Of all the techniques currently in use for assessing free-living patterns of total energy expenditure (TEE) and patterns of physical activity, some of the most promising advances have been in the area of heart rate (HR) methodology. It fulfils many of the criteria for providing continuous, indirect, and objective measures of TEE and patterns of physical activity, being relatively inexpensive, simple to use, robust and versatile in a wide variety of field settings, nonintrusive, and well tolerated by subjects. The relative merits of the method are particularly important when studying children because most of the available field techniques are likely to induce behavioral changes in their spontaneous and natural activity patterns.
A major shortcoming of the method is that HR monitoring is not a good predictor of energy expenditure at low levels of activity. However, more recent developments in minute-by-minute HR recorders now negate the need to rely on HR as an index of sedentary EE through identification of an individually predetermined threshold HR (Flex HR) that can be used to discriminate between resting and exercise HR. The Flex HR method, as it is known, has been shown to be valid for estimating group averages in HR TEE although individual estimates lack precision (4,12,13,20).
One of the key factors that affects the integrity of the Flex HR method is the representativeness of the derived HR versus oxygen consumption (V̇O2) regression equation to free-living activity patterns. The individual nature of the HR versus V̇O2 relationship makes it necessary to establish a regression equation for HR versus V̇O2 for each subject at several levels and intensity of activity, while recognizing that factors other than V̇O2, such as ambient temperature, food intake, body posture, and muscle groups exercised can influence HR. Although the need to individually calibrate subjects has been acknowledged and endorsed in successive studies, nevertheless the choice of calibration activities remains a matter of debate.
On the one hand, standard activities such as cycle ergometry and treadmill protocols, either singly or in combination, have been used in adult subjects (4,5,8,10,13,15,17). Alternatively, it has been argued that precision may be improved if calibration activities embrace or simulate the usual range of activities and postures of the subject because this will have greater relevance to daily life (5,6,11). The issue of the appropriateness of different calibration protocols in studies of HR TEE in children has not been addressed. Most calibration protocols in children have involved one type of exercise, and this is usually some type of activity on a treadmill (7,12,14,21–23), or cycle ergometry (2,9,25), or step-testing and jogging (16). It could be argued that such a limited range of activities may not be representative of the daily activity patterns of children particularly because their physical behavior patterns may show greater variation in posture and movement than does adult daily activity.
The aim of the present study is to determine the effect of applying different calibration equations derived from using various combinations of body positions and movements in the estimation of free-living TEE in children by the Flex HR method.
METHODS
Subjects.
Seven male subjects, aged 8–10 yr, volunteered to take part in the present study, which was carried out during the school summer vacation (July–August). Parents of the children were contacted by letter through local primary schools in the Coleraine area. Before the study, the children and the parents were informed of all aspects of the study, and parents gave their written informed consent in each case. The study was approved by the Ethical Committee of the University of Ulster.
Anthropometry.
Body weight, in light indoor clothing and no shoes, was measured to the nearest 0.1 kg (Weylux, Model 824/890, CMS Weighing Equipment, London, U.K.) and height to the nearest 0.1 cm by using a wall mounted stadiometer.
Subject calibration.
Each subject was individually calibrated under standardized conditions to establish the relationship between HR and V̇O2. Calibrations were carried out ≥ 2 h post prandially and after the subject had rested for 30 min after arriving at the laboratory. Calibration points were obtained by simultaneous measurement of HR and V̇O2 for the following eight activities carried out in sequence: lying in the supine position, sitting quietly, standing quietly (the children watched video cartoons during these activities), an arm and upper body exercise that involved reaching for tennis balls, a stooping-and-twisting exercise which involved picking up tennis balls, a stepping test, continuous graded exercising on a treadmill and cycle ergometry. These activities were selected to be representative of a variety of free-living body movements and postures and to give a range of upper-body, lower-body, and twisting trunk measurements.
For the arm exercise that involved reaching for tennis balls, the child sat on a step (22 cm high) facing a desk (72 cm high) and with a chair (48 cm high) positioned on either side of them. A box (18 × 18 × 9 cm) was placed on each chair. During the exercise the child picked up tennis balls from the desk and placed them in boxes using alternate arms. In the stooping-and-twisting exercise, the child stood beside a desk (72 cm high) on which was placed a box (18 × 18 × 9 cm). A similar box full of tennis balls was placed on the floor by their other side. The child was asked to stoop while bending only their knees, pick up a tennis ball, and then twist at the waist in order to place the ball in the container. In the stepping exercise children stepped at 20 steps·min−1 on a block (20 cm high). Exercising on the treadmill was carried out using a continuous protocol at speeds of 2.7 km·h−1 at a 10% gradient, 4.0 km·h−1 at a 12% gradient, and 5.5 km·h−1 at a 14% gradient respectively. Cycling ergometry was carried out at 50 rpm and 25 W. A metronome was used to achieve steady-state conditions during the arm-reaching, stooping-and-twisting, and stepping exercises.
Between each calibration activity, the subjects were allowed to relax while HR returned to basal levels. A 10-min relaxation period between the resting activities (lying, sitting, and standing) and a 15-min period between the exercise activities was allowed during which the subject sat quietly and watched video cartoons. A preliminary stabilization period of 3-min was allowed for each activity followed by a 3-min sampling period except for the treadmill activity where subjects exercised for 3 min at each speed/gradient in a continuous 9-min protocol.
The calibration point for a particular activity was computed as the mean of the 3-min sampling period of the HR and V̇O2. In the case of the treadmill exercise, three calibration points were calculated as the mean HR and V̇O2 from the last 2 min for each of the three speeds/gradients. Respiratory gas exchange in the supine position was measured with a ventilated hood system (Deltratrac Metabolic Monitor, Datex Instrumentation Corporation, Helsinki, Finland), whereas the other activities were measured at 30-s intervals with a Mijnhardt Oxycon-4 system interfaced with a dry-gas meter paramagnetic oxygen analyzer, infrared carbon dioxide analyzer, and electronic microprocessor system (Cardiokinetics Ltd, Medical Diagnostic Instrumentation, Salford, U.K.). Expired air was collected with a two-way nonrebreathing valve and nose clip (Hans Rudolph Inc., Kansas City, MO). Gas analyzers were calibrated with standard gases, and the system was cleared of room air before each measurement started. HR in the supine position was measured using a cardiofrequency meter (Polar Sporttester PE 4000, Polar Electro, Kempele, Finland), while HR during all other respiratory gas collections was monitored with a Cardiokinetics LifeTrace LT-ECG (Albury Instruments Ltd., London, U.K.).
Heart-rate monitoring.
Free-living HR was monitored in all subjects within 2 wk of the calibration procedure using the Polar Sporttester PE 4000 cardiofrequency meter. The location site on the chest was cleansed with an alcohol pad to decrease chances of detachment. The HR transmitter was attached to the chest with two disposable prejelled electrodes and further reinforced with elastoplast to prevent accidental dislodgement. The receiver, which is normally worn on the wrist, was adapted for wear in a waist belt in order to conceal it from direct view and prevent tampering with the function keys. Each child was fitted with the HR instrumentation early in the morning and it was worn continuously until removed by the parent after 12 h of recording time.
HR was recorded at 1-min intervals and data were retrieved in an IBM compatible Polar Advantage Interface System with Polar Precision Software for Windows, Version 5.04 (Polar Electro). Because HR was monitored during the school summer holidays, measurement days were not necessarily consecutive, and specific weekdays or weekend days were not preselected for monitoring. Due to malfunction of some HR recorders, a number of transmissions were lost or were too short to estimate TEE from HR. However, all children produced ≥ 2 d of complete HR recordings.
Derivation of calibration equations.
Seven regression equations were derived for each subject based on the calibration data from a variable combination of activities (Table 1). Calibration equation one (CE1) included all calibration data from the total range of activities undertaken, whereas CE2 was based on the three resting activities (lying, sitting, and standing) and the three exercise activities (stepping, treadmill, and cycle ergometry) that singly or in combination are the most frequently employed exercise activities used to derive calibration equations for the estimation of HR TEE. CE3–CE5 were selected to examine the relative merits of each of the three exercise activities when used in conjunction with the resting activities to derive calibration equations. CE6 was chosen to include a standard weight-dependent lower-body exercise (treadmill) in association with calibration points from upper body movement (arm-reaching, stooping-and-twisting) and resting activities. Finally, CE7 was based on calibration data from a non-weight-dependent lower body exercise (cycle ergometry), the two upper body and trunk exercises and the resting activities.
;)
Table 1:
Exercise combinations used to derive calibration equations.
Estimation of 24 h energy expenditure.
After preliminary editing to remove spurious HR data, TEE was calculated from HR using the Flex HR method, which has been fully described elsewhere (4,12,13,20). This requires the definition of a Flex HR for each subject, above which there is a good correlation between HR and V̇O2 but below which there is a poor correspondence between the two parameters. Flex HR was calculated as the mean of the highest HR for the resting activities (supine, sitting, and standing) and the lowest HR of the exercise activities (12). In the present study, a Flex HR was derived for each of the combination of activities used to establish the seven regression equations.
Energy expenditure for periods during which HR was monitored was calculated as follows. Energy expenditure for periods of daytime when HR ≤ Flex HR was calculated as equivalent to the mean of the resting V̇O2 for the resting activities (lying, sitting, and standing) and referred to as resting energy expenditure (REE). For the remainder of the time when HR was ≥ Flex HR, EE was calculated from the minute-by-minute recorded HR and each subject’s calibration curves and referred to as activity energy expenditure (AEE). Energy expenditure during sleep (SEE) was assumed to be equal to predicted basal metabolic rate (BMR) (19). Twenty-four-hour TEE was computed by summing the SEE, REE, and AEE.
Statistical analysis.
Data are presented as mean ± standard deviation (SD). Between-subject differences in estimates of TEE derived from the regression equations were assessed using ANOVA with least significant differences, at a 5% significance level.
RESULTS
The age and physical characteristics of the subjects are presented in Table 2.
Table 2:
Physical characteristics of the subjects.
Individual and summary estimates of Flex HR, REE, AEE, and TEE are presented in Tables 3 and 4, respectively. In absolute terms, mean TEE ranged from 6.07 ± 0.24 to 8.44 ± 0.57 MJ·d−1, whereas there was a three-fold difference in the range of observed AEE (1.50 ± 0.60 to 4.90 ± 0.79 MJ·d −1). However, when the TEE data are expressed in terms of kJ·kg·d−1 and compared with FAO/WHO/UNU (1985) recommendations (26) for this age group (301 kJ·kg·d−1), it is clear that under these measurement conditions, most of the children were largely sedentary and were expending considerably less energy than is deemed desirable for health.
Table 3:
Individual estimates of REE, AEE, and TEE from the derived calibration equations.
Table 4:
Summary estimates of SEE,a REE,b AEE,b and TEEb for individual subjects.
No significant differences were observed in mean TEE estimates derived from any of the regression equations (Table 5). The observed range (94 ± 8 to 111 ± 8 beats·min−1) in mean between-subject Flex HR derived from each of the calibration curves highlights the sensitivity of Flex HR thresholds to the choice of workload corresponding to the lowest intensity exercise. The lowest mean Flex HR (94 ± 8 beats·min−1) was observed in CE1, 6, and 7 due to the lower HR achieved in the early stages of the arm exercise relative to the lowest HR observed at the start of the stepping, treadmill, and/or cycle ergometry exercises. These exercises (CE2–5) induced a significant (P < 0.05) elevation in Flex HR ranging from approximately 10 beats·min−1 (stepping and/or treadmill exercises) to 17 beats·min−1 (cycle ergometry exercise) relative to the arm exercise.
Table 5:
Summary estimates of Flex HR, REE, AEE, and TEE from the derived calibration equations.
The variation in Flex HR thresholds had an important effect on the relative contribution which REE and AEE made to the estimates of TEE. A lower Flex HR threshold had the effect of increasing the contribution of AEE but decreasing the contribution of REE to the estimate of TEE and vice versa. However, although mean AEE ranged from 2.10 ± 1.18 MJ·d−1 (based on Flex HR 111 ± 8 beats·min−1) to 3.55 ± 1.44 MJ·d−1 (based on a Flex HR 94 ± 8 beats·min−1), none of the differences were significant. The range in observed REE (1.05 ± 0.60 to 1.89 ± 0.62 MJ·d−1) values, corresponding to the lowest and highest defined Flex HR respectively, was significantly different (P < 0.05) only between the minimum (CE1, 6, and 7) and maximum values for REE (CE4).
A typical example of the variation in the slopes and intercepts of the seven calibration curves for one subject (subject 1) is shown in Figure 1. The low CV (4%) between the estimates of TEE of subject 1 (Table 4) may be explained by the fact that the mean daytime HR of this subject was only 106 ± 6 beats·min−1 and 85% of the daytime HR was spent ≤ 140 beats·min−1 (Table 6), which is the HR at which the slopes of the regression lines begin to show marked divergence. Only if a substantially greater proportion of the daytime HR had been spent at ≥ 140 beats·min−1, and particularly in the higher heart rate ranges associated with vigorous physical activity, would larger discrepancies in derived estimates of TEE be expected between each of the regression equations.
Figure 1:
An example of calibration results for subject 1.
Table 6:
Observed heart rate parameters.
This was also the case with the majority of subjects (with the exception of subject 3) because mean daytime HR of the subjects ranged from 85 to 112 beats·min−1, and 82–98% of recorded daytime HR was ≤ 140 beats·min−1 (Table 6). In contrast, the greater variation (CV 9%) between the estimates of TEE observed in subject 3 was associated with a mean daytime HR of 122 beats·min−1 (20 beats·min−1 > mean Flex HR), and, in addition, approximately one third of the daytime HR of this subject was spent ≥ 140 beats·min−1.
DISCUSSION
Heart rate monitoring has become one of the most popular field techniques for assessing TEE and/or patterns of physical activity in children. However, there are a number of unresolved issues that affect the integrity of the Flex HR method, notably the definition of a reproducible Flex HR and the extent to which individually derived HR versus V̇O2 regression equations reflect cardiovascular responses to free-living activity patterns.
Between-subject variability in the slopes of the regression line is well documented, making it necessary to individually calibrate subjects (3,5,11). However, as this and earlier studies (5,11) have confirmed, the selection of activities used in calibration protocols can also have a significant effect on within-subject variability in slopes, and hence could affect estimates of TEE. Most studies in children have tended to favor relatively simple calibration procedures (7,12,14,20,21,23). Only one study to date has used a range of upper- and lower-body activities that were selected to be representative of children’s usual activity patterns (24). Unfortunately, this study was concerned with the validation of minute-by-minute HR recorders for use in children and the issue of the representativeness of the data for estimating TEE was not considered.
In the present study the calibration activities were deliberately selected to simulate a range of postures and movements of varying intensities and which embraced upper- and lower-body activities. However, the entire calibration procedure took approximately 2–2.5 h for each child to complete and, not surprisingly, constant encouragement from the investigator was required to maintain motivation and attention. For this reason, it is debatable whether a lengthy and physically and mentally demanding calibration procedure for subjects is justified, or indeed necessary, given the nonsignificant differences and low CV(%) in within- and between-subject estimates of TEE. As shown in Figure 1 it is only if HR was sustained ≥ 140 beats·min−1 (corresponding approximately to onset of moderate-and-vigorous physical activity) (1) for considerable periods of the day that variation in the slopes of the regression line would result in significant discrepancies in TEE estimates. Given that many children do not now voluntarily engage in moderate-to-high intensity activity for sustained periods (18), it is apparent that the accuracy of HR versus V̇O2 relationship at HR ≥ Flex HR and ≤ 140 beats·min−1 is the most critical if major inaccuracies in the estimation of TEE are to be avoided.
The identification of a reproducible Flex HR to distinguish between resting and activity HR also remains problematic since it is based on the tenuous assumption that one discrete pulse point can provide a clear-cut distinction between rest and exercise. As this study has clearly shown, the definition of Flex HR is very sensitive to the HR generated during the lowest-intensity exercise. In turn, the effect of this variation in Flex HR thresholds on estimates of TEE and the components of TEE is a function of the usual level of physical activity of the subjects. In the largely sedentary subjects involved in the present study, the impact of the observed range in Flex HR on estimates of TEE was negligible. Instead, the major effect was observed in the relative contributions made by REE and AEE to the estimates of TEE.
Consequently, it may be that accurate Flex HR definition in subjects following sedentary lifestyles may not be significant if the objective of the study is to estimate TEE only. In these cases, the accuracy of the REE values for estimation of EE during periods of the day when HR falls below Flex HR will assume critical importance. Similarly, in more active subjects where a greater proportion of daily HR will fall in the active part of the calibration curve, prediction errors in TEE are also likely to depend less on the defined Flex HR than on the representativeness of the derived regression lines. However, definition of an appropriate Flex HR will be important in epidemiological studies where the functional significance of patterns of physical activity is the focus of interest and where Flex HR is used as a lower cut-off for defining the onset of sedentary-to-moderate intensity activity. Spuriously high or low Flex HR will have the effect of under- or over-estimating, respectively, the time spent in sedentary-to-moderate intensity activity.
Unfortunately, due to the limited number of HR monitoring periods for the estimation of HR TEE and without a verification check for accuracy of the HR TEE data against doubly-labeled water estimates of TEE, it is not possible to state if the observed estimates of TEE are representative of usual TEE and/or which calibration activity(s), if any, provided the most appropriate regression equation for the estimation of HR TEE. Furthermore, the study is limited by its small sample size, and the relatively homogeneous patterns of physical activity of the group preclude definitive recommendations about the best way to calibrate HR versus V̇O2 across the range of physical activity patterns
For a number of reasons, the choice of calibration activities for derivation of regression equations will be difficult to resolve. First, even if any of the calibration protocols that were evaluated in this study could be shown to be superior for estimating TEE, within-subject day-to-day variability in the slopes and intercepts of the regression lines would inevitably lead to lack of agreement between duplicate estimates of TEE. Thus, it is more important that the chosen calibration curve should only be determined close to the time when HR is monitored under free-living conditions and repeated as necessary to account for suspected or real changes in nutritional, physiological, physical, medical, or environmental conditions. Second, although the merits of including a wide range of different activities in the calibration procedure are intuitively plausible, this should be tempered with the caution that the demands of an inevitably lengthy protocol may be intolerable for some subjects, particularly children. In any case, this is probably unnecessary because a calibration protocol based on three resting activities (lying, sitting, and standing) and one standard weight-dependent lower-body exercise such as treadmill walking/running or stepping is unlikely to generate significantly greater error than a more inclusive protocol.
Undoubtedly, efforts to improve the accuracy of HR TEE estimates should be ongoing, but the benefits of any additional complexity in the methodology should not detract from the fact that HR monitoring remains one of the most feasible and cost-effective techniques for assessing TEE and/or associated patterns of physical activity under difficult and remote field conditions. Under these circumstances, the selection of calibration activities will inevitably be dictated by the facilities available for measuring respiratory gas exchange and the homogeneity or heterogeneity of habitual physical activity patterns. Finally, even if opinions about the types of activities necessary to calibrate HR versus V̇O2 could be reconciled, it is most unlikely that HR versus V̇O2 relationships established in necessarily contrived situations could capture the cardiorespiratory dynamics associated with spontaneous and complex free-living EE patterns. For these reasons, the desired precision in individual estimates of HR TEE is likely to remain elusive and this constitutes the major, and perhaps intractable problem of this methodology.
Funding source: University of Ulster.
We thank the subjects and their parents for participating in the study. Current addresses of authors: Northern Ireland Center for Diet and Health.
REFERENCES
1. Armstrong, N., J. Balding, P. Gentle, and B. Kirby. Patterns of physical activity among 11 to 16 year old British children. Br. Med. J. 301:203–205, 1992.
2. Bitar, A., M. Vermorel, N. Fellmann, M. Bedu, A. Chamoux, and J. Coudert. Heart rate recording method validated by whole body indirect calorimetry in 10-yr-old children. J. Appl. Physiol. 81:1169–1173, 1996.
3. Bradfield, R. B., P. B. Huntzicker, and G. J. Fruehan. Errors in group regressions for prediction of individual energy expenditure. Am. J. Clin. Nutr. 8:1015–1016, 1970.
4. Ceesay, S. M., A. M. Prentice, K. C. Day, et al. The use of heart rate monitoring in the estimation of energy expenditure: a validation study using indirect whole-body calorimetry. Br. J. Nutr. 61:175–186, 1989.
5. Christensen, C. C., H. M. M. Frey, E. Foenstelien, E. Aadland, and H. E. Refsum. A critical evaluation of energy expenditure estimates based on individual O
2 consumption/heart rate curves and average daily heart rate. Am. J. Clin. Nutr. 37:468–472, 1983.
6. Dauncey, M. J., and W. P. T. James. Assessment of the heart-rate method for determining energy expenditure in man, using a whole-body calorimeter. Br. J. Nutr. 42:1–13, 1979.
7. Emons, H. J. G., D. C. Groenenboom, K. R. Westerterp, and W. H. M. Saris. Comparison of heart rate monitoring combined with indirect calorimetry and the doubly labelled water (
2H
218O) method for the measurement of energy expenditure in children. Eur. J. Appl. Physiol. 65:99–103, 1992.
8. Haskell, W. L., M. C. Yee, A. Evans, and P. J. Irby. Simultaneous measurement of heart rate and body motion to quantitate physical activity. Med. Sci. Sports Exerc. 25:109–115, 1993.
9. Janz, K. F., J. C. Golden, J. R. Hansen, and L. T. Mahoney. Heart rate monitoring of physical activity in children and adolescents: the Muscatine Study. Pediatrics 89:256–261, 1992.
10. Leonard, W. R., P. T. Katzmarzyk, M. A. Stephen, and A. G. P. Ross. Comparison of the heart-rate monitoring and factorial methods: assessment of energy expenditure in highland and coastal Ecuadoreans. Am. J. Clin. Nutr. 61:1146–1152, 1995.
11. Li, R., P. Deurenberg, and J. G. A. J. Hautvast. A critical evaluation of heart rate monitoring to assess energy expenditure in individuals. Am. J. Clin. Nutr. 58:602–607, 1993.
12. Livingstone, M. B. E., W. A. Coward, A. M. Prentice, et al. Daily energy expenditure in free-living children: comparison of heart-rate monitoring with the doubly labelled water (
2H
218O) method. Am. J. Clin. Nutr. 56:343–352, 1992.
13. Livingstone, M. B. E., A. M. Prentice, W. A. Coward, et al. Simultaneous measurement of free-living energy expenditure by the doubly labelled water method and heart-rate monitoring. Am. J. Clin. Nutr. 52:59–65, 1990.
14. Maffeis, C., L. Pinelli, M. Zaffanello, F. Schena, P. Iacumin, and Y. Schutz. Daily energy expenditure in free-living conditions in obese and non-obese children: comparison of doubly labelled water (
2H
218O) method and heart-rate monitoring. Int. J. Obes. 19:671–677, 1995.
15. Meijer, G. A., K. R. Westerterp, H. Koper, and F. Ten Hoor. Assessment of energy expenditure by recording heart rate and body acceleration. Med. Sci. Sports Exerc. 21:343–347, 1989.
16. Panter-Brick, C., A. Todd, R. Baker, and C. Worthman. Comparative study of flex heart rate in three samples of Nepali boys. Am. J. Hum. Biol. 8:653–660, 1996.
17. Racette, S. B., D. A. Schoeller, and R. F. Kushner. Comparison of heart rate and physical activity recall with doubly labelled water in obese women. Med. Sci. Sports Exerc. 27:126–133, 1995.
18. Riddoch, C. J., and C. A. G. Boreham. The health-related physical activity of children. Sports Med. 19:80–102, 1995.
19. Schofield, W. N., C. Schofield, and W. P. T. James. Basal metabolic rate. Hum. Nutr. Clin. Nutr. 39C(Suppl. 1):1–96, 1985.
20. Spurr, G. B., A. M. Prentice, P. R. Murgatroyd, G. R. Goldberg, J. C. Reina, and N. T. Christman. Energy expenditure from minute-by-minute heart-rate recording: comparison with indirect calorimetry. Am. J. Clin. Nutr. 48:552–559, 1988.
21. Spurr, G. B., and J. C. Reina. Marginal malnutrition in school-aged Colombian girls: dietary intervention and daily energy expenditure. Hum. Nutr. Clin. Nutr. 41C:93–104, 1987.
22. Spurr, G. B., and J. C. Reina. Energy expenditure/basal metabolic rate ratios in normal and marginally undernourished Colombian children 6–16 years of age. Eur. J. Clin. Nutr. 43:515–527, 1989.
23. Spurr, G. B., J. C. Reina, and M. Barac-Nieto. Marginal malnutrition in school-aged Colombian boys: metabolic rate and estimated daily energy expenditure.
Am. J. Clin. Nutr. 44: 113–126, 1986.
24. Treiber, F. A., L. Musante, S. Hartdagan, H. Davis, M. Levy, and W. B. Strong. Validation of a heart rate monitor with children in laboratory and field settings. Med. Sci. Sports Exerc. 21:338–342, 1989.
25. van den Berg-Emons, R. J. G., W. H. M. Saris, K.R. Westerterp, and M. A. van Baak. Heart rate monitoring to assess energy expenditure in children with reduced physical activity. Med. Sci. Sports Exerc. 28:496–501, 1996.
26. World Health Organisation. Energy and protein requirements, report of a joint FAO/WHO/UNU expert consultation. Geneva: Technical Report Series 724:92–98, 1985.
