Introduction

Sex hormone binding globulin (SHBG) is a glycoprotein primarily synthesized in the liver. Its main function has traditionally been considered a transport protein for serum sex steroids, controlling circulating concentrations of free hormones. In vitro studies have demonstrated that estradiol stimulates SHBG secretion in a dose-dependent manner, whereas testosterone has a biphasic effect lowering it at both low and high concentrations.

However, there is evidence suggesting that SHBG has other functions besides carrying sex hormones. Recent studies report that SHBG may contribute to the pathophysiology of insulin resistance and Type 2 diabetes mellitus (T2DM) (1,2).

Observational studies have associated low serum SHBG concentrations with T2DM (3)_, regardless the level of sex hormones in men or women (4). This suggests that SHBG may have a role in the pathogenesis of T2DM and could allow predicting the risk of developing this disease.

The role of SHBG in the pathogenesis of insulin resistance and T2DM has been mostly analyzed in adults (3,4). A recently published, longitudinal population-based study in adult women showed a significant association between low SHBG levels and an increased risk to metabolic syndrome and T2DM (5). However, only few studies have been carried out in children or adolescents (6).

Menstrual disturbances and hirsutism are common complaints in adolescent girls, the latter especially among women in the Mediterranean area. These features are characteristic of polycystic ovary syndrome (PCOS), the most frequent endocrine disorder in women of reproductive age and considered a polygenic condition partly dependent on genetic and environmental factors (7), more specifically overweight and obesity. However, there is no agreement on how to diagnose PCOS in adolescents and whether the criteria for the diagnosis of this condition in adult women would be suitable for adolescents (8,9).

Considering the few published reports on SHBG alterations in this age group population and the lack of consensus on diagnostic criteria of PCOS in adolescents, the aim of the study is to determine if there is an association of SHBG with insulin resistance and hyperandrogenism markers in adolescent women who complained of hirsutism and/or menstrual disturbances, and if there is an influence of weight on this relationship.

Materials and Methods

Observational, transversal, and descriptive study in which all adolescent women who visited the Pediatric Endocrinology Unit in a third level University Hospital due to hirsutism or/and menstrual abnormalities between 2009 and 2014 were evaluated.

The study was carried out according to the principles of the Helsinki Declaration of 1975 and the protocol was approved by the ethics committee of the hospital.

Initially, 40 patients who had a complete hormonal and biochemical analysis (including SHBG, sex hormones, and carbohydrate metabolism) with features of polycystic ovarian syndrome (10) were included in the study.

Subjects fulfilling the following criteria were included in the study: a) oligomenorrhea and/or anovulation or clinical/biochemical parameters for hyperandrogenism, b) at least two years of evolution since menarche, c) have performed analytical hormones in early follicular phase of the menstrual cycle (post-treatment with oral progesterone overload, 10 mg  of Progevera^® for 5 days, if amenorrhea).

Exclusion criteria included: a) history of thyroid disorders, b) current treatment with oral contraceptives, c) hyperprolactinemia (>30ng/ml), d) diabetes mellitus, e) Cushing’s syndrome, f) androgen secreting tumor, or g) congenital adrenal hyperplasia.

Five patients were excluded from the study due to underlying diseases: two had elevated HbA1c levels (9.1% and 6.3%); they underwent oral glucose tolerance tests and were diagnosed with T2DM and impaired glucose tolerance (IGT), respectively; one patient had non-classical congenital adrenal hyperplasia (serum 17-OH-progesterone concentration after ACTH stimulation of 1052 ng/dL); one patient suffered from neurofibromatosis type 1, and another had adrenal Cushing's syndrome waiting for surgery. Thus, 35 patients were included in the study.

Participants were seen by a pediatric endocrinologist who obtained the following information: patients´ medical history, age of menarche, regularity of the menstrual cycle, anthropometric measurements, Tanner stage, scoring of hirsutism (using Ferriman and Gallwey (F-G) score), and other relevant physical alterations (presence of goiter, acanthosis nigricans, etc.). Body mass index (BMI) was calculated as the weight in kilograms divided by the square of the height in meters. Standard deviation scores (SDS) for BMI were calculated according to Hernández et al. normative data (11).

Blood samples were collected by venipuncture at 07:00 and 08:00 h after an overnight fast. Plasma glucose concentrations were measured by routine clinical chemistry immediately after extraction. HbA_1c was measured in blood samples with EDTA by high-pressure liquid chromatography using a fully-automated Adams Menarini HI-AUTO A1c 8160 analyzer manufactured by Arkray (Kyoto, Japan); the inter-assay coefficient of variation (CV) was 1.8% and 1.5% at HbA_1c levels of 4.8% and 9.0%, respectively (reference range: 4-5.8 %). Serum samples for the remaining biochemical measurements were immediately frozen and stored at -20ºC until assayed, and were analyzed simultaneously under the same analytical conditions.

Serum glucose concentrations were measured with a routine clinical chemistry laboratory analyzer (Cobas c711, Roche Diagnostic, Spain).

Immunoreactive insulin, serum lutropin (LH), follitropin (FSH), estradiol, dehydroepiandrosterone (DHEAS), sex hormone binding globulin (SHBG) and testosterone levels were determined using an automated electrochemiluminescence immunoassay (Modular E, Roche Diagnostics GmbH Mannheim, Germany).

For insulin, the analytical sensitivity was 0.2 mUI/l and the inter-assay CV < 2.8%. Insulin resistance was assessed by using the homeostasis model assessment of insulin resistance (HOMA-IR) and calculated with fasting glucose and insulin using the following equation: HOMA-RI-index= fasting insulin (mUI/L) x fasting glucose (mmol/L)/ 22.5.

For LH and FSH, the analytical sensitivity was 0.10 mIU/mL for both and the inter-assay CV was <2.2% for LH and <5.3% for FSH.

For estradiol, the analytical sensitivity was 5 pg/mL and the functional sensitivity was 12 pg/mL; inter-assay CV was <6.2%.

For DHEAS, the analytical sensitivity was 1 ng/mL and the inter-assay CV was < 4.7%.

For SHBG, the analytical sensitivity was 0.350 nmol/l. and the inter-assay CV was < 5.6%.

For testosterone, the analytical sensitivity was 2.5 ng/dL and the inter-assay CV was <8.4%. Free androgen index (FAI) was calculated as the percentage ratio of testosterone to SHBG on a molar basis.

17-hydroxiprogesterone concentrations were measured by a solid-phase ¹²⁵I radioimmunoassay (Siemens Healthcare Diagnostics, Newark, USA). The analytical sensitivity was 7 ng/dL and the inter-assay CV was <10%.

Androstenedione concentrations were measured by a solid-phase ¹²⁵I radioimmunoassay (ZenTech, Angleur, Belgium). The analytical sensitivity was 5 ng/dL and the inter-assay CV was <8.7%. Genital ultrasound and other explorations were done when necessary.

In a second phase, in order to assess the effect of adiposity on insulin resistance and other hormonal parameters, the cohort was divided into two groups based on standardized BMI according to Hernández et al. normative data (11):
 BMI-SDS< 1.5 (normal weight) or BMI-SDS >1.5 (overweight or obesity).

Descriptive statistics were expressed as mean±SD. Departure from normality was assessed by the Kolmogorov-Smirnov test. Student’s t test was used for comparing between mean values of the groups. Associations between variables were evaluated by the Pearson’s correlation test, and multiple linear regression analyses were performed by using the stepwise method in order to identify which factors independently determine serum SHBG concentrations.

The statistical significance level was fixed at 5%. Data were analyzed using the SPSS 12.0 statistical package.

Results

Patients´ characteristics and measured parameters (means ± SD) of the whole group are shown in Table 1.

Table 2 shows the same anthropometric, clinical, and laboratory data classified according to their BMI-SDS: either normal weight (BMI-SDS<1.5) (n=19) or overweight/obese (BMI-SDS≥1.5) (n=16).

Mean values (obese vs. normal weight) for SHBG (22.1±11.8nmol/L vs. 35.0±16.9nmol/L, p=0.015) were significantly lower in overweight/obese girls. However, both insulin (25.6±13.7 vs. 12.8±5.3, p<0.05) and HOMA-IR (5.4±2.8 vs. 2.3±0.63, p <0.05) were significantly higher in the overweight/obese group.

No differences were found in androgen levels or degree of hirsutism between groups.

The univariate correlation analysis is summarized in Table 3. In all patients and for the overweight/obese group, SHBG inversely correlated with BMI-SDS (r=-0.528, p=0.001 and r=-0.509, p=0.044), insulin (r=-0.476, p=0.011 and r=-0.682, p=0.007), HOMA-IR (r=-0.438, p=0.025 and r=-0.656, p=0.011), and FAI (r=-0.656, p<0.001 and r=-0.639, p= 0.008), respectively.

On the other hand, for the group of patients with BMI-SDS<1.5, the negative correlation between SHBG, BMI-SDS, insulin and HOMA-IR was not significant. However, a significant positive association with FSH (r=0.589, p=0.021) and negative with testosterone (r=-0.476, p=0.039), the free androgen index (r=-0.736, p<0.001) and hirsutism (r=-0.491, p=0.033) was observed for this group.

A multiple linear regression analysis was performed in order to identify which anthropometric and hormonal variables were independent and could explain the changes in serum SHBG concentrations. Variables that had previously shown a significant correlation with SHBG in the univariate analysis were included in this analysis.

In the group with overweight/obese participants, BMI-SDS and HOMA-IR explained 43% of SHBG variability (p=0.045). However, only HOMA-IR was an independent factor for prediction (r=0.656, β=-0,522, p=0.011) (Table 4), because BMI-SDS lost statistical significance when corrected for HOMA-IR.

In the normal weight group of subjects, FSH and testosterone were selected as predictor variables for SHBG (r=0.681; p=0.024), but only FSH remained as an independent factor (r=0.589, β=7.562, p=0.021) (Table 4); thus, in the final model, each one unit increase in FSH was associated with a 7.562% increase in SHBG concentration.

The relationship between SHBG with both predictors (FSH and HOMA-IR) is shown in figures 1 and 2, respectively.

Discussion

To date, SHBG was thought to act as a carrier of sex steroids, although recent studies have reported and association between SHBG and increased risk to develop metabolic syndrome and T2DM (1,2,5).

Metabolic syndrome is a risk factor for T2DM, due to the increased insulin resistance that this implies.  It is thus necessary to rule out factors that associate to this syndrome, e.g. overweight or obesity, to better study the relationship between SHBG and insulin resistance.

The levels of SHBG depend on sex and age (12,13) and show an inverse correlation with androgens and positive with estrogens. As sex hormones modulate insulin resistance, a study in which all patients are in the same pubertal stage and of the same gender can provide useful information.

Our findings are consistent with results previously reported in adults. Recently, a population-based, longitudinal, observational study was published on potential associations between endogenous androgens and SHBG levels and cardiometabolic risk, which revealed an inverse association of SHBG levels in women with metabolic syndrome and T2DM, therefore low SHBG levels might represent a risk marker for this diseases in women (5). In our patients an inverse relationship between SHBG with markers of insulin resistance, BMI and hyperandrogenism is also observed. The fact that HOMA-IR behave as an independent factor of variability of SHBG would support the hypothesis that this protein could be used as an early marker of insulin resistance (5,14).

It seems that overweight/obese girls with PCOS have a greater decrease of SHBG values in comparison with subjects with normal weight. Some studies indicate that this occurs mainly because of the excess of body fat rather than secondary to hyperandrogenism and insulin resistance. However, in our study, multiple regression analyses showed that the relationship between SHBG concentrations and HOMA-IR, in overweight/obese girls, was in part independent from BMI as reported by Gascón F. et al. (6), who studied a cohort of 122 prepubertal children.

Sex hormones have different binding affinities for SHBG. e.g., the affinity of testosterone is two-fold greater than that of estradiol (15). In this study, all girls had some degree of hyperandrogenism but multiple linear regression showed the HOMA-IR as the only independent variable for SHBG variability (p <0.001). Thus, the relationship between SHBG and this insulin resistance marker in individuals of the overweight/obese group may not be explained by fluctuations in serum androgen concentrations or free sex steroid levels as observed in other studies. (15,16)

In the normal weight group, HOMA-IR does not exert any effect on SHBG concentrations, being FSH the principal predictor that explains its variability in blood. FSH is the main regulator of follicular growth and maturation, as well as the responsible for estrogen secretion; therefore, it is not surprising that it shares certain relationship with SHBG status.

No significant correlation was found between estrogen and SHBG, explained by the limitations of estradiol measurement. The precision of its determination has been reported to be insufficient, particularly at low concentrations (16).

Genetic studies suggest that the transmission of specific polymorphisms in the SHBG gene could have an influence on the risk of developing T2DM (3, 17,18).

There are several limitations in our study. First, the study was a cross-sectional survey and larger prospective studies are needed to assess the relationships. Secondly, we selected patients that complained for menstrual abnormalities or hirsutism and that these women might not be representative of female adolescents without features of PCOS. A study that would include a control group and a larger sample size should be carried out to determine how SHBG behaves in both groups.

Conclusions

In adolescent girls with PCOS features, serum SHBG is subjected to different regulations in a weight-dependent manner.

In overweight/obese subjects, HOMA-IR is an independent factor, which explains SHBG variability. This could support the hypothesis of this protein possibly being an early marker of insulin resistance for this population. However, in girls with normal weight, FSH, responsible of estrogen secretion, predominantly explains SHBG variability.

More observations are needed in the group of adolescent female to confirm the results reported in adults.

Conflicts of interest

Authors declare no potential Conficts of Interest.

1. Le TN, Nestler JE, Strauss III JF, Wickham III EP. Sex hormone-binding globulin and type 2 diabetes mellitus. Trends Endocrin Met 2012; 23: 32-40.

2. Wallace IR, McKinley MC, Bell PM, Hunter SJ. Sex hormone binding globulin and insulin resistance. Clin Endocrinol 2013; 78: 321-9.

3. Ding EL, Song Y, Manson JE, Hunter DJ, Lee CC, Rifai N, et al.. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med 2009; 361: 1152-63.[Pubmed]

4. Ding EL, Song Y, Malik VS, Liu S. Sex differences of endogenous sex hormones and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA 2006; 295: 1288-99.[Pubmed]

5. Fenske B, Kirsche H, Gross S, Wallaschofski H, Völzke H, Dörr M, et al. Endogenous androgens and sex hormone-binding globulin in women and risk of metabolic syndrome and type 2 diabetes. J Clin Endocrinol Metab 2015. http://dx.doi.org/10.1210/jc.2015-2546.

6. Gascón F, Valle M, Martos R, Ruz FJ, Ríos R, Montilla P, et al. R. Sex hormone-bindinng globulin as a marker for hyperinsulinemia and/or insulin resistance in obese children. Eur J Endocrinol 2000; 143: 85-9.[Pubmed]

7. Conway G, Dewailly D, Diamanti-Kandarakis E, Escobar-Morreale HF, Franks S, Gambineri A, et al; ESE PCOS Special Interest Group. The polycystic ovary syndrome: a position statement from the European Society of Endocrinology. Eur J Endocrinol 2014; 171:1-29.[Pubmed]

8. Hickey M, Doherty DA, Atkinson H, Sloboda DM, Franks S, Norman RJ, et al. Clinical, ultrasound and biochemical features of polycystic ovary syndrome in adolescents: implications for diagnosis. Hum Reprod 2011; 26: 1469-77.[Pubmed]

9. Carmina E, Oberfield SE, Lobo RA. The diagnosis of polycystic ovary syndrome in adolescents. Am J Obstet Gynecol 2013; 203: 201.e1-e5.

10. Rotterdam ESHRE/ASRM-Sponsored PCOS consensus workshop group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 2004; 19: 41-7.[Pubmed]

11. Hernández M, Castellet J, Narvaiza JL, Rincón JM, Ruiz I, Sánchez E, Sobradillo B, Zurimendi A. Curvas y tablas de crecimiento (0-18 años). Instituto de Investigación sobre Crecimiento y Desarrollo. Fundación Faustino Orbegozo, Bilbao, ISBN: 84-7391-177-6, 1988. pp. 1-32.

12. Konforte D, Shea JL, Kyriakopoulou L, Colantonio D, Cohen AH, Shaw J, et al.. Complex biological pattern of fertility hormones in children and adolescents: a study of healthy children from the CALIPER cohort and establishment of pediatric reference intervals. Clin Chem 2013; 59: 1215-27.[Pubmed]

13. Elmlinger MW, Kühnel W, Wormstall H, Döller PC. Reference intervals for testosterone, androstendione and SHBG levels in healthy females and males from birth until old age. Clin Lab 2005; 51: 625-32.[Pubmed]

14. Davis SR, Robinson PJ, Moufarege A, Bell RJ. The contribution of SHBG to the variation in HOMA-IR is not dependent on endogenous oestrogen or androgen levels in postmenopausal women. Clin Endocrinol (Oxford) 2012; 77: 541-7.

15. Dunn JF, Nisula BC, Rodbard D. Transport of steroid hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma. J Clin Endocrinol Met 1981; 53: 58-68.

16. Rosner W, Hankinson SE, Sluss PM, Vesper HW, Wierman ME. Challenges to the Measurement of Estradiol: An Endocrine Society Position Statement from the European Society of Endocrinology. Eur J Endocrinol 2014; 171: 1-29.[Pubmed]

17. Perry JR, Weedon MN, Langenberg C, Jackson AU, Lyssenko V, Sparsø T, et al. Genetic evidence that raised sex hormone binding globulin (SHBG) levels reduce the risk of type 2 diabetes. Hum Mol Genet 2010; 19: 535-44.[Pubmed]

18. Aydın B, Winters SJ. Sex Hormone Binding Globulin in Children and Adolescents. J Clin Res Pediatr Endocrinol. 2015 Jan 18. http://dx.doi.org/10.4274/jcrpe.2764.  PubMed PMID: 26761949;