点击显示 收起
1 From the Center for Human Nutrition, Johns Hopkins Bloomberg School of Public Health, Baltimore (MST); the US Department of Agriculture/Agricultural Research Service Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston (NFB); and the University of Maryland, Department of Medicine, Division of Gerontology, Baltimore VA Medical Center, Baltimore (JDS).
2 The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.
3 Supported by NIH grant R29 HD34029 (to MST), the USDA/ARS under Cooperative Agreement no. 6250-51000-023-00D/01 (to MST and NFB), and the University of Maryland GRECC and Claude D Pepper Older Americans Independence Center (to JDS).
4 Reprints not available. Address correspondence to MS Treuth, Center for Human Nutrition, Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe Street, Baltimore, MD 21205. E-mail: mtreuth{at}jhsph.edu..
See corresponding editorial on page 1051.
ABSTRACT
Background: Conflicting evidence exists on the causal factors underlying the development of excess adiposity in children.
Objective: We determined the effect of energy expenditure (EE), muscle energetics, and physical fitness on weight and fat gain in prepubertal girls with or without a predisposition to obesity.
Design: Normal-weight girls (n = 101) were recruited at 8 y of age according to parental body mass index. Eighty-eight girls completed the 2-y study, and the groups were as follows: LN, girls with 2 lean parents; LNOB, girls with 1 obese and 1 lean parent; and OB, girls with 2 obese parents. Measurements of weight, height, and body composition were taken 1 and 2 y after baseline. Girls underwent baseline measurements of EE by 24-h calorimetry and doubly labeled water, of muscle metabolism by 31P nuclear magnetic resonance, and of fitness.
Results: Fat mass (FM) and percentage body fat (%BF) differed significantly between the groups at years 1 and 2; the OB group had higher FM (P = 0.03) and %BF (P = 0.046) at year 1 and higher FM (P = 0.047) at year 2 than did the LN group. After adjustment for baseline weight, group, time, ethnicity, and Tanner stage, sleep EE, basal EE, 24-h EE, and peak oxygen uptake were negatively associated with FM and %BF (P < 0.04). After adjustment for the same variables, muscle oxidative capacity and free-living total EE were negatively and positively predictive, respectively, of changes in %BF between 8 and 10 y of age (both P = 0.04).
Conclusions: Nonobese girls with 2 obese parents have a significant risk of developing obesity. High free-living total EE and low muscle oxidative capacity predict high rates of fat gain.
Key Words: Obesity • children • energy expenditure • fitness • body composition • physical activity
INTRODUCTION
The prevalence of obesity in children in the United States has been increasing, and 22–27% of US children are currently estimated to be overweight or obese (1). Evidence suggests that both environmental and genetic factors play an important role in the development of obesity.
The relative contributions of energy intake and energy expenditure (EE) to the increased prevalence of childhood obesity are unclear. Conflicting evidence on the contribution of basal energy-requiring processes, physical activity, and total EE (TEE) to the etiology of childhood obesity has yet to be resolved. Several well-designed studies of children at risk of becoming obese by virtue of parental obesity yielded conflicting results on whether EE influences excess adiposity development. Two studies that showed an association between EE and later weight gain indicated that the observed differences in TEE between subjects were partially accounted for by differences in activity EE (2, 3). However, other studies in children reported that no component of EE accounted for the longitudinal changes in fat mass (FM) observed in infants (4) and in boys and girls (5–7). However, low fitness was shown to predict an increase in adiposity in young children (6).
To differentiate the causes of obesity from the consequences, children need to be studied before major increases in adiposity develop. Because it is known that children of obese parents are more likely to gain excess weight (8), we studied children who were initially nonobese so that we could examine factors that might contribute to the children becoming overweight. We enrolled normal-weight, 8-y-old girls with or without a predisposition to obesity [classified by parental body mass index, (BMI; in kg/m2)] into a longitudinal cohort study and followed them for 2 y. At baseline, we reported no significant differences in weight, body composition, EE (basal, sleep, 24-h, free-living, and activity) (9), fitness and physical activity (10), and muscle energy metabolism (11) in the 8-y-old girls with 2 lean parents, 1 lean and 1 obese parent, or 2 obese parents. This longitudinal study was designed to examine whether EE, skeletal muscle energetics, and physical fitness are related to attained values for and changes in weight, FM, and percentage body fat (%BF) in girls aged 8–10 y with or without a predisposition to obesity.
SUBJECTS AND METHODS
Subjects
Healthy prepubertal girls aged 8–9 y (n = 101) were recruited from the local Houston area to participate in the study. The ethnicity of the girls (47 whites, 27 African Americans, and 14 Hispanics) was determined from the self-reported ethnicity of the mothers. All the girls were at Tanner stage 1 at baseline. The girls were screened and had to be below the 90th percentile of their weight-for-height (12) and have a %BF of 12–30% (13) to be included in the study. Girls with cardiovascular disease, anemia, diabetes, significant renal or hepatic disease, hypothyroidism, or musculoskeletal problems were excluded. The girls’ parents provided written informed consent, and the girls provided assent to participate in this study, which was approved by the Institutional Review Board for Human Subject Research for Baylor College of Medicine and Affiliated Hospitals.
Study protocol
All measurements were performed at the US Department of Agriculture/Agricultural Research Service Children’s Nutrition Research Center. Measures of body composition, EE, skeletal muscle energy metabolism, fitness, and Tanner staging were taken at baseline when the girls were 8 y of age. Anthropometric, body composition, and Tanner staging measures were repeated at 9 and 10 y of age.
Body composition
The girls’ body weight was measured to the nearest 0.1 kg by using a digital balance (Scale-Tronix, Dallas), and their height was measured to the nearest 1 cm by using a stadiometer (Holtain Ltd, Crymych, Pembs, United Kingdom). BMI was defined as weight (kg)/height2 (m). The parents were measured in the same manner, and parental BMI was used to classify the girls into the following groups: LN, girls with 2 lean (BMI < 25) parents; LNOB, girls with 1 obese (BMI > 28) and 1 lean parent; and OB, girls with 2 obese parents.
Body composition was measured by dual-energy X-ray absorptiometry (DXA) (QDR 2000, pencil beam mode, software 5.56; Hologic, Madison, WI). For DXA, each subject lied on the bed in a supine position and was scanned from head to toe in 10–15 min. DXA allows for determination of bone mineral density and 3 compartments: lean tissue mass, FM, and bone mineral content. For the total body, fat-free mass (FFM) was defined as the sum of lean tissue mass and bone mineral content.
Tanner staging
The mothers were shown pictures of the various Tanner stages (14, 15) and identified their daughters’ stage at each yearly visit.
Energy expenditure, muscle energy metabolism, and fitness
Baseline values of and measurement procedures for EE, skeletal muscle energy metabolism, and fitness were described in detail previously (9–11), but brief descriptions are also provided here. The subjects’ fitness was measured in the morning after they arrived at the center in a fasted state. Next, the girls underwent nuclear magnetic resonance measurements. This was followed by 24-h calorimetry studies. During the children’s visit to the research center, which lasted 1.5 d, all their meals were provided to them.
The girls spent 24 h in a whole-room calorimeter in which basal, sleep, exercise, and 24-h EE values were measured. Exercise intensity was based on the child’s peak oxygen uptake (
Nuclear magnetic resonance measurements were performed on the Bruker Biospec Imaging/Spectroscopy System (Bruker Instruments, Billerica, MA) equipped with a 2.4-T, 40-cm, horizontal-bore superconducting magnet. After the baseline spectra were obtained, the child began plantar flexion exercise at the rate of 1 complete extension or flexion/4 s for 180 s. 31P spectra were obtained from raw data by using Fourier transformation of the free induction decay into the frequency domain, with 10-Hz line broadening to enhance signal-to-noise (NMR1 software; New Methods Research, Inc, Syracuse, NY). Changes in inorganic phosphate (Pi) and phosphocreatine (PCr) within the gastrocnemius and soleus muscles were assessed before, during, and after exercise. The mean Pi/PCr value during exercise was used in the analyses.
To assess physical fitness,
One-factor analysis of covariance was used to compare anthropometric and body composition measures between the groups of girls (based on the obesity phenotypes of their parents) after adjustment for the girls’ ethnicity and Tanner stage. Separate models were used for baseline and years 1 and 2 of the study. The Dunnett-Hsu procedure was used to adjust for multiple comparisons. As published previously (9–11), baseline values for EE, muscle oxidative capacity, and fitness in the subset of 88 girls who completed the longitudinal study (from the original 101 girls) are shown in Table 1.
View this table:
TABLE 1. . Baseline measures of energy expenditure (EE), muscle energy metabolism, and fitness in the prepubertal girls who completed the longitudinal study1
Mixed-effects analyses of variance (generalizations of repeated-measures analyses of variance) (SAS PROC MIXED, version 8.02; SAS Institute Inc, Cary, NC) were used to determine the relations of EE, muscle oxidative capacity, and fitness measured at baseline to weight and to FM and %BF over the course of the 3-y study (21). The models were adjusted for time of measurement (baseline, year 1, or year 2), ethnicity, Tanner stage, group (LN, LNOB, or OB), group x time interaction, and the predictor variable being examined. Use of the group x time interaction allowed us to examine each group’s passage through time separately, ie, it allowed us to examine changes in the dependent variable over time separately in each of the 3 groups of girls. Time, ethnicity, group, and Tanner stage were entered in the models as categorical variables. The models accounted for the serial autocorrelation of repeated measures on the same subject by modeling the within-subject covariance. Several models were run for each analysis, but the only difference between the models was the structure assumed for the covariance. The covariance structures that were considered included an unstructured one, compound symmetry structures, and first-order autoregressive structures. From the several models, a final model (and covariance structure) was chosen by reviewing each model’s residual plot and by comparing the models’ Akaike’s information criteria and Schwarz’s Bayesian information criteria. The Dunnett-Hsu procedure was used to adjust for multiple comparisons.
The relation of EE, muscle metabolism, and fitness measured at baseline to changes in FM and %BF was examined by using mixed-effects analysis of variance (a generalization of repeated-measures analysis of variance). A maximum of 2 changes were recorded for each girl: the difference between year 1 and baseline and the difference between years 2 and 1. The analysis included adjustment for group, ethnicity, time of initial measurement (baseline for the difference between year 1 and baseline, and year 1 for the difference between years 2 and 1), and Tanner stage at the time of initial measurement. Time, ethnicity, group, and Tanner stage were entered in the models as categorical variables as described above. In addition, the model included the value of the dependent variable at the time of the initial measurement, the predictor variable being examined, and an interaction between group and time of initial measurement. Use of the interaction allowed us to examine each group’s progress through time separately. Initial analyses included a second interaction: the interaction between group and the predictor being examined. This term allowed the 3 groups to respond differently to the predictor. This interaction was never significant and was dropped from the models. The models accounted for the serial autocorrelation between repeated measures on the same subject by modeling the within-subject covariance as described above. The Dunnett-Hsu procedure was used to adjust for multiple comparisons.
SAS version 8.2 (SAS Institute Inc) was used for all data analyses. Microsoft ACCESS (WINDOWS 95, version 7.0; Microsoft, Seattle, WA) was used for database management. Data are presented as means ± SDs, and a two-tailed P < 0.05 was taken as indicating significance.
RESULTS
Subject characteristics
The final sample at year 2 included 88 prepubertal girls (47 whites, 27 African Americans, and 14 Hispanics). Thirteen girls were lost to follow-up because their families moved from the area, and a few of the girls were missing DXA measures at year 1 because of equipment malfunctions. By year 1, 48%, 51%, and 1% of the girls were at Tanner stages 1, 2, and 3, respectively. By year 2, 20%, 56%, 16%, and 8% of the girls were at Tanner stages 1, 2, 3, and 4, respectively.
At baseline, the 3 groups of girls did not differ significantly in weight, FM, FFM, or %BF (Table 2). The 3 groups also did not differ significantly in weight at year 1 or 2. The girls who had 2 obese parents (the OB group) had significantly higher BMI (adjusted for ethnicity and Tanner stage) than did the girls who had 2 lean parents (the LN group) at baseline and at years 1 and 2. At year 1, the girls in the OB group had significantly more FM (P = 0.03) and significantly higher %BF (P = 0.046) than did the LN group. At year 2, the girls in the OB group had significantly more FM (P = 0.047) and tended to have higher %BF than did the girls in the LN group (P = 0.065). The FM and %BF values of the girls in the LNOB group were not significantly different from those of the girls in the LN group. FFM did not differ significantly between the groups across the study years. No significant correlations were observed between the girls’ changes in weight, FM, FFM, and %BF from baseline to year 2 and the corresponding maternal or paternal changes (data not shown).
View this table:
TABLE 2. . Comparison of anthropometric and body-composition measures at baseline and at years 1 and 2 between the 3 groups of girls1
Predictors of fat mass, %BF, and changes in fat mass and %BF
The relations of measures of EE, muscle energy metabolism, and fitness to weight are shown in Table 3, and the relations of those measures to FM and %BF are shown in Table 4. The measures of EE included basal EE, sleep EE, 24-h EE, nonprotein respiratory quotient, fat and carbohydrate oxidation (all of the foregoing were measured in a calorimeter), free-living EE (total EE, activity EE, and PAL), a muscle metabolism variable (Pi/PCr during exercise), and fitness (
View this table:
TABLE 3. . Measures of energy expenditure (EE), muscle oxidative capacity, and fitness at baseline and their association with weight in prepubertal girls1
View this table:
TABLE 4. . Measures of energy expenditure (EE), muscle energy metabolism, and fitness at baseline and their association with fat mass and percentage body fat (%BF) in prepubertal girls1
The measures of EE, muscle energy metabolism, and fitness were also examined to see whether they predicted changes in FM or %BF in the girls (Table 5). None of the interactions explored were significant. Only Pi/PCr during exercise and total EE were significant predictors of changes in %BF (both P = 0.04). Similar trends were observed for the prediction of changes in FM by basal EE (P = 0.09) and Pi/PCr during exercise (P = 0.07).
View this table:
TABLE 5. . Measures of energy expenditure (EE), muscle energy metabolism, and fitness at baseline as predictors of changes in fat mass and percentage body fat (%BF) in prepubertal girls1
DISCUSSION
In this study, we examined potential factors underlying the development of excess adiposity in prepubertal girls with a familial predisposition to obesity. At 8 y of age, the girls underwent measurements of EE by 24-h calorimetry and doubly labeled water, of muscle energy metabolism by 31P nuclear magnetic resonance, and of fitness by the
Changes in adiposity over time in prepubertal girls were also examined by Figueroa-Colon et al (5). The gains in FM (adjusted for FFM) were 1.2 and 3.3 kg after 1.6 and 2.7 y of follow-up, respectively. The girls in that study actually lost body fat (-0.8%) during the study, whereas the girls in the present study gained body fat. The differences in results between the 2 studies may have been due to the higher initial body fatness (26.8% compared with 21.7%), younger and wider age range (5–9 y compared with 8–10 y), and smaller sample size (n = 49 compared with n = 88) in the study by Figueroa-Colon than in the present study.
We found that EEs while sleeping, at rest, and over 24 h were associated with FM and %BF in the girls, whereas activity EE, PAL, substrate oxidation, and Pi/PCr were not. Total EE was associated with FM but not with %BF. The coefficients were negative, such that high sleep, rest, 24-h, and total EEs in the prepubertal girls were associated with less FM. Fitness was also negatively associated with FM and %BF in these girls. Similarly, low fitness predicted increasing adiposity in children aged 4.6–11 y at baseline (6).
These results are in contrast with those of 3 longitudinal studies that examined predictors of fat gain, although in different samples of children (5–7). In a study of premenarcheal girls (5), EE measured in a room calorimeter was not a predictor of changes in body fat; however, the fathers’ total FM and %BF were predictors. In 2 other studies using calorimetry and doubly labeled water (6, 7), no component of EE (rest, free living, or activity) predicted fat gains in white or African American boys and girls. Again, these differences may be due to differences in samples and study design; our study had a relatively large sample size for the types of unique measures involved, and we included only nonobese girls in a narrow age range at baseline. It is possible, however, that our criterion of including only nonobese girls may have limited our potential to find group differences.
Our results indicated that substrate oxidation did not predict increases in adiposity. Conflicting information exists in the literature regarding the role of substrate oxidation as a predictor of increasing adiposity. In studies of adult Pima Indians (23) and in the Baltimore Longitudinal Study on Aging (24), substrate oxidation (fat and carbohydrate oxidation) predicted weight gain. In contrast, in studies of children, respiratory quotients did not play a role in increases in adiposity (6, 7). In prepubertal children, exogenous fat oxidation was directly related to FM (25).
No previous studies of which we are aware examined the effect of Pi/PCr during exercise on fat gain in children, and therefore there is no comparative data in the literature. The Pi-PCr ratio is an indication of muscle oxidative capacity and could serve as a proxy variable influenced by physical activity or exercise training. Thus, both muscle energy metabolism and overall total EE are factors that are intertwined with physical activity, and the rate of weight gain is related to the metabolic economy of muscle (26). Clearly, public health messages should promote regular physical activity for children. Interventions to promote changes in adiposity should promote increases in overall daily EE and physical activity and focus on activities to increase fitness.
In conclusion, we found that normal-weight girls with 2 obese parents have a significant risk of developing obesity. Several factors underlying the development of excess adiposity in children with a familial predisposition to obesity were identified. Low EE during sleep, at rest, and over a 24-h period and low physical fitness were associated with high FM and %BF in the prepubertal girls. High free-living total EE and low muscle oxidative capacity predicted high rates of fat gain.
ACKNOWLEDGMENTS
We thank the children and parents who participated in the study, the Metabolic Research Unit staff and body composition staff for technical assistance, and B Kertz for subject recruitment.
MST contributed to the study design, data collection, data analysis, and the writing of the manuscript; NFB contributed to the study design, data collection, and the writing of the manuscript; and JDS contributed to data analysis. None of the authors had any financial or personal conflicts of interest in any company sponsoring the research.
REFERENCES