Endocrine DOI 10.1007/s12020-014-0200-7
ORIGINAL ARTICLE
Prevalence and impact of hyperandrogenemia in 1,218 women with polycystic ovary syndrome Sarantis Livadas • Christos Pappas • Athanasios Karachalios • Evangelos Marinakis • Nikoleta Tolia • Maria Drakou • Philippos Kaldrymides • Dimitrios Panidis • Evanthia Diamanti-Kandarakis
Received: 1 October 2013 / Accepted: 31 January 2014 Springer Science+Business Media New York 2014
Abstract Hyperandrogenemia modifies phenotypic characteristics of women with polycystic ovary syndrome (PCOS). The aim of the present study is to evaluate (a) the prevalence of hyperandrogenemia in PCOS women (Rotterdam criteria) and (b) the impact of either the degree or the type of hyperandrogenemia on phenotype. Anthropometric, clinical, hormonal, metabolic and ultrasound characteristics of 1,218 women with PCOS were analyzed in this cross-sectional study. The prevalence of hyperandrogenemia was 58.8 %. Women with hyperandrogenemia had higher luteinizing hormone (LH), follicle-stimulating hormone (FSH), free androgen index, lower sex-hormonebinding globulin (SHBG) and fasting glucose levels compared to women with normal androgens (p \ 0.001 for all comparisons; p = 0.001 for fasting glucose). Regarding the presence of isolated hyperandrogenemia, the group with only elevated testosterone levels was termed GT and an analogous categorization was made for dehydroepiandrosterone sulfate (GD) and androstenedione (D4) (GD4), respectively. GT, GD and GD4 comprised the 17.2, 7.6 and 4.1 % of total cohort, respectively. These groups differed significantly between them in LH, LH/FSH ratio, and SHBG (p \ 0.001). Hyperandrogenemia is found in almost S. Livadas C. Pappas A. Karachalios E. Marinakis N. Tolia M. Drakou P. Kaldrymides E. Diamanti-Kandarakis (&) Endocrine Unit, Third Department of Internal Medicine, Medical School, National and Kapodistrian University of Athens, Mesogeion 152, 11527 Athens, Greece e-mail: [email protected]; [email protected] D. Panidis Division of Endocrinology and Human Reproduction, Second Department of Obstetrics and Gynecology, Hippokration Hospital, Aristotle University of Thessaloniki, Thessalonı´ki, Greece
60 % of women with PCOS (Rotterdam criteria), and it affects hormonal characteristics of these women such as LH and SHBG values. Regarding the impact of isolated hyperandrogenemia on PCOS characteristics, it appears that D4 and testosterone elevations are associated with increased LH levels. Keywords Polycystic ovary syndrome Hyperandrogenemia Testosterone Ovarian volume
Introduction Polycystic ovary syndrome (PCOS) is characterized by a variety of clinical and laboratory manifestations (i.e., menstrual irregularities, infertility, clinical and biochemical hyperandrogenism, ovarian enlargement, abnormal gonadotropin secretory dynamics, obesity, insulin resistance), which do not coexist in all women diagnosed with PCOS [1–3]. It is intriguing, whereas these abnormalities, interacting with each other in different combinations, may affect and result in different phenotypes [4–6]. Specifically, concerning hyperandrogenemia, it is interesting to investigate the impact of specific type of hyperandrogenemia on PCOS phenotypes. Another question that rises is whether the adoption of Rotterdam criteria for the diagnosis of polycystic ovary syndrome modifies the prevalence of overall hyperandrogenemia, since these criteria encompass a broader spectrum of manifestations, including ovarian morphology, which differentiate the population as suffering from PCOS [7]. Regarding the impact of specific androgens on several characteristics of the syndrome, available literature data are limited. It has been found that testosterone levels are positively associated with luteinizing hormone (LH) values
123
Endocrine
as well as follicle number, ovarian volume and stroma amount [8, 9]. Additionally, androstenedione (D4) has been positively correlated with ovarian volume and is associated with better response to PCOS treatment with lifestyle intervention or metformin [10–13]. Finally, regarding dehydroepiandrosterone sulfate (DHEAS), increased levels have been associated with improved metabolic profile, namely better lipid profile and inflammatory markers’ insulin resistance and markers of endothelial function compared to women with lower levels of DHEAS [14–16]. However, these data arise mainly from the evaluation of PCOS women diagnosed according to 1990 NIH criteria, and additionally, the effect of isolated hyperandrogenemia on PCOS phenotype according to currently used Rotterdam criteria is pending. Furthermore, to the best of our knowledge, data derived from large cohorts reporting the prevalence and distribution of hyperandrogenemia in women with PCOS diagnosed with the use of Rotterdam criteria are not available. In an attempt to answer these clinical questions, we conducted the present study, analyzing data originated from a single center of 1,218 women with PCOS, with Caucasian origin and Greek ethnicity. Patients and methods The study group was comprised of 1,218 Greek women with PCOS of reproductive age attending the Human Reproduction Unit in Aristotle University of Thessaloniki. The median age of studied population was 23 years (20/ 28), and their median BMI was 24.59 (21.55/30.12), (numbers in parenthesis indicate the 25th and 75th percentiles, respectively). The inclusion of the patients in this cohort lasted for ten consecutive years. PCOS was defined according to Rotterdam criteria [7]. Participants were diagnosed with PCOS if they had at least two of the following three features present: oligomenorrhea and/or anovulation, clinical or biochemical hyperandrogenemia and morphology of polycystic ovaries in ultrasound. Hyperandrogenemia was defined as the finding of elevated androgens, and the specific cutoff values for testosterone, DHEAS and D4 were 65 ng/dL, 2,800 lg/L and 2.5 ng/ mL, respectively. All women were studied during the follicular phase of their menstrual cycle (days 3–8) or in case of oligomenorrhea if progesterone values confirmed anovulation. In all women, body height, weight, waist and hip circumference were measured and both body mass index (BMI) and waistto-hip ratio (WHR) were estimated. The presence of hirsutism was estimated by Ferriman–Gallwey scale (FG). None of the women studied had galactorrhea or any endocrine or systemic disease that could possibly affect reproductive physiology. Prolactin levels have been measured in all studied subjects. The cutoff points for prolactin,
123
thyroid-stimulating hormone (TSH) and free thyroxine were 30 ng/ml, 0.5–4.70 lIU/mL and 10–23 pmol/L, respectively. No woman reported use of any medication during the last semester that could interfere with the normal function of the hypothalamic–pituitary–gonadal axis, including metformin. When basic 17a-hydroxyprogesterone (17OHP) levels were [1.5 ng/mL, the Synacthen test (Synacthen 0.25 mg/1 ml; Novartis Pharma, Rueil-Malmaison, France) was conducted to rule out congenital adrenal hyperplasia. Other causes of hyperandrogenism, including prolactinoma, Cushing’s syndrome and androgen-secreting tumors, were also excluded. Plasma fasting glucose (mg/dl) was determined by the glucose oxidase color method (Glucose LR, GOD-PAP; Linear Chemicals, Barcelona, Spain). Total cholesterol (mg/dl) was determined by the enzymatic Cobas Mira method (Cholesterol LR, CHOD-PAP; Linear Chemicals). Insulin (lU/ml) was measured by a solid-phase enzymeamplified sensitivity immunoassay (INS-EASIA; Biosource Technologies, Nivelles, Belgium). Total testosterone (ng/dl) was measured by enzyme-linked immunosorbent assay (ELISA) (testosterone enzyme immunoassay test kit, LI7603; Linear Chemicals). Sexhormone-binding globulin (SHBG) serum levels (nmol/l) were measured by ELISA (SHBG ELISA, MX 520 11; IBL, Hamburg, Germany). DHEAS (ng/ml) serum levels were measured by DSL DHEAS radioimmunoassay kit (Diagnostic Systems Laboratories, Webster, TX, USA). LH (IU/l) and FSH (IU/l) were measured using the LHsp and FSH IRMA kits from Biosource Technologies Nivelles, Belgium. D4 (ng/ml) was measured by radioimmunoassay using active androstenedione-coated tube radioimmunoassay kit DSL 3800 (Diagnostic Systems Laboratories). Progesterone (ng/ml) was measured by chemiluminescence immunoassay (Architect, Abbott, IL, USA). The intra- and inter-assay coefficients of variation (CVs) for low and high levels, respectively, were (a) 3.0 and 5.3 % and 4.5 and 9.5 % for insulin, (b) 5.0 and 6.4 % and 4.4 and 8.4 % for total testosterone, (c) 3.0 and 5.3 % and 7.2 and 8.4 % for SHBG, (d) 6.5 and 8.8 % and 3.5 and 4.5 % for LH, (e) 2.7 and 5.3 % and 1.6 and 3.6 % for FSH, (f) 9.4 and 6.3 % and 9.6 and 9.9 % for DHEAS, (g) 5.6 and 2.8 % and 9.8 and 7.0 % for D4, and (h) 3.4 and 5.5 % and 4.7 and 5.6 % for progesterone, respectively. Free androgen index (FAI) was determined as follows: testosterone (nmol/L) 9 100/SHBG (nmol/L). Insulin resistance was estimated by the homeostasis model assessment (HOMA-IR). HOMA-IR is defined as follows: fasting serum insulin (lU/ml) 9 fasting plasma glucose (mmol/l)/22.5. Ultrasound imaging included right and left ovary volume, number of follicles in each ovary, uterus volume and endometrial thickness. It was performed by the same
Endocrine
investigator. Ovary volume was calculated as follows: ovarian volume = (p/6) 9 ovarian length 9 ovarian height 9 ovarian width. Mean ovary volume was estimated as follows: right ovary volume ? left ovary volume/ 2, and the sum of follicles was estimated as follows: the number of follicles in right ovary ? the number of follicles in left ovary. Informed consent was obtained from all women, and the study was approved by the institutional review board. The study met the requirements of the 1975 Helsinki guidelines. Statistical analysis All continuous variables (age, BMI, waist-to-hip ratio, Ferriman–Gallwey score, FSH, LH, LH/FSH, testosterone, D4, DHEAS, SHBG, FAI, glucose, insulin, HOMA-IR, mean ovary volume and sum of follicles) showed not normal distribution as it was documented by the use of the Kolmogorov–Smirnov test. Lann–Whitney test was used to evaluate the differences between hyperandrogenemic and normoandrogenemic PCOS women. The studied population was divided regarding their BMI status to underweight, lean, overweight and obese. In each of the aforementioned BMI groups, the characteristics of hyperandrogenemic women were compared with normoandrogenemic using Mann–Whitney test. Comparison of characteristics between the groups which derived regarding the presence of isolated hyperandrogenemia (GT, GD, GD4 and GP) was performed using Kruskal–Wallis test, and post hoc analysis between two groups was performed using Mann–Whitney test corrected for multiple comparisons (Bonferroni test). The statistic significance was set at p \ 0.05. Statistical analysis was conducted using the statistical package of SPSS 13.00 version.
Results The studied population of 1,218 women with PCOS was divided initially into two groups, regarding the presence of hyperandrogenemia, in women either with elevated or with normal androgens. Seven hundred and sixteen women (58.8 %) displayed hyperandrogenemia, whereas in 502 (41.2 %) subjects, circulating androgens were between normal values. The women with hyperandrogenemia had higher LH and LH/FSH ratio, lower SHBG, lower fasting glucose and higher FAI values. These findings are analytically depicted in Table 1. In order to overcome the impact of BMI on different PCOS characteristics, the subjects were subdivided regarding their BMI status to underweight (BMI \ 20), lean (BMI 20–25), overweight (BMI 25.001–30) and obese (BMI [ 30 kg/m2). In each group, hyperandrogenemic
Table 1 Comparison of PCOS women with elevated androgens (EA) versus normal androgens (NA) in all studied parameters EA (n = 716) BMI (kg/m2) WHR FG score
NA (n = 502)
p
24.71 (21.72/29.86) 24.41 (21.43/30.61) 0.77 (0.73/0.82) 9 (6/12)
0.78 (0.73/0.83) 9 (4/11)
0.843 0.533 \0.001
FSH (U/L)
5.80 (4.69/6.96)
5.75 (3.91/6.90)
0.608
LH (U/L)
7.08 (4.80/10.26)
5.57 (3.91/8.20)
\0.001
LH/FSH
1.24 (0.82/1.84)
0.97 (0.67/1.51)
\0.001
55.85 (41.13/65)
\0.001
Testosterone (ng/dL)
86.40 (75/105)
D4 (ng/ml)
3.19 (2.50/3.90)
2.08 (1.70/2.50)
\0.001
FAI
9.05 (6/13.99)
4.30 (2.71/6.58)
\0.001
41 (29/60.90)
\0.001
SHBG (nmol/l) DHEAS (lg/L)
32 (23/48.15)
3,477 (2,600/4,230) 2,159 (1,580/2,802) \0.001
Glucose (mg/dL)
95 (88/103)
99 (89/106)
0.001
Insulin (lU/ml)
9.54 (6.29/15)
9.38 (6/15.67)
0.891
HOMA-IR
2.18 (1.45/3.60)
2.18 (1.42/3.74)
0.634
Mean ovary volume (cm3)
7.50 (5.36/9.94)
7.50 (5.51/9.90)
0.514
Sum of follicles
22 (15/27)
21 (15/26)
0.138
Data are presented as median. Numbers in brackets indicate the 25th and 75th percentiles Bold values are statistically significant (p \ 0.05)
women were compared with normoandrogenemic women on the same anthropometric, hormonal, metabolic and ultrasound characteristics. It was found that hyperandrogenemic women displayed higher LH and LH/FSH and lower SHBG values, compared to normoandrogenemic women, regardless of their BMI status. It was also found that lean and obese hyperandrogenemic women displayed lower fasting glucose values. Overweight hyperandrogenemic women displayed lower mean ovary volume. These findings are analytically depicted in Table 2. Subsequently, the studied population was divided into four groups regarding the presence of isolated hyperandrogenemia and the following four groups are derived: GT [women with isolated increased testosterone serum levels (C75 ng/dL), n = 210], GD [women with isolated increased DHEAS serum levels (C3,600 lg/L), n = 92], GD4 [women with isolated increased D4 serum levels (C3.4 ng/mL), n = 50] and GP (women with normal D4, testosterone and DHEAS serum levels, n = 502). Three hundred and sixty-four women (n = 364) displayed mixed hyperandrogenemia, since more than one androgens were elevated in these subjects. Accordingly, the most frequent
123
123
0.95 (0.64/1.40)
1.47 (0.84/2.30)
1.44 (1/2.11)
6.25 (4.62/9.09)
1.54 (1.06/2.12)
6.62 (5.16/9.25)
22 (17/26.50)
HOMA-IR
Mean ovary volume (cm3)
Sum of follicles
0.165
0.619
0.66
0.90
0.871
\0.001
0.012
20 (14/27)
7.69 (5.41/9.90)
1.8 (1.26/2.54)
8 (5.68/10.86)
94 (87/101)
7.36 (5.12/11.17)
38.04 (28.80/55)
Bold values are statistically significant (p \ 0.05)
EA elevated androgens, NA normal androgens
21 (15/26)
7.16 (5.45/9.80)
1.76 (1.24/2.40)
7 (5.39/10.6)
98 (89/105)
3.47 (2.27/5.11)
52.32 (38.2/71.6)
Data are presented as median. Numbers in brackets indicate the 25th and 75th percentiles
21 (13/26)
6.30 (5/9)
6.80 (4.95/9.20)
Insulin (lU/ml)
96 (85.50/103)
3.07 (1.80/4.68)
96 (86/102)
6.26 (4.48/8.91)
Glucose (mg/dl)
FAI
58 (41.3/77.40)
46.70 (35.50/69.8)
2.1 (1.76/2.55)
54 (42/64.60)
1.01 (0.71/1.73)
6.30 (4.42/9.30)
5.95 (4.7/7.13)
9 (4/10)
0.75 (0.71/0.78)
NA (n = 207)
3,475 (2,653/4,081) 2,220 1,565/2,840
3.23 (2.57/4.04)
\0.001
SHBG (nmol/l)
2.2 (1.84/2.54)
85 (74.95/101)
1.24 (0.87/2)
\0.001 \0.001
5.93 (4.80/7.21) 7.50 (5.25/11.26)
0.192
9 (6/11)
0.75 (0.71/0.78)
\0.001
3,070 (2,216/3,973) 1,931 (1,513/2,420)\0.001
3 (2.40/3.87)
84.90 (74/105.50) 51.90 (34.50/66)
0.729 0.973
DHEAS (lg/l)
D4 (ng/ml)
Testosterone (ng/dl)
LH/FSH
5.54 (3.54/8.37)
LH (U/L)
5.80 (4.94/6.70)
6.37 (5.03/7.19)
8.48 (5.57/13.25)
FSH (U/L)
6 (0/10)
0.73 (0.71/0.75)
6 (1.5/10)
0.73 (0.70/0.76)
EA (n = 293)
p
EA (n = 85)
NA (n = 61)
20–25 kg/m2
\20 kg/m2
FG score
WHR
BMI range
0.942
29 (21.62/40)
\0.001
0.491
0.218
0.483
0.097
0.001
22 (15/28)
7.41 (5.12/9.68)
2.30 (1.52/3.70)
10.28 (7/15)
94 (87/104)
10.41 (7.11/15.31)
3,765 (2,810/4,520)
\0.001
\0.001
3.23 (2.60/3.86)
89.34 (75/109)
1.16 (0.80/1.69)
6.12 (4.50/9.40)
5.52 (4.53/6.70)
10 (7/13)
0.79 (0.76/0.82)
EA (n = 163)
\0.001
\0.001
0.009
0.001
0.518
\0.001
p
25.001–30 kg/m2
20 (15.5/26)
8.37 (6.93/10.66)
2.53 (1.86/3.76)
10.90 (7.21/15.02)
96 (87/104.35)
5.61 (3.57/7.51)
35.30 (27/48.35)
2,235 (1,615/2,818)
2.08 (1.71/2.54)
59 (43.50/65.50)
0.98 (0.65/1.52)
5.30 (3.90/8.15)
5.70 (4.61/6.95)
10 (4/12)
0.80 (0.76/0.85)
NA (n = 101) 0.05
0.026
0.012
0.591
0.139
0.325
0.004
0.372
0.431
0.741
\0.001
0.002
\0.001
\0.001
\0.001
p
22 (17/27)
7.92 (5.75/10.80)
4.32 (2.73/6.17)
18 (11.30/24.3)
98 (91/105)
13.40 (9.83/20.82)
22.40 (17.57/30.36)
3,441 (2,503/4,306)
3 (2.35/3.70)
92 (77/110)
1.23 (0.81/1.76)
6.52 (3.95/9.50)
5.70 (4.40/6.40)
10 (7/14)
0.84 (0.79/0.90)
EA (n = 175)
[30 kg/m2
20 (14/26)
6.85 (5.58/9.47)
4.38 (2.81/6.31)
17.90 (11.87/25.7)
101 (93.50/111)
6.53 (4.06/9.29)
28 (20.04/38.5)
2,100 (1,580/2,815)
1.98 (1.60/2.42)
58.50 (41/66.10)
0.87 (0.62/1.31)
5 (3.38/6.91)
5.50 (4.50/6.70)
9 (5.5/11)
0.83 (0.80/0.87)
NA (n = 133)
Table 2 Comparison of women with elevated androgens (EA) versus women with normal serum androgens (NA) according to their BMI status in all studied parameters
0.68
0.875
0.002
0.05
0.05
0.367
0.579
0.046
\0.001
\0.001
\0.001
\0.001
\0.001
\0.001
\0.001
p
Endocrine
Endocrine Fig. 1 The percentage of isolated and combined hyperandrogenemia in 1,218 women with PCOS. Normal serum androgens were found in 41.2, 29.9 % had combined hyperandrogenemia, and isolated elevated testosterone, DHEAS and D4 were found in 17.2, 7.6 and 4.1 %, respectively
abnormality as the sole marker of hyperandrogenemia was isolated elevated testosterone levels (17.2 %), followed by DHEAS (7.6 %) and D4 (4.1 %). In 364 women (29.9 %), combined hyperandrogenemia was disclosed, since more than one androgens were elevated. These findings are clearly illustrated in Fig. 1. From the comparisons between subgroups, it was found that LH values and LH/FSH ratio were higher in GD4 and GT groups (Table 3).
Discussion In the present study, it was found that in a very large cohort (n = 1,218) of women with PCOS diagnosed with Rotterdam criteria, (1) the overall prevalence of hyperandrogenemia is 58.8 %, and testosterone levels constituting the predominant form in women with isolated hyperandrogenemia, and (2) that LH values were higher in women with either elevated testosterone (GT) or D4 (GD4) levels compared to those with raised DHEAS (GD) levels or those with normal androgens (GP). Obesity, a key player of PCOS characteristics development [17], did not appear to be an important denominator of different phenotypes between normo- and hyperandrogenemic PCOS women. This finding suggests that hyperandrogenemia consists of an important determinant of different PCOS phenotypes, per se. Concerning biochemical hyperandrogenism prevalence, our findings are different from studies using the other available criteria for PCOS diagnosis. Huang et al. estimated that the overall prevalence of hyperandrogenemia in
women diagnosed with PCOS according to NIH criteria was 75.3 %, and Androgen Excess Society guideline reported that elevated circulating androgens levels are observed in about 60–80 % of population [18, 19]. In our population, we estimated that the prevalence of hyperandrogenemia is 58.8 %, lower than the aforementioned rates. However, in the aforementioned study except the fact that diagnosis was based on NIH criteria, the study group was comprised of obese women (mean BMI 33.39 ± 9.26 kg/m2), with different ethnicity and race (14 % of them were Afro-American), while in the present study, women were divided according to the degree of obesity and belonged to the same ethnicity. Regarding the prevalence of isolated hyperandrogenemia in women with PCOS, available data are scanty. In the present study, isolated elevated testosterone levels were found in a rate of 17.2 %. In a study carried out in Brazil, 44 % of women with PCOS according to Rotterdam criteria displayed elevated testosterone values [20]. Although the studied population was comparable to the one evaluated in the present study, regarding age and BMI status, there was a mixed population on ethnicity and race. In addition, it has not indicated whether increased testosterone levels constitute the sole hyperandrogenemic abnormality or whether there was a combination of any kind between elevated testosterone and other androgens. In the aforementioned NIH cohort, elevated levels of testosterone as the sole abnormality were found in only 2 % of patients [18]. Regarding LH values, we have noticed that LH levels were higher in women with isolated testosterone or D4
123
Endocrine Table 3 Comparison of women with isolated increased testosterone (GT), DHEAS (GD), D4 (GD4) and women with normal androgens (GP)
BMI (kg/m2) WHR FG score
GT (n = 210)
GD (n = 92)
GD4 (n = 50)
GP (n = 502)
24.78 (21.27/31.89)
24.46 (21.48/29.08)
23.49 (21.65/29.87)
24.41 (21.43/30.61)
0.77 (0.73/0.82)
0.78 (0.73/0.83)
0.79 (0.75/0.84)
0.78 (0.73/0.83)
9 (4.75/11)
10 (4/12)
FSH (U/L)
5.89 (4.67/6.80)
5.77 (4.45/6.98)
LH* (U/L)
7.13 (4.80/9.90)
5.37 (3.77/7.62)
LH/FSH^
1.23 (0.84/1.81)
Testosterone (ng/dl)# D4 (ng/ml)[
2.64 (2.20/2.99)
DHEAS (lg/l)\
2,565 (2,063/3,095)
##
FAI
7.93 (5.61/11.89)
SHBG** (nmol/l) Glucose (mg/dl)^^
36.81 (26.54/55) 94.5 (85/105)
Insulin (lU/ml)
9 (5.93/14.81)
HOMA-IR 3
Ovary volume (cm ) Follicle sum
86 (80/96)
9 (5.75/12) 6 (4.88/7.43) 9.39 (6.15/12.58)
p 0.737 0.447
9 (4/11)
0.192
5.75 (4.68/6.90)
0.764
5.57 (3.91/8.20)
\0.001 \0.001
0.97 (0.66/1.24)
1.57 (0.96/2.31)
0.97 (0.67/1.51)
63.60 (61/68.78)
64.50 (52.75/70)
55.85 (41.13/65)
\0.001
2.43 (2.1/2.82)
3.90 (3.56/4.24)
2.08 (1.70/2.50)
\0.001
2,770 (2,148/3,131)
2,159 (1,580/2,802)
\0.001
7.11 (5.30/9.10)
5.08 (3.66/7.33)
4.30 (2.71/6.58)
\0.001
31.60 (23.73/40.30) 96 (88/103)
37.50 (26.88/51.14) 97 (91/103.50)
41 (29/60.90) 99 (89/106)
\0.001 0.03
4,058 (3,835/4,646)
9.27 (5.90/15.65)
9.18 (7/15.51)
9.38 (6/15.67)
0.82
2.06 (1.28/3.55)
2.21 (1.38/3.93)
2.22 (1.48/3.98)
2.18 (1.42/3.74)
0.519
7.52 (5.40/10.21)
7.15 (5.19/8.22)
7.40 (4.67/10.30)
7.5 (5.51/9.90)
0.109
21 (15/26)
0.513
21 (15/27)
21 (15/26)
22 (17.75/28.50)
Data are presented as median. Numbers in brackets indicate the 25th and 75th percentiles Post hoc analysis with Mann–Whitney test corrected for multiple comparisons (Bonferroni test) * All groups differ statistically significantly among them (p \ 0.001; p = 0.027 in GD4 vs. GT group), except GD versus GP group (p = 1.353) ^
All groups differ statistically significantly among them (p \ 0.001), except GD4 versus GT group (p = 0.117) and GD versus GP group (p = 0.957)
#
All groups differ statistically significantly among them (p \ 0.001. GD4 vs. GP p = 0.006), except GD4 versus GD group (p = 1.77)
[
All groups differ statistically significantly among them (p \ 0.001), except GT versus GD group (p = 0.147)
\
All groups differ statistically significantly among them (p \ 0.001), except GT versus GD4 group (p = 0.795)
##
All groups differ statistically significantly among them (p \ 0.001, GD4 vs. GD p = 0.003), except GT versus GD group and GD4 versus GP group (p = 0.06) ** Only GD versus GP group (p \ 0.001) and GT versus GD group (p = 0.009) differ statistically significantly among four groups
^^
Only GT versus GP group (p = 0.015) differs statistically significantly among four groups
Bold values are statistically significant (p \ 0.05)
elevated values compared to women with isolated elevated DHEAS or with women without hyperandrogenemia. This constitutes a finding not encountered during our literature review. It is known that in women with PCOS, a high prevalence of abnormal gonadotropin secretion has been reported, which consists of an increase in both the amplitude and frequency of LH secretion. It has been hypothesized that the frequency of GnRH pulses determines which gonadotropin hormone is synthesized and secreted with rapid GnRH pulses favoring LH secretion and this disturbed machinery has been implicated in the pathogenesis of the syndrome, through the enhanced androgen secretion from the stimulated ovarian theca cell [21–23]. However, recent data suggest a vicious circle between LH amplitude and androgen ovarian secretion, since Solorzano et al. [24] reported that hyperandrogenemia may adversely affect LH pulse regulation during pubertal maturation leading to persistent hyperandrogenemia and the development of
123
PCOS. Rather, an intrinsic dysregulation in the steroidogenic pathways of ovarian cells may result in a condition in which the set point for desensitization to LH is set higher [25–27]. Our findings suggest that the specific type of hyperandrogenemia may contribute to this disturbed cycle between androgens and LH secretion, since women with either testosterone or D4 elevated levels and not those with elevated DHEAS levels or normal androgens displayed higher values of LH. Interestingly, SHBG values were lower in subgroups of women with hyperandrogenemia, regardless of their BMI status. This fascinating finding cannot be attributed to the well-described insulin action on SHBG liver synthesis, since insulin concentrations were comparable among subgroups (Table 3), thus enhancing the notion that SHBG regulation is multifactorial, with hyperandrogenemia consisting of a possible factor of SHBG regulation [28–30]. It is also interesting that different phenotypes between
Endocrine
normoandrogenemic and hyperandrogenemic PCOS women do not seem to be affected by BMI, since syndrome’s manifestation such as LH and SHBG levels differs significantly between them—either they were normal, overweight or obese. The finding of lower fasting glucose levels in women with hyperandrogenemia in comparison with women with PCOS with normal circulating androgens (Table 1) is intriguing. Although in absolute values the difference is very small (95 vs. 99 mg/dl), further research is required to elucidate any potential role of androgens in glycemia intraPCOS group, since this aspect has not been reported before. This study has two major limitations: the cross-sectional setting and the long time needed to collect these data. On the other hand, the large number of studied subjects and their categorization according to BMI status strengthen the findings of the present study. In conclusion, our findings suggest that in women with polycystic ovary syndrome diagnosed with Rotterdam criteria, the prevalence of overall hyperandrogenemia is lower and the prevalence of isolated elevated testosterone is higher compared with women diagnosed with 1990 NIH criteria. Additionally, elevations of specific circulating androgens (specifically D4 and testosterone elevations) may modify LH levels. The above findings as a whole suggest that the type of hyperandrogenemia may play a clinically significant role in PCOS diagnosis. Conflict of interest
All authors have nothing to disclose.
References 1. E. Diamanti-Kandarakis, Polycystic ovarian syndrome: pathophysiology, molecular aspects and clinical implications. Expert Rev. Mol. Med. 3010, e3 (2008) 2. E. Diamanti-Kandarakis, A. Dunaif, Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications. Endocr. Rev. 33, 981–1030 (2012) 3. E. Kandaraki, C. Christakou, E. Diamanti-Kandarakis, Metabolic syndrome and polycystic ovary syndrome… and vice versa. Arq. Bras. Endocrinol. Metabol. 53, 227–237 (2009) 4. E. Diamanti-Kandarakis, D. Panidis, Unravelling the phenotypic map of polycystic ovary syndrome (PCOS): a prospective study of 634 women with PCOS. Clin. Endocrinol. (Oxf.) 67, 735–742 (2007) 5. E. Diamanti-Kandarakis, C. Christakou, E. Marinakis, Phenotypes and environmental factors: their influence in PCOS. Curr. Pharm. Des. 18, 270–282 (2012) 6. S.F. Witchel, S.E. Recabarren, F. Gonza´lez, E. Diamanti-Kandarakis, K.I. Cheang, A.J. Duleba, R.S. Legro, R. Homburg, R. Pasquali, R.A. Lobo, C.C. Zouboulis, F. Kelestimur, F. Fruzzetti, W. Futterweit, R.J. Norman, D.H. Abbott, Emerging concepts about prenatal genesis, aberrant metabolism and treatment paradigms in polycystic ovary syndrome. Endocrine 42, 526–534 (2012)
7. 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). Fertil. Steril. 81, 19–25 (2004) 8. M.A. Lawson, S. Jain, S. Sun, K. Patel, P.J. Malcolm, R.J. Chang, Evidence for insulin suppression of baseline luteinizing hormone in women with polycystic ovarian syndrome and normal women. J. Clin. Endocrinol. Metab. 93, 2089–2096 (2008) 9. T.D. Pache, F.H. de Jong, W.C. Hop, B.C. Fauser, Association between ovarian changes assessed by transvaginal sonography and clinical and endocrine signs of the polycystic ovary syndrome. Fertil. Steril. 59, 544–549 (1993) 10. A.A. Kyei-Mensah, S. LinTan, J. Zaidi, H.S. Jacobs, Relationship of ovarian stromal volume to serum androgen concentrations in patients with polycystic ovary syndrome. Hum. Reprod. 13, 1437–1441 (1998) 11. Z. Puzigac´a, G.M. Prelevic´, Z. Stretenovic´, L. Balint-Peric´, Ovarian enlargement as a possible marker of androgen activity in polycystic ovary syndrome. Gynecol. Endocrinol. 5, 167–174 (1991) 12. R. Pasquali, A. Gambineri, C. Cavazza, D. Ibarra Gasparini, W. Ciampaglia, G.E. Cognigni, U. Pagotto, Heterogeneity in the responsiveness to long-term lifestyle intervention and predictability in obese women with polycystic ovary syndrome. Eur. J. Endocrinol. 164, 53–60 (2011) 13. P. Moghetti, R. Castello, C. Negri, F. Tosi, F. Perrone, M. Caputo, E. Zanolin, M. Muggeo, Metformin effects on clinical features, endocrine and metabolic profiles, and insulin sensitivity in polycystic ovary syndrome: a randomized, double-blind, placebo controlled 6-month trial, followed by open long-term. J. Clin. Endocrinol. Metab. 85, 139–146 (2000) ¨ . Turhan, Impact 14. N. Ko¨s¸ u¨s¸ , A. Ko¨s¸ u¨s¸ , Z. Kamalak, D. Hızlı, N.O of adrenal versus ovarian androgen ratio on signs and symptoms of polycystic ovarian syndrome. Gynecol. Endocrinol. 28, 611–614 (2012) 15. K. Brennan, A. Huang, R. Azziz, Dehydroepiandrosterone sulfate and insulin resistance in patients with polycystic ovary syndrome. Fertil. Steril. 91, 1848–1852 (2009) 16. A. Dagre, J. Lekakis, C. Mihas, A. Protogerou, L. Thalassinou, D. Tryfonopoulos, G. Douridas, C. Papamichael, M. Alevizaki, Association of dehydroepiandrosterone-sulfate with endothelial function in young women with polycystic ovary syndrome. Eur. J. Endocrinol. 154, 883–890 (2006) 17. D. Panidis, K. Tziomalos, E. Papadakis, C. Vosnakis, P. Chatzis, I. Katsikis, Lifestyle intervention and anti-obesity therapies in the polycystic ovary syndrome: impact on metabolism and fertility. Endocrine 44, 583–590 (2013) 18. A. Huang, K. Brennan, R. Azziz, Prevalence of hyperandrogenemia in the polycystic ovary syndrome diagnosed by the NIH 1990 criteria. Fertil. Steril. 93, 1938–1941 (2010) 19. R. Azziz, E. Carmina, D. Dewailly, E. Diamanti-Kandarakis, H.F. Escobar-Morreale, W. Futterweit, O.E. Janssen, R.S. Legro, R.J. Norman, A.E. Taylor, S.F. Witchel, Task force on the phenotype of the polycystic ovary syndrome of the androgen excess and PCOS Society. Criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: an androgen excess society guideline. Fertil. Steril. 9, 456–488 (2009) 20. L. Gabrielli, E.M. Aquino, Polycystic ovary syndrome in Salvador, Brazil: a prevalence study in primary healthcare. Reprod. Biol. Endocrinol. 10, 96 (2012) 21. E. Diamanti-Kandarakis, J. Papailiou, S. Palimeri, Hyperandrogenemia: pathophysiology and its role in ovulatory dysfunction in PCOS. Pediatr. Endocrinol. Rev. 3, 198–204 (2006) 22. S.K. Blank, C.R. McCartney, J.C. Marshall, The origins and sequelae of abnormal neuroendocrine function in polycystic ovary syndrome. Hum. Reprod. Update 12, 351–361 (2006)
123
Endocrine 23. A.E. Taylor, B. McCourt, K.A. Martin, E.J. Anderson, J.M. Adams, D. Schoenfeld, J.E. Hall, Determinants of abnormal gonadotropin secretion in clinically defined women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 82, 2248–2256 (1997) 24. C.M. Solorzano, J.P. Beller, M.Y. Abshire, J.S. Collins, C.R. McCartney, J.C. Marshall, Neuroendocrine dysfunction in polycystic ovary syndrome. Steroids 77, 332–337 (2012) 25. A.J. Jakimiuk, S.R. Weitsman, A. Navab, D.A. Magoffin, Luteinizing hormone receptor, steroidogenesis acute regulatory protein, and steroidogenic enzyme messenger ribonucleic acids are overexpressed in thecal and granulosa cells from polycystic ovaries. J. Clin. Endocrinol. Metab. 86, 1318–1323 (2001) 26. V.L. Nelson, R.S. Legro, J.M. McAllister, Augmented androgen production is a stable steroidogenic phenotype of propagated theca cells from polycystic ovaries. Mol. Endocrinol. 13, 946–957 (1999)
123
27. R.L. Rosenfield, B. Bordini, Evidence that obesity and androgens have independent and opposing effects on gonadotropin production from puberty to maturity. Brain Res. 1364, 186–197 (2010) 28. V. Toscano, R. Balducci, P. Bianchi, R. Guglielmi, A. Mangiantini, F. Sciarra, Steroidal and non-steroidal factors in plasma sex hormone binding globulin regulation. J. Steroid Biochem. Mol. Biol. 43, 431–437 (1992) 29. M.G. Forest, A. Lecoq, M. David, M. Pugeat, Effects of human chorionic gonadotropin, androgens, adrenocorticotropin hormone, dexamethasone and hyperprolactinemia on plasma sex steroid-binding protein. Ann. N. Y. Acad. Sci. 538, 214–234 (1988) 30. F. Gonza´lez, C.L. Sia, F.Z. Stanczyk, H.E. Blair, M.E. Krupa, Hyperandrogenism exerts an anti-inflammatory effect in obese women with polycystic ovary syndrome. Endocrine 42, 726–735 (2012)
ORIGINAL ARTICLE
Prevalence and impact of hyperandrogenemia in 1,218 women with polycystic ovary syndrome Sarantis Livadas • Christos Pappas • Athanasios Karachalios • Evangelos Marinakis • Nikoleta Tolia • Maria Drakou • Philippos Kaldrymides • Dimitrios Panidis • Evanthia Diamanti-Kandarakis
Received: 1 October 2013 / Accepted: 31 January 2014 Springer Science+Business Media New York 2014
Abstract Hyperandrogenemia modifies phenotypic characteristics of women with polycystic ovary syndrome (PCOS). The aim of the present study is to evaluate (a) the prevalence of hyperandrogenemia in PCOS women (Rotterdam criteria) and (b) the impact of either the degree or the type of hyperandrogenemia on phenotype. Anthropometric, clinical, hormonal, metabolic and ultrasound characteristics of 1,218 women with PCOS were analyzed in this cross-sectional study. The prevalence of hyperandrogenemia was 58.8 %. Women with hyperandrogenemia had higher luteinizing hormone (LH), follicle-stimulating hormone (FSH), free androgen index, lower sex-hormonebinding globulin (SHBG) and fasting glucose levels compared to women with normal androgens (p \ 0.001 for all comparisons; p = 0.001 for fasting glucose). Regarding the presence of isolated hyperandrogenemia, the group with only elevated testosterone levels was termed GT and an analogous categorization was made for dehydroepiandrosterone sulfate (GD) and androstenedione (D4) (GD4), respectively. GT, GD and GD4 comprised the 17.2, 7.6 and 4.1 % of total cohort, respectively. These groups differed significantly between them in LH, LH/FSH ratio, and SHBG (p \ 0.001). Hyperandrogenemia is found in almost S. Livadas C. Pappas A. Karachalios E. Marinakis N. Tolia M. Drakou P. Kaldrymides E. Diamanti-Kandarakis (&) Endocrine Unit, Third Department of Internal Medicine, Medical School, National and Kapodistrian University of Athens, Mesogeion 152, 11527 Athens, Greece e-mail: [email protected]; [email protected] D. Panidis Division of Endocrinology and Human Reproduction, Second Department of Obstetrics and Gynecology, Hippokration Hospital, Aristotle University of Thessaloniki, Thessalonı´ki, Greece
60 % of women with PCOS (Rotterdam criteria), and it affects hormonal characteristics of these women such as LH and SHBG values. Regarding the impact of isolated hyperandrogenemia on PCOS characteristics, it appears that D4 and testosterone elevations are associated with increased LH levels. Keywords Polycystic ovary syndrome Hyperandrogenemia Testosterone Ovarian volume
Introduction Polycystic ovary syndrome (PCOS) is characterized by a variety of clinical and laboratory manifestations (i.e., menstrual irregularities, infertility, clinical and biochemical hyperandrogenism, ovarian enlargement, abnormal gonadotropin secretory dynamics, obesity, insulin resistance), which do not coexist in all women diagnosed with PCOS [1–3]. It is intriguing, whereas these abnormalities, interacting with each other in different combinations, may affect and result in different phenotypes [4–6]. Specifically, concerning hyperandrogenemia, it is interesting to investigate the impact of specific type of hyperandrogenemia on PCOS phenotypes. Another question that rises is whether the adoption of Rotterdam criteria for the diagnosis of polycystic ovary syndrome modifies the prevalence of overall hyperandrogenemia, since these criteria encompass a broader spectrum of manifestations, including ovarian morphology, which differentiate the population as suffering from PCOS [7]. Regarding the impact of specific androgens on several characteristics of the syndrome, available literature data are limited. It has been found that testosterone levels are positively associated with luteinizing hormone (LH) values
123
Endocrine
as well as follicle number, ovarian volume and stroma amount [8, 9]. Additionally, androstenedione (D4) has been positively correlated with ovarian volume and is associated with better response to PCOS treatment with lifestyle intervention or metformin [10–13]. Finally, regarding dehydroepiandrosterone sulfate (DHEAS), increased levels have been associated with improved metabolic profile, namely better lipid profile and inflammatory markers’ insulin resistance and markers of endothelial function compared to women with lower levels of DHEAS [14–16]. However, these data arise mainly from the evaluation of PCOS women diagnosed according to 1990 NIH criteria, and additionally, the effect of isolated hyperandrogenemia on PCOS phenotype according to currently used Rotterdam criteria is pending. Furthermore, to the best of our knowledge, data derived from large cohorts reporting the prevalence and distribution of hyperandrogenemia in women with PCOS diagnosed with the use of Rotterdam criteria are not available. In an attempt to answer these clinical questions, we conducted the present study, analyzing data originated from a single center of 1,218 women with PCOS, with Caucasian origin and Greek ethnicity. Patients and methods The study group was comprised of 1,218 Greek women with PCOS of reproductive age attending the Human Reproduction Unit in Aristotle University of Thessaloniki. The median age of studied population was 23 years (20/ 28), and their median BMI was 24.59 (21.55/30.12), (numbers in parenthesis indicate the 25th and 75th percentiles, respectively). The inclusion of the patients in this cohort lasted for ten consecutive years. PCOS was defined according to Rotterdam criteria [7]. Participants were diagnosed with PCOS if they had at least two of the following three features present: oligomenorrhea and/or anovulation, clinical or biochemical hyperandrogenemia and morphology of polycystic ovaries in ultrasound. Hyperandrogenemia was defined as the finding of elevated androgens, and the specific cutoff values for testosterone, DHEAS and D4 were 65 ng/dL, 2,800 lg/L and 2.5 ng/ mL, respectively. All women were studied during the follicular phase of their menstrual cycle (days 3–8) or in case of oligomenorrhea if progesterone values confirmed anovulation. In all women, body height, weight, waist and hip circumference were measured and both body mass index (BMI) and waistto-hip ratio (WHR) were estimated. The presence of hirsutism was estimated by Ferriman–Gallwey scale (FG). None of the women studied had galactorrhea or any endocrine or systemic disease that could possibly affect reproductive physiology. Prolactin levels have been measured in all studied subjects. The cutoff points for prolactin,
123
thyroid-stimulating hormone (TSH) and free thyroxine were 30 ng/ml, 0.5–4.70 lIU/mL and 10–23 pmol/L, respectively. No woman reported use of any medication during the last semester that could interfere with the normal function of the hypothalamic–pituitary–gonadal axis, including metformin. When basic 17a-hydroxyprogesterone (17OHP) levels were [1.5 ng/mL, the Synacthen test (Synacthen 0.25 mg/1 ml; Novartis Pharma, Rueil-Malmaison, France) was conducted to rule out congenital adrenal hyperplasia. Other causes of hyperandrogenism, including prolactinoma, Cushing’s syndrome and androgen-secreting tumors, were also excluded. Plasma fasting glucose (mg/dl) was determined by the glucose oxidase color method (Glucose LR, GOD-PAP; Linear Chemicals, Barcelona, Spain). Total cholesterol (mg/dl) was determined by the enzymatic Cobas Mira method (Cholesterol LR, CHOD-PAP; Linear Chemicals). Insulin (lU/ml) was measured by a solid-phase enzymeamplified sensitivity immunoassay (INS-EASIA; Biosource Technologies, Nivelles, Belgium). Total testosterone (ng/dl) was measured by enzyme-linked immunosorbent assay (ELISA) (testosterone enzyme immunoassay test kit, LI7603; Linear Chemicals). Sexhormone-binding globulin (SHBG) serum levels (nmol/l) were measured by ELISA (SHBG ELISA, MX 520 11; IBL, Hamburg, Germany). DHEAS (ng/ml) serum levels were measured by DSL DHEAS radioimmunoassay kit (Diagnostic Systems Laboratories, Webster, TX, USA). LH (IU/l) and FSH (IU/l) were measured using the LHsp and FSH IRMA kits from Biosource Technologies Nivelles, Belgium. D4 (ng/ml) was measured by radioimmunoassay using active androstenedione-coated tube radioimmunoassay kit DSL 3800 (Diagnostic Systems Laboratories). Progesterone (ng/ml) was measured by chemiluminescence immunoassay (Architect, Abbott, IL, USA). The intra- and inter-assay coefficients of variation (CVs) for low and high levels, respectively, were (a) 3.0 and 5.3 % and 4.5 and 9.5 % for insulin, (b) 5.0 and 6.4 % and 4.4 and 8.4 % for total testosterone, (c) 3.0 and 5.3 % and 7.2 and 8.4 % for SHBG, (d) 6.5 and 8.8 % and 3.5 and 4.5 % for LH, (e) 2.7 and 5.3 % and 1.6 and 3.6 % for FSH, (f) 9.4 and 6.3 % and 9.6 and 9.9 % for DHEAS, (g) 5.6 and 2.8 % and 9.8 and 7.0 % for D4, and (h) 3.4 and 5.5 % and 4.7 and 5.6 % for progesterone, respectively. Free androgen index (FAI) was determined as follows: testosterone (nmol/L) 9 100/SHBG (nmol/L). Insulin resistance was estimated by the homeostasis model assessment (HOMA-IR). HOMA-IR is defined as follows: fasting serum insulin (lU/ml) 9 fasting plasma glucose (mmol/l)/22.5. Ultrasound imaging included right and left ovary volume, number of follicles in each ovary, uterus volume and endometrial thickness. It was performed by the same
Endocrine
investigator. Ovary volume was calculated as follows: ovarian volume = (p/6) 9 ovarian length 9 ovarian height 9 ovarian width. Mean ovary volume was estimated as follows: right ovary volume ? left ovary volume/ 2, and the sum of follicles was estimated as follows: the number of follicles in right ovary ? the number of follicles in left ovary. Informed consent was obtained from all women, and the study was approved by the institutional review board. The study met the requirements of the 1975 Helsinki guidelines. Statistical analysis All continuous variables (age, BMI, waist-to-hip ratio, Ferriman–Gallwey score, FSH, LH, LH/FSH, testosterone, D4, DHEAS, SHBG, FAI, glucose, insulin, HOMA-IR, mean ovary volume and sum of follicles) showed not normal distribution as it was documented by the use of the Kolmogorov–Smirnov test. Lann–Whitney test was used to evaluate the differences between hyperandrogenemic and normoandrogenemic PCOS women. The studied population was divided regarding their BMI status to underweight, lean, overweight and obese. In each of the aforementioned BMI groups, the characteristics of hyperandrogenemic women were compared with normoandrogenemic using Mann–Whitney test. Comparison of characteristics between the groups which derived regarding the presence of isolated hyperandrogenemia (GT, GD, GD4 and GP) was performed using Kruskal–Wallis test, and post hoc analysis between two groups was performed using Mann–Whitney test corrected for multiple comparisons (Bonferroni test). The statistic significance was set at p \ 0.05. Statistical analysis was conducted using the statistical package of SPSS 13.00 version.
Results The studied population of 1,218 women with PCOS was divided initially into two groups, regarding the presence of hyperandrogenemia, in women either with elevated or with normal androgens. Seven hundred and sixteen women (58.8 %) displayed hyperandrogenemia, whereas in 502 (41.2 %) subjects, circulating androgens were between normal values. The women with hyperandrogenemia had higher LH and LH/FSH ratio, lower SHBG, lower fasting glucose and higher FAI values. These findings are analytically depicted in Table 1. In order to overcome the impact of BMI on different PCOS characteristics, the subjects were subdivided regarding their BMI status to underweight (BMI \ 20), lean (BMI 20–25), overweight (BMI 25.001–30) and obese (BMI [ 30 kg/m2). In each group, hyperandrogenemic
Table 1 Comparison of PCOS women with elevated androgens (EA) versus normal androgens (NA) in all studied parameters EA (n = 716) BMI (kg/m2) WHR FG score
NA (n = 502)
p
24.71 (21.72/29.86) 24.41 (21.43/30.61) 0.77 (0.73/0.82) 9 (6/12)
0.78 (0.73/0.83) 9 (4/11)
0.843 0.533 \0.001
FSH (U/L)
5.80 (4.69/6.96)
5.75 (3.91/6.90)
0.608
LH (U/L)
7.08 (4.80/10.26)
5.57 (3.91/8.20)
\0.001
LH/FSH
1.24 (0.82/1.84)
0.97 (0.67/1.51)
\0.001
55.85 (41.13/65)
\0.001
Testosterone (ng/dL)
86.40 (75/105)
D4 (ng/ml)
3.19 (2.50/3.90)
2.08 (1.70/2.50)
\0.001
FAI
9.05 (6/13.99)
4.30 (2.71/6.58)
\0.001
41 (29/60.90)
\0.001
SHBG (nmol/l) DHEAS (lg/L)
32 (23/48.15)
3,477 (2,600/4,230) 2,159 (1,580/2,802) \0.001
Glucose (mg/dL)
95 (88/103)
99 (89/106)
0.001
Insulin (lU/ml)
9.54 (6.29/15)
9.38 (6/15.67)
0.891
HOMA-IR
2.18 (1.45/3.60)
2.18 (1.42/3.74)
0.634
Mean ovary volume (cm3)
7.50 (5.36/9.94)
7.50 (5.51/9.90)
0.514
Sum of follicles
22 (15/27)
21 (15/26)
0.138
Data are presented as median. Numbers in brackets indicate the 25th and 75th percentiles Bold values are statistically significant (p \ 0.05)
women were compared with normoandrogenemic women on the same anthropometric, hormonal, metabolic and ultrasound characteristics. It was found that hyperandrogenemic women displayed higher LH and LH/FSH and lower SHBG values, compared to normoandrogenemic women, regardless of their BMI status. It was also found that lean and obese hyperandrogenemic women displayed lower fasting glucose values. Overweight hyperandrogenemic women displayed lower mean ovary volume. These findings are analytically depicted in Table 2. Subsequently, the studied population was divided into four groups regarding the presence of isolated hyperandrogenemia and the following four groups are derived: GT [women with isolated increased testosterone serum levels (C75 ng/dL), n = 210], GD [women with isolated increased DHEAS serum levels (C3,600 lg/L), n = 92], GD4 [women with isolated increased D4 serum levels (C3.4 ng/mL), n = 50] and GP (women with normal D4, testosterone and DHEAS serum levels, n = 502). Three hundred and sixty-four women (n = 364) displayed mixed hyperandrogenemia, since more than one androgens were elevated in these subjects. Accordingly, the most frequent
123
123
0.95 (0.64/1.40)
1.47 (0.84/2.30)
1.44 (1/2.11)
6.25 (4.62/9.09)
1.54 (1.06/2.12)
6.62 (5.16/9.25)
22 (17/26.50)
HOMA-IR
Mean ovary volume (cm3)
Sum of follicles
0.165
0.619
0.66
0.90
0.871
\0.001
0.012
20 (14/27)
7.69 (5.41/9.90)
1.8 (1.26/2.54)
8 (5.68/10.86)
94 (87/101)
7.36 (5.12/11.17)
38.04 (28.80/55)
Bold values are statistically significant (p \ 0.05)
EA elevated androgens, NA normal androgens
21 (15/26)
7.16 (5.45/9.80)
1.76 (1.24/2.40)
7 (5.39/10.6)
98 (89/105)
3.47 (2.27/5.11)
52.32 (38.2/71.6)
Data are presented as median. Numbers in brackets indicate the 25th and 75th percentiles
21 (13/26)
6.30 (5/9)
6.80 (4.95/9.20)
Insulin (lU/ml)
96 (85.50/103)
3.07 (1.80/4.68)
96 (86/102)
6.26 (4.48/8.91)
Glucose (mg/dl)
FAI
58 (41.3/77.40)
46.70 (35.50/69.8)
2.1 (1.76/2.55)
54 (42/64.60)
1.01 (0.71/1.73)
6.30 (4.42/9.30)
5.95 (4.7/7.13)
9 (4/10)
0.75 (0.71/0.78)
NA (n = 207)
3,475 (2,653/4,081) 2,220 1,565/2,840
3.23 (2.57/4.04)
\0.001
SHBG (nmol/l)
2.2 (1.84/2.54)
85 (74.95/101)
1.24 (0.87/2)
\0.001 \0.001
5.93 (4.80/7.21) 7.50 (5.25/11.26)
0.192
9 (6/11)
0.75 (0.71/0.78)
\0.001
3,070 (2,216/3,973) 1,931 (1,513/2,420)\0.001
3 (2.40/3.87)
84.90 (74/105.50) 51.90 (34.50/66)
0.729 0.973
DHEAS (lg/l)
D4 (ng/ml)
Testosterone (ng/dl)
LH/FSH
5.54 (3.54/8.37)
LH (U/L)
5.80 (4.94/6.70)
6.37 (5.03/7.19)
8.48 (5.57/13.25)
FSH (U/L)
6 (0/10)
0.73 (0.71/0.75)
6 (1.5/10)
0.73 (0.70/0.76)
EA (n = 293)
p
EA (n = 85)
NA (n = 61)
20–25 kg/m2
\20 kg/m2
FG score
WHR
BMI range
0.942
29 (21.62/40)
\0.001
0.491
0.218
0.483
0.097
0.001
22 (15/28)
7.41 (5.12/9.68)
2.30 (1.52/3.70)
10.28 (7/15)
94 (87/104)
10.41 (7.11/15.31)
3,765 (2,810/4,520)
\0.001
\0.001
3.23 (2.60/3.86)
89.34 (75/109)
1.16 (0.80/1.69)
6.12 (4.50/9.40)
5.52 (4.53/6.70)
10 (7/13)
0.79 (0.76/0.82)
EA (n = 163)
\0.001
\0.001
0.009
0.001
0.518
\0.001
p
25.001–30 kg/m2
20 (15.5/26)
8.37 (6.93/10.66)
2.53 (1.86/3.76)
10.90 (7.21/15.02)
96 (87/104.35)
5.61 (3.57/7.51)
35.30 (27/48.35)
2,235 (1,615/2,818)
2.08 (1.71/2.54)
59 (43.50/65.50)
0.98 (0.65/1.52)
5.30 (3.90/8.15)
5.70 (4.61/6.95)
10 (4/12)
0.80 (0.76/0.85)
NA (n = 101) 0.05
0.026
0.012
0.591
0.139
0.325
0.004
0.372
0.431
0.741
\0.001
0.002
\0.001
\0.001
\0.001
p
22 (17/27)
7.92 (5.75/10.80)
4.32 (2.73/6.17)
18 (11.30/24.3)
98 (91/105)
13.40 (9.83/20.82)
22.40 (17.57/30.36)
3,441 (2,503/4,306)
3 (2.35/3.70)
92 (77/110)
1.23 (0.81/1.76)
6.52 (3.95/9.50)
5.70 (4.40/6.40)
10 (7/14)
0.84 (0.79/0.90)
EA (n = 175)
[30 kg/m2
20 (14/26)
6.85 (5.58/9.47)
4.38 (2.81/6.31)
17.90 (11.87/25.7)
101 (93.50/111)
6.53 (4.06/9.29)
28 (20.04/38.5)
2,100 (1,580/2,815)
1.98 (1.60/2.42)
58.50 (41/66.10)
0.87 (0.62/1.31)
5 (3.38/6.91)
5.50 (4.50/6.70)
9 (5.5/11)
0.83 (0.80/0.87)
NA (n = 133)
Table 2 Comparison of women with elevated androgens (EA) versus women with normal serum androgens (NA) according to their BMI status in all studied parameters
0.68
0.875
0.002
0.05
0.05
0.367
0.579
0.046
\0.001
\0.001
\0.001
\0.001
\0.001
\0.001
\0.001
p
Endocrine
Endocrine Fig. 1 The percentage of isolated and combined hyperandrogenemia in 1,218 women with PCOS. Normal serum androgens were found in 41.2, 29.9 % had combined hyperandrogenemia, and isolated elevated testosterone, DHEAS and D4 were found in 17.2, 7.6 and 4.1 %, respectively
abnormality as the sole marker of hyperandrogenemia was isolated elevated testosterone levels (17.2 %), followed by DHEAS (7.6 %) and D4 (4.1 %). In 364 women (29.9 %), combined hyperandrogenemia was disclosed, since more than one androgens were elevated. These findings are clearly illustrated in Fig. 1. From the comparisons between subgroups, it was found that LH values and LH/FSH ratio were higher in GD4 and GT groups (Table 3).
Discussion In the present study, it was found that in a very large cohort (n = 1,218) of women with PCOS diagnosed with Rotterdam criteria, (1) the overall prevalence of hyperandrogenemia is 58.8 %, and testosterone levels constituting the predominant form in women with isolated hyperandrogenemia, and (2) that LH values were higher in women with either elevated testosterone (GT) or D4 (GD4) levels compared to those with raised DHEAS (GD) levels or those with normal androgens (GP). Obesity, a key player of PCOS characteristics development [17], did not appear to be an important denominator of different phenotypes between normo- and hyperandrogenemic PCOS women. This finding suggests that hyperandrogenemia consists of an important determinant of different PCOS phenotypes, per se. Concerning biochemical hyperandrogenism prevalence, our findings are different from studies using the other available criteria for PCOS diagnosis. Huang et al. estimated that the overall prevalence of hyperandrogenemia in
women diagnosed with PCOS according to NIH criteria was 75.3 %, and Androgen Excess Society guideline reported that elevated circulating androgens levels are observed in about 60–80 % of population [18, 19]. In our population, we estimated that the prevalence of hyperandrogenemia is 58.8 %, lower than the aforementioned rates. However, in the aforementioned study except the fact that diagnosis was based on NIH criteria, the study group was comprised of obese women (mean BMI 33.39 ± 9.26 kg/m2), with different ethnicity and race (14 % of them were Afro-American), while in the present study, women were divided according to the degree of obesity and belonged to the same ethnicity. Regarding the prevalence of isolated hyperandrogenemia in women with PCOS, available data are scanty. In the present study, isolated elevated testosterone levels were found in a rate of 17.2 %. In a study carried out in Brazil, 44 % of women with PCOS according to Rotterdam criteria displayed elevated testosterone values [20]. Although the studied population was comparable to the one evaluated in the present study, regarding age and BMI status, there was a mixed population on ethnicity and race. In addition, it has not indicated whether increased testosterone levels constitute the sole hyperandrogenemic abnormality or whether there was a combination of any kind between elevated testosterone and other androgens. In the aforementioned NIH cohort, elevated levels of testosterone as the sole abnormality were found in only 2 % of patients [18]. Regarding LH values, we have noticed that LH levels were higher in women with isolated testosterone or D4
123
Endocrine Table 3 Comparison of women with isolated increased testosterone (GT), DHEAS (GD), D4 (GD4) and women with normal androgens (GP)
BMI (kg/m2) WHR FG score
GT (n = 210)
GD (n = 92)
GD4 (n = 50)
GP (n = 502)
24.78 (21.27/31.89)
24.46 (21.48/29.08)
23.49 (21.65/29.87)
24.41 (21.43/30.61)
0.77 (0.73/0.82)
0.78 (0.73/0.83)
0.79 (0.75/0.84)
0.78 (0.73/0.83)
9 (4.75/11)
10 (4/12)
FSH (U/L)
5.89 (4.67/6.80)
5.77 (4.45/6.98)
LH* (U/L)
7.13 (4.80/9.90)
5.37 (3.77/7.62)
LH/FSH^
1.23 (0.84/1.81)
Testosterone (ng/dl)# D4 (ng/ml)[
2.64 (2.20/2.99)
DHEAS (lg/l)\
2,565 (2,063/3,095)
##
FAI
7.93 (5.61/11.89)
SHBG** (nmol/l) Glucose (mg/dl)^^
36.81 (26.54/55) 94.5 (85/105)
Insulin (lU/ml)
9 (5.93/14.81)
HOMA-IR 3
Ovary volume (cm ) Follicle sum
86 (80/96)
9 (5.75/12) 6 (4.88/7.43) 9.39 (6.15/12.58)
p 0.737 0.447
9 (4/11)
0.192
5.75 (4.68/6.90)
0.764
5.57 (3.91/8.20)
\0.001 \0.001
0.97 (0.66/1.24)
1.57 (0.96/2.31)
0.97 (0.67/1.51)
63.60 (61/68.78)
64.50 (52.75/70)
55.85 (41.13/65)
\0.001
2.43 (2.1/2.82)
3.90 (3.56/4.24)
2.08 (1.70/2.50)
\0.001
2,770 (2,148/3,131)
2,159 (1,580/2,802)
\0.001
7.11 (5.30/9.10)
5.08 (3.66/7.33)
4.30 (2.71/6.58)
\0.001
31.60 (23.73/40.30) 96 (88/103)
37.50 (26.88/51.14) 97 (91/103.50)
41 (29/60.90) 99 (89/106)
\0.001 0.03
4,058 (3,835/4,646)
9.27 (5.90/15.65)
9.18 (7/15.51)
9.38 (6/15.67)
0.82
2.06 (1.28/3.55)
2.21 (1.38/3.93)
2.22 (1.48/3.98)
2.18 (1.42/3.74)
0.519
7.52 (5.40/10.21)
7.15 (5.19/8.22)
7.40 (4.67/10.30)
7.5 (5.51/9.90)
0.109
21 (15/26)
0.513
21 (15/27)
21 (15/26)
22 (17.75/28.50)
Data are presented as median. Numbers in brackets indicate the 25th and 75th percentiles Post hoc analysis with Mann–Whitney test corrected for multiple comparisons (Bonferroni test) * All groups differ statistically significantly among them (p \ 0.001; p = 0.027 in GD4 vs. GT group), except GD versus GP group (p = 1.353) ^
All groups differ statistically significantly among them (p \ 0.001), except GD4 versus GT group (p = 0.117) and GD versus GP group (p = 0.957)
#
All groups differ statistically significantly among them (p \ 0.001. GD4 vs. GP p = 0.006), except GD4 versus GD group (p = 1.77)
[
All groups differ statistically significantly among them (p \ 0.001), except GT versus GD group (p = 0.147)
\
All groups differ statistically significantly among them (p \ 0.001), except GT versus GD4 group (p = 0.795)
##
All groups differ statistically significantly among them (p \ 0.001, GD4 vs. GD p = 0.003), except GT versus GD group and GD4 versus GP group (p = 0.06) ** Only GD versus GP group (p \ 0.001) and GT versus GD group (p = 0.009) differ statistically significantly among four groups
^^
Only GT versus GP group (p = 0.015) differs statistically significantly among four groups
Bold values are statistically significant (p \ 0.05)
elevated values compared to women with isolated elevated DHEAS or with women without hyperandrogenemia. This constitutes a finding not encountered during our literature review. It is known that in women with PCOS, a high prevalence of abnormal gonadotropin secretion has been reported, which consists of an increase in both the amplitude and frequency of LH secretion. It has been hypothesized that the frequency of GnRH pulses determines which gonadotropin hormone is synthesized and secreted with rapid GnRH pulses favoring LH secretion and this disturbed machinery has been implicated in the pathogenesis of the syndrome, through the enhanced androgen secretion from the stimulated ovarian theca cell [21–23]. However, recent data suggest a vicious circle between LH amplitude and androgen ovarian secretion, since Solorzano et al. [24] reported that hyperandrogenemia may adversely affect LH pulse regulation during pubertal maturation leading to persistent hyperandrogenemia and the development of
123
PCOS. Rather, an intrinsic dysregulation in the steroidogenic pathways of ovarian cells may result in a condition in which the set point for desensitization to LH is set higher [25–27]. Our findings suggest that the specific type of hyperandrogenemia may contribute to this disturbed cycle between androgens and LH secretion, since women with either testosterone or D4 elevated levels and not those with elevated DHEAS levels or normal androgens displayed higher values of LH. Interestingly, SHBG values were lower in subgroups of women with hyperandrogenemia, regardless of their BMI status. This fascinating finding cannot be attributed to the well-described insulin action on SHBG liver synthesis, since insulin concentrations were comparable among subgroups (Table 3), thus enhancing the notion that SHBG regulation is multifactorial, with hyperandrogenemia consisting of a possible factor of SHBG regulation [28–30]. It is also interesting that different phenotypes between
Endocrine
normoandrogenemic and hyperandrogenemic PCOS women do not seem to be affected by BMI, since syndrome’s manifestation such as LH and SHBG levels differs significantly between them—either they were normal, overweight or obese. The finding of lower fasting glucose levels in women with hyperandrogenemia in comparison with women with PCOS with normal circulating androgens (Table 1) is intriguing. Although in absolute values the difference is very small (95 vs. 99 mg/dl), further research is required to elucidate any potential role of androgens in glycemia intraPCOS group, since this aspect has not been reported before. This study has two major limitations: the cross-sectional setting and the long time needed to collect these data. On the other hand, the large number of studied subjects and their categorization according to BMI status strengthen the findings of the present study. In conclusion, our findings suggest that in women with polycystic ovary syndrome diagnosed with Rotterdam criteria, the prevalence of overall hyperandrogenemia is lower and the prevalence of isolated elevated testosterone is higher compared with women diagnosed with 1990 NIH criteria. Additionally, elevations of specific circulating androgens (specifically D4 and testosterone elevations) may modify LH levels. The above findings as a whole suggest that the type of hyperandrogenemia may play a clinically significant role in PCOS diagnosis. Conflict of interest
All authors have nothing to disclose.
References 1. E. Diamanti-Kandarakis, Polycystic ovarian syndrome: pathophysiology, molecular aspects and clinical implications. Expert Rev. Mol. Med. 3010, e3 (2008) 2. E. Diamanti-Kandarakis, A. Dunaif, Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications. Endocr. Rev. 33, 981–1030 (2012) 3. E. Kandaraki, C. Christakou, E. Diamanti-Kandarakis, Metabolic syndrome and polycystic ovary syndrome… and vice versa. Arq. Bras. Endocrinol. Metabol. 53, 227–237 (2009) 4. E. Diamanti-Kandarakis, D. Panidis, Unravelling the phenotypic map of polycystic ovary syndrome (PCOS): a prospective study of 634 women with PCOS. Clin. Endocrinol. (Oxf.) 67, 735–742 (2007) 5. E. Diamanti-Kandarakis, C. Christakou, E. Marinakis, Phenotypes and environmental factors: their influence in PCOS. Curr. Pharm. Des. 18, 270–282 (2012) 6. S.F. Witchel, S.E. Recabarren, F. Gonza´lez, E. Diamanti-Kandarakis, K.I. Cheang, A.J. Duleba, R.S. Legro, R. Homburg, R. Pasquali, R.A. Lobo, C.C. Zouboulis, F. Kelestimur, F. Fruzzetti, W. Futterweit, R.J. Norman, D.H. Abbott, Emerging concepts about prenatal genesis, aberrant metabolism and treatment paradigms in polycystic ovary syndrome. Endocrine 42, 526–534 (2012)
7. 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). Fertil. Steril. 81, 19–25 (2004) 8. M.A. Lawson, S. Jain, S. Sun, K. Patel, P.J. Malcolm, R.J. Chang, Evidence for insulin suppression of baseline luteinizing hormone in women with polycystic ovarian syndrome and normal women. J. Clin. Endocrinol. Metab. 93, 2089–2096 (2008) 9. T.D. Pache, F.H. de Jong, W.C. Hop, B.C. Fauser, Association between ovarian changes assessed by transvaginal sonography and clinical and endocrine signs of the polycystic ovary syndrome. Fertil. Steril. 59, 544–549 (1993) 10. A.A. Kyei-Mensah, S. LinTan, J. Zaidi, H.S. Jacobs, Relationship of ovarian stromal volume to serum androgen concentrations in patients with polycystic ovary syndrome. Hum. Reprod. 13, 1437–1441 (1998) 11. Z. Puzigac´a, G.M. Prelevic´, Z. Stretenovic´, L. Balint-Peric´, Ovarian enlargement as a possible marker of androgen activity in polycystic ovary syndrome. Gynecol. Endocrinol. 5, 167–174 (1991) 12. R. Pasquali, A. Gambineri, C. Cavazza, D. Ibarra Gasparini, W. Ciampaglia, G.E. Cognigni, U. Pagotto, Heterogeneity in the responsiveness to long-term lifestyle intervention and predictability in obese women with polycystic ovary syndrome. Eur. J. Endocrinol. 164, 53–60 (2011) 13. P. Moghetti, R. Castello, C. Negri, F. Tosi, F. Perrone, M. Caputo, E. Zanolin, M. Muggeo, Metformin effects on clinical features, endocrine and metabolic profiles, and insulin sensitivity in polycystic ovary syndrome: a randomized, double-blind, placebo controlled 6-month trial, followed by open long-term. J. Clin. Endocrinol. Metab. 85, 139–146 (2000) ¨ . Turhan, Impact 14. N. Ko¨s¸ u¨s¸ , A. Ko¨s¸ u¨s¸ , Z. Kamalak, D. Hızlı, N.O of adrenal versus ovarian androgen ratio on signs and symptoms of polycystic ovarian syndrome. Gynecol. Endocrinol. 28, 611–614 (2012) 15. K. Brennan, A. Huang, R. Azziz, Dehydroepiandrosterone sulfate and insulin resistance in patients with polycystic ovary syndrome. Fertil. Steril. 91, 1848–1852 (2009) 16. A. Dagre, J. Lekakis, C. Mihas, A. Protogerou, L. Thalassinou, D. Tryfonopoulos, G. Douridas, C. Papamichael, M. Alevizaki, Association of dehydroepiandrosterone-sulfate with endothelial function in young women with polycystic ovary syndrome. Eur. J. Endocrinol. 154, 883–890 (2006) 17. D. Panidis, K. Tziomalos, E. Papadakis, C. Vosnakis, P. Chatzis, I. Katsikis, Lifestyle intervention and anti-obesity therapies in the polycystic ovary syndrome: impact on metabolism and fertility. Endocrine 44, 583–590 (2013) 18. A. Huang, K. Brennan, R. Azziz, Prevalence of hyperandrogenemia in the polycystic ovary syndrome diagnosed by the NIH 1990 criteria. Fertil. Steril. 93, 1938–1941 (2010) 19. R. Azziz, E. Carmina, D. Dewailly, E. Diamanti-Kandarakis, H.F. Escobar-Morreale, W. Futterweit, O.E. Janssen, R.S. Legro, R.J. Norman, A.E. Taylor, S.F. Witchel, Task force on the phenotype of the polycystic ovary syndrome of the androgen excess and PCOS Society. Criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: an androgen excess society guideline. Fertil. Steril. 9, 456–488 (2009) 20. L. Gabrielli, E.M. Aquino, Polycystic ovary syndrome in Salvador, Brazil: a prevalence study in primary healthcare. Reprod. Biol. Endocrinol. 10, 96 (2012) 21. E. Diamanti-Kandarakis, J. Papailiou, S. Palimeri, Hyperandrogenemia: pathophysiology and its role in ovulatory dysfunction in PCOS. Pediatr. Endocrinol. Rev. 3, 198–204 (2006) 22. S.K. Blank, C.R. McCartney, J.C. Marshall, The origins and sequelae of abnormal neuroendocrine function in polycystic ovary syndrome. Hum. Reprod. Update 12, 351–361 (2006)
123
Endocrine 23. A.E. Taylor, B. McCourt, K.A. Martin, E.J. Anderson, J.M. Adams, D. Schoenfeld, J.E. Hall, Determinants of abnormal gonadotropin secretion in clinically defined women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 82, 2248–2256 (1997) 24. C.M. Solorzano, J.P. Beller, M.Y. Abshire, J.S. Collins, C.R. McCartney, J.C. Marshall, Neuroendocrine dysfunction in polycystic ovary syndrome. Steroids 77, 332–337 (2012) 25. A.J. Jakimiuk, S.R. Weitsman, A. Navab, D.A. Magoffin, Luteinizing hormone receptor, steroidogenesis acute regulatory protein, and steroidogenic enzyme messenger ribonucleic acids are overexpressed in thecal and granulosa cells from polycystic ovaries. J. Clin. Endocrinol. Metab. 86, 1318–1323 (2001) 26. V.L. Nelson, R.S. Legro, J.M. McAllister, Augmented androgen production is a stable steroidogenic phenotype of propagated theca cells from polycystic ovaries. Mol. Endocrinol. 13, 946–957 (1999)
123
27. R.L. Rosenfield, B. Bordini, Evidence that obesity and androgens have independent and opposing effects on gonadotropin production from puberty to maturity. Brain Res. 1364, 186–197 (2010) 28. V. Toscano, R. Balducci, P. Bianchi, R. Guglielmi, A. Mangiantini, F. Sciarra, Steroidal and non-steroidal factors in plasma sex hormone binding globulin regulation. J. Steroid Biochem. Mol. Biol. 43, 431–437 (1992) 29. M.G. Forest, A. Lecoq, M. David, M. Pugeat, Effects of human chorionic gonadotropin, androgens, adrenocorticotropin hormone, dexamethasone and hyperprolactinemia on plasma sex steroid-binding protein. Ann. N. Y. Acad. Sci. 538, 214–234 (1988) 30. F. Gonza´lez, C.L. Sia, F.Z. Stanczyk, H.E. Blair, M.E. Krupa, Hyperandrogenism exerts an anti-inflammatory effect in obese women with polycystic ovary syndrome. Endocrine 42, 726–735 (2012)
