| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Research Institute for Endocrinology, Reproduction and Metabolism, Division of Reproductive Endocrinology and Fertility (M.H.A.v.H., M.B.H.K., J.S.) and Department of Epidemiology and Medical Statistics (F.J.V.), Vrije Universiteit Medical Center, 1081 HV Amsterdam; Netherlands Organization for Applied Scientific Research (TNO), Prevention and Health, Child Health Division (R.A.H.), 2301 CE Leiden; and Department of Youth Health Care of the Public Health Care Service Amstelland-de Meerlanden (C.K.), 1185 JC Amstelveen, The Netherlands
Address correspondence and requests for reprints to: M. H. A. van Hooff, Vrije Universiteit Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
Body Mass Index, waist circumference, hip circumference, LH, androstenedione, testosterone, and dehydro-epiandrosterone sulphate increased significantly (linear regression, P < 0.05) by gynecological age in all menstrual cycle pattern groups. For PRL and estradiol a significant increase with gynecological age was only documented in the regular menstrual cycle group and for waist to hip ratio only in the irregular menstrual cycle group. No significant correlation could be documented between gynecological age and overnight fasting insulin concentrations or glucose to insulin ratio.
We found no significant correlation between insulin concentrations or glucose to insulin ratio and androgen concentrations. Significant positive correlations were found between LH and androgens.
LH and androgen levels increase during the first years after menarche, and reference values should be adjusted for gynecological age. In these years, no significant correlation between hyperinsulinemia and hyperandrogenemia could be documented.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
Puberty is a state of relative insulin insensitivity and compensating hyperinsulinemia (11, 12, 13, 19, 20, 21). To examine whether the physiological, relative hyperinsulinemia during this period of life goes together with hyperandrogenism or other sequelae as seen in adult patients with the polycystic ovary syndrome (PCOS) (22, 23, 24, 25), we studied the correlation between insulin concentrations and glucose to insulin ratio (GI ratio), on one hand, and body measurements, gonadotropins, and androgens, on the other hand.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
The methodology of the POMP study has been described in detail elsewhere (17, 18). In summary, the POMP study can be divided into three major phases. In the first phase, 2248 white, ninth grade schoolgirls filled out a questionnaire and were interviewed on their menstrual cycle pattern. The definitions of the menstrual cycle patterns were: 1) regular menstrual cycles: an average length of the cycle between 22 and 41 days, either none or a single cycle with a length less than 22 or more than 41 days during the past year; 2) irregular menstrual cycles: an average length of the cycle between 22 and 41 days, two or more cycles with a length less than 22 or more than 41 days during the past year; 3) oligomenorrhea: an average length of the cycle between 42 and 180 days; 4) secondary amenorrhea: the absence of menstruation for 180 days or more; and 5) polymenorrhea: an average length of the cycle of 21 days or less.
All girls with oligomenorrhea, secondary amenorrhea, or polymenorrhea and a sample of the girls with irregular menstrual cycles were invited for a physical examination and vena puncture. For every girl with oligomenorrhea or secondary amenorrhea who agreed to vena puncture, two girls with regular menstrual cycles were enrolled in the control group. The girls who gave blood were representative samples of the menstrual cycle groups from which they were a member (18).
The adolescents who were interviewed in the first phase of the study were asked approval to have a follow-up questionnaire mailed to them after 18 months. In the second and third phases of the POMP study, adolescents with regular menstrual cycles who participated in vena puncture in the first phase, and all girls with oligomenorrhea, secondary amenorrhea, polymenorrhea, and irregular menstrual cycles, and those who had not yet, or less than 6 months earlier, experienced menarche at the time of the first questionnaire received a second and third questionnaire after 18 and 36 months, respectively. The responders were categorized according to menstrual cycle pattern and were invited for venapuncture.
Three hundred thirty-nine girls gave 490 blood samples. Forty-nine
girls gave three blood samples, 98 girls gave two samples, and 147
girls gave one sample. Table 1
shows the
number of adolescents who gave blood in the different phases of the
study categorized by menstrual cycle pattern. For the statistical
analysis, girls with normogonadotropic secondary amenorrhea were
recorded as oligomenorrhea (26). Polymenorrhea was rare, prevalence
0.8% (17). This subgroup was too small for statistical analysis.
|
Methods
Physical examination.Weight was measured to the nearest 0.5 kg using a mobile spring scale (Seca, Hamburg, Germany), and height was measured to the nearest 0.5 cm using a mobile measuring rod (Microtoise; Stanley mabo, Poissy, France). Waist circumference was measured to the nearest 0.5 cm with a plastic tape at the smallest frontal waist diameter, usually at the level of the umbilicus. Hip circumference was measured at the broadest part of the lower body, usually at the level of the trochanters (27). Body hair grading was assessed according to the Ferriman and Gallwey (F&G) score (28). Abnormal body hair was defined as a F&G score of 1 or more, and hirsutism as a F&G score of 8 or more.
Blood sampling and hormone measurements.In the first phase of the study, blood was taken at school between 1200 and 1700 h. In the second and third phases of the study, blood was taken after an overnight fast between 0800 and 1000 h at school or at home. In girls with regular or irregular menstrual cycles the blood was taken between the 1st and 10th day of the menstrual cycle. To exclude the influence of a midcycle LH peak, blood from these girls should, in retrospect, have been taken at least 18 days before the next menstruation. In oligomenorrheic girls the blood sample should have been taken at least 2 weeks after the 1st day of a period and at least 3 weeks before the next period. This procedure excludes the possible influence of periovulatory hormonal changes and postovulatory progesterone production on the LH and androgen concentrations, which extends itself into the follicular phase of the next menstrual cycle (29, 30, 31, 32, 33, 34). In secondary amenorrhea, blood samples were taken at random.
All hormones were determined by commercially available kits. LH and FSH
were determined by immunofluorometric assays (Amerlite; Amersham Pharmacia Biotech UK, Amersham, UK). PRL concentrations
were measured by an immunoradiometric assay (Medgenix Diagnostics,
Fleurus Belgium); estradiol (E2) concentrations were
measured by a RIA (Estradiol-2; Sorin Biomedica, Saluggia, Italy);
androstenedione (
A) and dehydroepiandrosterone-sulphate (DHEAS)
concentrations were measured by a RIA; and testosterone (T) was
measured by a double antibody RIA (Coat a Count; DPC, Los Angeles, CA).
Insulin concentrations were measured by RIA (Insik-5; Sorin Biomedica).
Glucose was measured by automatic hexokinase method.
Androgens and E2 are expressed in standard international
units. The conversion factor to metric units is described by the
following equations: for A µg/L * 3.49 = nmol/L; for T ng/mL
* 3.47 = nmol/L; for DHEAS ng/ml * 0.00271 = µmol/L; and
E2 pg/ml * 3.67 = pmol/L.
LH, FSH, A, T, and DHEAS were measured in all samples.
E2 and PRL concentrations were measured in the first sample
of girls with regular and irregular menstrual cycles and in all samples
of girls with oligo- or amenorrhea. Insulin and glucose concentrations
were only measured in samples taken after an overnight fast in the
second and third phases of the study.
Statistics.The data were analyzed with BMDP statistical software package (BMDP statistical software, Cork, Ireland). Means ± SEM are presented. Although our study was prospective, we were not able to collect usable endocrine data or body measurements of all adolescents at all phases of the study. A major problem in the study was drop out due to the start of oral contraceptive pills, about 60% of all participants. This made the number in the various menstrual cycle pattern groups too small for a longitudinal analysis. The data were analyzed in a cross-sectional mode. In each analysis, cases with incomplete data were excluded for that specific analysis. Multiple linear regression analysis (Module 2R) was used to estimate the influence of the number of samples (1, 2, or 3) a participant gave or the phase of the study in which the blood sample was taken on the analysis.
The gynecological age (months) was calculated by subtracting the age at menarche (months) from the calendar age (months). The least squares linear regression equation between gynecological age and body measurements or hormones was calculated by linear regression analysis (module 2R). The effect of logarithmic or polynomial transformation of the dependent variables was evaluated. The equality of the regression lines across the menstrual cycle pattern groups was tested by linear regression by groups (module 1R). Differences in the slopes of the regression lines in the various menstrual cycle groups were evaluated by multiple linear regression, with determinants describing the interaction between menstrual cycle pattern and gynecological age in the regression model (module 2R). The mean hormone levels of girls with regular menstrual cycles vs. those with other menstrual cycle patterns were compared by analysis of covariance to adjust for differences in gynecological age between the various menstrual cycle pattern groups.
Correlation coefficients between the various hormones and body measurements were calculated (Module 6D). The effect of the number of samples per individual on the correlation coefficients was estimated by calculating the partial correlation coefficient adjusted for these determinants (Module 2R).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
Table 2 shows the least squares
regression equations of the BMI, waist circumference, hip
circumference, WH ratio, pituitary hormones, androgens,
E2, insulin, and GI ratio by gynecological age
and menstrual cycle pattern. Fig. 1
illustrates the mean body measurements and hormones by gynecological
age in years. In all menstrual cycle pattern subgroups, BMI and waist
and hip circumference showed a significant increase by gynecological
age (F-ratio, P < 0.05). In the irregular menstrual
cycle group, a significant increase by gynecological age was also found
for WH ratio. In all menstrual cycle pattern subgroups, LH, T, A,
and DHEAS showed a significant increase (F-ratio, P <
0.05) during the first 6 yr after menarche. FSH showed no increase, and
PRL and E2 showed only a significant increase in
the regular menstrual cycle group. No significant association could be
documented between gynecological age and insulin or GI ratio.
|
|
Because the regular menstrual cycle group was defined as the reference group including normal subjects, a possible effect of obesity or hirsutism of subjects in this group was evaluated. The prevalences of obesity (BMI >25.0 kg/m2, 4.4%), abnormal body hair (F&G score >0, 10%) or hirsutism (F&G score >7, 1.3%) in the regular menstrual cycle group were low. Exclusion of obese or hirsute girls from the analysis had no effect on the relation between gynecogical age and androgen levels. Exclusion of girls with regular menstrual cycles in one phase of the study and irregular menstrual cycles or oligomenorrhea in another phase of the study also had no significant effect.
Body measurements and hormones by menstrual cycle pattern
Table 3 shows the mean ±
SEM of the body measurements and hormones by menstrual
cycle pattern. No significant difference was found in BMI, waist and
hip circumference, or WH ratio between the various menstrual cycle
pattern groups with analysis of covariance adjusting for differences in
gynecological age.
|
A levels were significantly higher in the oligomenorrhea
group compared with the regular menstrual cycle group. DHEAS and PRL
levels were higher in the irregular menstrual cycle group compared with
the regular menstrual cycle group. PRL concentrations in the irregular
menstrual cycle group were also significantly higher than those levels
in the oligomenorrhea group. The difference in these levels was rather
small: 0.21 U/mL in the regular menstrual cycle group vs.
0.24 U/mL in the irregular menstrual cycle subgroup.
Insulin concentrations did not differ between the menstrual cycle
pattern groups. The GI -ratio was significantly lower
(P = 0.04) in the regular menstrual cycle group
compared with the oligomenorrhea group. The proportion of adolescents
with BMI <18 kg/m2 was significantly higher in
the oligomenorrhea group (28%) than in the regular menstrual cycle
group (7%). After exclusion of these adolescents with low BMI, no
significant difference was found in insulin concentrations (10.0
± 0.4 mU/L vs. 10.1 ± 0.8 mU/L) or GI ratio (9.6
± 0.4 ng/10-4 U vs. 10.4 ± 0.8
ng/10-4 U) between the regular menstrual cycle
group and the oligomenorrhea group. Mean LH, A, and T concentrations
of oligomenorrheic girls increased by about 5–10% after exclusion of
girls with low BMI.
Correlation between androgen or insulin concentrations and BMI and body fat distribution
Table 4 shows the correlation
coefficients between the various body measurements and androgens,
insulin, and GI ratio by menstrual cycle pattern. Weak, positive
correlations were found between BMI, waist and hip circumference, on
one hand, and androgens and insulin concentration, on the other hand.
Weak, negative correlations were found between these body measurements
and GI ratio. The correlation coefficients in the irregular menstrual
cycle group and the oligomenorrhea group did not differ significantly
from those in the regular menstrual cycle group.
|
Correlation between insulin or LH concentrations and androgen concentrations
Table 5 shows the correlation
coefficients between insulin, GI ratio and LH, on one hand, and
androgens, on the other hand. Except for an unexpected negative
correlation between T and insulin in the regular menstrual cycle group,
no significant correlation could be documented between insulin
concentrations or GI ratio and androgen concentrations. LH was
significantly correlated with A and T in the irregular menstrual
cycle group and the oligomenorrhea group. All correlations were low. At
most, 22% (0.472) of the variation in androgen
levels was explained by variation in LH levels.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
A, and T concentrations were significantly higher in the
oligomenorrhea group. The relations between gynecological age and mean
hormone concentrations or body measurements were best described by
linear equations. Although Fig. 1
suggests that the increase in these
parameters by gynecological age expressed in years seems bigger in the
first 3 yr after menarche, we were not able to identify a gynecological
age after which the increase significantly diminished. The hormone
levels in the 5th and 6th yr after menarche are comparable with the
reference values given by the manufactures of the assays for adults.
Although insulin levels are described to be higher at the end of
puberty (i.e. Tanner stage 4 and 5) than during prepuberty
and adulthood (11, 12, 13, 19, 20, 21), we could not document a significant
decrease in basal insulin concentrations or an increase in GI ratio by
gynecological age. Because insulin levels were only determined in the
second and third phases of the study, a decrease in the first 2 years
after menarche may have remained undetected due to a small number of
participants (n = 14) with a gynecological age < 2
years. Limitations of our study are that not all participants had equal numbers of blood samples and the data are mixed cross-sectional and longitudinal. The influence on the results of the number of samples or the phase of the study in which a sample was taken was examined, and no significant influence could be documented. The main reason for drop out was the start of oral contraceptive pills. An influence of this on the analysis cannot be excluded, however, within the menstrual cycle pattern subgroups the baseline characteristics of pill users vs. non-pill users were not different. Another limitation is that we did not document ovulation. Hormone concentrations may differ in ovulatory and anovulatory cycles in adolescents (9, 10, 14, 15). Furthermore, anovulation in the cycle preceding the study cycle may result in higher LH and androgen levels in the follicular phase of the study cycle (31, 32, 33, 34). The prevalence of ovulatory cycles increases from 14–38% in the 1st year after menarche to 65–87% 4–6 yr after menarche (10, 14, 16, 35) Thus, the increase in LH and androgen levels in the various menstrual cycle groups seems to coincide with an increase in prevalence of ovulatory cycles. LH and androgen levels may have been relatively high due to anovulation during first years after menarche and the increase of these hormones may have been stronger if only ovulatory cycles had been studied
No clear explanation can be given for the significant differences in PRL and DHEAS concentrations between girls with regular and irregular menstrual cycles (18). The hormonal pattern in the oligomenorrhea group resembles that seen in adult PCOS and may be a stage in the development of this condition (14, 15, 16, 18, 36). In adults, several observations suggest an important role for hyperinsulinemia in the pathophysiology of hyperandrogenism and the PCOS (22, 23, 24, 25). The insulin-IGF-I hypothesis postulates that progressively rising insulin and IGF-I levels during puberty cause excessive ovarian stimulation and induce a polycystic ovary syndrome in predisposed girls.(37) Our data are not in line with this hypothesis. We found no association between basal insulin concentrations or GI ratio with androgens or menstrual cycle pattern; however, other well-known relationships such as between basal insulin concentrations and BMI or between menstrual cycle pattern and androgen or LH concentrations were confirmed. Apter et al. (38) found insulin insensitivity in six extremely obese girls (BMI >30 kg/m2) with oligomenorrhea and hyperandrogenemia (38). Our group of oligomenorrheic girls was representative for all oligomenorrheic girls in a geographically defined area. Only 10% of these oligomenorrheic girls were obese. Extreme obesity as described in the study of Apter et al. (38) is very rare in our population and in The Netherlands.
Our oligomenorrheic group will be a mixture of girls in whom this menstrual cycle pattern is a stage in their maturation to a regular menstrual cycle pattern and girls who have or will develop PCOS, characterized by oligo- or anovulation with high LH or androgen concentrations. Earlier longitudinal studies on oligomenorrhea or hyperandrogenemia showed that adolescents maintain these characteristics in adulthood (36, 39), suggesting that about 50% of our oligomenorheic girls will develop PCOS as adults. We think a primary, eliciting role for insulin in the development of the PCOS is unlikely. However, an excessive production of androgens by the ovaries in reaction to normal insulin or IGF-I stimulations or an eliciting role in extreme obese girls remains possible. Our data suggest a more prominent role for hyperandrogenism due to enzyme dysregulation or inappropriate LH secretion in the pathophysiology of PCOS. It has also been suggested that adolescent menstrual irregularities are a result of an inoperative LH response to estrogens due to immaturity of the estrogen-induced positive feedback mechanism (40). The fairly high frequency of ovulation in oligomenorrheic girls (65.9%) is in contradiction with this hypothesis (16).
Despite the limitations we mentioned, we think our findings give a true description of these changes in the first years after menarche and of the correlations between these determinants. Because LH and androgens in the years after menarche are lower than in adults, reference values to diagnose hyperandrogenism or inappropriate LH secretion should be adjusted for gynecological age. As a correlation between basal insulin levels or GI ratio and androgen levels or oligomenorrhea is missing, the insulin-IGF-I hypothesis as an explanation for the development of PCOS in puberty is not supported.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received August 25, 1999.
Revised December 15, 1999.
Accepted December 30, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
-dihydrotestosterone during female puberty. J Clin Endocrinol
Metab. 45:1039–1048.[Medline]
This article has been cited by other articles:
 |
|
 |
|
  K. Zhang, S. Pollack, A. Ghods, C. Dicken, B. Isaac, G. Adel, G. Zeitlian, and N. Santoro Onset of Ovulation after Menarche in Girls: A Longitudinal Study J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1186 - 1194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. J. Baer, G. A. Colditz, W. C. Willett, and J. F. Dorgan Adiposity and Sex Hormones in Girls Cancer Epidemiol. Biomarkers Prev., September 1, 2007; 16(9): 1880 - 1888. [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]
|
|
|
|
|
|
S. S. Tworoger, A. H. Eliassen, S. A. Missmer, H. Baer, J. Rich-Edwards, K. B. Michels, R. L. Barbieri, M. Dowsett, and S. E. Hankinson Birthweight and Body Size throughout Life in Relation to Sex Hormones and Prolactin Concentrations in Premenopausal Women Cancer Epidemiol. Biomarkers Prev., December 1, 2006; 15(12): 2494 - 2501. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. McCartney, K. A. Prendergast, S. Chhabra, C. A. Eagleson, R. Yoo, R. J. Chang, C. M. Foster, and J. C. Marshall The Association of Obesity and Hyperandrogenemia during the Pubertal Transition in Girls: Obesity as a Potential Factor in the Genesis of Postpubertal Hyperandrogenism J. Clin. Endocrinol. Metab., May 1, 2006; 91(5): 1714 - 1722. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Chhabra, C. R. McCartney, R. Y. Yoo, C. A. Eagleson, R. J. Chang, and J. C. Marshall Progesterone Inhibition of the Hypothalamic Gonadotropin-Releasing Hormone Pulse Generator: Evidence for Varied Effects in Hyperandrogenemic Adolescent Girls J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2810 - 2815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A.M. de Boer, C. B. Lambalk, H. H. Hendriks, C. van Aken, E. A. van der Veen, and J. Schoemaker Growth hormone secretion is impaired but not related to insulin sensitivity in non-obese patients with polycystic ovary syndrome Hum. Reprod., March 1, 2004; 19(3): 504 - 509. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.H.A. van Hooff, F.J. Voorhorst, M.B.H. Kaptein, R.A. Hirasing, C. Koppenaal, and J. Schoemaker Predictive value of menstrual cycle pattern, body mass index, hormone levels and polycystic ovaries at age 15 years for oligo-amenorrhoea at age 18 years Hum. Reprod., February 1, 2004; 19(2): 383 - 392. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Clavel-Chapelon Evolution of age at menarche and at onset of regular cycling in a large cohort of French women Hum. Reprod., January 1, 2002; 17(1): 228 - 232. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, Slope of the
regression line of the oligomenorrhea group significantly different
from the slope in the regular menstrual cycle group. 
, Mean
significantly different (P < 0.05) compared with
girls with regular menstrual cycles by analysis of covariance to adjust
for differences in gynecological age.