- Email: [email protected]
Fat oxidation in nonobese and obese adolescents: Effect of body composition and pubertal development
F
Fat oxidation in nonobese and obese adolescents: Effect of body composition and pubertal development
Dénes Molnár, MD, PhD, and Yves Schutz, PhD Objectives: To measure postabsorptive fat oxidation (Fox) and to assess its association with body composition (lean body mass [LBM] and body fat mass [BFM]) and pubertal development.
Design: We studied 235 control (male/female ratio = 116/119; age [mean ± SD]: 13.1 ± 1.7 years; weight: 45.3 ± 10.5 kg; LBM: 34.3 ± 7.1 kg; BFM: 11.0 ± 4.5 kg) and 159 obese (male/female ratio = 93/66; age: 12.9 ± 2.1 years; weight: 76.2 ± 19.1 kg; LBM: 47.4 ± 10.9 kg; BFM: 28.8 ± 9.2 kg) adolescents. Postabsorptive Fox was calculated from oxygen consumption, carbon dioxide production, and urinary nitrogen as measured by indirect calorimetry and Kjeldahl’s method, respectively. Body composition was determined by anthropometry. Results: Postabsorptive Fox (absolute value and percentage of resting metabolic rate) was significantly (p < 0.001) higher in the obese adolescents (76.7 ± 26.3 gm/24 hours, 42.3% ± 18.7%) than in the control subjects (40.0 ± 26.3 gm/24 hours, 28.7% ± 17.0%), even if adjusted for LBM. Fox corrected for BFM was similar in control and in obese children, but was significantly lower in girls compared with boys (control male subjects: 62.1 ± 29.1 gm/24 hours, control female subjects: 51.6 ± 28.4 gm/24 hours, obese male subjects: 57.3 ± 29 gm/24 hour, obese female subjects: 45.0 ± 28.4 gm/24 hours). BFM and LBM showed a significant positive correlation with Fox. By stepwise regression analysis the most important determinant of Fox was BFM in obese and LBM in control children. There was a significant rise in Fox during puberty; however, it was mainly explained by changes in body composition. Conclusions: Obese adolescents have higher Fox rates than their normal-weight counterparts. Both LBM and fat mass are important determinants of Fox. (J Pediatr 1998;132:98-104)
It has been proposed that adipose tissue plays a significant role in the regulation of energy intake and balance.1 This concept was supported by the observation that in
obese women subjected to energy restriction, resistance to further weight loss developed when fat-cell size reached that of lean subjects.2 A positive fat balance does not en-
From the Department of Pediatrics, University Medical School of Pécs, Pécs, Hungary, and the Institute of Physiology, University of Lausanne, Lausanne, Switzerland. Supported by a research grant from Nestle Nutrition and Hungarian National Research grant (OKTA 5267 and 16065). Portions of the data have been reported in abstract form (Int J Obes 1995;19(suppl 12):42 and Pediatr Res 1995;38:445). Submitted for publication Feb. 13, 1996; accepted April 29, 1997. Reprint requests: Dénes Molnár, MD, PhD, Department of Pediatrics, University Medical School of Pécs, József A. u. 7., H-7623 Pécs, Hungary. Copyright © 1998 by Mosby, Inc. 0022-3476/98/$5.00 + 0 9/21/83009
98
hance fat oxidation acutely,3 but rather through a slower increase in fat stores, which in turn raises the Fox rate by causing insulin resistance.4 According to this theory the accumulation of body fat may be regarded as a compensatory mechanism by which the body increases Fox and reaches new energy and fat balance. Studies in support of this hypothesis in adults5,6 and in prepubertal children7,8 demonstrated higher lipid oxidation in obese than in lean subjects, and positive correlation was found between body fat mass and fasting Fox, thus supporting the theory that enlargement of body fat store is accompanied by an increased Fox. More recent studies in aging women9 and in healthy white subjects,10 however, failed to find a relationship between Fox and fat mass. Moreover, Nagy et al.10 suggested that the positive correlation between Fox and fat mass demonstrated earlier was related to the much greater range of fat mass compared with that of fat-free mass of the cohorts investigated in these studies.5,6 BFM Fox LBM RMR RQ WHR
Body fat mass Fat oxidation Lean body mass Resting metabolic rate Respiratory quotient Waist/hip ratio
The aims of the present study were to (1) investigate substrate utilization in obese and nonobese prepubertal and adolescent children, (2) analyze the effect of body composition and body fat distribution on Fox, (3) determine how puberty, when dramatic changes in lean body mass and BFM occur,11-13 affect Fox. We have studied a carefully characterized large cohort of obese and normal-weight, prepubertal and adolescent children of both genders with a wide range of both BFM and LBM.
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
MOLNÁR AND SCHUTZ
Table I. Physical characteristics of subjects
Group Control (n = 235) Male (n = 116) Female (n = 119) Obese (n = 159) Male (n = 93) Female (n = 66)
Age (yr)
Weight (kg)
Height (cm)
Relative weight (%)
BFM (kg)
BF (%)
LBM (kg)
WHR
CPR
13.1 ± 1.7 44.5 ± 11.6 13.1 ± 1.7 46.0 ± 9.3
157.6 ± 13.2 157.3 ± 8.9
94.0 ± 10.4 95.9 ±11.2
9.6 ± 3.9 12.4 ± 4.6
21.2 ± 4.6 26.2 ± 5.5
34.9 ± 8.6 33.6 ± 5.4
0.82 ± 0.05 0.74 ± 0.05
0.62 ± 0.2 0.64 ± 0.2
12.8 ± 1.8 75.4 ± 18.8* 13.2 ± 1.9 77.4 ± 19.5*
160.0 ± 11.9 153.4 ± 23.4* 26.7 ± 7.8* 158.7 ± 10.2 155.0 ± 22.8* 31.7 ± 10.3*
35.1 ± 3.0* 48.7 ± 11.5* 0.89 ± 0.05* 40.3 ± 4.4* 45.6 ± 9.8* 0.81 ± 0.06*
0.99 ± 0.2* 0.92 ± 0.2*
CPR, Central/peripheral skinfold ratio. *p < 0.001 obese versus control group (Student’s test).
METHODS Subjects The study included the data of 235 nonobese (116 male) and 159 obese (93 male) Hungarian white children whose ages ranged from 9.5 to 16.5 years. Most of the subjects participated in a previous study exploring the validity of equations currently used for the prediction of resting metabolic rate.14 Anthropometric measurements and the determination of RMR and fuel oxidation were carried out in all subjects. Blood samples for the analysis of plasma insulin and glucose concentrations were obtained only from a subset of patients (58 control and 94 obese subjects) who volunteered to have venipuncture. The age, weight, height, and body composition of the subset were similar to that of the whole cohort. The cohort of normal-weight children (body weight less than 120% of the expected weight for height15; triceps and subscapular skin-fold thickness <90th percentile of the reference values for age and sex15) was recruited from primary and high schools. Obese children (body weight exceeding predicted body weight for height by 20% or more; triceps and subscapular skinfold thickness >90th percentile of the reference values for age and sex15) were recruited from patients referred to the obesity clinic of the Department of Pediatrics, University Medical School of Pécs. Pubertal stage was assessed according to Tanner.16 Subjects with diabetes mellitus or any other metabolic or endocrine disease were excluded. Physical examination and routine laboratory tests documented the
absence of any health problems. The subjects included in the study did not take medicine, consume alcohol, or smoke cigarettes regularly. These data were obtained from a questionnaire administered by a trained assistant. Descriptive data are listed in Table I. The study was approved by the ethical committee of the University of Pécs, and the investigations were carried out in accordance with the Declaration of Helsinki II.
Body Composition Anthropometric measurements were carried out by the same investigator and included weight, height, waist and hip circumference, and skin-fold thickness. The child’s weight was obtained in light clothing (shorts or dress and tee-shirt, without shoes) to the nearest 0.1 kg on a standard beam scale. Height was measured to the nearest 0.1 cm by a Holtain stadiometer without shoes. Body mass index was calculated by dividing weight (in kilograms) by height (in square meters). Skin-fold thicknesses (triceps, biceps, suprailiac, subscapular, and calf skin folds) were measured in triplicate, to the nearest millimeter, on the left side of the body by using Holtain caliper. The formula derived by Parizkova and Roth17 was used to calculate the percentage of body fat from the sum of five skin folds. LBM was calculated by subtracting fat mass from body weight. Body fat distribution was assessed by the waist/hip ratio and by the ratio of central skin folds (subscapular, suprailiac) to extremity skin folds (biceps, triceps and calf). WHR was measured as the minimal waist circumference to the
circumference of the maximal gluteal protuberance.
Experimental Design Before the gas exchange measurements children received an unrestricted diet. The day before the test they did not perform any intense physical exercise. The patients arrived at the laboratory approximately at 7:30 AM after an overnight fast (from 8 PM the day before). The lifestyle questionnaire was completed and a blood sample was obtained from the antecubital vein for the determination of plasma insulin and glucose. Subjects rested on a comfortable bed, under the canopy in a temperaturecontrolled room (25° to 27° C) for at least 30 minutes before the commencement of the gas exchange measurements to accommodate to the environment. During the investigation the children were watching nonviolent cartoons to avoid spontaneous movements and to reduce unnecessary nervousness that might have increased energy expenditure.
Energy Expenditure Measurements and Calculation of Fuel Oxidation Rate Resting metabolic rate was measured for 45 minutes with the help of a Deltatrac metabolic cart (Datex Instrumentarium Corp., Helsinki, Finland), using the ventilated hood technique as described earlier.14 Before each test the calorimeter was calibrated with a reference gas mixture (95% oxygen, 5% carbon dioxide). Ethanol burning tests were performed to assess the precision of the indirect calorimeter. The mean ± SD of measured respiratory quotient was 0.668 ± 0.008 99
MOLNÁR AND SCHUTZ
THE JOURNAL
OF PEDIATRICS JANUARY 1998
Table II. Resting metabolic rate, respiratory quotient, and serum insulin and glucose levels in obese and control adolescents
Group Male subjects Control (n = 116) Obese (n = 93) Female subjects Control (n = 119) Obese (n = 66)
Resting metabolic rate (kJ/24 hr)
Resting metabolic rate* (kJ/24 hr)
Respiratory quotient
Nonprotein respiratory quotient
Immunoreactive insulin† (µU/ml)‡
Glucose (mmol/L)†
5600 ± 973 7357 ± 1242§
6210 ± 521 6390 ± 522
0.88 ± 0.05 0.85 ± 0.05§
0.9 ± 0.06 0.85 ± 0.05§
10.4 ± 4.9 19.9 ± 10.2§
4.5 ± 0.7 4.8 ± 0.7
5112 ± 632 6736 ± 1211§
5746 ± 521 5882 ± 522
0.89 ± 0.05 0.84 ± 0.06§
0.9 ± 0.05 0.85 ± 0.07§
10.9 ± 5.0 17.0 ± 9.1§
4.5 ± 0.7 4.8 ± 0.7
Data are shown as mean ± SD. Group means comparisons were by two-way analysis of variance. *Adjusted by analysis of covariance with lean body mass and body fat mass as covariates. †n = 39 in control and 55 in obese males, n = 19 in control and 39 in obese female subjects. ‡7.9 × µU/ml = pmol/L. §p < 0.001 control versus obese group.
Table III. Postabsorptive fat, carbohydrate, and protein oxidation rates in obese and nonobese adolescents
Male subjects Control (n = 116) Obese (n = 93) Fox (gm/24 hr) Fox adjusted for LBM (gm/24 hr) Fox adjusted for BFM (gm/24 hr) Fox (%) Fox adjusted for LBM (%) Fox adjusted for BFM (%) Cox (gm/24 hr) Cox (%) Pox (gm/24 hr) Pox adjusted for LBM (gm/24 hr) Pox (%) Pox adjusted for LBM (%)
42.1 ± 28.9 49.9 ± 29.0 62.1 ± 29.1 28.9 ± 17.3 31.1 ± 17.2 35.0 ± 17.2 178.0 ± 54.3 56.7 ± 17.8 47.3 ± 12.1 52.7 ± 6.1 14.4 ± 1.9 14.9 ± 1.7
77.0 ± 38.8‡ 61.0 ± 28.9† 57.3 ± 29.0 41.4 ± 16.4‡ 37.1 ± 17.4* 35.4 ± 17.4 184.0 ± 78 44.3 ± 17.7‡ 61.6 ± 14.8‡ 51.2 ± 6.1 14.4 ± 1.9 13.4 ± 1.7
Female subjects Control (n = 119) Obese (n = 66) 38.1 ± 23.4 48.0 ± 29.0 51.6 ± 28.4*,§ 28.6 ± 16.4 31.3 ± 17.4 35.4 ± 17.5 181.7 ± 44.6 61.8 ± 17.8 28.6 ± 5.6 35.4 ± 6.1‡ 9.6 ± 1.9 10.2 ± 1.6‡
76.0 ± 41.0‡ 66.1 ± 28.9† 45.0 ± 28.4*,¶ 43.6 ± 19.9‡ 41.0 ± 17.1† 34.1 ± 17.1 168.0 ± 69.0 44.7 ± 17.8‡ 46.0 ± 11.0‡ 39.1 ± 4.9‡ 11.7 ± 1.9‡ 11.1 ± 1.7‡
Data are expressed as mean ± SD; group means comparisons were by two-way analysis of variance. Analysis of covariance was used to adjust Fox for LBM or BFM as covariates. Cox, Carbohydrate oxidation; Pox, protein oxidation. *p < 0.05. †p < 0.01. ‡p <0.001. §Obese versus control. ¶Female versus male.
(and was the same as that of ethanol = 0.666) with a mean error of 1.1% ± 0.6%. The coefficient of variation on RMR of a 1-day and a 1-week interval was less than 3%. Nonprotein RQ values higher than 1 were observed in 12 (5%) of 235 control children and in 4 (2.5%) of 159 obese children. In all these cases the results were confirmed by a second independent measurement. Furthermore the RQ values of 20 young adults (17- to 19-year-olds), measured by the same calorimeter were 100
within the normal range for adults (mean ± SD = 0.83 ± 0.03, range: 0.76 to 0.88). The macronutrient oxidation rate was calculated from oxygen consumption, carbon dioxide production, and urinary nitrogen excretion by means of equations published by Ferrannini.18 Because patients had fasted overnight, it was assumed that most of their plasma glucose turnover was derived from liver glycogenolysis and consequently the equation appropriate for the calculation of carbohydrate oxidation de-
rived from glycogen (equation 6) was used, except in children with net lipid synthesis, where equation 18 was applied.18 Urine was collected for 12 hours (8 PM to 8 AM) in a randomly selected sample of 160 patients (8 obese and 8 normal-weight male and female patients in each of the 5 pubertal stages). Urine was collected in a plastic container and stored at 4° C. The patients brought the container to the laboratory where the last urine sample was voided at 8 AM. The mean urinary nitro-
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
MOLNÁR AND SCHUTZ
gen values (in grams per kilogram per hour) were calculated for each of the subgroups and the average values were used for patients in whom the urine collection was not performed.
Table IV. Correlation coefficients between anthropometric variables, age, and postabsorptive fat oxidation
Laboratory Methods
Body weight LBM BFM WHR CPR Age
Plasma immunoreactive insulin was measured by a commercially available radioimmunoassay kit (Institute of Isotopes of the Hungarian Academy of Sciences). Plasma glucose concentration was determined by the glucose oxidase method. Urinary nitrogen was determined according to the method of Kjeldahl.
Statistical Analysis All results given are expressed as mean ± SD, unless stated otherwise. Statistical differences were assessed by using the unpaired Student t test in which overweight children were compared with control children. Analysis of variance followed by the Fisher least significant difference comparison test were used to compare the mean values of subgroups. Analysis of covariance was used to standardize RMR, Fox for LBM, BFM, and so on. The rationale for the adjustment procedure was described by Ravussin and Bogardus.19 Simple and stepwise regression analysis were used for the evaluation of the relationship between Fox and other parameters.
RESULTS Age and height of both genders were comparable between the obese and normal-weight children. Weight, relative body weight, BFM, body fat percentage, and LBM were significantly higher in the obese groups (Table I). Body fat was more centrally distributed in the obese children, as shown by the higher WHR and central/peripheral skin-fold ratio. Postabsorptive RMR was significantly higher in obese than in nonobese children. However, RMR adjusted for LBM and BFM by means of analysis of covariance (LBM and BFM as covariates) was comparable in obese and nonobese children of both genders (Table II). The postabsorptive RQ and nonprotein RQ were significantly lower in obese than in nonobese children. In
Male
Female
Control
Obese
Control
Obese
0.44‡ 0.48‡ 0.47‡ –0.12 0.23* 0.31‡
0.45‡ 0.45‡ 0.52‡ 0.1 –0.02 0.24*
0.27† 0.29† 0.25‡ –0.09 0.13 0.16
0.54‡ 0.54‡ 0.56‡ 0.14 0.08 0.15
CPR, Central/peripheral skin-fold ratio. Correlation coefficients obtained by simple linear regression analysis. *p < 0.05. †p < 0.01. ‡p < 0.001.
obese children plasma immunoreactive insulin level was elevated, whereas blood glucose concentration was normal compared with control subjects (Table II). Fat oxidation (in grams per 24-hour period) was considerably higher in the obese subjects of both genders than in the normal-weight groups (Table III). The difference could be explained by the higher energy expenditure of obese children (Table II); because of their increased metabolically active body mass (LBM) they combusted larger amounts of macronutrients. This was indicated by the significant positive correlation between Fox and RMR (kilojoules per 24 hours) (r = 0.69, p < 0.001). To account for this relationship, Fox was also expressed as a percentage of RMR, oxidative fat energy percent, or it was adjusted for LBM (LBM as covariate); both oxidative fat energy percent and Fox adjusted for LBM remained significantly higher in the obese groups. Absolute and relative values of Fox adjusted for BFM were not different between obese and normal-weight children (Table III). The absolute values of carbohydrate oxidation were similar in the obese and nonobese children; however, its relative values were significantly lower in the obese groups. Compared with control subjects, the absolute value of protein oxidation was significantly higher in obese girls and boys, whereas the relative contribution of protein oxidation to the energy expenditure was slightly higher only in the obese female subjects (Table III).
There was a significant sex difference in fat and protein oxidation. The absolute value of Fox was significantly lower in girls compared with boys, when adjusted for BFM (Table III). This was similar when Fox was expressed as a function of BFM (control male: 4.5 ± 0.2, control female: 3.3 ± 0.2 gm/24 hours per kilogram; obese male: 3.0 ± 0.2, obese female: 2.4 ± 0.3 gm/24 hours per kilogram; p < 0.05), suggesting that female subjects oxidize less fat per unit of body fat than boys. Protein oxidation was also significantly lower in girls in relative and absolute terms and remained so even when adjusted for the difference in LBM (Table III). The correlation coefficients between body weight, LBM, BFM, indexes of body fat distribution, age, and Fox are shown in Table IV. There were statistically significant positive associations between body weight, BFM, LBM, and Fox (in grams per 24 hours). The correlations between indexes of fat distribution (WHR, central/peripheral skin-fold ratio) and Fox were not significant in most of the analyzed subgroups. The association between Fox (in grams per 24 hours) and age was statistically significant only in male subjects (Table IV). Carbohydrate oxidation did not correlate with any of the anthropometric variables. The most important factor influencing the variance in protein oxidation was LBM (in the whole cohort R2=0.7). To assess whether the correlation between Fox and BFM was related to the 101
MOLNÁR AND SCHUTZ
close covariation between fat mass, LBM, and body weight, multiple stepwise linear regression analysis was done (body weight, LBM, BFM, age WHR, and central/peripheral skin-fold ratio as independent variables). In obese male and female subjects BFM explained 27% (y = 2.6 BFM + 8.2) and 31% (y = 2.2 BFM + 5.1) of the variance of Fox (in grams per 24 hours), respectively; the other independent variables had no significant effect. In male control subjects LBM explained 23%, LBM + BFM 26.3% (y = 1.8 BFM + 1.0 LBM – 11); in female control subjects LBM explained 8.5% (y = 1.3 LBM – 4.6) of the variance in Fox. In the whole cohort BFM explained 39% and BFM + LBM 41% (y = 1.3 BFM + 0.99 LBM – 7.4) of the variance in Fox. In the whole cohort the sum of five skin-fold thicknesses showed slightly lower correlation with Fox (r = 0.56, p < 0.001) than BFM (r = 0.62, p < 0.001). Fasting serum insulin also showed a significant positive correlation with Fox in both obese (y = 1.2x + 58.8; r = 0.3; p < 0.01) and nonobese children (y = 4.8x – 9.0; r = 0.84; p < 0.001), and remained so when Fox was adjusted for LBM (control: r = 0.8; obese: r = 0.2), or for BFM (control: r = 0.81; obese: r = 0.21). However, the correlation coefficient of fasting insulin and Fox was significantly higher and the slope of the regression line significantly steeper (p < 0.001) in the control group than in the obese subjects. There was no sex difference in the association between insulin and Fox. The correlation between fasting blood glucose levels and Fox was not significant. Fox increased significantly with onset of puberty in all subgroups of children (Figure). This increase leveled off during the advance of puberty except in male control subjects. The changes were similar, except in obese male subjects, when Fox was expressed as oxidative fat energy percent (Figure). Fox adjusted for BFM and LBM did not change significantly either in obese or in control children during puberty and the significant difference between obese and control groups disappeared completely (Figure). A similar pattern of changes was seen when Fox was delineated in the function of age. There was a progressive increase in Fox from 40.8 gm/24 hours (28.6%) in 102
THE JOURNAL
OF PEDIATRICS JANUARY 1998
Figure. Postabsorptive fat oxidation during pubertal development. Upper panel: Fat oxidation in absolute value; middle panel: oxidative fat energy percent; lower panel: fat oxidation adjusted for LBM and BFM. Number of observations in nonobese male subjects (pubertal stage I, II-III, IV-V): 39, 45, 32; in obese male subjects: 33, 39, 21; in nonobese female subjects: 15, 49, 55; in obese female subjects: 10, 10, 46. Vertical bars represent ±SE. *p < 0.01, pubertal stage II-III and IV-V versus pubertal stage I; #p < 0.01 obese versus nonobese. Groups means were compared by threeway analysis of variance.
THE JOURNAL OF PEDIATRICS Volume 132, Number 1 the 9.5- to 11-year-old children to 59.3 gm/24 hours (36%) in the 15- to 16.5year-old group. The increase in Fox with age also vanished when adjusted for BFM and LBM.
DISCUSSION This study constitutes a large number of observations on postabsorptive Fox in carefully characterized prepubertal and pubertal white children as measured by the same investigator, using the same method under the same environmental conditions. Moreover, the study cohort had a wide and similar range of both LBM and BFM. Although the estimation of body composition on the basis of skin-fold measurements is not the most accurate method, it is still widely used in clinical pediatric practice. The measurement errors were kept as low as possible by using the mean of triplicate measurements carried out by a single investigator. In spite of this, the errors of skinfold measurements were probably greater in obese patients than in control subjects. The use of regression equations for the calculation of body composition may constitute a source of further errors. From this respect it is quite reassuring, however, that correlations of body fat and the sum of skin folds with Fox were comparable. The expected error of assessment of percentage body fat from skin-fold thicknesses ranges from 3% to 5% compared with body fat measured by densitometry.17,20 The more accurate methods such as underwater weighing, dual x-ray absorptiometry, deuterium dilution, and computed tomography are either slightly invasive, impracticable in pediatric practice, or too expensive. The bioelectric impedance method is simple and relatively cheap, but it may not have advantages over anthropometric methods.21 Nevertheless, the limitations of the anthropometric method used in the present study should be taken into consideration when the results are evaluated. The major finding of the study was that Fox of obese adolescents exceeded that of normal-weight control subjects and remained so after adjustment for LBM. However, when Fox was adjusted for BFM the difference vanished between the obese and normal-weight groups. The Fox
MOLNÁR AND SCHUTZ
rate was significantly lower in girls than in boys when adjusted for BFM. These results confirm the observations in adults.10 This effect cannot be explained by the lower RMR of girls compared with boys after correction for LBM (Table II), since the absolute value of Fox was comparable for the two genders. The influence of the different hormonal status on Fox is also unlikely, because the sexual dimorphism in Fox was already present in prepubertal children. The higher amount of gluteofemoral adipose tissue with low lipolytic activity in girls than in boys, as indicated by the lower WHR of girls, or the difference in the fat-free mass composition (muscle and non-muscle) may be an explanation for this gender difference. The relationship between Fox and BFM is unclear; the previously published results are contradictory. According to our results both fat-free mass and fat mass were positively correlated with Fox in normal-weight and overweight boys and girls. The main determinants of Fox were BFM in obese children, LBM in control subjects, and BFM + LBM in the whole cohort. This observation supports the physiologic point of view that both the fat mass, as the producer of free fatty acids, and the LBM as the user of free fatty acids, are important determinants of Fox. Because less than half of the lipid oxidation can be explained by differences in body composition, other factors not evaluated in the present study, such as 24-hour energy balance, macronutrient composition of the diet, and insulin resistance must be considered to explain the residual variance in postabsorptive Fox. In the present study fasting plasma insulin concentrations, which to a certain extent reflect insulin resistance, showed a positive association with postabsorptive Fox, supporting earlier observations in adults that insulin insensitivity leads to decreased carbohydrate, increased Fox, and to lower rates of weight gain in Pima Indians.4 Nagy et al.10 reported significant inverse correlation between insulin and Fox; however, the body composition of their cohort was totally different from that of the population of our study. Fat mass was inversely related to LBM in their population, whereas they had a significant positive correlation in our cohort (r = 0.8).
Positive fat balance enhances Fox through a slow, progressive increase in fat stores, which in turn raises Fox rate by causing insulin resistance.3,4 The high flux of free fatty acids to the liver and muscle reduces carbohydrate oxidation and stimulates Fox by further aggravating insulin insensitivity22 and by increasing carnitine accumulation in the liver and muscle.23 The growth in LBM tends to parallel increases in fat mass during the development of obesity, and may provide the increasing capacity for Fox to meet the needs. The observed increase in Fox by the onset of puberty leveled off in female control and in male and female obese subjects, whereas it continued to increase in a linear fashion in the male control subjects. When Fox was adjusted for LBM and BFM, no change was observed during pubertal development, indicating that the effect of puberty on Fox was related to changes in body composition.
REFERENCES 1. Kennedy GC. The role of depot fat in the hypothalamic control of food intake in the rat. Proc R Soc Lond 1953;140B:578-92. 2. Björntorp P, Carlagen G, Isaksson B, Krotkiewski M, Larsson B, Sjöström L. Effect of energy-reduced dietary regimen in relation to adipose tissue cellularity in obese women. Am J Clin Nutr 1975; 28:445-52. 3. Flatt JP, Ravussin E, Acheson KJ, Jequier E. Effects of dietary fat on postprandial substrate oxidation and on carbohydrate and fat balances. J Clin Invest 1985;76:1019-24. 4. Swinburn BA, Nyomba BL, Saad MF, et al. Insulin resistance associated with lower rates of weight gain in Pima indians. J Clin Invest 1991;88:168-73. 5. Schutz Y, Tremblay A, Weinsier RL, Nelson KM. Role of fat oxidation in the long-term stabilization of body weight in obese women. Am J Clin Nutr 1992;55:670-4. 6. Astrup A, Buermann B, Western P, Toubro S, Raben A, Christensen NJ. Obesity as an adaptation to a high-fat diet: evidence from a cross-sectional study. Am J Clin Nutr 1994;59:350-5. 7. Maffeis C, Pinelli L, Schutz Y. Increased fat oxidation in prepubertal obese children: a metabolic defense against further weight gain? J Pediatr 1995;126:15-20. 8. Le Stunff C, Bougneres P-F. Time course of increased lipid and decreased glucose oxi-
103
MOLNÁR AND SCHUTZ
9.
10.
11.
12.
13.
104
dation during early phase of childhood obesity. Diabetes 1993;42:1010-6. Calles-Escandón J, Arciero PJ, Gardner AW, Bauman C, Poehlman ET. Basal fat oxidation decreases with aging in women. J Appl Physiol 1995;78:266-71. Nagy TR, Goran MI, Weinsier RL, Toth MJ, Schutz Y, Poehlman ET. Determinants of basal fat oxidation in healthy caucasians. J Appl Physiol 1996;80:1743-8. Parizková J. The development of subcutaneous fat in adolescents and the effect of physical training and sport. Physiologia Bohemoslovenica 1959;8:112-7. Parizková J. Age trends in fat in normal and obese children. J Appl Physiol 1961;16:173-4. Novak LP. Changes of total body water during adolescent growth. Anthrop Közl 1986;30:181-6.
THE JOURNAL
14. Molnár D, Jeges S, Erhardt É, Schutz Y. Measured and predicted resting metabolic rate in obese and non-obese adolescents. J Pediatr 1995;127:571-7. 15. Eiben OG, Barabás A, Panto E. The Hungarian National Growth Study I. Reference data on the biological developmental status and physical fitness of 3-18 year-old Hungarian youth in the 1980s. Humanbiol Budapest 1991;21:16-118. 16. Tanner JM. Growth at adolescence. 2nd ed. Oxford: Blackwell; 1962. 17. Parizkova J, Roth Z. Assessment of depot fat in children from skinfold thickness measurements by Holtain caliper. Hum Biol 1972;44:613-6. 18. Ferrannini E. The theoretical bases of indirect calorimetry: a review. Metabolism 1988;37:287-301. 19. Ravussin E, Bogardus C. Relationship of genetics, age, and physical fitness to daily
20.
21.
22.
23.
OF PEDIATRICS JANUARY 1998
energy expenditure and fuel utilization. Am J Clin Nutr 1989;49:968-75. Deurenberg P, Pieters JL, Hautvast JGA. The assessment of the body fat percentage by skinfold thickness measurements in childhood and young adolescence. Br J Nutr 1990; 63:293-303. Deurenberg P, Smit HE, Kusters CSL. Is the bioelectrical impedance method suitable for epidemiological field studies? Eur J Clin Nutr 1989;43:647-54. Randle PJ, Garland PB, Hales CN, Newsholme EA. The glucose fatty-acid cycle: its role in insulin sensitivity and metabolic disturbances of diabetes mellitus. Lancet 1963;1:785-9. Harper P, Wadström C, Backman L, Cederbald G. Increased liver carnitine content in obese women. Am J Clin Nutr 1995;61:18-25.
Fat oxidation in nonobese and obese adolescents: Effect of body composition and pubertal development
Dénes Molnár, MD, PhD, and Yves Schutz, PhD Objectives: To measure postabsorptive fat oxidation (Fox) and to assess its association with body composition (lean body mass [LBM] and body fat mass [BFM]) and pubertal development.
Design: We studied 235 control (male/female ratio = 116/119; age [mean ± SD]: 13.1 ± 1.7 years; weight: 45.3 ± 10.5 kg; LBM: 34.3 ± 7.1 kg; BFM: 11.0 ± 4.5 kg) and 159 obese (male/female ratio = 93/66; age: 12.9 ± 2.1 years; weight: 76.2 ± 19.1 kg; LBM: 47.4 ± 10.9 kg; BFM: 28.8 ± 9.2 kg) adolescents. Postabsorptive Fox was calculated from oxygen consumption, carbon dioxide production, and urinary nitrogen as measured by indirect calorimetry and Kjeldahl’s method, respectively. Body composition was determined by anthropometry. Results: Postabsorptive Fox (absolute value and percentage of resting metabolic rate) was significantly (p < 0.001) higher in the obese adolescents (76.7 ± 26.3 gm/24 hours, 42.3% ± 18.7%) than in the control subjects (40.0 ± 26.3 gm/24 hours, 28.7% ± 17.0%), even if adjusted for LBM. Fox corrected for BFM was similar in control and in obese children, but was significantly lower in girls compared with boys (control male subjects: 62.1 ± 29.1 gm/24 hours, control female subjects: 51.6 ± 28.4 gm/24 hours, obese male subjects: 57.3 ± 29 gm/24 hour, obese female subjects: 45.0 ± 28.4 gm/24 hours). BFM and LBM showed a significant positive correlation with Fox. By stepwise regression analysis the most important determinant of Fox was BFM in obese and LBM in control children. There was a significant rise in Fox during puberty; however, it was mainly explained by changes in body composition. Conclusions: Obese adolescents have higher Fox rates than their normal-weight counterparts. Both LBM and fat mass are important determinants of Fox. (J Pediatr 1998;132:98-104)
It has been proposed that adipose tissue plays a significant role in the regulation of energy intake and balance.1 This concept was supported by the observation that in
obese women subjected to energy restriction, resistance to further weight loss developed when fat-cell size reached that of lean subjects.2 A positive fat balance does not en-
From the Department of Pediatrics, University Medical School of Pécs, Pécs, Hungary, and the Institute of Physiology, University of Lausanne, Lausanne, Switzerland. Supported by a research grant from Nestle Nutrition and Hungarian National Research grant (OKTA 5267 and 16065). Portions of the data have been reported in abstract form (Int J Obes 1995;19(suppl 12):42 and Pediatr Res 1995;38:445). Submitted for publication Feb. 13, 1996; accepted April 29, 1997. Reprint requests: Dénes Molnár, MD, PhD, Department of Pediatrics, University Medical School of Pécs, József A. u. 7., H-7623 Pécs, Hungary. Copyright © 1998 by Mosby, Inc. 0022-3476/98/$5.00 + 0 9/21/83009
98
hance fat oxidation acutely,3 but rather through a slower increase in fat stores, which in turn raises the Fox rate by causing insulin resistance.4 According to this theory the accumulation of body fat may be regarded as a compensatory mechanism by which the body increases Fox and reaches new energy and fat balance. Studies in support of this hypothesis in adults5,6 and in prepubertal children7,8 demonstrated higher lipid oxidation in obese than in lean subjects, and positive correlation was found between body fat mass and fasting Fox, thus supporting the theory that enlargement of body fat store is accompanied by an increased Fox. More recent studies in aging women9 and in healthy white subjects,10 however, failed to find a relationship between Fox and fat mass. Moreover, Nagy et al.10 suggested that the positive correlation between Fox and fat mass demonstrated earlier was related to the much greater range of fat mass compared with that of fat-free mass of the cohorts investigated in these studies.5,6 BFM Fox LBM RMR RQ WHR
Body fat mass Fat oxidation Lean body mass Resting metabolic rate Respiratory quotient Waist/hip ratio
The aims of the present study were to (1) investigate substrate utilization in obese and nonobese prepubertal and adolescent children, (2) analyze the effect of body composition and body fat distribution on Fox, (3) determine how puberty, when dramatic changes in lean body mass and BFM occur,11-13 affect Fox. We have studied a carefully characterized large cohort of obese and normal-weight, prepubertal and adolescent children of both genders with a wide range of both BFM and LBM.
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
MOLNÁR AND SCHUTZ
Table I. Physical characteristics of subjects
Group Control (n = 235) Male (n = 116) Female (n = 119) Obese (n = 159) Male (n = 93) Female (n = 66)
Age (yr)
Weight (kg)
Height (cm)
Relative weight (%)
BFM (kg)
BF (%)
LBM (kg)
WHR
CPR
13.1 ± 1.7 44.5 ± 11.6 13.1 ± 1.7 46.0 ± 9.3
157.6 ± 13.2 157.3 ± 8.9
94.0 ± 10.4 95.9 ±11.2
9.6 ± 3.9 12.4 ± 4.6
21.2 ± 4.6 26.2 ± 5.5
34.9 ± 8.6 33.6 ± 5.4
0.82 ± 0.05 0.74 ± 0.05
0.62 ± 0.2 0.64 ± 0.2
12.8 ± 1.8 75.4 ± 18.8* 13.2 ± 1.9 77.4 ± 19.5*
160.0 ± 11.9 153.4 ± 23.4* 26.7 ± 7.8* 158.7 ± 10.2 155.0 ± 22.8* 31.7 ± 10.3*
35.1 ± 3.0* 48.7 ± 11.5* 0.89 ± 0.05* 40.3 ± 4.4* 45.6 ± 9.8* 0.81 ± 0.06*
0.99 ± 0.2* 0.92 ± 0.2*
CPR, Central/peripheral skinfold ratio. *p < 0.001 obese versus control group (Student’s test).
METHODS Subjects The study included the data of 235 nonobese (116 male) and 159 obese (93 male) Hungarian white children whose ages ranged from 9.5 to 16.5 years. Most of the subjects participated in a previous study exploring the validity of equations currently used for the prediction of resting metabolic rate.14 Anthropometric measurements and the determination of RMR and fuel oxidation were carried out in all subjects. Blood samples for the analysis of plasma insulin and glucose concentrations were obtained only from a subset of patients (58 control and 94 obese subjects) who volunteered to have venipuncture. The age, weight, height, and body composition of the subset were similar to that of the whole cohort. The cohort of normal-weight children (body weight less than 120% of the expected weight for height15; triceps and subscapular skin-fold thickness <90th percentile of the reference values for age and sex15) was recruited from primary and high schools. Obese children (body weight exceeding predicted body weight for height by 20% or more; triceps and subscapular skinfold thickness >90th percentile of the reference values for age and sex15) were recruited from patients referred to the obesity clinic of the Department of Pediatrics, University Medical School of Pécs. Pubertal stage was assessed according to Tanner.16 Subjects with diabetes mellitus or any other metabolic or endocrine disease were excluded. Physical examination and routine laboratory tests documented the
absence of any health problems. The subjects included in the study did not take medicine, consume alcohol, or smoke cigarettes regularly. These data were obtained from a questionnaire administered by a trained assistant. Descriptive data are listed in Table I. The study was approved by the ethical committee of the University of Pécs, and the investigations were carried out in accordance with the Declaration of Helsinki II.
Body Composition Anthropometric measurements were carried out by the same investigator and included weight, height, waist and hip circumference, and skin-fold thickness. The child’s weight was obtained in light clothing (shorts or dress and tee-shirt, without shoes) to the nearest 0.1 kg on a standard beam scale. Height was measured to the nearest 0.1 cm by a Holtain stadiometer without shoes. Body mass index was calculated by dividing weight (in kilograms) by height (in square meters). Skin-fold thicknesses (triceps, biceps, suprailiac, subscapular, and calf skin folds) were measured in triplicate, to the nearest millimeter, on the left side of the body by using Holtain caliper. The formula derived by Parizkova and Roth17 was used to calculate the percentage of body fat from the sum of five skin folds. LBM was calculated by subtracting fat mass from body weight. Body fat distribution was assessed by the waist/hip ratio and by the ratio of central skin folds (subscapular, suprailiac) to extremity skin folds (biceps, triceps and calf). WHR was measured as the minimal waist circumference to the
circumference of the maximal gluteal protuberance.
Experimental Design Before the gas exchange measurements children received an unrestricted diet. The day before the test they did not perform any intense physical exercise. The patients arrived at the laboratory approximately at 7:30 AM after an overnight fast (from 8 PM the day before). The lifestyle questionnaire was completed and a blood sample was obtained from the antecubital vein for the determination of plasma insulin and glucose. Subjects rested on a comfortable bed, under the canopy in a temperaturecontrolled room (25° to 27° C) for at least 30 minutes before the commencement of the gas exchange measurements to accommodate to the environment. During the investigation the children were watching nonviolent cartoons to avoid spontaneous movements and to reduce unnecessary nervousness that might have increased energy expenditure.
Energy Expenditure Measurements and Calculation of Fuel Oxidation Rate Resting metabolic rate was measured for 45 minutes with the help of a Deltatrac metabolic cart (Datex Instrumentarium Corp., Helsinki, Finland), using the ventilated hood technique as described earlier.14 Before each test the calorimeter was calibrated with a reference gas mixture (95% oxygen, 5% carbon dioxide). Ethanol burning tests were performed to assess the precision of the indirect calorimeter. The mean ± SD of measured respiratory quotient was 0.668 ± 0.008 99
MOLNÁR AND SCHUTZ
THE JOURNAL
OF PEDIATRICS JANUARY 1998
Table II. Resting metabolic rate, respiratory quotient, and serum insulin and glucose levels in obese and control adolescents
Group Male subjects Control (n = 116) Obese (n = 93) Female subjects Control (n = 119) Obese (n = 66)
Resting metabolic rate (kJ/24 hr)
Resting metabolic rate* (kJ/24 hr)
Respiratory quotient
Nonprotein respiratory quotient
Immunoreactive insulin† (µU/ml)‡
Glucose (mmol/L)†
5600 ± 973 7357 ± 1242§
6210 ± 521 6390 ± 522
0.88 ± 0.05 0.85 ± 0.05§
0.9 ± 0.06 0.85 ± 0.05§
10.4 ± 4.9 19.9 ± 10.2§
4.5 ± 0.7 4.8 ± 0.7
5112 ± 632 6736 ± 1211§
5746 ± 521 5882 ± 522
0.89 ± 0.05 0.84 ± 0.06§
0.9 ± 0.05 0.85 ± 0.07§
10.9 ± 5.0 17.0 ± 9.1§
4.5 ± 0.7 4.8 ± 0.7
Data are shown as mean ± SD. Group means comparisons were by two-way analysis of variance. *Adjusted by analysis of covariance with lean body mass and body fat mass as covariates. †n = 39 in control and 55 in obese males, n = 19 in control and 39 in obese female subjects. ‡7.9 × µU/ml = pmol/L. §p < 0.001 control versus obese group.
Table III. Postabsorptive fat, carbohydrate, and protein oxidation rates in obese and nonobese adolescents
Male subjects Control (n = 116) Obese (n = 93) Fox (gm/24 hr) Fox adjusted for LBM (gm/24 hr) Fox adjusted for BFM (gm/24 hr) Fox (%) Fox adjusted for LBM (%) Fox adjusted for BFM (%) Cox (gm/24 hr) Cox (%) Pox (gm/24 hr) Pox adjusted for LBM (gm/24 hr) Pox (%) Pox adjusted for LBM (%)
42.1 ± 28.9 49.9 ± 29.0 62.1 ± 29.1 28.9 ± 17.3 31.1 ± 17.2 35.0 ± 17.2 178.0 ± 54.3 56.7 ± 17.8 47.3 ± 12.1 52.7 ± 6.1 14.4 ± 1.9 14.9 ± 1.7
77.0 ± 38.8‡ 61.0 ± 28.9† 57.3 ± 29.0 41.4 ± 16.4‡ 37.1 ± 17.4* 35.4 ± 17.4 184.0 ± 78 44.3 ± 17.7‡ 61.6 ± 14.8‡ 51.2 ± 6.1 14.4 ± 1.9 13.4 ± 1.7
Female subjects Control (n = 119) Obese (n = 66) 38.1 ± 23.4 48.0 ± 29.0 51.6 ± 28.4*,§ 28.6 ± 16.4 31.3 ± 17.4 35.4 ± 17.5 181.7 ± 44.6 61.8 ± 17.8 28.6 ± 5.6 35.4 ± 6.1‡ 9.6 ± 1.9 10.2 ± 1.6‡
76.0 ± 41.0‡ 66.1 ± 28.9† 45.0 ± 28.4*,¶ 43.6 ± 19.9‡ 41.0 ± 17.1† 34.1 ± 17.1 168.0 ± 69.0 44.7 ± 17.8‡ 46.0 ± 11.0‡ 39.1 ± 4.9‡ 11.7 ± 1.9‡ 11.1 ± 1.7‡
Data are expressed as mean ± SD; group means comparisons were by two-way analysis of variance. Analysis of covariance was used to adjust Fox for LBM or BFM as covariates. Cox, Carbohydrate oxidation; Pox, protein oxidation. *p < 0.05. †p < 0.01. ‡p <0.001. §Obese versus control. ¶Female versus male.
(and was the same as that of ethanol = 0.666) with a mean error of 1.1% ± 0.6%. The coefficient of variation on RMR of a 1-day and a 1-week interval was less than 3%. Nonprotein RQ values higher than 1 were observed in 12 (5%) of 235 control children and in 4 (2.5%) of 159 obese children. In all these cases the results were confirmed by a second independent measurement. Furthermore the RQ values of 20 young adults (17- to 19-year-olds), measured by the same calorimeter were 100
within the normal range for adults (mean ± SD = 0.83 ± 0.03, range: 0.76 to 0.88). The macronutrient oxidation rate was calculated from oxygen consumption, carbon dioxide production, and urinary nitrogen excretion by means of equations published by Ferrannini.18 Because patients had fasted overnight, it was assumed that most of their plasma glucose turnover was derived from liver glycogenolysis and consequently the equation appropriate for the calculation of carbohydrate oxidation de-
rived from glycogen (equation 6) was used, except in children with net lipid synthesis, where equation 18 was applied.18 Urine was collected for 12 hours (8 PM to 8 AM) in a randomly selected sample of 160 patients (8 obese and 8 normal-weight male and female patients in each of the 5 pubertal stages). Urine was collected in a plastic container and stored at 4° C. The patients brought the container to the laboratory where the last urine sample was voided at 8 AM. The mean urinary nitro-
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
MOLNÁR AND SCHUTZ
gen values (in grams per kilogram per hour) were calculated for each of the subgroups and the average values were used for patients in whom the urine collection was not performed.
Table IV. Correlation coefficients between anthropometric variables, age, and postabsorptive fat oxidation
Laboratory Methods
Body weight LBM BFM WHR CPR Age
Plasma immunoreactive insulin was measured by a commercially available radioimmunoassay kit (Institute of Isotopes of the Hungarian Academy of Sciences). Plasma glucose concentration was determined by the glucose oxidase method. Urinary nitrogen was determined according to the method of Kjeldahl.
Statistical Analysis All results given are expressed as mean ± SD, unless stated otherwise. Statistical differences were assessed by using the unpaired Student t test in which overweight children were compared with control children. Analysis of variance followed by the Fisher least significant difference comparison test were used to compare the mean values of subgroups. Analysis of covariance was used to standardize RMR, Fox for LBM, BFM, and so on. The rationale for the adjustment procedure was described by Ravussin and Bogardus.19 Simple and stepwise regression analysis were used for the evaluation of the relationship between Fox and other parameters.
RESULTS Age and height of both genders were comparable between the obese and normal-weight children. Weight, relative body weight, BFM, body fat percentage, and LBM were significantly higher in the obese groups (Table I). Body fat was more centrally distributed in the obese children, as shown by the higher WHR and central/peripheral skin-fold ratio. Postabsorptive RMR was significantly higher in obese than in nonobese children. However, RMR adjusted for LBM and BFM by means of analysis of covariance (LBM and BFM as covariates) was comparable in obese and nonobese children of both genders (Table II). The postabsorptive RQ and nonprotein RQ were significantly lower in obese than in nonobese children. In
Male
Female
Control
Obese
Control
Obese
0.44‡ 0.48‡ 0.47‡ –0.12 0.23* 0.31‡
0.45‡ 0.45‡ 0.52‡ 0.1 –0.02 0.24*
0.27† 0.29† 0.25‡ –0.09 0.13 0.16
0.54‡ 0.54‡ 0.56‡ 0.14 0.08 0.15
CPR, Central/peripheral skin-fold ratio. Correlation coefficients obtained by simple linear regression analysis. *p < 0.05. †p < 0.01. ‡p < 0.001.
obese children plasma immunoreactive insulin level was elevated, whereas blood glucose concentration was normal compared with control subjects (Table II). Fat oxidation (in grams per 24-hour period) was considerably higher in the obese subjects of both genders than in the normal-weight groups (Table III). The difference could be explained by the higher energy expenditure of obese children (Table II); because of their increased metabolically active body mass (LBM) they combusted larger amounts of macronutrients. This was indicated by the significant positive correlation between Fox and RMR (kilojoules per 24 hours) (r = 0.69, p < 0.001). To account for this relationship, Fox was also expressed as a percentage of RMR, oxidative fat energy percent, or it was adjusted for LBM (LBM as covariate); both oxidative fat energy percent and Fox adjusted for LBM remained significantly higher in the obese groups. Absolute and relative values of Fox adjusted for BFM were not different between obese and normal-weight children (Table III). The absolute values of carbohydrate oxidation were similar in the obese and nonobese children; however, its relative values were significantly lower in the obese groups. Compared with control subjects, the absolute value of protein oxidation was significantly higher in obese girls and boys, whereas the relative contribution of protein oxidation to the energy expenditure was slightly higher only in the obese female subjects (Table III).
There was a significant sex difference in fat and protein oxidation. The absolute value of Fox was significantly lower in girls compared with boys, when adjusted for BFM (Table III). This was similar when Fox was expressed as a function of BFM (control male: 4.5 ± 0.2, control female: 3.3 ± 0.2 gm/24 hours per kilogram; obese male: 3.0 ± 0.2, obese female: 2.4 ± 0.3 gm/24 hours per kilogram; p < 0.05), suggesting that female subjects oxidize less fat per unit of body fat than boys. Protein oxidation was also significantly lower in girls in relative and absolute terms and remained so even when adjusted for the difference in LBM (Table III). The correlation coefficients between body weight, LBM, BFM, indexes of body fat distribution, age, and Fox are shown in Table IV. There were statistically significant positive associations between body weight, BFM, LBM, and Fox (in grams per 24 hours). The correlations between indexes of fat distribution (WHR, central/peripheral skin-fold ratio) and Fox were not significant in most of the analyzed subgroups. The association between Fox (in grams per 24 hours) and age was statistically significant only in male subjects (Table IV). Carbohydrate oxidation did not correlate with any of the anthropometric variables. The most important factor influencing the variance in protein oxidation was LBM (in the whole cohort R2=0.7). To assess whether the correlation between Fox and BFM was related to the 101
MOLNÁR AND SCHUTZ
close covariation between fat mass, LBM, and body weight, multiple stepwise linear regression analysis was done (body weight, LBM, BFM, age WHR, and central/peripheral skin-fold ratio as independent variables). In obese male and female subjects BFM explained 27% (y = 2.6 BFM + 8.2) and 31% (y = 2.2 BFM + 5.1) of the variance of Fox (in grams per 24 hours), respectively; the other independent variables had no significant effect. In male control subjects LBM explained 23%, LBM + BFM 26.3% (y = 1.8 BFM + 1.0 LBM – 11); in female control subjects LBM explained 8.5% (y = 1.3 LBM – 4.6) of the variance in Fox. In the whole cohort BFM explained 39% and BFM + LBM 41% (y = 1.3 BFM + 0.99 LBM – 7.4) of the variance in Fox. In the whole cohort the sum of five skin-fold thicknesses showed slightly lower correlation with Fox (r = 0.56, p < 0.001) than BFM (r = 0.62, p < 0.001). Fasting serum insulin also showed a significant positive correlation with Fox in both obese (y = 1.2x + 58.8; r = 0.3; p < 0.01) and nonobese children (y = 4.8x – 9.0; r = 0.84; p < 0.001), and remained so when Fox was adjusted for LBM (control: r = 0.8; obese: r = 0.2), or for BFM (control: r = 0.81; obese: r = 0.21). However, the correlation coefficient of fasting insulin and Fox was significantly higher and the slope of the regression line significantly steeper (p < 0.001) in the control group than in the obese subjects. There was no sex difference in the association between insulin and Fox. The correlation between fasting blood glucose levels and Fox was not significant. Fox increased significantly with onset of puberty in all subgroups of children (Figure). This increase leveled off during the advance of puberty except in male control subjects. The changes were similar, except in obese male subjects, when Fox was expressed as oxidative fat energy percent (Figure). Fox adjusted for BFM and LBM did not change significantly either in obese or in control children during puberty and the significant difference between obese and control groups disappeared completely (Figure). A similar pattern of changes was seen when Fox was delineated in the function of age. There was a progressive increase in Fox from 40.8 gm/24 hours (28.6%) in 102
THE JOURNAL
OF PEDIATRICS JANUARY 1998
Figure. Postabsorptive fat oxidation during pubertal development. Upper panel: Fat oxidation in absolute value; middle panel: oxidative fat energy percent; lower panel: fat oxidation adjusted for LBM and BFM. Number of observations in nonobese male subjects (pubertal stage I, II-III, IV-V): 39, 45, 32; in obese male subjects: 33, 39, 21; in nonobese female subjects: 15, 49, 55; in obese female subjects: 10, 10, 46. Vertical bars represent ±SE. *p < 0.01, pubertal stage II-III and IV-V versus pubertal stage I; #p < 0.01 obese versus nonobese. Groups means were compared by threeway analysis of variance.
THE JOURNAL OF PEDIATRICS Volume 132, Number 1 the 9.5- to 11-year-old children to 59.3 gm/24 hours (36%) in the 15- to 16.5year-old group. The increase in Fox with age also vanished when adjusted for BFM and LBM.
DISCUSSION This study constitutes a large number of observations on postabsorptive Fox in carefully characterized prepubertal and pubertal white children as measured by the same investigator, using the same method under the same environmental conditions. Moreover, the study cohort had a wide and similar range of both LBM and BFM. Although the estimation of body composition on the basis of skin-fold measurements is not the most accurate method, it is still widely used in clinical pediatric practice. The measurement errors were kept as low as possible by using the mean of triplicate measurements carried out by a single investigator. In spite of this, the errors of skinfold measurements were probably greater in obese patients than in control subjects. The use of regression equations for the calculation of body composition may constitute a source of further errors. From this respect it is quite reassuring, however, that correlations of body fat and the sum of skin folds with Fox were comparable. The expected error of assessment of percentage body fat from skin-fold thicknesses ranges from 3% to 5% compared with body fat measured by densitometry.17,20 The more accurate methods such as underwater weighing, dual x-ray absorptiometry, deuterium dilution, and computed tomography are either slightly invasive, impracticable in pediatric practice, or too expensive. The bioelectric impedance method is simple and relatively cheap, but it may not have advantages over anthropometric methods.21 Nevertheless, the limitations of the anthropometric method used in the present study should be taken into consideration when the results are evaluated. The major finding of the study was that Fox of obese adolescents exceeded that of normal-weight control subjects and remained so after adjustment for LBM. However, when Fox was adjusted for BFM the difference vanished between the obese and normal-weight groups. The Fox
MOLNÁR AND SCHUTZ
rate was significantly lower in girls than in boys when adjusted for BFM. These results confirm the observations in adults.10 This effect cannot be explained by the lower RMR of girls compared with boys after correction for LBM (Table II), since the absolute value of Fox was comparable for the two genders. The influence of the different hormonal status on Fox is also unlikely, because the sexual dimorphism in Fox was already present in prepubertal children. The higher amount of gluteofemoral adipose tissue with low lipolytic activity in girls than in boys, as indicated by the lower WHR of girls, or the difference in the fat-free mass composition (muscle and non-muscle) may be an explanation for this gender difference. The relationship between Fox and BFM is unclear; the previously published results are contradictory. According to our results both fat-free mass and fat mass were positively correlated with Fox in normal-weight and overweight boys and girls. The main determinants of Fox were BFM in obese children, LBM in control subjects, and BFM + LBM in the whole cohort. This observation supports the physiologic point of view that both the fat mass, as the producer of free fatty acids, and the LBM as the user of free fatty acids, are important determinants of Fox. Because less than half of the lipid oxidation can be explained by differences in body composition, other factors not evaluated in the present study, such as 24-hour energy balance, macronutrient composition of the diet, and insulin resistance must be considered to explain the residual variance in postabsorptive Fox. In the present study fasting plasma insulin concentrations, which to a certain extent reflect insulin resistance, showed a positive association with postabsorptive Fox, supporting earlier observations in adults that insulin insensitivity leads to decreased carbohydrate, increased Fox, and to lower rates of weight gain in Pima Indians.4 Nagy et al.10 reported significant inverse correlation between insulin and Fox; however, the body composition of their cohort was totally different from that of the population of our study. Fat mass was inversely related to LBM in their population, whereas they had a significant positive correlation in our cohort (r = 0.8).
Positive fat balance enhances Fox through a slow, progressive increase in fat stores, which in turn raises Fox rate by causing insulin resistance.3,4 The high flux of free fatty acids to the liver and muscle reduces carbohydrate oxidation and stimulates Fox by further aggravating insulin insensitivity22 and by increasing carnitine accumulation in the liver and muscle.23 The growth in LBM tends to parallel increases in fat mass during the development of obesity, and may provide the increasing capacity for Fox to meet the needs. The observed increase in Fox by the onset of puberty leveled off in female control and in male and female obese subjects, whereas it continued to increase in a linear fashion in the male control subjects. When Fox was adjusted for LBM and BFM, no change was observed during pubertal development, indicating that the effect of puberty on Fox was related to changes in body composition.
REFERENCES 1. Kennedy GC. The role of depot fat in the hypothalamic control of food intake in the rat. Proc R Soc Lond 1953;140B:578-92. 2. Björntorp P, Carlagen G, Isaksson B, Krotkiewski M, Larsson B, Sjöström L. Effect of energy-reduced dietary regimen in relation to adipose tissue cellularity in obese women. Am J Clin Nutr 1975; 28:445-52. 3. Flatt JP, Ravussin E, Acheson KJ, Jequier E. Effects of dietary fat on postprandial substrate oxidation and on carbohydrate and fat balances. J Clin Invest 1985;76:1019-24. 4. Swinburn BA, Nyomba BL, Saad MF, et al. Insulin resistance associated with lower rates of weight gain in Pima indians. J Clin Invest 1991;88:168-73. 5. Schutz Y, Tremblay A, Weinsier RL, Nelson KM. Role of fat oxidation in the long-term stabilization of body weight in obese women. Am J Clin Nutr 1992;55:670-4. 6. Astrup A, Buermann B, Western P, Toubro S, Raben A, Christensen NJ. Obesity as an adaptation to a high-fat diet: evidence from a cross-sectional study. Am J Clin Nutr 1994;59:350-5. 7. Maffeis C, Pinelli L, Schutz Y. Increased fat oxidation in prepubertal obese children: a metabolic defense against further weight gain? J Pediatr 1995;126:15-20. 8. Le Stunff C, Bougneres P-F. Time course of increased lipid and decreased glucose oxi-
103
MOLNÁR AND SCHUTZ
9.
10.
11.
12.
13.
104
dation during early phase of childhood obesity. Diabetes 1993;42:1010-6. Calles-Escandón J, Arciero PJ, Gardner AW, Bauman C, Poehlman ET. Basal fat oxidation decreases with aging in women. J Appl Physiol 1995;78:266-71. Nagy TR, Goran MI, Weinsier RL, Toth MJ, Schutz Y, Poehlman ET. Determinants of basal fat oxidation in healthy caucasians. J Appl Physiol 1996;80:1743-8. Parizková J. The development of subcutaneous fat in adolescents and the effect of physical training and sport. Physiologia Bohemoslovenica 1959;8:112-7. Parizková J. Age trends in fat in normal and obese children. J Appl Physiol 1961;16:173-4. Novak LP. Changes of total body water during adolescent growth. Anthrop Közl 1986;30:181-6.
THE JOURNAL
14. Molnár D, Jeges S, Erhardt É, Schutz Y. Measured and predicted resting metabolic rate in obese and non-obese adolescents. J Pediatr 1995;127:571-7. 15. Eiben OG, Barabás A, Panto E. The Hungarian National Growth Study I. Reference data on the biological developmental status and physical fitness of 3-18 year-old Hungarian youth in the 1980s. Humanbiol Budapest 1991;21:16-118. 16. Tanner JM. Growth at adolescence. 2nd ed. Oxford: Blackwell; 1962. 17. Parizkova J, Roth Z. Assessment of depot fat in children from skinfold thickness measurements by Holtain caliper. Hum Biol 1972;44:613-6. 18. Ferrannini E. The theoretical bases of indirect calorimetry: a review. Metabolism 1988;37:287-301. 19. Ravussin E, Bogardus C. Relationship of genetics, age, and physical fitness to daily
20.
21.
22.
23.
OF PEDIATRICS JANUARY 1998
energy expenditure and fuel utilization. Am J Clin Nutr 1989;49:968-75. Deurenberg P, Pieters JL, Hautvast JGA. The assessment of the body fat percentage by skinfold thickness measurements in childhood and young adolescence. Br J Nutr 1990; 63:293-303. Deurenberg P, Smit HE, Kusters CSL. Is the bioelectrical impedance method suitable for epidemiological field studies? Eur J Clin Nutr 1989;43:647-54. Randle PJ, Garland PB, Hales CN, Newsholme EA. The glucose fatty-acid cycle: its role in insulin sensitivity and metabolic disturbances of diabetes mellitus. Lancet 1963;1:785-9. Harper P, Wadström C, Backman L, Cederbald G. Increased liver carnitine content in obese women. Am J Clin Nutr 1995;61:18-25.