Comparison of diabetes mellitus and insulin resistance screening methods for women with polycystic ovary syndrome

Transcription

1 Comparison of diabetes mellitus and insulin resistance screening methods for women with polycystic ovary syndrome William W. Hurd, M.D., M.Sc., M.P.H., a Mohamed Y. Abdel-Rahman, M.D., a,d Salah A. Ismail, M.D., d Mostafa A. Abdellah, M.D., d Christine L. Schmotzer, M.D., c and Ajay Sood, M.D. b a Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology, b Department of Internal Medicine, Division of Endocrinology, and c Department of Pathology, University Hospitals Case Medical Center, Case Western Reserve University, Cleveland, Ohio; and d Department of Obstetrics and Gynecology, Sohag University Hospitals, Sohag, Egypt Objective: To compare screening strategies for type 2 diabetes mellitus (DM), impaired glucose tolerance (pre-dm), and insulin resistance (IR) in women with polycystic ovary syndrome (PCOS). Design: Prospective study. Setting: Academic reproductive endocrinology practice. Patient(s): Adult women with PCOS (n ¼ 111). Intervention(s): None. Main Outcome Measure(s): Subjects were screened for pre-dm and DM using a 2-hour glucose tolerance test (GTT), hemoglobin A1c (HbA1c), or fasting plasma glucose (FPG) and for IR using homeostasis model assessment (HOMA), insulin levels (fasting and 2 hours after 75 glucose load), or obesity. Screening approaches were compared using positive and negative percent agreement and Cohen s kappa (k). Result(s): DM and pre-dm were diagnosed by GTT in 4% and 20% of subjects, respectively. Screening with FPG failed to identify 41% of pre-dm and 20% of DM subjects. GTT and HbA1c had only fair agreement (k ¼ 0.29). IR was diagnosed in 24% of subjects with pre-dm or DM and in 56% of the remaining subjects using HOMA and insulin levels. HOMA and elevated insulin levels demonstrated substantial agreement for detecting IR (k ¼ ). Obesity demonstrated fair to slight agreement (k ¼ ). Conclusion(s): Women with PCOS should be screened for Pre-DM and DM using GTT or HbA1c, and those with Pre-DM or DM are presumed to have IR. In the rest, IR can be detected using either HOMA or insulin levels. (Fertil Steril Ò 2011;96: Ó2011 by American Society for Reproductive Medicine.) Key Words: Polycystic ovarian syndrome, type 2 diabetes mellitus, insulin resistance, oral glucose tolerance test, hemoglobin A1c, blood glucose Polycystic ovary syndrome (PCOS) is a well-defined constellation of symptoms and physical findings that has significant implications for a woman s health, appearance, and fertility. PCOS has become the most common endocrinopathy of reproductive-age women in the world. The prevalence of PCOS in the United States has been estimated to be as high as 7% and appears to be rising (1). PCOS is believed to have several possible underlying causes. The most common of these is insulin resistance (IR), which is estimated to be present in at least 50% of women with PCOS (2). As many as 90% of obese women with PCOS will be found to have IR, although lean women with PCOS can also have IR (3, 4). There is good evidence that IR causes or exacerbates the most clinically significant manifestations of PCOS, most notably ovulation dysfunction and hirsutism (5, 6). Both of these conditions Received February 15, 2011; revised June 30, 2011; accepted July 5, 2011; published online August 3, W.W.H. has nothing to disclose. M.Y.A.-R. has nothing to disclose. S.A.I. has nothing to disclose. M.A.A. has nothing to disclose. C.L.S. has nothing to disclose. A.S. has nothing to disclose. Presented in part at the 8th Annual Meeting of the Androgen Excess and Polycystic Ovary Syndrome Society, which was held in Munich, Germany, on September 11 12, Reprint requests: William W. Hurd, M.D., M.Sc., M.P.H., Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology, University Hospitals Case Medical Center, Euclid Avenue MAC 7007, Cleveland, Ohio ( William.hurd@uhhospitals. org). have been shown to respond to treatment with insulin sensitizers in women with IR (7). The presence of IR also puts women with PCOS at increased risk for type 2 diabetes mellitus (DM). The prevalence of DM in women with PCOS is approximately 10% and increases by 2% annually (8). Because women with PCOS are at increased risk for DM, screening for glucose intolerance in all patients newly diagnosed with PCOS is recommended (9 11). The most commonly recommended initial screening test is the 2-hour glucose tolerance test (GTT) (9 11). Although the American Diabetes Association (ADA) has long recommended fasting plasma glucose (FPG) for periodic DM screening in the general population, this test is not recommended as a screening test for women with PCOS since it is relatively insensitive in these women (12). Recently, the ADA has recommended using hemoglobin A1c (HbA1c) for DM screening in the general population (12). However, the accuracy of this test in women with PCOS remains uncertain. In addition, the most appropriate interval for periodic rescreening of women with PCOS remains controversial; some organizations do not recommend periodic rescreening, while others recommend rescreening every 1 3 years depending on coexisting risk factors (9 13). Routine evaluation of women with PCOS for IR is even more controversial. Despite the fact that IR increases the risk of developing of DM, some organizations recommend against routine IR screening for women with PCOS, while others suggest that IR screening is optional and of potential benefit only in selected women /$36.00 Fertility and Sterility â Vol. 96, No. 4, October doi: /j.fertnstert Copyright ª2011 American Society for Reproductive Medicine, Published by Elsevier Inc.

2 (10, 11, 14). The primary rationales for not screening women with PCOS for IR are [1] at least half of women with PCOS will have some degree of IR; [2] clinically available tests for IR are relatively inaccurate, particularly in women with impaired glucose tolerance; and [3] IR assessment has not been shown to identify women who will respond to therapy or to predict clinical outcome (10, 11, 13). Despite these limitations, a number of investigators recommend using one of a number of strategies to identify which women with PCOS have IR to guide diagnostic and therapeutic management (7, 11, 15, 16). Several methods for detecting IR are currently being used. For clinical studies, one of the most common methods is using fasting insulin and glucose levels to calculate the homeostasis model assessment (HOMA) (15). Since this method does not evaluate insulin during glucose load, an alternative method to identify IR is to determine whether insulin levels are above the normal reference range in the fasting state or 2 hours after a 75-g glucose load (11, 14). A relatively simple method is to assume that all obese women with PCOS have IR (7, 11, 16). The relative accuracy of these three approaches has not been compared, and only the accuracy of HOMA has been verified using the hyperinsulinemic-euglycemic clamp test (17). We thus designed this study to compare various strategies for diagnosing pre-dm and DM and IR in a cohort of women with PCOS. To diagnose pre-dm and DM, we used GTT, FPG, and HbA1c. All women with pre-dm or DM were assumed to have IR (9, 18). To diagnose IR in the remaining women, we compared the results of HOMA and fasting and 2-hour insulin levels, or we used obesity as an indicator of IR. MATERIALS AND METHODS We prospectively studied 125 women presenting with apparent PCOS at University Hospitals Case Medical Center reproductive endocrinology outpatient unit at Case Western Reserve University between June 2009 and October The study was approved by the Institutional Review Board of University Hospitals, and all patients gave informed consents before participation. The study cohort included women between the age of 18 and 45 years referred to our clinic for oligo- or amenorrhea and/or hyperandrogenism who were selected at random and agreed to take part in this study. The clinical diagnosis of PCOS was made using the Rotterdam criteria when women had at least two of three criteria: ovulatory dysfunction (oligoovulation or anovulation), excess androgen activity (hirsutism, acne, or elevated serum androgens), and/or polycystic ovaries by ultrasound (13). Oligomenorrhea was defined as fewer than eight menses per year or cycles longer than 35 days in length. Hirsutism was defined as a Ferriman- Gallwey score R8 (19). Polycystic ovarian morphology was diagnosed using transvaginal ultrasound when at least one ovary had R12 antral follicles with a mean diameter <9 mm and/or a total ovarian volume >10 cm 3. Women with medical conditions that mimic PCOS were excluded from this study using appropriate tests, including those with hypothyroidism, hyperprolactinemia, Cushing syndrome, nonclassic congenital adrenal hyperplasia, and androgen-secreting tumors of the ovaries or adrenal glands (6). Women with PCOS were also excluded from the study if they had taken any medication within 3 months before the study that could affect glucose or sex hormone metabolism, such as metformin or hormonal contraceptives. Diagnosis of Pre-DM and DM Subjects were screened for pre-dm and DM using both GTT and HbA1c according to ADA guidelines (12). Pre-DM was diagnosed when FPG >99 mg/dl, 2-hour glucose >139 mg/dl, or HbA1c > 5.6% (i.e., increased risk for diabetes ) (12). DM was diagnosed when FPG R126 mg/dl, 2-hour glucose >200 mg/dl, or HbA1c R6.5%. To examine the accuracy of using FPG as an initial screen, the rates of pre-dm and DM were recalculated after considering GTT results only in subjects with FPG >99 mg/dl (9, 20). Diagnosis of IR For the purposes of this study, we used as a reference group women with apparent IR diagnosed by [1] abnormal GTT, [2] abnormal HOMA, or [3] elevated insulin levels, either fasting or 2 hours after 75 g oral glucose load. All patients with PCOS with abnormal GTT (pre-dm or DM) were assumed to be IR, since IR is the overwhelming etiology of glucose intolerance in these women and HOMA is not accurate in the presence of hyperglycemia (9, 18). HOMA was calculated using the formula HOMA ¼ (fasting insulin miu/ ml fasting glucose mmol/l)/22.5, and IR was diagnosed when HOMA was >2.67 (21, 22). Insulin levels were considered elevated when they were above the laboratory reference range based on the 95th percentile for the US population, and IR was diagnosed if either fasting insulin >19 miu/ml or 2-hour (after 75 g oral glucose) insulin was >79 miu/ml 2 hours. Obesity was defined as body mass index (BMI) >30 kg/m 2, according to National Institutes of Health guidelines (23). Laboratory Methods Serum glucose was measured on the Siemens Dimension series chemistry analyzers. Serum insulin and total T were measured using the Advia Centaur chemiluminescent immunoassay system (Siemens Healthcare Diagnostics); 17-hydroxyprogesterone was measured using the DPC Coat-A-Count radioimmunoassay (Diagnostic Products); DHEAS was measured using the Immulite 2000 chemiluminescent immunoassay (Siemens); HbA1c was measured by ion exchange chromatography with a National Glycohemoglobin Standardization Program certified Variant Turbo II analyzer (Bio-Rad Laboratories). Statistical Analysis Two approaches for diagnosing pre-dm and DM and five approaches for diagnosing IR were compared using percent agreement and Cohen s kappa (k) coefficient (24, 25). Positive and negative percent agreement (PPA and NPA, respectively) are calculated identically to sensitivity and specify and used when none of the diagnostic tests compared are the gold standard (24). Cohen s k coefficient is a standard statistical measure of overall agreement that corrects for chance agreement and uses the following agreement scale: none, <0; slight, ; fair, ; moderate, ; substantial, ; and excellent, (25). Results were expressed as mean SD or n (%). Analyses were conducted using PASW version 18 statistical software (SPSS). RESULTS A total of 125 patients with apparent PCOS were recruited for this study. Of these, 14 were excluded because they were found to have congenital adrenal hyperplasia (n ¼ 2), sickle cell anemia (n ¼ 2), hyperprolactinemia (n ¼ 4), or hypothyroidism (n ¼ 4) or because they were taking metformin (n ¼ 2). Characteristics for TABLE 1 Clinical and hormonal characteristics of subjects. Characteristic Data No. of subjects 111 Age 28 7 BMI, kg/m Obese, BMI >30 kg/m 2 (%) 80 (72) Ovulatory dysfunction (%) 82 (74) Hirsutism (%) 70 (63) T, ng/dl DHEAS, mg/dl hydroxyprogesterone, pmol/l Note: Data are mean SD or n (%) Hurd et al. Diabetes and insulin resistance screening in PCOS Vol. 96, No. 4, October 2011

3 FIGURE 1 Comparison of diagnostic approaches for IR, DM, and pre-dm. Vertical lines delineate classifications using a combination of 2- hour GTT, HOMA, and insulin levels, both fasting and 2 hours after 75 g oral glucose load (Insulin). Bars demonstrate the shifts in classifications according to diagnostic approach. Obesity ¼ equating obesity (BMI >30 kg/m 2 ) with IR. the remaining 111 women included in the analysis are described in Table 1. Diagnosis of Pre-DM and DM According to GTT results, 20% (22/111) of the subjects were identified as pre-dm and 4% (5/111) as DM (Fig. 1 and Table 2). HbA1c had fair agreement with GTT in diagnosing Pre-DM or DM (k ¼ 0.29). However, HbA1c classified two subjects as pre-dm whom GTT had identified as DM, and vice versa. In addition, HbA1c classified 44% (12/27) subjects as normal who had pre-dm according to GTT, all of whom had IR according to HOMA. HbA1c classified an additional 24% (20/84) subjects with normal GTTS as having pre- DM, and 95% (19/20) of these subjects had evidence of IR. FPG was elevated (>99 mg/ml) in 63% (17/27) of those with pre-dm or DM according to GTT (Fig. 1 and Table 2). As a result, using elevated FPG as the initial screening failed to identify 41% (9/22) of subjects with pre-dm and 20% (1/5) with DM because their FPG values were normal. Although not as sensitive, FPG had substantial agreement with GTT (k ¼ 0.74). Diagnosis of IR The 24% (27/111) of subjects with Pre-DM or DM according to GTT results were considered to have IR. Using a combination of HOMA and elevated insulin levels, IR was detected in an additional 74% (62/84) of the remaining subjects with normal GTTs (Fig. 1 and Table 3). Thus, a total of 80% (89/111) of subjects were identified as having IR using a combination of GTT, HOMA, and elevated insulin levels. The five approaches for diagnosing IR varied greatly in their agreement with a combination of GTT, HOMA, and insulin levels (Fig. 1 and Table 3). Using either HOMA or elevated insulin levels (fasting or 2 hour) in combination with GTT had substantial agreement (k ¼ 0.70 and 0.73, respectively) and detected approximately 85% (76/89 and 77/89, respectively) of patients with PCOS with apparent IR. Although the IR rates using HOMA and insulin levels were similar, these were not all the same subjects: 13 subjects with borderline IR identified by HOMA (normal fasting glucose and insulin) were not detected using insulin levels, while 14 subjects with normal HOMA had elevated 2-hour insulin levels consistent with IR. Not surprisingly, elevated 2-hour insulin levels were found in 81% (22/27) of subjects with abnormal GTTs. Using a combination of HbA1c and HOMA to detect IR also showed substantial agreement (k ¼ 0.73). Obesity, particularly when used with FPG, was relatively inaccurate in detecting apparent IR (Fig. 1 and Table 3). Obesity plus GTT had fair agreement (k ¼ 0.33), and obesity plus FPG had slight agreement (k ¼ 0.18). Both approaches designated almost 10% of subjects as having IR who had no evidence by other screening methods and missed 14% 25% who had apparent IR by other methods. When used with FPG, this approach missed the 47% (9/19) of nonobese women who had laboratory evidence of IR. Using obesity as an IR marker misclassified 23% 36% (26/111 and 38/111, respectively) of subjects. DISCUSSION This study provides insights into the relative accuracy of the various methods used clinically to detect IR, pre-dm, and DM in women with PCOS. We found that HbA1c, which reflects average plasma glucose values during the preceding months, has fair agreement with GTT as to which patients have either pre-dm or DM. Our findings are consistent with previous studies of the accuracy of HbA1c in the general population (26). HbA1c apparently misses women with abnormal GTTs whose average plasma glucose levels remain TABLE 2 Comparison of pre-dm and DM diagnostic strategies for women with PCOS. GTT Pre-DM DM normal PPA NPA k (95% CI) HbA1c Pre-DM/DM 15 (13) 20 (18) Normal 12 (11) 64 (58) ( ) FPG Pre-DM/DM 17 (16) 0 Normal 9 (8) 84 (76) ( ) Note: Values are n (%) unless otherwise noted. Pre-DM ¼ fasting glucose FPG >99 mg/dl, 2-hour glucose >139 mg/dl, or HbA1c >5.6%; GTT ¼ 2-hour GTT. Fertility and Sterility â 1045

4 TABLE 3 Comparison of IR diagnostic strategies for women with PCOS. GTT D HOMA D insulin IR Normal PPA NPA k (95% CI) GTT þ HOMA IR 76 (68) 0 Normal 13 (12) 22 (20) ( ) GTT þ insulin IR 77 (69) 0 Normal 12 (11) 22 (20) ( ) HbA1c þ HOMA IR 79 (71) 1 (1) Normal 10 (9) 21 (19) ( ) GTT þ obesity IR 73 (66) 10 (9) Normal 16 (14) 12 (11) ( ) FPG þ obesity IR 61 (55) 10 (9) Normal 28 (25) 12 (11) ( ) Note: Values are n (%) unless otherwise indicated. Women with type 2 DM or pre-dm were considered to have IR. GTT ¼ 2-hour GTT; insulin ¼ elevated insulin levels, fasting or 2 hours after 75 g oral glucose load; obesity ¼ BMI >30 kg/m 2. normal because of diet and/or activity but detect a few additional women at increased risk for DM despite normal GTT and HOMA, most of whom had elevated 2-hour insulin levels. Because <10% of women with PCOS will have DM, a larger study is required to determine the agreement between GTT and HbA1c for diagnosing DM in women with PCOS. As other investigators have reported, using FPG as a screening test for women with PCOS has high specificity but is relatively insensitive (27). In the present study, relying on FPG would have missed more than one-third of subjects with pre-dm or DM. Both laboratory methods for detecting IR (HOMA and insulin levels) identified approximately the same number of women with IR in our population. All subjects with pre-dm or DM (24% of the study population) were assumed to have IR. Of the remaining patients, HOMA identified more than half as having IR, and approximately 25% of those identified had normal fasting insulin and glucose levels. Using elevated fasting or 2-hour insulin levels as criteria for IR missed subjects with normal fasting insulin levels but identified additional subjects with IR who manifested as elevated 2-hour insulin levels undetected by HOMA. Using the combination of HOMA and insulin levels identified the greatest number of women with IR at both ends of the IR spectrum. Obesity was a relatively imprecise marker for IR in women with PCOS without pre-dm or DM compared with using HOMA and insulin levels. In our population, this approach inappropriately classified 10% of obese women as having IR who had no evidence of IR by HOMA or insulin levels and missed approximately half of nonobese women who had laboratory evidence of IR. Our findings are consistent with previous controlled studies of IR in women with PCOS that found that approximately 50% 70% of both obese and nonobese women have IR (3, 28). Aweakness of this study was that no definitive test for IR was performed on our subjects, such as the hyperinsulinemic-euglycemic clamp test (17). For this reason, the diagnosis of IR had to be inferred. It is accepted that women with abnormal glucose tolerance (i.e., abnormal GTT or elevated HbA1c) are extremely likely to have IR (27). The 80% IR rate we found in our subjects using a combination of GTT, HOMA, and elevated insulin levels is consistent with previously published studies of women with PCOS (29). The results of this study are applicable to more than one related clinical issue. The first is the issue of how to screen for pre-dm and DM in women with PCOS. GTT has been the standard for a number of years, since FPG is relatively insensitive for women with PCOS (27). The use of HbA1c for DM screening rather than GTT, as recently recommended by the ADA, simplifies the screening process and might ultimately increase the rate of periodic rescreening (12). Our data suggest that, although results for HbA1c and GTT are not identical in women with PCOS, both tests identify a similar group of patients with impaired glucose tolerance. It is important to remember that when HbA1c R6.5%, the diagnosis of diabetes should be confirmed by repeating the HbA1c in those without clinical symptoms of DM or glucose levels >200 mg/dl (20). HbA1c and GTTwere not as congruent when identifying those with Pre-DM, an important group because they are candidates for increased frequency of DM screening according to ADA guidelines (12).More than 10% of subjects with Pre-DM according to GTT were not detected by HbA1c, apparently because their diet and/or activity level maintained average glucose levels in the normal range (30). However, almost 20% additional subjects with normal GTTs were identified by HbA1c as having Pre-DM (i.e., increased risk for diabetes ), and 95% of these subjects had laboratory evidence of IR. The second clinical issue to which this study is applicable is the comparability of different screening tests for IR in women with PCOS. Regardless of the approach used, women with PCOS determined to be Pre-DM and DM should be considered to have IR (9). For the remaining women with PCOS, equating obesity with IR is relatively inaccurate, since this approach will misclassify approximately one-third of women with PCOS and normal glucose tolerance. Using HOMA to detect IR had the advantage of simplicity but misses almost one-fifth of women with PCOS and normal glucose tolerance whose only evidence of IR are elevated 2-hour insulin levels. Using fasting and 2-hour insulin levels to detect IR will 1046 Hurd et al. Diabetes and insulin resistance screening in PCOS Vol. 96, No. 4, October 2011

5 identify these women but requires the equivalent of a GTT. On the basis of these results, a case can be made in women with PCOS to examine fasting and 2-hour insulin levels when a GTT is performed but to rely on HOMA alone when HbA1c is used as a DM screen. The question remains, is there value in screening women with PCOS for IR? If clinicians increase the frequency of DM screening in the presence of risk factors in addition to PCOS, IR screening would identify women with PCOS at increased risk for DM who are neither obese nor have pre-dm (12, 31). However, a case could be made for frequent screening of all women with PCOS for DM, since most (89% in our study) will have additional risk factors, including obesity, pre-dm, and/or IR. Since HbA1c is a simpler screening method than GTT, it is likely that more frequent screening for all women with PCOS will be recommended in the future. Some clinicians recommend screening for IR to guide therapy (7, 11, 31, 15, 16). Although insulin level assessment has not been shown to identify women who will respond to therapy, no study has demonstrated that women without IR benefit from taking insulin sensitizers (10). When insulin sensitizers such as metformin are being considered to treat symptoms in women with PCOS (e.g., ovulation dysfunction or hirsutism), evaluation for IR might help differentiate patients more or less likely to benefit from prolonged drug treatment. In summary, both GTT and HbA1c appear to be useful screens for DM in women with PCOS. Although the results of these two tests are not identical, HbA1c has an advantage in terms of convenience and expense since it is a single blood test that does not require fasting. Current guidelines recommend repeat DM screening at least every 3 years for women with PCOS. However, it might be justifiable to screen these women more frequently, since most will have additional risk factors for developing DM, including obesity, glucose intolerance, and IR. For clinicians who wish to evaluate women with PCOS for IR to guide diagnostic approach or therapy, women with pre-dm or DM should be considered to have IR. For the remaining women with PCOS, both elevated HOMA and elevated insulin levels (fasting or 2 hours after oral glucose load) appear to identify the majority of those with IR. Both FPG as a DM screen and obesity as a marker for IR are relatively inaccurate for women with PCOS. REFERENCES 1. 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: Ovalle F, Azziz R. Insulin resistance, polycystic ovary syndrome, and type 2 diabetes mellitus. Fertil Steril 2002;77: Dunaif A, Segal KR, Futterweit W, Dobrjansky A. Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 1989;38: Wild RA, Carmina E, Diamanti-Kandarakis E, Dokras A, Escobar-Morreale HF, Futterweit W, et al. Assessment of cardiovascular risk and prevention of cardiovascular disease in women with the polycystic ovary syndrome: a consensus statement by the Androgen Excess and Polycystic Ovary Syndrome (AE-PCOS) Society. J Clin Endocrinol Metab 2010;95: 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: Azziz R, Carmina E, Dewailly D, Diamanti- Kandarakis E, Escobar-Morreale HF, Futterweit W, et al. Positions statement: criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: an Androgen Excess Society guideline. J Clin Endocrinol Metab 2006;91: Diamanti-Kandarakis E, Economou F, Palimeri S, Christakou C. Metformin in polycystic ovary syndrome. Ann N Y Acad Sci 2010;1205: Ehrmann DA, Barnes RB, Rosenfield RL, Cavaghan MK, Imperial J. Prevalence of impaired glucose tolerance and diabetes in women with polycystic ovary syndrome. Diabetes Care 1999;22: Salley KE, Wickham EP, Cheang KI, Essah PA, Karjane NW, Nestler JE. Glucose intolerance in polycystic ovary syndrome a position statement of the Androgen Excess Society. J Clin Endocrinol Metab 2007;92: ACOG Committee on Practice Bulletins Gynecology. ACOG Practice Bulletin No. 108: Polycystic ovary syndrome. Obstet Gynecol 2009;114: American Association of Clinical Endocrinologists Polycystic Ovary Syndrome Writing Committee. American Association of Clinical Endocrinologists position statement on metabolic and cardiovascular consequences of polycystic ovary syndrome. Endocr Pract 2005;11: American Diabetes Association. Standards of medical care in diabetes Diabetes Care 2010;33:S 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: American Diabetic Association. Consensus Development Conference on Insulin Resistance. Diabetes Care 1998;21: Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care 2004;27: Nestler JE, Stovall D, Akhter N, Iuorno MJ, Jakubowicz DJ. Strategies for the use of insulin-sensitizing drugs to treat infertility in women with polycystic ovary syndrome. Fertil Steril 2002;77: Kauffman RP, Baker VM, Dimarino P, Gimpel T, Castracane VD. Polycystic ovarian syndrome and insulin resistance in white and Mexican American women: a comparison of two distinct populations. Am J Obstet Gynecol 2002;187: Kang ES, Yun YS, Park SW, Kim HJ, Ahn CW, Song YD, et al. Limitation of the validity of the homeostasis model assessment as an index of insulin resistance in Korea. Metabolism 2005;54: Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 1961;21: Gillett MJ. International Expert Committee report on the role of the A1c assay in the diagnosis of diabetes. Diabetes Care 2009;32: Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28: Bonora E, Kiechl S, Willeit J, Oberhollenzer F, Egger G, Targher G, et al. Prevalence of insulin resistance in metabolic disorders: the Bruneck Study. Diabetes 1998;47: Clinical Guidelines on the Identification. Evaluation, and Treatment of Overweight and Obesity in Adults The Evidence Report. National Institutes of Health. Obes Res 1998;6:S Food and Drug Administration. Statistical guidance on reporting results from studies evaluating diagnostic tests. Rockville, MD: Food and Drug Administration, DeviceRegulationandGuidance/GuidanceDocuments/ ucm htm. 25. Vach W. The dependence of Cohen s kappa on the prevalence does not matter. J Clin Epidemiol 2005;58: Bennett CM, Guo M, Dharmage SC. HbA(1c) as a screening tool for detection of type 2 diabetes: a systematic review. Diabet Med 2007;24: Legro RS, Kunselman AR, Dodson WC, Dunaif A. Prevalence and predictors of risk for 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: 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: Carmina E, Lobo RA. Use of fasting blood to assess the prevalence of insulin resistance in women with polycystic ovary syndrome. Fertil Steril 2004;82: Rodbard HW, Blonde L, Braithwaite SS, Brett EM, Cobin RH, Handelsman Y, et al. American Association of Clinical Endocrinologists medical guidelines for clinical practice for the management of diabetes mellitus. Endocr Pract 2007;13(Suppl 1): M ohlig M, Fl oter A, Spranger J, Weickert MO, Schill T, Schl osser HW, et al. Predicting impaired glucose metabolism in women with polycystic ovary syndrome by decision tree modelling. Diabetologia 2006;49: Fertility and Sterility â 1047