Insulin resistance in polycystic ovary syndrome: hyperandrogenemia versus normoandrogenemia

European Journal of Obstetrics & Gynecology and Reproductive Biology 100 (2001) 62±66

Insulin resistance in polycystic ovary syndrome: hyperandrogenemia versus normoandrogenemia BuÈlent Okan Yõldõz*, Olcay Gedik Department of Internal Medicine, Division of Endocrinology, Hacettepe University School of Medicine, Ankara, Turkey Received 5 May 2000; accepted 14 June 2001

Abstract Objective: The goal of this study was to evaluate the insulin resistance and glucose tolerance in hyperandrogenemic and normoandrogenemic groups of patients with polycystic ovary syndrome (PCOS). Study design: In this cross-sectional study, 17 hyperandrogenemic and 14 normoandrogenemic, age and weight-matched non-obese women with PCOS were studied. All patients had clinical hyperandrogenism and chronic anovulation with polycystic ovaries on ultrasound. Insulin resistance and glucose tolerance were determined by measuring insulin and glucose concentrations following a 75 g oral glucose tolerance test (OGTT). Fasting glucose to insulin ratio (FG:I ratio), insulin area under the curve (AUCinsulin) during OGTT, and homeostasis model assessment for insulin resistance (HOMA-IR) were calculated. Results: Hyperandrogenemic group of patients had fasting hyperinsulinemia, lower FG:I ratio, higher AUCinsulin, and HOMA-IR compared with normoandrogenemic group. The differences between two groups were statistically signi®cant. Conclusion: PCOS has variable biochemical features. Hyperandrogenemia associated with insulin resistance differs from normoandrogenemia in this syndrome. Fasting insulin concentrations, FG:I ratio, AUCinsulin, and HOMA-IR are convenient markers for determining insulin resistance in PCOS. # 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Polycystic ovary syndrome; Insulin resistance; Normoandrogenemia

1. Introduction Polycystic ovary syndrome (PCOS) is probably the most common endocrine disorder, affecting women of reproductive age with 5±10% prevalence estimate. Its etiology remains unknown. The most widely accepted diagnostic criteria for PCOS are (1) ovulatory dysfunction (2) clinical evidence of hyperandrogenism and/or hyperandrogenemia and (3) exclusion of other known disorders, such as nonclassic adrenal 21-hydroxylase de®ciency, hyperprolactinemia, or androgen-secreting neoplasms [1]. The alternative diagnostic approach is primarily based on ovarian ultrasonographic appearances and combined with certain characteristic symptoms such as menstrual disturbances, infertility, hirsutism and obesity [2]. Polycystic ovaries are de®ned on ultrasound by the presence of eight or more subcapsular follicular cysts 10 mm or less in diameter and increased ovarian stroma; however, these changes can be present in women who are entirely endocrinologically normal [3]. Hyperandrogenemia and clinical hyperandrogenism are *

Corresponding author. Present address: Emekli Subay Evleri, 2. Cad 23 C Blok No: 15/2, YuÈcetepe, Ankara, Turkey. Tel.: ‡90-312-2320399. E-mail address: [email protected] (B.O. Yõldõz).

included in the diagnostic criteria of PCOS separately [1]. Not all patients present simultaneously with all features and various combinations are possible. Women with PCOS are at increased risk of developing impaired glucose tolerance (IGT) and Type 2 DM. Insulin resistance, defects in insulin secretion, familial and environmental factors are considered as possible components in the predisposition to Type 2 DM in PCOS [4]. In 1980, Burghen et al. reported that women with PCOS had basal and glucose stimulated hyperinsulinemia compared with weight-matched control women, suggesting the presence of insulin resistance. They noted signi®cant positive linear correlations between insulin and androgen [5]. An association between basal hyperinsulinemia and elevated androgen levels has been consistently demonstrated in both obese and non-obese women with PCOS [6±8]. However, the differences in insulin secretion and sensitivity in hyperandrogenemic and normoandrogenemic groups of patients with PCOS has not been evaluated yet. A variety of procedures have been used to detect the presence of insulin resistance. Although the most widely accepted research method is the euglycemic insulin clamp technique, the complexity and the cost of the procedure limits its use [9]. There is signi®cant correlation between

0301-2115/01/$ ± see front matter # 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 0 1 - 2 1 1 5 ( 0 1 ) 0 0 4 5 8 - 4

B.O. Yõldõz, O. Gedik / European Journal of Obstetrics & Gynecology and Reproductive Biology 100 (2001) 62±66

fasting insulin levels and insulin action as measured by clamp technique. The estimate of insulin resistance obtained by homeostasis model assessment (HOMA) also is correlated with estimates obtained by the euglycemic clamp technique [10]. Legro et al. reported that fasting glucose to insulin ratio (FG:I ratio) is a useful measure of insulin sensitivity in women with PCOS [11]. Yeni-Komshian et al. recently de®ned the degree of correlation between a quantitative assessment of insulin resistance and various surrogate estimates in 490 non-diabetic individuals. The results of this study showed that the total insulin response during an OGTT provides the best surrogate measure of insulin resistance [12]. Some PCOS patients with clinical hyperandrogenism and chronic anovulation do not have elevated serum androgen levels. In this study, we attempted to determine whether the insulin secretion and sensitivity differs in hyperandrogenemic and normoandrogenemic groups of patients with PCOS by measuring serum insulin responses to oral glucose load, FG:I ratio and performing HOMA. One aim of the study was to evaluate whether these tests are convenient to evaluate insulin resistance. 2. Materials and methods 2.1. Subjects A total of 31 patients with PCOS were studied. The diagnosis of PCOS was based on clinical features of menstrual irregularity (oligomenorrhea 6 weeks-6 months, or amenorrhea >6 months), hirsutism (Ferriman±Gallwey score >7), elevated LH/FSH ratio (>2), chronic anovulation (progesterone <5 ng/ml) and ovarian ultrasonographic appearances de®ned by Adams et al. [3]. Because the patients were divided into two groups based on their androgenic hormone pro®les, these were not included in the diagnostic criteria. All patients were euthyroid and for at least 3 months before the study, were not taking any medication known to affect sex hormone or carbohydrate metabolism. Nonclassical 21-hydroxylase de®ciency, hyperprolactinemia and androgen-secreting tumors were excluded by appropriate tests. Patients were divided into two groups based on their testosterone and free testosterone levels. Taking into account the bioavailability of androgens, free testosterone was chosen to be measured as the most representative ovarian androgen. The body mass index (BMI: weight (kg)/height2 (m2)) and the waist to hip ratio (WHR) were measured. Waist to hip ratio was calculated from the circumferences measured in duplicate in the supine position (waist, midway between the lower rib margin and the iliac crest; hip, widest circumference over the great trochanters). Group 1, hyperandrogenemic group (HA) consisted of 17 women. The mean age (S.D.) was 24:1  4:2 (range 19± 36), the mean BMI was 22:5  2:1 kg/m2, the mean WHR

63

was 0:81  0:03 and seven patients in this group, had history of diabetes mellitus in the ®rst degree relatives. Group 2, normoandrogenemic group (NA) consisted of 14 women. The mean age (S.D.) was 23:3  1:8 (range 20±27), the mean BMI was 22:3  1:6 kg/m2, the mean WHR was 0:75  0:03 and four patients in this group, had history of diabetes mellitus in the ®rst degree relatives. All patients gave informed consent according to the Helsinki Committee requirements. 2.2. Hormone and blood glucose assays The LH, FSH, testosterone, and DHEAS were measured by chemiluminescent enzyme immunoassay (Immulite 2000 Diagnostic Products Corporation CA USA) with an average interassay coef®cient of variation (CV) 8% and intraassay CV 7%. Free testosterone and insulin were measured by radioimmunoassay (Diagnostic Systems Lab. Inc. Texas, USA). Interassay CV's were 8.35 and 5%, intraassay CV's were 8.8 and 6.4%, respectively. Glucose was measured by the glucose oxidase technique (Boehringer±Mannheim, Germany). 2.3. Study protocol All the studies were carried out within the ®rst 7 days of menstruation or at random if 3 or more months of amenorrhea were present. The studies were performed after a 3 day 300 g carbohydrate diet and an overnight fast. A 75 g oral glucose load was administered and blood was obtained for glucose and insulin determinations at 0, 30, 60, 90 and 120 min through an intravenous catheter. Using glucose and insulin data from the oral glucose tolerance test (OGTT), following calculations are performed: glucose and insulin response areas under the glucose and insulin curves during OGTT (AUCglucose and AUCinsulin), FG:I ratio (mg/10 4U), the ratio of the 30 min insulin increment to the 30 min glucose increment. DI30 30 min insulin ˆ DG30 30 min glucose

0 min insulin …pmol=l† 0 min glucose …mmol=l†

The HOMA model [12] is applied by using the following formulas: insulin resistance …HOMA IR† fasting insulin …mU=ml†  fasting glucose …mmol=l† ˆ 22:5 b-cell function …HOMA b-cell† ˆ

20  fasting insulin …mU=ml† fasting glucose …mmol=l† 3:5

2.4. Statistical analysis Data analysis was performed by using the SPSS 7.5 PC Package (SPSS Inc. IL, USA). All parameters were given as

64

B.O. Yõldõz, O. Gedik / European Journal of Obstetrics & Gynecology and Reproductive Biology 100 (2001) 62±66

Table 1 The clinical, hormonal and metabolic features of the patientsa HA n Age BMI WHR* LH/FSH Testosterone (nmol/l) (normal: 0.63±1.2)* Free testosterone (pmol/l) (normal: 0.29±3.18)* DHEAS (mg/dl) (normal: 35±430)* Fasting glucose (mg/dl) Fasting insulin (mU/ml)* AUCglucose (mg/dl/min) AUCinsulin (mU/ml/min)** FG:I (mg/10 4 U)*** HOMA IR* HOMA b-cell DI30/DG30 (pmol/mmol)

17 24.1 22.5 0.81 2.6 1.49

NA     

4.2 2.1 0.03 0.4 0.24

3.52  0.28 343 88 7.57 13956 7737 12.6 1.64 100.06 136.6

        

62 5 2.7 2039 775 3.4 0.65 27.7 39.7

14 23.3 22.3 0.75 2.3 0.80

    

1.8 1.6 0.03 0.2 0.17

1.54  0.53 213 82 4.2 12773 4238 21.8 0.77 137.06 107.1

        

64 10 1.9 1079 1131 6.9 0.28 80.08 45.2

a

Values are means  S:D. * P < 0:01 (HA vs. NA). ** P < 0:001 (HA vs. NA). *** P < 0:05 (HA vs. NA).

mean  S:D. The data were normally distributed and the non-parametric two samples Mann±Whitney test was used to determine signi®cant differences between the groups. Spearman correlations were used in correlation analysis. 3. Results The clinical, hormonal and metabolic features of the patients are shown in Table 1. The age and BMI were well matched in the two groups. The mean WHR was signi®cantly higher in HA group (0:81  0:03 versus 0:75  0:03 P < 0:01). The hyperandrogenemic group appeared to have higher DHEAS levels in normal limits (P < 0:01). All subjects had normal blood glucose responses in the OGTT. The fasting insulin, AUCinsulin, and HOMA IR were signi®cantly higher in HA group (P < 0:01, P < 0:001, P < 0:01, respectively). The fasting glucose to insulin ratio was signi®cantly lower in HA group (P < 0:05). Although higher fasting glucose, AUCglucose, DI30/DG30 levels and

lower HOMA b-cell values were found in HA group, there were no statistically signi®cant differences between the two groups. Spearman correlations between metabolic variables are shown in Table 2 (data is given for the study group as a whole). Fasting glucose was correlated with HOMA b-cell negatively and with DI30/DG30 positively (r ˆ 0:454, P < 0:05; r ˆ 0:492, P < 0:01, respectively). Fasting insulin was correlated with fasting glucose to insulin ratio negatively and with HOMA IR positively (r ˆ 0:962, P < 0:01; r ˆ 0:936, P < 0:01, respectively). Fasting glucose to insulin ratio was correlated with HOMA b-cell and HOMA IR negatively (r ˆ 0:314, P < 0:05; r ˆ 0:860, P < 0:01, respectively). The markers of insulin secretion (HOMA b-cell and DI30/DG30) were correlated negatively (r ˆ 0:467, P < 0:01). 4. Comment In this comparative cross-sectional study, we evaluated the insulin secretion and sensitivity in age and weightmatched non-obese hyperandrogenemic and normoandrogenemic two groups of patients with PCOS. It was ®rst reported in 1980 that women with PCOS had hyperinsulinemia, suggesting that they were insulin resistant. Burghen et al. evaluated plasma insulin, androstenedione and testosterone in obese women (eight with PCOS, six controls) and they found that hyperandrogenism correlated with hyperinsulinism [5]. Subsequently, insulin resistance is observed frequently in women with PCOS. Both obese and non-obese PCOS women have b-cell dysfunction as well as insulin resistance. However, this is not associated with glucose intolerance in the majority of PCOS women [1]. In a recent prospective study of 254 PCOS patients, the prevalence of IGT and diabetes were 10.3 and 1.5%, respectively, in nonobese PCOS women [13]. Defects in insulin action in PCOS appear to represent intrinsic abnormalities that are independent of obesity, metabolic derangements, body fat topography and sex hormone levels [14]. Hyperandrogenism and insulin resistance are independent features of PCOS, obesity exacerbates the clinical expression of hyperandrogenism [1]. There are many trials in the literature evaluating the insulin resistance, androgen levels and obesity relationship

Table 2 Spearman correlations between metabolic variables

Fasting glucose Fasting insulin Fasting glucose to insulin ratio HOMA b-cell HOMA IR *

P < 0:05. P < 0:01.

**

Fasting glucose

Fasting insulin

± ± ± ± ±

0.229 ± ± ± ±

Fasting glucose to insulin ratio 0.007 0.962** ± ± ±

HOMA b-cell 0.454* 0.151 0.314* ± ±

HOMA IR

DI30/DG30

0.345 0.936** 0.860** 0.095 ±

0.492** 0.332 0.223 0.467** 0.349

B.O. Yõldõz, O. Gedik / European Journal of Obstetrics & Gynecology and Reproductive Biology 100 (2001) 62±66

in PCOS. Hyperandrogenism in PCOS may contribute to the associated insulin resistance. Moghetti et al. found that insulin resistance in women with hyperandrogenism is partially reversed by antiandrogen treatment [15]. Studies in which insulin levels have been lowered by a variety of modalities indicate that hyperinsulinemia augments androgen production in PCOS. Furthermore, in many women with PCOS, lowering insulin levels ameliorates but does not abolish hyperandrogenism [1]. Most of the trials in the literature have compared insulin sensitivity and secretion in hyperandrogenemic PCOS patients and control groups. In a study by Ke et al., hyperandrogenemic PCOS patients showed higher insulin and C-peptide responses to OGTT than controls. This study suggested that early insulin release after ingestion of glucose is enhanced in PCOS [16]. In our study, markers of insulin secretion ((HOMA b-cell and DI30/DG30) did not differ between hyperandrogenemic and normoandrogenemic groups of patients with PCOS. Sharp et al. studied 47 overweight PCOS and 21 control women with a similar study design. They found that basal insulin and sum of insulin levels during OGTT in women with PCOS were not signi®cantly different form the control group. Within the PCOS group, basal insulin levels were higher in women with oligo-amenorrhea than in those with regular ovulatory menses despite similarly raised androgen levels [17]. There is one trial that compared subjects with polycystic ovaries (PCO) without hyperandrogenemia and PCOS patients. This study showed that abnormalities of glucose and insulin metabolism following OGTT were equally prevalent among women diagnosed having PCO compared to those with PCOS [18]. We compared the hyperandrogenemic and normoandrogenemic PCOS patients. All of our patients had PCOS diagnosis with clinical hyperandrogenism and chronic anovulation. They also had the polycystic ovary morphology consistent with the diagnosis of the syndrome. Patients with hyperandrogenemia showed higher insulin levels, higher WHR when compared to normoandrogenemic ones. Pasquali et al. showed that women with PCOS have high WHR compared to control group independent of obesity and increased amount of truncal-abdominal fat is closely associated with hyperinsulinemia [19]. Our data suggest that hyperandrogenemia with increased amount of truncal-abdominal fat (high WHR) is associated with hyperinsulinism and insulin resistance in patients with PCOS. Some PCOS patients do not develop signs of androgen excess because of genetic differences in target tissue number and sensitivity to androgens [20]. Although our preliminary study has small sample size, it suggests that the opposite may be true. Presence of normal serum androgen levels in PCOS patients with clinical hyperandrogenism is associated with lower insulin levels and lower WHR. Obesity is not a confounding factor in our study because BMI values of all patients are in normal range. None of the patients had IGT or diabetes in both groups. It can be explained by young age, low BMI values and

65

ethnicity. All patients, especially the women with family history of diabetes are candidates to develop diabetes in the future. In conclusion, although all PCOS patients in this study have clinical hyperandrogenism, chronic anovulation and polycystic ovary morphology on ultrasound, hyperandrogenemic group differs from normoandrogenemic group in relation to insulin levels, insulin resistance and WHR. However, we accept that the small numbers in each subgroup will inevitably contribute to reduced power for our study. For practical purposes, the insulin response to OGTT is suf®ciently sensitive to show glucose tolerance. The diagnosis of insulin resistance in PCOS can be easily determined by the insulin response to OGTT. Fasting hyperinsulinemia, the FG:I ratio and HOMA IR can be considered adequate and sensitive markers to determine insulin resistance in PCOS women in association with hyperandrogenemia. Further studies with large sample sizes are needed to test the suggestions of our study. References [1] Dunaif A. Insulin resistance and the polycystic ovary syndrome: mechanism and implications for pathogenesis. Endocr Rev 1997; 18(6):774±800. [2] Conway GS, Honour JW, Jacobs HS. Heterogeneity of the polycystic ovary syndrome: clinical, endocrine and ultrasound features in 556 patients. Clin Endocrinol 1989;30:459±70. [3] Polson DW, Wadsworth J, Adams J, Franks S. Polycystic ovariesÐa common finding in normal women. Lancet 1988;1:870±2. [4] Ehrmann DA, Sturis J, Byrne MM, Karrison T, Rosenfield RL. Insulin secretory defects in polycystic ovary syndrome. Relationship to insulin sensitivity and family history of non-insulin dependent diabetes mellitus. J Clin Invest 1995;96:520±7. [5] Burghen GA, Givens JR, Kitabchi AE. Correlation of hyperandrogenism with hyperinsulinism in polycystic ovarian disease. J Clin Endocrinol Metab 1980;50:113±6. [6] Holte J. Disturbances in insulin secretion and sensitivity in women with the polycystic ovary syndrome. Baillieres Clin Endocrinol Metab 1996;10(2):221±48. [7] Rittmaster RS, Deshwal N, Lehman L. The role of adrenal hyperandrogenism, insulin resistance, and obesity in the pathogenesis of polycystic ovary syndrome. J Clin Endocrinol Metab 1993;76:1295± 300. [8] Dunaif A, Segal KR, Futterweit W, Dobrjansky A. Profound peripheral insulin resistance, independent of obesity in polycystic ovarian syndrome. Diabetes 1989;38:1165±74. [9] Defronzo RA, Ferrannini E. Insulin resistance: a multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 1991;14(3):173±94. [10] Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and b-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412±9. [11] Legro RS, Finegood D, Dunaif A. A fasting glucose to insulin ratio is a useful measure of insulin sensitivity in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1998;83:2694±8. [12] Yeni-Komshian H, Carantoni M, Abbasi F, Reaven G. Relationship between several surrogate estimates of insulin resistance and quantification of insulin-mediated glucose disposal in 490 healthy nondiabetic volunteers. Diabetes Care 2000;23:171±5.

66

B.O. Yõldõz, O. Gedik / European Journal of Obstetrics & Gynecology and Reproductive Biology 100 (2001) 62±66

[13] Legro RS, Kunselman AR, Dodson WC, Dunaif A. Prevalence and predictors of risk fot type 2 diabetes mellitus and impaired glucose tolerance in polycystic ovary syndrome: a prospective controlled study in 254 affected women. J Clin Endocrinol Metab 1999;84:165±9. [14] Dunaif A, Segal RK, Shelley DR, Green G, Dobrjansky A, Licholai T. Evidence for distinctive and intrinsic defects in insulin action in polycystic ovary syndrome. Diabetes 1992;41:1257±66. [15] Moghetti P, Tosi F, Castello R, et al. The insulin resistance in women with hyperandrogenism is partially reversed by antiandrogen treatment: evidence that androgens impair insulin action in women. J Clin Endocrinol Metab 1996;81:952±960. [16] Ke WX, Shan GQ, Hua SY. Different responses of insulin, C-peptide, and testosterone to an oral glucose tolerance test in two groups of women with polycystic ovarian syndrome. Acta Obstet Gynecol Scand 1996;75(2):166±9.

[17] Sharp PS, Kiddy DS, Reed MJ, Anyaoku V, Johnston DG, Franks S. Correlation of plasma insulin and insulin-like growth factor-I with indices of androgen transport and metabolism in women with polycystic ovary syndrome. Clin Endocrinol 1991;35:253±7. [18] Norman RJ, Hague WM, Masters SC, Wang XJ. Subjects with polycystic ovaries without hyperandrogenaemia exhibit similar disturbances in insulin and lipid profiles as those with polycystic ovary syndrome. Hum Reprod 1995;10(9):2258±61. [19] Pasquali R, Casimirri F, Venturoli S. Body fat distribution has weight-independent effects on clinical, hormonal and metabolic features of women with polycystic ovary syndrome. Metabolism 1994;43:706±13. [20] Lobo RA, Goebelsmann U, Hortan R. Evidence for the importance of peripheral tissue events in the development of hirsutism in polycystic ovary syndrome. J Clin Endocrinol Metab 1983;57:393±7.