International Journal of Obesity (1999) 23, 678±685
ß 1999 Stockton Press All rights reserved 0307±0565/99 $12.00
http://www.stockton-press.co.uk/ijo
Serum leptin concentration, body composition,
and gonadal hormones during puberty
EW Demerath1*, B Towne1, W Wisemandle1, J Blangero2, W Cameron Chumlea1 and RM Siervogel1
1
Wright State University School of Medicine, Department of Community Health, Division of Human Biology, 1005 Xenia Avenue, Yellow
Springs, OH 45387-1695 USA and 2Southwest Foundation for Biomedical Research, Department of Genetics, P. O. Box 28147, San
Antonio, TX 78228-0147 USA
BACKGROUND: Recent evidence has suggested that leptin concentration is associated with gonadal hormone levels,
and that changes in leptin concentration may trigger the onset of reproductive function in children. However, the
concurrent changes in body composition during puberty make the independent associations between leptin and
gonadal hormone concentrations in children dif®cult to resolve.
METHODS: To investigate the nature of associations between leptin levels and pubertal maturation, serum
concentrations of leptin, estradiol, and testosterone and body composition measures were examined in a sample
of 152 healthy pre-pubertal, pubertal, and post-pubertal children.
RESULTS: Leptin concentration was nearly three-fold higher in post-pubertal girls than in pre-pubertal girls, but was
relatively similar in pre- and post-pubertal boys. Signi®cant sex differences in leptin concentration existed in prepubertal, pubertal and post-pubertal children, and these remained signi®cant after controlling for adiposity. After
adjusting for total body fat, fat-free mass and age, testosterone concentration was negatively associated with leptin
levels in pubertal boys, while estradiol concentration was positively associated with leptin level in pubertal girls.
CONCLUSIONS: Girls have higher serum leptin concentration before, during, and after puberty than boys, even after
accounting for the development of greater female adiposity. Although other factors may be involved, sexual
dimorphism in leptin concentrations during puberty appears to be partly due to a stimulatory effect of estradiol on
leptin concentration in females and a suppressive effect of testosterone on leptin concentration in males.
Keywords: Leptin; puberty; testosterone; estradiol; body composition
Introduction
Since the discovery of the ob gene and its gene
product, leptin, in 1994,1 evidence from experimental
studies has shown that leptin functions as a marker of
peripheral body fat stores and that it is involved in the
control of energy balance through the inhibition of
food intake and the stimulation of energy expenditure.2 ± 5 Leptin is secreted by the adipocyte,6 and its
concentration in the blood is highly correlated with
the amount of adipose tissue in the body. Correlations
between measures of body fat and leptin concentration
range from 0.7 to 0.9 in both adults and children.7 ± 10
However, leptin levels are higher in females than
males, even after adjusting for greater female fatness,
and this has been shown in both adults and children.7,9,11 ± 14
Although the biological basis for the sex difference
in leptin concentration is unclear at this time, it is
likely that leptin expression is in¯uenced by sex-
speci®c, or sex-in¯uenced, traits such as levels of
gonadal hormones or the relative proportion of fat
and fat-free mass. Studies in male and female
rodents15 ± 18 and in boys19 suggest that leptin may
be a trigger for the initiation of sexual maturation.
Independent effects of testosterone20,21 and estradiol22
on leptin concentration have been reported, while
other studies have suggested that the level of fatfree mass may contribute to the lower leptin concentration of males compared to females.11,23
Therefore, if leptin concentration is related to
hypothalamic-pituitary-gonadal function, then sexual
dimorphism in leptin may emerge during puberty, but
the emergent sex differences in body composition that
occur during that time also must be considered. This
study examined the relationships between leptin,
gonadal hormones (testosterone and estradiol) and
body composition in pre-pubertal, pubertal, and
post-pubertal children.
Subjects and Methods
*Correspondence: Dr. E. W. Demerath, Division of Human
Biology, Department of Community Health, Wright State
University School of Medicine, 1005 Xenia Avenue, Yellow
Springs, OH 45387-1695 USA.
Subjects
We studied 152 healthy white children (79 boys and
73 girls) aged 8 ± 18 y who were enrolled in the Fels
Longitudinal Study24 and who were examined
Leptin concentration during puberty
EW Demerath et al
between March, 1990 and December, 1996. This is a
community-based, longitudinal study of growth and
aging. Data from this study has been used to construct
the National Center for Health Statistics' growth
charts for children 0 to 36 months.25 Most current
participants are descendants of subjects randomly
recruited from the local community during the
1930's, 1940's, and 1950's. The families were and
are randomly ascertained; that is, they are not selected
on the presence or absence in a family member of any
particular condition or disease. To characterize the
nutritional status of the group, mean weight and
stature within each chronological age fell between
the 50th and 75th percentiles of the U.S. population
distribution.26,27
A fasting blood sample was collected from each
participant between 0800 and 0900 hours and aliquoted for different purposes, including a serum
sample archived at 7 80 C until analysis. Body
composition measurements were taken on the same
day following the collection of serum. These data
were collected following protocols and using
informed consent forms approved by the Institutional
Review Board of Wright State University. Legal
consent was provided by parents for participants
under age 18, and written and=or verbal consent was
obtained from each child before testing.
Pubertal development
Pubertal development ratings were determined by
self-assessment of breast development in girls and
genitalia development in boys every six months
between the ages of 10 and 17, inclusive. The purpose
of the self-assessment was explained in private to each
participant who was then given sex-appropriate sets of
line drawings and verbal descriptions matching those
®rst described by Reynolds and Wines28 and later
popularized by Tanner,29 and that are often referred
to, therefore, as Tanner stages. The line drawings
depicting each Tanner stage provide templates of
®ve stages of breast and genitalia development.
After any questions had been answered, the participant undressed completely in a private cubicle with a
full-length mirror to assist self-examination. Given the
longitudinal nature of the study, which necessitates
ongoing participation, and the reluctance of partici-
pants, particularly adolescents, to undergo examination, we rely on self-assessment of Tanner stages.
Self-assessed pubertal development estimation in clinical settings has been estimated to have a concordance
rate with physician assessment of 60 ± 85%.30,31
For children less than 10 years of age, Tanner stage
1 was assigned, while for those greater than 17 years
of age, Tanner stage 5 was assigned. Approximately
95% of boys less than 10 years of age would be
expected to be in Tanner stage 1, according to data
collected on the mean and standard deviation of the
age on reaching genital stage 2.32 Approximately 85%
of girls less than 10 would be expected to be in Tanner
stage 1, according to data collected on the age upon
reaching breast stage 2.33,34 Given the error introduced
by this assumption, particularly in girls, we examined
individual estradiol values in girls classi®ed as Tanner
stage 1. All values were less than 15 pg=ml, the
reference level for pre-pubertal children,35 indicating
that our assumption that girls less than 10 years of age
were ``pre-pubertal'' was reasonable.
These ®ve Tanner stages were reduced to three
pubertal status groups to increase sample sizes and to
simplify the presentation of results. These groups are:
`pre-pubertal', corresponding to Tanner stage 1; `pubertal', corresponding to Tanner stage 2, 3, or 4; and
`post-pubertal', corresponding to Tanner stage 5. The
distribution of chronological ages within each pubertal status group is shown in Table 1.
Body Composition
Four measures of body composition were obtained:
the body mass index (BMI, kg=m2), total body fat
(TBF, kg), subcutaneous fat estimated from the sum
of six skinfolds (Sum of skinfolds, mm), and total fatfree mass (FFM, kg). Stature was measured to 0.1 cm
on a Holtain stadiometer (Seritex, Carlstadt, NJ),
weight was measured to 0.1 kg on an electronic
scale (Seca, Munich, Germany), and skinfolds were
measured to 0.5 mm with Holtain calipers (Holtain
Ltd., Crymych, U.K.). The skinfolds included three
trunk sites (subscapular, suprailiac, and midaxillary)
and three extremity sites (triceps, biceps, and medial
calf) and were measured, along with stature and
weight, using standard anthropometric methods.36
Table 1 Distribution of chronological age within each pubertal status group
8
9
10
11
12
13
Stage
Age (y)
14
15
16
17
18
All
0
8
4
0
2
4
0
2
9
0
0
12
0
0
8
29
42
39
0
6
3
0
3
1
0
3
3
0
0
16
0
0
8
27
42
32
Boys
Pre-pubertal
Pubertal
Post-pubertal
14
0
0
10
0
0
1
3
0
2
11
0
1
9
0
1
7
2
Pre-pubertal
Pubertal
Post-pubertal
13
0
0
9
0
0
3
8
0
1
8
0
0
9
0
1
5
1
Girls
679
Leptin concentration during puberty
EW Demerath et al
680
Measures of TBF and FFM were obtained from a
multicomponent body composition model using
underwater weighing and residual volume to calculate
total body density. Residual volume was estimated
using a helium-dilution method on a computerized
pulmonary function analyzer (Sensormedics, Yorma
Linda, CA). Age-and sex-speci®c equations were used
to calculate the density of the fat-free mass.37 Using
these ®gures and total body density from hydrodensitometry, percentage body fat (%BF) was determined.36,37 Total body fat and FFM were calculated
from body weight and %BF.
Leptin and gonadal hormone assays
Serum concentrations of leptin (ng=ml) were measured using a radioimmunoassay method (Linco
Research, Inc., St. Charles, MO) from the serum
aliquots stored at 7 80 C, using the manufacturer's
protocol. This assay's limit of sensitivity is 0.5 ng=ml,
and the intra-assay and inter-assay coef®cients of
variation were 6% and 7%, respectively.38 These
values are within the range reported by the manufacturer. No signi®cant degradation of the protein under
storage was detected; the correlation between length
of storage (up to six years) and leptin concentration
was not signi®cant (r ˆ 0.04, P ˆ 0.41).
The concentrations of total testosterone and estradiol
were measured in serum stored at 4 C for 1 ± 5 d until
analyzed using the appropriate DPC Coat-A-Count
radioimmunoassay packages (Diagnostic Products Corporation, Los Angeles, CA). Detection limits for estradiol and total testosterone were 8 ± 3600 pg=ml and
0.04 ± 16 ng=ml, respectively. The intra-assay coef®cient of variation was 14% for estradiol and 12% for
total testosterone. The inter-assay coef®cient of variation
was 15% for estradiol and 13% for total testosterone.
There were 12 `pubertal' and 29 `post-pubertal'
girls who reported having reached menarche and who
indicated the number of days since their last menstruation. The distribution of visits that occurred 0 ± 7 d,
8 ± 14 d, 15 ± 21 d, 22 ± 28 d, and more than 28 d since
the ®rst day of the last menstrual period is provided in
Table 2. Visits occurred randomly throughout the
menstrual cycle for these girls, and leptin concentration was not signi®cantly related to the timing of
blood collection during the menstrual cycle.
Statistical analysis
Multiple measures were available for some individuals; all available data were used in the analysis,
with the restriction that an individual was represented
only once within each of the three pubertal status
groups. Data from the most recent visit within each of
three pubertal status groups were chosen, yielding a
total of 110 observations from 79 boys and 101
observations from 73 girls. Tests of statistical signi®cance were performed within, not between, pubertal
status groups to avoid violating the assumption of
independence of observations.
Leptin values were not normally distributed, but
were positively skewed in both sexes; therefore, either
non-parametric tests, or natural logarithm (ln) transformed leptin values were used in statistical analyses,
as indicated in the text.
The signi®cance of sex differences in leptin concentration was tested within each pubertal status
group using the Wilcoxon rank-sum test. Second, a
subset of single values from the most recent visit of
each participant was used in an analysis of variance to
assess the signi®cance of mean differences in leptin,
body composition and hormones across pubertal status
groups. Third, an analysis of covariance was performed on all data within each pubertal status group
to assess sex differences in leptin concentration
adjusting for total body fat. This subset of values
was used for the calculation of Pearson correlation
coef®cients to examine overall associations between
leptin and body composition and gonadal hormone
concentrations, and correlations also were calculated
within each pubertal status group. In pre-pubertal
children, most values for estradiol and testosterone
were less than the detection limit of the assays, and
therefore the relationships between leptin and these
hormones could not be tested in that group. Finally,
stepwise multiple regression was used to determine
the relative contribution of testosterone and estradiol
levels in boys and girls, respectively, to variation in
leptin concentration in pubertal and post-pubertal
children. Variables were allowed to enter the stepwise
model at a signi®cance level of 0.10 or less, but were
allowed to stay in the model only if reaching a
signi®cance level of 0.05. All statistical analyses
were performed using SAS for Windows (version
6.12, SAS Institute, Cary, NC, 1989 ± 1996).
Results
Characteristics of pre-pubertal, pubertal, and postpubertal children
Means and standard deviations are presented in Table
3 for hormone levels and body composition, by sex
Table 2 Distribution of visits to study center over the menstrual cycle
Number of days since first day of last menstrual period
Pubertal Stage
Missing*
0^7
8 ^ 14
15 ^ 21
22 ^ 28
> 28
Pre-pubertal
Pubertal
Post-pubertal
27 (100%)
30 (71%)
2 (6%)
0
2 (5%)
5 (16%)
0
1 (2%)
7 (23%)
0
5 (12%)
6 (19%)
0
2 (5%)
8 (26%)
0
2 (5%)
3 (10%)
*All pre-pubertal girls, and 29 out of 30 pubertal girls with missing data, had missing data because they were pre-menarcheal.
Leptin concentration during puberty
EW Demerath et al
681
and pubertal status group. Both sexes exhibited the
expected increases in BMI, FFM, and gonadal hormones over the course of pubertal development;
however, TBF and sum of skinfolds increased signi®cantly only in girls.
Effects of sex and pubertal stage on leptin concentration
Serum leptin concentration nearly tripled in females
during the course of pubertal development. Mean
leptin concentration ( standard deviation) was
6.7 4.6 ng=ml in pre-pubertal girls, 12.7 8.1 ng=ml in pubertal girls, and 19.6 11.9 ng=ml in
post-pubertal girls (Figure 1) In contrast, leptin concentration changed little in males during this period;
mean levels were 4.7 3.9 ng=ml in pre-pubertal
boys, 4.1 4.1 ng=ml in pubertal boys, and 4.2 5.2 ng=ml in post-pubertal boys. Leptin concentration
was signi®cantly higher in girls than boys (P < 0.01)
in all groups.
An analysis of covariance was used to simultaneously test the effects of sex and TBF on leptin
concentration to determine whether or not the sex
differences in leptin level could be explained by
greater female fatness. When leptin concentration
was adjusted for TBF, girls still had signi®cantly
higher mean leptin concentrations than did boys in
each group (Table 4). In the post-pubertal status
group, there was a four-fold difference in leptin
concentration between the sexes.
Figure 1 Serum leptin concentration in boys (open bars) and
girls (hatched bars). Means and standard deviations are indicated. Asterisks denote signi®cance of sex differences by Wilcoxon two-sample rank-sum test, *P < 0.01; **P < 0.0001.
Table 4 Sex differences in leptin after adjusting for TBF
Pre-pubertal
n ˆ 56
Variables in model
Sex
6.5*
TBF
21.8**
Model F-ratio
11.4**
Boys
Girls
3.7
6.1
Pubertal
n ˆ 84
Post-pubertal
N ˆ 71
F-Ratio
47.5**
76.8**
84.6**
Adjusted means{ (ng/ml)
3.7
9.0
209.5**
123.9**
295.1**
3.3
12.3
*P < 0.01; ** P < 0.0001;
{Mean leptin concentration adjusted for TBF (least squares
means). Tests were performed on In-transformed values.
Adjusted means shown were back-transformed.
Associations between leptin concentration, body
composition, and gonadal hormones
To provide an overview of the relationships between
leptin, body composition and gonadal hormones
during childhood and adolescence, correlations were
calculated for a sub-sample of values (one observation
per subject, 79 boys and 73 girls). In boys, leptin was
positively correlated with BMI (r ˆ 0.61, P < 0.0001),
sum of skinfolds (r ˆ 0.85, P < 0.0001), and TBF
(r ˆ 0.83, P < 0.0001), and was negatively associated
with testosterone (r ˆ 7 0.32, P < 0.05). In girls,
leptin was positively correlated with BMI (r ˆ 0.72,
P < 0.0001), sum of skinfolds (r ˆ 0.73, P < 0.0001),
and TBF (r ˆ 0.79, P < 0.0001). Girls also exhibited
positive relationships between leptin and age
(r ˆ 0.43, P < 0.0001), FFM (r ˆ 0.44, P < 0.001),
and estradiol (r ˆ 0.34, P < 0.05).
Table 3 Descriptive statistics
Boys
Mean s.d.
Age (y)
Weight (kg)
Stature (cm)
BMI (kg/m2)
Sum of Skinfolds (mm)
TBF (kg)
FFM (kg)
Testosterone (ng/ml)**
Estradiol (pg/ml)**
Girls
Mean s.d.
Pre-Pubertal
(n ˆ 29)
Pubertal
(n ˆ 42)
Post-Pubertal
(n ˆ 39)
Pre-Pubertal
(n ˆ 27)
Pubertal
(n ˆ 42)
Post-Pubertal
(n ˆ 32)
9.8 1.5
34.0 7.2
140.3 8.7
17.1 2.1
49 21
8.2 3.7
27.4 5.1
0.1 0.2
12.7 3.8
13.6 1.8
51.6 16.6
162.5 13.1
19.1 3.8
53 32
9.8 6.0
42.6 12.7
2.8 2.2
24.6 21.4
17.7 0.8*
71.7 16.7*
178.9 7.5*
22.4 5.1*
54.5 31
12.2 10.2
59.8 10.0*
5.8 1.9*
27.5 12.4
9.5 1.3
30.9 5.8
136.2 8.7
16.5 2.0
50 20
5.9 2.2
24.4 4.4
0.3 0.6
10.7 2.0
13.8 1.9
50.1 10.8
159.1 8.9
19.6 3.0
64 28
11.7 5.5
38.1 6.9
0.3 0.2
48.4 48.8
17.6 0.9*
60.3 9.9*
165.0 6.6*
22.1 2.8*
74 21*
16.3 7.7*
44.5 6.2*
0.3 0.2
70.3 51.9*
* Indicates a signi®cant difference across pubertal status groups from analysis of variance in a subset of independent measures (one
per subject: Boys N ˆ 79; Girls N ˆ 73), P < 0.05. Tests for testosterone and estradiol were performed on In-transformed values.
**Values for testosterone and estradiol that were below the detection limits of the assay were set to missing. Sample sizes for these
variables were N ˆ 9, N ˆ 34, and N ˆ 34 in pre-pubertal, pubertal, and post-pubertal boys, respectively, and N ˆ 13, N ˆ 32, and N ˆ 30
in pre-pubertal, pubertal, and post-pubertal girls, respectively.
Leptin concentration during puberty
EW Demerath et al
682
To provide more details of these associations with
regard to pubertal development, Table 5 shows the
correlations between leptin concentration, measures of
body composition and gonadal hormone concentrations by sex and pubertal status group. Strong positive
correlations between leptin concentration and measures of adiposity (r ˆ 0.6 to 0.8) were found in both
sexes, and in both pubertal and post-pubertal children,
but these associations tended to be somewhat weaker
among pre-pubertal girls. Negative correlations were
found between leptin concentration and both age
(r ˆ 7 0.37, P < 0.05) and testosterone concentration
(r ˆ 7 0.51, P < 0.01) in post-pubertal boys. Positive
correlations were found between leptin concentration
and FFM (r ˆ 0.37, P < 0.05), testosterone (r ˆ 0.41,
P < 0.01), and estradiol concentration (r ˆ 0.37,
P < 0.05) in pubertal girls.
Relationships between serum leptin concentration
and gonadal hormones also were tested using stepwise
multiple linear regression in pubertal and post-pubertal boys and girls (Table 6). The independent variables in the model included age, TBF, FFM,
testosterone (in boys) and estradiol (in girls). While
TBF accounted for a substantial portion of the variation in leptin concentration in pubertal and postpubertal boys R2 ˆ 0.54 and 0.66, respectively), testosterone also had a small but signi®cant independent
negative effect (partial R2 ˆ 0.10, P < 0.05) on leptin
concentration in pubertal boys. Effects of the other
variables tested in this model (age and FFM) were not
signi®cant in boys.
Similarly, total body fat explained a large portion of
variation in leptin concentration in pubertal and postpubertal girls (R2 ˆ 0.50 and 0.68, respectively).
Additionally, a small but signi®cant positive association was found between leptin and estradiol in pubertal girls (partial R2 ˆ 0.07, P < 0.05).
Discussion
Leptin concentration increases during pubertal development in girls; leptin concentrations were three-fold
higher in post-pubertal girls than pre-pubertal girls. In
contrast, leptin concentrations were similar across
pubertal status groups in boys. These ®ndings corroborate those from most other studies that have examined leptin during puberty;11,39 ± 44 there is one study,
however, that has reported a decline in leptin level
with advancing pubertal stage in obese children of
both sexes.9
In our study, girls had signi®cantly higher leptin
levels than boys in each pubertal status group even
after adjusting for total body fat. The magnitude of the
sex difference in leptin concentration was accentuated
during puberty and post-puberty. Reports are divided
on the presence of sexual dimorphism in leptin in prepubertal children. Several have found that sex differences in leptin do not emerge until puberty.11,12,39,45
Others, however, have reported signi®cant sex differ-
Table 5 Correlations between leptin concentration and body composition measures and gonadal hormone concentrations
Boys
Mean s.d.
Age (years)
BMI (kg/m2)
Sum of Skinfolds (mm)
TBF (kg)
FFM (kg)
Testosterone (ng/ml)***
Estradiol (pg/ml)***
Girls
Mean s.d.
Pre-Pubertal
(n ˆ 29)
Pubertal
(n ˆ 42)
Post-Pubertal
(n ˆ 39)
Pre-Pubertal
(n ˆ 27)
Pubertal
(n ˆ 42)
Post-Pubertal
(n ˆ 32)
0.12
0.9**
0.91**
0.72**
0.34
Ð
Ð
7 0.20
0.60**
0.72**
0.67**
0.03
7 0.27
7 0.16
7 0.37*
0.63**
0.85**
0.86**
7 0.16
7 0.51**
7 0.29
7 0.27
0.44*
0.48*
0.21
0.22
Ð
Ð
0.30
0.69**
0.72**
0.74**
0.37*
0.41*
0.37*
7 0.13
0.71**
0.76**
0.81**
0.09
0.20
7 0.07
*P < 0.05; **P < 0.01.
***See Table 3 for sample sizes for these variables.
Table 6 Stepwise multiple linear regression analysis of variables associated with leptin
Males
Pubertal
n
order*
Variable**
b
SE (b)
P-value
Partial R2
Model R2
28
1
2
1
TBF
Testosterone
TBF
1.10
7 0.13
1.19
0.19
0.05
0.16
< 0.0001
0.02
< 0.0001
0.54
0.10
0.66
0.64
0.66
1
2
1
TBF
Estradiol
TBF
1.24
0.27
1.38
0.22
0.06
0.18
< 0.0001
0.04
< 0.0001
0.50
0.07
0.68
0.57
0.68
Post-pubertal
Females
Pubertal
33
30
Post-Pubertal
27
Variables tested: TBF, FFM, age, estradiol (for girls), and testosterone (for boys).
*The order in which the variables entered the stepwise model.
**Leptin, TBF, testosterone and estradiol values were In-transformed for analysis.
Leptin concentration during puberty
EW Demerath et al
ences in leptin level in pre-pubertal children similar to
those shown here.7,21,41 Further, female newborns
have been found to have higher cord blood concentrations of leptin than male infants,46 suggesting that
sexual dimorphism in leptin level exists not only in
children and adults but throughout life.
In our study, the sexual dimorphism in leptin in
children was explained not only by sex differences in
adiposity, but also by the opposite effects of testosterone and estradiol in pubertal (Tanner stages 2 ± 4)
boys and girls. Without consideration of total fat,
leptin was negatively correlated with testosterone in
post-pubertal boys, and positively correlated with
testosterone and estradiol in pubertal girls. After
adjusting for TBF in a stepwise regression model,
the leptin-testosterone association was apparent only
in pubertal boys, indicating that the unadjusted correlation seen in post-pubertal boys was probably due to
the effect of body fat. In pubertal girls, the strength of
the association between leptin and estradiol was
reduced but not eliminated by the adjustment for
TBF. These results suggest that the level of gonadal
hormones relates to leptin concentration at particular
points in pubertal development but that the association
is modi®ed by the level of body fat. The sexual
dimorphism in leptin concentration among pre-pubertal children requires further investigation, as it
appears to exist after controlling for differences in
body fat and before the pubertal increase in gonadal
hormone concentration begins.
We have recently found (unpublished data) that
average yearly increments in leptin concentration are
positive in 8 ± 11 year old boys and negative in 12 ± 14
and 15 ± 17 year old boys, while average yearly
increments in leptin concentration are positive in 8 ±
10 and 11 ± 13 year old girls, and close to zero in 14 ±
17 year old girls. In boys, increments in leptin concentration were negatively associated with serial
changes in testosterone concentration (r ˆ 7 0.3 to
7 0.6, P < 0.05) and dehydroepiandrosterone concentration (r ˆ 7 0.4 to 7 0.8, P < 0.05). In girls,
leptin increments were positively associated with
serial changes in estradiol concentration (r ˆ 0.5,
P < 0.001), but only among 11 ± 13 year olds. Older
and younger girls did not exhibit this association. These
unpublished results bolster the cross-sectional associations between leptin concentration and gonadal hormone concentrations described in the present analysis.
There is substantial evidence from in vitro and
animal studies to support a functional relationship
between leptin and gonadal hormones. Leptin receptors have been found in the hypothalamus47 and in the
ovary and prostate,48 indicating that leptin may affect
the production of gonadal hormones directly or indirectly, via changes in gonadotropin releasing hormone
from the hypothalamus. In experimental studies, a
dose-dependent decline in leptin expression has been
documented in cultured human adipocytes exposed to
testosterone21 and testosterone substitution has been
found to reduce leptin levels in hypogonadal men.20 In
the former case, signi®cant effects of testosterone on
leptin production were detected at very low concentrations of testosterone (1 ng=ml), similar to those
seen in early puberty in males. Con¯icting evidence
exists pertaining to the role of estrogen in the regulation of leptin production. Leptin has been shown to
raise depressed estradiol levels in ovariectomized
rats.22 However, Saad et al found no difference in
leptin levels between post-menopausal and pre-menopausal females,14 casting doubt on the biological
relationship between leptin and estrogens.
Compared to the large effect of adiposity on leptin
level, the effect of gonadal hormones in the present
study was relatively small and limited to only pubertal
children, and accounted for approximately 7 ± 10% of
leptin variation. Because of the relatively small effect
size of these hormonal factors, and the small sample
sizes of children typically observed in these studies,
there remains some disagreement as to whether or not
testosterone and=or estradiol have biologically significant relationships with leptin once other covariates
have been considered. Wabitsch et al21 found linear
associations between estradiol (r ˆ 0.2) and testosterone (r ˆ 7 0.2), similar to those reported here, in
their sample of obese children. Blum et al39 and
Ambrosius et al49 have reported a strong negative
relationship between leptin and testosterone in boys
(r ˆ 7 0.4), which, as was found in our study,
accounted for approximately 10% of leptin variation,
but neither study reported an effect of estradiol on
leptin concentration in girls. In contrast, a small study
of healthy prepubertal and pubertal children found
no relationship between testosterone and leptin in
boys, and a signi®cant positive association between
estradiol and leptin in girls,44 but when sexes were
combined, a negative association between free
testosterone and leptin was found.
The data shown here indicate that the nature of the
adipocyte response to increases in gonadal hormones
during puberty is sex-speci®c and is strongly affected
by differences in body composition. The interpretation
of study results is confused when boys and girls are
considered together in regression models, as was done
by Roemmich et al.44 In that case it appeared that in
both sexes, leptin was negatively associated with
testosterone and positively associated with estradiol.
However, our results suggest this is not the case;
leptin was positively correlated with testosterone in
pubertal girls and tended to be negatively correlated
with estradiol in boys. It appears that the leptin
response to speci®c circulating gonadal hormones
differs between the sexes. Without the separate analysis of males and females, or consideration of interaction effects between sex and hormonal or other
factors, the precise nature of these associations is
obscured.
The potentially confounding effect of changes in
body composition on the associations between leptin
and gonadal hormones in pubertal children was carefully considered in our study. It has been suggested
683
Leptin concentration during puberty
EW Demerath et al
684
that in addition to the known relation with fat mass,
serum leptin concentration may be in¯uenced by fatfree mass.11,23 However, there was no independent
association between leptin and FFM in either boys or
girls, nor did this relationship appear in a stepwise
regression model when the effect of adiposity was
simultaneously considered.
In a comparison of the relationships between leptin
and various measures of adiposity (BMI, TBF from
hydrodensitometry, and skinfolds), it was found that
TBF and skinfolds measures were more strongly
correlated with leptin than was BMI. For this
reason, TBF was used in the multiple regression
analysis. Most studies analyzing the relationships
between leptin and pubertal development in humans
have used BMI or other weight-dependent measures
of adiposity. However, BMI and weight, which re¯ect
lean body mass as well as fat mass, may not be good
proxies of total adiposity in children. In adults, weight
gain is almost entirely due to increases in adiposity,
whereas in growing children, increases in bone and
muscle mass account for a signi®cant proportion of
weight gain. The use of alternative methods of body
composition assessment, such as the 18O-dilution
method used by Arslanian et al50 and Salbe et al,45
may explain why these authors found no signi®cant
sex differences in leptin concentration in children
after adjusting for body fat. Therefore, especially
when studying children, the particular technique
used to measure body fat stores may affect the
detection of other factors associated with leptin.
Conclusion
Leptin changed relatively little from pre- to postpuberty in males compared to females. The rise in
leptin concentration in girls from pre- to post-puberty
occurred simultaneously with a signi®cant rise in the
level of body fat. Although total body fat was the
primary determinant of leptin concentration, small
independent effects of gonadal hormones were
found. Testosterone was negatively associated with
leptin in pubertal boys, and estradiol was positively
associated with leptin in pubertal girls. These ®ndings
support the notion that while sexual dimorphism in
leptin exists in pre-pubertal children, sex differences
are magni®ed during the process of sexual maturation
due to increases in fat mass in girls and sex-speci®c
responses to circulating gonadal hormones. Future
analyses of long-term serial data in these children
will determine if leptin directly in¯uences the production of, or is directly in¯uenced by, circulating sex
hormones during puberty, or whether their association
is the product of other maturational processes.
Acknowledgements
This study was supported by grants from the National
Institutes of Health (HD12252 and HD31621). The
excellent technical assistance of Merita Mof®tt, Edgar
Rodriguez, and Chandra Greenberg is greatly appreciated.
References
1 Zhang Y, Proenca R, Maffei M, Barone M, Leopold L,
Friedman JM. Positional cloning of the mouse obese gene
and its human homologue. Nature 1994; 372: 425 ± 432.
2 Camp®eld LA, Smith FJ, Guisez Y, Devos R, Burn P.
Recombinant mouse OB protein: evidence for a peripheral
signal linking adiposity and central neural networks. Science
1995; 269: 546 ± 549.
3 Halaas JL, Gajiwala KS, Maffei M, Cohen S, Chait BT,
Rabinowitz D, Lallone RL, Burley SK, Friedman JM.
Weight-reducing effects of the plasma protein encoded by
the obese gene. Science 1995; 269: 543 ± 546.
4 Hwa JJ, Ghibaudi L, Compton D, Fawzi AB, Strader CD.
Intracerebroventricular injection of leptin increases thermogenesis and mobilizes fat metabolism in ob=ob mice. Horm
Metab Res 1996; 29: 659 ± 663.
5 Pellymounter MA, Cullen MJ, Baker MB, Hecht R, Winters
D, Boone T, Collins F. Effects of the obese gene product on
body weight regulation in ob=ob mice. Science 1995; 269:
540 ± 543.
6 Masuzaki H, Ogawa Y, Isse N, Satoh N, Okazaki T, Shigemoto M, Mori K, Tamura N, Hosoda K. Human obese gene
expression: adipocyte-speci®c expression and regional differences in the adipose tissue. Diabetes 1995; 44: 855 ± 858.
7 Caprio S, Tamborlane WV, Silver D, Robinson C, Leibel R,
Mccarthy S, Grozman A, Belous A, Maggs D. Hyperleptinemia: an early sign of juvenile obesity.Relations to body fat
depots and insulin concentrations. Am J Physiol 1996; 271:
626 ± 630.
8 Considine RV, Sinha MK, Heiman ML, Kriauciunas A,
Stephens TW, Nyce MR, Ohannesian JP, Marco CC, McKee
LJ. Serum immunoreactive-leptin concentrations in normalweight and obese humans. New Engl J Med 1996; 334:
292 ± 295.
9 Hassink SG, Sheslow DV, Delancey E, Opentanova I, Considine RV, Caro JF. Serum leptin in children with obesity:
Relationship to gender and development. Pediatrics 1996; 98:
201 ± 203.
10 Maffei M, Halaas J, Ravussin E, Pratley RE, Lee GH, Zhang
Y, Fai Y, Kim S, Lallone R. Leptin levels in human and
rodent: measurement of plasma leptin and Ob RNA in obese
and weight-reduced subjects. Nat Med 1995; 1: 1155 ± 1161.
11 Clayton PE, Gill MS, Hall CM, Tillmann V, Whatmore AJ,
Price DA. Serum leptin through childhood and adolescence.
Clin Endocrinol 1997; 46: 727 ± 733.
12 Lahlou N, Landais P, De Boissieu D, Bougneres P-F. Circulating leptin in normal children and during the dynamic phase
of juvenile obesity. Diabetes 1997; 46: 989 ± 993.
13 Nagy TR, Gower BA, Trowbridge CA, Dezenberg C, Shewchuk RM, Goran MI. Effects of gender, ethnicity, body
composition, and fat distribution on serum leptin concentrations in children. J Clin Endocrinol Metab 1997; 82:
2148 ± 2152.
14 Saad MF, Damani S, Gingerich RL, Riad-Gabriel MG, Khan
A, Boyadjian R, Jinagouda SD, El-Tawil K, Rude RK. Sexual
dimorphism in plasma leptin concentration. J Clin Endocrinol
Metab 1997; 82: 579 ± 584.
15 Ahima RS, Dushay J, Flier SN, Prabakaran D, Flier JS. Leptin
accelerates the onset of puberty in normal female mice. J Clin
Invest 1997; 99: 391 ± 395.
16 Chehab FF, Lim ME, Lu R. Correction of the sterility defect in
homozygous obese female mice by treatment with the human
recombinant leptin. Nat Gen 1996; 12: 318 ± 320.
17 Chehab FF, Mounzih L, Lu R, Lim ME. Early onset of
reproductive function in normal female mice treated with
leptin. Science 1997; 275: 88 ± 90.
Leptin concentration during puberty
EW Demerath et al
18 Cheung CC, Thornton JE, Kuijper JL, Weigle DS, Clifton
DK, Steiner RA. Leptin is a metabolic gateway for the onset
of puberty in the female rat. Endocrinol 1997; 138:
855 ± 858.
19 Mantzoros CS, Flier JS, Rogol AD. A longitudinal assessment
of hormonal and physical alterations during normal puberty in
boys. V. Rising leptin levels may signal the onset of puberty.
J Clin Endocrinol Metab 1997; 82: 1066 ± 1070.
20 JockenhoÈvel F, Blum WF, Vogel E, Englaro P, MuÈllerWieland D, Reinwein D, Rascher W, Krone W. Testosterone
substitution normalizes elevated serum leptin levels in hypogonadal men. J Clin Endocrinol Metab 1997; 82: 2510 ± 2513.
21 Wabitsch M, Blum WF, Muche R, Braun M, Hube F, Rascher
W, Heinze E, Teller W, Hauner H. Contribution of androgens
to the gender difference in leptin production in obese children
and adolescents. J Clin Invest 1997; 100: 808 ± 813.
22 Shimizu H, Shimomura Y, Nakanishi Y, Futawatari T, Ohtani
K, Sato N, Mori M. Estrogen increases in vivo leptin production in rats and human subjects. J Endocrinol 1997; 154:
285 ± 292.
23 Gill MS, Toogood AA, ONeill PA, Adams JE, Thorner MO,
Shalet SM, Clayton PE. Relationship between growth hormone (GH) status, serum leptin and body composition in
healthy and GH de®cient elderly subjects. Clin Endocrinol
1997; 47: 161 ± 167.
24 Roche AF. Growth, maturation and body composition: The
Fels Longitudinal Study 1929 ± 1991. Cambridge University
Press: Cambridge, 1992.
25 Hamill, PVV, Drizd T, Johnson C, Reed R, Roche AF, Moore
W. Physical growth: National Center for Health Statistics
percentiles. Am J Clin Nutr 1979; 32: 607 ± 629.
26 Hamill PVV, Johnston FE, Grams W. Height and weight of
children, United States. Vital and Health Statistics 1970; 11
No. 104: 1 ± 49.
27 Hamill PVV, Johnston FE, Lemeshow S. Height and weight of
youths 12 ± 17 years, United States. Vital and Health Statistics
1973; 11 No. 104: 1 ± 81.
28 Reynolds EL and Wines JV. Individual differences in physical
changes associated with adolescence in girls. Am J Dis Child
1948; 75: 329 ± 50.
29 Tanner JM. Growth at adolescence. Blackwell Scienti®c
Publications: Oxford, 1962.
30 Duke PM, Litt IF, Gross RT. Adolescents self-assessment of
sexual maturation. Pediatrics 1980; 6: 918 ± 920.
31 Neinstein LS. Adolescent self-assessment of sexual maturation. Clin Ped 1982; 21: 482 ± 484.
32 Marshall WA, Tanner JM. Variations in the pattern of pubertal
changes in boys. Arch Dis Child 1970; 45: 13 ± 23.
33 Marshall WA, Tanner JM. Variations in pattern of
pubertal changes in girls. Arch Dis Child 1969; 44: 291 ±
303.
34 Largo R, Prader A. Pubertal development in Swiss girls. Helv
Paediatr Acta 1983; 38: 229 ± 243.
35 Endocrine Sciences. Endocrine Sciences Directory. Endocrine
Sciences: Calabasas Hills, CA, 1996.
36 Lohman TG, Roche AF, Martorell R. Anthropometric standardization reference manual. Human Kinetics Books: Champaign, II, 1988.
37 Guo S, Roche AF, Houtkooper L. Fat-free mass in children
and young adults predicted from bioelectrical impedance and
anthropometric variables. Am J Clin Nutr 1989; 50: 435 ± 443.
38 Comuzzie AG, Hixson JE, Almasy L, Mitchell BD, Mahaney
MC, Dyer TD, Stern MP, MacCluer JW, Blangero J. A major
quantitative trait locus determining serum leptin levels and
fat mass is located on chromosome 2. Nat Gen 1997; 15:
273 ± 276.
39 Blum WF, Englaro P, Hanitsch S, Juul A, Hertel NT, Muller J,
Skakkebaek NE, Heiman ML, Birkett M. Plasma leptin levels
in healthy children and adolescents: dependence on body mass
index, body fat mass, gender, pubertal stage, and testosterone.
J Clin Endocrinol Metab 1997; 82: 2904 ± 2910.
40 Carlsson B, Ankarberg C, Rosberg S, Norjavaara E, Albersson-Wilkland K. Carlsson LMS.Serum leptin concentrations in
relation to pubertal development. Arch Dis Child 1997; 77:
396 ± 400.
41 Ellis KJ, Nicolson M. Leptin levels and body fatness in
children: effects of gender, ethnicity, and sexual development.
Pediatr Res 1997; 42: 484 ± 488.
42 Falorni A, Bini V, Molinari D, Papi F, Celi F, DiSteano G,
Berioli MG, Bacosi ML, Contessa G. Leptin serum levels in
normal weight and obese children and adolescents: relationship with age, sex, pubertal development, body mass index and
insulin. Int J Obes 1997; 21: 881 ± 890.
43 Garcia-Mayor RV, Andrade MA, Rios M, Lage M, Dieguez C,
Casanueva FF. Serum leptin levels in normal children: relationship to age, gender, body mass index, pituitary gonadal
hormones, and pubertal stage. J Clin Endocrinol Metab 1997;
82: 2849 ± 2855.
44 Roemmich JN, Clark PA, Berr SS, Mai V, Mantzoros CS,
Flier JS, Weltman A, Rogol AD. Gender differences in leptin
levels during puberty are related to the subcutaneous fat depot
and sex steroids. Am J Physiol 1998; 275: 543 ± 551.
45 Salbe AF, Nicolson M, Ravussin E. Total energy expenditure
and the level of physical activity correlate with plasma leptin
concentrations in ®ve-year-old children. J Clin Invest 1997;
99: 592 ± 595.
46 Hassink SG, de Lancey E, Sheslow DV, Smith-Kirwin SM,
O'Connor DM, Considine RV, Opentanova l, Dostai K, Spear
ML, Leef K, Ash M, Spitzer AR, Funanage VL. Placental
leptin: An important new growth factor in intrauterine and
neonatal development? Pediatrics 1997; 100: 11 ± 16.
47 Schwartz MW, Seeley RJ, Camp®eld LA, Burn P, Baskin DG.
Identi®cation of targets of leptin action in rat hypothalamus.
J Clin Invest 1996; 98: 1101 ± 1106.
48 Ciof® JA, A. W.S, Zupancic TJ, Smith-Gbur J, Mikhail A,
Platika D, Snodgrass HR. Novel B219=OB receptor isoforms:
possible role of leptin in hematopoiesis and reproduction. Nat
Med 1996; 2: 585 ± 588.
49 Ambrosius WT, Compton JA, Bowsher RR, Pratt JH. Relation
of race, age, and sex hormone differences to serum leptin
concentrations in children and adolescents. Horm Res 1998;
49: 240 ± 246.
50 Arslanian S, Suprasongsin C, Kalhan SC, Drash AL, Brna R,
Janosky JE. Plasma leptin in children: Relationship to puberty,
gender, body composition, insulin sensitivity, and energy
expenditure. Metabolism 1998; 47: 309 ± 312.
685