SEARCH

SEARCH BY CITATION

Keywords:

  • exercise;
  • energy expenditure;
  • obesity;
  • heart rate;
  • accelerometry

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Summary/Future Directions
  5. Acknowledgments
  6. References

The number of physical activity measures and indexes used in the human literature is large and may result in some difficulty for the average investigator to choose the most appropriate measure. Accordingly, this review is intended to provide information on the utility and limitations of the various measures. Its primary focus is the objective assessment of free-living physical activity in humans based on physiological and biomechanical methods.

The physical activity measures have been classified into three categories:

  • 1
    . Measures based on energy expenditure or oxygen uptake, such as activity energy expenditure, activity-related time equivalent, physical activity level, physical activity ratio, metabolic equivalent, and a new index of potential interest, daytime physical activity level.
  • 2
    . Measures based on heart rate monitoring, such as net heart rate, physical activity ratio heart rate, physical activity level heart rate, activity-related time equivalent, and daytime physical activity level heart rate.
  • 3
    . Measures based on whole-body accelerometry (counts/U time).

Quantification of the velocity and duration of displacement in outdoor conditions by satellites using the Differential Global Positioning System may constitute a surrogate for physical activity, because walking is the primary activity of man in free-living conditions.

A general outline of the measures and indexes described above is presented in tabular form, along with their respective definition, usual applications, advantages, and shortcomings. A practical example is given with typical values in obese and non-obese subjects. The various factors to be considered in the selection of physical activity methods include experimental goals, sample size, budget, cultural and social/environmental factors, physical burden for the subject, and statistical factors, such as accuracy and precision.

It is concluded that no single current technique is able to quantify all aspects of physical activity under free-living conditions, requiring the use of complementary methods. In the future, physical activity sensors, which are of low-cost, small-sized, and convenient for subjects, investigators, and clinicians, are needed to reliably monitor, during extended periods in free-living situations, small changes in movements and grade as well as duration and intensity of typical physical activities.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Summary/Future Directions
  5. Acknowledgments
  6. References

Between 1991 and 1998, the prevalence of obesity, defined as body mass index (BMI) ≥ 30 kg/m2, increased in the United States by nearly 50% (1), and 55% of adults are now classified as overweight or obese (BMI ≥ 25 kg/m2) (2). The etiology of the rising prevalence of obesity is unclear, although there is increasing evidence to suggest that reduced physical activity may play a major role (3, 4, 5, 6, 7). Accordingly, the Surgeon General has strongly emphasized the importance of and the need to increase the proportion of the general U. S. population who spontaneously engage in physical activity (8). Part of the problem in understanding the role of physical activity in the rising prevalence of obesity is that longitudinal trends survey data provide only self-reported leisure-time physical activity, rather than total daily physical activity (8).

A recent National Institutes of Health Expert Panel expressed concern about failure of past studies to obtain reliable assessments of physical activity (9) due, in part, to difficulties in assessing physical activity under free-living conditions (10). There is clearly a need for new and better tools to assess routine physical activity patterns to assess the impact of environmental changes and of exercise intervention programs on public health. Recognizing the state-of-the-art, the purpose of this overview is two-fold: to describe those measures and indexes currently available for objective assessment of free-living physical activity in humans and to propose new parameters that may be worthy of additional study and refinement in future research.

Presentation and Discussion of the Various Measures

Various systems have been suggested to classify physical activities. In this report, we have chosen to categorize physical activity measures according to whether their method of assessment is based on energy cost, heart rate (HR), or accelerometry. Nonobjective measures based on self-report and questionnaires are not discussed. Currently available measures, which are reported in the literature, are presented below and described in outline form in Table 1. In addition, we have proposed several parameters that may be useful to assess physical activity but that, to our knowledge, have not yet been reported.

Table 1.  Measures currently used to assess physical activity and activity-related EE under free-living and controlled conditions
Method of assessmentAcronymParameter (units)FormulaDefinition/applicationAdvantages/disadvantages
EE or oxygen uptakeAEEActivity EE, net (kcal/d)TEE (kcal/d)− REE (kcal/d)Definition: Activity-related EE. Application: Useful for energy-balance studies to calculate energy requirements due to physical activity using an objective assessment of TEE.TEE usually derived from chamber calorimetry or DLW, relatively accurate but expensive procedures. Provides measure of activity-related EE but is only indirect measure of physical activity because AEE is influenced by subject's body mass and energy economy of movement.
 PALeePhysical activity level (ratio)
  • image
Definition: Index of physical activity-related EE over typical 24-h period. Application: Enables comparisons of average daily physical activity EE levels among individuals or populations.(See above regarding TEE.) Can be used for population studies using estimates of TEE from HR data or average energy intake to maintain body weight. Individual variability in exercise economy not taken into account. Because REE is measured over a short period relative to TEE, small variations in REE may result in disproportionately large variations in estimates of PALee. Sleep duration can alter the ratio (Figure 1).
 METeeMetabolic equivalent (ratio)
  • image
Definition: Relative intensity of a specific physical activity performed in the steady state. Application: Useful for describing intensities of various activities and for prescribing physical activities to patients.Vo2 of specific physical activity is measured by indirect calorimetry, whereas resting Vo2 is usually estimated from standard reference value (i.e., 3.5 mL O2/kg · min).
 PAReePhysical activity ratio
  • image
Definition: Per-minute energy cost of specific activity, relative to per-minute REE. Application: Enables comparisons of energy costs of tasks performed by different persons.Identical to METS except REE is usually measured rather than estimated.
 ARTEeeActivity-related time equivalent (min/d)
  • image
Definition: Index of amount of time spent at EE level equivalent to that of reference activity. 0.9 provides adjustment for average thermic effect of food (10%). Application: Enables comparisons of duration of physical activity between subjects who have different energy costs of movement due to differences in body mass and/or energy economy of exercise.TEE usually derived from chamber calorimetry or DLW, relatively accurate but expensive procedures. Assumes that the reference activity tasks are typical for the study population.
HRHRHR (beats/min)HR response to activityDefinition: Measure of absolute HR response to selected activities. Application: Enables comparisons of absolute HR responses and related energy cost of various activities.HR monitoring is relatively easy, inexpensive, and noninvasive. HR is influenced by factors other than energy demand. To obtain energy cost of activity, HR-Vo2 relationship must be established for each subject for each exercise.
 HRnetNet HR (beats/min)HRnet = (daily average HR [beats/min]− resting HR [beats/min])× 1440 (min/d)Definition: Measure of above-rest HR response to free-living activities. Application: Enables comparisons of average above-resting HR level and related EE of free-living activities.(As above). Resting HR is influenced by nonphysical activity factors, reducing the specificity of above-resting HR responses. HRnet will be influenced by individual differences in economy of performing physical activities; does not reflect intensity or duration of activities performed.
 PARhrPhysical activity ratio
  • image
Definition: Measure of cardiac response to selected activities as a multiple of resting HR. Application: Enables comparisons of relative HR responses of subjects to various activities.(As above.) External stimuli may raise resting HR, giving spuriously low estimates of PAR.
AccelerometrynoneAccelerometry counts (counts/unit time)Average accelerometry counts during activityDefinition: Measure of accelerometric response to selected activities in terms of intensity and duration. Application: Enables comparisons of intensities and duration of locmotion when raw data are provided. Enables comparisons of average EE of locomotion using combination of accelerometry data and EE estimated from standard regression equations based on subject charateristics.Procedure is relatively convenient, inexpensive, and noninvasive. Measures acceleration in 2 to 3 planes. Provides imprecise distinction of small vs. large movements of limbs and movements against a grade. Does not measure static work.

We found that most indices of physical activity describe relationships of two variables as a ratio. It is important to be aware that use of a ratio to adjust biological parameters poses some risk that it may result in a mathematical error. This can occur when the relationship of the numerator to denominator is not linear and does not regress to a zero intercept (11, 12). In such cases, it may be more appropriate to adjust the variable using an approach such as regression analysis.

Measures Based on Energy Expenditure or Oxygen Uptake

Activity energy expenditure (AEE) is a direct measure of the energy cost of physical activity, which is particularly useful in studies of energy balance. AEE may be used to measure the energy cost of a specific exercise task (as kcal/min or kJ/min), or it may be used to estimate a subject's average EE during nonresting periods, using the equation: AEE (kcal/d) = total daily EE (TEE; kcal/d) − resting EE (REE; kcal/d). AEE can also be estimated taking into account an average thermic effect of food of 10%, which is considered in the formula by some investigators: AEE (kcal/d) = 0.9 · TEE (kcal/d) − REE (kcal/d).

AEE is influenced by body weight (impacting the energy cost of moving one's body mass) and by the economy or efficiency of performing exercise tasks. Thus, AEE does not necessarily reflect the intensity of activity performed, and it does not permit comparison of the amounts or duration of activity performed by different individuals.

Because it is difficult to compare values of AEE between individuals of different body mass, in some analyses AEE has been adjusted for body weight (13). A recent refinement of this approach is activity-related time equivalent (ARTEEE), which corrects AEE not only for body weight but also for the economy of performing physical activities (14). ARTEEE is an index of the amount of time a person spends at a level of EE equivalent to that of a reference activity or set of activities, using the equation: ARTEEE index (min/d) = (TEE [kcal/d] × 0.9 − REE [kcal/d])/(reference activity EE [kcal/min] − REE [kcal/min]). The numerator is AEE per day, except that TEE is multiplied by 0.9 to adjust for an average thermic effect of food of 10%. TEE can be assessed using doubly labeled water (DLW) or chamber calorimetry. The denominator is AEE per minute for the reference activity task(s), i.e., the average, above-rest energy cost of performing standardized exercises. Because the energy cost of exercise is measured on each subject, the denominator takes into account the contribution of the subject's weight and economy of movement to the energy cost of performing the tasks. The exercise tasks are performed in the laboratory and are selected to reflect typical activities of subjects in free-living conditions, e.g., walking, stair climbing, walking carrying a small load, etc. ARTEEE is particularly useful for expressing the amount of time each day one expends an amount of energy equivalent to that of the reference activity and enables comparisons among subjects with different body weights and different exercise energy economies, as has been found in black vs. white women (14). The index is expensive to assess and, hence, its usefulness is limited to studies of small groups.

A commonly used alternative to AEE is to express, as a ratio, the energy cost of a sustained exercise task relative to REE. Examples of such EE-based indices include physical activity level (PALEE), metabolic equivalent (METEE), and physical activity ratio (PAREE). PALEE expresses total daily EE relative to basal or REE, thereby providing an index of the average relative excess energy output consequent to physical activity (i.e., intensity × duration) over a typical 24-hour period. TEE is commonly derived using the DLW technique (15). Based on this objective measurement of EE, the validity of PALEE has been tested and confirmed in a large subject sample (16). When derived from DLW data, the index is relatively accurate but expensive to assess and, hence, is best suited for small-group studies dependent on objective measures of EE. A less expensive approach is to estimate TEE using HR data (discussed in the next section). Individual variability in the thermic effect of food and in exercise energy economy is not taken into account in the PALEE index. Furthermore, the denominator, REE, is generally measured over a period of 30 minutes or less, whereas the numerator, TEE, is measured over periods of either 24 hours (using chamber calorimetry) or 1 to 2 weeks (using DLW). Hence, small day-to-day variations in REE will result in disproportionately large variations in the estimates of physical activity level.

For the purposes of large population studies, PALEE has been calculated by dividing total energy intake by an estimate of REE (17). In this approach, energy intake serves as a surrogate of TEE and, hence, assumes that the study subjects are in energy balance when energy intake is assessed. REE is predicted from subject characteristics, including sex, weight, height, and age. PALEE can also be used to estimate total daily energy requirements of a population by assuming an average physical activity level for the group under study (18). For example, a light physical activity level equivalent to a PALEE of 1.56 would predict an average energy requirement of ∼2000 kcal/d for women weighing 55 kg (18). In a related manner, the index has been used to verify the accuracy of self-reported energy intake (19), i.e., below a certain threshold (physical activity level × 1.2), the index suggests that energy intake is underestimated.

Although previously unreported in the literature, a potentially useful variation of the PALEE index is daytime physical activity level (PALEEday), which provides a description of the average intensity of physical activity over the nonsleeping period. PALEEday is based on assessment of EE during daytime awake hours only, excluding the EE of sleep, because the duration and quality of sleep may vary among subjects. PALEE and PALEEday provide potentially useful but different information. Consider an example of an individual with the same level of daytime physical activity but different periods of sleep (case 1 vs. case 2; Figure 1). With the shorter daytime period of activity and the longer duration of sleep, the value of PALEE drops from 1.5 to 1.4. Although accurately reflecting the lower 24-hour average activity level, the reduced PALEE value makes the individual seem to have been less physically active while awake. By contrast, PALEEday gives an activity level that is unchanged at 1.7. For an individual with a longer period of sleep, but who is more active for a shorter daytime period (case 1 vs. case 3), PALEE remains unchanged at 1.5, whereas PALEEday is increased from 1.7 to 1.8, reflecting the more active daytime period. An obvious limitation of the PALEEday index is obtaining data on daytime EE. This would be best derived from chamber calorimetry, using only the daytime period, or from HR monitoring (see PALHRday below). However, whole-body calorimeters are not suited for assessing free-living EE and small portable indirect calorimeters using a face mask to collect expired gas impose a restriction for EE measurements of several hours duration.

image

Figure 1. Examples of how PALEE can provide different types of information than PALEEday. In a situation in which the period of sleep is longer but daytime physical activity is unchanged (case 1 vs. case 2), PALEE indicates a lower activity level, whereas PALEEday remains unchanged. In a situation in which the period of sleep is longer but daytime physical activity is increased (case 1 vs. case 3), PALEE indicates no change in activity level, whereas PALEEday is increased.

Download figure to PowerPoint

The METEE is defined as the ratio of work metabolic rate to a standard resting metabolic rate and describes the relative intensity of exercise tasks performed in the steady state. One METEE is typically considered the resting metabolic rate of a person while sitting quietly (20). The index is generally used by exercise physiologists to express the oxygen uptake or intensity of various activities as multiples of the resting METEE level. It is useful for describing and prescribing exercises of different intensities (21). The Compendium of Physical Activities (22) describes various daily physical activities according to their METEE levels, with activity intensities ranging from 0.9 METEE (sleeping) to 18 METEE (running at 10.9 mph). Such a classification system is particularly useful for epidemiological studies, wherein METEE scores can be ascribed to individuals according to their self-reported physical activity levels and then related to particular health risk outcomes (23). The actual energy cost of an activity will vary between individuals due to a number of factors, such as body mass, adiposity, age, sex, and environmental conditions (22). To allow for variations in body weight, the METEE is generally expressed in terms of oxygen uptake per unit body mass, with 1 METEE equivalent to ∼3.5 mL O2/kg × min (21). Because 5 kcal is approximately equal to 1 L of oxygen consumed, 1 METEE is equivalent to ∼1 kcal/kg × h or 4.184 kJ/kg × h. A more accurate estimate of oxygen uptake can be obtained by indirect calorimetry.

The index PAREE is defined as the per-minute energy cost of performing a physical activity relative to the person's per-minute REE (18). Thus, it is essentially identical to METEE, except that REE is usually measured rather than estimated. Typically, PAREE is used in nutrition studies to compare the energy cost of various physical activities (24).

To demonstrate the types of information and typical values obtained from various physical activity measures, we have ascribed values to each measure based on two hypothetical subjects, one lean and one obese (Table 2). Although the subjects are hypothetical, their physiological patterns are based on data obtained from never-obese and overweight subjects with comparable characteristics studied in our laboratory (14, 25). Among the EE-based physical activity indexes, several results are noteworthy. Unadjusted for body weight, AEE was found to be similar in the two subjects despite the fact that subject 2 is less physically active. In fact, it seems that the majority of obese subjects are moderately active and that an increase in the activity level of obese subjects is limited by the ability to perform exercise of higher intensity.

Table 2.  Representative values obtained for currently available and proposed new measures to assess physical activity: values are based on two hypothetical subjects of the same sex, age, and height, but different weights and activity levels*
  Subject 1 (lighter, more active)Subject 2 (heavier, less active)
Parameter (units) Female, age = 40 yr BMI = 25 kg/m2 (67 kg, 1.63 m) REE = 1480 kcal/d (1.0 kcal/min) 24-hr EE = 2150 kcal/dFemale, age = 40 yr BMI = 30 kg/m2 (80 kg, 1.63 m) REE = 1550 kcal/d (1.1 kcal/min) 24-hr EE = 2200 kcal/d
  • *

    Subject characteristics and physiological data, although hypothetical, are based on actual data of similar subjects studied in our laboratory (14, 25). The reference exercise task used for each parameter = steady-state response to bicycle ergometry at 60 rpm, 50-W workload, 4 minutes. Calculation of METee is based on assumption that 1 MET = 3.5 mL O2/kg ·min.

Parameters based on EE or O2 uptake   
AEEActivity EE (kcal/d or kcal/kg · d)450 kcal/d450 kcal/d
  6.7 kcal/kg/d5.6 kcal/kg/d
PALeePhysical activity level (24-hr EE/REE) (ratio)1.451.42
  (2150/1480)(2200/1550)
PALeedayDaytime physical activity level (daytime EE/REE) (ratio)1.651.58
  (2450/1480)(2450/1550)
METeeMetabolic equivalent [exercise O2 uptake (O2/kg · min)/standard resting O2 uptake (O2/kg · min)] (ratio)3.4 (12.0./3.5)3.4 (12.0/3.5)
PAReePhysical activity ratio (reference exercise EE [kcal/min]/REE [kcal/min]) (ratio)4.04.5
  (4.0/1.0)(5.0/1.1)
ARTEeeActivity-related time equivalent (min/d) (24-hr EE [kcal/d] · 0.9− REE [kcal/d])/(reference exercise EE [kcal/min]− REE [kcal/min])152 min/d110 min/d
  (2150 · 0.9− 1480)/(4.0− 1.0)(2200 · 0.9− 1550)/(5.0− 1.1)
Parameters based on HR   
HRnetNet HR (beats/d) (average 24-hr HR [beats/min]− resting HR [beats/min]) · 1440 min/d21,600 beats/d (80− 65) · 144021,660 beats/d (85− 70) · 1440
PALhrPhysical activity level (24-hr HR/resting HR) (ratio)1.231.21
  (80/65)(85/70)
PALhrdayDaytime physical activity level (daytime HR/resting HR) (ratio)1.311.29
  (85/65)(90/70)
PARhr (METhr)Physical activity ratio (exercise HR/resting HR) (ratio)1.851.86
  (120/65)(130/70)
ARTEhrActivity-related time equivalent (min/d) (24-hr HR [beats/d)− resting HR [beats/d])/(reference exercise HR [beats/min]− resting HR [beats/min])393 min/d360 min/d
  (115,200− 93,600)/(120− 65)(122,400− 100,800)/(130− 70)

The reason for the finding that the AEE is similar in the two subjects is that subject 2, although less active, has a greater body mass. Although this limitation of AEE may be readily apparent, it is a reminder that AEE is best used to compare activity-related EE rather than physical activity levels. Other parameters, such as PALEE and PALEEday, provide a better comparison of activity levels and, as shown in Table 2, they reflected a greater activity level in the more active subject.

METEE and PAREE were used to compare the responses of the hypothetical subjects to the standardized cycle ergometer exercise task. Estimated METEE levels, expressed relative to body weight, were comparable at 3.4 times the standard resting oxygen uptake. By contrast, PAREE levels, which were based on actual measured responses to the cycle ergometry task performed in the laboratory but unadjusted for body weight, indicated that the heavier subject had a greater energy cost of performing the exercise task than the lighter subject and, hence, had a higher PAREE value. Finally, the ARTEEE index indicated that the lighter more active subject spent an average of 42 more minutes each day (152 vs. 110 min/d) at an EE level equivalent to riding the cycle ergometer in the laboratory (∼3 METEE). This type of information may be useful if an investigator wishes to determine how much time a person spends on an average day maintaining an EE of physical activity comparable to that prescribed by an investigator or health agency.

Measures Based on HR

Although variable from one person to the next, within individuals HR and oxygen uptake tend to be linearly related throughout a wide range of aerobic exercise tasks (26). When this relationship is known for a particular subject, recordings of HR can be used to estimate the person's oxygen consumption and, in turn, EE in free-living conditions (27, 28). HR and oxygen uptake have been shown to be moderately correlated during field and laboratory activities, with HR accounting for nearly 50% of the variability in oxygen uptake (29). Use of HR monitors to estimate EE is relatively inexpensive, convenient, noninvasive, and versatile. HR monitoring has been used with increased frequency in recent years, facilitated by availability of low-cost, portable monitors that are capable of measuring and storing minute-by-minute data over several hours and averaged data over days or weeks. An advantage of minute-by-minute HR monitoring is that it permits preselection of a threshold HR above the sedentary level, obviating a well-recognized limitation of HR data, i.e., that HR is not a good predictor of EE at low levels of physical activity (26). In addition, minute-by-minute HR data provide information about the frequency, intensity, and duration of free-living physical activities (28, 30).

Estimates of PALEE can be obtained relatively inexpensively by obtaining total daily EE derived from HR monitoring data in free-living conditions, based on the individual regression of HR against oxygen uptake measured in the laboratory (30, 31). Differences in levels of fitness can substantially affect the slope of the regression between HR and EE among individuals. For example, a trained person can be expected to have a lower HR for the same level of oxygen consumption. However, such differences should not preclude use of this measure to estimate physical activity-related EE. Perhaps the major limitation of assessing physical activity based on HR data is the lack of established relationships between HR and energy cost of the wide variety of activities encountered in daily living. An additional limitation is the confounding effect of factors other than energy demand on the HR response to exercise. Confounding factors include ambient conditions, time of day, emotional state, hydration status, food and caffeine intake, smoking, previous activity, body position, muscle groups used, and the static vs. dynamic use of limbs (26, 32, 33, 34). As a result, HR data obtained within individuals may yield large variability and unreliable estimates of EE (35, 36, 37, 38). However, when applied to groups of individuals, the HR-monitoring technique provides an acceptable estimate of TEE and associated patterns of physical activity (26).

Another simple and relatively accurate HR-based method to assess EE in free-living conditions is net HR (HRnet) (29). HRnet parallels AEE in that resting HR is subtracted from total daily average HR using an equation such as: HRnet (beats/d) = (daily average HR [beats/min] − resting HR [beats/min]) × 1440 (min/d). Obtained using a HR monitor, HRnet provides a means to compare individuals’ average, above-rest HR level during routine daily activities. HRnet would be expected to have the same limitations as AEE. That is, HRnet will be influenced by individual differences in HR responses to the same activities and will not necessarily reflect the duration or intensity of the activities performed.

The physical activity ratio based on HR data (PARHR) is an index of a subject's HR response to a selected physical activity divided by the resting HR and enables comparisons among subjects of the intensities of their responses to specific exercise tasks. Other indices based on HR responses, which, to our knowledge, have not yet been reported in the literature but which are potentially useful and worthy of additional study, are physical activity level (PALHR), daytime physical activity level (PALHRday), and activity-related time equivalent (ARTEHR). As a parallel measure to PALEE, which is based on EE data, PALHR is a ratio of a person's average daily HR and resting HR. Similarly, PALHRday is a measure comparable to PALEEday, i.e., a person's average HR is measured during daytime awake hours and expressed relative to resting HR. To assess physical activity during the daytime, Wareham et al. (30) used an approach somewhat similar to PALHRday by expressing the percentage of daytime hours in which the physical activity level was at a predetermined level above basal EE ARTEHR is a measure parallel to ARTEEE and, potentially, can be used as an index of the amount of time a person sustains a HR equivalent to that of a reference exercise task. ARTEHR refines the HRnet parameter by comparing one's average daily above-rest HR to the above-rest HR response to a standard reference exercise task. That is:

  • image

where total daily HR = the subject's average HR throughout a 24-hour period and exercise HR = the subject's mean, steady-state HR response to a standardized exercise task performed in the laboratory (e.g., walking on a treadmill).

Shown in Table 2 are values for the two hypothetical subjects for parameters based on HR data. Several findings are noteworthy. As in the case of AEE, HRnet was found to be similar in the two subjects, despite the fact that one is more physically active. Although a measure of each subject's above-rest average daily HR pattern in free-living conditions, this parameter does not necessarily reflect the level of physical activity because it is influenced by nonactivity-related factors such as body mass. PALHR and PALHRday described average 24-hour activity levels in the range of 1.23 to 1.31, which were considerably below the levels of 1.42 to 1.65 described by their EE-based counterparts PALEE and PALEEday. This is most likely explained by the fact that there is a considerably lower ceiling for relative intensity parameters based on HR data than those based on EE data, because HR cannot vary as much as EE. For example, in a young individual the range of HR responses to an exercise task is a maximum of ∼4–5-fold that of resting HR. The range of EE responses to the same task in a very fit individual may be ∼20-fold that of basal EE. In part, this is reflected in the responses of the hypothetical subjects to the cycle ergometry task, wherein the PAREE value was more than twice the value obtained using PARHR.

Measures Based on Accelerometry

Accelerometry devices measure the rate (i.e., intensity) of body movement in up to three planes (i.e., anterior–posterior, lateral, and vertical). When the devices provide raw data, investigators can estimate the relative intensities as well as duration of various physical activities. Because laboratory studies have demonstrated linear relationships between accelerometry counts and EE during physical activities such as walking and running (39, 40), the energy cost of such activities can be obtained from accelerometry data. Various activities then can be classified according to their intensity and can be expressed using METEE levels (40). To estimate TEE of individuals in free-living situations, many manufacturers incorporate into their devices computer programs that convert into EE the sum of the measured accelerometry counts and predict REE from standard regression equations based on the subject's characteristics of age, height, weight, and sex.

Advantages of accelerometry devices include their small size (permitting subjects to wear the monitors without interfering with normal movement) and their ability to record data continuously for periods of days, weeks, and even months (41). Models with internal real-time clocks also help discriminate activity patterns (42). A notable limitation of most currently available accelerometers is their inability to detect the additional energy cost of upper body movement (unless sensors are placed on the upper limbs), load carriage (static work), or moving on soft or graded terrain (39, 43, 44). Basically, their accuracy is limited to assessing the energy cost of dynamic work, such as locomotion on the level. Not surprisingly, then, accelerometry data demonstrate a better relationship with walking than with other common household activities, such as house cleaning and yard work, or recreational activities, such as playing golf (43). Finally, because the relationship of accelerometry data to EE is dependent on the type of activity performed, estimates of the energy cost of tasks performed in the laboratory may not apply to activities performed under free-living conditions (43).

In summary, ambulatory accelerometers enable objective assessment of total physical activity and can well distinguish differences in activity levels (even of low magnitude) among individuals and within given individuals. In addition, it can assess the effect of lifestyle interventions on physical activity. Clinical prescriptions for increasing physical activity level and/or duration can also benefit from inconspicuous assessment of physical activity using accelerometers. Finally, a change in AEE (or TEE) can be also picked up quantitatively, provided the acceleration signals can be properly converted by an adequate algorithm into EE.

As a new index that has not yet been described in the literature, ARTEaccel is a potentially useful extrapolation of activity-related time equivalent (ARTE)EE, the difference being that ARTEaccel is based on accelerometry-derived data. ARTEaccel is an index of the amount of time a person spends with accelerometric patterns equivalent to that of a reference exercise task. That is:

  • image

where total daily ACCL = the subject's average number of accelerometry counts during a 24-hour period and exercise ACCL = the subject's mean, steady-state accelerometry count response to a standardized exercise task in the laboratory (e.g., level treadmill walking).

The ARTEaccel index has the potential advantage over an unadjusted accelerometry reading in that it would enable comparison of physical activity duration among subjects who differ in their biomechanical efficiency of movement. The index may be worthy of investigation to demonstrate whether it appropriately and usefully corrects for interindividual variations in accelerometric responses to specified activity tasks.

Summary/Future Directions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Summary/Future Directions
  5. Acknowledgments
  6. References

In view of the large number of measures and indexes currently used to estimate physical activity in humans, it is important to maintain a perspective of their advantages and disadvantages according to their derivation (i.e., based on EE, HR, or accelerometry data), their convenience and cost, and their intended use (e.g., to assess impact of activity on cardiovascular fitness or on energy balance, in small-scale or large-scale studies). Generally, it should be noted that there is a definite paucity of studies about the validity and reliability of these indexes but this depends on the circumstances under which they are used. Surprisingly, there is very little information on objectively measured changes in physical activity after changes in body weight. In fact, misleading information may be obtained even with some currently available measures. For example, from recent findings, it seems that across BMIs from 25 to 35 kg/m2 the PAL index remains essentially constant (1.65 to 1.73) in women but tends to decrease in men (45). The interpretation that this reflects similar levels of physical activity in women and lower levels of activity in men at higher BMIs seems logical (45). However, it could be challenged on the premise that changes in body weight exert disproportionate influences on the denominator of the index (i.e., REE, which is essentially determined by fat-free mass) and the numerator of the index (i.e., TEE, which is dependent on body weight and amount of activity). In a situation in which body weight is lost without significant loss of fat-free mass, REE will decrease only modestly, if at all, whereas the reduced body weight will lower total EE (i.e., AEE) despite an identical activity level. As a result, the PAL ratio will fall after weight loss despite no actual change in physical activity. The reverse situation may occur if large losses in fat-free mass were to occur, such as during rapid weight loss. Although these are theoretical situations, conceivably they could mask real changes in physical activity expected with weight change. Available measures should provide data that are not only valid, but also are representative of activities performed in real life. Each index provides different information, and no single parameter covers all of these aspects simultaneously. For example, if it is important to quantitate the average level of physical activity and the amount of work performed by a person involved in a weight-control program, a parameter that estimates average daily activity-related EE may suffice (i.e., PAL day). To compare activity-related EE among subjects of different body sizes may require a parameter that adjusts the data for individual differences in REE and, perhaps, economy of movement. By contrast, if it is important to know if a person is achieving threshold levels of daily exercise to improve cardiovascular fitness, then a parameter that provides a minute-by-minute profile of the intensity and duration of free-living physical activities may be required.

The energy expended by a person during a selected physical activity depends on the duration, the intensity, and the economy or work with which the activity is performed. Thus, EE represents a composite value influenced by several factors. To our knowledge, currently, there is no single device that can provide objective, accurate, noninvasive, minute-by-minute profiles of EE (not physical activity), without carrying a portable calorimeter and wearing a face mask or a mouthpiece.

The method of choice will depend on how the measurement will be used. For example, if a rough estimate of the physical activity level of a population in an epidemiological study is the desired outcome, then use of a simple pedometer (46) or electronic pedometer (47) may be sufficient. If patterns and intensity of activity are needed, then an accelerometer may be better suited for the study. Misperception of habitual physical activity could provide invalid data with the measures and indexes discussed. This particular situation can occur primarily with questionnaires and nonobjective measurements. However, we cannot exclude the possibility that accelerometers may not be worn properly during the course of the day and that erroneous data may indeed be obtained if the subject were to interfere with the sensor (i.e., shake it for a few minutes). This kind of behavior is likely to be rare, but such possibilities cannot be excluded.

Indirect estimates of free-living EE based on HR data are affected by confounding factors (i.e., nonexercise factors that influence HR responses), and estimates based on acceleration data do not include the energy cost of work not measured by accelerometry (e.g., static work or walking on an incline). Because the most common free-living physical activity is walking on various grades, a proxy measure of the intensity of locomotion may be derived from the velocity of displacement (walking and running) assessed by the satellites’ global positioning system in differential mode (48). Used in conjunction with new portable devices, it can be used to assess the duration and intensity of typical free-living activities on nonlevel surfaces (49) and to enable study of the biomechanical aspects of movement (50).

In the future, efforts should be directed toward developing motion sensors and methods that are less expensive, such as pedometers (51), more convenient, and more reliable to assist researchers and clinicians in assessing and prescribing free-living physical activity. It is hoped that this review will also serve as a call for more research in this important area.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Summary/Future Directions
  5. Acknowledgments
  6. References

This work was supported by National Institutes of Health Grant R01-DK51684 and Clinical Nutrition Research Unit Grant P30-DK56336. Y.S. was supported by Swiss National Science Foundation Grant 32-55928.98.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Summary/Future Directions
  5. Acknowledgments
  6. References
  • 1
    Mokdad, A. H., Serdula, M. K., Dietz, W. H., Bowman, B. A., Marks, J. S., Koplan, J. P. (1999) The spread of the obesity epidemic in the United States, 1991–1998. JAMA 282: 15191522.
  • 2
    Flegal, K. M., Carrol, M. D., Kuczmarski, R. J., et al (1998) Overweight and obesity in the United States: prevalence and trends, 1960–1994. Int J Obes Relat Metab Disord 22: 3947.
  • 3
    Rissanen, A. M., Heliovaara, M., Knekt, P., Reunanen, A., Aromaa, A. (1991) Determinants of weight gain and overweight in adult Finns. Eur J Clin Nutr 45: 419430.
  • 4
    Poehlman, E. T., Toth, M. J., Bunyard, L. B., et al (1995) Physiological predictors of increasing total and central adiposity in aging men and women. Arch Intern Med 155: 24432448.
  • 5
    Weinsier, R. L., Heini, A. F., Hunter, G. R., Goran, M. I., Sell, S. M. (1998) Etiology of obesity: relative contribution of metabolic factors, diet and physical activity. Am J Med 105: 145150.
  • 6
    Hill, J. O., Melanson, E. L. (1999) Overview of the determinants of overweight and obesity: current evidence and research issues. Med Sci Sports Exerc 31(suppl): S515SS21.
  • 7
    Samaras, K., Kelly, P. J., Chiano, M. N., Spector, T. D., Campbell, L. V. (1999) Genetic and environmental influences on total-body and central abdominal fat: the effect of physical activity in female twins. Ann Intern Med 130: 873882.
  • 8
    U. S. Department of Health and Human Services. (1996) Physical Activity and Health: A Report of the Surgeon General. Atlanta, GA: U. S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion.
  • 9
    Wing, R. R. (1999) Physical activity in the treatment of the adulthood overweight and obesity: current evidence and research issues. Med Sci Sports Exerc 31(suppl): S547S552.
  • 10
    Meijer, G. A., Westerterp, K. R., Verhoeven, F. M., Koper, H. B., Ten Hoor, F. (1991) Methods to assess physical activity with special reference to motion sensors and accelerometers. IEEE Trans Biomed Eng 38: 2219.
  • 11
    Tanner, J. M. (1949) Fallacy of per-weight and per-surface area standards, and their relation to spurious correlation. J Appl Physiol 2: 115.
  • 12
    Katch, V. L. (1973) Use of the oxygen/body weight ratio in correlation analyses: spurious correlations and statistical considerations. Med Sci Sports Exerc 5: 253257.
  • 13
    Esparza, J., Fox, C., Harper, I. T., et al (2000) Daily energy expenditure in Mexican and USA Pima Indians: low physical activity as a possible cause of obesity. Int J Obes Relat Metab Disord 24: 5559.
  • 14
    Weinsier, R. L., Hunter, G. R., Zuckerman, P. A., et al (2000) Energy expenditure and free-living physical activity in black and white women: comparison before and after weight loss. Am J Clin Nutr 71: 11381146.
  • 15
    Westerterp, K. R. (1998) Alterations in energy balance with exercise. Am J Clin Nutr 68: 970S974S.
  • 16
    Black, A. E., Coward, W. A., Cole, T. J., Prentice, A. M. (1996) Human energy expenditure in affluent societies: an analysis of 574 doubly-labeled water measurements. Eur J Clin Nutr 50: 7292.
  • 17
    Stam-Moraga, M. C., Kolanowski, J., Dramaix, M., de Backer, G., Kornitzer, M. D. (1999) Sociodemographic and nutritional determinants of obesity in Belgium. Int J Obes Relat Metab Disord 23 (suppl): 19.
  • 18
    James, W. P. T., Schofield, E. C., eds. (1990) Human Energy Requirements: A Manual for Planners and Nutritionists Oxford University Press Oxford, England.
  • 19
    Black, A. E. (2000) Critical evaluation of energy intake using the Goldberg cut-off for energy intake: basal metabolic rate: a practical guide to its calculation, use and limitations. Int J Obes Relat Metab Disord 24: 11191130.
  • 20
    McArdle, W. D., Katch, F. I., Katch, V. L., eds.. (1996) Exercise Physiology: Energy, Nutrition, and Human Performance 4th ed., Williams & Wilkins Baltimore, MD.
  • 21
    Jette, M., Sidney, K., Blumchen, G. (1990) Metabolic equivalents (METS) in exercise testing, exercise prescription, and evaluation of functional capacity. Clin Cardiol 13: 555565.
  • 22
    Ainsworth, B. E., Haskell, W. L., Whitt, M. C., et al (2000) Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 9(suppl): S498S516.
  • 23
    Manson, J. E., Hu, F. B., Rich-Edwards, J. W., et al (1999) A prospective study of walking as compared with vigorous exercise in the prevention of coronary heart disease in women. N Engl J Med 341: 650658.
  • 24
    McGandy, R. B., Barrows, C. H. Jr, Spanias, A., Meredith, A., Stone, J. L., Norris, A. H. (1966) Nutrient intakes and energy expenditure in men of different ages. J Gerontol 21: 581587.
  • 25
    Hunter, G. R., Weinsier, R. L., Darnell, B. E., Zuckerman, P., Goran, M. I. (2000) Racial differences in energy expenditure and aerobic fitness in premenopausal women. Am J Clin Nutr 71: 500506.
  • 26
    Livingstone, M. B. E. (1997) Heart-rate monitoring: the answer for assessing energy expenditure and physical activity in population studies?. Br J Nutr 78: 869871.
  • 27
    Payne, P. R., Wheeler, E. F., Salvosa, C. B. (1971) Prediction of daily energy expenditure from average pulse rate. Am J Clin Nutr 24: 11641170.
  • 28
    Spurr, G. B., Prentice, A. M., Murgatroyd, P. R., Goldberg, G. R., Reina, J. C., Christman, N. T. (1988) Energy expenditure from minute-by-minute heart-rate recording: comparison with indirect calorimetry. Am J Clin Nutr 48: 552559.
  • 29
    Strath, S. J., Swartz, A. M., Bassett, D. R. Jr, O'Brien, W. L., King, G. A., Ainsworth, B. E. (2000) Evaluation of heart rate as a method for assessing moderate intensity physical activity. Med Sci Sports Exerc 32(suppl): S465SS70.
  • 30
    Wareham, N. J., Hennings, S. J., Prentice, A. M., Day, N. E. (1997) Feasibility of heart-rate monitoring to estimate total level and pattern of energy expenditure in a population-based epidemiological study: the Ely Young Cohort Feasibility Study 1994–1995. Brit J Nutr 78: 889900.
  • 31
    Ekelund, U., Sjöström, M., Yngve, A., Nilsson, A. (2000) Total daily energy expenditure and pattern of physical activity measured by minute-by-minute heart rate monitoring in 14–15 year old Swedish adolescents. Eur J Clin Nutr 54: 195202.
  • 32
    Montoye, H. J., Taylor, H. L. (1984) Measurement of physical activity in population studies: a review. Hum Biol 56: 195216.
  • 33
    Parker, S. B., Hurley, B. F., Hanlon, D. P., Vaccaro, P. (1989) Failure of target heart rate to accurately monitor intensity during aerobic dance. Med Sci Sports Exerc 21: 230234.
  • 34
    Maas, S., Kok, M. L., Westra, H. G., Kemper, H. C. (1989) The validity of the use of heart rate in estimating oxygen consumption in static and in combined static/dynamic exercise. Ergonomics 32: 141148.
  • 35
    Maffeis, C., Pinelli, L., Zaffanello, M., Schena, F., Iacumin, P., Schutz, Y. (1995) Daily energy expenditure in free-living conditions in obese and non-obese children: comparison of doubly labeled water (2H218O) method and heart-rate. Int J Obes Relat Metab Disord 19: 671677.
  • 36
    Rutgers, C. J., Klijn, M. J., Deurenberg, P. (1997) The assessment of 24-hour energy expenditure in elderly women by minute-by-minute heart rate monitoring. Ann Nutr Metab 41: 8388.
  • 37
    Davidson, L., McNeill, G., Haggarty, P., Smith, J. S., Franklin, M. F. (1997) Free-living energy expenditure of adult men assessed by continuous heart-rate monitoring and doubly-labeled water. Br J Nutr 78: 695708.
  • 38
    McCrory, M. A., Molé, P. A., Nommsen-Rivers, L. A., Dewey, K. G. (1997) Between-day and within-day variability in the relation between heart rate and oxygen consumption: effect on the estimation of energy expenditure by heart-rate monitoring. Am J Clin Nutr 66: 1825.
  • 39
    Bouten, C. V., Westerterp, K. R., Verduin, M., Janssen, J. D. (1994) Assessment of energy expenditure for physical activity using a triaxial accelerometer. Med Sci Sports Exerc 26: 15161523.
  • 40
    Freedson, P. S., Melanson, E., Sirard, J. (1998) Calibration of the Computer Science and Applications, Inc. accelerometer. Med Sci Sports Exerc 30: 77781.
  • 41
    Westerterp, K. R. (1999) Assessment of physical activity in relation to obesity: current evidence and research issues. Med Sci Sports Exerc 31(suppl): S522S5S5.
  • 42
    Melanson, E. L. Jr, Freedson, PS. (1996) Physical activity assessment: a review of methods. Crit Rev Food Sci Nutr 36: 385396.
  • 43
    Hendelman, D., Miller, K., Baggett, C., Debold, E., Freedson, P. (2000) Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Med Sci Sports Exerc 32(suppl): S442S449.
  • 44
    Sherman, W. M., Morris, D. M., Kirby, T. E., et al (1998) Evaluation of a commercial accelerometer (Tritrac-R3 D) to measure energy expenditure during ambulation. Int J Sports Med 19: 4347.
  • 45
    Prentice, A. M., Black, A. E., Coward, W. A., Cole, T. J. (1996) Energy expenditure in overweight and obese adults in affluent societies: an analysis of 319 doubly-labeled water measurements. Eur J Clin Nutr 50: 9397.
  • 46
    Sequeira, M. M., Rickenbach, M., Wietlisbach, V., Tullen, B., Schutz, Y. (1995) Physical activity assessment using a pedometer and its comparison with a questionnaire in a large population survey. Am J Epidemiol 142: 989999.
  • 47
    Bassett, D. R., Ainsworth, B. E., Swartz, A. M., Strath, S. J., O'Brien, W. L., King, G. A. (2000) Validity of four motion sensors in measuring moderate intensity physical activity. Med Sci Sports Exerc 32(suppl 9): S471SS80.
  • 48
    Schutz, Y., Herren, R. (1999) Assessment of speed of human locomotion using a differential satellite global positioning system. Med Sci Sports Exerc 32: 642646.
  • 49
    Perrin, O., Terrier, P., Ladetto, Q., Merminod, B., Schutz, Y. (2000) Improvement of walking speed prediction by accelerometry and altimetry, validated by satellite positioning. Med Biol Eng Comput 38: 164168.
  • 50
    Terrier, P., Ladetto, Q. Merminod, B., Schutz, Y. (2000) High-precision satellite positioning system as a new tool to study the biomechanics of human locomotion. J Biomech 33: 17171722.
  • 51
    Freedson, P. S., Miller, K. (2000) Objective monitoring of physical activity using motion sensors and heart rate. Res Q Exerc Sport 71(suppl 2): S21SS9.