The measurement of energy expenditure in humans dates back to the early 1900’s when Atwater developed and implemented the first human calorimeter (Atwater and Benedict, 1903). During the 90 years since this discovery it has been clearly established that both direct calorimetry and indirect calorimetry are viable and accurate methods of measuring energy expenditure (Webb et al., 1980). Direct calorimetry refers to the direct measurement of the heat production generated from energy utilization, while indirect calorimetry measures heat production from knowledge of respiratory gas exchange.
The capacity to measure energy expenditure has proven to be a powerful research tool for the understanding of various energy needs under a variety metabolic conditions such as pregnancy and lactation (Roberts et al., 1986; Heini et al., 1991; Singh et al., 1989), aging (Goran and Poehlman, 1992), trauma (Goran et al., 1990; Novick et al., 1988) over and underfeeding (Pasquet et al., 1992; Roberts et al.,1990) obesity (Welle et al., 1992; Prentice et al., 1986) and exercise (Westerterp et al., 1986; Goran and Poehlman 1992).
Measurement of human energy expenditure has been a focus of human biological research since Harris and Benedict completed their classic study in which they were interested in determining the factors influencing basal metabolic rate in the early 1900’s (Harris and Benedict, 1917). From this work these two investigators formulated prediction equations for basal requirements that still remain the standard equations used today.
Although technical advances in direct and indirect calorimetry have made it easier to measure both resting and thermic effect of food components of total expenditure under laboratory conditions, the quantification of exercise related expenditure and total expenditure in free-living humans has been somewhat more elusive. This was due primarily to the inability to measure energy expenditure by an unrestricted and unemcumbracing means in a free-living setting.
Components of Total Energy Expenditure
Total daily energy expenditure can be divided into three components: Resting metabolic rate, thermic response of a meal and energy expenditure of physical activity.
1) Resting Metabolic Rate
Resting metabolic rate accounts for between 60-75% of total daily energy expenditure and is associated with the energy cost of maintaining physiological homeostasis which includes the energy cost of maintaining body temperature, cardiac output, respiration, nervous system function and other non-voluntary activities. This component of energy expenditure is determined primarily by fat-free mass (Ravussin et al., 1988; Arciero et al., 1993) and is also influenced by: body fatness (Poehlman et al., 1991), gender (Astrup et al., 1992; Fukagawa et al., 1990) and physical fitness (Poehlman & Horton, 1990; Bingham et al., 1989; Westerterp et al., 1991).
Resting metabolic rate is traditionally measured 12 hours post-absorptive, under highly controlled basal conditions while the subject is lying comfortably in a room temperature environment with a clear plastic, ventilated hood covering their head. This measurement requires the subject to lie still while both inspired and expired respiratory gases (O2, CO2) are collected for approximately 30-45 minutes. The method is dependent on the fact that oxygen and carbon dioxide gas are the basic energy substrate and end product of metabolism that fuels the human body.
2) Thermic Response of a Meal
The thermic response of a meal is the energy cost of the digestion, absorption, mobilization and storage of the foodstuffs. On average this component comprises approximately 10% of total daily energy expenditure and is measured under similar testing conditions for resting energy expenditure for a period of approximately three to four hours after a standard test meal (Tuttle et al.,1953). The thermic response of a meal was originally reported as the specific energy cost for the digestion, mobilization and storage of the protein portion of the meal (Swift et al.,1957). It has been shown recently that 100 grams of protein increases the thermic response more than isocaloric loads of fat and carbohydrate (Welle et al., 1981).
However, these investigators noted an increase in sympathetic nervous activity, using plasma norepinephrine levels, in the carbohydrate meal and no change with both the protein and fat. This finding was confirmed by Schultz et al who found that dietary fat did not significantly increase the respiratory quotient (the ratio of expired carbon dioxide gas to inspired oxygen gas, CO2 / O2 = respiratory quotient) while carbohydrate increased both the respiratory quotient and substrate oxidation rates however, no change in urinary catacholamines was seen within any of the sub-groups (Schultz et al., 1989). This same group had earlier reported a rise in metabolic rate of 20% following the ingestion of a 500 gram meal of polysaccharide and had concluded that dietary carbohydrate may play a vital role in the maintenance of energy balance. It is therefore clear that the thermic effect of a meal varies according to the mixture of macronutrients ingested.
3) Energy Expenditure of Physical Activity
The energy expended in physical activity is the most variable component of total daily energy expenditure. It is composed of both involuntary (shivering) and voluntary (fidgeting and intentional physical activity) muscular activity (Poehlman et al., 1991). Recently, it has been shown that even under sedentary and controlled living conditions total energy expenditure can fluctuate by as much as + 9% (Goran et al., 1993). In this particular study, 5 males (23 + 4 years; 66.1 + 4.6 kg) were studied over two similar 10 day periods. Although there were no significant differences between total energy expenditure or resting energy expenditure between the two phases, (resting metabolism variation between phases, 3.7% + 3.9% kcal/day), it is evident by the non-significant fluctuation (change between phases) in both resting metabolic rate and total energy expenditure that physical activity is very spontaneous in nature even under controlled conditions (Goran et al., 1993).
The energy expended in physical activity is influenced somewhat by body weight and composition (heavier persons will require more energy than lighter persons and leaner persons will require more energy than fatter counterparts of the same weight for the same activity and intensity) but, is primarily driven by an individual’s desire and ultimate performance of physical activity (Welle et al.,1992). It has been shown that in healthy free-living adults this component of daily expenditure can be as low as 264 kilocalories per day in a 65 year old female (Goran & Poehlman, 1992) to over 5000 kilocalories per day in a group of three Toure de France bicycle racers (Westerterp et al., 1986).
Field Techniques For Measuring Total Energy Expenditure
The measurement of energy expenditure for the determination of energy requirements has been the focus of much attention in the field of human energy metabolism. Traditionally, energy requirements have been suggested from data that have been generated using food intake data (FAO-WHO-UNU, 1985). However, with the advancement of measurement techniques, namely the doubly labeled water technique, it is possible to establish energy requirements based on measurement of free-living energy expenditure.
Much of the work in this area has been aimed at specific populations so that proper energy requirements can be established for differing environmental conditions or specific nutritional and metabolic needs. There are several techniques that are used for these types of measurements: energy intake / energy balance, the factorial method, heart rate monitoring, respiratory gas exchange and doubly labeled water.
1) Energy Intake Recording
A common method of measuring total energy expenditure in humans is the energy intake recording method (Lichtman, 1992). Using this procedure, energy intake is recorded over a period of days and total expenditure is estimated to be the energy cost of the food stuffs consumed minus the caloric cost of the change in body energy stores (Schoeller and Van Santen, 1982). This procedure is heavily dependent on the ability of either the research staff or the subjects to accurately record the type and amount of food eaten during the study period. It also requires an accurate estimation of body composition at the beginning and at the end of the study period so that any change in energy stores can be noted.
The measurement of the initial and final body composition are very critical in this method since over or under feeding will directly impact the energy stores and thus, the accuracy of the measurement. Therefore, poor estimations of body composition can significantly alter total energy expenditure (If a + 2% error in estimating body composition is made in a 70 kg person an error of approximately 12,600 kcal is introduced in to the estimation of energy balance which, over a period of 14 days would introduce an error of + 900 kcal/day). While this method yields relatively accurate predictions of total energy expenditure under strictly controlled conditions, it is has been shown to be less accurate in free-living conditions. It has been reported in the literature that free-living persons in certain sub-groups are likely to grossly underestimate the amount (therefore, energy content) of food that they consume or simply grow inpatient with the detail needed to keep such records (Stunkard and Waxman, 1981; Schoeller et al., 1986).
Using the doubly labeled water method to estimate total energy expenditure, Goran and Poehlman (1992) have reported an underestimation of energy intake by using a three day diet record in a cohort of free-living elderly persons. They observed differences from the actual measured expenditure of 31% in the females and 12% in the males respectively, using comparisons from doubly labeled water (females, intake = 1432 kcal/day, doubly labeled water = 2092 kcal/day; males, intake = 2326 kcal/day, doubly labeled water = 2675 kcal/day). An underestimation in intake was also reported in a group of elite female runners was also found by Haggarty & McGaw (1990) and has been repeatedly shown in obese populations (Prentice et al., 1986, Welle et al., 1992; Lichtman et al., 1981).
2) The Factorial Method
The factorial method of estimating total energy expenditure involves the recording of specific activities performed over the course of a day. The energy cost of activities are derived from published tables and summed at the end of the study period, usually a period of several days (Durnin & Passmore, 1955; Forbes-Ewan et al., 1988). The factorial approach is a relatively accurate technique when preformed under controlled laboratory conditions; however, when applied to free-living conditions there is a greater likelihood for under-reporting of activity due to the time involved filling out the record or forgetfulness (Forbes-Ewan et al., 1988). This error can be experienced when either the duration or the intensity is over or underestimated by the subject. It is also subject to the availability of proper, up-to-date activity charts that include a vast variety of currently preformed activities.
Schulz et al. (1989) used the factorial method in a validation study with the doubly labeled water technique for 14 days in 6 subjects (four males, two females). In this protocol subjects were asked to record their activities every 15 minutes for the time that they were awake. The subjects were given a table of various activities and asked to record them using a scale of 1 to 12 (1 low demand, 12 high demand) to determine the relative energy demands of the task which was used to calculate total energy expenditure. In this particular study the investigators noted little difference between the two methods (TEE doubly labeled water = 3150 kcal/day, factorial method = 2983 kcal/day).
3) Heart Rate Monitoring
Heart rate monitoring is a method of estimating energy expenditure that is based on the strong positive correlation that exists between heart rate and oxygen consumption (Payne et al., 1971; Bradfield, 1971). This relationship between heart rate and oxygen consumption can be explained by the demand for oxygen, which is used by the respiring tissues as the fuel substrate, as either the intensity or duration of the activity increases. Therefore, there is an inherent correlation between increased oxygen consumption and cardiac output. It has also been shown that heart rate will respond to changes in food intake and environmental temperature (Webster, 1967).
These findings lead to the study of heart rate monitoring as a possible tool for the measurement of free-living energy expenditure in the field. The technique requires establishing individual regression lines relating heart rate and oxygen consumption for each subject under laboratory conditions. This is achieved by monitoring both heart rate and oxygen consumption while subjects perform step increases in intensity of activity (usually walking on a treadmill or stepping up on a block) from resting levels to a heart rate greater than 160 beats per minute. Heart rate and oxygen consumption are then plotted as a regression so that a given heart rate can be used to an determine an approximate oxygen consumption (Bradfeild, 1971). Heart rate is then monitored using a transmitter worn around the chest and a watch-like receiver on the wrist for periods of 24 hours or greater in free-living subjects. The recorded heart rates are then used in conjunction with the regression plot to determine approximate oxygen consumptions rates which can be converted to energy expenditure.
Three laboratories have recently examined the accuracy and precision of heart rate monitoring in a field setting with simultaneous measurement of total energy expenditure using doubly labeled water, a technique which measures total energy expenditure using heavy isotopes of water (Schulz et al., 1989; Livingstone et al., 1990; Goran et al., 1992). Schulz et al. (1989) investigated the use of heart rate monitoring in six young adults (4 male, 2 female) and found no significant difference between total energy expenditure measured with heart rate and that measured with doubly labeled water (mean heart rate = 3413 kcal/day, mean doubly labeled water = 3174 kcal/day).
Livingstone et al. (1990) carried out a very similar experiment comparing FLEX heart rate (measureable heart rate changes between resting and exercising levels) and doubly labeled water in 14 subjects (9 male, 5 female, mean age 32 years). Though there was a wide of range of deviation between the results of doubly labeled water and heart rate results on an individual subject basis (-22.2% to +52.1%), they noted only a 10% difference in nine of the subjects studied. In a cohort of 31 four to six year old children, Goran et al. (1992) found that a measure of random heart rate had strong explainatory power of total energy expenditure. This was found after adjusting the data for fat-free mass and resting metabolic rate. By using these three independent variables (fat free mass, resting metabolic rate and heart rate) 86% of the variance in individual total energy expenditure was explained.
However, there is still some problematic error in using heart rate monitoring for the prediction of total energy expenditure on an individual basis. It has been observed in some individuals that slight changes in heart rate does not affect energy expenditure and therefore, a linear relationship between heart rate and oxygen consumption does not exist (Payne et al., 1971). Even with this in mind it does appears promising that with further study heart rate monitoring may become somewhat of an accurate and acceptable measurement tool for expenditure recording in free-living humans.
That does it for Part I of The Science of Human Energy Expenditure. In Part II of this series we will turn more closely examine the measurement of free living energy expenditure using doubly labeled water.