Physical activity is defined as any bodily movement achieved by contraction of skeletal muscles that increases energy expenditure above resting level (3). Hence, there is a number of different movement patterns that can be performed by humans defined as physical activity. Even if the number of movement patterns are reduced to those commonly performed (walking, running, bicycling, sporting, working), there is still a challenge for the physical activity assessment method to discriminate between the different patterns. Children are recommended to reach 60 min of physical activity, preferably on all days of the week (28). There are a number of ways to reach this recommendation, and, thus, the method trying to assess the compliance in the population has to identify and discriminate between all physical activities.
Objective assessment of physical activity has expanded as a consequence of technical developments and has made it possible for continuous assessment of physical activity for longer periods (7-14 d). There are two ways that objective assessment of physical activity has been performed, either by registration of body movements (pedometers and accelerometers), or the physiological consequences of them (heat loss, oxygen consumption/carbon dioxide, and heart rate). Direct calorimetry, indirect calorimetry, and the doubly labeled water method are considered as the most accurate methods to assess physical activity by measuring heat loss or oxygen consumption/carbon dioxide and translating it into energy expenditure. However, because of their cost and technical demands, they are limited to small, experimental studies, but they can serve as criterion methods in validation studies. Pedometers, accelerometers, and heart rate monitors are commonly used for the assessment of physical activity because of their lower cost, ease of use, and availability. Pedometers register the number of steps taken, accelerometers register acceleration of the center of body mass, and heart rate monitors register the heart rate response to the intensity of the physical activity. However, pedometers and accelerometers are limited to a few movement patterns, and at low intensities, the heart rate could be raised by emotions like anxiety, increase in body temperature, or as a postexercise response lag without an associated increase in energy expenditure. Combining accelerometry with heart rate monitoring has increased the accuracy in assessing physical activity (4,25).
A new generation of monitors that either combines multiple accelerometers on different body segments or that combines accelerometry with other physiological signals in a single device has contributed to a progress in the physical activity assessment field. The SenseWear Armband (BodyMedia, Inc., Pittsburgh, PA) combines five different sensors into one device attached as an armband around the upper arm. This design has increased the wearability, which is especially important when assessing physical activity in children. A computer software, Innerview Professonal software, applies activity-specific algorithms for the calculation of energy expenditure based on analysis of the pattern of signals from the sensors. Validation studies in adults have resulted in promising results (8,12,13), where the SenseWear Armband showed higher accuracy compared with uniaxial and multiaxial accelerometers (13). No study has yet been published showing the validity of SenseWear Armband in children. Originally, the algorithms were based on adult data. However, the latest version of the software, InnerView Professional software (version 5.1), includes algorithms for children based on data from the activities resting, walking, and stationary biking (children and adolescents 6-17 yr old).
Metabolic equivalents (METs) represent the energy cost of physical activities as multiples of resting metabolic rate. A comprehensive list of activities and their MET values have been compiled for adults (1), promoting comparison of activity intensity levels across studies. There is a scarcity of energy cost data of physical activities in children. MET values assessed in children differ from those assessed in adults for many physical activities, and the energy value of 1 MET is greater in children than adults (10). A list of physical activities and their MET values in children may improve the accuracy and comparability when assessing energy expenditure in children using subjective methods (direct observation, interviews, diaries, and questionnaires).
The primary aim of the study was to examine the validity of SenseWear Pro2 Armband to assess energy cost during rest and during different physical activities representing both simple and more complex movement patterns performed by children. A secondary aim was to contribute with values of energy cost and MET in an overview of physical activities in children.
Study Design and Subjects
Energy cost assessed during rest and during different physical activities was compared between SenseWear Pro2 Armband (SWA) and a portable metabolic system, Oxycon Mobile (OM), measuring V˙O2 and V˙CO2. OM was used as criterion method for energy expenditure in this study. Twenty healthy children, 11 boys and 9 girls ages 11-13, were recruited from an ongoing study of physical activity and aerobic fitness in children and adolescents in Göteborg, Sweden. They were selected because they had acceptable aerobic fitness to be able to perform all activities in the present study, at least 46 mL·kg−1 for boys and 41 mL·kg−1 for girls (23). Aerobic fitness (V˙O2peak) was assessed from the results of a cardiopulmonary exercise test on a stationary bike performed in the ongoing study of physical activity and fitness in children and adolescents in Göteborg. Written informed consent was received from the children and their parents. The study was approved by the ethics research committee of the Göteborg University.
Assessment of Energy Cost
SenseWear Pro2 Armband.
SenseWear Pro2 Armband is worn on the back of the upper arm and is a multiple-sensor device collecting data from a skin temperature sensor, near-body temperature sensor, heat flux sensor, galvanic skin response sensor, and a biaxial accelerometer. The skin temperature sensor and near-body temperature sensor (a vent on the side of the armband) consist of sensitive thermistors in contact to the skin relying on change in resistance with changing temperature. The heat flux sensor uses the difference between skin temperature and near-body temperature to assess heat loss. The galvanic skin response sensor measures the conductivity of the skin between two electrodes in contact to the skin. The conductivity of the skin varies according to physical and emotional stimuli. The biaxial accelerometer registers the movement of the upper arm and provides information about body position. The information from the sensors, together with gender, age, height, and weight, are incorporated into proprietary algorithms to estimate energy expenditure. These algorithms are activity specific and are automatically applied on the basis of an analysis of the pattern of signals from the sensors. SWA (version 6.03) was worn on the right arm over the triceps muscle. It was placed on the arm of the subject while in a seated position 10 min before data collection. Energy expenditure was calculated at 1-min intervals, using InnerView Professional software (version 5.1), including data from all sensors, together with gender, age, body weight, and height.
Oxycon Mobile (VIASYS Healthcare, Conshohocken, PA) is a battery-operated, portable, wireless metabolic system measuring gas exchange breath-by-breath and attached to the body in a vest system (18). A flow sensor unit is connected to a face mask (Hans Rudolf, Inc., Kansas City, MO) detecting the air flow by the rotation of a low-resistance, bidirectional turbine; it allows determination of ventilation. Via a sampling line connected to the flow sensor unit, the expired air is analyzed for O2 and CO2 concentrations in a sensor box using a microfuel cell and thermal conductivity, respectively. A data-exchange unit collects the data and sends them telemetrically to a base station connected to a computer. After 30 min of warm-up time and immediately before data collection, a two-point (0.2 and 2.0 L·s−1) air flow calibration was performed using the automatic flow calibrator, and the gas analyzers were calibrated against a certified gas mixture of 5% O2, 16% CO2, and 79% N2 (Reissner-Gase Gmbh and Co, Lichtenfels, Germany), together with determination of measurement delay time. Energy expenditure was calculated from the gas-exchange data using the Weir equation (30): EE (kJ) = 4.184 (3.9V˙O2 + 1.1V˙CO2).
The subject arrived to the laboratory in a motor vehicle in an at least 3 h postprandially. Body weight was determined to the nearest 0.1 kg using a digital scale, and body height was determined to the nearest centimeter, using a horizontal headboard with an attached, wall-mounted metric rule. SWA and OM were time synchronized and attached to the body. After a short instruction and adaptation to the equipment, the subject lay down on a bed, listening to soft music for 30 min, for assessment of resting energy expenditure (REE) (Table 1). During a 90-min break when the equipment was detached from the body, the subjects were offered a small meal, followed by instructions and time for getting acquainted to the activities included in the study. SWA and OM were again attached to the body. After 10 min in sitting position, the subject performed five different activities of 5-min duration each, separated by 5 min in sitting position: 1) sitting, playing games on a mobile phone; 2) stepping up and down on an 18-cm-high step board at a pace of 30 step-ups and step-downs per minute; 3) bicycling on a stationary bike (Cardio Care 827E, Monark Exercise AB, Sweden) at a pace of 60 rpm and with a resistance corresponding to moderate bicycling intensity defined by the children; 4) jumping on a trampoline at a self-set pace; and 5) playing basketball, dribbling through a marked track, and ending each round by scoring a goal; encouraged to score as many goals as possible during the 5-min period (Table 1).
After a 10-min break in sitting position, the subject mounted a precalibrated treadmill (Cardionics 2111, Cardionics AB, Sweden) to perform walking and running at 2, 3, 4, 5, 6, 7, 8, and 10 km·h−1. Each speed had duration of 5 min and was increased without breaks until the highest speed was reached or the subject chose to stop (Table 1). The subjects were encouraged to complete all speeds.
For the assessment of REE from SWA and OM, the average of minutes 16-25 of the 30 min lying down was used. Minutes 1-15 were used for the subject to become relaxed and as an equilibrium period for OM. For the assessment of energy cost of all other activities, the average of minutes 3-5 was used, considering the first 2 min as an equilibrium period. To establish the MET values of the activities in the study, total energy cost of each activity was divided by REE, all values assessed by OM. To test the validity of SWA, the energy cost assessed by SWA and OM was compared for each activity separately, using a paired, two-sided t-test. For the treadmill activity, a paired, two-sided t-test was used for each speed. A Bland-Altman plot (2) was used to study the agreement between the methods and how it was related to the intensity of the physical activities (MET) assessed by OM. Pearson's product-moment correlation coefficient was used to assess a relation between the intensity of the physical activities and the difference between the methods. Values of the energy cost and MET for all activities are presented as means (SD). All statistical analysis was performed using SPSS version 14.0 (SPSS Inc., Chicago, IL).
Complete results for all activities and all speeds on the treadmill were achieved from 17 out of 20 children. One of the children performed only REE and the treadmill activities, and the other two did not perform the highest speed on the treadmill. Subject characteristics for all 20 children are presented in Table 2. There was no statistically significant gender difference in any of the subject characteristics and REE. The highest intensity achieved was during basketball playing, 8.9 (1.3) METs, and was comparable with running at a speed of 10 km·h−1, 8.7 (1.1) METs, assessed by OM (Table 3). Walking up and down on a step board resulted in almost the same intensity as walking 6 km·h−1, 5.0 (0.6) and 4.9 (0.7) METs, and jumping on a trampoline was comparable with walking/running 7 km·h−1, 7.7 (1.1) and 7.5 (1.0) METs. The intensity of stationary biking, 5.5 (1.0) METs, was between that of walking 6 km·h−1 and walking/running 7 km·h−1. Three ofthe subjects started to run at 7 km·h−1. Eight of the subjects tried to run at 7 km·h−1, but the speed was not comfortable for running, so they alternated between walking and running. All subjects were running at 8 km·h−1.
Except for jumping on a trampoline and walking 2 and 3 km·h−1, SWA significantly underestimated energy cost (Table 3). Jumping on a trampoline showed the largest variation in the difference between SWA and OM, −2.7 (11.9) kJ·min−1. The largest relative mean difference was achieved during bicycling, where SWA underestimated energy cost by 51% (13). The underestimation by SWA increased by increasing intensity, and between running at 8 and 10 km·h−1, SWA did not assess any further increase in energy cost (Table 3). The energy cost of lying down (REE) and sitting and playing games on a mobile phone did not differ when assessed by SWA, 3.5 (0.4) and 3.6 (0.5) kJ·min−1 (P = 0.60), but it differed when assessed by OM, 4.3 (0.5) and 5.6 (0.6) kJ·min−1 (P < 0.001).
The plots in Figure 1 further illuminate the increase in underestimation of energy cost by SWA with increasing intensity (MET) of the physical activity. Whereas the correlation between the intensity of the physical activity and the difference between the methods for treadmill walking/running was −0.71 (P < 0.001), with a uniform variation of the difference between methods across speeds (Fig. 1B and Table 3), the inclusion of all physical activities in the analysis decreased the correlation to −0.58 (P < 0.001), affected by the results from jumping on a trampoline, bicycling, and also sitting and playing mobile game (Fig. 1A). These physical activities protruded from the shape of the relation between the intensity of the physical activity and the difference between the methods settled by treadmill walking/running, whereas walking up and down on a step board and playing basketball were aligned with this relation.
Validation of SenseWear Pro2 Armband.
This is the first study to examine the validity of the SenseWear Armband during different physical activities in children representing both simple and more complex movement patterns. SWA underestimated energy cost for all activities except for jumping on a trampoline and slow-to-normal walking. However, for jumping on a trampoline, there was a large individual variation in the assessment accuracy. Overall, the underestimation by SWA increased with increasing physical activity intensity.
Other studies validating SWA have been performed in healthy adults (8,12,13). One of them presented the results of validating SWA during rest, showing good agreement between SWA and energy cost assessed by indirect calorimetry; both methods reached a value of 5.4 (0.4) kJ·min−1 (8). In our study in children, SWA underestimated energy cost during rest by 0.7 (0.5) kJ·min−1. The REE assessed by SWA, 3.5 (0.4) kJ·min−1, was closer to Schofield's equation for estimating BMR using body weight and height (22), 3.8 (0.3) kJ·min−1, or Molnar's equation for estimating RMR using body weight, height and, age (17), 3.5 (0.3) kJ·min−1, than REE assessed by OM, 4.3 (0.5) kJ·min−1. When using the mean values of body weight and height from the validation study in adults presented above (8) in the BMR equation by Schofield (22), it resulted in an energy cost of 5.3 kJ·min−1, to compare with 5.4 kJ·min−1 assessed by both SWA and indirect calorimetry. In that study, they used a more strict definition of REE. In our study, we assessed REE in an at least a 3-h-postprandial condition and after 15 min in a supine position. Hence, the REE in our subjects probably attained a higher value than if the stricter REE protocol had been used. According to the manufacturer, an algorithm for resting based on children's data is included in the software for calculating energy expenditure. It seems that this algorithm does not discriminate between different resting conditions in lying position, and the error may be larger the farther away an individual is from a BMR condition. It may be difficult, or perhaps even unnecessary for reaching a close approximation of TEE, for an activity monitor to discriminate between different resting conditions in lying position. Of more importance could be to discriminate between resting in lying, sitting, or standing position. One function of the bidirectional accelerometer in SWA is to provide information about body position. OM detected a difference in energy cost between lying and sitting playing games on the mobile phone with a factor of 1.3. This difference was not detected by SWA. This could generate a greater error because children spend a lot of time sitting during the day (9). Hence, SWA does not cover resting in different body positions in children accurately.
SWA has been validated in healthy adults during stair stepping (12) and cycle ergometry (8,12). The stair stepping was performed on a 20-cm bench at a pace of 20 and 35 step-ups and step-downs per minute, compared with an 18-cm step board at a pace of 30 step-ups and step-downs in our study. In adults, SWA generated mean differences in energy cost of approximately −0.4 and −0.8 kJ·min−1 for the two paces, compared with −6.6 kJ·min−1 in our study. The cycle ergometry in the two studies were performed at a rate of 60 rpm on a load resulting in an intensity representing 60% V˙O2peak (8) and at a rate of 50 rpm during two different loads (12), compared with a rate of 60 rpm with a load resulting in moderate intensity in our study. SWA generated mean differences in energy cost of approximately −1.6 kJ·min−1 at the rate of 60 rpm and 0.8 and −0.8 kJ·min−1 for the two different loads at the rate of 50 rpm, compared with −12.0 kJ·min−1 in our study. A possible explanation to the difference in result between children and adults for stair stepping may be that in the adult study they used an algorithm developed for stair stepping. According to the manufacturer, there is no corresponding algorithm for children. Walking up and down on a step board and bicycling on a stationary bike belong to the more standardized physical activities in our study. Despite the lack of an algorithm for stair stepping in children, this physical activity showed almost the same measurement accuracy as treadmill walking at the same intensity (MET). Walking on a treadmill is a standardized physical activity used in calibration studies (7,19,21,20,27) and represents a simple movement pattern. Surprisingly, bicycling on a stationary bike showed a larger underestimation compared with walking on a treadmill at the same intensity, even though an algorithm for this physical activity has been developed for children. Because bicycling is a common physical activity in children, it is important in the future development of SWA that adjustments be made for this.
Jumping on a trampoline and playing basketball were the least standardized physical activities in our study, which may have contributed to the higher variation in intensity level. Also, these physical activities were of more complex characters and probably not covered properly by the children algorithms (resting, stationary biking, walking). However, the underestimation of energy cost by SWA for playing basketball was approximately the same as running on the treadmill at the same intensity. The more complex movement pattern when playing basketball did not have any major effect on the accuracy of SWA in addition to the intensity of the physical activity. Jumping on a trampoline differed markedly from the other physical activities in that it generated a large variation in energy cost assessed by SWA, independently of how intense the physical activity was performed. Although there was no statistically significant difference between SWA and OM, the movement pattern of this physical activity could be interpreted quite differently by SWA. However, jumping on a trampoline is not a common physical activity in children, and, hence, it would not affect the accuracy of SWA for longer periods.
SWA accurately assessed energy cost of slow-to-normal walking, which could be explained by that SWA has an algorithm for walking based on children studies. However, SWA did not accurately assess energy cost of higher walking intensities and during running. In the future development of the algorithm, this could be easily adjusted for as long as the sensors continue to respond to increased intensity and the variation in assessment accuracy does not change across the intensity span. This was the case in our study, up to the speed of 8 km·h−1. Then, the energy cost assessed by SWA leveled off. In a study where different activity monitors were compared during treadmill exercise in adults (13), the slope of the energy cost assessed by SWA was decreasing at speeds above 8 km·h−1. However, the energy cost assessed by SWA still increased above 8 km·h−1. In men, it reached a plateau at 12 km·h−1, but in women it increased even further. In our study, the leveling off of energy cost assessed by SWA coincided with the point where the subjects started to run. Because the treadmill exercise in our study was terminated at 10 km·h−1, the response by SWA at higher speeds in children is not made clear. A difference in the relationship between speed and energy cost assessed by indirect calorimetry has been reported between walking and running (29). Treadmill walking and running may not accurately describe the actual energy cost of walking and running in children. Hence, the future algorithm should also be based on field studies.
The design of SWA may make it a useful device in the assessment of physical activity and energy expenditure in both clinical and epidemiological settings. However, SWA has to be improved for assessment of intensity, and more activity-specific algorithms are needed to be able to cover the energy costs of children's physical activities. Also, more studies are needed to assess its validity in children during different physical activities and during longer periods of time under free-living conditions, including several physical activity monitors simultaneously. There is also a need for validation studies in other age groups.
Energy costs and MET values of physical activities.
A common procedure to assess energy cost of physical activities has been to use the MET values from the compendium of physical activities published by Ainsworth et al. (1) and multiply it with time spent in different physical activities. The assumption has been that 1 MET is 1 kcal·kg−1·h−1. This is a good approximation for adults, but not for children. The energy value of 1 MET in our study was 1.6 (0.2) kcal·kg−1·h−1, and this has been reported in another study to range from 1.7 (0.4) for children to 1.2 (0.2) kcal·kg−1·h−1 in adolescents (10), reflecting the degree of maturity. The authors conclude that the MET values of different physical activities in children and adolescents in their study were only slightly lower than adult values taken from the compendium of physical activities (1), and they suggest that a closer estimate of the energy cost of activity in children would be to use their adjusted estimates of the energy cost of 1 MET and then multiply it with MET values found in the compendium of physical activities (1). However, there were MET values for some of the physical activities in the children and adolescents that differed notably from those in the compendium of physical activities (10). For example, the MET value for sitting playing video game was 1.2-1.3 compared with 1.5 in the compendium; for walking 5.6 km·h−1, 4.3-4.7 compared with 3.8 in the compendium; and 6.7-8.2 for running 8.0 km·h−1 compared with 8 in the compendium. Hence, we suggest the creation of a physical activity compendium for children based on measured values only, and we want to contribute with our results, together with other published results, in initiating this compilation. We are aware that energy cost (MET) of physical activities in children changes with age, and, hence, the future compendium should preferably present MET values for separate age groups. This takes studies with sample sizes large enough for age separation or studies focusing on a limited age interval. Our results represent children at the age interval of 11-13 yr of age.
The REE in our study, expressed with different units to be comparable with other studies, was 6132 (830) kJ·d−1, 1466 (198) kcal·d−1, 1.57 (0.25) kcal·kg−1·h−1, or 0.026 (0.004) kcal·kg−1·min−1. It is comparable with REE in other studies for the same age group (10,19,24). Harrell et al. (10) report REE in three different age groups. Our subjects belong to both the two first age groups (age group 1 = 8-12 for boys, 8-11 for girls; age group 2 = 13-15 for boys, 12-14 for girls). The REE in their study was 1.71 (0.41) kcal·kg−1·h−1 for the younger group and 1.34 (0.36) kcal·kg−1·h−1 for the older group. The REE in our subjects were somewhere in between. In 11-yr-old children (6-16 yr old), Puyah et al. (19) report an REE of 0.026 (0.006) kcal·kg−1·min−1. Spadano et al. (24) found an REE of 5942 (657) kJ·d−1 in 12-yr-old girls, compared with 5878 (827) kJ·d−1 in the girls in our study.
Sitting and playing games on a mobile phone reached an MET value of 1.3 in our study. Three other studies have reported MET values in children while sitting and playing video/computer games of 1.1 (26), 1.2 (19), and 1.2-1.3 (10). Climbing stairs has been reported to result in an MET value of 5.4-6.5 (10) and 6.6 (26), compared with 5.0 when walking up and down on a step board in our study and 5.7 during step aerobics on a step bench (26). Energy cost, but not MET value, for basketball playing has been reported to 7.4 kcal·min−1 (6), compared with 9.3 kcal·min−1 in our study. The difference in results may be explained by the fact that the subjects in our study received encouragement to score as many goals as possible during 5 min, whereas in the other study the subjects were allowed to perform the activity at their own pace. Shooting and retrieving a basketball generated an MET value of 6.5 (26) compared with 8.9 when playing basketball in our study. Bicycling outside on level ground has been reported to generate an MET value of 5.9 (26) compared with 5.5 on a stationary bike in our study.
Several studies have reported MET values of treadmill walking and/or running in children (10,15,19,24,26) (Table 4). The MET values from Spadano et al. (24), Harrell et al. (10), Treuth et al. (26), and in our own study seem to be higher compared with Maffeis et al. (15) and Puyah et al. (19). Spadano et al. (24) used the stationary metabolic system OCM-1 (Ametek, Pittsburgh, PA); Harrell et al. (10) and Treuth et al. (26) used the portable Cosmed K4b2 (Cosmed, Rome, Italy); and we used the portable Oxycon Mobile. Maffeis et al (15) used a Douglas bag system (O2/CO2 concentrations: Oxinos 100 and Binos C gas analyzers, Lwybold Heraeus Gmbh, Hinau, Germany. Gas volume: Gas Meter MCG, SIM, Rome, Italy), and Puyah et al. (19) used a room respiratory calorimeter. Comparisons between different devices and techniques measuring V˙O2 and V˙CO2 for the assessment of energy costs should be done with caution because there is a variation in accuracy and precision (5,11,14,16). This is important when averaging results from different studies. In a validation study of OM during bicycling on a stationary bike using the stationary system Oxycon Pro (VIASYS Healthcare) as criterion method, OM showed statistically significant lower values of V˙O2 at high workloads above 200 W but statistically nonsignificant higher values of V˙CO2 for all workloads (18). This could theoretically produce lower values of energy cost at higher intensities and, hence, does not explain the higher values for walking and running in our study compared with Maffeis et al. (15) and Puyah et al. (19).
On the basis of the results from our study and other studies where energy costs of physical activities and REE in children and adolescents have been assessed by measuring V˙O2 and V˙CO2, we present a summary table of MET values (Table 5). These MET values derive from healthy boys and girls at an age span of 6-18 yr. Some of the activities have received a range of MET values because different results have been reported. We chose not to replace the range of MET values with one average MET value, because different techniques were used to assess the energy cost, or there could be age differences. The first step in calculating the energy cost of physical activity is to multiply the MET value of the activity performed with time spent in this activity (minutes or hours), resulting in MET-minutes or MET-hours. The energy cost value is then derived by multiplying MET-minutes or MET-hours with measured or estimated REE (kcal·min−1 or kcal·h−1). For example, if a 12-yr-old boy is running at 9 km·h−1 for 0.5 h, he will reach 3.0 MET·h (6.1 × 0.5 h). If using the age-adjusted REE by Harrell et al. (1.71 kcal·kg−1·h−1) (10), the energy cost will be 5.1 kcal·kg−1 (1.71 kcal·kg−1·h−1 × 3.0 MET·h), or 204 kcal with a body weight of 40 kg. Complemented with results from more studies, we hope that this table of MET values will aid researchers with future assessments of energy expenditure in children.
The SenseWear Pro2 Armband underestimated the energy cost of most activities in children, an underestimation that increased with increased physical activity intensity. A table of energy costs (MET values) of physical activities in children, measured by indirect calorimetry, is presented as an initiation of the creation of a compendium of physical activities in children. The energy cost data are collected from the current study and other reported studies.
The authors are grateful to the children who, with great patience, performed all the activities in this study, and to the personnel at the physiotherapy section at The Queen Silvia Children's Hospital, in whose premises we performed the activities. We are also grateful for the financial support from the Swedish Heart and Lung Association and the Swedish Heart and Lung Foundation.
No support was provided for this study by any manufacturer of the instruments used.
1. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc.
2. Bland, J. M., and D. G. Altman. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
3. Caspersen, C. J., K. E. Powell, and G. M. Christenson. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep.
4. Corder, K., S. Brage, N. J. Wareham, and U. Ekelund. Comparison of PAEE from combined and separate heart rate and movement models in children. Med. Sci. Sports Exerc.
5. Duffield, R., B. Dawson, H. C. Pinnington, and P. Wong. Accuracy and reliability of a Cosmed K4b2 portable gas analysis system. J. Sci. Med. Sport
6. Eisenmann, J. C., S. J. Strath, D. Shadrick, P. Rigsby, N. Hirsch, and L. Jacobson. Validity of uniaxial accelerometry during activities of daily living in children. Eur. J. Appl. Physiol.
7. Ekelund, U., J. Aman, and K. Westerterp. Is the ArteACC index a valid indicator of free-living physical activity in adolescents? Obes. Res.
8. Fruin, M. L., and J. W. Rankin. Validity of a multi-sensor armband in estimating rest and exercise energy expenditure
. Med. Sci. Sports Exerc.
9. Hardy, L. L., S. L. Bass, and M. L. Booth. Changes in sedentary behavior among adolescent girls: a 2.5-year prospective cohort study. J. Adolesc. Health
10. Harrell, J. S., R. G. McMurray, C. D. Baggett, M. L. Pennell, P.F. Pearce, and S. I. Bangdiwala. Energy costs of physical activities in children and adolescents. Med. Sci. Sports Exerc.
11. Hodges, L. D., D. A. Brodie, and P. D. Bromley. Validity and reliability of selected commercially available metabolic analyzer systems. Scand. J. Med. Sci. Sports
12. Jakicic, J. M., M. Marcus, K. I. Gallagher, et al. Evaluation of the SenseWear Pro Armband to assess energy expenditure
during exercise. Med. Sci. Sports Exerc.
13. King, G. A., N. Torres, C. Potter, T. J. Brooks, and K. J. Coleman. Comparison of activity monitors to estimate energy cost of treadmill exercise. Med. Sci. Sports Exerc.
14. Macfarlane, D. J. Automated metabolic gas analysis systems: a review. Sports Med.
15. Maffeis, C., Y. Schutz, F. Schena, M. Zaffanello, and L. Pinelli. Energy expenditure
during walking and running in obese and nonobese prepubertal children. J. Pediatr.
16. McNaughton, L. R., R. Sherman, S. Roberts, and D. J. Bentley. Portable gas analyser Cosmed K4b2 compared to a laboratory based mass spectrometer system. J. Sports Med. Phys. Fitness
17. Molnar, D., S. Jeges, E. Erhardt, and Y. Schutz. Measured and predicted resting metabolic rate in obese and nonobese adolescents. J. Pediatr.
18. Perret, C., and G. Mueller. Validation of a new portable ergospirometric device (Oxycon Mobile) during exercise. Int. J. Sports Med.
19. Puyau, M. R., A. L. Adolph, F. A. Vohra, and N. F. Butte. Validation and calibration of physical activity monitors in children. Obes. Res.
20. Rowlands, A. V., P. W. Thomas, R. G. Eston, and R. Topping. Validation of the RT3 triaxial accelerometer for the assessment of physical activity. Med. Sci. Sports Exerc.
21. Sallis, J. F., M. J. Buono, J. J. Roby, D. Carlson, and J. A. Nelson. The Caltrac accelerometer as a physical activity monitor
for school-age children. Med. Sci. Sports Exerc.
22. Schofield, W. N. Predicting basal metabolic rate, new standards and review of previous work. Hum. Nutr. Clin. Nutr.
39(Suppl. 1):5-41, 1985.
23. Shvartz, E., and R. C. Reibold. Aerobic fitness norms for males and females aged 6 to 75 years: a review. Aviat. Space Environ. Med.
24. Spadano, J. L., A. Must, L. G. Bandini, G. E. Dallal, and W. H. Dietz. Energy cost of physical activities in 12-y-old girls: MET values and the influence of body weight. Int. J. Obes. Relat. Metab. Disord.
25. Treuth, M. S., A. L. Adolph, and N. F. Butte. Energy expenditure
in children predicted from heart rate and activity calibrated against respiration calorimetry. Am. J. Physiol.
26. Treuth, M. S., K. Schmitz, D. J. Catellier, et al. Defining accelerometer thresholds for activity intensities in adolescent girls. Med. Sci. Sports Exerc.
27. Trost, S. G., D. S. Ward, S. M. Moorehead, P. D. Watson, W.Riner, and J. R. Burke. Validity of the computer science and applications (CSA) activity monitor
in children. Med. Sci. Sports Exerc.
28. U.S. Department of Health and Human Services. Dietary Guidelines for Americans, 2005. Washington, DC: U. S. Government Printing Office, 2005. 001-000-4719-1.
29. Walker, J. L., T. D. Murray, A. S. Jackson, J. R. Morrow Jr., and T. J. Michaud. The energy cost of horizontal walking and running in adolescents. Med. Sci. Sports Exerc.
30. Weir, J. B. New methods for calculating metabolic rate with special reference to protein metabolism. J. Physiol.