| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Division of Endocrinology (S.H.T.), Department of Pediatrics, The Children’s Hospital, Division of Biostatistics (B.W.J.), and Department of Medicine (R.H.E.), The Center for Human Nutrition and Division of Endocrinology, Metabolism, and Diabetes, University of Colorado Health Sciences Center, Denver, Colorado 80262
Address all correspondence and requests for reprints to: Dr. Sharon H. Travers, Division of Endocrinology, B-265, The Children’s Hospital, 1056 East 19th Avenue, Denver, Colorado 80218. E-mail: . travers.sharon{at}tchden.org
Abstract
Insulin resistance is a known sequela of obesity; however, the relationship of insulin resistance to future weight gain remains unclear. In several studies, insulin resistance has been associated with weight stabilization. For the most part, this relationship has been found in adults who are overweight. To evaluate the relationship of insulin resistance to future fat accumulation in pubertal children, a 3-yr prospective study was carried out. Insulin sensitivity (Si) was determined by Bergman’s minimal model in 111 healthy children, aged 9.7–14.5 yr. All children were Tanner stage II or III pubertal development at baseline. Body composition was assessed by body mass index, skinfold thickness, hydrodensitometry, and bioelectric impedance analysis at baseline and annually thereafter for an additional 3 yr. A repeated-measures analysis showed that the change in percentage body fat estimated from skinfold thickness [%BF(SF)] over time changed with increasing Si (P < 0.0001). Si was divided into tertiles for each gender, with the lowest tertile representing the most insulin-resistant children. For girls, those in the lowest tertile maintained their %BF(SF) over 3 yr, whereas those girls in the middle and upper tertile had an increase in their %BF(SF). For boys, those in the lowest tertile showed a decrease in their %BF(SF), whereas those boys in the middle and upper tertile maintained their %BF(SF). These results suggest that during puberty, children who are more insulin resistant have decreased sc fat gain.
OBESITY IS ONE of the most serious nutritional problems in the United States, and it is a problem that continues to become more prevalent. The National Health and Nutrition Examination Survey III data report that more than 30% of adults in the United States are currently obese (1). Similarly, childhood and adolescent obesity has steadily increased with a prevalence in 1991 of 22% (2). Obesity in adolescents has been shown to be associated with increased blood pressure (3), adverse lipoprotein profiles (4), and noninsulin-dependent diabetes (5). Overweight adolescents also appear to have an increased risk of obesity-related mortalities and morbidities later in life independent of their adult weight status (6). Most importantly, adolescent obesity is an antecedent of adult obesity with up to 80% of obese adolescents becoming obese adults (7). Furthermore, the severity of obesity in adults is greater in those who were obese as adolescents (8). Consequently, the identification of those children who are at high risk of becoming obese during puberty is important for the development of successful strategies for obesity prevention.
It has been proposed that insulin resistance is a physiological adaptation to obesity that limits fat deposition and leads to weight stabilization (9). The reverse of this, enhanced insulin sensitivity, may be permissive for weight gain. Several longitudinal studies have shown that individuals who were more insulin sensitive, measured by glucose clamp or fasting insulin levels, had greater rates of weight gain than those individuals who were more insulin resistant (10, 11, 12, 13, 14). In all of these studies, the subjects were adults and the relationship between insulin sensitivity and weight gain was most significant in the subjects with higher body mass indices (BMIs). In two other populations, weight gain was found to be highest in the most insulin-resistant subjects (13, 14). Interestingly, these subjects tended to be leaner than those in the previous studies.
Thus far, there has been only one study in children examining the relationship between insulin sensitivity and weight gain. In prepubertal Pima Indian children, fasting hyperinsulinemia was associated with higher rates of weight gain over 10 yr (15). To further explore this issue and look specifically at body composition changes, we assessed insulin sensitivity in early pubertal children using the frequently sampled iv glucose tolerance test and prospectively followed them for 3 yr.
Subjects and Methods
Subjects
The study population consisted of 111 healthy, nondiabetic children, ranging in age from 9.7–14.5 yr. There were no first-degree relatives with type I insulin-dependent diabetes mellitus. Pubertal development was assessed by the criteria of Marshall and Tanner (16) according to pubic hair and breast or genital development. Of the initial 111 subjects, 16 were excluded from analysis for the following reasons: nine subjects moved or were lost to follow-up before yr 1, two patients had delayed progression of pubertal development, one patient was diagnosed with Kallman syndrome, and one patient with gonadal dysgenesis. All subjects were Caucasian except for three African-Americans; consequently, these three were also excluded from longitudinal analysis. There were 47 boys and 48 girls. Attrition from the study left 86 subjects at yr 2 and 73 subjects at yr 3 for analysis. None of the subjects was on any medications known to influence glucose metabolism. None of the subjects was taking part in a physical fitness training program. The study protocol was approved by the Colorado Multiple Institutional Review Board, and informed consent was obtained from the participating subjects and their parents.
Frequently sampled iv glucose tolerance test
At baseline, insulin sensitivity (Si) was calculated from a frequently sampled iv glucose tolerance test using the modified minimal model of Bergman et al. (17) and Cutfield et al. (18). All subjects were admitted to the Pediatric General Clinical Research Center at The Children’s Hospital in Denver after a 10-h overnight fast. An iv catheter was inserted into each arm. After 30 minutes, 0-min blood samples were drawn. At zero time, 25% dextrose (0.3 g/kg) was infused over 60 sec. Twenty minutes after the dextrose infusion, a bolus of tolbutamide (300 mg/1.73 m2) was infused over 60 sec. Additional blood samples were drawn from the contralateral arm at 2, 4, 8, 19, 22, 25, 30, 35, 40, 50, 70, 90, and 180 min. Each sample was placed in a chilled heparinized tube for the measurement of glucose and insulin.
Anthropometric measurements
Anthropometric and body composition measurements were done at baseline and yearly for three follow-up visits. Height and weight were measured with subjects barefoot and wearing a hospital gown. Height was measured with a Harpendon stadiometer (Seritex, East Rutherford, NJ). BMI was calculated (weight in kilograms divided by the square of the height in meters) (19).
Skinfold thickness measurements were made in triplicate on the nondominant side of the body. Measurements were made to the nearest 0.1 mm with Lange skinfold calipers (Cambridge Scientific Industries, Cambridge, MA) at the triceps, halfway between the acromion process and the olecranon process, and subscapular, 20 mm below the tip of the scapula, at an angle of 45 degrees to the lateral side of the body. The mean of the three measurements was used as the representative value for each site. The equations used for the prediction of percentage body fat were developed by Slaughter et al. (20 20) and take into account the effects of age and gender on body density.
Underwater weighing
Percentage body fat was calculated from hydrostatic determination of body density. Body weights in air and underwater were measured to the nearest 25 g using Heath platform and Chatillon spring scales (Kew Gardens, NY), respectively. Residual lung volume was determined (simultaneously with underwater weighing) using a closed-circuit nitrogen-dilution method. Nitrogen concentration during rebreathing was measured with a Med-Science 505-D Nitralizer (St. Louis, MO). Percentage body fat was estimated from body density using a modified Siri equation as described by Lohman (21). This modified equation takes into account gender and age differences in the density of fat-free mass. Reproducibility tests showed an average difference of 2–4% between measurements for the same subject.
Bioelectric impedance analysis (BIA)
Each subject also underwent BIA to estimate body composition. Resistance and reactance were measured with the subject lying supine by use of an impedance analyzer with two body surface electrodes and a conduction current of less than 1 µA and 50 kHz (model 410, Biodynamics, Bellevue, WA). One surface electrode was placed on the dorsal surface of the right hand and one on the dorsal surface of the right foot. Percentage body fat and fat-free mass were calculated using the prediction equations of Houtkooper et al. (22).
Assays
Immediately after collection, the heparinized blood samples were centrifuged and plasma glucose measurements were made by the glucose oxidase method (model 2300 STAT glucose analyzer, Yellow Springs Instrument Co., Yellow Springs, OH). Additional plasma was stored at -20 C until insulin assay. Insulin was determined by a double-antibody RIA technique (23). The intraassay coefficient of variation for insulin was 5.8% at 287 pmol/liter and 6.4% at 718 pmol/liter. The interassay coefficient of variation for insulin was 5.8% at 287 pmol/liter and 5.4% at 718 pmol/liter.
Calculations
All Si determinations were calculated from glucose and insulin data from the frequently sampled iv glucose tolerance test by the Bergman modified minimal model program (17). In this method, mathematical models of glucose and insulin kinetics are implemented on the computer and are used to analyze the plasma glucose and insulin dynamics following iv glucose and tolbutamide injection.
Statistical analysis
The statistical analysis software system (SAS Institute, Inc., Cary, NC) was used for all statistical analyses. When looking for differences in baseline characteristics within and across gender and Tanner stages, two statistical procedures were used. When comparing two groups at baseline (e.g. boys vs. girls or Tanner II vs. Tanner III), a two-sample t test was used. A one-way ANOVA with four groups (Tanner stage II boys, Tanner stage III boys, Tanner stage II girls, Tanner stage III girls) was used to compare baseline characteristics across these four groups. When the variable of interest was nonnormally distributed, nonparametric analyses or parametric analyses on a suitable transformed variable were performed. When looking at various measures of body fat over time, a repeated-measures analysis was undertaken. This analysis used a linear mixed model taking into account the patients’ gender, Tanner stage, and baseline Si. Furthermore, this linear mixed model analysis was also performed adjusting for the baseline weight of the subjects. All values, unless noted otherwise, are reported as means ± SD.
Results
Baseline characteristics of the study population are presented in Table 1
. Of the 47 boys, pubertal development was Tanner stage II in 28 and stage III in 19. Of the 48 girls, pubertal development was Tanner stage II in 25 and stage III in 23. Comparisons of the baseline data among the four groups have been previously published (24). Of note is that mean Si was significantly lower in girls than in boys (P < 0.05). When Tanner stage was taken into consideration, Si was significantly lower in Tanner stage III girls, compared with each of the three other groups (P < 0.02).
|
depicts the high correlations of each of the four measures of body fat [BMI, percentage body fat estimated from skinfold thickness (SF), underwater weighing (UWW), and BIA]. This table indicates that the correlation coefficients for each of these measures ranged from 0.65–0.92 with the large majority being greater than 0.80. All pairwise correlation coefficients were highly statistically significant (P < 0.0001). Furthermore, these correlation coefficients remained stable over time.
|
Table 3 presents the average %BF(SF) measurements over time for each gender, divided into Si tertiles. Si tertile group 1 consists of the lowest Si values (so the most insulin-resistant subjects), group 2 the middle values, and group 3 the highest Si values. This table shows that in girls, those in the lowest Si tertile are maintaining their %BF(SF) over the 3 yr, whereas those girls in the middle and upper tertiles have an increase in their %BF(SF) by about 5%. For boys, those in the lowest Si tertile show a significant decrease in %BF(SF) by approximately 4%, whereas those boys in the middle and upper tertiles maintained their %BF(SF). This is illustrated graphically in Fig. 1, A and B, which show the change from baseline %BF(SF) over time for each Si tertile group for girls and boys, respectively. Note that these figures plot the mean change in %BF(SF) ± 1 SEM to get a better comparison of the Si tertile groups.
|
|
shows the mean absolute triceps measurements over time for gender, divided by Si tertiles. This table gives similar conclusions as were seen in the %BF(SF) measurements. Discussion
The results of this study showed that insulin resistance in early puberty was associated with maintenance or decreased percentage body fat over a 3-yr period. This relationship was observed only for %BF(SF) measurements and not by other methods. For analysis, Si at baseline was divided into tertiles for each gender. The most insulin-resistant boys had a decrease in their percentage body fat over 3 yr, and the middle and most insulin-sensitive boys (second and third tertiles) maintained their percentage body fat. In girls, the most insulin-resistant group maintained their percentage body fat, and the other two groups had increases in their percentage body fat. Because at baseline the most insulin-resistant subjects were the heaviest, as shown in a previous publication (24), the analyses controlled for initial body weight, and this did not change the relationship observed between Si and changes in percentage body fat.
Over the past few years, several prospective studies have been done evaluating the relationship between Si and future weight gain. Unfortunately, these studies used only BMI as an assessment of body fat, and for the most part, Si was estimated from fasting insulin concentrations. The results from these studies are conflicting. The first of these studies was in a population of Pima Indians, which showed that the most insulin-resistant subjects, as measured by the euglycemic clamp, gained less weight over 3.5 yr than those who were more insulin sensitive (10). The San Antonio Heart Study and San Valley Diabetes Study both used fasting insulin levels as a measure of insulin resistance and found that an increase in fasting insulin predicted lower rates of weight gain over a period of 8 yr and 4.3 yr, respectively (11, 12). A study of Mauritians, which have a multiethnic population, also found a tendency for weight gain to be lowest in Asian Indian and Creole subjects who had higher fasting insulin levels (13). The Atherosclerosis Risk in Communities Study also showed a negative association between fasting insulin and weight change after 6 yr (14). In all of these studies, the relationship between Si and weight change was most pronounced or only significant in individuals with the highest BMIs. Conflicting studies include the Coronary Artery Risk Development in Young Adult Study in which there was a positive association between fasting insulin and weight change (14). In the Mauritian study, weight gain was highest in the Chinese men who were most insulin resistant (13). Additionally, low insulin secretion and high fasting insulin predicted intraabdominal fat accumulation in a group of Japanese-American men (25).
A recent study from the Joslin Clinic evaluated offspring of two parents with noninsulin-dependent diabetes mellitus and looked at acute insulin secretion in addition to insulin sensitivity assessed by a frequently sampled intravenous glucose tolerance test (26). They found that Si was associated with greater weight gain, but there was a strong interdependence of acute insulin secretion with Si. For subjects with low acute insulin secretion, weight gain was the same, regardless of whether Si was low or high. However, for subjects with high insulin secretion, weight gain rate was strongly influenced by Si such that rate of weight gain was low in insulin-resistant subjects and high in insulin-sensitive subjects. From this study, it appears that Si plays a permissive role for weight gain if insulin secretion is sufficient.
There has been only one previous study looking at the relationship between Si and weight gain in children. This study was done in prepubertal Pima Indian children aged 5–9 yr and was designed to understand the predictors of obesity (15). However, most of these children were already overweight with a range of relative weight for height of 70–215% (average 119%). Their results indicated that fasting hyperinsulinemia was associated with higher rates of weight gain over 10 yr. In the majority of the above studies, the outcome was on weight rather than body composition. Consequently, our study is the first to demonstrate a negative relationship between insulin resistance and fat accumulation in children.
There are several mechanisms for how insulin resistance may limit body fat gain. In the insulin-resistant state, there is an increase in hepatic glycogenolysis and gluconeogenesis and decreases in insulin-mediated glucose transport, glycogen synthesis, and glucose oxidation in muscle. In fat metabolism, insulin resistance is associated with increases in lipolysis and very low-density lipoprotein secretion. There is also a decrease in responsiveness to insulin in sc adipose tissue lipoprotein lipase (27). All of these changes create a metabolic environment that favors less effective fat storage and increased fat oxidation that could limit fat accumulation.
In this study, insulin sensitivity related only to changes in %BF(SF). Insulin sensitivity did not relate to changes in weight, BMI, or percentage body fat when calculated from UWW or BIA. Both weight and BMI take into account stature and lean body mass as well as body fatness. Consequently, in children who are actively growing, changes in weight and BMIs will not solely reflect changes in body fatness. Changes in percentage body fat is thus a better way to assess for body fat accumulation. Insulin sensitivity was not related to changes in percentage body fat when calculated UWW or BIA. There are several possible explanations for this finding. First, to adjust for the chemical immaturity of fat-free mass in children, age- and gender-specific constants were used to calculate body composition from body density (21). However, these constants are age and not pubertal or bone age specific. Because children enter puberty at different chronological ages and their development is more related to their stage of puberty and skeletal maturation, constants derived according to pubertal stage or bone age may be more accurate. Second, UWW can be technically difficult, especially for children. For both of these reasons, the percentage body fat measurements from UWW may not be totally reliable for longitudinal analysis. Dual-energy x-ray absorptiometry analysis may be a preferred way of assessing body composition in children; however, at the time of this study, this technique was not widely available.
In addition to Si being related to changes in %BF(SF), it was also related to changes in the absolute value of SF thicknesses. Because all of our other measurements of percentage body fat estimated total body fat, it is possible that Si relates more to changes in sc fat stores than total or visceral fat. At present, evidence suggests that insulin action is greater in sc than visceral adipose tissue. This is true for both the antilipolytic effects of insulin (28, 29) and the stimulatory effects of insulin on lipoprotein lipase (30). These differences relate, at least in part, to decreases in insulin receptor affinity and postreceptor signaling in visceral adipocytes (31).
In summary, this study shows that in early pubertal children, insulin resistance may serve to limit body fat changes. This study, consequently, provides further evidence that insulin resistance may be a physiologic adaptation to obesity
Acknowledgments
We thank The Energy Balance Laboratory of The Colorado Clinical Nutrition Research Unit for performing the underwater weight measurements, and Shayne Bland for help with some of the statistical analyses. We also appreciate The Children’s Hospital and the General Clinical Research Center and the staff for tremendous effort in helping to complete this project.
Footnotes
This work was supported in part by Grants M01 RR00069 and RR00051 from the General Clinical Research Centers Program, National Center for Research Resources, NIH.
Abbreviations: %BF(SF), Percentage body fat estimated from skinfold thickness; BIA, bioelectric impedance analysis; BMI, body mass index; SF, skinfold thickness; Si, insulin sensitivity; UWW, underwater weighing.
Received August 29, 2001.
Accepted May 9, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  J. A Morrison, C. J Glueck, P. S Horn, G. B Schreiber, and P. Wang Pre-teen insulin resistance predicts weight gain, impaired fasting glucose, and type 2 diabetes at age 18-19 y: a 10-y prospective study of black and white girls Am. J. Clinical Nutrition, September 1, 2008; 88(3): 778 - 788. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. Fleisch, N. Agarwal, M. D. Roberts, J. C. Han, K. R. Theim, A. Vexler, J. Troendle, S. Z. Yanovski, and J. A. Yanovski Influence of Serum Leptin on Weight and Body Fat Growth in Children at High Risk for Adult Obesity J. Clin. Endocrinol. Metab., March 1, 2007; 92(3): 948 - 954. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. H Eckel, T. L Hernandez, M. L Bell, K. M Weil, T. Y Shepard, G. K Grunwald, T. A Sharp, C. C Francis, and J. O Hill Carbohydrate balance predicts weight and fat gain in adults. Am. J. Clinical Nutrition, April 1, 2006; 83(4): 803 - 808. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. T. Rhodes, J. I. Wolfsdorf, D. D. Cuthbertson, H. A. Feldman, D. S. Ludwig, and the DPT-1 Study Group Effect of Low-Dose Insulin Treatment on Body Weight and Physical Development in Children and Adolescents at Risk for Type 1 Diabetes Diabetes Care, August 1, 2005; 28(8): 1948 - 1953. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Tong, W. Y. Fujimoto, S. E. Kahn, D. S. Weigle, M. J. McNeely, D. L. Leonetti, J. B. Shofer, and E. J. Boyko Insulin, C-Peptide, and Leptin Concentrations Predict Increased Visceral Adiposity at 5- and 10-Year Follow-Ups in Nondiabetic Japanese Americans Diabetes, April 1, 2005; 54(4): 985 - 990. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. P. Araujo, C. T. De Souza, A. L. Gasparetti, M. Ueno, A. C. Boschero, M. J. A. Saad, and L. A. Velloso Short-Term in Vivo Inhibition of Insulin Receptor Substrate-1 Expression Leads to Insulin Resistance, Hyperinsulinemia, and Increased Adiposity Endocrinology, March 1, 2005; 146(3): 1428 - 1437. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. H. Slyper The Pediatric Obesity Epidemic: Causes and Controversies J. Clin. Endocrinol. Metab., June 1, 2004; 89(6): 2540 - 2547. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
