Image Map

The Science of Human Energy Expenditure – Part III

In Part II of this series on energy expenditure, we turned our attention to the measurement of free living energy expenditure using doubly labeled water. In Part III, we will conclude by looking at validation of this method’s efficacy for use in humans, as well as potential issues around its use.

Validations of Doubly Labeled Water in Humans

The doubly labeled water method has been used in humans for the last decade and its ability to accurately estimate total energy expenditure in free-living populations has increased its use by laboratories interested in human energy expenditure. Since the original validation study in 1982 (Schoeller and Van Santen, 1982) there have been numerous other investigators who have validated the technique in their laboratories to estimate energy expenditure in human beings in various environments (Appendix I).

Validations of Doubly Labeled Water with Exercise

The doubly labeled water technique has been compared against respiratory gas analysis via a metabolic chamber at both high and low activity levels to determine if the method is valid at high levels of expenditure. Westerterp et al (1988) studied the difference in carbon dioxide production rates in nine subjects measured over six days with doubly labeled water while the subjects lived in a respiratory chamber. Five persons measured at low activity levels had a standard deviation between means of 3.9% while four persons performing daily on a bicycle ergometer had a standard deviation of 7.0% between the two methods (Westerterp et al., 1988). Klein et al. observed a 44 year old male (70.5 kg, 171.5 cm and 26.5% body fat) during a five day inpatient stay in a metabolic chamber to determine total energy expenditure with simultaneous measurement with doubly labeled water (Klein et al., 1984). Included in this protocol were four one-half hour sessions on a bicycle ergometer (1 kp load). These authors reported excellent agreement between the two methods (2143 kcal/d vs. 2105 kcal/d), validating the method with exercise in a metabolic chamber.

Estimation of Physical Activity with Doubly Labeled Water

Deriving estimates for the energy cost of physical activity under free-living conditions has traditionally been challenging due to the lack of an accurate, unrestrictive and convenient measurement device. Despite the lack of free-living data in humans, exercise is often prescribed as a way for individuals to deplete excess energy stores, increase lean body mass and also to increase overall fitness. It has been well documented that activity will increase expenditure during the exercising bout and for a period following an exercise bout (Poehlman et al., 1991; Bielinski et al., 1985). However, the possibility of studying the overall metabolic changes have been limited to information obtained using the more subjective methods, such as heart rate monitoring and activity recording, as previously discussed. The doubly labeled water technique is unique in the sense that it is the only known methodology which can be used to non-invasively estimate the energy expenditure of physical activity. This is accomplished by measuring resting metabolic rate and measuring the thermic response of a meal or more commonly assume a value (usually 10% of total energy expenditure) while the measurement of total energy expenditure is taking place (Goran & Poehlman, 1992). Deriving the equation:

EEPA = (0.9 * TEE) – RMR

Where EEPA = energy expenditure of physical activity, TEE = total energy expenditure, RMR = resting metabolic rate and 0.9 is the correction factor for the thermic response of a meal

Two of the published exercise studies using doubly labeled water were conducted by Westerterp et al. (1986) and Goran and Poehlman (1992). Westerterp and his colleagues used the method in athletes that were competing in the Toure de France Bicycle race. The subjects were measured over three different phases of the race and the method was validated using the energy intake / energy balance method. During this protocol the highest levels of energy expenditure to date were recorded in these participants, exceeding 8000 kilocalories per day. This was at a level that was over five times their measured resting metabolic rates. It was also found that these athletes had a tendency to underreport their energy intake. However, by utilizing this method in other areas of high performance athletics it may be possible for exact nutrient intakes to be established to create a diet that will help increase optimal performance.

Goran and Poehlman (1992) used the doubly labeled water technique in a cohort of 13 elderly persons (age range 56 – 78, six females, seven males) to evaluate the changes in energy utilization in response to a period of exercise. Total energy expenditure was measured for 10 days during a control period and during a prescribed exercise training protocol. Total energy expenditure of this population was not significantly different during the last 10 days of training compared to the baseline period (TEE, 2408 + 478 to 2479 + 497 kcal/day).

This finding was observed in light of the fact that an increase of 11% in resting metabolic rate (equivalent to approximately 167 kcal/day) was calculated and the additional caloric cost of the exercise training (150 kcal/day). From these data it may be concluded that elderly persons become less active in their leisure time when participating in an exercise program (energy expenditure of physical activity 571 + 386 to 340 + 452 kcal/day).

Total Energy Expenditure in Disease and Trauma

Human trauma is a state in which there are usually special energy needs to promote optimal recovery in a patient. Traumatic states have been described as conditions in which there are both increased energy needs (i.e. burn victims and surgery patients) and decreased energy needs (i.e. anorexia nervosa and bulimia) For some time burn patients were treated nutritionally with hypercaloric diets to help combat the hypermetabolism associated with the healing process (Caldwell et al., 1981; Long, 1979).

Goran et al. (1990) used the doubly labeled water technique to show that the energy requirements of burned children and adolescents were actually much lower then previously thought (TEE / RMR = 1.18 + 0.17, n = 8) and resting metabolic rate was only elevated 29% above basal metabolic rate (TEE = 1.55 * BMR), (Goran et al., 1990). This indicates that the traditional hypercaloric diets used in the treatment of burn patients may not be necessary to maintain actual energy balance.

Novick et al. (1988) measured the energy cost of elective surgery by using the doubly labeled water technique in 7 female patients (34 + 6 years, 53 + 2 kg). They measured both a baseline period, the surgical procedure and the recuperation period totaling 10 consecutive days. They reported an 11.9% increase in carbon dioxide production rate which translated into an increase of total energy expenditure of 18% (267 kcal/day). These findings were in agreement with those found by Long et al. (1979) and Askanazi et al. (1981) who used near continuous gas exchange to measure metabolic rate in postoperative surgery patients. As discussed, the method is not a good indicator of the actual energetic cost of the surgery since it measures expenditure over a period of days but, the method could be used more successfully during the post-operative stage to determine energy expenditure during recuperation.

In a study of the energy expenditure of anoretic females, Casper et al. (1991) measured total energy expenditure using doubly labeled water. A total of twelve females were studied, six of which were defined as clinically anoretic and six of who served as controls. The subjects had similar ages and heights (mean 25 years, 163.5 cm) but, had significant differences in weight (42.5 vs. 56.5 kg), body mass index (15.7 vs. 21.6) and a test of eating attitude (47.0 vs. 10.2). The subjects had similar values of total energy expenditure (1972 vs. 1985 kcal/d) but, displayed significantly different resting energy expenditure (997 vs. 1319 kcal/d). The authors concluded that anoretic females have a decreased resting expenditure, probably related to starvation, and a higher then average physical activity level.

Obesity and Energy Expenditure

Obesity is defined as the inability to maintain energy balance where energy intake exceeds energy expenditure. It has been theorized that obese people had a metabolic adaptation which make which make them more energy efficient, therefore, having a better ability to store excess intake as body fat (James, 1978). With the ability of the doubly labeled water technique to measure free-living total energy expenditure, some of the possible mechanisms of obesity have been examined.

In a study conducted by Welle et al. (1992) 12 normal weight females (59.6 + 4.0 kg) were compared to 26 obese females (85.2 + 10.8 kg). They found that on an absolute scale the obese women had higher values of both resting energy expenditure (1342 vs. 1178 kcal/day) and total energy expenditure (2323 vs. 1962 kcal/day). However, after adjusting the values of total energy expenditure for weight and lean body mass, using analysis of covariance, there was no significant difference between the two groups therefore, concluding that obese women must consume more energy to maintain their obese state (Welle et al., 1992).

In a similar study, Prentice et al. (1986) reported the differences between energy expenditure in 13 lean (57.5 + 6.3 kg) and 9 obese (87.9 + 14.3 kg) females. As in the previous study the obese groups had significantly higher values of both resting (1605 vs. 1352 kcal/d) and total energy expenditure (2146 vs. 1775 kcal/d). However, when resting energy expenditure was expressed per kilogram of fat free mass the values between the groups were similar. After adjusting total energy expenditure for resting energy expenditure using the TEE:RMR ratio, the obese group still had a slightly higher level of expenditure (1.336 vs. 1.297). This study reinforces previously reported data in which Prentice et al. (1985) showed that healthy women had lower than expected levels of total energy expenditure (expressed as 1.38 * RMR) when measured by the doubly labeled water technique.

Statement of the Problem: Expression of Data

The doubly labeled water technique has revolutionized the way which modern energy expenditure research is being conducted. Since its use in human populations a decade ago, we have gained information on the energy utilization in obesity (Welle et al. 1992; Prentice et al., 1986), aging (Goran and Poehlman, 1992) pregnancy and lactation (Roberts et al., 1986), exercise (Goran and Poehlman, 1992; Westerterp et al., 1986) and trauma (Goran et al.,1990, Novick et al.,1988). However, an appropriate method for the expression of total energy expenditure data has yet to be identified.

At present there has not been an investigation in the expression of total energy expenditure data generated by the doubly labeled water technique which allows for accurate, inter-individual comparison between persons with different metabolic body size (body weight and height) and different body compositions (proportions of fat-free mass).

Traditionally total energy expenditure data has been expressed in terms of a ratio with resting metabolic rate, expressed as TEE: RMR, yielding the following equation:

TEE = b * RMR Equation 1

Where TEE = total energy expenditure, RMR = resting metabolic rate and b is the activity factor

The activity factor which is generated from this equation has historically been used to estimate total energy expenditure and energy requirements from resting metabolic rate when more appropriate methods of measurement were not available (FAO-WHO-UNU, 1985; National Research Council, 1987). However, since the doubly labeled water method has been validated in the human population as an accurate means of estimating free-living, total energy expenditure, the use of the activity factor for the expression of total energy expenditure data is still currently used in the literature. It seems possible that the TEE:RMR ratio remains in the literature due to a matter of convenience and the lack of an alternative method for the expression of data.

Recently, Bingham et al. (1989) studied the effects of exercise on resting metabolic rate in three women and three men (weight range 51 – 69 kg). Total energy expenditure was measured using doubly labeled water for both a control period and treatment period for 13 days each. Although the subjects spanned a considerable weight range and were of mixed gender the total energy expenditure data was expressed in terms of an activity factor with resting metabolic rate (control period, 1.58 * RMR, treatment period, 1.99 * RMR).

Roberts et al. (1991) studied the energy requirements of fourteen young men over 10 days using the doubly labeled water technique (mean age 22.3, weight range 56 – 99 kg, percent fat range 2.7% – 16.2%). The subjects were free living and allowed to carry on with their typical daily routines and activities. Measured total energy expenditure data was again expressed in terms of a ratio with resting metabolic rate even though there was a wide range of body sizes, compositions and energy expenditures (mean value, 1.98, range 1.59 – 2.60).

The problem in using the traditional activity factor to express data is that it has never been tested in a large enough data set to demonstrate that it satisfies the underlying assumptions on which it is based. The use of the activity factor assumes: 1) that the relationship between total energy expenditure and resting metabolic rate is significant, 2) linear in nature, 3) passes through the origin at zero and 4) that total energy expenditure is a single compartment model (i.e. total energy expenditure is a multiple of resting expenditure) ruling out the contribution of other physiological variables (eg body weight). Total energy expenditure is the sum of resting metabolic rate, thermic effect of food and energy expenditure of physical activity, modeled by the more complex equation:

TEE = b1 * RMR + b2 * TEM + b3 * EEPA + i Equation 2

Where TEE is total energy expenditure, RMR is resting metabolic rate, TEM is the thermic response to a meal, EEPA is the energy expenditure physical activity, bi is the relevant regression coefficient and i is the intercept.

Therefore, the activity factor is not a proper method of expressing data from a physiological stand-point. Thus, the purpose of the following study was to: 1) Compile total energy expenditure data derived from the doubly labeled water method from the literature and determine the strongest biological markers. 2) To test the underlying assumptions of the traditional ratio method. 3) To propose a regression equation for the normalization of total energy expenditure data for the purpose of data comparison.

In Part II of this series on energy expenditure, we turned our attention to the measurement of free living energy expenditure using doubly labeled water. In Part III, we will conclude by looking at validation of this method’s efficacy for use in humans, as well as potential issues around its use.

Validations of Doubly Labeled Water in Humans

The doubly labeled water method has been used in humans for the last decade and its ability to accurately estimate total energy expenditure in free-living populations has increased its use by laboratories interested in human energy expenditure. Since the original validation study in 1982 (Schoeller and Van Santen, 1982) there have been numerous other investigators who have validated the technique in their laboratories to estimate energy expenditure in human beings in various environments (Appendix I).

Validations of Doubly Labeled Water with Exercise

The doubly labeled water technique has been compared against respiratory gas analysis via a metabolic chamber at both high and low activity levels to determine if the method is valid at high levels of expenditure. Westerterp et al (1988) studied the difference in carbon dioxide production rates in nine subjects measured over six days with doubly labeled water while the subjects lived in a respiratory chamber. Five persons measured at low activity levels had a standard deviation between means of 3.9% while four persons performing daily on a bicycle ergometer had a standard deviation of 7.0% between the two methods (Westerterp et al., 1988). Klein et al. observed a 44 year old male (70.5 kg, 171.5 cm and 26.5% body fat) during a five day inpatient stay in a metabolic chamber to determine total energy expenditure with simultaneous measurement with doubly labeled water (Klein et al., 1984). Included in this protocol were four one-half hour sessions on a bicycle ergometer (1 kp load). These authors reported excellent agreement between the two methods (2143 kcal/d vs. 2105 kcal/d), validating the method with exercise in a metabolic chamber.

Estimation of Physical Activity with Doubly Labeled Water

Deriving estimates for the energy cost of physical activity under free-living conditions has traditionally been challenging due to the lack of an accurate, unrestrictive and convenient measurement device. Despite the lack of free-living data in humans, exercise is often prescribed as a way for individuals to deplete excess energy stores, increase lean body mass and also to increase overall fitness. It has been well documented that activity will increase expenditure during the exercising bout and for a period following an exercise bout (Poehlman et al., 1991; Bielinski et al., 1985). However, the possibility of studying the overall metabolic changes have been limited to information obtained using the more subjective methods, such as heart rate monitoring and activity recording, as previously discussed. The doubly labeled water technique is unique in the sense that it is the only known methodology which can be used to non-invasively estimate the energy expenditure of physical activity. This is accomplished by measuring resting metabolic rate and measuring the thermic response of a meal or more commonly assume a value (usually 10% of total energy expenditure) while the measurement of total energy expenditure is taking place (Goran & Poehlman, 1992). Deriving the equation:

EEPA = (0.9 * TEE) – RMR
Where EEPA = energy expenditure of physical activity, TEE = total energy expenditure, RMR = resting metabolic rate and 0.9 is the correction factor for the thermic response of a meal

Two of the published exercise studies using doubly labeled water were conducted by Westerterp et al. (1986) and Goran and Poehlman (1992). Westerterp and his colleagues used the method in athletes that were competing in the Toure de France Bicycle race. The subjects were measured over three different phases of the race and the method was validated using the energy intake / energy balance method. During this protocol the highest levels of energy expenditure to date were recorded in these participants, exceeding 8000 kilocalories per day. This was at a level that was over five times their measured resting metabolic rates. It was also found that these athletes had a tendency to underreport their energy intake. However, by utilizing this method in other areas of high performance athletics it may be possible for exact nutrient intakes to be established to create a diet that will help increase optimal performance.

Goran and Poehlman (1992) used the doubly labeled water technique in a cohort of 13 elderly persons (age range 56 – 78, six females, seven males) to evaluate the changes in energy utilization in response to a period of exercise. Total energy expenditure was measured for 10 days during a control period and during a prescribed exercise training protocol. Total energy expenditure of this population was not significantly different during the last 10 days of training compared to the baseline period (TEE, 2408 + 478 to 2479 + 497 kcal/day).

This finding was observed in light of the fact that an increase of 11% in resting metabolic rate (equivalent to approximately 167 kcal/day) was calculated and the additional caloric cost of the exercise training (150 kcal/day). From these data it may be concluded that elderly persons become less active in their leisure time when participating in an exercise program (energy expenditure of physical activity 571 + 386 to 340 + 452 kcal/day).

Total Energy Expenditure in Disease and Trauma

Human trauma is a state in which there are usually special energy needs to promote optimal recovery in a patient. Traumatic states have been described as conditions in which there are both increased energy needs (i.e. burn victims and surgery patients) and decreased energy needs (i.e. anorexia nervosa and bulimia) For some time burn patients were treated nutritionally with hypercaloric diets to help combat the hypermetabolism associated with the healing process (Caldwell et al., 1981; Long, 1979).

Goran et al. (1990) used the doubly labeled water technique to show that the energy requirements of burned children and adolescents were actually much lower then previously thought (TEE / RMR = 1.18 + 0.17, n = 8) and resting metabolic rate was only elevated 29% above basal metabolic rate (TEE = 1.55 * BMR), (Goran et al., 1990). This indicates that the traditional hypercaloric diets used in the treatment of burn patients may not be necessary to maintain actual energy balance.

Novick et al. (1988) measured the energy cost of elective surgery by using the doubly labeled water technique in 7 female patients (34 + 6 years, 53 + 2 kg). They measured both a baseline period, the surgical procedure and the recuperation period totaling 10 consecutive days. They reported an 11.9% increase in carbon dioxide production rate which translated into an increase of total energy expenditure of 18% (267 kcal/day). These findings were in agreement with those found by Long et al. (1979) and Askanazi et al. (1981) who used near continuous gas exchange to measure metabolic rate in postoperative surgery patients. As discussed, the method is not a good indicator of the actual energetic cost of the surgery since it measures expenditure over a period of days but, the method could be used more successfully during the post-operative stage to determine energy expenditure during recuperation.

In a study of the energy expenditure of anoretic females, Casper et al. (1991) measured total energy expenditure using doubly labeled water. A total of twelve females were studied, six of which were defined as clinically anoretic and six of who served as controls. The subjects had similar ages and heights (mean 25 years, 163.5 cm) but, had significant differences in weight (42.5 vs. 56.5 kg), body mass index (15.7 vs. 21.6) and a test of eating attitude (47.0 vs. 10.2). The subjects had similar values of total energy expenditure (1972 vs. 1985 kcal/d) but, displayed significantly different resting energy expenditure (997 vs. 1319 kcal/d). The authors concluded that anoretic females have a decreased resting expenditure, probably related to starvation, and a higher then average physical activity level.

Obesity and Energy Expenditure

Obesity is defined as the inability to maintain energy balance where energy intake exceeds energy expenditure. It has been theorized that obese people had a metabolic adaptation which make which make them more energy efficient, therefore, having a better ability to store excess intake as body fat (James, 1978). With the ability of the doubly labeled water technique to measure free-living total energy expenditure, some of the possible mechanisms of obesity have been examined.

In a study conducted by Welle et al. (1992) 12 normal weight females (59.6 + 4.0 kg) were compared to 26 obese females (85.2 + 10.8 kg). They found that on an absolute scale the obese women had higher values of both resting energy expenditure (1342 vs. 1178 kcal/day) and total energy expenditure (2323 vs. 1962 kcal/day). However, after adjusting the values of total energy expenditure for weight and lean body mass, using analysis of covariance, there was no significant difference between the two groups therefore, concluding that obese women must consume more energy to maintain their obese state (Welle et al., 1992).

In a similar study, Prentice et al. (1986) reported the differences between energy expenditure in 13 lean (57.5 + 6.3 kg) and 9 obese (87.9 + 14.3 kg) females. As in the previous study the obese groups had significantly higher values of both resting (1605 vs. 1352 kcal/d) and total energy expenditure (2146 vs. 1775 kcal/d). However, when resting energy expenditure was expressed per kilogram of fat free mass the values between the groups were similar. After adjusting total energy expenditure for resting energy expenditure using the TEE:RMR ratio, the obese group still had a slightly higher level of expenditure (1.336 vs. 1.297). This study reinforces previously reported data in which Prentice et al. (1985) showed that healthy women had lower than expected levels of total energy expenditure (expressed as 1.38 * RMR) when measured by the doubly labeled water technique.

Statement of the Problem: Expression of Data

The doubly labeled water technique has revolutionized the way which modern energy expenditure research is being conducted. Since its use in human populations a decade ago, we have gained information on the energy utilization in obesity (Welle et al. 1992; Prentice et al., 1986), aging (Goran and Poehlman, 1992) pregnancy and lactation (Roberts et al., 1986), exercise (Goran and Poehlman, 1992; Westerterp et al., 1986) and trauma (Goran et al.,1990, Novick et al.,1988). However, an appropriate method for the expression of total energy expenditure data has yet to be identified.

At present there has not been an investigation in the expression of total energy expenditure data generated by the doubly labeled water technique which allows for accurate, inter-individual comparison between persons with different metabolic body size (body weight and height) and different body compositions (proportions of fat-free mass).

Traditionally total energy expenditure data has been expressed in terms of a ratio with resting metabolic rate, expressed as TEE: RMR, yielding the following equation:

TEE = b * RMR Equation 1
Where TEE = total energy expenditure, RMR = resting metabolic rate and b is the activity factor

The activity factor which is generated from this equation has historically been used to estimate total energy expenditure and energy requirements from resting metabolic rate when more appropriate methods of measurement were not available (FAO-WHO-UNU, 1985; National Research Council, 1987). However, since the doubly labeled water method has been validated in the human population as an accurate means of estimating free-living, total energy expenditure, the use of the activity factor for the expression of total energy expenditure data is still currently used in the literature. It seems possible that the TEE:RMR ratio remains in the literature due to a matter of convenience and the lack of an alternative method for the expression of data.

Recently, Bingham et al. (1989) studied the effects of exercise on resting metabolic rate in three women and three men (weight range 51 – 69 kg). Total energy expenditure was measured using doubly labeled water for both a control period and treatment period for 13 days each. Although the subjects spanned a considerable weight range and were of mixed gender the total energy expenditure data was expressed in terms of an activity factor with resting metabolic rate (control period, 1.58 * RMR, treatment period, 1.99 * RMR).

Roberts et al. (1991) studied the energy requirements of fourteen young men over 10 days using the doubly labeled water technique (mean age 22.3, weight range 56 – 99 kg, percent fat range 2.7% – 16.2%). The subjects were free living and allowed to carry on with their typical daily routines and activities. Measured total energy expenditure data was again expressed in terms of a ratio with resting metabolic rate even though there was a wide range of body sizes, compositions and energy expenditures (mean value, 1.98, range 1.59 – 2.60).

The problem in using the traditional activity factor to express data is that it has never been tested in a large enough data set to demonstrate that it satisfies the underlying assumptions on which it is based. The use of the activity factor assumes: 1) that the relationship between total energy expenditure and resting metabolic rate is significant, 2) linear in nature, 3) passes through the origin at zero and 4) that total energy expenditure is a single compartment model (i.e. total energy expenditure is a multiple of resting expenditure) ruling out the contribution of other physiological variables (eg body weight). Total energy expenditure is the sum of resting metabolic rate, thermic effect of food and energy expenditure of physical activity, modeled by the more complex equation:

TEE = b1 * RMR + b2 * TEM + b3 * EEPA + i Equation 2
Where TEE is total energy expenditure, RMR is resting metabolic rate, TEM is the thermic response to a meal, EEPA is the energy expenditure physical activity, bi is the relevant regression coefficient and i is the intercept.

Therefore, the activity factor is not a proper method of expressing data from a physiological stand-point. Thus, the purpose of the following study was to: 1) Compile total energy expenditure data derived from the doubly labeled water method from the literature and determine the strongest biological markers. 2) To test the underlying assumptions of the traditional ratio method. 3) To propose a regression equation for the normalization of total energy expenditure data for the purpose of data comparison.

×