Relation of Leptin to Insulin Resistance Syndrome in Children


MMC 94, 420 Delaware St. SE, Minneapolis, MN 55455. E-mail:


Objectives: To examine the relation of leptin to insulin resistance, as measured by euglycemic insulin clamp, and insulin resistance syndrome factors in thin and heavy children.

Research Methods and Procedures: Anthropometrics, insulin, blood pressure, and leptin were measured in 342 11- to 14-year-old children (189 boys, 153 girls, 272 white, 70 black). Insulin sensitivity (M) was determined by milligrams glucose uptake per kilogram per minute and expressed as M/lean body mass (Mlbm). Children were divided by median BMI (boys = 20.5 kg/m2; girls = 21.4 kg/m2) into below-median (thin) and above-median (heavy) groups. Correlation coefficients between log-leptin and components of insulin resistance syndrome were adjusted for Tanner stage, gender, and race.

Results: BMI was related to leptin in boys (r = 0.70, p < 0.001) and girls (r = 0.75, p < 0.001). Leptin was higher in girls than boys (32.6 vs. 12.3 ng/mL, p = 0.0001). Leptin levels increased in girls and decreased in boys during puberty, paralleling the changes in body fat. Leptin was significantly correlated with insulin, Mlbm, triglycerides, and blood pressure in heavy children and only with insulin in thin children. After adjustment for body fat, the correlations remained significant for insulin and Mlbm in heavy children and with insulin in thin children.

Discussion: Significant associations were found between leptin and insulin resistance in children, and these associations were attenuated by adjustment for adiposity. These findings at age 13 likely have long-term consequences in the development of the obesity-insulin resistance-related cardiovascular risk profile.


Insulin resistance syndrome, a clustering of obesity, hypertension, hyperinsulinemia, and dyslipidemia, is strongly associated with type 2 diabetes and cardiovascular disease (1, 2, 3). Obesity has a central role in insulin resistance syndrome. Leptin, a product of the ob gene, is a peptide produced by differentiated adipocytes (4, 5, 6), and it controls energy metabolism at the level of the hypothalamus by suppressing food intake and stimulating energy expenditure (7, 8, 9, 10, 11). Leptin levels are elevated in obese adults (12, 13, 14) and children (15), indicating resistance to the effects of leptin in these individuals (16).

Leptin has been associated with insulin resistance in adults; some studies report this association to be dependent on body fatness (12) and others as independent of body fatness (17, 18). In contrast, there has been only one study in children using direct measures of insulin resistance to study the relation of leptin to adiposity and insulin resistance (19). However, components of insulin resistance syndrome are present in children long before the development of overt cardiovascular disease (20, 21, 22). Because both leptin and insulin resistance are strongly related to adiposity and other cardiovascular risk factors (20), studying these relations in childhood may help clarify some aspects in the development of the insulin resistance syndrome. A cross-sectional study of Taiwanese school children reported independent associations for obesity and leptin levels with fasting insulin, blood pressure, and triglycerides (23). In a study of prepubertal lean and obese children and in children with insulin-dependent diabetes, fasting insulin explained a significant portion of the variance of leptin (24). More recently, a study in 4- to 17-year-old children showed that BMI explained most of the variability in leptin and concluded that in obesity, total adiposity, but not insulin, is the main determinant of leptin levels (25). Fasting insulin is an indirect measure of insulin resistance. Moreover, the use of insulin to assess the relation between leptin and insulin resistance may not be optimal, because fasting insulin is highly related to adiposity (26).

The overall long-term objective of this research was to determine the influence of insulin resistance and fatness on the development of type 2 diabetes and cardiovascular risk. In this study, we investigated the relations between leptin and insulin resistance in children using euglycemic insulin clamp, which is recognized as being a more precise measure of insulin resistance than fasting insulin (27). The goal was to describe the leptin-obesity-insulin resistance relations in this age group in an attempt to better understand their influence on the development of the insulin resistance syndrome.

Research Methods and Procedures

The study was approved by the Institutional Review Board Human Subjects Committee of the University of Minnesota. Consent was obtained from all children and their parents/guardians.

The participants in this study were recruited after blood pressure screening of 12, 043 fifth to eighth grade Minneapolis Public Schools students (4216 white, 3819 black, 1383 southeast Asian, 569 Native American, 397 Hispanic, and 1659 mixed ethnicity children; 6035 boys and 6008 girls), representing 93% of all eligible students in those grades. As previously described (20), recruitment letters were mailed to randomly selected black and non-Hispanic white children with stratification according to sex, race, and blood pressure (one-half from the upper 25 percentiles and one-half from the lower 75 percentiles to enrich the study population with potentially higher-risk children). Euglycemic insulin clamps were completed in 357 participants as previously described (20). Serum samples obtained at the time of the clamps and stored at −70 °C were available for leptin measurement in 342 children (189 boys and 153 girls; 70 black and 272 white), forming the cohort for the present report.

All participants had a physical examination performed by a board-certified pediatrician. Children were divided into Tanner stages according to pubic hair development in boys and breast and pubic hair development in girls. For the purposes of these analyses, leptin and insulin sensitivity data for Tanner stages 2 to 4 (T2 to T4; mid-puberty)1 were combined and compared with data from Tanner 1 (T1; prepuberty) and Tanner 5 (T5; end-puberty). Body composition measurements were obtained by trained research personnel. Height was measured using a wall-mounted stadiometer, and weight was determined using a balance scale. Triceps and subscapular skinfold thicknesses were measured in duplicate to the nearest millimeter with Lange calipers (test-retest CV: for triceps = 2.6%; for subscapular skinfold = 3.7%), and the mean values were used in the analyses. Waist circumference was calculated as the average of two measurements and measured at the smallest circumference of the waist (test-retest CV = 0.5%). Percentage body fat (%BF) for the boys and girls was calculated from the equations devised by Slaughter et al. (28) based on triceps and subscapular skinfold values, taking into account gender, ethnicity, and stage of maturation. Based on these equations, fat body mass by skinfold (FBM) was calculated as %BF × weight (kilograms), and lean body mass by skinfold (LBM) was calculated as weight − FBM.

Blood pressure was measured twice using random-zero sphygmomanometers, with participants seated after a 5-minute rest and 1 minute between measures.

Euglycemic insulin clamp studies were conducted in the University of Minnesota Clinical Research Center, as previously described (26). Participants were admitted after a 10-hour overnight fast. Blood samples for serum insulin levels were obtained at baseline (−10, −5, and 0 minutes before starting the insulin infusion) and at steady state during the clamp (+140, +160, and +180 minutes). Plasma glucose was also measured at baseline (−10, −5, and 0 minutes) and every 5 minutes during the clamp. The insulin infusion was started at time 0 and continued at a rate of 1 mU/kg per minute for 3 hours. An infusion of 20% dextrose was started at time 0 and was adjusted based on the plasma glucose levels to maintain euglycemia, i.e., plasma glucose at 100 mg/dL (5.6 mM). Insulin sensitivity, M, was determined from the amount of glucose required to maintain euglycemia over the final 40 minutes of the euglycemic clamp study and was expressed as Mlbm (i.e., glucose use per kilogram lean body mass per minute).

The blood samples for insulin were collected on ice and centrifuged within 20 minutes. Insulin levels were measured by radioimmunoassay (Diagnostic Products Corporation, Los Angeles, CA). The intra-assay CV was 2% to 9%, the interassay CV was 4.9% to 7.3%, and its cross reactivity to proinsulin was 20%. Lipid levels were determined in the University of Minnesota laboratory. Plasma glucose was measured immediately at the bedside with a Beckman Glucose Analyzer II (Beckman Instruments, Fullerton, CA). Plasma samples used for leptin determination were stored at −70 °C and were not thawed until the leptin assay was performed. Leptin levels were measured in duplicate by the DSL-10–23, 100 ACTIVE Human Leptin ELISA, an enzymatically amplified “two-step” immunoassay. The sensitivity (minimum detection limit) of the assay is 0.05 ng/mL. The intra-assay CV was 1.5% to 6.2%, and the interassay CV was 3.3% to 5.3%. Because the distribution of leptin is skewed, natural log-leptin + 1 was used for all analyses.

All data are expressed as mean ± SEM. Pearson correlation analysis and multiple linear regression were used for analyses. Comparisons were adjusted by linear regression for race, gender, and Tanner score. A p value of 0.05 or less was considered to be statistically significant. All statistical analyses were conducted using the SAS statistical package, version 8.2 (SAS Institute, Cary, NC).


The clinical and laboratory characteristics of the participants are described in Table 1. Of the 342 children, 55% were boys and 20% were black. The mean age of the participants was 13 ± 0.1 years. As expected for this age, girls were at a more advanced Tanner stage than boys. The boys were taller than the girls and had a greater LBM, but %BF, FBM, and skinfold thickness were greater in the girls. Mlbm was significantly lower and fasting insulin was significantly higher in girls. Leptin was higher in girls. All children were normotensive, and there were no significant differences in lipids between genders.

Table 1.  Clinical and laboratory data (mean ± SEM) on 342 children
  • *

    p ≤ 0.005 between boys and girls.

Age (years)13.0 ± 0.312.9 ± 0.1
Tanner score3.2 ± 0.13.7 ± 0.1*
Height (cm)163.5 ± 0.7160.2 ± 0.6*
Weight (kg)58.4 ± 0.357.7 ± 1.0
BMI (kg/m2)21.5 ± 0.122.1 ± 0.3
%BF (%)25.4 ± 0.930.5 ± 0.6*
FBM (kg)15.8 ± 0.818.1 ± 0.8*
LBM (kg)42.3 ± 0.638.8 ± 0.4*
Triceps skinfold (cm)19.6 ± 0.724.1 ± 0.7*
Waist (cm)78.3 ± 0.875.3 ± 0.9*
Systolic blood pressure (mm Hg)108.0 ± 0.6106.2 ± 0.7
Diastolic blood pressure (mm Hg)54.9 ± 0.957.6 ± 1.0
Total cholesterol (mg/dL)151.0 ± 2.3153.4 ± 1.9
Triglycerides (mg/dL)87.9 ± 3.794.4 ± 4.2
Low-density lipoprotein-cholesterol (mg/dL)88.4 ± 2.089.3 ± 1.7
High-density lipoprotein-cholesterol (mg/dL)44.9 ± 0.645.1 ± 0.8
Fasting insulin (mU/L)13.8 ± 0.715.7 ± 0.8*
Mlbm (mg/kg/min)13.3 ± 0.311.6 ± 0.3*
Leptin (ng/ml)12.3 ± 1.132.6 ± 1.7*

As shown in Figure 1, A and B, boys had higher leptin levels than girls at prepuberty (T1, p = 0.05), but leptin became significantly higher in girls than boys at T2 to T4 (p = 0.0001) and at T5 (p = 0.0001). Leptin levels increased significantly and steadily in girls over the course of puberty (T1 to T2–T4, p = 0.01; T2–T4 to T5, p = 0.02), and this was paralleled by an increase in %BF (T1 to T2–T4, p = 0.02; T2–T4 to T5, p = 0.01). Leptin levels in boys decreased significantly from T1 to T2–T4 (p = 0.003) and leveled off thereafter (T2–T4 to T5, p = 0.6). A similar decrease was noted in %BF (T1 to T2–T4, p = 0.03; T2–T4 to T5, p = 0.1). Despite these differences in %BF and leptin levels between boys and girls, the change in insulin resistance during puberty was similar, increasing significantly from T1 to T2–T4 (p = 0.003) and decreasing, although not significantly, from T2–T4 to T5 (p = 0.7).

Figure 1.

Changes in log-leptin, percentage body fat, and insulin resistance (Mlbm) from Tanner stages 1 to 5 in boys (A) and girls (B).

Leptin was significantly correlated with all measures of body fatness in boys and girls and with LBM in girls. It was also significantly correlated with fasting insulin and Mlbm (inversely) before and after adjustment for %BF for the entire cohort (data not shown) and for boys and girls separately (Table 2). Associations of leptin with other components of insulin resistance syndrome are also shown in Table 2. Correlations were significant with triglycerides in both genders, with systolic blood pressure in boys, and with high-density lipoprotein-cholesterol in girls. After adjustment for %BF, the correlation became insignificant in both genders for blood pressure and lipids, with the exception of triglycerides in girls. The results were similar when FBM or BMI were substituted for %BF in the analyses.

Table 2.  Correlations of leptin with body composition and components of insulin resistance syndrome
 UnadjustedAdjusted for percentage body fat
  1. HDL-C, high-density lipoprotein-cholesterol.

Fasting insulin    
Systolic blood pressure    

In an attempt to determine whether the relation between leptin and insulin resistance risk factors differed in thin and heavy children, we divided the cohort by median BMI (boys = 20.5 kg/m2, girls = 21.4 kg/m2) and examined the relations of leptin to fasting insulin and Mlbm in the group with below-median BMI and the group with above-median BMI (Table 3). Leptin levels were significantly higher in the above-median BMI group than below-median BMI group (33.89 ± 1.43 ng/mL vs. 12.36 ± 1.30 ng/mL, p = 0.0001). Leptin was significantly related to fasting insulin, Mlbm, triglycerides, and systolic blood pressure in the above-median BMI group. In the below-median BMI group the correlation was only significant between leptin and fasting insulin. After adjustment for %BF, leptin remained significantly associated with fasting insulin and Mlbm in the above-median BMI group and continued to be significantly correlated with fasting insulin in the below-median BMI group.

Table 3.  Correlations of leptin with components of insulin resistance syndrome in thin (below median BMI) and heavy (above median BMI) children
 UnadjustedAdjusted for percentage BF
 Above median BMIBelow median BMIAbove median BMIBelow median BMI
  1. HDL-C, high-density lipoprotein-cholesterol.

Fasting insulin    
Systolic blood pressure    


This cross-sectional study in 342 children describes the relations of leptin with insulin resistance and factors associated with insulin resistance syndrome, using the euglycemic insulin clamp. The results confirm the significant association between leptin and obesity. However, these results also show a significant relation between leptin and insulin resistance that is independent of body fatness and found primarily in heavy, as opposed to thin, children.

Insulin resistance syndrome (also known as metabolic syndrome or syndrome X), a well-recognized constellation of risk factors for type 2 diabetes and atherosclerotic cardiovascular disease, has its roots in childhood. Autopsy studies revealed that early evidence of the disease processes can be found in children (29, 30). Clinical studies in children have shown a significant relation of fasting insulin and insulin resistance to cardiovascular risk factors. Studies have also shown that obesity has a central role in the syndrome (20, 22, 31, 32). Because of its strong relation to adiposity, it is reasonable to suggest that leptin may play a role in the development of insulin resistance syndrome. With recent marked increase in childhood obesity (33, 34) and type 2 diabetes (35, 36), questions regarding the role of obesity-related factors, such as leptin, and insulin resistance in children have increased in importance.

Strong relations between adiposity and leptin were associated with the gender differences in leptin levels. Leptin was higher in girls than boys at all Tanner stages of development after the initiation of puberty, in agreement with previous studies in children (23, 37). This is probably related to greater accretion of fat mass in girls in contrast to the greater increase in lean mass in boys during adolescence (38, 39), as recently reported in 27 pubertal children (19). Thus, it is apparent from both the current results and the existing literature that puberty is a state of increasing leptin resistance in girls. The increase in leptin levels observed between prepuberty (T1) and puberty (T2 to T4) suggests that this surge in leptin in females may be under gonadal and hormonal influence and important for adolescent development. This hypothesis is supported by a recent report showing that the pattern for leptin levels through puberty in girls is correlated with estradiol levels, whereas the lower leptin in boys is correlated with testosterone levels (15). Animal studies have shown that the reproductive development in females is preceded by an increase in leptin levels, suggesting that plasma leptin concentration may be a signal for onset of sexual maturity and fertility (40, 41).

The heavier children in this cohort (i.e., the group with BMI above the median) had higher leptin levels and a significant correlation between leptin and components of insulin resistance syndrome, in contrast to children with BMI below the median. Although fatness seems to play a prominent role in the relations of leptin to insulin resistance, fatness must also introduce another component, because the leptin-Mlbm relation remains significant in the heavy children, even after adjusting for %BF. It is not clear why this should be present only in heavy children; however, it is possible that the significantly higher levels of leptin and insulin resistance represent as yet unidentified metabolic disturbances associated with an increase in adiposity, which is known to be generally associated with an increased cardiovascular risk.

In a study of 47 children, of which 27 were in pubertal stages similar to the present cohort, the relationship between leptin and insulin resistance disappeared after adjustment for body fatness determined by H2180 (19). In the present study, the significant associations between leptin and insulin resistance were maintained after adjustment for adiposity (%BF, FBM). The explanation for this difference is not clear but may be related to either differences in the number of children studied or to potentially reduced accuracy in this study because of the use of skinfold measurements to determine body fatness. Positive associations of leptin with hyperinsulinemia have been previously reported in adults and children (17, 18, 19) and with insulin sensitivity (euglycemic clamp) independent of adiposity in adults (17), suggesting that there may be an increased resistance to the effects of leptin in insulin-resistant states. Data from studies in obese (ob/ob) and normal insulin-resistant and lipoatrophic diabetic mice suggest that leptin can improve insulin resistance when leptin deficiency is the probable cause of insulin resistance (42, 43). However, administration of leptin has no effect on insulin resistance in leptin-resistant (db/db) mice. Thus, the coexistence of hyperleptinemia and hyperinsulinemia in humans may be explained by leptin resistance. This study extends these results to children by demonstrating that leptin and insulin resistance measured by the euglycemic clamp are correlated independent of adiposity. Based on these reports, we hypothesize that the independent association between insulin resistance and leptin resistance, beginning in childhood, may become stronger with aging and the long-term insulin resistance induced by chronic obesity (44). Thus, it is reasonable to speculate that weight management in the young could potentially modify the relation between leptin and insulin resistance and modify the development of insulin resistance syndrome and cardiovascular risk.

Associations of leptin with other components of insulin resistance syndrome were dependent on adiposity, underscoring the importance of obesity in the link between leptin and components of insulin resistance syndrome. Relations between leptin and triglycerides were stronger in the presence of fatness but were also present independent of body fatness in boys only. A similar pattern was suggested in a study in 203 obese children, where triglycerides and fasting insulin explained most of the variation in leptin in boys but not girls (25). It is known that leptin lowers tissue triglyceride levels (45), and it has been suggested that the triglyceride content of tissues sets the level of both insulin resistance and insulin production (46). This finding in adolescent males may be relevant to the increased incidence of premature adult cardiovascular risk in men. Following these factors longitudinally, as children make the transition into young adulthood, should help clarify the mechanisms involved with these associations.

In summary, this study has confirmed the significant difference in leptin levels between adolescent boys and girls and documented that this difference develops during puberty in association with gender-specific changes in body fat. In addition, the study has extended previous observations by showing a significant relation between leptin and insulin resistance (Mlbm) independent of body fat and in heavy, as opposed to thin, children. These results suggest that a clearer understanding of the leptin-obesity-insulin resistance relationship will require natural history studies, beginning at birth, and that the role of insulin resistance relative to the development of risk factors and type 2 diabetes needs to be contrasted between and within heavy and thin children.


This study was supported by Grants HL 52851, 1K23 HL04000-03, and M01 RR00400 from the National Institutes of Health.


  • 1

    Nonstandard abbreviations: T2, Tanner stage 2; T4, Tanner stage 4; T1, Tanner stage 1; T5, Tanner stage 5; %BF, percentage body fat; FBM, fat body mass; LBM, lean body mass.