Clinical and endocrine characteristics of the main polycystic ovary syndrome phenotypes

POLYCYSTIC OVARY SYNDROME Clinical and endocrine characteristics of the main polycystic ovary syndrome phenotypes Ettore Guastella, M.D., a Rosa Alba Longo, M.D., b and Enrico Carmina, M.D. b a Department of Obstetrics and Gynecology, and b Department of Biomedical Sciences, University of Palermo, Palermo, Italy Objective: To evaluate the clinical and endocrine differences between main polycystic ovary syndrome (PCOS) phenotypes. Design: To evaluate clinical and hormone parameters in a large group of consecutive women with PCOS diagnosed according Rotterdam criteria and divided according their phenotype. Setting: University department of medicine. Patient(s): Three hundred eighty-two consecutive women with PCOS and 85 ovulatory controls. Intervention(s): Evaluation of clinical and hormone parameters. Main Outcome Measure(s): Blood levels of gonadotropins, testosterone, sex-hormone binding globulin, dehydroepiandrosterone sulfate, 17a-hydroxyprogesterone, progesterone, glucose, and insulin, and calculation of the free androgen index and insulin sensitivity. Result(s): The severe PCOS phenotype (hyperandrogenism, chronic anovulation, and polycystic ovaries: type I classic PCOS) was the most common phenotype in 53.9% of the patients. The phenotype of 8.9% of patients was characterized by hyperandrogenism and chronic anovulation but normal ovaries (type II classic PCOS). The two phenotypes of classic PCOS had similar clinical and endocrine characteristics, but the patients with polycystic ovaries had a higher luteinizing hormone/follicle-stimulating hormone (LH/FSH) ratio. Ovulatory PCOS was relatively common (28.8% of PCOS patients) and presented milder clinical and endocrine alterations than the classic PCOS phenotypes. The normoandrogenic phenotype was relatively uncommon. These patients had a normal body mass index, insulin sensitivity, and free androgen index but showed increased levels of LH and LH/FSH ratio. Conclusion(s): Ovulatory PCOS represents the mild form of classic PCOS, but the normoandrogenic phenotype, although part of the spectrum, may represent a different disorder or have a different pathogenetic pathway. (Fertil Steril Ò 2010;94:2197 201. Ó2010 by American Society for Reproductive Medicine.) Key Words: PCOS, hyperandrogenism, anovulation Polycystic ovary syndrome is a very heterogeneous disorder that may include different phenotypes. For many years, only the classic phenotype of chronic anovulation and hyperandrogenism was included in the PCOS diagnosis (1). The Rotterdam criteria (2) are more extensive, including patients with very different clinical appearances. In fact, according to these guidelines, the diagnosis of PCOS may present in patients with four different phenotypes: [1] hyperandrogenism, chronic anovulation, and polycystic ovaries; [2] hyperandrogenism and chronic anovulation but normal ovaries; [3] hyperandrogenism and polycystic ovaries but ovulatory cycles; and [4] chronic anovulation and polycystic ovaries but no clinical or biochemical hyperandrogenism. An expert committee of the Androgen Excess and PCOS Society (AEPCOS) suggested including only the first three phenotypes, excluding the fourth from the PCOS spectrum (3). However, few studies have addressed the clinical and endocrine difference between Received November 10, 2009; revised February 2, 2010; accepted February 4, 2010; published online March 19, 2010. G.E. has nothing to disclose. L.R.A. has nothing to disclose. C.E. has nothing to disclose. Reprint requests: Enrico Carmina, M.D., Department of Biomedical Sciences, University of Palermo, Via delle Croci 47, 90139 Palermo, Italy (FAX: 390916555995; E-mail: enricocarmina@libero.it). these phenotypes (4 6), and disagreement on the limits of the disorder persists. We previously analyzed the differences between classic (chronic anovulation and hyperandrogenism) and ovulatory PCOS (hyperandrogenism and polycystic ovaries but ovulatory cycles) (7) and found that patients with classic PCOS have a more severe form of the disorder but both phenotypes share the same alterations. In this study, we report on the relative prevalence in a medical setting and the clinical and endocrine characters of all four Rotterdam PCOS phenotypes. We studied 382 consecutive women in whom PCOS had been diagnosed according to the Rotterdam criteria (2). We classified them according to the four phenotypes, and analyzed the relative prevalence and the clinical and endocrine characteristics of each phenotype. A true prevalence study would survey a community, and the setting of our study (an endocrinology clinic) may have influenced the reported prevalence of different phenotypes. However, our department represents a regional reference center for women with all types of menstrual irregularities and/or clinical signs of androgen excess. MATERIALS AND METHODS Between 2004 and 2009, among the women referred to the endocrine unit of the Department of Clinical Medicine, University of Palermo, for clinical 0015-0282/$36.00 Fertility and Sterility â Vol. 94, No. 6, November 2010 2197 doi:10.1016/j.fertnstert.2010.02.014 Copyright ª2010 American Society for Reproductive Medicine, Published by Elsevier Inc.

TABLE 1 Clinical data of 85 normal ovulatory controls and 382 PCOS patients divided according their phenotype. n Age (y) BMI Waist (cm) Ferriman-Gallwey scores Controls 85 25 4 22 2 77 3 2 1 Type I classic PCOS (A, HA, PCO) 206 24 5 28 6 a,b,c 92.3 13 a,b,c 11.4 4 a,b,c Type II classic PCOS (A, HA) 34 24 5 31 6 a,b,c 97.5 17 a,b,c 11.5 5 a,b,c Ovulatory PCOS (HA, PCO) 110 25 5 24.5 4 a,c 85 9 a,c 9.4 3.6 a,c Normoandrogenic PCOS (A, PCO) 32 25 5 22 3 80 4 3 1 Note: A ¼ anovulation (chronic); BMI ¼ body mass index; HA ¼ hyperandrogenism (clinical or biochemical); PCO ¼ polycystic ovaries; PCOS ¼ polycystic ovary syndrome. a P<.01 compared with controls. b P<.01 compared with ovulatory PCOS. c P<.01 compared with normoandrogenic PCOS. hyperandrogenism and/or menstrual irregularities, 382 consecutive women were found to have PCOS, diagnosed according to Rotterdam criteria (2). None of the patients enrolled in our study had taken any medications for at least 3 previous months. Our procedures were in agreement with the Helsinki Declaration of 1975 as revised in 1983, and the study was approved by the local ethics council. All participants gave their informed consent to participate to the study. Clinical hyperandrogenism was defined as presence of hirsutism, acne, or androgenic alopecia. Hirsutism was assessed by Ferriman-Gallwey- Lorenzo scores (8), and patients with scores R6 were considered hirsute; acne was graded by a scoring system from 0 to 3 (9), and alopecia was evaluated by the Ludwig scoring system (10). Menstrual cycles shorter than 25 days and longer than 34 days were considered abnormal. The clinical study included assessment of clinical hyperandrogenism, body mass index (BMI), age, waist circumference, and characteristics of the menstrual cycle. In all patients and in 85 normal ovulatory controls, the blood levels of luteinizing hormone (LH), follicle-stimulating hormone (FSH), testosterone (T), dehydroepiandrosterone sulfate (DHEAS), estradiol, sex-hormone binding globulin (SHBG), 17a-hydroxyprogesterone (17-OHP), progesterone, glucose, and insulin were measured. The free androgen index (FAI) was calculated (11), and insulin sensitivity was evaluated by the quantitative insulin-sensitivity check index (QUICKI) method (12). Concentrations of LH, FSH, estradiol, testosterone, SHBG, DHEAS, and 17-OHP were determined during the follicular phase (days 5 to 8) of a spontaneous or progestininduced cycle. Serum progesterone was determined on days 21 to 24 of a spontaneous or induced menstrual cycle. Serum insulin and blood glucose were measured in the fasting state, in the morning. Serum hormone levels were quantified by well-established methods validated previously in our laboratory. All steroids were measured by specific radioimmunoassay (RIA) after extraction by use of methods previously described elsewhere (13). In all assays, the intra-assay and interassay coefficients of variation did not exceed 6% and 15%, respectively. In some patients, clinical suspicion also prompted measurement of urinary free cortisol, serum prolactin, and thyroid-stimulating hormone by commercial RIA methods. Anovulation was defined as serum progesterone <3 ng/ml (<9.54 nmol/ L). In patients with normal menses, at least two consecutive menstrual cycles were studied and a finding of low levels of serum progesterone (<3 ng/ml) in both cycles indicated the presence of chronic anovulation. Biochemical hyperandrogenism was defined as serum testosterone >60 ng/dl (>2.08 nmol/ L) and/or serum DHEAS R3 mg/ml (R7.8 mmol/l) and/or FAI R4. These values of hyperandrogenism had been previously calculated in our population by use of the same assays (14). Increased serum 17-OHP was defined as serum 17-OHP >3 mg/l (>9.1 nmol/l). In patients with mildly increased serum 17-OHP (<10 mg/l to >3 mg/l), increased 17-OHP response to adrenocorticotropic hormone (ACTH) administration (1 mg IV with blood samples at 0, 30, and 60 minutes) was required for diagnosis of nonclassic 21-hydroxylase deficiency (15). In all patients and controls, ovarian morphologic features were assessed by pelvic ultrasound. Ovarian size (by measurement of the main three ovarian diameters in both ovaries), and the presence, size, and number of ovarian microcysts were determined. The data obtained by abdominal and transvaginal ultrasounds were considered useful for the diagnosis. Polycystic ovaries were established by the presence of 10 or more peripherally oriented cystic structures in one ultrasonographic plane, each of which measured 2 to 10 mm in diameter, arranged around a dense stroma (16, 17). Ovarian volume was calculated by the formula p/6 (D 1 D 2 D 3 ), where the dimensions (D) of length, width, and thickness were used. The size of both ovaries was assessed, and the mean ovarian size was calculated. According to our data on normal women, increased ovarian size indicates that the mean ovarian size is larger than 7.5 cc 3 (17). The women with PCOS diagnosed according to Rotterdam criteria (2) were divided in four different phenotypes: [1] hyperandrogenism, chronic anovulation, and polycystic ovaries; [2] hyperandrogenism and chronic anovulation but normal ovaries; [3] hyperandrogenism and polycystic ovaries but ovulatory cycles; or [4] chronic anovulation and polycystic ovaries but no clinical or biochemical hyperandrogenism. The following arbitrary names were assigned to these phenotypes, respectively: [1] type I classic PCOS; [2] type II classic PCOS; [3] ovulatory PCOS; and [4] normoandrogenic PCOS. Statistical Analysis Group means were compared using analysis of variance (ANOVA) with post hoc least squares means pair wise comparisons (after log transformation of the values). All data are presented as mean standard deviation (SD). RESULTS Relative Prevalence of Different PCOS Phenotypes Of the study participants, 206 patients (53.9%) presented with all characteristics of the Rotterdam guidelines: hyperandrogenism, chronic anovulation, and polycystic ovaries (type I classic PCOS). Thirty-four patients (8.9%) exhibited hyperandrogenism and chronic anovulation but normal ovaries (type II classic PCOS). In total, 240 patients (62.8%) had PCOS according to U.S. National Institutes of Health criteria (1). We found that 110 patients (28.8%) presented with hyperandrogenism and polycystic ovaries but normal ovulatory cycles (ovulatory PCOS). In total, 350 patients (91.6%) had PCOS according AEPCOS guidelines (3). Finally, 32 patients (8.4%) presented chronic anovulation and polycystic ovaries but showed no clinical or biologic signs of hyperandrogenism (normoandrogenic PCOS). 2198 Guastella et al. PCOS phenotypes Vol. 94, No. 6, November 2010

TABLE 2 Concentrations of LH, LH/FSH ratio, and androgens in 85 normal ovulatory controls and 382 PCOS patients divided according their phenotype. LH (miu/ml) LH/FSH ratio T (ng/dl) SHBG (nmol/l) FAI DHEAS (mg/ml) Controls 6.4 3 1 0.2 36 16 62 16 2 0.8 1.6 0.6 Type I classic PCOS (A, HA, PCO) 10.0 5 a,b,c 1.8 0.9 a,b,c 82 29 a,b,d 29 16 a,b,d 9.7 4.6 a,b,d 2.7 1.2 a,d Type II classic PCOS (A, HA) 7.3 3 1.5 0.8 a,b 82 32 a,b,d 30 20 a,b,d 9.9 4.5 a,b,d 2.6 1.1 a,d Ovulatory PCOS (HA, PCO) 6.8 3.1 1.1 0.5 66 23 a,d 41 25 a,b 7 4.2 a,d 2.7 1.2 a,d Normoandrogenic PCOS (A, PCO) 10.7 5 a,b,c 1.7 0.5 a,b 46 17 a 65 35 2.4 1.8 1.5 0.6 Note: A ¼ anovulation (chronic); HA ¼ hyperandrogenism (clinical or biochemical); DHEAS ¼ dehydroepiandrosterone sulfate; FAI ¼ free androgen index; FSH ¼ follicle-stimulating hormone; LH ¼ luteinizing hormone; PCO ¼ polycystic ovaries; PCOS ¼ polycystic ovary syndrome; SHBG ¼ sex-hormone binding globulin; T, testosterone. a P<.01 versus controls. b P<.01 versus ovulatory PCOS. c P<.01 versus type II classic PCOS. d P<.01 versus normoandrogenic PCOS. Clinical and Endocrine Characters of Different PCOS Phenotypes No difference in age was observed between the different PCOS phenotypes (Table 1). Patients with both type I and type II classic PCOS had statistically significantly (P<.01) higher values for BMI, waist circumference, Ferriman-Gallwey-Lorenzo scores, levels of testosterone, FAI, and insulin, and lower QUICKI values than the controls and the patients with ovulatory or normoandrogenic PCOS (Tables 1, 2, and 3). Both phenotypes had also statistically significantly (P<.01) higher DHEAS values than the controls and women with normoandrogenic PCOS (see Table 2). The two phenotypes of classic PCOS differed only in LH and LH/FSH ratios, which were statistically significantly higher in type I than in type II classic PCOS (see Table 2). Type I classic PCOS had statistically significantly higher values for LH levels and LH/FSH ratio than the controls and patients with ovulatory PCOS (see Table 2). Patients with type II classic PCOS had LH values that were similar to the controls and patients with ovulatory PCOS but a higher LH/FSH ratio than those two groups (see Table 2). Patients with ovulatory PCOS had statistically significantly higher values for BMI, waist circumference, Ferriman-Gallwey- Lorenzo score, levels of testosterone, FAI, DHEAS, and insulin, and lower QUICKI values than controls and patients with normoandrogenic PCOS (see Tables 1, 2, and 3). Patients with ovulatory PCOS had LH and LH/FSH ratios similar to normal controls but lower than patients with classic or normoandrogenic PCOS (see Table 2). In spite of the differences in LH concentration, all of the PCOS phenotypes had similar levels of estradiol. Finally, patients with normoandrogenic PCOS had statistically significantly higher values for LH and the LH/FSH ratio than the controls (see Table 2). Also, although their testosterone values were in the normal range, their mean testosterone was statistically significantly higher compared with controls (see Table 2). However, there was no statistically significant difference between patients with normoandrogenic PCOS and controls in terms of concentrations of SHBG, FAI, or DHEAS (see Table 2). Normoandrogenic PCOS patients and controls had also similar BMI, waist circumference, Ferriman-Gallwey-Lorenzo scores, levels of DHEAS and insulin, and QUICKI (see Tables 1, 2, and 3). DISCUSSION After evaluation of a sufficiently large number of consecutive women with a diagnosis of PCOS according to the Rotterdam criteria (2), our study provides information on the relative prevalence in TABLE 3 Insulin, insulin sensitivity, and ovarian volume in 85 normal ovulatory controls and 382 PCOS women divided according their phenotype. Insulin (miu/ml) QUICKI Ovarian volume (cm 3 ) Controls 9.8 3 0.354 0.020 4.7 1.2 Type I classic PCOS (A, HA, PCO) 15.6 6.6 a,b,c 0.324 0.019 a,b,c 12.1 3 a,d Type II classic PCOS (A, HA) 16.5 6.7 a,b,c 0.320 0.024 a,b,c 5 1.5 Ovulatory PCOS (HA, PCO) 12.5 6.3 a,c 0.336 0.020 a,c 11 4 a,d Normoandrogenic PCOS (A, PCO) 10.1 3.9 0.348 0.025 12.4 3 a,d Note: A ¼ anovulation (chronic); HA ¼ hyperandrogenism (clinical or biochemical); PCO ¼ polycystic ovaries; PCOS ¼ polycystic ovary syndrome; QUICKI ¼ quantitative insulin-sensitivity check index. a P<.01 versus controls. b P<.01 versus ovulatory PCOS. c P<.01 versus normoandrogenic PCOS. d P<.01 versus type II classic PCOS. Fertility and Sterility â 2199

a medical setting and the clinical and endocrine differences among the main PCOS phenotypes, as determined according Rotterdam criteria (2). These diagnostic guidelines (2) have included in the PCOS group four different phenotypes. The names we have given to these phenotypes (type I and type II classic PCOS, ovulatory PCOS, normoandrogenic PCOS) may be arbitrary and not accepted by all experts, but they reflect the main characteristics of the different phenotypes. For a long time, the diagnosis of PCOS was based only on chronic anovulation and hyperandrogenism (1), and we have classified these patients as classic PCOS, differentiating between type I and type II depending on the finding of polycystic ovaries. The other phenotypes are characterized by the finding of normal ovulation in spite of hyperandrogenism and polycystic ovaries and by the absence of clinical or biochemical hyperandrogenism in spite of chronic anovulation and polycystic ovaries. Although the Rotterdam criteria have been accepted by many experts, almost no information was available regarding the differences between the included phenotypes at the time the criteria were created. In particular, there were no data on normoandrogenic PCOS or on patients with classic PCOS but normal ovaries (the phenotype that we have defined as type II classic PCOS). Thus, the Androgen Excess and PCOS Society has suggested not including patients with normal androgen levels among the PCOS phenotypes (3). More recently, Dewailly and al. (4) reported that patients with chronic anovulation and polycystic ovaries but normal androgens have higher values of testosterone than controls, which suggests that these patients have a mild form of androgen excess. Welt et al. (5), studying a large heterogeneous group that included patients from New England and Iceland, concluded that ovulatory and normoandrogenic PCOS both represent a mild form of PCOS. In fact, they observed that these patients presented similar but milder abnormalities compared with patients with PCOS as diagnosed according U.S. National Institutes of Health criteria (5). Our study has the advantage of evaluating a large, homogenous group of patients under the same environmental conditions. Although we used a slightly different, U.S. variation on the Rotterdam criteria (16), it should be noted that some of the authors of Rotterdam guidelines also have corrected their definition of increased ovarian volume and use the limit of 7.5 cc (the same that we use) (18). 1. Type I classic PCOS represents the most common and severe form of PCOS. These patients presented as group with abdominal obesity, increased levels of LH and LH/FSH ratio, increased androgens, and elevated insulin and insulin resistance. In our setting, these patients represented almost 60% of all PCOS patients. 2. Type II classic PCOS patients are very similar to other classic PCOS patients, but they are less common (only 8.4% of the total number of PCOS patients). The main difference between type I and type II classic PCOS is the LH levels and LH/FSH ratio. All patients with classic PCOS have increased LH and LH/ FSH ratios when compared with controls, but patients with polycystic ovaries have statistically significantly higher LH levels (and LH/FSH ratios) than patients with normal ovaries. This suggests that LH is important or in some way is related to the morphologic appearance of the ovaries. 3. Ovulatory PCOS seems to be the mild form of classic PCOS. These patients presented with the most characteristics of PCOS but in a milder form. In fact, patients with ovulatory PCOS had intermediate values (between classic PCOS and controls) of BMI, waist circumference, testosterone, insulin, and QUICKI. Only the LH level was generally normal, and it represents the main difference from the other phenotypes of PCOS. On the other hand, it is well known that many patients with classic PCOS can attain ovulation by reducing their body weight (19). Therefore, we suggest that ovulatory PCOS may be the mild form of classic PCOS. Passage from classic to ovulatory PCOS and vice versa may be possible and related to sociocultural and environmental influences (20, 21). 4. Normoandrogenic PCOS represents a unique group. We have confirmed the report by Dewailly et al. (4) that these patients have a mild testosterone excess (although within the normal range), but these patients do not present with hyperandrogenism in fact, both their FAI and DHEAS levels are normal. Nor do these patients present with abdominal obesity they have a normal BMI and normal waist circumference. From an endocrine point of view, this group of women is mainly characterized by increased LH and LH/FSH ratio, which is similar to classic PCOS and may be related in some way to their anovulation. In our opinion, normoandrogenic PCOS is part of the PCOS spectrum but may be a different disorder or have a different pathogenetic pathway. These women do not exhibit the two main components of PCOS: hyperandrogenism (at least, clinically relevant hyperandrogenism) and insulin resistance. Thus, we believe that it would be correct to keep these patients separated from other PCOS patients, as suggested by the Androgen Excess and PCOS Society committee (3). The different phenotypes of PCOS present similarities but also important differences in their clinical and endocrine pattern, and research studies should keep these differences in mind. We suggest that, although the Rotterdam criteria may be useful for understanding the complexity and the heterogeneity of the syndrome, it is important that patients are well characterized according to their phenotype. Patients with completely different phenotypes and clinical and endocrine patterns should not be indiscriminately grouped together. This is particularly important for patients with normoandrogenic PCOS, who may develop PCOS by a different pathogenetic pathway or have a different disorder. REFERENCES 1. Zawadzki JK, Dunaif A. Diagnostic criteria for polycystic ovary syndrome: towards a rationale approach. In: Dunaif A, Givens JR, Haseltine F, Merriam GR, eds. Polycystic ovary syndrome. Boston: Blackwell Scientific, 1992:377 84. 2. Rotterdam ESHRE/ASRM Sponsored PCOS Consensus Workshop Group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19 25. 3. Azziz R, Carmina E, Dewailly D, Diamanti- Kandarakis E, Escobar-Morreale HF, Futterweit W, et al. Position statement: criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: an androgen excess society guideline. J Clin Endocrinol Metab 2006;91:4237 45. 4. Dewailly D, Catteau-Jonard S, Reyss AC, Leroy M, Pigny P. Oligoanovulation with polycystic ovaries but not overt hyperandrogenism. J Clin Endocrinol Metab 2006;91:3922 7. 5. Welt CK, Gudmundsson JA, Arsson 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 Endocrinol Metab 2006;91:4842 8. 6. 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. 7. Carmina E. The spectrum of androgen excess disorders. Fertil Steril 2006;85:1582 5. 8. Hatch R, Rosenfield RL, Kim MH, Tredway D. Hirsutism: implications, etiology and management. Am J Obstet Gynecol 1981;140:815 30. 2200 Guastella et al. PCOS phenotypes Vol. 94, No. 6, November 2010

9. Lookingbill DP, Egan N, Santen RJ, Demers LM. Correlation of serum 3a-androstanediol glucuronide with acne and chest hair density in men. J Clin Endocrinol Metab 1988;67:986 91. 10. Ludwig E. Classification of the types of androgenic alopecia (common balding occurring in the female sex). Br J Dermatol 1977;97:247 54. 11. Morley JE, Patrick P, Perry HM III. Evaluation of assays available to measure free testosterone. Metabolism 2002;51:554 9. 12. Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, Quon MJ. Quantitative insulin sensitivity check index: a simple, accurate method of assessing insulin sensitivity in humans. J Clin Endocrinol Metab 2000;85:2402 10. 13. Carmina E, Stanczyk F, Chang L, Miles RA, Lobo RA. The ratio of androstenedione: 11b-hydroxyandrostenedione is an important marker of adrenal androgen excess in women. Fertil Steril 1992;58: 148 52. 14. Carmina E. Prevalence of idiopathic hirsutism. Eur J Endocrinol 1998;139:421 3. 15. Carmina E, Lobo RA. Ovarian suppression reduces clinical and endocrine expression of late onset congenital adrenal hyperplasia due to 21- hydroxylase deficiency. Fertil Steril 1994;62: 738 43. 16. Balen AH, Laven JS, Tan SL, Dewailly D. Ultrasound assessment of the polycystic ovary: international consensus definitions. Hum Reprod Update 2003;9: 505 14. 17. Carmina E, Orio F, Palomba S, Longo RA, Lombardi G, Lobo RA. Ovarian size and blood flow in women with polycystic ovary syndrome (PCOS) and their correlations with some endocrine parameters. Fertil Steril 2005;84:413 9. 18. Jonard S, Robert Y, Dewailly D. Revisiting the ovarian volume as a diagnostic criterion for polycystic ovaries. Hum Reprod 2005;20:2893 8. 19. Huber-Buckholz MM, Carey DG, Norman RJ. Restoration of reproductive potential by lifestyle modification in obese polycystic ovary syndrome: role of insulin sensitivity and luteinizing hormone. J Clin Endocrinol Metab 1999;84: 1470 4. 20. Carmina E, Legro RS, Stamets K, Lowell J, Lobo RA. Difference in body weight between American and Italian women with polycystic ovary syndrome: influence of the diet. Hum Reprod 2003;11: 2289 93. 21. Di Fede G, Mansueto P, Longo RA, Rini G, Carmina E. Influence of sociocultural factors on the ovulatory status of polycystic ovary syndrome. Fertil Steril 2009;91:1853 6. Fertility and Sterility â 2201