点击显示 收起
1 From the Medical Research Council Human Nutrition Research, Cambridge, United Kingdom (KLR, WAC, and SAJ); the Northern Ireland Centre for Diet and Health, University of Ulster, Coleraine, United Kingdom (MBEL, AM, and KLR); the Medical Research Council Childhood Nutrition Research Centre, Institute of Child Health, London, United Kingdom (JCKW); and the Medical Research Council International Nutrition Group, Nutrition and Public Health Intervention Research Unit, London School of Hygiene and Tropical Medicine, London, United Kingdom (AMP)
2 Supported by a grant for data collection from The Sugar Bureau, Kellogg's, Coca Cola, and Masterfoods and for data analysis by the Medical Research Council. 3 Address reprint requests to K Rennie, Northern Ireland Centre for Diet and Health, University of Ulster, Coleraine BT52 1SA, United Kingdom. E-mail: klr1000{at}cam.ac.uk.
ABSTRACT
Background: Physical inactivity increases the risk of obesity, but the relations between reported levels of physical activity (PA) and measures of body fatness (BF) in children are remarkably inconsistent.
Objective: We examined the relation between objective measures of PA and body-composition indexes in nonobese children.
Design: A cross-sectional study was conducted in 100 children aged 6–8 y who were recruited according to their risk of future obesity: high-risk children had 1 obese parent [body mass index (BMI; in kg/m2): >30] and low-risk children had 2 nonobese biological parents (BMI: <30). Free-living activity energy expenditure (AEE) and PA level were calculated from 7-d doubly labeled water measurements, time spent in light-intensity activity was assessed by heart rate monitoring, and body composition was determined from isotopic dilution. To adjust for body size, fat mass and fat-free mass were normalized for height and expressed as fat mass index (FMI) and lean mass index (LMI), respectively.
Results: High-risk children had significantly higher BMI, LMI, and FMI than did low-risk children, but no group differences in PA were found. AEE and PA level were positively associated with LMI and, after adjustment for sex and fat-free mass, negatively associated with FMI but not with BMI. Boys who spent more than the median time in light-intensity activities had significantly higher FMI than did less sedentary boys. This difference was not observed in girls.
Conclusions: AEE and PA level were negatively associated with BF in nonobese children. Accurate measures of body composition are essential to appropriate assessment of relations between PA and obesity risk.
Key Words: Children • physical activity • energy expenditure • body composition • obesity
INTRODUCTION
The prevalence of childhood obesity continues to increase, and established obesity is difficult to resolve. Therefore, an urgent need exists to identify modifiable risk factors for obesity to mitigate the emerging epidemic. It is now well established that children with 1 obese parent are at high risk of excess weight gain in both childhood and adolescence and that the risk increases with age (1, 2). However, it is not clear which environmental factors, in particular which components of physical activity and diet, are the most important contributors to this risk.
Although the increase in childhood obesity is frequently attributed to a decline in physical activity (PA), a remarkable lack of consistency exists in the relations between reported levels of PA and degrees of fatness (3, 4). This inconsistency could be due to methodologic flaws in assessing activity, inactivity, and body composition. The cross-sectional nature of many of the studies that have used objective measures of activity in children and adolescents to compare activity levels in lean and obese children do not allow the establishment of cause-and-effect relations (5–11). Studies in healthy children that examined relations between activity measures–derived estimates of energy expenditure from doubly labeled water (DLW) and measures of fatness reported mixed results; some reported a negative relation (7, 12–16), and others reported no relation (17–19). Furthermore, few studies have examined associations between objective measures of the intensity of activity and fatness. One study that used heart rate (HR) monitoring in 28 boys aged 9 y found a positive association between time spent on sedentary activities and percentage body fat (%BF) but not between activity and %BF (20), whereas a relation between vigorous activity, assessed by accelerometry, and %BF was observed in studies of prepubescent children (16, 21, 22). However, no studies have examined associations between PA and fatness in lean children at different risks of obesity.
This study was designed to establish the relation between PA and body composition in nonobese children by simultaneously using 2 independent and objective measures of PA and by using an approach to optimally normalize body-composition variables for body size. Furthermore, we sought to determine in healthy nonobese children, some of whom could be considered preobese on the basis of parental characteristics, whether different associations were apparent between PA, activity energy expenditure (AEE), and sensitive measures of body composition in prepubescence.
SUBJECTS AND METHODS
Subjects
One hundred healthy children (60 boys, 40 girls) participated. They were recruited into 2 groups according to their risk (high or low) of future obesity (Table 1). High-risk children had 1 biological parent with a body mass index (BMI; in kg/m2) >30, and low-risk children had 2 nonobese biological parents (BMI: <30).
View this table:
TABLE 1. Physical characteristics of the participants1
The children were recruited from schools in the Coleraine area of Northern Ireland (United Kingdom). The town of Coleraine has a mixed socioeconomic background, and the population of 55 000 is predominantly white European. Parents of eligible children (ie, those aged 6–8 y and living with biological parents) were first contacted by letter, after which those who expressed interest in participating in the study were interviewed in their homes to explain the study in detail. Parents were informed that the study concerned the measurement of energy expenditure and food intake of children. No direct reference was made to obesity because this could have biased the recruitment of subjects. All measurements took place during the school term and were conducted over a 3-y period.
The parents of each subject gave written informed consent to their child's participation in the study, and no subject who agreed to participate was subsequently excluded. The Ethics Committee of the University of Ulster approved the study.
Anthropometry
In the metabolic suite at the University of Ulster, body weight in a swimsuit was measured to the nearest 0.1 kg (Weylux Model 824/890; CMS Weighing Equipment, London, United Kingdom), and height was measured to the nearest 0.1 cm with the use of a stadiometer (CMS Weighing Equipment) and standardized procedures. The weight and height of both parents, in light clothing and not wearing shoes, were also measured under standardized conditions. BMI was calculated for the children and the parents. For the children, the international reference standard was used to define overweight and obesity (23). This standard is based on average centiles, which equate to a BMI of 25 or 30 at age 18 y for overweight or obesity, respectively. Here, obesity is defined as a BMI on or above the obesity cutoff, whereas overweight is a BMI on or above the overweight cutoff and below the obesity cutoff. Body fat mass in children was measured by isotope dilution during the measurements of energy expenditure by using the DLW method.
Body composition and activity energy expenditure
Total energy expenditure (TEE) was measured over 10 d by the DLW method. After collection of a predose urine sample, every child was given oral doses of 0.05 g 2H2O/kg body wt and 0.125g H218O/kg body wt. Further samples were collected 8 h after dose and, thereafter, at a known time daily for 10 d. Samples were stored at –4 °C before measurement. A detailed description of the measurement was previously published (24).
Besides being used to calculate TEE, the intercepts of the isotope disappearance curves were used to provide estimates of total body water. Body water was calculated as the mean of the time zero H218O distribution/1.01 and of the time zero 2H2O distribution space/1.04. Fat-free mass (FFM) was calculated from total body water by dividing the water content of fat-free tissue with age- and sex-specific values (25). Fat mass (FM) was calculated as the difference between body weight and FFM. For comparison between subjects, FM and FFM were divided by height squared and expressed as FM index (FMI) and lean mass index (LMI) to adjust for body size (26, 27).
It is difficult to measure basal metabolic rate (BMR) in young children, because they are often unable to lie at rest for the prolonged periods needed to make accurate measurements. Therefore, we used predictive equations based on the weight, height, sex, and age of the child to estimate BMR; these equations were previously shown to have good agreement with measured BMR (28–30). AEE was calculated as
RESULTS
The physical characteristics of the study participants are presented in Table 1. No significant interactions between sex and risk group were found. Children in the high-risk group (31 boys, 19 girls) had significantly greater weight (P < 0.05), BMI (P < 0.01), waist circumference (P < 0.05), %BF (P < 0.05), LMI (P < 0.01), and FMI (P < 0.05) than did children in the low-risk group (29 boys, 21 girls), but the former group did not differ significantly in age or height. In the high-risk group, 44% of fathers and 46% of mothers were classified as obese, whereas 10% of the high-risk children had both parents classified as obese. With the use of the International Obesity Task Force BMI cutoffs, none of the children were defined as obese, but significantly(P < 0.01) more children in the high-risk group (n = 11) than in the low-risk group (n = 1) were classified as overweight. Boys were significantly taller and had a significantly higher LMI than did girls (P < 0.05).
Physical activity intensity levels and patterning
Complete HR data were collected from 42 boys and 19 girls for a median of 8.75 h/d for a median of 6 d. No differences were observed in the body-composition characteristics of those who completed HR monitoring and those who did not, except that the girls who completed HR monitoring had a significantly (P = 0.009) higher FMI.
Activity energy expenditure and physical activity
On all tests of sex x risk group interactions for measures of energy expenditure and PA (P > 0.05), TEE and PAL were significantly higher in boys than in girls (P < 0.01). Although TEE was significantly higher in the high-risk group than in the low-risk group (P < 0.05), no differences in PAL were observed between the risk groups (Table 2). With a median of 388 kJ/d (IQR: –977, 79 kJ/d), TEE was significantly (P < 0.001) lower than the estimated energy requirements for an active lifestyle (31) but significantly (P < 0.001) higher than the estimate for low activity by 570 kJ/d (IQR: 58, 1117 kJ/d); however, this difference was not attributable to differences between the sexes or between the risk groups.
View this table:
TABLE 2. Measures of energy expenditure and physical activity in prepubertal children by sex and risk group1
Boys expended significantly more energy in activity than did girls, and 43% of the difference in AEE between the sexes was accounted for by FFM. No difference in AEE or PAL between the risk groups was found when adjusted for sex or when further adjusted for FFM.
No significant differences were observed between boys and girls or between the risk groups in either the median or IQR total amounts of time spent in light-intensity activity (boys: 25.9%; IQR: 11.1–55.8%; girls: 31.9%; IQR: 20.9–54.7%) (Table 2) or between the time spent in light-intensity activity on weekends and on weekdays. Although time spent in vigorous activities did not differ significantly between risk groups, the boys spent significantly more time overall in vigorous activities during the week (median: 6.1%; IQR: 3.2–9.7) than on the weekends (median: 1.0%; IQR: 0–4.1; P < 0.001). No differences in vigorous activities between weekdays and weekends were observed in girls.
PAL was not associated with time spent in either vigorous or light-intensity activity, nor was AEE associated with time spent in vigorous activity in boys or girls. Although AEE was positively associated with time spent in light-intensity activity (Spearman's ñ = 0.32, P = 0.01), this association was no longer significant after AEE was normalized for fat-free body mass (Spearman's ñ = 0.09) (Figure 1).
FIGURE 1.. Activity energy expenditure (AEE) calculated from doubly labeled water data for total energy expenditure and estimated basal metabolic rate and time spent in light-intensity activity (in %) in 100 prepubertal boys () and girls () [(A) r = 0.32, P = 0.01] and AEE adjusted for fat-free mass (in kg) [(B) r = 0.09, NS]. AEE was normalized for fat-free mass (AEE/fat-free mass1.3).
Physical activity and body composition
Boys who spent more than the median time in light-intensity activity (36.6%) had significantly higher FMI than did boys who spent less than the median time in light-intensity activities (P = 0.04 for sex x light-intensity activity category interaction) (Table 3). However, this association was not observed in girls. No associations between vigorous activity category (median: <6.6% and 6.6% of time per day) and body composition were found.
View this table:
TABLE 3. Body-composition measures by activity category measured by heart rate monitoring in prepubescent girls and boys1
Both AEE and PAL were significantly (P < 0.001) positively associated with FFM after adjustment for sex (ß-coefficient: 5.88 and 11.5; SE: 0.89 and 3.26, respectively) and were negatively associated with FMI after adjustment of the latter for sex and fat-free body mass (ß-coefficient: –1.41 and –5.25; SE: 0.44 and 1.19; P = 0.002 and < 0.001, respectively). The associations of AEE and PAL with FMI were further strengthened after adjustment for risk group (ß-coefficient: –1.53 and –5.57; SE: 0.43 and 1.16; P = 0.001 and < 0.001, respectively) (Table 4). The final AEE model adjusted for sex, FFM, and risk group explained 17% of the variance in FMI; AEE explained 59% of this variance (Figure 2). The final PAL model explained 24% of the variance in FMI, and PAL itself explained 71% of the variance (Figure 3). All analyses were repeated with FM and FFM that were normalized with the use of more specific powers for height obtained with the use of log-log regression analysis (P = 2.9 for FM and P = 2.2 for FFM). All the relations observed with the body-composition indexes remained essentially unchanged. These results are reproducible when weight is used instead of FFM as a covariate in the analyses.
View this table:
TABLE 4. Regression analyses of energy expenditure measures of activity with fat mass index models1
FIGURE 2.. Fat mass index and activity energy expenditure (AEE) assessed by the doubly labeled water method adjusted for sex, risk group, and fat-free mass (in kg) in prepubertal children (n = 100). AEE was log-transformed; ß-coefficient = –1.53, P = 0.001.
FIGURE 3.. Fat mass index and physical activity level (PAL) assessed by the doubly labeled water method adjusted for sex, risk group, and fat-free mass (in kg) in prepubertal children (n = 100). TEE, total energy expenditure; BMR, basal metabolic rate. PAL was log-transformed; ß-coefficient: –5.57, P < 0.001.
A higher BMI was significantly associated with higher AEE (ß-coefficient: 1.27; SE: 0.57; P = 0.02), but this association was attenuated, becoming nonsignificant and negative after control for FFM (ß-coefficient: –0.69; SE: 0.58). No association of BMI with PAL was observed, with or without adjustment for FFM.
DISCUSSION
In this study, we measured activity by 2 separate and objective methods that allowed both the accurate quantification of energy expenditure (ie, DLW) and the patterning and intensity of activity (ie, HR monitoring). In both boys and girls, consistent negative associations were observed between FMI and AEE and PAL. In addition, in boys, negative associations between a measure of fatness and light-intensity activity were found. TEE was significantly lower (5%) than the estimated energy requirements for children with an active lifestyle but significantly higher (8%) than the estimated requirements for children with low activity levels. Two other studies reported that mean TEE in prepubescent children is 9–13% lower than estimated energy requirements (13, 37). However, both of those studies used the older (1985) FAO/WHO/UNU recommendations for comparison, and those recommendations were based on reported dietary intakes rather than on DLW measures of TEE (38).
Children are at increased risk of future obesity when they have parents who are obese (1). We have shown that, even in nonobese prepubescent children, consistent body-composition differences exist between those with 1 obese parent and those with nonobese parents. In our study, significantly higher BMI, FMI, and LMI and significantly larger waist circumferences were found in the high-risk group than in the low-risk group. In a smaller study of 49 children aged 6–10 y, those with an obese mother had higher measurements of abdominal fat (in %) and lower measurements of FFM (in kg) as assessed by dual-energy X-ray absorptiometry than did the children with a lean mother. However, when values were expressed as %BF, no difference in fatness was found (39). A larger study in 88 girls aged 8 y also reported no difference in %BF (measured by dual-energy X-ray absorptiometry) between girls with lean parents, 1 obese parent, or 2 obese parents at baseline (40). After a 2-y follow-up, girls with 2 obese parents had significantly higher BMI and more FM (in kg) than did girls with 2 lean parents, but no differences in body composition from the girls with 1 obese parent were apparent. The variation in results between those studies could reflect real differences between the populations or in the assessment of body fatness, particularly the approach used for normalizing body size.
The association between PAL and FMI was stronger than that observed between AEE and FMI, after adjustment for sex, risk group, and FFM. Some (7, 12–16) but not all (17–19) studies have reported cross-sectional inverse associations between PAL and body fat. In a prospective study, Goran et al (41) found no inverse relation between AEE and change in FM, after adjustment for FFM, over a 4-y period in prepubescent children. These discrepant results could reflect real differences between populations, but, again, they are more likely to be a consequence of the different approaches used for normalizing body size. For example, when using PAL, it is often assumed that the potential influence of body size is resolved by adjustment for BMR, but, in our study, PAL was found not to be independent of body size. In fact, it was significantly correlated with both body weight and FFM in boys and girls (weight: r = 0.52 and 0.35; P < 0.001 and = 0.03; FFM: r = 0.59 and 0.46; P < 0.001 and <0.001, respectively). However, after adjustment for fat-free mass, the association between PAL and FMI was stronger (ß-coefficient: –2.16 and –3.03; SE: 0.69 and 0.69; P = 0.002 and < 0.001, respectively, without and with adjustment). Spadano et al (42) also recently concluded that PAL is not independent of body weight in children. This negative confounding could possibly explain why some studies found no differences in PAL between lean and obese children (5, 10, 11). It is well established that obese children expend more energy in a given activity than do nonobese children (9, 43). Consequently, when body size in energy-derived activity measures is not properly adjusted for, the higher energy cost of weight-bearing activities in heavier children could mask real differences in activity levels between lean and overweight children.
Similarly, in examinations of associations between activity and FM, adjustment for FFM is critical to preventing inappropriate conclusions. Interindividual variability in FFM is important, because, after adjustment for size, children vary substantially in FFM as well as in fatness (44). In the current study, no association was found between AEE and FMI before adjustment for FFM. In the case of BMI, if we had not adjusted AEE for FFM, we would have concluded from our data that there was a significant positive association with AEE. However, after that adjustment, the opposite conclusion was justified. This illustrates the real dilemma in interpreting BMI as a measure of body fatness, without proper recognition that weight is a mixed measure of FFM and FM.
With the use of accelerometry, Reilly et al (37) reported recently that young children spend a considerable time in sedentary activities. The number of hours of television viewing, as a marker of sedentary behavior, has been the most consistent risk factor associated with childhood obesity (45, 46). We showed that high levels of light-intensity activity are associated with higher fat tissue mass in boys, but no associations with body composition were found in girls. This finding could be due to the fact that a smaller number of girls completed HR monitoring or could reflect a real sex difference in the effect of light-intensity activity on fat accretion. A study of 28 boys aged 9 y also found an association between the time spent in sedentary activity measured by the HR flex method and the %BF measured by skinfold thicknesses (r = 0.46, P < 0.05) (20). However, the investigators did not find an association between AEE and body fat.
We observed that children aged 6–8 y preferred to spend most of their time in low-to-moderate activity, particularly on weekends, and this finding is supported by other studies (47). Data from 4-d accelerometry showed that boys and girls aged 4–6 y spent a mean of only 32 and 25 min, respectively, in vigorous activity per day (21). Despite this modest time in vigorous activity, a positive association with %BF was observed in some studies (16, 21, 22). Conversely, we found no associations between time spent in vigorous activity and measures of body fat, which indicates that, in this age group, an undue emphasis may be placed on vigorous activity in relation to risk of obesity. Although young children may be active, their activities tend to be of low intensity and typically are not sustained over extended periods. In adults, high PALs can be achieved if there is sufficient input of moderate activity (48). Perhaps it is not surprising, therefore, that we found no relation between vigorous activity and either AEE or PAL. In the current study, children spent a median of 55% and 37% of their waking hours in activities of moderate and light intensity, respectively. After adjustment was made for FFM, no relation was observed between time spent in light-intensity activities and PAL or AEE. There are several reasons why AEE might not correlate significantly with the time spent in light-intensity activity: such an association might be confounded by variations in the proportion of time spent in other categories of activity or by differences in the energetic efficiency with which specific activities are carried out.
This study shows the critical importance of adjustment for body composition and normalization for body size in energy-derived activity measures when assessing relations between measures of PA and body fatness. Although no significant differences in activity were apparent between prepubescent children at various risks of future obesity, even at this young age, distinct differences in measures of body fatness were found. Prospective studies are now needed to examine the relation between physical activity, both in terms of energy expenditure and activity patterns, and risk of excess fat gain.
ACKNOWLEDGMENTS
We thank Antony Wright for the stable isotope analysis.
MBEL, AMP, and SAJ were responsible for the study design; MBEL and AM were responsible for the acquisition of the data; and WAC was responsible for the stable isotope analysis. KLR conducted the statistical analysis. KLR, JCKW, and MBEL were responsible for the interpretation of the data and drafting the manuscript. All authors contributed to the revision of the manuscript. KLR is the guarantor of the study. None of the authors had any personal or financial conflicts of interest.
REFERENCES