Characteristics of different phenotypes of polycystic ovary syndrome based on the Rotterdam criteria in a large-scale Chinese population

DOI: 10.1111/j.1471-0528.2009.02347.x www.bjog.org Fertility and assisted reproduction Characteristics of different phenotypes of polycystic ovary syndrome based on the Rotterdam criteria in a large-scale Chinese population HY Zhang, FF Zhu, J Xiong, XB Shi, SX Fu Department of Obstetrics and Gynecology, Second Xiangya Hospital of Central South University, Hunan, China Correspondence: Dr HY Zhang, Department of Obstetrics and Gynecology, Second Xiangya Hospital, 139 Ren Min Zhong Lu, Changsha, Hunan 410011, China. Email zhanghongyuan830@hotmail.com Accepted 16 July 2009. Published Online 14 September 2009. Objective To analyse the phenotypic spectrum of polycystic ovary syndrome (PCOS) and determine the association between metabolic, hormonal and new ultrasonographic criteria. Design Clinical cross-sectional study. Setting University teaching hospital. Population A total of 804 Chinese women, among whom 719 cases were diagnosed as PCOS based on the 2003 Rotterdam criteria. Eighty-five women with regular menstrual cycles and without hyperandrogenism were recruited as controls. Methods PCOS patients were divided into four subgroups: (i) oligoand/or anovulation (O), hyperandrogenism (H), and polycystic ovary morphology (P); (ii) O + H; (iii) H + P; and (iv) O + P. Main Outcome Measurements Clinical history, ultrasonographic (ovarian follicle number and volume), hormonal and metabolic parameters. Results The composition of the two new phenotypes created by the European Society for Human Reproduction and Embryology/ The American Society for Reproductive Medicine (ESHRE/ASRM) 2003 was 65.6% (O + P and H + P). BMI and F-G scores were highest in the O + H + P group and lowest in O + P and controls. Serum testosterone concentrations and insulin resistance were highest in cases with O + H + P and O + H, intermediate in cases with H + P, and lowest in cases with O + P and controls. The prevalence of metabolic syndrome in the five groups was 28.5% (O + H + P), 25.5% (O + H), 8.3% (H + P), 7.2% (O + P) and 3.5% (controls), respectively. Conclusions Nonclassic phenotypes for PCOS (O + P, H + P and O + H + P) were more frequent than the classic phenotype (O + H). The nonhyperandrogenic PCOS phenotype (O + P), one of the new phenotypes created by the Rotterdam criteria, may represent a form of PCOS associated with milder metabolic profile compared with the other phenotypes. Keywords Chinese women, hyperandrogenism, metabolic syndrome, polycystic ovaries, polycystic ovary syndrome. Please cite this paper as: Zhang H, Zhu F, Xiong J, Shi X, Fu S. Characteristics of different phenotypes of polycystic ovary syndrome based on the Rotterdam criteria in a large-scale Chinese population. BJOG 2009;116:1633 1639. Introduction The polycystic ovary syndrome (PCOS) was first reported in 1935 1 and has been considered to be the most common endocrinopathy in women, presenting in at least 5 7% of women of reproductive age. 2,3 Comparing the phenotypes of PCOS around the world has been extremely difficult because of the lack of agreement on standardised criteria to make the diagnosis. An international consensus workshop in Rotterdam, sponsored by The American Society for Reproductive Medicine (ASRM) and European Society for Human Reproduction and Embryology (ESHRE) attempted to reconcile these differences by expanding the diagnostic criteria for PCOS with the addition of the ultrasound assessment of ovarian morphology. 4,5 According to these new criteria, PCOS can be defined when at least two of the following three features are present: (i) oligo- and/or anovulation (O), (ii) clinical (hirsutism) and/or biochemical signs of hyperandrogenism (H), and (iii) polycystic ovaries on pelvic ultrasound (P) and exclusion of other identifiable endocrine disorders such as late-onset congenital adrenal hyperplasia, hyperprolactinemia, thyroid dysfunction, ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology 1633

Zhang et al. neoplastic androgen secretion, or drug-induced androgen excess. Using these Rotterdam criteria, four phenotypes of PCOS cases can be identified, (i.e.) O + H + P, O + H, H + P, and O + P. Currently there are limited data on clinical characteristics and endocrine metabolic features in women belonging to the novel PCOS phenotypes defined by the Rotterdam criteria 6 and documenting differences in subjects within these four PCOS groupings is important to determine whether their phenotypes are distinct or overlapping with each other or with unaffected women. 7,8 Moreover, there are significant ethnic and racial variations in the clinical presentation of PCOS, the frequency of hirsutism, acne, polycysticappearing ovaries, obesity, and insulin resistance. 9 For these reasons, we have studied a large population of women with PCOS in China. We therefore set out to contrast women with PCOS defined by (i) O + H + P; (ii) O + H with normal ovarian morphology; (iii) H + P with regular menses; and (iv) O + P in the absence of hyperandrogenism. Materials and methods Subjects Women with PCOS were ascertained from infertility and endocrine clinics, predominantly at the Second Xiangya Hospital, Central South University in Changsha. Women were considered affected if they had polycystic ovaries on ultrasound, after presentation with menstrual disturbances (oligo- or amenorrhoea) and/or hyperandrogenism. A diagnosis of hyperandrogenism required clinical and/or biochemical evidence (see below). Other potential endocrine and neoplastic causes of hyperandrogenemia were excluded. This diagnosis is consistent with the Rotterdam consensus criteria. 4 Subjects were divided into four PCOS subgroups based on the criteria outlined in the Rotterdam PCOS consensus workshop. 4 The groupings included (i) O (fewer than nine menstrual periods per year) + H (elevated Ferriman-Gallwey score or androgen level) + P (at least one ovary >10 ml or at least 12 follicles 2 9 mm in diameter); (ii) O + H without polycystic ovaries; (iii) H + P with regular menstrual cycles of 21 35 days; and (iv) O + P with no hyperandrogenemia. None of these 719 women were taking medication that could confound the clinical and endocrine presentation and therefore complicate the assignment of phenotypic subgroups (A total of 212 women from a total of 931 women with PCOS were excluded on this basis). We also included a control group comprising 85 Chinese women without PCOS. These women were either healthy volunteers or recruits from an infertility clinic (in whom the cause of infertility had been established due to male factors and/or tubal disease). All control women had ultrasound-proven normal ovarian morphology and regular menstrual cycles and none had hirsutism. Therefore, none of the control women had PCOS according to the 2003 Rotterdam criteria. None of them received oral contraceptives or other drugs that could interfere with the hormonal and metabolic studies. Control subjects were not sisters of subjects fulfilling the Rotterdam criteria for the diagnosis of PCOS. 4 This study was based on routine clinical practice. Institutional review board approval was obtained for this study, and informed consent was obtained from individuals who participated in this study. Methods Clinical assessment Personal medical history was obtained from every woman according to a customised prepared questionnaire. Menstrual cycle history was carefully documented and included a general review since menarche and a detailed recall of the last 2 to 3 year interval. Ovulatory dysfunction was defined as less than eight cycles per year, and regular menstrual cycle as 21 35 days in length. Physical examination was performed in each person by two doctors. Body mass index (BMI) was defined as body weight in kilograms divided by body height in metres square (kg/m 2 ). Obesity was defined as BMI 25 kg/m 2, according to the Asia-Pacific definition. 10 Body fat distribution was assessed by measurements of the waist to hip ratio (WHR). 11 Transvaginal ultrasound (LogIQ200 Pro series ultrasonic machine, Ge Company, Niskayuna, NY, China) was used to detect polycystic ovaries, defined as the presence of at least one ovary >10 ml or at least 12 follicles 2 9 mm in diameter. 12 Laboratory tests The fasting blood serum hormone levels were obtained during day 2 6 of the menstrual cycles (natural or bleeding after progestin withdrawal). If the patient had amenorrhoea, the examination was performed following the confirmation of dominant follicle absence with ultrasound. Serum endocrinology included measurement of folliclestimulating hormone (FSH), luteinising hormone (LH), prolactin (PRL), testosterone (T), and estradiol (E 2 ), as detected by chemiluminescence immunisation (Beckman Access Health Company, Chaska, MN, USA). Blood samples were collected for fasting insulin and fasting glucose examination. The level of insulin-resistance was calculated using the homeostasis model assessment (HOMA), which was calculated according to the formula [plasma glucose (mmol/l) insulin (lu/ml)]/22.5. Plasma glucose was detected by the glucose oxidase method (AU640 automation biochemistry analyzer and their relevant reagent, Olympus Company, Hamburg, Germany), and serum insulin levels were measured by chemiluminescence. 1634 ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology

Characteristics of different polycystic ovary syndrome phenotypes Fasting blood samples were also obtained for measurement of total cholesterol, HDL cholesterol, LDL cholesterol, and triglycerides (TG). Additionally, serum PRL, TSH, and 17-hydroxyprogesterone levels were detected in blood samples of cases with oligomenorrhoea to exclude other causes of menstrual disorders. Women were also assessed for the presence of metabolic syndrome, as defined by the recent 2005 international diagnostic consensus (International Diabetes Federation, IDF). 13 In women, an IDF-proscribed diagnosis of metabolic syndrome requires the presence of central obesity (waise circumference 80 cm), in addition to at least two of the following criteria: (i) elevated triglycerides ( 1.7 mmol/l); (ii) reduced HDL cholesterol (<1.29 mmol/l in women) or specific treatment for lipid abnormalities; (iii) elevated blood pressure (systolic BP 130 mmhg or diastolic BP 85 mmhg) or specific treatment of previously diagnosed hypertension; and (iv) impaired fasting plasma glucose 5.6 mmol/l or previously diagnosed type II diabetes. Hirsutism was defined by a modified Ferriman-Gallwey score >6. 14 Hyperandrogenism was defined when serum total testosterone was 0.6 ng/ml or higher, free testosterone was 2.5 pg/ml or greater. 15 Normal values for androgens were previously determined from a database analysis of nonobese, nonhirsute, ovulatory women living in southern China. Statistical analysis Data analysis was performed using the SPSS (SPSS, Inc., Chicago, IL, USA), Stata Version 13.0 (Stata Corp., College Station, TX, USA). Laboratory and clinical data were compared between groups using analysis of covariance (ANCOVA) to control for BMI. The Student s t-test was used for inter-group comparisons of continuous variables. Fisher s least significant difference (LSD) post hoc test was used to determine significant differences between groups. Categorical variables were compared using chi-square tests. Results are expressed as mean ± SEM and statistical significance for all analyses was defined as a two-tailed P-value of <0.05. Results Of the total number of subjects, 375 (52.2%) fulfilled the criteria for O + P, 193 (26.8%) for O + H + P, 96 (13.4%) for H + P, and 55 (7.6%) for O + H. Of the 719 PCOS patients, 248 (34.4%) met both the diagnosis criteria of NIH 1990 and ESHRE/ASRM 2003 (O + H + P and O + H), and the composition of the two new phenotypes created by the ESHRE/ASRM 2003 consensus was 65.6% (O + P and H + P). Subjects in each subset differed slightly in age (Table 1). BMI and WHR were increased in women with O + H and O + H + P compared with those with H + P, O + P, and controls. As expected, based on Table 1. Comparison of demographics and clinical endocrine features in women with polycystic ovary syndrome (PCOS) and control subjects O + H + P O + H H + P O + P Control n (%) 193 (26.8) 55 (7.6) 96 (13.4) 375 (52.2) 85 Age (years) 26 ± 4.9 25 ± 5.1 27 ± 3.7 26 ± 4.5 27 ± 5.3 BMI 36.5 ± 8.6 a 35.8 ± 9.3 a 30.9 ± 8.3 b 28.6 ± 6.5 b 27.3 ± 5.2 b WHR 0.85 ± 0.04 a 0.84 ± 0.03 a 0.82 ± 0.02 b 0.81 ± 0.03 b 0.82 ± 0.03 b F-G score 11.2 ± 6.7 a 8.4 ± 4.1 b 3.2 ± 1.8 c 4.9 ± 2.2 b 3.6 ± 1.5 c Number with acne (%) 82 a 66 b 62 b 23 c 19 c Ovarian volumes 13.7 ± 5.3 a 12.9 ± 4.7 a 11.2 ± 3.5 b 13.2 ± 4.8 c 6.9 ± 3.3 c Follicle count 12.2 ± 3.1 a 13.6 ± 4.3 a 11.5 ± 2.8 b 12.7 ± 3.6 c 8.5 ± 3.4 c Smoking (%) 21 (10.9) 7 (12.7) 14 (14.6) 34 (9.1) 5 (5.9) Alcohol (%) 6 (3.1) 4 (7.3) 5 (5.2) 18 (4.8) 3 (3.5) Family history of CAD (%) 42 (21.8) 15 (27.3) 14 (14.6) 64 (17.1) 27 (31.8) Family history of DM (%) 67 (34.7) 22 (40) 40 (41.7) 146 (38.9) 18 (21.2) Total testosterone (ng/dl) 89 ± 29 a 88 ± 33 a 79 ± 16 b 38 ± 16 c 37 ± 15 c Free testosterone 1.37 ± 0.85 a 1.29 ± 0.79 a 0.86 ± 0.37 b 0.43 ± 0.16 c 0.39 ± 0.23 c LH (miu/ml) 9.1 ± 4.9 a 8.9 ± 5.7 a 6.5 ± 2.3 c 7.3 ± 1.6 b 6 ± 1.3 c FSH (miu/ml) 6.4 ± 2.2 6.3 ± 2.6 6.1 ± 2.6 6.2 ± 2.3 6.2 ± 2.4 LH/FSH 1.7 ± 1.5 a 1.6 ± 1.3 a 1.2 ± 0.6 c 1.4 ± 0.4 b 1 ± 0.5 c BMI, body mass index; WHR, waist to hip ratio; CAD, coronary artery disease; DM, diabetes mellitus; H, hyperandrogenism; O, oligomenorrhoea; P, polycystic ovaries on ultrasound; FSH, follicle-stimulating hormone; LH, luteinising hormone. a c Significant differences between any two groups (P < 0.05), as demonstrated by Fisher least significant difference (LSD) post hoc tests, are indicated with different alphabets in superscript. ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology 1635

Zhang et al. the criteria that defined the categories of PCOS, there was a gradient in F-G scores with the highest score in the O + H + P group and the lowest score in O + P and controls. The prevalence of acne in PCOS cases was greatest in O + H + P group, intermediate in O + H and O + P group, and lowest in H + P group and controls. Subjects with PCOS defined as O + H and O + P had the largest ovarian volumes and the greatest number of follicles in a single plane on ultrasound, whereas the subjects with H + P had intermediate values, and all groups had greater volume and follicle number than controls. The endocrine parameters (gonadotrophins and androgens) in the four phenotypes and the control group are also depicted in Table 1. LH and the LH/FSH ratio were higher in cases with O + H + P, O + H and O + P compared with that in H + P group and controls, and were higher in cases with O + H + P and O + H than those with O + P. However, FSH levels have not significant difference among groups. Testosterone levels were highest in women with O + H + P and O + H, intermediate in women with H + P, and lowest in those with O + P and controls. The total cholesterol, LDL-cholesterol, and triglycerides levels were highest in cases with O + H and O + H + P, intermediate in cases with H + P, and lowest in cases with O + P and controls. There were no significant differences in the fasting glucose levels among the five groups. Insulin levels and HOMA for insulin resistance were highest in women with O + H and O + H + P compared with those with O + P and controls. Prevalence rates of metabolic syndrome (as defined by the 2005 IDF consensus) in the different subgroups were tracked with the insulin sensitivity evaluation (HOMA-IR, fasting insulin, triglycerides) described above. Notably, 28.5% of women in the O + H + P subgroup met these criteria (Table 2), as compared to 25.5% of O + H group, 8.3% of H + P group, 7.2% of O + P group and 3.5% of controls (chi-square comparisons of the four PCOS subgroups and the control group, P < 0.001.) Discussion PCOS is a heterogeneous disorder, and the Rotterdam criteria may have expanded the prevalence of this disease in the reproductive age female population by as much as 50% according to some estimates. 16 Identification of specific clinical patterns among women with PCOS has diagnostic implications in the search for specific genetic determinants and the development of apposite treatment regimens for various disease manifestations. 17 Based on the results of the present study of 719 PCOS patients, the composition of the two new phenotypes created by the Rotterdam criteria was 70.8%. The results were different from a previous study, 10 which found that of 406 PCOS patients enrolled based on the Rotterdam criteria, about 60% satisfied both the two diagnosis criteria, and only 30% were of the two new phenotypes. Thus, our study showed that the Rotterdam diagnostic criteria had enlarged the extent of PCOS particularly in our series of Chinese women. Androgen levels observed in this investigation bear a striking similarity to those reported by Welt et al. 18 in a PCOS population from Boston and Iceland. In that study, free and total testosterone also demonstrated a large variation with the NIH-defined population (O + H) demon- Table 2. Serum metabolic profile in women with polycystic ovary syndrome (PCOS) and control subjects O + H + P O + H H + P O + P Control Cholesterol (mg/dl) 180 ± 38 a 177 ± 40 a 178 ± 31 b 163 ± 26 c 161 ± 16 c HDL (mg/dl) 47.3 ± 9.8 a 48.4 ± 10.1 a 51.9 ± 10 c 54.5 ± 4.3 c 51.5 ± 3.5 c LDL (mg/dl) 118 ± 27 a 111 ± 36 a 107 ± 35 b 95 ± 14 c 98 ± 10 c TG (mg/dl) 95 ± 35 a 93 ± 38 a 76 ± 28 b 80 ± 11 c 82 ± 12 c Obesity (BMI 25) (%) 91 (43.3) 5 (33.3) 18 (32.7) 108 (24.6) 14 (16.5) WHR 0.85 (%) 68 (32.4) 4 (26.7) 13 (23.6) 79 (18) 3 (3.5) IFG (%) 27 (12.9) 2 (13.3) 5 (9.1) 31 (7.1) 4 (4.7) Fasting glucose (mg/dl) 93 ± 12 96 ± 25 96 ± 30 89 ± 10 90 ± 12 Fasting insulin (lu/ml) 17.1 ± 13.8 a 17.2 ± 13.6 a 16.3 ± 12.8 b 12.9 ± 10.3 c 11.8 ± 9.8 c HOMA-IR 5.86 ± 3.23 a 5.76 ± 6.6 a 4.63 ± 5.19 b 3.65 ± 2.97 c 3.24 ± 3.5 c Metabolic syndrome (2005 IDF criteria) (%) 55 (28.5) 14 (25.5) 8 (8.3) 27 (7.2) 3 (3.5) Values are mean ± SD. TG, triglycerides; HDL, high density lipoprotein; LDL, low density lipoprotein; BMI, body mass index; IFG, impaired fasting plasma glucose; WHR, waist to hip ratio; HOMA, homeostasis model assessment; H, hyperandrogenism; O, oligomenorrhoea; P, polycystic ovaries on ultrasound. a c Significant differences between any two groups (P < 0.05), as demonstrated by Fisher least significant difference (LSD) post hoc tests, are indicated with different alphabets in superscript. 1636 ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology

Characteristics of different polycystic ovary syndrome phenotypes strating the highest levels, O + P population the lowest and the H + P population an intermediate level. Dewailly et al. 19 (in a French population) and Barber et al. 20 (in a British population) also reported transitional levels of testosterone in the H + P patients, compared with O + H and O + P groups. Using the HOMA calculation as an index of insulin sensitivity, the three PCOS phenotypes with higher testosterone levels (O + H + P, O + H and H + P) demonstrated greater insulin resistance, compared with the O + P phenotype and control group. This contrasts with the Boston/ Iceland study in which only the O + H group was more insulin resistant. 21,22 The reasons behind the differences among studies are not readily apparent, but possible explanations for the disparities include genetic variation among ethnic populations, environmental factors (including obesity), and subtle variations in inclusion criteria (women with frank diabetes mellitus were excluded from our analysis). The metabolic syndrome is defined by both lipid and nonlipid criteria that identify individuals at increased risk for coronary heart disease and type II diabetes. 23 25 An increased risk of metabolic syndrome in women with hyperandrogenic PCOS has been reported. 26 The present study shows a trend toward higher risk of IFG and type II diabetes and higher prevalence of metabolic syndrome in all PCOS phenotypes, including women with the nonhyperandrogenic phenotype (O + P), compared with control subjects. Insulin resistance and compensatory hyperinsulinemia appears to play a significant role in the pathophysiology of metabolic syndrome. It is therefore to be expected that the BMI and WHR were increased in women with O + H and O + H + P compared with those with H + P, O + P, and controls. There are some authors 27 who suggest that the clinical phenotype may arise, at least in part, from differences in the degree of metabolic dysfunction. Of note, women with O + H and O + H + P had increased levels of nearly all parameters measured and there were no significant differences between these two groups. This suggests that the classic PCOS phenotype (O + H), the phenotype created by the NIH criteria and the complete phenotype (O + H + P) created by the Rotterdam criteria, may represent the forms of PCOS associated with greater metabolic profile compared with other phenotypes. The data of the metabolic complications in the ovulatory PCOS phenotype (H + P) are conflicting. One study reported significantly higher serum insulin levels, lower glucose/insulin ratios, and an atherogenic lipid profile in overweight women with hyperandrogenism and polycystic ovarian morphology (n = 22) compared with control subjects. 28 In another study, in the H + P group (defined by biochemical hyperandrogenism) the lipid levels and glucose and insulin concentrations were not significantly different from those in control subjects. 6 In the present study, we found that patients with H + P had an intermediate metabolic profile compared with the O + H phenotype and control subjects. This may partly result from the conflicting viewpoints regarding the association of hyperandrogenemia with increased risk for metabolic syndrome in women with PCOS. 29,30 The O + P group has been the most controversial of the four groups identified in the Rotterdam workshop 7,8,31 because hyperandrogenemia has been considered historically by many as a defining feature of PCOS. 32 The current data suggests that the PCOS group defined by O + P is distinguished from control women with irregular menstruation, greater ovarian volume and follicle number on ultrasound, which has expanded the previous findings. 33 The LH/FSH ratio, 34 was also high in the O + P group and was similar with that in the O + H group. Additionally, with the lowest prevalence of metabolic syndrome, the O + P phenotype had similar mean TG and LDL levels compared with the control group, and the mildest degrees of hyperandrogenemia and hyperinsulinemia This is similar to the findings of an Australian study, which showed that women with PCO and cycle irregularity had a similar metabolic profile as control subjects. 6 These results suggest that the nonhyperandrogenic PCOS phenotype (O + P), one of the new phenotypes created by the Rotterdam criteria, may represent a form of PCOS associated with a milder metabolic profile compared with other phenotypes. Larger studies are needed to determine the precise metabolic risks in this specific phenotype. Additionally, in our study, the mean F-G score of the PCOS group was 5.56, and 3.38 of the control group. The prevalence of PCOS patients whose F-G score is greater than the mean value of controls was 97.2%, significantly higher than the proportion that met the criterion of Western definition of 6. Thus, it may be necessary to consider ethnic difference in the criteria of hirsutism. Future studies are required to confirm these data by recruiting more Chinese patients to make a compatible criteria of hirsutism for Chinese women. The present study has some limitations. For example, the majority of patients presented in our clinic are of oriental skin, and the distribution of the PCOS phenotypes in other ethnic groups, especially those with higher metabolic complications (for example African Americans and Hispanic Americans) needs to be evaluated. We conclude that the most frequent PCOS phenotype is the group with O + P (nonhyperandrogenic phenotype). Nonclassic phenotypes (O + P, H + P and O + H + P) were more frequent than the classic phenotype (O + H). The nonhyperandrogenic PCOS phenotype (O + P), one of the new phenotypes created by the Rotterdam criteria, may represent a form of PCOS associated with a milder metabolic profile compared with the other phenotypes. Women ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology 1637

Zhang et al. with O + H (classic PCOS) may have the highest risk of metabolic syndrome. Subjects with PCOS defined by hyperandrogenemia are the most severely affected, with the highest androgen levels. Women with O + P are clearly distinguished from controls by their ovarian volume, follicle number, and irregular menstruation. Disclosure of interests We declare that we have no conflicts of interests. Contribution to authorship Zhang Hongyuan initiated the study, analysed the data and drafted the manuscript. All other authors were involved in data provision and in the critical revision and finalisation of the manuscript. Details of ethics approval All registries have ethical approval appropriate to their national and local ethics guidelines. Funding No funding was sought or obtained to undertake this specific study. Individual authors were supported through their institutional or grant funded employment to participate. No funding source had any role in the conduct, analysis or interpretation of the study. Acknowledgements The authors are grateful to the staff of the Division of Obstetrics and Gynaecology at the Second Xiangya Hospital of Central South University for their assistance and collaboration in data collection. j References 1 Stein I, Leventhal M. Amenorrhea associated with bilateral polycystic ovaries. Am J Obstet Gynecol 1935;29:181 5. 2 Azziz R, Woods KS, Reyna R, Key TJ, Knochenhauer ES, Yildiz BO. The prevalence and features of the polycystic ovary syndrome in an unselected population. J Clin Endocrinol Metab 2004;89:2745 9. 3 Knochenhauer ES, Key TJ, Kahsar-Miller M, Waggoner W, Boots LR, Azziz R. Prevalence of the polycystic ovary syndrome in unselected black and white women of the southeastern United States: a prospective study. J Clin Endocrinol Metab 1998;83:3078 82. 4 The Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 2004;19:41 7. 5 Zawadzki JK, Dunaif A. Diagnostic criteria for polycystic ovary syndrome: towards a rational approach. In: Dunaif A, Givens JR, Haseltine FP, Merriam GR, editors. Polycystic Ovary Syndrome. Boston: Blackwell Scientific; 1992. pp. 377 84. 6 Norman RJ, Hague WM, Masters SC, Wang XJ. Subjects with polycystic ovaries without hyperandrogenemia exhibit similar disturbances in insulin and lipid profiles as those with polycystic ovary syndrome. Hum Reprod 1995;10:2258 61. 7 Azziz R. Controversy in clinical endocrinology: diagnosis of polycystic ovarian syndrome: the Rotterdam criteria are premature. J Clin Endocrinol Metab 2006;91:781 5. 8 Franks S. Controversy in clinical endocrinology: diagnosis of polycystic ovarian syndrome: in defense of the Rotterdam criteria. J Clin Endocrinol Metab 2006;91:786 9. 9 Chen X, Yang D, Li L, Feng S, Wang L. Abnormal glucose tolerance in Chinese women with polycystic ovary syndrome. Hum Reprod 2006;21:2027 32. 10 The Asia-Pacific Perspective. Redefining Obesity and Its Treatment. Melbourne: International Diabetes Institute, 2000. 11 Kissebah A, Vydelingum N, Murray R, Evans DJ, Hartz AJ, Kalkhoff RK, et al. Relation of body fat distribution to metabolic complications of obesity. J Clin Endocrinol Metab 1982;54:254 60. 12 Balen AH, Laven JSE, Tan SL, Dewailly D. Ultrasound assessment of the polycystic ovary: international consensus definitions. Hum Reprod Update 2003;9:505 14. 13 Alberti KG, Zimmet P, Shaw J. Metabolic syndrome-a new worldwide definition: a consensus statement from the International Diabetes Federation. Diabet Med 2006;23:469 80. 14 Hatch R, Rosenfield RL, Kim MH, Tredway D. Hirsutism: implications, etiology, and management. Am J Obstet Gynecol 1981;140:815 30. 15 Robert PK, Teresa EB, Vicki MB, Pamela D, Daniel C. Endocrine and metabolic differences among phenotypic expressions of polycystic ovary syndrome according to the 2003 Rotterdam consensus criteria. Am J Obstet Gynecol 2008;198:670.e1 e10. 16 Broekmans FJ, knauff EA, Valkenburg O, Laven JS, Eijikemans MJ, Fauser BC. PCOS according to the Rotterdam consensus criteria: change in prevalence among WHO-h anovulation and association with metabolic factors. BJOG 2006;113:1210 7. 17 Urbanek M. The genetics of the polycystic ovary syndrome. Nat Clin Pract Endocrinol Metab 2007;3:103 11. 18 Welt CK, Gudmundsson JA, Arason G, Adams J, Palsdottir H, Gudlaugsdottir G, et al. Characterizing discrete subsets of polycystic ovary syndrome as defined by the Rotterdam criteria: the impact of weight on phenotype and metabolic features. J Clin Ecdocrinol Metab 2006;91:4842 8. 19 Dewailly D, Catteau-Jonard S, Reyss AC, Leroy M, Pigny P. Oligoanvulation with polycystic ovaries but not overt hyperandrogenism. J Clin Endocrinol Metab 2006;91:3922 7. 20 Barber TM, Wass JA, McCarthy MI, Franks S. Metabolic characteristics of women with polycystic ovaries and oligo-amenorrhoea but normal androgen levels: implications for the management of polycystic ovary syndrome. Clin Endocrinol (Oxf) 2007;66:513 7. 21 Adams JM, Taylor AE, Crowley WF Jr, Hall JE. Polycystic ovarian morphology with regular ovulatory cycles: insights into the pathophysiology of polycystic ovarian syndrome. J Clin Endocrimol Metab 2004;89:4343 50. 22 Polson DW, Adams J, Wadsworth J, Franks S. Polycystic ovaries-a common finding in normal women. Lancet 1988;1:870 2. 23 Ford ES. The metabolic syndrome and mortality from cardiovascular disease and all-causes: findings from the National Health and Nutrition Examination Survey II Mortality Study. Atherosclerosis 2004;173:307 12. 24 Haffner SM, Valdez RA, Hazuda HP, Mitchell BD, Morales PA, Stern MP. Prospective analysis of the insulin-resistance syndrome (syndrome X). Diabetes Care 1992;41:715 22. 25 Isomaa B, Almgren P, Tuomi T, Forsen B, Lahti K, Nissen M, et al. Cardiovascular morbidity and mortality associated with the metabolic syndrome. Diabetes Care 2001;24:683 9. 26 Ehrmann D, Barnes R, Rosenfield R, Cavaghan M, Imperial J. Prevalence and predictors of impaired glucose tolerance and diabetes in 1638 ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology

Characteristics of different polycystic ovary syndrome phenotypes women with polycystic ovary syndrome. Diabetes Care 1999; 22:141 6. 27 Chang WY, Knochenhauer ES, Bartolucci AA, Azziz R. Phenotypic spectrum of polycystic ovary syndrome: clinical and biochemical characterization of the three major clinical subgroups. Fertil Steril 2005;83:1343 6. 28 Carmina E, Lobo RA. Polycystic ovaries in hirsute women with normal menses. Am J Med 2001;111:602 6. 29 Dokras A, Bochner M, Hollinrake E, Markham S, Vanvoorhis B, Jagasia DH. Screening women with polycystic ovary syndrome for metabolic syndrome. Obstet Gynecol 2005;106:131 7. 30 Apridonidze T, Essah PA, Iuorno MJ, Nestler JE. Prevalence and characteristics of the metabolic syndrome in women with polycystic ovary syndrome. J Clin Endocrinol Metab 2005;90:1929 35. 31 Azziz R. Diagnostic criteria for polycystic ovary syndrome: a reappraisal. Fertil Steril 2005;83:1343 6. 32 Zawadzki JK, Dunaif A. Diagnostic criteria for polycystic ovary syndrome: towards a rational approach. In: Dunaif A, Givens JR, Haseltine FP, Merriam GR, et al., editors. Polycystic Ovary Syndrome. Boston: Blackwell Scientific; 1992. pp. 131 46. 33 Carmina E, Chu MC, Longo RA, Rini GB, Lobo RA. Phenotypic variation in hyperandrogenic women influences the findings of abnormal metabolic and cardiovascular risk parameters. J Clin Endocrinol Metab 2005;90:2545 9. 34 Fox R, Corrigan E, Thomas PA, Hull MG. The diagnosis of polycystic ovaries in women with oligo-amenorrhoea: predictive power of endocrine tests. Clin endocrinol (Oxf) 1991;34:127 31. ª 2009 The Authors Journal compilation ª RCOG 2009 BJOG An International Journal of Obstetrics and Gynaecology 1639