Morrissey BL, Gregg GC, Waters DR. Use of doubly labeled water technique in soldiers training for jungle warfare. J Appl Physiol. 1989; 67 (1): 14–8.
8. Fudge BW, Westerterp KR, Kiplamai FK, et al.. Evidence of negative energy balance using doubly labelled water in elite Kenyan endurance runners prior to competition. Br J Nutr. 2006; 95 (1): 59–66.
9. Hammond KA, Diamond J. Maximal sustained energy budgets in humans and animals. Nature. 1997; 386 (6624): 457–62.
10. Hayes M, Chustek M, Heshka S, Wang Z, Pietrobelli A, Heymsfield SB. Low physical activity levels of modern Homo sapiens
among free-ranging mammals. Int J Obes (Lond). 2005; 29 (1): 151–6.
11. Hill RJ, Davies PS. Energy expenditure during 2 wk of an ultra-endurance run around Australia. Med Sci Sports Exerc. 2001; 33 (1): 148–51.
12. Hill RJ, Davies PS. Energy intake and energy expenditure in elite lightweight female rowers. Med Sci Sports Exerc. 2002; 34 (11): 1823–9.
13. Hoyt RW, Jones TE, Baker-Fulco CJ, et al.. Doubly labeled water measurement of human energy expenditure during exercise at high altitude. Am J Physiol. 1994; 266 (3 Pt 2): R966–71.
14. Hoyt RW, Jones TE, Stein TP, et al.. Doubly labeled water measurement of human energy expenditure during strenuous exercise. J Appl Physiol. 1991; 71 (1): 16–22.
15. Jones PJ, Jacobs I, Morris A, Ducharme MB. Adequacy of food rations in soldiers during an arctic exercise measured by doubly labeled water. J Appl Physiol. 1993; 75 (4): 1790–7.
16. Jones PJ, Leitch CA. Validation of doubly labeled water for measurement of caloric expenditure in collegiate swimmers. J Appl Physiol. 1993; 74 (6): 2909–14.
17. Mudambo KS, Scrimgeour CM, Rennie MJ. Adequacy of food rations in soldiers during exercise in hot, day-time conditions assessed by doubly labelled water and energy balance methods. Eur J Appl Physiol Occup Physiol. 1997; 76 (4): 346–51.
18. Myers SD, Leamon SM, Nevola VR, Llewellyn MG. Energy expenditure during a single-handed transatlantic yacht race. Br J Sports Med. 2008; 42 (4): 285–8.
19. Peterson CC, Nagy KA, Diamond J. Sustained metabolic scope. Proc Natl Acad Sci U S A. 1990; 87: 2324–8.
20. Ruby BC, Shriver TC, Zderic TW, Sharkey BJ, Burks C, Tysk S. Total energy expenditure during arduous wildfire suppression. Med Sci Sports Exerc. 2002; 34 (6): 1048–54.
21. Schofield WN. Predicting basal metabolic rate, new standards and review of previous work. Hum Nutr Clin Nutr. 1985; 39 (1 suppl): 5–41.
22. Schulz LO, Alger S, Harper I, Wilmore JH, Ravussin E. Energy expenditure of elite female runners measured by respiratory chamber and doubly labeled water. J Appl Physiol. 1992; 72 (1): 23–8.
23. Sjödin AM, Andersson AB, Högberg JM, Westerterp KR. Energy balance in cross-country skiers: a study using doubly labeled water. Med Sci Sports Exerc. 1994; 26 (6): 720–4.
24. Speakman J. Doubly Labelled Water—Theory and Practice. New York (NY): Springer; 1997. p. 65–76.
25. Stroud MA, Ritz P, Coward WA, et al.. Energy expenditure using isotope-labelled water (2
O), exercise performance, skeletal muscle enzyme activities and plasma biochemical parameters in humans during 95 days of endurance exercise with inadequate energy intake. Eur J Appl Physiol Occup Physiol. 1997; 76 (3): 243–52.
26. Trappe TA, Gastaldelli A, Jozsi AC, Troup JP, Wolfe RR. Energy expenditure of swimmers during high volume training. Med Sci Sports Exerc. 1997; 29 (7): 950–4.
27. Westerterp KR, Bryant DM. Energetics of free existence in swallows and martins (hirundinidae) during breeding: a comparative study using doubly labeled water. Oecologia. 1984; 62: 376–81.
28. Westerterp KR, Plasqui G. Physical activity and human energy expenditure. Curr Opin Clin Nutr Metab Care. 2004; 7 (6): 607–13.
29. Westerterp KR, Saris WH, van EM, ten HF. Use of the doubly labeled water technique in humans during heavy sustained exercise. J Appl Physiol. 1986; 61 (6): 2162–7.
30. Westerterp-Plantenga MSFEWSAB. Food Intake and Energy Expenditure. Boca Raton (FL): CRC Press; 1994. p. 311–9.