Energy Expenditure in Lean and Obese Prepubertal Children

Transcription

1 Energy Expenditure in Lean and Obese Prepubertal Children James P. DeLany*f, David W. Harsha*, James C. Kime*, Julie Kumler*, Louis Melancon*, George A. Bray*f Abstract DELANY, JAMES P, DAVID W HARSHA, JAMES C KIME, JULffi KUMLER, LOUIS MELANCON AND GEORGE A BRAY. Energy expenditure in lean and obese prepubertal children. Obes Res. 1995;3[suppl1]: The relationship between energy expenditure and obesity was examined in prepubertal children. Consenting fifth graders underwent Tanner Staging, weight, height and skinfold measurements. Subjects were selected for further study to obtain equal numbers of girls and boys with a wide range of body composition. Weight, total daily energy expenditure (TDEE) by doubly labeled water (DL W), resting metabolic rate (RMR), and body composition were measured. Children were grouped into level of obesity based on tertues of subscapular plus triceps skinfolds. The sklnfold tertiles did quite wed in grouping subjects by degree of obesity, as differences in percent fat in each tertile were significantly different. There were no differences in fat-free mass between the groups, while the highest tertile group weighed 14 kg more than the lowest. For DL W, energy expenditure was calculated using day 8 and day 9 urine samples as the final time point to examine precision. Mean energy expenditure using either day was nearly Identical (222 ± 4 vs. 23 ± 37 kcalld), with a CV of the difference of 5.5%. No differences in RMR, energy expended In activity, or TDEE between the three groups were observed. A reduction in RMR or TDEE could not explain differences in obesity in these prepubertal children. However, the fact that the heaviest children expended the same amount of energy in activity and had the same TDEE as the leanest, while weighing 14 kg more, indicates From tlle *Pennington Biomedical Research Center, Baton Rouae. LA 788, and tlle tlsu Medical Center, New Orleans, LA Reprint requesta to Dr. DeLany, Pennington Biomedical Reaearch Center, 64 Perk:ins Road, Baton Rouge, LA 788. Copyright ()1995 NAASO. that the obese children had a reduced activity level. Key words: doubly labeled water, resting metabolic rate, physical activity Introduction The prevalence of obesity among US adults continues to increase, reaching 33.4% in 1991 (11). An increase in childhood obesity has also been reported (9), and investigators have suggested that the amount of time spent watching television is related to the degree of obesity (6). A recent 4-year follow-up of 54 obese children has shown that the degree of obesity in parents and grandparents as well as the degree of overweight in puberty were the most important factors related to weight in adult life (12). It is therefore important to study obesity in children as they go through puberty, to determine risk factors so that interventions can be initiated to prevent later obesity. In the past it has been difficult to determine exactly which portion of the energy balance equation was leading to obesity, whether an increase in intake or decrease in some parameter of energy expenditure. Studies using reported intake have suggested that daily energy intake is the same or less in obese compared to non-obese individuals, indicating that the obese had a greater energy efficiency or a lower physical activity than normal. However, using the doubly labeled water (DL W) method, it has been shown that obese individuals have a higher daily energy expenditure than lean controls, and that the obese underestimated their self-reported dietary intakes (1,13).. The introduction of DL W for the study of energy expenditure has added an important new tool for the study of obesity. The only requirement of subjects is to give urine and saliva specimens after drinking an initial do~ of 2 H2 18 o and then to give periodic urine samples dunng the next 1 to 2 weeks. During the period when energy expenditure is measured, subjects carry out their OBESITY RESEARCH Vol. 3 Suppl. 1 Mar

2 normal activities and are not required to maintain extensive diaries. The method bas been extensively validated under laboratory conditions (14) and is accurate with a precision of2% to 8%. When the DLW measurement of total daily energy expenditure (IDEE) is combined with indirect calorimetry for measurement of resting metabolic rate (RMR) and the thermic effect of food (1EF), a complete picture of energy expenditure is possible. Energy expended in physical activity is also obtained by subtracting RMR and 1EF from IDEE. We have begun a study to determine if reduced energy expenditure in prepubertal children is a marker for later weight gain using this approach. Energy expenditure has been studied in lean and obese adolescents using DL W to measure IDEE, and indirect calorimetry to measure RMR (2). In these 12- to 18-year-old adolescents, energy expenditure was higher in obese than nonobese children, indicating that reduced energy expenditure was not responsible for maintenance of obesity. This study did not rule out the possibility that at a younger age the obese children had a lower energy expenditure, or that reduced energy expenditure is a risk for later weight gain. We have begun a study in younger, prepubertal children, in which we will measure energy expenditure and body composition, with follow-up measurements 2 years later. Results reported herein are initial measurements in the frrst 46 children studied. Methods and Procedures All consenting fifth-grade children (age 1 to 11 years) were screened at local elementary schools for height, weight, Tanner Stage, subscapular and triceps skinfolds. Of the 92 children screened, 46 were selected for further study, with the final goal of obtaining equal numbers of boys and girls with the highest and lowest subscapular plus triceps skinfolds. Groups of children were brought to the Pennington Center for measures of body composition by dual energy x-ray absorptiometry (DEXA, Hologic QDR-2). The QDR-2, which employs low-dose alternating energy beams of 7 kilovolts peak (kvp) and 14 kvp beams, bas been shown to give accurate measures of body composition compared with underwater weighing and whole-body potassium-4 counting (15,16). Energy expenditure measurements were carried out at the school in a mobile lab with two metabolic carts with canopies (Sensormedics 29Z), and two beds, each with a television and video cassette player. Each child underwent a practice RMR determination to become familiar with the procedures to be carried out. Children were then scheduled for the RMR and 1EF measurement session and instructed to come to school fasted for 12 hours. The children were escorted to the mobile lab, rested quietly for 2 minutes, then underwent a 3-minute RMR. Subjects consumed a meal and Energy Expenditure, kcalld 4 D day 8 day oe!lo CiJ!±J ~ ~!] o o 1 Subject Figure 1: Total daily energy expenditure using day 8 and day 9 as final time point. 68 OBESITY RESEARCH Vol. 3 Suppl. 1 Mar. 1995

3 Table 1. Subject characteristics by ethnic group and gender Ethnic group African American White Gender M F M F n Age (years) 1.3 ± ± ± ±.5 Height (em) Weight (kg) Body Mass Index (kg/m 2) 142± 6 39±7 19±2 145 ± 6 41 ± 11 19±5 145 ± 5 37±6 19±6 139 ± 6 34± 8 19±6 then underwent a 3-hour measurement of TEF. The meal for the 1EF was 35% of the measured RMR, consisting of liquid Ensure. For DL W, children provided a urine sample for background 18o and 2H isotope abundances. The children then drank water containing.18 g of deuterium and.275 g 18o per kg total body water (TBW). Saliva samples were collected 2 and 3 hours post-dose for measurement of TBW. Urine collections were obtained the next morning for the initial time point, and then 8 and 9 days following dosing for the final time point. The 18o and 2H isotope abundances were measured in duplicate on a Finnigan MAT 252 gas-inlet Isotope Ratio Mass Spectrometer (5). Energy expenditure was calculated by the two-point method using the day 8 or day 9 urine samples as the final time point (5). Data were analyzed using ANOV A with the general linear models procedure because of unbalanced data (SAS 6.8 System for OS/2, SAS Institute, Cary, N.C ). Data are presented as least squares means± the standard error of the least squares mean. Post hoc t-tests were used to determine differences between least squares means. Similar results were obtained using means and the Least Significant Difference, Duncan's Multiple Range Test, or the Student-Newman-Keuls tests. Table 2. Subject characteristics by subscapular plus triceps skinfolds tertile Subscapular plus triceps tertile Results Collecting two final time point samples helped insure that there would be at least one good sample for analyses. In addition, this allowed us to examine the reproducibility of the DL W measurement. The mean IDEE obtained when comparing the 8 and 9 day urine samples as the final time point were very close (222 ± 4 vs. 23 ± 37 kcal/d). The coefficient of variation of the difference between the two time points was 5.5%. When the expenditure using the two time points is plotted for each subject, one can see the very close agreement between the two values, and in many cases the two numbers overlap (Figure 1). However, there are a few cases in which there is up to a 5 kcalld difference between the two. Of the 46 children studied to date, there are 24 boys and 22 girls. Fewer black children have been enrolled than white children (19 vs. 27). When examining the race by gender groups, although the subject characteristics are not significantly different (Table 1), there is an imbalance in the numbers in each group, and there are differences that nearly reached significance. Therefore, we cannot make comparisons between these groups for energy expenditure. Subjects were divided into groups by level of obesity based on the sum of the triceps and subscapular skinfold thicknesses. Children in the highest tertile weighed more than those in the lower tertiles (Table 2). The percent body fat was significantly different between each of the three tertile groups. Although BMI followed the same trend as percent body fat, differences between tertiles were not significant. There were no significant differences in fat-free mass (FFM) between tertile groups, although the children in the highest tertile had 2 kg FFM more than the lowest tertiles. There was no significant difference in TDEE between t11e three tertile groups, even though those in the highest tertile group weighed 14 kg more than those in the lowest tertile, and 11 kg more than those in the second tertile (Table 3). There were no significant differences between the groups in RMR, although the highest tertile group had a 75 kcalld greater RMR, consistent with the nonsignificant difference in FFM. The 1EF was Table 3. Energy expenditure by subscapular plus triceps tertile Height (em) Weight (kg~ BMI(kglm ) body fat(%) FFM (kg) 141 ± ± 1 17 ± 1 15 ± 1* 27 ± ± ± 1 18 ± 1 22 ± 1* 27 ± ± ± 1* 22 ± 1 29 ± 1* 29 ± 1 *Least squares means significantly different from other tertile groups. IDEE (kcavd) RMR(kcaVd) TEF (kcavd) Activity (kcavd) Subscapular plus triceps tertile ± ±5 1 ± 2 76 ± ± ±5 13 ± 2 95 ± ± ±5 16 ± 2 81 ± 8 OBESITY RESEARCH Vol. 3 Suppl. 1 Mar

4 : : _:.!:---- : -'-.,.... "r t : ~ ~------~ :f. 11t'-,f"'"~- : :- * : Total Daily EE /::;. 1st Tertile 2nd Tertile 3rd Tertile RMR Regression 1&2 Regression, All Tertiles weight, kg Figure 2: Total daily energy expenditure and resting metabolic rate by weight. enrichment of the final sample introduce a considerably greater error in energy expenditure than those of the ini tial enrichment sample. In addition, the CV observed in the present study includes real differences that may have occurred in energy expenditure between the 2 days (although they would be averaged out over 8 days). Even when using the two-point DL W method, collec tion of at least two final collection samples is very important, in case there is a bad sample, for whatever also similar between the three tertile groups. There was also no difference in energy expended in activity, as calculated by subtracting RMR and TEF from TDEE (Table 3). To further examine energy expenditure in the three groups, the data were plotted against weight (Figure 2). For TDEE, a regression line based on data from the first two tertiles was produced. When this is done, one can clearly see that most of the children in the highest tertile fell below the regression line (11 out of 15), indicating a reduced energy expenditure for their weight. No such trend was observed for RMR, as children in all groups were equally distributed around the overall regression line (Figure 2). Discussion There was very close agreement in TDEE observed when using the 8- and 9-day urine samples as the fmal time point, with a mean difference between the two values of only 8 kcalld. It was somewhat disappointing, however, that in a few cases the energy expenditure was different by as much as 5 kcalld. The 5.5% coefficient of variation of the difference is in agreement with previous findings. The CV from the present study only involves errors associated with the final time point, and not those associated with the baseline and initial enrichment samples. However, errors in measurement of. 7 OBESITY RESEARCH Vol. 3 Suppl. 1 Mar Table 4. Adjustments in energy expenditure for differences in body size Adjustment none /kg body weight lkgffm /RMR Covariate FFM Body weight Subscapular plus triceps tertile IDEE (kcavd) 214 ± ± ± ± ±9 229 ± ± ± ± ±.7 237±9 241 ± ± ± ± ± ±9 21 ± 1 *Least squares means significantly different from other tertile groups

5 reason (such as sample mislabeling, contamination or fractionation in the field or lab). There is also a backup sample for the initial time point built into the method, since we have the 3-hour saliva sample that can be used for the initial enrichment if there is a problem with the 24-hour urine sample. The fmding that there was no significant difference in RMR between the three groups was not surprising. Previous reports have demonstrated that absolute RMR is greater in obese children, but equal to that in lean individuals when expressed per kg of lean body mass or per unit surface area (7,8). Since there were no significant differences in FFM in the three groups of children by level of body fat in the present study, we did not expect any differences in RMR ~tween the groups. The fact that there were no significant differences in IDEE or activity between the heaviest children, who weighed 14 kg more than the leanest group, was quite surprising. These results are in contrast to our findings in slightly older children, in which the heaviest children had a higher RMR (due to higher FFM) and IDEE than the leanest children (4). The results for the current study suggest that the children with higher body fat spent less time engaging in physical activity than the leaner children, since it would take more energy for the heaviest children to carry out the same activity as a lighter child. That the heaviest children were less active for their body size is further illustrated in Figure 2, where the majority of the heaviest children fell below the regression line of energy expenditure by weight for the leaner children, while there was no difference for RMR. In a study of activity patterns measured by filming 5 fifth-grade students, obese children were shown to be significantly less active than nonobese children (3). One parameter that may be leading to inactivity in children is television viewing (6). Television viewing may reduce IDEE by reducing the amount of time spent in activity, but also by decreasing metabolic rate (1). In addition, there was a trend, although not significant, for obese children to show a larger decrease in RMR while watching television (1). When comparing energy expenditure between individuals of differing body size and composition, one must make adjustments. What the appropriate adjustment is has been a question for many years, and except for RMR, remains unresolved. For RMR, it has recently been established that the metabolically active FFM is the appropriate parameter to make adjustments for comparisons, even though FFM accounts for only about 4% to 6% of the variance in RMR. One reason it took so long for FFM to be used in comparing RMR is that FFM has not been routinely measured until recently. There is also fairly good agreement that one cannot simply divide RMR by FFM, since the intercept does not go through zero. The question of the appropriate adjustment to make when comparing IDEE between groups is complicated by the fact that this parameter also contains RMR. Therefore, simply dividing by body weight or FFM may not be appropriate. In addition, whereas fat mass is generally considered not to be metabolically active for RMR, fat mass is added weight, which will increase energy requirements during locomotion. Therefore, adjusting for body weight may be more appropriate than FFM for IDEE. When IDEE is divided by body weight, the heaviest children had a significantly lower energy expenditure than the leaner two groups, while dividing by FFM or RMR indicated no differences (Table 4). Using FFM or body weight as a covariate in analysis of variance, which would get around potential problems of non-zero intercepts, also gave no differences in IDEE between the three groups. In summary, in this study of prepubertal children, levels of IDEE, RMR, energy expended in physical activity, and TEF were similar in groups based on tertiles of fatness. No difference in RMR was expected, since the children in each group had similar levels of fat-free mass. However, the fact that physical activity and IDEE were similar in all groups was surprising, since the children in the highest tertile group weighed 14 kg more than those in the leanest group, and 11 kg more than the intermediate group. This indicates that the obese children either spent less time engaging in physical activity, or engaged in less vigorous activity, since it would take more energy for the heavier children to carry out the same activity as the leanest children. Acknowledgments Supported by NIH, NICHHD #ROt HD Rererences 1. Bandlnl LG, Schoeller DA, Cyr HN, Dietz WH. Validity of reported intake in obese and nonobese adolescents. Am J Clin Nutr. 199;52: Bandlnl LG, Schoeller DA, Dietz WH. Energy expenditure in obese and nonobese adolescent. Pediatr Res. 199;27: Corbin CB, Pletcher P. Diet and physical activity patterns of obese and nonobese elementary school children. Res Quart. 1968;39: DeLany JP, Bray GA. Energy expenditure of black and white 6th grade boys and girls. FASEB J. 1993;7:A DeLany JP, Schoeller DA, Hoyt RW, Askew EW, Sharp MA. Field use of n 2 18o to measure energy expenditure of soldiers at different energy intake. J Appl Physiol. 1989;67: Dietz WH, Gortmaker SL. Do we fatten our children at the 1V set? Television viewing and obesity and adolescents. Pediatrics. 1985;75: OBESITY RESEARCH Vol. 3 Suppl. 1 Mar

6 7. Epstein LH, Wing RR, Cluss P, et al. Resting metabolic rate in lean and obese children: relationship to child and parent weight and percent-overweight change. Am J Clin Nutr. 1989;49: Freymond D, Larson K, Bogardus C, Ravussln E. Energy expenditure during normo- and overfeeding in peripubertal children of lean and obese Pima Indians. Am J Physiol. 1989;257:E647-E Gortmaker SL, Dietz WH, Sobol AM, Wehler CA. Increasing pediatric obesity in the United States. Am J Disab Child. 1987;141 : Klesges RC, Shelton ML, Klesges LM. Effects of television on metabolic rate: potential implications for childhood obesity. Pediatrics. 1993;91: Kuczmarskl RJ, Flegal KM, Campbell ST, Johnson CL. Increasing prevalence of overweight among US adults. lama. 1994;272: Mossberg H-. 4-year follow-up of overweight children. lancet. 1989;2: Prentice, A, Black AE, Coward WA, et al. High levels of energy expenditure in obese women. BMJ. I 1986;292: Schoeller DA. Measurement of energy expenditure in freeliving humans by using doubly labeled water. J Nutr. 1988;118: Slosman DO, Casez JP, Plchard C, et al. Assessment of whole-body composition with dual-energy x-ray absorptiometry. Radiology. 1992;185: Snead DB, Birge SJ, Kohrt WM. Age-related differences in body composition by hydrodensitometry and dual-energy X-ray absorptiometry. J Appl Physiol. 1993;74: OBESITY RESEARCH Vol. 3 Suppl. 1 Mar. 1995