| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Göteborg Pediatric Growth Research Center, Department of Pediatrics, University of Göteborg, S-416 85 Göteborg, Sweden
Address all correspondence and requests for reprints to: Dr. Ensio Norjavaara, Sahlgrenska University Hospital/Östra, University of Göteborg, Department of Pediatrics, Göteborg Pediatric Growth Research Center, S-416 85 Göteborg, Sweden. E-mail: ensio.norjavaara{at}pediat.gu.se
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The fact that many of the endocrine changes that occur during puberty also occur, but to a greater degree, in PCOS, has led to the hypothesis that PCOS may originate from abnormal pubertal development (1, 2). Indeed, a period of multicystic ovaries occurs in early pubertal development (3, 4), with an increased LH/FSH ratio, LH hyperpulsatility (5, 6), and increased insulin secretion (7, 8). Of special interest in this respect are the studies by Apter and Vihko (9) and Ibanez et al. (10). Apter and Vihko found that girls with the highest androgen levels during adolescence had the lowest fertility rates in the third decade of life, and Ibanez and co-workers reported that premature adrenarche is a risk factor for ovarian hyperandrogenism. These two studies indicate that androgen levels during childhood/adolescence may influence ovarian function in later life.
These new concepts, that PCOS may originate from abnormal pubertal development and that androgen production during childhood/adolescence may influence ovarian function in later life, call for more detailed analysis of androgen dynamics during puberty in girls. Androgen concentrations in girls during puberty have been determined in both cross-sectional (11, 12, 13, 14) and longitudinal studies (15), but no report exists on detailed 24-h androgen profiles during puberty. In the present study of healthy girls of different heights at different pubertal stages, an attempt was made to describe the spontaneous secretory pattern of testosterone in relation to the spontaneous secretory pattern of estradiol and serum levels of dehydroepiandrosterone sulfate (DHEAS) and to relate these to adrenarche, gonadarche, and menarche.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study group consisted of 56 healthy Swedish girls. The girls
were investigated as voluntary healthy controls or for short or tall
stature. Four of the girls were followed longitudinally throughout
puberty, 11 participated 2–3 times during their pubertal development,
whereas 41 of the girls participated only once during the study. All
children were healthy and well nourished, and had normal thyroid,
liver, and kidney functions. Coeliac disease was excluded. In girls
shorter than -2 SD score, GH deficiency was
excluded with an arginine-insulin tolerance test, and Turner’s
syndrome was excluded by karyotype analysis. Puberty was assessed
according to Tanner and Whitehouse (16) for breast and pubic hair
development. Breast development was given priority when assessing
gonadarche. Further details of the girls are given in Table 1
, which shows their Tanner
classification for breast development and the time interval from
menarche. Table 1 also gives heights and weights converted to
SD score, using the Swedish growth reference values for
healthy children (17) and weight/height SD score calculated
according to Albertsson-Wikland et al. (18).
|
Samples for measurement of steroids were taken from 28 girls when prepubertal (breast stage 1, pubic hair development 1), from 14 girls in early puberty (breast stage 2, all premenarcheal), from 31 girls in midpuberty [12 at breast stage 3 (8 premenarcheal and 4 postmenarcheal), 17 at breast stage 4 (6 premenarcheal and 11 postmenarcheal), and 2 at breast stage 5 (both postmenarcheal)], and from 11 girls in late puberty (4 at breast stage 4, and 7 at breast stage 5; all postmenarcheal).
Informed consent was obtained from each girl and her parents. The protocol was approved by the ethical committee of the Medical Faculty, University of Göteborg.
Study protocol
The study was conducted at the Children’s Hospital in Goteborg, Sweden. The children stayed in the hospital for at least 2 days while each profile was taken. A heparinized needle was inserted during the first evening or morning. Serum samples (2 mL) for steroid measurements were taken at 1000, 1400, 1800, 2200, 0200, 0400, and 0600 h. One to five 24-h serum profiles were determined for each girl for steroid measurements during the investigation time, and a total of 84 serum profiles was obtained. Parallel with the serum profiles, single samples were also taken between 1000–1400 h for measurement of sex hormone-binding globulin (SHBG) and DHEAS. DHEAS was chosen because it fluctuates less during the day than dehydroepiandrosterone (DHEA) (20).
The girls were divided into groups according to their pubertal development (breast, time after menarche, and pubic hair). The first samples were chosen for the girls who had had samples taken more than once at any pubertal stage.
Measurement of testosterone
Serum testosterone concentrations were determined in duplicate by RIA using direct coated tube technology (Spectria Testosterone, Orion Diagnostics, Espoo, Finland). In addition to the kit standards, three standards of lower concentrations (0.025, 0.05, and 0.15 nmol/L) were prepared by dilution with zero standard. Twice the serum volume (50 µL) was used; otherwise, the RIA was conducted according to the manufacturer’s instructions.
Detection limit
The detection limit of a RIA is usually defined as the apparent concentration 2 or 3 SD below the counts at maximum binding. The highest sensitivity was obtained with twice the serum volume as that recommended in the assay, which resulted in half the concentration compared with the original assay. Double the serum volume was therefore used for the RIA. Twenty zero standards were measured in a single assay. The detection limit was 0.03 nmol/L for 3 SD below the counts at maximum binding.
As a check, two patient samples around the detection limit were measured three or four times in different assays. The sample with a median value of 0.029 nmol/L was measured four times and had an interassay coefficient of variation (CV) of 7.6%. The sample with a median value of 0.051 nmol/L was measured three times and had a CV of 1.9%. The detection limit of the RIA was therefore considered to be 0.03 nmol/L.
Intra- and interassay CVs
The intraassay CV for the RIA was 10.6% for 0.21 nmol/L, 6.6% for 0.42 nmol/L, 6.6% for 0.90 nmol/L, 5.7% for 4.5 nmol/L, and 4.0% for 16.6 nmol/L, calculated from 20 replicates. The interassay CV for the RIA was 15.5% for 0.23 nmol/L, 9.9% for 0.88 nmol/L, 5.0% for 5.0 nmol/L, and 7.7% for 16.2 nmol/L, calculated from 20 assays.
Linearity
Five different serum samples (testosterone concentrations, 0.2–2.5 nmol/L) were serially diluted seven times in steps of two. For each sample, linearity was obtained down to the detection limit.
We also mixed two samples with high levels of testosterone (4.4 and 18.8 nmol/L) with five samples with low levels of testosterone (0.03–0.37 nmol/L) and obtained a recovery of 101.7% (range, 87.7–111.5%).
Free testosterone
Free testosterone was calculated as: free testosterone (pmol/L) = testosterone determined by RIA (nmol/L)/(K x SHBG (nmol/L) + 1) x 1000, according to Ekins (21), where K is the equilibrium constant for testosterone binding to SHBG (1.6 x 109 L/mol).
Measurement of DHEAS
Serum concentrations of DHEAS were determined in duplicate by RIA using coated tube technology (Coat-A-Count DHEA-SO4; Diagnostic Products Corp., Los Angeles, CA). The intraassay CV for the RIA was 10.9% for values below 1.0 µmol/L and below 7.5% for values over 1.0 µmol/L. The interassay CV was 21% for 0.7 µmol/L, 14% for 4.4 µmol/L, and 11% for 13.6 µmol/L.
Measurement of 17ß-estradiol
Serum 17ß-estradiol concentrations were determined in duplicate by a modified RIA using coated tube technology (Spectria, Orion Diagnostics) after diethyl ether extraction, as previously described (22). The detection limit for the RIA is 7.8 pmol/L. The intraassay CV for extracted serum was below 16% for concentrations between 8–35 pmol/L and below 10% for concentrations above 35 pmol/L. In unextracted serum, the intraassay CV was below 7% for concentrations between 50–200 pmol/L and below 3% for concentrations above 200 pmol/L. The interassay CV for extracted serum was 27% for concentrations between 8–15 pmol/L and 17% for concentrations between 15–30 pmol/L. The interassay CV in unextracted serum was below 11%.
Free 17ß-estradiol
Free 17ß-estradiol was calculated as: free 17ß-estradiol (pmol/L) = 17ß-estradiol determined by RIA (pmol/L)/(K x SHBG (nmol/L) + 1), according to Ekins (21), with an equilibrium constant (K) for binding to SHBG of 0.68 x 109 L/mol.
Measurement of SHBG
Serum concentrations of SHBG were determined by an immunoradiometric assay from Orion Diagnostics. The analysis was performed at the Department of Clinical Chemistry at Sahlgrenska University Hospital (Göteborg, Sweden; accredited laboratory 1240 according to European norm 45001). The interassay CV is less than 6% over a concentration range between 38–70 nmol/L.
Statistical procedures
Values are given as medians together with the ranges or 95% confidence intervals. P < 0.05 was considered significant.
The serum levels of testosterone are presented as individual curves with 95% confidence intervals for the median (figures) or as the median with 95% confidence intervals for the median (tables) for each pubertal stage. Nonparametric statistical methods were used for analyses: the Wilcoxon signed rank sum test for analysis of diurnal variations and the Mann-Whitney test for comparison of different pubertal stages (23). Values of testosterone below the detection limit were considered to be 0.03 nmol/L. Linear regression analyses are given with 95% confidence limits. As the serum 17ß-estradiol levels were not normally distributed, linear regression analyses were performed on log-transformed data.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Prepuberty (Tanner breast stage 1, pubic hair stage 1). Serum
testosterone concentrations in 28 prepubertal girls are shown in Table 2 and Fig. 1. One of the 28 girls had a testosterone
concentration under the detection limit in all samples taken during the
24 h. In 10 of the 28 girls, serum testosterone levels were under
the detection limit at certain times (in 6 girls at 2200 h, in 2
girls at 0200 h, and in 2 girls at both 2200 and 0200 h), but
above the detection limit at least once during the early morning (0600
or 1000 h). The Wilcoxon signed rank sum test indicated a diurnal
rhythm, with testosterone levels at 0600 and 1000 h significantly
higher than those at 1400–0400 h (P < 0.02–0.001)
and testosterone levels significantly lower at 2200 h compared
with those at 0200–1800 h (P < 0.01–0.001).
|
|
). A significant correlation was also found between chronological age and the mean serum testosterone concentrations over 24 h in prepubertal girls (r = 0.45; P < 0.02).
As there was a wide variation in height in the prepubertal girls (-3.0
to +3.7 SD score), a linear regression analysis was
performed on height SD score and weight/height
SD score vs. serum testosterone (Fig. 2), 17ß-estradiol, and DHEAS
concentrations. No correlation between height SD score or
weight/height SD score and serum sex steroid levels was
found.
|
, and statistical analyses
are given in Table 2. All of the girls had detectable serum levels of
testosterone. There was a diurnal rhythm, with testosterone levels at
0600 and 1000 h significantly higher than those at 1800–0400 h
(P < 0.05–0.01) and testosterone concentrations
significantly lower at 2200 and 0200 h compared with those at
0400–1800 h (P < 0.05–0.01).
|
).
Midpuberty (Tanner breast stage 3–4, pubic hair stage 3–5).
The 24-h profiles of serum testosterone in girls in midpuberty, but
still before menarche, are shown in Fig. 3, and statistical analyses
are given in Table 2. There was a diurnal rhythm, with testosterone
levels at 0600 h significantly higher than those at 1000–0400 h
(P < 0.05–0.01) and testosterone levels significantly
lower at 2200 h compared with those at 0200–1800 h
(P < 0.05–0.01).
Postmenarche. The diurnal rhythm was also present after
menarche. The 24-h profiles of serum testosterone in girls in
midpuberty, but up to 1 yr after menarche (Tanner breast stage 3–5),
are shown in Fig. 3, and statistical analyses are given in Table 2.
Testosterone levels at 0600 and 1000 h were significantly higher
than those at 2200–0400 h (P < 0.01–0.001), and
testosterone levels were significantly lower at 2200 and 0200 h
compared with those at 0400–1800 h (P < 0.01–0.001).
The 24-h profiles of serum testosterone in girls in late puberty (1–2
yr after menarche) are shown in Fig. 3, and statistical analyses are
given in Table 2. Testosterone levels at 0600 and 1000 h were
significantly higher than those at 2200–0200 h (P <
0.01), and testosterone levels were significantly lower at 2200 and
0200 h compared with those at 0400–1800 h (P <
0.05–0.01).
Differences in the serum testosterone concentrations between different pubertal stages were also analyzed. Serum testosterone concentrations in prepubertal girls were significantly lower than those in pubertal girls at all sampling times (P < 0.05–0.001). Serum testosterone concentrations in girls in early puberty were significantly lower than those in girls in mid- or late puberty (P < 0.05–0.001) at all sampling times except at 1000 h for premenarcheal girls in midpuberty and at 0200 h for postmenarcheal girls in midpuberty. No statistical differences were found in serum testosterone concentrations between pre- and postmenarcheal girls in midpuberty or between midpuberty and late puberty (1–2 yr postmenarche).
Calculated free serum testosterone in relation to gonadarche
Calculated free serum testosterone concentrations in the 24-h
profiles from prepubertal and pubertal girls are given in Table 3. Statistical analysis indicated a
diurnal rhythm before puberty and during pubertal development, with
calculated free serum testosterone levels at 0600 and 1000 h being
significantly higher than those at 2200 and 0200 h
(P < 0.02–0.001, respectively).
|
No significant differences were found in free serum testosterone concentrations between pre- and postmenarcheal girls in midpuberty or between girls in midpuberty and those in late puberty (1–2 yr postmenarche).
Puberty in relation to adrenarche
The 24-h profiles of serum testosterone in girls in different
stages of puberty, classified according to pubic hair development, are
shown in Table 4.
|
Linear regression analyses were performed between DHEAS and chronological age and between DHEAS and bone age. A significant correlation between DHEAS and bone age was found in prepubertal girls (r = 0.49; P < 0.01), whereas no correlation was found between DHEAS and bone age in pubertal girls. No correlation was found between DHEAS and chronological age in prepubertal or pubertal girls.
Diurnal variation in serum testosterone concentrations, calculated free serum testosterone, and serum 17ß-estradiol in relation to menarche
The pubertal girls were reclassified according to menarche.
Relationships between serum testosterone, calculated free testosterone,
and 17ß-estradiol concentrations vs. time to menarche are
shown in Fig. 4. The girls who were
followed longitudinally are plotted with connected lines in
the left panel.
|
. The 24-h
profiles of serum 17ß-estradiol, according to the girls’ pubertal
development, have been published previously (22). There was a marked
diurnal rhythm, with high levels in the morning (1000 h) and low levels
in the evening (2200–0200 h) before menarche, but this rhythm was lost
1 yr after menarche. For the linear regression analyses, serum levels
of testosterone at 0600 and 2200 h were chosen, as these sampling
times represent the highest and lowest levels (present study), and
serum levels of 17ß-estradiol were analyzed at 1000 and 2200 h
for the same reason (22). A significant correlation was found between serum testosterone and menarche (r = 0.56 and 0.48 for 2200 and 0600 h, respectively). This analysis highlights the diurnal pattern for the period studied (up to 2 yr postmenarche), with the regression line for 0600 h parallel to the regression line for 2200 h, but with higher serum levels. A significant correlation was also found between calculated free testosterone and menarche (r = 0.55 and 0.49 for 2200 and 0600 h, respectively).
A significant correlation was found between log serum 17ß-estradiol and menarche (r = 0.71 and 0.47 for 2200 and 1000 h, respectively). The diurnal pattern for 17ß-estradiol diminished after menarche and was lost between 1–2 yr postmenarche, as indicated by the regression line for 1000 h being well above the line for 2200 h before menarche, but converging around 2 yr postmenarche.
Diurnal variation in serum testosterone and calculated free serum testosterone concentrations in relation to serum 17ß-estradiol
Linear regression analyses between serum testosterone
concentrations and calculated free serum testosterone in relation to
serum 17ß-estradiol concentrations and calculated free serum
17ß-estradiol are shown in Fig. 5. A
significant correlation was found between serum testosterone and log
serum 17ß-estradiol for all pubertal stages (r = 0.40–0.76;
P < 0.001), with the exception of calculated free
serum testosterone and log free serum 17ß-estradiol in late puberty.
As most of the serum 17ß-estradiol levels were under the detection
limit in prepubertal girls, a regression analysis of mean
17ß-estradiol vs. testosterone is not shown. However,
there does appear to be a correlation between 17ß-estradiol and
testosterone in prepuberty. In performing a linear regression analyses
where the values under the detection limit were set at the detection
limit or the values under the detection limit were excluded, a
significant correlation (r = 0.50; P < 0.001 and
r = 0.44; P < 0.05, respectively) was obtained in
both cases.
|
). The
ratio was significantly higher in prepuberty and in early puberty than
in the later stages of puberty (P < 0.05–0.001).
Furthermore, the ratio was significantly higher in premenarcheal girls
in midpuberty than in postmenarcheal girls in midpuberty at 2200 and
0600 h (P < 0.05).
|
Levels of serum DHEAS in girls at different pubertal stages are
given in Table 1. The results of linear regression analyses of serum
testosterone levels and serum DHEAS are shown in Fig. 6. A significant correlation was found
between serum testosterone and serum DHEAS for all the girls (r =
0.56 and 0.58 for 2200 and 0600 h, respectively; P
< 0.001). However, when the different pubertal stages were analyzed
separately, a significant correlation was found only in prepubertal
girls (r = 0.67 and 0.89 for 2200 and 0600 h, respectively;
P < 0.001).
|
As peripheral tissue is able to interconvert androgens and also has some aromatas capacity, linear regression analyses were performed on weight/height vs. DHEAS, testosterone, and estradiol. In prepubertal girls, a significant correlation was found between weight/height and mean serum testosterone over 24 h (r = 0.45; P < 0.05). No correlation was found between weight/height and DHEAS. As most of the serum 17ß-estradiol levels were under the detection limit, a regression analysis of mean 17ß-estradiol vs. weight/height was not possible to perform. In pubertal girls, a significant correlation was found between mean serum 17ß-estradiol and weight/height (r = 0.52; P < 0.001), a week correlation was found between serum DHEAS and weight/height (r = 0.30; P < 0.05), and no correlation was found between weight/height and mean serum testosterone.
To further illustrate changes in sex steroids over time, linear regression analyses were performed between the studied sex steroids and bone age. When including all the girls, significant correlations (P < 0.001) were found between bone age (TW II) and mean serum testosterone over 24 h (r = 0.71), mean serum 17ß-estradiol (r = 0.74) and serum DHEAS (r = 0.47). However, when excluding the prepubertal girls, significant correlations remained between bone age and testosterone (r = 0.38; P < 0.01) and 17ß-estradiol (r = 0.59; P < 0.001), but no correlation was found between bone age and DHEAS.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In adult woman, androgens (i.e. DHEA, DHEAS, androstenedione, and testosterone) are produced in the adrenals and ovaries and by interconversion in other tissues. The interconversion can take place in peripheral tissue and liver, where androstenedione and testosterone are interconvertible and in equilibrium with each other, whereas the aromatization of these steroids to estrone and 17ß-estradiol is irreversible (24). Circulating levels of testosterone are thus the result of ovarian secretion of testosterone together with the peripheral interconversion of androstenedione to testosterone. In adult females, equal amounts of androstenedione are secreted from the adrenals and ovaries (24, 25, 26). The source of testosterone in girls during puberty is complicated due to marked changes in the steroidogenic capacity occurring during adrenarche and gonadarche and to changes in body composition (increased amounts of fat and muscle and less water), probable changes in enzymes involved in the metabolism of steroid hormones, and decreased levels of SHBG.
In the present study there was a good correlation between serum testosterone and DHEAS in prepubertal girls, consistent with the adrenals being the source of androgens before gonadarche (these androgens are converted to testosterone in peripheral tissues). It is well known that adrenarche starts in girls at 6–8 yr of age, with an increased synthesis of adrenal androgens (27, 28, 29, 30) well before gonadarche. An indication that adrenarche starts before gonadarche was also seen in the present study. First, DHEAS correlated with bone age in prepubertal girls; secondly, prepubertal girls with a bone age less than 6.5 yr had serum levels of testosterone below the detection limit at midnight, whereas all prepubertal girls with a bone age over 9 yr had detectable levels of testosterone. Chronological age was not a specific indication of adrenarche in this study due to the fact that the prepubertal girls showed a wide variance in maturation (-3.0 to +3.7 SD score in height). At gonadarche, the correlation between testosterone and DHEAS was lost, whereas the correlation between testosterone and 17ß-estradiol remained until 1 yr after menarche, consistent with the ovary being the major producer of testosterone during gonadarche. However, androstenedione of adrenal or ovarian origin is converted to testosterone in the ovary, and it is therefore not possible to separate these two sources completely.
One possible reason for the lack of correlation between serum levels of DHEAS and testosterone during gonadarche despite increased synthesis of androgens by both the adrenals and ovaries is the complicated nature of their synthesis of these steroids. DHEA is produced in both adrenals and ovaries. Approximately 90% of the DHEA and 99% of the DHEAS found in the circulation are secreted from the adrenals (26). Testosterone is synthesized in two ways: from interconversion of DHEA and androstenedione in peripheral tissue and by conversion of DHEA and androstenedione to testosterone in the ovary.
The aromatization of testosterone to 17ß-estradiol is an irreversible process, which explains the correlation between testosterone and 17ß-estradiol during gonadarche. In late puberty (1–2 yr postmenarche), no correlation was observed between testosterone and either DHEAS or calculated free 17ß-estradiol, which is a reflection of the hormonal milieu during this period, with some girls having more or less regular menstrual cycles in combination with circadian androgen secretion from the adrenals. In comparison, androgen secretion changes throughout the menstrual cycle in adult women, with a diurnal pattern of testosterone secretion in the early follicular phase (31, 32), but not in the periovulatory period (31, 33, 34) or in the luteal phase (31), and with the highest levels in the periovulatory phase (31, 35). In the present study, it was not determined whether the girls had irregular or regular menstrual cycles or what the phase of the cycle was when the samples were taken. It is therefore not surprising that no correlation was found among DHEAS, 17ß-estradiol, and testosterone in late puberty.
The present study also highlights the peripheral (extraglandular) interconversion of androgens to testosterone and aromatization to estrogens. A correlation between estradiol and weight/height in pubertal girls was found. One interpretation of this could be that adipose tissue has a role in the synthesis of estrogen. However, the situation during puberty is complicated, as estrogen also induces the development of secondary sexual characteristics, which involves increasing amounts of adipose tissue (36, 37). Adipose tissue also has limited capacity to interconvert weak androgens to testosterone. In the present study, a correlation between testosterone and weight/height was found only in prepubertal girls, possibly indicating that adipose tissue does not have a role in testosterone synthesis in pubertal children.
The results of the present study demonstrate a diurnal rhythm of testosterone secretion during puberty in girls, which is already present before puberty. This diurnal rhythm is most obvious before puberty and in early puberty. The rhythm before puberty is similar to that of adrenal steroids and probably reflects an adrenal source for testosterone. In pubertal girls, there is a diurnal rhythm of testosterone, which is caused by a mixture of circadian androgen secretion and possibly also diurnal ovarian androgen secretion. This is based on the observation reported in the literature that the circadian rhythm of ACTH remains during puberty (38) in combination with the observation of a similar diurnal rhythm of LH during puberty (5, 6, 39, 40).
In the present study a close correlation between testosterone and 17ß-estradiol (serum levels or calculated free steroid) was found during puberty and for up to 1 yr postmenarche. The 17ß-estradiol determined in the present study was of ovarian origin, as the adrenals have a very low aromatization capacity and produce only minute amounts of estrone, whereas the aromatizing capacity of the ovaries is considerably greater than that of adipose tissue and muscle (41). However, the adrenals are still, at least partly, involved in this process, as androstenedione of adrenal origin can be aromatized peripherally or in the ovary.
In the present study girls in late puberty still had a diurnal rhythm of testosterone secretion. This is not surprising considering that the girls’ menstrual cycles were probably in different phases, and that some of the girls probably had anovulatory cycles. It has been reported that more than half of the cycles are anovulatory during the first 2 yr after menarche (42, 43), and that girls and women with anovulatory menstrual cycles also have a diurnal rhythm of testosterone (2, 32), probably due to the cycles remaining in the follicular phase. Apter et al. (2) also found a diurnal rhythm of testosterone in girls in late puberty.
The serum levels of testosterone in the present study are consistent with levels reported previously from girls before and during puberty (2, 6, 12, 13, 14, 15, 42, 43). An interesting observation in the present study is that those prepubertal girls who had serum levels of testosterone below the detection limit had the lowest bone age. When different pubertal stages were analyzed, significant differences in testosterone levels were found only between prepuberty or early puberty and midpuberty or late puberty, but not between stages in midpuberty and late puberty. Serum testosterone is thus a marker of the start of gonadarche, but may not indicate the tempo of pubertal progression. However, when serum testosterone levels were analyzed with regard to the time to or after menarche, a good correlation was found for both the lowest level and the highest level of serum testosterone over 24 h. A significant correlation was also found during puberty between testosterone and bone age. Obviously, girls in mid-/late puberty increase their serum levels of testosterone as they mature, which has also been reported in other studies (43).
The present results indicate a relative transient hyperandrogenism and an increased ratio between testosterone and 17ß-estradiol (serum levels or calculated free steroid) before and during early puberty compared with those in mid- and late puberty. Ratios between steroids cannot be related directly to androgenism or estrogenism. It is clear, however, that gonadarche in it earliest phase starts in an androgen-dominated state, but during pubertal development the hormonal changes are markedly estrogenic compared with the relatively minor changes in androgens. It is an interesting hypothesis that PCOS may develop from abnormal pubertal development, and a critical point in pubertal development could be in the transition stage from the early pubertal androgen-dominated state to the estrogenic state later in puberty. This is speculation and was not proven by the results of the present study, but it is an interesting observation that gonadarche starts in an androgenic state.
In conclusion, the present study has demonstrated a diurnal rhythm of testosterone that is present both before and during puberty. This is presumably a result of the changing patterns of GnRH and gonadotropin secretion during puberty in combination with the circadian secretion of ACTH. The pattern has its nadir in the late evening or just after midnight, with the highest levels in the morning (0600–1000 h). Furthermore, the study shows a relative hyperandrogenicity in early puberty, with high levels of androgens in relation to estrogens. The prepubertal androgens originate from the adrenals. From the onset of puberty, there is a good correlation between testosterone and 17ß-estradiol, but no correlation between testosterone and DHEAS or weight/height, consistent with the ovaries being the major source of testosterone.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received April 28, 1998.
Revised December 2, 1998.
Accepted December 10, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-dihydrotestosterone
during female puberty. J Clin Endocrinol Metab. 45:1039–1048.[Medline]
This article has been cited by other articles:
 |
|
 |
|
  C. R. McCartney, K. A. Prendergast, S. K. Blank, K. D. Helm, S. Chhabra, and J. C. Marshall Maturation of Luteinizing Hormone (Gonadotropin-Releasing Hormone) Secretion across Puberty: Evidence for Altered Regulation in Obese Peripubertal Girls J. Clin. Endocrinol. Metab., January 1, 2009; 94(1): 56 - 66. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Allvin, C. Ankarberg-Lindgren, H. Fors, and J. Dahlgren Elevated Serum Levels of Estradiol, Dihydrotestosterone, and Inhibin B in Adult Males Born Small for Gestational Age J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1464 - 1469. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Ankarberg-Lindgren and E. Norjavaara A purification step prior to commercial sensitive immunoassay is necessary to achieve clinical usefulness when quantifying serum 17 -estradiol in prepubertal children Eur. J. Endocrinol., January 1, 2008; 158(1): 117 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. McCartney, S. K. Blank, K. A. Prendergast, S. Chhabra, C. A. Eagleson, K. D. Helm, R. Yoo, R. J. Chang, C. M. Foster, S. Caprio, et al. Obesity and Sex Steroid Changes across Puberty: Evidence for Marked Hyperandrogenemia in Pre- and Early Pubertal Obese Girls J. Clin. Endocrinol. Metab., February 1, 2007; 92(2): 430 - 436. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hero, C. Ankarberg-Lindgren, M.-R. Taskinen, and L. Dunkel Blockade of oestrogen biosynthesis in peripubertal boys: effects on lipid metabolism, insulin sensitivity, and body composition Eur. J. Endocrinol., September 1, 2006; 155(3): 453 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Fahrenkrug, B. Georg, J. Hannibal, P. Hindersson, and S. Gras Diurnal Rhythmicity of the Clock Genes Per1 and Per2 in the Rat Ovary Endocrinology, August 1, 2006; 147(8): 3769 - 3776. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.K. Blank, C.R. McCartney, and J.C. Marshall The origins and sequelae of abnormal neuroendocrine function in polycystic ovary syndrome Hum. Reprod. Update, July 1, 2006; 12(4): 351 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Aksglaede, A. Juul, H. Leffers, N. E. Skakkebaek, and A.-M. Andersson The sensitivity of the child to sex steroids: possible impact of exogenous estrogens Hum. Reprod. Update, July 1, 2006; 12(4): 341 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hero, E. Norjavaara, and L. Dunkel Inhibition of Estrogen Biosynthesis with a Potent Aromatase Inhibitor Increases Predicted Adult Height in Boys with Idiopathic Short Stature: A Randomized Controlled Trial J. Clin. Endocrinol. Metab., December 1, 2005; 90(12): 6396 - 6402. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Xita, A. Tsatsoulis, I. Stavrou, and I. Georgiou Association of SHBG gene polymorphism with menarche Mol. Hum. Reprod., June 1, 2005; 11(6): 459 - 462. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. van Anders and E. Hampson Waist-to-Hip Ratio Is Positively Associated With Bioavailable Testosterone but Negatively Associated With Sexual Desire in Healthy Premenopausal Women Psychosom Med, March 1, 2005; 67(2): 246 - 250. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Pardee, J. Reinking, and H. Krause Nuclear Hormone Receptors, Metabolism, and Aging: What Goes Around Comes Around Sci. Aging Knowl. Environ., November 24, 2004; 2004(47): re8 - re8. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R. McCartney, A. B. Bellows, M. B. Gingrich, Y. Hu, W. S. Evans, J. C. Marshall, and J. D. Veldhuis Exaggerated 17-hydroxyprogesterone response to intravenous infusions of recombinant human LH in women with polycystic ovary syndrome Am J Physiol Endocrinol Metab, June 1, 2004; 286(6): E902 - E908. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
