Habitual physical activity and physical activity intensity: their relation to ...

European Journal of Clinical Nutrition (2004) 58, 285–291

& 2004 Nature Publishing Group All rights reserved 0954-3007/04 $25.00 www.nature.com/ejcn

ORIGINAL COMMUNICATION Habitual physical activity and physical activity intensity: their relation to body composition in 5.0–10.5-y-old children RA Abbott1,2* and PSW Davies2,3 1 School of Human Movement Studies, University of Queensland, St Lucia, Australia; 2School of Human Movement Studies, Queensland University of Technology, Kelvin Grove, Australia; and 3Children’s Nutrition Research Centre, Department of Paediatrics and Child Health, University of Queensland, Royal Children’s Hospital, Herston, Australia

Background: Concerns of a decrease in physical activity levels (PALs) of children and a concurrent increase in childhood obesity exist worldwide. The exact relation between these two parameters however has as yet to be fully defined in children. Objective: This study examined the relation in 47 children, aged 5–10.5 y (mean age 8.470.9 y) between habitual physical activity, minutes spent in moderate, vigorous and hard intensity activity and body composition parameters. Design: Total energy expenditure (TEE) was calculated using the doubly labelled water technique and basal metabolic rate (BMR) was predicted from Schofield’s equations. PAL was determined by PAL ¼ TEE/BMR. Time spent in moderate, vigorous and hard intensity activity was determined by accelerometry, using the Tritrac-R3D. Body fatness and body mass index (BMI) were used as the two measures of body composition. Results: Body fat and BMI were significantly inversely correlated with PAL (r ¼ 0.43, P ¼ 0.002 and r ¼ 0.45, P ¼ 0.001). Times spent in vigorous activity and hard activity were significantly correlated to percentage body fat (r ¼ 0.44, P ¼ 0.004 and r ¼ 0.39, P ¼ 0.014), but not BMI. Children who were in the top tertiles for both vigorous activity and hard activity had significantly lower body fat percentages than those in the middle and lowest tertiles. Moderate intensity activity was not correlated with measures of body composition. Conclusions: As well as showing a significant relation between PAL and body composition, these data intimate that there may be a threshold of intensity of physical activity that is influential on body fatness. In light of world trends showing increasing childhood obesity, this study supports the need to further investigate the importance of physical activity for children. European Journal of Clinical Nutrition (2004) 58, 285–291. doi:10.1038/sj.ejcn.1601780 Keywords: physical activity; doubly labelled water; intensity; children; body composition; accelerometry

Introduction Over the past two decades, the prevalence of obesity in childhood has dramatically increased (Freedman et al, 1997; Troiano & Flegal, 1998; Chinn & Rona, 2001). For example, in the United States, the prevalence of overweight increased

*Correspondence: RA Abbott, School of Human Movement Studies, Connell Building, University of Queensland, St Lucia QLD 4072, Australia. E-mail: [email protected] Guarantor: RA Abbott. Contributors: RAA was principally involved in all aspects of the research: study design, data collection, data analyses and writing of the manuscript. PSWD was coresponsible for the study design, consulted on data analysis and contributed to the preparation of the manuscript. Received 22 January 2003; revised 21 March 2003; accepted 8 April 2003

two-fold between 1973 and 1994 (Freedman et al, 1997) and increased by 60–70% in young Australians between 1985 and 1997, with obesity increasing two to four-fold (Booth et al, 2003). This increase in overweight prevalence clearly reflects a shift towards a positive energy balance. Physical activity and dietary intake are the two primary modifiable behaviours controlling this balance. Data from large dietary studies in the United States and the United Kingdom suggest that calorie intake has not increased concurrently with the increasing obesity, and therefore point to increasing sedentary behaviours and reduced physical activity in both adults and children, as being the causative factors in the increased rate of obesity (Harsha, 1995; Kennedy & Goldberg, 1995; Heini & Weinsier, 1997; Prentice, 1997). Declining physical activity levels (PALs) in children have also been suggested

Physical activity in young children RA Abbott and PSW Davies

286 from studies assessing total energy expenditure (TEE) (Fontvieille et al, 1993; Davies, 1996). The relation between physical activity and body composition, in particular body fatness, has been explored in both cross-sectional and longitudinal studies, with inconsistent findings and whether there is a relation remains, at this point, controversial. High daytime activity, measured by a motion sensor, of children aged 4–8 y of age has been associated with reduced childhood adiposity (Berkowitz et al, 1985). Furthermore, increasing PALs have been associated with reduced levels of body fat in preschool children and young school-aged children (Davies et al, 1995; Deheeger et al, 1997; Ball et al, 2001). In contrast, free-living energy expenditure was shown to be positively related to fatness and not predictive of the development of body fat in young children (Goran et al, 1998). As Goran (2001) recently noted, the hypothesis that reduced energy expenditure is related to increased fat gain in childhood remains contentious and, to date, unproven. The inconsistent findings are due, in part, to the different methodologies used to assess and define physical activity. Differing ages of children involved and small sample sizes have also limited the comparison of studies. In a recent meta-analysis on the effect of type of physical activity measure on the relation between body fatness and habitual physical activity, the authors concluded that the size of the relation found was highly dependent on the activity measure chosen (Rowlands et al, 2000). Studies assessing habitual activity in terms of energy expenditure, including the PAL index, however, were not included in this meta-analysis. The PAL index is used to reflect habitual physical activity and is calculated by dividing TEE by basal metabolic rate (BMR). TEE has four principal components: BMR, the energy cost of growth, the thermic effect of food and the energy cost of activity. Dividing TEE by BMR, to provide an estimate of the energy cost of activity, or PAL, assumes that the majority of energy expended over a 24 hr period, above that of basal metabolism, is due to the energy cost of activity, and that dividing by BMR, fully accounts for BMR in individuals of all sizes, sexes and body composition. These assumptions have been validated and are widely accepted (Black et al, 1996; Coward, 1998). For comparative purposes, in a meta-analysis by Torun et al (1996), of studies using doubly labelled water, children aged 6–13 y have been shown to have a mean PAL of 1.79. Habitual physical activity is an important dimension of childhood activity, especially since the spontaneity of children’s movement makes assessing specific activities difficult. However, other dimensions of physical activity, such as duration and intensity, may also be influential on parameters of body composition. In a recent study by Ekelund et al (2002), the intensity and duration of physical activity and the total amount of physical activity was significantly lower in obese adolescents, compared to controls, whereas TEEs and activity-related energy expenditures were no different. Janz et al (2002) have recently European Journal of Clinical Nutrition

reported an association between low levels of vigorous activity and increasing body fat levels in children; however, there was no such association with moderate activity. In contrast, time devoted to low-to-moderate intensity physical activity appeared to influence TEE to a greater extent than activity of high intensity in adults (Westerterp, 2001). To date, there has been little research in this field, especially in young children, most likely due to the difficulty in its accurate measurement. The aim therefore for this study was to combine two methods of assessing physical activity and explore the relation between physical activity and body composition in a group of young children, with emphasis on physical activity intensity.

Methods Children attending a local Brisbane state school (aged 5– 10.5 y) and those of staff at Queensland University of Technology were invited to take part in the study. All measurements were made during school term time. A home visit was made to confirm all aspects of the study and obtain written informed consent from the parent and assent from the child. In total, 47 children (23 boys, 24 girls), consented to be recruited into the study. Ethical approval was obtained from Queensland University of Technology Human Research Ethics Committee and formal permission was obtained from the state school in conjunction with Education Queensland.

Anthropometry and body composition Height was recorded to the last completed millimetre using a fixed Holtain stadiometer (Holtain Ltd, Crymch, UK). Weight was measured, in minimal clothing, using digital scales (Tanita BWB 600, Tanita Corp, Tokyo) and recorded to the nearest 0.1 kg. Body mass index (BMI) (Wt/Ht2) was used as the index of relative adiposity. Percentage body fat was calculated from the measurement of the 18O dilution space, which had previously been calculated as part of the doubly labelled water technique (see below) for the assessment of energy expenditure. Taking into account the fact that 18O overestimates total body water by 1% (Schoeller, 1983) and using published total body hydration constants from children of different ages (Fomon et al, 1982) fat-free mass can then be calculated from total body water. The percentage of fat-free mass was then determined. Fat mass was then calculated as the difference between total body weight and fat-free mass and expressed as a percentage of body weight.

Habitual physical activity: determination of PAL PAL was calculated as the ratio of TEE to BMR. TEE was determined using the doubly labelled water technique. This method, described in detail elsewhere (Davies et al, 1997) relies on the differential elimination over 10 days of the two


287 stable isotopes, deuterium (2H) and oxygen 18 (18O) from the body. The difference in the rates of elimination of the isotopes is directly proportional to the carbon dioxide production rate and hence energy expenditure. In practical terms, at a home visit, a baseline urine sample was collected. Children were then given 2H2O and H18 2 O, at a dose regimen of 0.125 g/ 18O/kg body weight and 0.05g/2H/kg body weight to drink. A urine sample was collected after approximately 5 h and subsequently daily, for 10 days, and stored either in the fridge or freezer until collection on the final study visit. Sample preparation allowed for determination of both 2H and 18O enrichment by equilibration with both hydrogen and carbon dioxide, respectively. Isotopic enrichment of the urine samples was measured via isotope ratio mass spectrometry (PDZ Europa, Crewe, UK). Energy expenditure was calculated by the slope-intercept method, with isotope dilution spaces calculated by extrapolation of the enrichments to time zero. A respiratory quotient of 0.85 was assumed in the calculation of TEE. Due to the nature of the study with children attending the measurement centre at different times of the day, in fasted and nonfasted conditions, measured RMR was not deemed suitable. BMR was therefore predicted from the subjects sex, age, weight and height using Schofield’s equations for children aged 3–10 y of age (Schofield, 1985). Predicted BMR has been shown to have good agreement with measured RMR by indirect calorimetry in children of all ages (Firouzbakhsh et al, 1993; Allen et al, 1995).

Intensity of physical activity: accelerometry ‘Minute-by-minute’ accelerometry was recorded using the Tritrac-R3D tri-axial accelerometer (Professional Products, Madison, USA). The Tritrac-R3D was wrapped in bubble-wrap and fixed in position within a ‘bum-bag’, worn around the waist, for a 4-day period during the 10-day urine collection phase for the DLW technique. The accelerometers were worn for 2 week days and 2 weekend days and all data collection were performed during school term time. The children were asked to wear the bum-bag for as long as possible, from when they first got up, to when they went to bed. The Tritrac-R3D measures minute-by-minute acceleration in three dimensions: antero-posterior (x), medio-lateral (y) and vertical (z). Movement counts are produced for each minute for all dimensions, as well as a summary for all three planes, called the total vector magnitude, and calculated as (x2 þ y2 þ z2)0.5. The mean daily movement count for each plane and the mean vector magnitude count were calculated. The minutes spent in moderate (MODTT), vigorous (VIGTT), and ‘hard and above’ (HARDTT) according to vector magnitude counts, were also calculated. Moderate intensity activity was defined as counts in the range of 1000–1999, vigorous intensity activity as 2000–3499, ‘hard and above’ intensity activity as above 3500. These zones have been described and validated by Rowlands et al (2000).

Statistical methods Descriptive data are expressed as mean 7s.d. Differences in body composition between genders were assessed by Student’s t-test. Pearson-product moment correlation was used to explore the association between PAL with percentage body fat and BMI and Spearman’s rank correlation was used for the association between time spent in intensity of activity with body composition. The relation of body composition across PAL tertiles was investigated further using a two-factor analysis-of-variance, specifically testing for gender interaction. A one-way analysis of variance was also used to explore the relation between tertiles of vigorous activity and hard activity with body composition, with post hoc Bonferroni analysis to identify specific group differences. The statistical package used was SPSS for Windows (SPSS Inc, USA).

Results Anthropometry and body composition The general descriptives of the body composition of the children, according to gender, are shown in Table 1. The boys and girls were similar in age, weight, height and BMI. However, as expected, girls had a higher fat mass and percentage body fat compared to boys. In addition, boys had significantly greater levels of fat-free mass than girls. Using the international cutoff points for BMI, proposed by Cole et al (2000), three boys and seven girls were classified overweight, but none of the children were classified as obese. Correlation analyses showed the expected significant increases in height and weight with increasing age group, but there was no correlation between age and percentage body fat or BMI (data not shown).

Physical activity PAL ranged from 1.32 to 2.18, with a mean and s.d. of 1.7270.19. There was no significant difference in PAL values between boys and girls (see Table 2). Four days of accelerometry recordings were obtained from 19 boys and 21 girls, with the loss of data from seven children occurring either

Table 1 Characteristics of body composition variables between boys and girls

Age (y) Height (m) Weight (kg) BMI (kg/m2) Fat mass (kg) Percentage body fat (%) Fat-free mass (kg) Percentage fat-free mass (%)

Boys (n=23)

Girls (n=24)

Significance (boys vs girls) P

8.570.9 1.3370.09 30.076.7 16.771.9 7.873.9 25.277.8 22.273.9 74.877.8

8.470.9 1.3170.06 30.175.8 17.472.4 9.973.4 32.276.2 20.272.4 67.876.2

NS NS NS NS NS 0.001 0.047 0.001

Values are mean7s.d.

European Journal of Clinical Nutrition


288 50

Table 2 Descriptives of energy expenditure measurements and TritracR3D recordings for both boys and girls Significance (boys vs girls) P

19887253 1.7170.14 34427548

18887185 1.7270.19 34217656

254776 276771 3587116 5107128

201761 229742 301775 429778

0.021 0.017 NS 0.024

3647126 149783 22718

299794 84741 876

NS 0.005 0.006

NS NS NS

30

20 GENDER Girls

10

Boys

Values are mean7s.d.

due to data transfer problems with specific accelerometers or failure to wear the monitor for the required time due to practical constraints. An average of 3431 min (57 h) of movement recording was obtained per child, approximately 14 h/day. These data are shown in Table 2. Boys had higher mean values of movement counts in both the x and y planes as well as a total higher vector magnitude. This translated, in turn, to boys spending significantly longer periods in activity classed as vigorous or hard, compared to girls: 149 vs 84 min/ day for vigorous activity and 22 vs 8 min/day for hard activity. Comparison of PAL and TEE with the measures from accelerometry showed significant correlations between PAL and the mean vector magnitude (r ¼ 0.31, P ¼ 0.049), the raw y count (r ¼ 0.34, P ¼ 0.032) and the raw z count (r ¼ 0.42, P ¼ 0.008). TEE was correlated with only the raw y count (r ¼ 0.33, P ¼ 0.036) and raw z count (r ¼ 0.40, P ¼ 0.011). Neither PAL nor TEE, however, were correlated with minutes spent in MODTT, VIGTT or HARDTT.

PAL and body composition PAL was significantly negatively correlated with both percentage body fat (r ¼ 0.43, P ¼ 0.002) and with BMI (r ¼ 0.45, P ¼ 0.001). The relation between PAL and percentage body fat is shown in Figure 1. The relation between PAL and both percentage body fat and BMI was also explored by comparing both body composition variables across PAL tertiles. Although for both percentage body fat and BMI, increasing tertiles of PAL produced lower mean parameter values, one-way ANOVA showed there to be significant differences at the 5% level between PAL tertiles and BMI only. A box plot depicting the mean plus 95% confidence intervals for percentage body fat for each PAL tertile is shown in Figure 2. Knowing there to be a marked difference in percentage body fat between boys and girls, the issue of whether gender European Journal of Clinical Nutrition

40 Percentage body fat

Girls (n=21)

0 1.2

Total Population 1.4

1.6

1.8 PAL

2.0

2.2

Figure 1 Correlation of PAL with percentage body fat (r ¼ 0.43, P ¼ 0.002) with boys and girls highlighted.

50

7F/8M

9F/6M

8F/9M

15 1.32-1.63

15 1.64-1.80

17 1.81-2.18

40 Percentage body fat

TEE (kcal/day) PAL Total accelerometry recording (min) Mean x (counts) Mean y (counts) Mean z (counts) Total vector magnitude (min/day) MODTT (min/day) VIGTT (min/day) HARDTT (min/day)

Boys (n=19)

30

20

10

0 N=

Tertiles of PAL

Figure 2 Box-bar plot comparing percentage body fat across PAL tertiles. Plots show the mean plus 95% confidence intervals, with ANOVA showing no significant differences between the tertiles (P ¼ 0.073). Distribution of gender between tertiles is shown.

was significant in the relation between PAL and percentage body fat was examined with a two-factor ANOVA. Despite the fact that in univariate analysis, there was a greater correlation between PAL and body fat in the girls (r ¼ 0.68, P ¼ 0.001), compared to the boys (r ¼ 0.39, P ¼ 0.067), gender was shown to be not significant in this relation within the study. As with percentage body fat, the association between PAL and BMI was more marked in girls than boys (r ¼ 0.62, P ¼ 0.001 vs r ¼ 0.26, P ¼ 0.212), but gender was not shown to interact significantly. For subsequent analyses, therefore, the children were treated as a single group.


289 50 10F/4M

7F/6M

4F/ 9M

Percentage body fat

40

30

20

10

0 N=

p=0.052

p=0.021

13 < 71 mins

13 71-125 mins

13 > 125 mins

Tertiles of vigorous activity 50 8F/5M

11F/3M

2F/11M

40

Percentage body fat

Intensity of activity and body composition Intensity of activity, as assessed by accelerometry, was significantly correlated with percentage body fat, but not BMI. Spearman’s rank correlation analyses showed that VIGTT and HARDTT were statistically negatively correlated with percentage body fat, with coefficients of r ¼ 0.44 (P ¼ 0.004) and r ¼ 0.39 (P ¼ 0.014). There was no significant association with MODTT (r ¼ 0.17, P ¼ 0.300). Further exploration of the association between intensity of activity and percentage body fat was achieved by dividing the time spent in activity intensities into tertiles. For vigorous intensity activity, the tertiles were: less than 71 min/day, 71–125 min/day and greater than 125 min/day. One-way ANOVA analysis showed there to be a significant difference between the tertiles. Post hoc analysis showed that those children in the highest tertile for time spent in vigorous movement had significantly lower percentage body fat levels than those in either the middle (22.9 vs 30.6%) or the lowest tertile (22.9 vs 29.5%). There was no difference between the lowest and middle tertile. These data are shown in the box-plot in Figure 3. Knowing that boys accumulated more time in greater intensity activities than girls, gender was tested for interaction, but was not significant, thus justifying analysing the group as a whole. The same analysis was performed for HARDTT. For hard and above intensity activity, the tertiles were: less than 8 min/day, 8 to 15 min/day and greater than 15 min/day. One-way ANOVA analysis showed there to be a significant difference between the tertiles. Post hoc analysis showed that those children in the highest tertile for time spent in hard and above movement had significantly lower percentage body fat levels than those in either the middle (22.7 vs 29.8%) or the lowest tertile (22.7 vs 30.5%). There was no difference between the lowest and middle tertile. These data are shown as a box-plot in Figure 3.

30

20 p=0.016

p=0.027

10

0 N=

13

14

13

< 8 mins

8-15 minutes

> 15 minutes

Tertiles of daily 'hard and above' activity

Discussion Body composition, assessed in terms of either percentage fat or BMI, was significantly inversely correlated with habitual physical activity in this study. The strength of association was similar for both measures of body composition: the more active the children, the lower their percentage fat or smaller their weight for height. The most active children, those in the top tertile of physical activity, had statistically significantly lower BMI scores and lower percentage body fat, than those children in the lowest tertile of physical activity. There has been considerable interest in the relation between physical activity and body composition in children. As stated, the recent meta-analysis by Rowlands et al (2000) concluded that there was a weak to moderate relation between body fat and activity in children; however, studies that had used energy expenditure ratios or indices, as measures of physical activity, were not included. Their rationale for this was two-fold. Firstly they cited Bar-Orr et al (1994), who stated ‘when determining the relation of

Figure 3 Box-plots of tertiles of time spent, per day, in vigorous movement activity and ‘hard and above’ movement activity and associated percentage body fat. The box-plot presents the median value, the inter-quartile range and the smallest and largest value for each vigorous activity tertile. Significant differences between tertiles, from Bonferroni post hoc analyses after one-way ANOVA, are shown. Distribution of gender between tertiles is shown.

physical activity with adiposity, physical activity should be expressed as body movement, not as energy expended’. Secondly they suggest that physical activity and energy expenditure are not synonymous and the terms should not be used interchangeably. However, PAL, predominantly reflects the energy cost of habitual physical activity and it would be amiss to exclude studies that are assessing a different, but by no means unimportant, component of habitual physical activity. It may indeed be that the energy cost of habitual activity is equally as important, in terms of body composition, as the extent of physical movement. Indeed, in this study PAL was shown to inversely correlate with percentage body fat. European Journal of Clinical Nutrition


290 Of the studies looking at activity and body composition that have assessed activity by some measure of energy expenditure, there have been mixed findings. In agreement with the findings from this study, PAL has been shown to be inversely associated with percentage body fat in both younger children (1.5–4.5 y) and adolescents (Davies et al, 1995; Ekelund et al, 2002) with correlations of –0.52 and –0.53, respectively. These compare to the correlation of –0.42 observed in this study. In contrast however, Salbe et al (1997) showed there to be no difference in PAL values between lean and obese 5-y-old children and Goran et al (1997) found no association between activity energy expenditure and fat mass in children aged 4–8 y. Despite having measured both TEE and REE, Goran et al (1997) did not explore whether PAL itself was related to fat mass. Evidently, there needs to be more research in this area. In this study, there was no significant effect of gender on the relation between physical activity and body composition. A significant negative association between activity, measured by movement counts, and fatness in girls and not boys has been found previously (Almeida et al, 1999). Rowlands et al (1999) also found that the relation between body fat and activity was stronger in girls than boys, though the relation only became significant when the two were grouped together. In contrast, Ball et al (2001) showed the association of PAL with percentage body fat to be significant in 6–9-y-old boys and not girls. The majority of the studies assessing this relation do not show the data split by gender and this could be attributable to authors finding no difference between boys and girls, as in this study, and therefore not presenting the data, or the gender effect was not explored. For this study, in particular, the relatively small sample size may have precluded finding any true-effect, if present. In the meta-analysis by Rowlands et al, no effect of gender on the strength of the relation between activity level and body fat was observed when all activity measures were considered together, or when the activity measures were considered separately. They did however acknowledge that with the variety of methods that had been used for determining both activity and fat mass itself, more standardised studies are needed to fully explore the effect of gender and age on this relation. The question of whether intensity of activity influences body composition also remains in debate. In this study, increasing amounts of both vigorous intense and hard intense activity, measured by accelerometry, were associated with reduced percentage body fat. Moreover, children who accumulated over 2 h of vigorous movement activity per day were significantly less fat than those who failed to accumulate on average about 1.0–1.5 h of vigorous movement activity. Time spent in moderately intense levels of physical activity was not associated with body fatness. Furthermore, minutes spent in either vigorous or hard activity were not correlated with overall TEE or PAL, suggesting that these activity parameters do not represent the same dimension of activity and may independently affect body fat levels. European Journal of Clinical Nutrition

Indeed, these data are suggestive of a threshold of intensity of physical activity above which activity interacts with body composition. There have been few studies to date, that have assessed intensity of activity by accelerometry and related it to body composition. Rowlands et al (1999) also found a significant negative association with both moderate and vigorous movement activity with percentage body fat. They reported a significant correlation of –0.42 for combined vigorous and hard movement activity with body fat, which compares well with the significant correlations of –0.44 and –0.39 for vigorous and hard activity from our study. However in this study, a significant correlation between body fat and moderate intensity activity was not observed. Our data are however in agreement with that of Janz et al (2002) who showed correlation between percentage body fat with time spent in vigorous activity (r ¼ 0.26–0.30), but not moderate activity, in 436 children aged 4–6 y. In addition, Trost et al (2001) found that obese children exhibited significantly lower daily accumulation of total accelerometer counts, but in particular, lower counts in the moderate and moderate-tovigorous physical activity count range than nonobese children. This would inherently suggest less time spent in activities of higher intensity. Clearly, the benefit of time spent in higher intensities of physical activity needs to be more fully investigated. The limitations of this study are clear. The design was cross-sectional which limits the ability to determine any causality in the results and data was collected over a specified 10-day period for energy expenditure and 4 days for accelerometry. Increasing data collection for accelerometry to 7 days might reflect more accurately usual intensity patterns (Trost et al, 2000). There is also a question as to whether findings from this group of children, who were a voluntary convenience sample, would be representative of findings in children in general. Nevertheless, with regular reports of increasing childhood obesity, the results from this study are noteworthy and indeed timely. They lend support to the hypothesis that it is the alteration of childhood activity patterns, namely a reduction in PALs, that is at least partly attributable to increasing levels of childhood fatness and also suggest that intensity of physical activity may be of influence. Further research into this area is indeed warranted.

Acknowledgements This research was supported by a PDZ Europa scholarship in conjunction with an Overseas Postgraduate Research Scholarship from the Australian Government. Neither author have any affiliation with Europa Scientific (partfunders of this research).

References Allen JR, McCauley JC, Waters DL, O’Connor J, Roberts DCK & Gaskin KJ (1995): Resting energy expenditure in children with phenylketonuria. Am. J. Clin. Nutr. 62, 797–801.


291 Almeida MJ, Fox KR & Boutcher S (1999): Physical activity levels and fatness in male and female adolescents. Int. J. Obes. Relat. Metab. Disord. 23(Suppl 3), S78. Ball EJ, O’Connor J, Abbott R, Steinbeck KS, Davies PSW, Wishart C, Gaskin KJ & Baur L (2001): Total energy expenditure, body fatness and physical activity in children aged 6–9 years. Am. J. Clin. Nutr. 74, 524–528. Bar-Orr O & Baranowski T (1994): Physical activity, adiposity and obesity among adolescents. Pediatr. Exerc. Sci. 6, 348–360. Berkowitz RI, Agras S, Korner AF, Kraemer HC & Zeanah CH (1985): Physical activity and adiposity: a longitudinal study from birth to childhood. J. Pediatr. 106, 734–738. Black AE, Coward WA, Cole TJK & Prentice AM (1996): Human energy expenditure in affluent societies: an analysis of 574 doubly labelled water measurements. Eur. J. Clin. Nutr. 50, 72–92. Booth ML, Chey T, Wake M, Norton K, Hesketh K, Dollman J & Robertson I (2003): Change in the prevalence of overweight and obesity among young Australians, 1969–1997. Am. J. Clin. Nutr. 77, 29–36. Chinn S & Rona RJ (2001): Prevalence and trends in overweight and obesity in three cross-sectional studies of British Children, 1974–1994. BMJ 322, 24–26. Cole TJ, Bellizi MC, Flegal KM & Dietz WH (2000): Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 320, 1240–1243. Coward WA (1998): Contributions of the doubly labelled water method to studies of energy balance in the Third World. Am. J. Clin. Nutr. 68(Supp 1), S962–S969. Davies PSW, Gregory J & White A (1995): Physical activity and body fatness in pre-school children. Int. J. Obes. Relat. Metab. Disord. 19, 6–10. Davies PSW (1996): Total energy expenditure in young children. Am. J. Hum. Biol. 8, 183–188. Davies PSW, Wells JCK, Hinds S, Day JME & Laidlaw A (1997): Total energy expenditure in nine and twelve month old infants. Eur. J. Clin. Nutr. 51, 249–252. Deheeger M, Rolland-Cachera MF & Fontvieille AM (1997): Physical activity and body composition in 10 year old French children: linkages with nutritional intake. Int. J. Obes. Relat. Metab. Disord. 21, 372–379. Ekelund U, Aman J, Yngve A, Renman C, Westerterp K & Sjostrom M (2002): Physical activity but not energy expenditure is reduced in obese adolescents: a case–control study. Am. J. Clin. Nutr. 76, 935–941. Firouzbakhsh S, Mathis RK, Dorchester WL, Oseas RS, Groncy PK, Grant KE & Finklestein JZ (1993): Measured resting energy expenditure in children. J. Pediatr. Gastroenterol. Nutr. 16, 136–142. Fomon SJ, Haschke F, Ziegler EE & Nelson SE (1982): Body composition of reference children from birth to age 10 years. Am. J. Clin. Nutr. 35, 1169–1175. Fontvieille AM, Harper IT, Ferraro RT, Sproul M & Ravussin E (1993): Daily energy expenditure by five year old children, measured by doubly labelled water. J. Pediatr. 123, 200–207.

Freedman DS, Srinivasan SR, Valdez RA & Berenson GS (1997): Secular increases in relative weight and adiposity among children over two decades: the Bogalusa Heart Study. Pediatrics 99, 420–425. Goran MI (1997): Energy expenditure, body composition, and disease risk in children and adolescents. Proc. Nutr. Soc. 56, 195–209. Goran MI (2001): Metabolic precursors and effects of obesity in children: a decade of progress 1990–1999. Am. J. Clin. Nutr. 73, 158–171. Goran MI, Shewchuk R, Gower BA, Nagy TR, Carpenter WH & Johnson R (1998): Longitudinal changes in fatness in white children: no effect of childhood energy expenditure. Am. J. Clin. Nutr. 67, 309–316. Harsha DW (1995): The benefits of physical activity in childhood. Am. J. Med. Sci. 310(Supp 1), S109–S113. Heini AF & Weinsier RL (1997): Divergent trends in obesity and fat intake patterns: the American paradox. Am. J. Med. 102, 259–264. Janz KF, Levy SM, Burns TL, Torner JC, Willing MC & Warren DDS (2002): Fatness, physical activity, and television viewing in children during the adiposity rebound period: the iowa bone development study. Prev. Med. 35, 563–571. Kennedy E & Goldberg J (1995): What are American children eating? Implications for public policy. Nutr. Rev. 53, 111–126. Prentice AM (1997): ObesityFthe inevitable penalty of civilisation? Br. Med. Bull. 53, 229–237. Rowlands AV, Eston RG & Ingledew DK (1999): Relationship between activity levels, aerobic fitness, and body fat in 8 to 10 year old children. J. Appl. Physiol. 86, 1428–1435. Rowlands AV, Ingledew DK & Eston RG (2000): The effect of type of physical activity measure on the relationship between body fatness and habitual physical activity in children: a meta analysis. Ann. Hum. Biol. 27, 479–497. Salbe A, Fontvieille AM, Harper I & Ravussin E (1997): Low levels of physical activity in 5 year old children. J. Pediatr. 131, 423–429. Schoeller DA (1983): Energy expenditure from doubly labelled water: some fundamental considerations in humans. Am. J. Clin. Nutr. 38, 999–1005. Schofield WN (1985): Predicting basal metabolic rate, new standards and review of previous work. Hum. Nutr. Clin. Nutr. 39C, S5–S41. Torun B, Davies PSW, Livingstone MBE, Paolisso M, Sackett R & Spurr GB (1996): Energy requirements and dietary recommendations for children and adolescents 1 to 18 years old. Eur. J. Clin. Nutr. 50, S37–S81. Troiano RP & Flegal KM (1998): Overweight children and adolescents: description, epidemiology and demographics. Pediatrics 101, 497–504. Trost SG, Pat RR, Freedson PS, Sallis JF & Taylor WC (2000): Using objective physical activity measures with youth: how many days of monitoring are needed? Med. Sci. Sports Exerc. 32, 426–431. Trost SG, Kerr LM, Ward DS & Pate RR (2001): Physical activity and determinants of physical activity in obese and non-obese children. Int. J. Obes. Relat. Metab. Disord. 25, 822–829. Westerterp KR (2001): Pattern and intensity of physical activity. Nature 410, 539.

European Journal of Clinical Nutrition