Polycystic ovary syndrome

Transcription

1 ORIGINAL ARTICLES: REPRODUCTIVE ENDOCRINOLOGY Opposing effects of dehydroepiandrosterone sulfate and free testosterone on metabolic phenotype in women with polycystic ovary syndrome Elisabeth Lerchbaum, M.D., a Verena Schwetz, M.D., a Albrecht Giuliani, M.D., b Thomas R. Pieber, M.D., a and Barbara Obermayer-Pietsch, M.D. a a Department of Internal Medicine, Division of Endocrinology and Metabolism, and b Department of Obstetrics and Gynecology, Medical University of Graz, Graz, Austria Objective: To study the association of adrenal and ovarian androgen levels with metabolic parameters in a large cohort of women with polycystic ovary syndrome (PCOS). Design: Cross-sectional study. Setting: Outpatient clinic of an academic hospital. Patient(s): Six hundred twenty-two women with PCOS. Intervention(s): None. Main Outcome Measure(s): Analysis of the association of endocrine dehydroepiandrosterone sulfate (DHEAS) and free testosterone (FT) parameters with metabolic measurements. Result(s): In multivariate adjusted logistic regression analyses, the odds ratio (OR) for insulin resistance was statistically significantly higher (4.42, : ) for women with PCOS who had elevated FT levels compared with the women with normal DHEAS and FT levels (reference group). We found no statistically significant differences when women with PCOS with elevated DHEAS or a combined elevation of DHEAS and FT levels were compared with the reference group. Women with PCOS and a high DHEAS/FT ratio had a more beneficial metabolic profile compared with the women with a low DHEAS/FT ratio. In multivariate adjusted binary logistic regression analyses, we found a statistically significantly lower risk for insulin resistance in the women with PCOS in the highest DHEAS/FT-ratio quartile compared with women with PCOS in the lowest quartile (OR 0.35, : ). Conclusion(s): Our results suggest that the distinction between adrenal and ovarian hyperandrogenism is important when evaluating metabolic risk in PCOS. (Fertil Steril Ò 2012;98: Ó2012 by American Society for Reproductive Medicine.) Key Words: DHEAS, free testosterone, hyperandrogenemia, metabolic disturbances, polycystic ovary syndrome Discuss: You can discuss this article with its authors and with other ASRM members at fertstertforum.com/lerchbaume-dheas-free-testosterone-metabolic-phenotype-pcos/ Use your smartphone to scan this QR code and connect to the discussion forum for this article now.* * Download a free QR code scanner by searching for QR scanner in your smartphone s app store or app marketplace. Received March 29, 2012; revised July 2, 2012; accepted July 3, 2012; published online July 14, E.L. has nothing to disclose. V.S. has nothing to disclose. A.G. has nothing to disclose. T.R.P. has nothing to disclose. B.O.-P. has nothing to disclose. Supported by BioPersMed (COMET K-project ), which is funded by the Federal Ministry of Transport, Innovation, and Technology (BMVIT) and the Federal Ministry of Economics and Labour/the Federal Ministry of Economy, Family, and Youth (BMWA/BMWFJ) and the Styrian Business Promotion Agency (SFG). Reprint requests: Elisabeth Lerchbaum, M.D., Medical University of Graz, Department of Internal Medicine, Division of Endocrinology and Metabolism, Auenbruggerplatz 15, 8036 Graz, Austria ( elisabeth.lerchbaum@medunigraz.at). Fertility and Sterility Vol. 98, No. 5, November /$36.00 Copyright 2012 American Society for Reproductive Medicine, Published by Elsevier Inc. Polycystic ovary syndrome (PCOS) is the most common endocrine disorder, affecting women of reproductive age at a prevalence of approximately 7% (1). Women with PCOS present with clinical and/ or biochemical hyperandrogenism, oligomenorrhea or amenorrhea, and polycystic ovaries on ultrasound. Moreover, women with PCOS frequently exhibit insulin resistance and obesity VOL. 98 NO. 5 / NOVEMBER 2012

2 Fertility and Sterility Although PCOS is a complex disorder presenting with various endocrine and metabolic abnormalities, there has been accumulating evidence that high androgen levels are the key factor in the disease pathogenesis (2). This notion is supported by the fact that hyperandrogenism may lead to polycystic ovary morphology and ovulatory dysfunction in animal models as well as in women with PCOS (3, 4). Those androgens are produced in the ovary as well as in the adrenal gland. It has been estimated that 25% of androstenedione and testosterone production is of ovarian origin, 25% is of adrenal origin, and 50% is produced in peripheral tissues, while the adrenal cortex accounts almost uniquely for the synthesis of dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS) as well as androstenediol and 11b-hydroxy androstenedione. In PCOS, the ovaries produce up to 60% of androgens, while the adrenals contribute the remaining 40% (5). The literature reports that 20% to 50% of women with PCOS have elevated adrenal androgens such as DHEAS (6 8). It has been suggested that ovarian and adrenal androgens have opposing effects on body weight and insulin metabolism in women with PCOS (7, 9, 10). On the one hand, high total testosterone (TT) and free testosterone (FT) levels are associated with an adverse metabolic phenotype in women with PCOS (2). High testosterone levels are linked with obesity, in particular with an abdominal fat distribution (11), as well as with insulin resistance and a higher prevalence of glucose intolerance (2). On the other hand, the role of adrenal androgens in metabolic disturbances is less clear. One study suggested a positive association of high DHEAS levels with metabolic disturbances such as hypertension (12). Conversely, Brennan et al. (10) found an independent association of high DHEAS levels with decreased insulin resistance in a cohort of 352 women with PCOS. Moreover, Chen et al. (13) demonstrated in 318 women with PCOS that despite a positive correlation of DHEAS levels with testosterone, high DHEAS levels were positively associated with a beneficial metabolic phenotype, including parameters such as abdominal obesity, insulin resistance, and dyslipidemia. A study including 280 Turkish women with PCOS reported that a high DHEAS/TT ratio, suggesting a higher adrenal androgen production, is associated with a beneficial metabolic phenotype (14). There were, however, no data available on glucose tolerance tests in those studies (10, 14). We studied the association of adrenal and ovarian androgen levels with metabolic parameters in a large cohort of women with PCOS in whom oral glucose tolerance tests were performed. MATERIALS AND METHODS Patients The study cohort consisted of 622 women with PCOS, aged 16 to 45 years, who had been routinely referred to our outpatient clinic for PCOS evaluation between 2006 and These patients had consulted our outpatient clinic for PCOSrelated symptoms such as obesity, hirsutism, acne, or menstrual irregularities. Moreover, 195 (31.3%) of the 622 women with PCOS were affected by infertility. We diagnosed PCOS by use of the Rotterdam criteria (15), with which two out of the following three characteristics are required to confirm the diagnosis: clinical and/or biochemical signs of hyperandrogenism, oligoovulation and/or anovulation, and/or polycystic ovaries (by ultrasound). Disorders with a similar clinical presentation, such as hyperprolactinemia, congenital adrenal hyperplasia, Cushing syndrome, or androgen-secreting tumors, must be excluded. Hyperandrogenism was defined by the clinical presence of hirsutism (modified Ferriman-Gallwey score R6), acne, or alopecia, and/or elevated androgen levels (normal of TT: <0.77 ng/ml and DHEAS: <2.75 mg/ml). Oligoovulation and/or anovulation were defined by the presence of oligomenorrhea or amenorrhea. Polycystic ovarian morphology was examined by ultrasound. Polycystic ovaries were defined as the presence of 12 or more follicles in each ovary measuring 2 9 mm in diameter and/or increased ovarian volume (>10 ml, calculated by use of the formula 0.5 Length Width Thickness) (15). Hyperprolactinemia, Cushing syndrome, congenital adrenal hyperplasia, and androgen-secreting tumors were excluded by specific laboratory analysis (prolactin, cortisol, adrenocorticotropic hormone [ACTH], and 17a-OH progesterone). The PCOS group had not taken any medication known to affect endocrine parameters, carbohydrate metabolism, or serum lipid profile for at least 3 months before entering the study. The study protocol was approved by the ethics committee of the Medical University of Graz, and written informed consent was obtained from each patient. Procedures Standard anthropometric data were obtained from each subject: height, weight, waist circumference (WC), hip circumference (HC), and blood pressure. We measured WC in the standing position, midway between the lower costal margin and the iliac crest. We measured HC in a standing position, at the maximum circumference over the buttocks. The BMI was calculated as the weight in kilograms divided by the square of height in meters. Hirsutism was quantified with the modified Ferriman-Gallwey score. The basal blood samples for hormone TT, sex hormonebinding globulin (SHBG), androstenedione, DHEAS, free triiodothyronine (ft 3 ), free thyroxine (ft 4 ), thyroid-stimulating hormone (TSH), 17a-hydroxyprogesterone (17-OHP), prolactin, cortisol, and ACTH and metabolic parameters (glucose and insulin) were collected between 8.00 and 9.00 AM after an overnight fast. Assays were performed in the same laboratory as they came in, and the laboratory kits and assays did not change between 2006 and Both TT and SHBG were measured on a daily basis and were stored at 4 C until analysis. The DHEAS levels were measured on a weekly basis, and blood samples were frozen and stored at 40 C until analysis. The FT values were calculated from TT, SHBG, and albumin according to Vermeulen et al. (16). Elevated FT levels were defined as the upper quartile of FT levels (>0.013 ng/ml). All participants underwent a fasting 75 g oral glucose tolerance test. Blood samples were drawn after 30, 60, and VOL. 98 NO. 5 / NOVEMBER

3 ORIGINAL ARTICLE: REPRODUCTIVE ENDOCRINOLOGY 120 minutes for glucose and insulin determination. Impaired glucose tolerance (IGT) was identified by 2-hour glucose levels between 140 mg/dl and 199 mg/dl (between 7.8 mmol/l and 11.1 mmol/l), as defined by the American Diabetes Association. The metabolic syndrome was defined by the National Cholesterol Education Program and the Adult Treatment Panel-III in women presenting at least three of the following criteria: WC >88 cm, high-density lipoprotein (HDL) cholesterol <50 mg/dl, triglyceride level >150 mg/dl, raised blood pressure (systolic >130 mm Hg, diastolic >85 mm Hg), and raised fasting glucose (>110 mg/dl) or prevalent type 2 diabetes (17). Insulin resistance was estimated using the homeostatic model assessment-insulin resistance (HOMA-IR) and was assumed for levels >2.5. We calculated HOMA-IR as the product of the fasting plasma insulin value (miu/ml) and the fasting plasma glucose value (mg/dl) divided by 405. Biochemical Analyses We measured DHEAS (LDN Labor Diagnostika Nord GmbH), androstenedione, and 17-OHP (DiaMetra; BioVendor) via enzyme-linked immunosorbent assay (ELISA), with intraassay and interassay coefficients of variation (CV) of <10%. Insulin was measured by ELISA (Siemens) with intra-assay and interassay CV of 4.0 and 2.6, respectively. We measured SHBG by luminescence immunoassay (Roche) with an intraassay and interassay CV of 1.3% and 2.1%, respectively. We measured TT as well as prolactin, TSH, ft 3,fT 4, and cortisol (Siemens) by luminescence immunoassay; the intra-assay and interassay CV for the TT measurements were 6.2% and 4.7%, respectively. Albumin was measured by photometric assay (Roche). Statistical Analysis Data are presented as median with interquartile, unless otherwise stated. The Kolmogorov-Smirnov test and descriptive statistics were used to evaluate the distribution of data. All continuous parameters following a non-normal distribution were logarithmically transformed when parametric tests were performed. The PCOS group was stratified into four groups according to their androgen status: normal DHEAS and FT levels, elevated DHEAS levels, elevated FT levels, and elevated DHEAS and FT levels. Binary logistic regression analysis was performed to examine the association of insulin resistance with age, BMI, WHR, and androgen groups. We calculated DHEAS/TT and the DHEAS/FT ratio to examine adrenal versus ovarian androgen synthesis. Analysis of variance (ANOVA) and the chi-square test were used to compare groups. Pearson correlation analyses were used to determine relationships between variables. Multivariable stepwise linear regression analysis was performed with area under the curve (AUC) insulin as the dependent variable and with age, BMI, WHR, and DHEAS/FT ratio (or DHEAS and FT) as independent variables. Binary logistic regression analyses were performed to examine the associations of insulin resistance (yes/no, defined as HOMA-IR >2.5) and metabolic syndrome (yes/no) with age, BMI, WHR, and DHEAS/FT ratio. P value of <.05 was considered statistically significant. All analyses were performed with SPSS 18.0 (SPSS Inc.). RESULTS Androgen Groups In the women with PCOS, we observed elevated levels of DHEAS, FT, and TT in 26.4%, 24.4%, and 27.2%, respectively. In particular, 31.8% of women with PCOS with elevated DHEAS levels also had elevated FT levels. In women with PCOS with elevated FT levels, 35.0% of all women were affected by elevated DHEAS levels. The metabolic and endocrine parameters according to androgen status are shown in Table 1. We found statistically significant differences in endocrine as well as metabolic parameters between the androgen groups. Of note, the women with PCOS who had elevated FT levels had an adverse metabolic profile compared with the women with PCOS who had elevated DHEAS levels. Moreover, several parameters such as HOMA-IR, fasting and stimulated glucose, and fasting and AUC insulin as well as the serum lipids and HbA1c levels were higher in the women with PCOS who had elevated FT levels compared with the women with PCOS with both elevated FT and DHEAS levels. Further, the prevalence of insulin resistance, IGT, and the metabolic syndrome was higher in women with PCOS who had elevated FT levels compared with the other androgen groups. In the binary logistic regression analyses, adjusting for age, BMI, and WHR, the odds ratio (OR) for insulin resistance was 4.42 ( ) for women with PCOS who had elevated FT levels compared with the women with PCOS who had normal DHEAS and FT levels (reference) (Fig. 1B). We found no statistically significant differences when the women with PCOS who had elevated DHEAS levels or a combined elevation of DHEAS and FT levels were compared with the reference group. Similarly, in the binary logistic regression analyses, adjusting for age, BMI, and WHR, we observed a statistically significantly increased risk for prevalent IGT (OR 2.29 [ ]) (see Fig. 1C) in women with PCOS who had elevated FT levels compared with the women with PCOS who had normal DHEAS and FT levels. We found again no statistically significant differences when women with PCOS who had elevated DHEAS levels or a combined elevation of DHEAS and FT levels were compared with the reference group. Elevated DHEAS versus a Combined Elevation of DHEAS and FT Levels To further analyze the prospective protective role of DHEAS over FT, we performed subgroup analyses of the women with elevated DHEAS levels as compared with the women who had a combined elevation of DHEAS and FT levels. We observed statistically significantly higher BMI (P¼.006) as well as higher 1-hour insulin (P¼.007), 2-hour insulin (P¼.007), and AUC insulin (P<.001) levels in women with PCOS who had a combined elevation of DHEAS and FT levels than in the women with elevated DHEAS levels. These differences lost their statistical significance when the analyses were adjusted for age and BMI (P¼.139, P¼.113, and P¼.070 for 1320 VOL. 98 NO. 5 / NOVEMBER 2012

4 Fertility and Sterility TABLE 1 Characteristics of women with polycystic ovary syndrome according to androgen groups. DHEAS and FT normal (n [ 360, 57.9%) DHEAS elevated (n [ 110, 17.1%) FT elevated (n [ 98, 15.8%) DHEAS and FT elevated (n [ 54, 8.9%) P value Age (y) <.001 BMI (kg/m 2 ) <.001 WC (cm) <.001 WHR BP (mm Hg) Systolic <.001 Diastolic Glucose Fasting (mg/dl) h (mg/dl) h (mg/dl) AUC Fasting insulin (miu/ml) <.001 Insulin 1h(mIU/mL) <.001 2h(mIU/mL) <.001 HOMA <.001 AUC <.001 TT (ng/ml) <.001 FT (ng/ml) <.001 SHBG (nmol/l) <.001 DHEAS (mg/ml) <.001 A (ng/ml) <.001 Triglycerides (mg/dl) <.001 TC (mg/dl) HDL (mg/dl) TC/HDL LDL (mg/dl) HbA1c (%) FG score Insulin resistance (%) <.001 IGT (%) <.001 MS (%) Note: Data were analyzed by analysis of variance (ANOVA) and chi-square test. A ¼ androstenedione; AUC ¼ area under the curve; BMI ¼ body mass index; BP ¼ blood pressure; DHEAS ¼ dehydroepiandrosterone sulfate; FG score ¼ Ferriman-Gallwey score; FT ¼ free testosterone; HDL ¼ high-density lipoprotein; HOMA-IR ¼ homeostatic model assessment-insulin resistance; IGT ¼ impaired glucose tolerance; LDL ¼ low-density lipoprotein; MS ¼ metabolic syndrome; PCOS ¼ polycystic ovary syndrome; SHBG ¼ sex hormone-binding globulin; TC ¼ total cholesterol; TT ¼ total testosterone; WC ¼ waist circumference; WHR ¼ waist-to-hip ratio. 1-hour insulin, 2-hour insulin, and AUC insulin, respectively). We found no statistically significant differences between the two groups regarding fasting insulin, fasting and stimulated glucose levels, HOMA-IR, and serum lipids. The prevalence of insulin resistance, IGT, and the metabolic syndrome was higher in women with PCOS who had a combined elevation of DHEAS and FT compared with the women with PCOS who had elevated DHEAS levels (P<.001 for all). These analyses lost their statistical significance after adjustments for age and BMI (P¼.176, P¼.091, and P¼.246 for insulin resistance, IGT, and the metabolic syndrome, respectively). DHEAS/FT Ratio The characteristics of the women with PCOS according to their DHEAS/FT ratios are shown in Table 2. The women with PCOS who were within the higher DHEAS/FT quartiles had statistically significantly higher SHBG levels and statistically significantly lower BMI, WC, WHR, systolic and diastolic blood pressure, fasting and stimulated glucose levels, fasting and stimulated insulin levels, HOMA-IR, TT, FT, DHEAS, TG, total cholesterol (TC), and HbA1c levels. Further, the prevalence of insulin resistance and the metabolic syndrome was statistically significantly lower among women with PCOS who were within the higher DHEAS/FT quartiles, and IGT tended to be less prevalent in the women with PCOS within the higher quartiles. We found statistically significant positive correlations of the DHEAS/FT ratio with SHBG and statistically significant negative correlations with TT, age, BMI, WC, WHR, systolic and diastolic blood pressure, fasting and stimulated glucose levels, fasting and stimulated insulin levels, HOMA-IR, TG, TC, TC/HDL, low-density lipoprotein (LDL), and HbA1c levels (Supplemental Table 1, available online). Further, the DHEAS/ TT ratio was statistically significantly negatively correlated with age, diastolic blood pressure, fasting glucose, and TC levels. Moreover, we observed statistically significant positive correlations of DHEAS with FT, TT, and Ferriman-Gallwey scores, and statistically significant negative correlations VOL. 98 NO. 5 / NOVEMBER

5 ORIGINAL ARTICLE: REPRODUCTIVE ENDOCRINOLOGY FIGURE 1 FIGURE 1 Continued regression analyses used impaired glucose tolerance as the dependent variable and age, body mass index, waist-to-hip ratio, and androgen groups as independent variables. P¼.033 for FT elevated versus DHEAS and FT normal (reference). DHEAS and FT normal: n ¼ 360; DHEAS elevated: n ¼ 110; FT elevated: n ¼ 98; DHEAS and FT elevated: n ¼ 54. (A) Odds ratio with 95% confidence interval (CI) for insulin resistance according to dehydroepiandrosterone sulfate/free testosterone (DHEAS/FT) levels in women with polycystic ovary syndrome (PCOS) (n ¼ 622). The binary logistic regression analyses used insulin resistance as the dependent variable and age, body mass index, waist-to-hip ratio, and DHEAS/FT ratio as independent variables. P¼.027 for quartile 4 versus quartile 1 (reference). (B) Odds ratio with 95% CI for insulin resistance according to androgen groups in women with PCOS. The binary logistic regression analyses used insulin resistance as the dependent variable and age, body mass index, waist-to-hip ratio, and androgen groups as independent variables. P<.001 for FT elevated versus DHEAS and FT normal (reference). DHEAS and FT normal: n ¼ 360; DHEAS elevated: n ¼ 110; FT elevated: n ¼ 98; DHEAS and FT elevated: n ¼ 54. (C) Odds ratio with 95% CI for impaired glucose tolerance according to androgen groups in women with PCOS. The binary logistic with SHBG, age, fasting glucose, and TC. The FT levels were statistically significantly positively correlated with TT, BMI, WC, WHR, systolic and diastolic blood pressure, fasting and stimulated glucose levels, fasting and stimulated insulin levels, TG, TC/HDL, HbA1c levels, and Ferriman-Gallwey scores and were inversely related to SHBG, age, and HDL levels. As the DHEAS/FT ratio was more closely associated with metabolic disturbances than the DHEAS/TT ratio, we performed the following analyses with the DHEAS/FT ratio as an explanatory variable. In crude binary logistic regression analysis, the women with PCOS in the highest DHEAS/FT quartile had a statistically significantly lower OR for prevalent insulin resistance (0.41 ( ), P¼.034) than the women with PCOS in the lowest DHEAS/FT quartile (reference). Similar results were obtained in multivariate binary logistic regression analyses adjusting for age, BMI, and WHR (OR 0.35 [ ]) (see Fig. 1A). In crude binary logistic regression analysis, women with PCOS in the highest DHEAS/FT quartile had a statistically significantly lower risk of metabolic syndrome (OR 0.20 [ ], P¼.004) compared with the women in the lowest quartile (reference). These results remained statistically significant after adjusting for age but were attenuated after additional adjustment for BMI and WHR (data not shown). We performed stepwise linear regression analyses using AUC insulin as the dependent variable and age, BMI, WHR, and DHEAS/FT ratio as independent variables. We found that all variables were statistically significantly associated with AUC insulin (BMI: beta ¼ 0.268, P<.001; WHR: beta ¼ 0.300, P<.001; age: beta ¼ 0.233, P<.001; DHEAS/ FT-ratio: beta ¼ 0.124, P¼.005). When DHEAS and FT were entered separately in stepwise linear regression analyses, AUC insulin was negatively correlated with DHEAS (beta ¼ 0.132, P¼.005) and positively with FT (beta ¼ 0.230, P<.001). DISCUSSION We present evidence that women with PCOS and elevated FT levels had an adverse metabolic profile compared with women with PCOS who had elevated DHEAS levels. As we found no independent difference between women with elevated DHEAS levels compared with women with elevated DHEAS and FT levels, one might speculate that DHEAS neutralizes the adverse effect of high FT levels. Moreover, a high ratio of adrenal to ovarian androgens was associated with a beneficial metabolic phenotype VOL. 98 NO. 5 / NOVEMBER 2012

6 Fertility and Sterility TABLE 2 Characteristics of women with polycystic ovary syndrome according to DHEAS/FT-ratio quartiles. Quartile 1 (<138) n [ 156 Quartile 2 ( ) n [ 155 DHEAS/FT-ratio quartiles Quartile 3 ( ) n [ 155 Quartile 4 (>331) n [ 156 P value Age (y) BMI (kg/m 2 ) <.001 WC (cm) <.001 WHR BP (mm Hg) Systolic Diastolic <.001 Glucose Fasting (mg/dl) < h (mg/dl) h (mg/dl) <.001 AUC Insulin Fasting (miu/ml) <.001 1h(mIU/mL) <.001 2h(mIU/mL) <.001 HOMA <.001 AUC <.001 TT (ng/ml) <.001 FT (ng/ml) <.001 SHBG (nmol/l) <.001 DHEAS (mg/ml) <.001 A (ng/ml) Triglycerides (mg/dl) TC (mg/dl) HDL (mg/dl) TC/HDL LDL (mg/dl) HbA1c (%) FG score Prevalent acne (%) Insulin resistance (%) <.001 IGT (%) MS (%) Note: Data were analyzed by analysis of variance (ANOVA) and chi-square test. A ¼ androstenedione; AUC ¼ area under the curve; BMI ¼ body mass index; BP ¼ blood pressure; CRP ¼ C-reactive protein; DHEAS ¼ dehydroepiandrosterone sulfate; FG score ¼ Ferriman-Gallwey score; FT ¼ free testosterone; FTv ¼ free testosterone according to Vermeulen; HDL ¼ high-density lipoprotein; HOMA-IR ¼ homeostatic model assessment-insulin resistance; IGT ¼ impaired glucose tolerance; LDL ¼ low-density lipoprotein; MS ¼ metabolic syndrome; PCOS ¼ polycystic ovary syndrome; SHBG ¼ sex hormone-binding globulin; TC ¼ total cholesterol; TT ¼ total testosterone; WC ¼ waist circumference; WHR ¼ waist-to-hip ratio. We analyzed endocrine and metabolic parameters according to the presence of elevated DHEAS and/or FT levels. As expected, women with PCOS and elevated FT levels had an adverse metabolic risk profile compared with women with PCOS and elevated DHEAS levels, which is in line with previous studies (2, 10). It has been postulated that DHEAS opposes testosterone with respect to the risk of obesity, insulin resistance, and metabolic disturbances (9, 10). This notion is supported by our results showing that several metabolic parameters such as fasting and stimulated glucose, HOMA-IR, and serum lipids were higher in women with PCOS and elevated FT levels compared with women with PCOS who had both elevated FT and DHEAS levels. Moreover, in the subgroup analyses of women with elevated DHEAS levels versus women with a combined DHEAS and FT elevation, we found no independent differences between groups regarding the metabolic phenotype. Thus, DHEAS might exert some beneficial effect in this group and might protect women against the adverse effects of testosterone. These results support the hypothesis that high DHEAS levels have a beneficial effect on glucose and insulin metabolism. Thus, women with PCOS who have elevated DHEAS in combination with elevated FT levels might be protected against metabolic disturbances compared with women who have a discrete FT elevation. This notion is supported by the fact that no independent difference was found between women with a discrete DHEAS elevation compared with women with a combined DHEAS and FT elevation. We are aware of the speculative nature of this hypothesis, but nevertheless we believe it to be an interesting finding that deserves further investigation in prospective studies. It has been shown that DHEAS levels decrease with age, as does insulin action (18, 19). Accordingly, we found a negative VOL. 98 NO. 5 / NOVEMBER

7 ORIGINAL ARTICLE: REPRODUCTIVE ENDOCRINOLOGY association of DHEAS levels with age. The association of low DHEAS levels with insulin resistance and consecutive hyperinsulinemia might be driven by this factor. However, our results show an inverse association of DHEAS with hyperinsulinemia and insulin resistance that remains statistically significant after adjusting for age. We found elevated DHEAS levels in 26% of our women with PCOS, which is in line with previous studies showing that serum DHEAS is increased in 20% to 30% of women with PCOS (8). One explanation for elevated DHEAS levels in PCOS might be found in an altered cortisol metabolism. In particular, it has been shown that the peripheral metabolism of cortisol is increased in PCOS due to the enhanced inactivation of cortisol by 5a-reductase or impaired reactivation of cortisol from cortisone by 11b-hydroxysteroid dehydrogenase type 1 (20, 21). This could lead to a decreased negative feedback on ACTH, resulting in increased pituitary adrenal axis activity, including androgen synthesis to maintain normal cortisol levels. Moreover, increased peripheral sulfatase activity has been observed in women with PCOS, which might increase circulating DHEAS levels (22). A recent study in a Turkish cohort of women with PCOS demonstrated a favorable association of high DHEAS/TT ratios with insulin resistance, obesity, and dyslipidemia (14). Our study confirms the hypothesis that high levels of circulating androgens derived from the adrenal gland and lower levels of ovarian androgens exert a beneficial effect on metabolic disturbances in a large cohort of women with PCOS. Furthermore, it extends the results of data derived from oral glucose tolerance tests. We found, however, a closer association of the DHEAS/FT ratio with metabolic disturbances, when compared with the DHEAS/TT ratio. The Turkish study (14), however, did not report data on FT levels. Of note, the assessment of FT has been suggested to be more sensitive than the assessment of TT levels in hyperandrogenic disorders (2). We have to acknowledge that it remains to be determined whether the inverse relationship of the DHEAS/FT ratio and hyperinsulinemia is caused by a beneficial effect of DHEAS on insulin action, or whether DHEAS is simply a biomarker reflecting high levels of insulin. As the women with PCOS and elevated DHEAS levels had a similar metabolic profile compared with the women with elevated DHEAS and FT levels, one might speculate that DHEAS is protective against FT effects in those women. There are several mechanisms that might explain this purported protective effect of DHEAS in our PCOS cohort. First, DHEA might have an effect on insulin action. This notion is supported by the fact that DHEA decreases gluconeogenesis by suppressing the activity and expression of glucose- 6-phosphatase and phosphoenolpyruvate carboxykinase (23). Further, DHEA increases glucose uptake in hepatocytes and increases insulin binding to its receptor (9, 23). Second, high circulating levels of glucose or insulin might impact DHEAS synthesis in the adrenal gland, which might be mediated via insulin action on the latter (24). It is interesting that it has been shown that DHEAS levels increase after weight loss induced by gastric banding in premenopausal morbidly obese women (25) and a decrease of DHEAS levels over time is associated with the development of diabetes (26). Given our interesting results showing an association of DHEAS levels with a beneficial metabolic phenotype, existing evidence regarding DHEAS treatment should be discussed briefly. It is interesting that DHEAS supplementation in rats resulted in a reduction in food intake, body weight, total body fat content, and adipocyte size (27). Moreover, DHEAtreated rats had lower insulin levels (28). Several studies in humans that were mainly focused on patients with adrenal insufficiency investigated the effect of DHEAS treatment (29). Morales et al. (30) reported that DHEAS supplementation in women aged 40 to 70 years leads to an increase in insulinlike growth factor (IGF-1) levels as well as an improvement of physical and psychological well-being but without a significant change in insulin sensitivity. Others, however, speculate that oral DHEA supplementation, which is converted to DHEAS, might have a beneficial impact on the prevention and treatment of insulin resistance (31). It remains to be determined whether the therapeutic use of DHEAS results in metabolic improvements in women. As accumulating evidence suggests that hyperandrogenemia is the key mechanism in the pathogenesis of PCOS, our study might contribute to the understanding of the complex involvement of ovarian as well as adrenal androgens in PCOS. Our results suggest that the measurement of both FT and DHEAS is important in women affected by PCOS. Moreover, the discrimination between a distinct FT or DHEAS elevation compared with a combined FT and DHEAS elevation might allow some risk stratification regarding metabolic consequences. Thus, women with PCOS and discrete FT elevations might constitute a high-risk group who are in need of a close metabolic follow-up and intensive treatment, including lifestyle intervention. Our study has limitations that should be noted. First, the observational character of our study does not allow final conclusions as to a potential causal role of adrenal and ovarian androgens in the development of metabolic disturbances in women with PCOS. Second, we did not measure androgen levels with mass spectrometry, which is considered the gold standard for measuring androgen concentrations in women (32). However, given the case that measurements of androgen levels are inaccurate in our cohort, our results would underestimate rather than overestimate the observed association between androgen levels and metabolic disturbances. The strengths of our study are a relatively large sample size of women with PCOS as well as the detailed metabolic characterization, which includes glucose and insulin levels derived from oral glucose tolerance tests. We have presented evidence that women with PCOS and elevated FT levels have an adverse metabolic phenotype compared with women who have elevated DHEAS levels, which suggests that the distinction between adrenal and ovarian hyperandrogenism is important when evaluating metabolic risk in PCOS. Further studies are warranted to elucidate whether DHEAS is a simple biomarker reflecting hyperinsulinemia or whether there is a causal effect of high DHEAS levels on improved insulin sensitivity, which might have an impact on new therapeutic strategies in women with PCOS VOL. 98 NO. 5 / NOVEMBER 2012

8 Fertility and Sterility Moreover, longitudinal studies investigating the long-term metabolic and cardiovascular health of women affected by PCOS should pay attention to the source of the androgens contributing to hyperandrogenemia. REFERENCES 1. Diamanti-Kandarakis E, Kouli CR, Bergiele AT, Filandra FA, Tsianateli TC, Spina GG, et al. A survey of the polycystic ovary syndrome in the Greek island of Lesbos: hormonal and metabolic profile. J Clin Endocrinol Metab 1999; 84: Azziz R, Carmina E, Dewailly D, Diamanti-Kandarakis E, Escobar- Morreale HF, Futterweit W, et al. Criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: an androgen excess society guideline. J Clin Endocrinol Metab 2006;91: Jonard S, Dewailly D. The follicular excess in polycystic ovaries, due to intraovarian hyperandrogenism, may be the main culprit for the follicular arrest. Hum Reprod Update 2004;10: Chen MJ, Yang WS, Chen CL, Wu MY, Yang YS, Ho HN. The relationship between anti-m ullerian hormone, androgen and insulin resistance on the number of antral follicles in women with polycystic ovary syndrome. Hum Reprod 2008;23: Baptiste CG, Battista MC, Trottier A, Billargeon JP. Insulin and hyperandrogenism in women with polycystic ovary syndrome. J Steroid Biochem Mol Biol 2010;122: Carmine E. Ovarian and adrenal hyperandrogenism. Ann NY Acad Sci 2006; 1092: Carmina E, Lobo RA. Prevalence and metabolic characteristics of adrenal androgen excess in hyperandrogenic women with different phenotypes. J Endocrinol Invest 2007;30: Kumar A, Woods KS, Bartolucci AA, Azziz R. Prevalence of adrenal androgen excess in patients with the polycystic ovary syndrome (PCOS). Clin Endocrinol 2005;62: Buffington CK, Givens JR, Kitabchi AE. Opposing actions of dehydroepiandrosterone and testosterone on insulin sensitivity: in vivo and in vitro studies of hyperandrogenic females. Diabetes 1991;40: Brennan K, Huang A, Azziz R. Dehydroepiandrosterone sulfate and insulin resistance in patients with polycystic ovary syndrome. Fertil Steril 2009;91: Wehr E, M oller R, Horejsi R, Giuliani A, Kopera D, Schweighofer N, et al. Subcutaneous adipose tissue topography and metabolic disturbances in polycystic ovary syndrome. Wien Klin Wochenschr 2009;121: Schunkert H, Hense HW, Andus T, Riegger GA, Straub RH. Relation between dehydroepiandrosterone sulfate and blood pressure levels in a populationbased sample. Am J Hypertens 1999;12: Chen MJ, Chen CD, Yang JH, Chen CL, Ho HN, Yang WS, et al. High serum dehydroepiandrosterone sulfate is associated with phenotypic acne and a reduced risk of abdominal obesity in women with polycystic ovary syndrome. Hum Reprod 2011;26: K oş uş N, K oş uş A, Kamalak Z, Hızlı D, Turhan NO. Impact of adrenal versus ovarian androgen ratio on signs and symptoms of polycystic ovarian syndrome. Gynecol Endocrinol 2012;28: Rotterdam ESHRE/ASRM-Sponsored Consensus Workshop Group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 2004;19: Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 1999;84: National Cholesterol Education Program (NCEP) Expert Panel on Detection E, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel III). Circulation 2002;106: Orentreich N, Brind JL, Rizer RL, Vogelman JH. Age changes and sex differences in serum dehydroepiandrosterone sulfate concentrations throughout adulthood. J Clin Endocrinol Metab 1984;59: Paolisso G, Ammendola S, Rotondi M, Gambardella A, Rizzo MR, Mazziotti GT, et al. Insulin resistance and advancing age: what role for dehydroepiandrosterone sulfate? Metabolism 1997;46: Walker BR, Rodin A, Taylor NF, Clayton RN. Endogenous inhibitors of 11b-hydroxysteroid dehydrogenase type 1 do not explain abnormal cortisol metabolism in polycystic ovary syndrome. Clin Endocrinol 2000;52: Tsilchorozidou T, Honour JW, Conway GS. Altered cortisol metabolism in polycystic ovary syndrome: insulin enhances 5a-reduction but not the elevated adrenal steroid production rates. J Clin Endocrinol Metab 2003;88: Azziz R, Nestler JE, Dewailly D. Androgen excess disorders in women. Philadelphia: Lippincott-Raven; Yamashita R, Saito T, Satoh S, Aoki K, Kaburagi Y, Sekihara H. Effects of dehydroepiandrosterone on gluconeogenic enzymes and glucose uptake in human hepatoma cell line, HepG2. Endocr J 2005;52: Farah-Eways L, Reyna R, Knochenhauer ES, Bartolucci AA, Azziz R. Glucose action and adrenocortical biosynthesis in women with polycystic ovary syndrome. Fertil Steril 2004;81: Savastano S, Belfiore A, Guida B, Angrisani L, Orio F Jr, Cascella T, et al. Role of dehydroepiandrosterone sulfate levels on body composition after laparoscopic adjustable gastric banding in pre-menopausal morbidly obese women. J Endocrinol Invest 2005;28: Kameda W, Daimon M, Oizumi T, Jimbu Y, Kimura M, Hirata A, et al. Association of decrease in serum dehydroepiandrosterone sulfate levels with the progression to type 2 diabetes in men of a Japanese population: the Funagata study. Metabolism 2005;54: de Heredia FP, Cerezo D, Zamora S, Garaulet M. Effect of dehydroepiandrosterone on protein and fat digestibility, body protein and muscular composition in high-fat-diet-fed old rats. Br J Nutr 2007;97: Sanchez J, Perez-Heredia F, Priego T, Portillo MP, Zamora S, Garaulet M, et al. Dehydroepiandrosterone prevents age-associated alterations, increasing insulin sensitivity. J Nutr Biochem 2008;19: Arlt W, Allolio B. Adrenal insufficiency. Lancet 2003;361: Morales AJ, Nolan JJ, Nelson JC, Yen SS. Effects of replacement dose of dehydroepiandrosterone in men and women of advancing age. J Clin Endocrinol Metab 1994;78: Allolio B, Arlt W. DHEA treatment: myth or reality? Trends Endocrinol Metab 2002;13: Rosner W, Vesper H, Endocrine Society; American Association for Clinical Chemistry; American Association of Clinical Endocrinologists; Androgen Excess/PCOS Society; American Society for Bone and Mineral Research; American Society for Reproductive Medicine; American Urological Association; Association of Public Health Laboratories; Endocrine Society; Laboratory Corporation of America; North American Menopause Society; Pediatric Endocrine Society. Toward excellence in testosterone testing: a consensus statement. J Clin Endocrinol Metab 2010;95: VOL. 98 NO. 5 / NOVEMBER

9 ORIGINAL ARTICLE: REPRODUCTIVE ENDOCRINOLOGY SUPPLEMENTAL TABLE 1 Correlation analyses of DHEAS/FT-ratio, DHEAS, and FT levels with endocrine and metabolic parameters in women with polycystic ovary syndrome (n [ 622). Parameter DHEAS/FT-ratio DHEAS Free testosterone Correlation coefficient P value Correlation coefficient P value Correlation coefficient P value TT < < <.001 SHBG < < <.001 Age < <.001 BMI < <.001 WC < <.001 WHR <.001 BP Systolic <.001 Diastolic < <.001 Glucose Fasting < <.001 1h h < <.001 AUC <.001 Insulin Fasting < <.001 1h < <.001 2h <.001 HOMA-IR < <.001 AUC < <.001 TC Triglycerides < <.001 HDL <.001 TC/HDL <.001 LDL HbA1c <.001 FG score <.001 Note: Data were analyzed by Pearson correlation analyses. AUC ¼ area under the curve; BMI ¼ body mass index; BP ¼ blood pressure; CRP ¼ C-reactive protein; DHEAS ¼ dehydroepiandrosterone sulfate; FG score ¼ Ferriman-Gallwey score; FT ¼ free testosterone; HDL ¼ high-density lipoprotein; HOMA-IR ¼ homeostatic model assessment-insulin resistance; LDL ¼ low-density lipoprotein; PCOS ¼ polycystic ovary syndrome; SHBG ¼ sex hormone-binding globulin; TC ¼ total cholesterol; TT ¼ total testosterone; WC ¼ waist circumference; WHR ¼ waist-to-hip ratio e1 VOL. 98 NO. 5 / NOVEMBER 2012