Main

It is thought that leptin, the protein secreted from adipose tissue(1), is part of the regulatory feedback loop that contributes to the control of body weight, particularly that of body fat. Considine et al.(2) reported higher serum concentrations of leptin for obese adults when compared with normal-weight adults, and the serum levels were correlated with%Fat. When male and female subjects with equivalent fatness were compared, there was no difference in serum leptin levels between the sexes. Recently, Rosenbaum et al.(3) reported the effects of body composition and gender on plasma leptin levels of adults, including pre- and postmenopausal female subjects. Gender differences in leptin concentration were observed, as were differences in female subjects based on their menopause status. Hassink et al.(4) recently reported serum leptin values for children. As in the studies of adults, the children with excessively high BMI values (>95th percentile) had significantly higher serum leptin concentrations than children within the normal BMI range. Leptin levels were found to decrease with advancing Tanner stage, and girls had higher leptin levels than boys.

In this study, we measured serum leptin levels and body composition of children and young adults. We wanted to determine: 1) the relation of serum leptin concentrations with body FM and relative fatness (%Fat and BMI); 2) the effects of gender, ethnicity, and sexual development on these relationships; and 3) whether these relationships were different between the prepubertal children, the adolescents at Tanner stage 5, and the young adults.

METHODS

Subjects. The children examined were participants in a larger study focused on the diversity of growth patterns among European-, African-, and Mexican-American populations(5). All the subjects were in good health, were not taking medications that would alter body composition, and provided a 24-h dietary record. For this study, a subset was selected on the basis of body fatness (defined as fat/body weight) to provide a full range of fatness values representing the total population. This subset consisted of 183 children (91 male, 92 female) spanning the age range of 3-18 y, and 27 young adults (20-29 y of age). A midday blood sample, taken after a 3-h fast, but before a lunch meal, was available for each subject. Sexual development was based on Tanner staging for the children(6). All subjects in the adult group were assigned a Tanner stage of 5. This study was approved by the Institutional Review Board of Baylor College of Medicine. An informed written consent was obtained for each participant before enrollment in the study.

Procedures. Body composition was measured using DXA. The instrument was a Hologic QDR-2000 operated in the single beam mode (software version 5.56). This technique provides a measure of the subject's FM, bone mineral mass, and lean tissue mass(7, 8). The DXA measurement also provides a measure of body fatness defined as%Fat = 100× (FM/weight).

The frozen serum samples were used for the measurement of leptin concentration. The solid phase immunoassay used affinity-purified polyvalent antibodies to recombinant human leptin(3). Serum samples and recombinant human leptin standards were bound to antibody-coated microtiter wells. The bound leptin was detected with horseradish peroxidase-conjugated antibody and quantitated with tetramethylbenzidine/peroxide. The assay detection limit was 0.020 ng/mL. The intra- and interassay variance were 3.5 and 8%, respectively.

It is well known that serum leptin concentration correlates with various indices of FM for adults(2, 3), so that subjects with higher FM usually have higher serum leptin concentrations. However, two children with the same FM may have quite different serum leptin levels depending on their overall body size. For example, a 6-y-old and a 14-y-old girl may have the same FM, yet there would be substantial differences in the size of the lean body mass. Furthermore, PV tends to be a relatively constant fraction of the lean body mass(9). Thus, dividing total circulating leptin (presumably produced by adipose tissue mass) by PV should result in a lower plasma leptin concentration for the older girl with the same FM, but larger PV. Although it is relatively easy to obtain a serum sample, it may be more physiologically relevant to relate the total leptin content of the PV to total FM. We used an algorithm(9) based on the individual's weight and height to calculate PV (PV) and defined LTP =[serum leptin] × PV.

Statistical analysis. Data presented in the tables are mean± SD. Comparisons between gender groups at the same Tanner stage and between Tanner stages within the same gender were performed using a t test. ANOVA was used to test for the significance of the covariates of gender, age, race, and Tanner stage on the relationships between serum leptin concentrations and FM,%Fat, and BMI. Linear regression analyses were performed to determine the relationships between serum leptin levels and body FM,%Fat, and BMI for each gender. As the distribution of FM can be skewed in the general population, linear regressions among these parameters were also performed after a log transformation of these data. For all statistical analyses, a p value <0.05 was considered statistically significant.

RESULTS

The anthropometric and body composition data for the children and young adults are presented in Table 1. The results were categorized by gender and Tanner stage. Table 2 contains the corresponding values for serum leptin, LTP, and serum leptin/FM for each group. The relationship between serum leptin and FM for the children in this study is presented in Figure 1. As is evident in this figure, this relationship is gender-dependent. The results of linear regression analysis for serum leptin as a function of FM,%Fat, and BMI for each gender are presented in Table 3. Although serum leptin concentration was significantly correlated with the two indices of body fatness (%Fat and BMI), the highest correlations were with the absolute FM. As seen in Figure 1, the girls generally had higher serum leptin values than the boys for the same amount of FM. This difference was confirmed by ANOVA in which gender was a significant covariate (p< 0.0001) for the relationships between serum leptin concentrations and FM,%Fat, and BMI. ANOVA also indicated that sexual development, based on the Tanner classification, had a significant effect (p < 0.05) on these relationships, whereas race did not (p > 0.85). The results of the t test analyses indicated that the most significant different between adjacent Tanner groups was for Tanner stage 1 versus 2 for both gender groups. When serum leptin values were normalized for variations in FM, there was no effect related to Tanner classification for female or male subjects.

Table 1 Anthropometric and body composition values for children and young adults (mean ± SD)
Full size table
Table 2 Leptin values for children and young adults(mean ± SD)
Full size table
Figure 1
figure 1

Relationship between serum leptin concentrations and total body FM. The solid line is the linear function for female subjects(○); the dashed line is for male subjects (▪).

Full size image
Table 3 Results of regression analyses for leptin vs fat,% Fat, and BMI
Full size table

Because the FM values do not fit a normal distribution, these data and that for the serum leptin concentration and LTP were log-transformed for further analyses. The log-log relationship between LTP and FM for females had the highest correlation (r = 0.93, p < 0.0001); the equation was: LTP (μg) = 0.259 FM (kg)1.65. More than 87% of the variation in the estimated total plasma leptin level could be accounted for by the variation in FM. The LTP versus FM relationship for the male subjects was also significant (r = 0.70,p < 0.001), similar to that obtained for the female subjects, but there was increased scatter in the LTP values for a given value of FM.

The range of serum leptin/FM at each Tanner stage is presented in Figure 2 for each gender. The mean values for leptin/FM are presented in Table 2. Two young boys had extremely high leptin/FM ratios, which could be attributed to very low FM values rather than high serum leptin values. There was significant overlap in the range of leptin/FM values between girls and boys at each Tanner stage. However, the lowest values observed for the girls were generally higher than those detected for the boys at each Tanner stage. No difference could be attributed to ethnicity in the range of leptin/FM values within each gender. There was a statistical difference for leptin/FM for boys between Tanner stage 1 and 2. However, the slight decrease in this ratio for increasing Tanner stages was not significant. The mean values for the females remained relatively invariant across all Tanner stages, although there was a statistical difference between stage 1 and 2 as observed for the boys. There were gender differences at each Tanner stage and for the young adults; the females had a significantly higher mean leptin/FM value than the males (p < 0.05-0.001).

Figure 2
figure 2

Relationship between the serum leptin/FM ratio and Tanner stage for female (○) and male subjects (▪). The mean values for each gender and Tanner stage are shown.

Full size image

The serum leptin concentrations, the estimated total circulating amount of leptin (LTP), and the serum leptin/FM ratio for the young adults were not statistically different from that for the Tanner stage 5 group.

DISCUSSION

We observed that serum leptin concentrations increased with FM in children. We also observed that serum leptin levels were higher in female than in male subjects at all pubertal stages, even when adjusted for differences in FM. Furthermore, the total circulating plasma leptin content was a nonlinear function of FM. Although these relationships differed on the basis of gender, there was no effect related to ethnicity.

There is a paucity of serum leptin data for children with which to compare the findings of this study. Hassink et al.(4) reported a mean ± SD value of 7.8 ± 6.5 ng/mL for serum leptin for normal weight children defined as having a BMI classification below the 85th percentile for age, race, and gender(10). The children in our study with this same BMI classification had mean serum leptin values of 3.4 ± 3.4 ng/mL for boys, and 7.8 ± 5.8 ng/mL for girls. For an obesity classification, Hassink et al.(4) used BMI values above the 95th percentile; the corresponding mean serum leptin value was 38.6 ± 21 ng/mL. The boys in our study with BMI values above the 95th percentile had a mean serum leptin concentration of 14.3 ± 10.7 ng/mL, whereas the girls with a similar BMI classification had 19.9 ± 12.2 ng/mL.

A number of studies have reported serum leptin values for adults. Considine et al.(2) reported a mean value of 7.5± 9.3 ng/mL for normal weight adults, and 31.3 ± 24.1 ng/mL for obese adults. Rosenbaum et al.(3) reported mean serum leptin values of 14.9 ± 25.5 ng/mL for men, 46.3 ± 37.5 ng/mL for premenopausal women, and 23.0 ± 14.7 ng/mL for postmenopausal women. Rosenbaum et al.(3) also provided regression equations that described the linear relationships between serum leptin concentrations and FM for adult male and female subjects. If we substitute the FM values observed for our young girls into the equations of Rosenbaum et al.(3), the predicted serum leptin values were consistently higher (approximately 55%) than those we observed for the girls with the same FM. It is hard to attribute this difference to the methods used to assess body composition. Rosenbaum et al.(3) used hydrodensitometry, whereas we used whole-body DXA measurements. The reported 2-3% difference for the estimates of FM between these methods(11, 12) is not enough to account for the larger differences in predicted serum leptin concentration for the same FM. Differences due the leptin assay are also unlikely as the samples for this study and those of Rosenbaum et al.(3) were assayed in the same laboratory. There is the possibility that some of the teenage female subjects in our study were restricting food intake voluntarily(13), and Considine et al.(1) have reported that intentional efforts at weight loss can be associated with a substantial reduction in serum leptin concentration. Malmström et al.(14), on the other hand, have demonstrated that infusion of insulin increases serum leptin levels, whereas Muscelli et al.(15) observed no change in plasma leptin levels for obese and lean subjects studied in the fasting state and during two hours of euglycemic hyperinsulinemia. Both of these investigators, however, concluded that serum leptin levels may increase after a prolonged period of elevated insulin even at concentrations that are comparable to those observed in slightly overweight normal subjects. That is, obesity-induced chronic hyperinsulinemia may contribute to raising leptin levels in adults.

The mean leptin/FM value for men reported by Rosenbaum et al.(3) was consistent with our findings for the male subjects in Tanner stages 3-5, but was higher than the 0.2 (ng/mL)/kg observed for the younger male subjects in our study. A comparison of our results for male children with the serum leptin versus FM relationship reported by Rosenbaum et al.(3) for adult male subjects do not produce a systematic difference in the predicted serum leptin levels for the same FM. However, only about 41% of the variation in serum leptin levels could be attributed to FM, and adjusting for differences in body size by converting serum leptin concentrations to LTP did not improve the correlation with FM. It has been proposed that androgens may have a suppressive effect on serum leptin levels(16). If true, then this effect could be evident in boys during the time of most rapid muscle growth. Alternately, the decreased serum leptin concentrations during this period of intense pubertal development may reflect mainly only the loss of excess body fat during these years(17).

We did not observe a significant change in the mean leptin/FM ratio with advancing sexual maturation for either gender. Although there was significant overlap in the range of leptin/FM values between boys and girls, the mean leptin/FM values for the total group of boys was significantly lower(p < 0.0001) than that for the girls. Rosenbaum et al.(3) also observed a lower leptin/FM ratio for adult male subjects when compared with the value for adult female subjects. These differences were evident at each Tanner stage. Unfortunately, Hassink et al.(4) did not report leptin/FM ratios for their study of children.

We observed the lowest mean serum leptin concentrations and LTP for those children in the Tanner stage 1 group. The highest leptin concentration for boys was at Tanner stage 2, followed by a slight decrease, which did not reach statistical significance. For the girls, the mean leptin concentration remained relatively constant for Tanner stage 3-5 groups. Hence, our findings do not support the general concept of “leptin resistance” during puberty as proposed by Hassink et al.(4) for all children. The neuroendocrine role of leptin in the regulation of body weight and its composition is as yet only partly known(18, 19). Our study suggests that gender differences in the serum leptin concentrations and leptin/FM ratio are already evident at prepubertal ages. It is possible that there are differences in the clearance of leptin from the blood, and/or variations in the transport system to the brain's leptin receptor site(20–24) that are gender-dependent. Longitudinal studies of children may help to better define the potential role of leptin during growth and development.