Received

2011; pu

Y.-H.L. ha

has noth

nothing

The first tw

Supported

038-001

(98TMU

Reprint req

cology,

Xinglong

FC

230

Antim€ullerian hormone and polycystic ovary syndrome

Yi-Hui Lin, M.D.,a Wan-Chun Chiu, Ph.D.,c Chien-Hua Wu, Ph.D.,b,e Chii-Ruey Tzeng, M.D.,d

Chun-Sen Hsu, M.D.,a and Ming-I Hsu, M.D.a

a Department of Obstetrics and Gynecology, and b Center of Excellence for Clinical Trials and Research, Wan Fang Hospital,c School of Nutrition and Health Sciences, College of Public Health and Nutrition, and d Department of Obstetrics and

Gynecology, Taipei Medical University Hospital, Taipei Medical University, Taipei; and e Department of Applied

Mathematics, Chung-Yuan Christian University, Chung-Li, Taiwan

Objective: To assess the relationship between antim€ullerian hormone (AMH) and parameters related to polycysticovary syndrome (PCOS).Design: Prospective study.Setting: Academic tertiary care center.Patient(s): A total of 290 women.Intervention(s): None.Main Outcome Measure(s): Parameters related to insulin resistance and metabolic syndrome.Result(s): Women with polycystic ovary morphology had significantly higher AMH levels than women in the con-trol group. The prevalence of PCOS increased from 21% in the low-AMH (<4 ng/mL) group to 37% in themoderate-AMH (4–11 ng/mL) group and 80% in the high-AMH (>11 ng/mL) group. However, significantdifferences in insulin resistance parameters were not observed among groups. The results of the correlationanalysis revealed that AMH levels were positively correlated with LH, total T, A, and total cholesterol content;however, AMH levels were negatively correlated with age, body mass index, and the number of menstrualcycles per year. AMH levels were not correlated with insulin resistance parameters.Conclusion(s): Elevated serumAMH levels increase the risk of PCOS but do not affect the risk of insulin resistanceor metabolic syndrome. (Fertil Steril� 2011;96:230–5. �2011 by American Society for Reproductive Medicine.)

Key Words: AMH, PCOS, insulin resistance, metabolic syndrome

Antim€ullerian hormone (AMH) levels indicate the quantity of theovarian follicle pool and may be a useful marker of ovarian reserves(1). The measurement of AMH levels can be useful in assisted repro-ductive technologies for predicting poor response, cycle cancella-tion, hyperresponse, and ovarian hyperstimulation syndrome (2).In addition, serum AMH levels are an important marker of polycys-tic ovary syndrome (PCOS). Women with PCOS have 2- to 3-foldhigher levels of AMH than healthy women (3). Serum AMH levelsare correlated with three diagnostic hallmarks of PCOS, includinghyperandrogenism (HA), oligoanovulation (ANOV), and polycysticovary morphology (PCOM) (4, 5). The differences in circulatingAMH levels between the main phenotypic groups of women withPCOS appear to reflect the severity of the syndrome (6).

Do high AMH levels increase the risk of PCOS, insulin resis-tance, and/or metabolic abnormalities? Are the serum levels ofAMH associated with various PCOS phenotypes? Are the clinicaland/or biochemical presentations between high and lowAMH levelsin women with and without PCOS different?

November 28, 2010; revised March 28, 2011; accepted April 5,

blished online May 5, 2011.

s nothing to disclose. W.-C.C. has nothing to disclose. C.-H.W.

ing to disclose. C.-R.T. has nothing to disclose. C.-S.H. has

to disclose. M.-I.H has nothing to disclose.

o authors contributed equally to this work.

by grants from the National Science Council (NSC 98-2629-B-

-MY3) and Taipei Medical University and Wan Fang Hospital

-WFH-05–03 and DOH100-TD-B-111–003).

uests: Ming-I Hsu, M.D., Department of Obstetrics and Gyne-

Wan Fang Hospital, Taipei Medical University, No.111, Sec. 3,

Rd., Taipei 11696, Taiwan (E-mail: hsumingi@yahoo.com.tw).

ertility and Sterility� Vol. 96, No. 1, July 2011opyright ª2011 American Society for Reproductive Medicine, P

The prospective study described in the present report wasconducted to answer these questions.

MATERIALS AND METHODSThe study was approved by the Institutional Review Board of the Wan Fang

Medical Center at Taipei Medical University, Taipei, Taiwan, and was con-

ducted at the Reproductive Endocrinology Clinic from June 1, 2009, to Sep-

tember 31, 2010. The study population consisted of patients and healthy

volunteers. Patients who complained of infertility, menstrual irregularity,

and hyperandrogenismwere recruited from the gynecologic outpatient clinic.

Each of the studied subjects gave informed consent. The following subjects

were excluded from the study and control populations: 1) women who had

been diagnosed with other etiologies that should be excluded when diagnos-

ing PCOS (7); 2) women who experienced menarche <3 years before the

beginning of the study; 3) women who received hormones or drugs for major

medical diseases; 4) women who presented ovarian cysts or ovarian tumors;

and 5) women who were >45 years old. In total, 290 women were recruited.

The subjects’ medical histories were obtained, and the number of men-

strual cycles during the previous year was recorded. The waist-to-hip ratio

(WHR) was defined as the waist girth/hip girth, and the body mass index

(BMI) was defined as the body weight in kilograms divided by the body

height in meters squared (kg/m2). Obesity was defined as a BMIR25 kg/m2.

The following components were measured and calculated: 1) total testos-

terone (T), androstenedione (A), DHEAS, and 17a-OH-progesterone levels

and the free androgen index (FAI); 2) fasting insulin, fasting glucose,

2-hour oral glucose tolerance test (OGTT) glucose level, the homeostasis

model assessment of insulin resistance index (HOMA-IR), and the presence

of diabetes; 3) serum FSH, LH, and PRL; 4) total cholesterol, triglycerides

(TG), high-density lipoprotein (HDL), and low-density lipoprotein (LDL);

and 5) glutamic oxaloacetic transaminase (GOT), glutamic pyruvic transam-

inase (GPT), and high-sensitivity C-reactive protein (hsCRP) levels.

AMH levels were measured with an ELISA kit (Diagnostic Systems Lab-

oratories). The FAI was calculated by using the formula FAI¼ T (nmol/L)�

0015-0282/$36.00ublished by Elsevier Inc. doi:10.1016/j.fertnstert.2011.04.003

FIGURE 1

The mean serum antim€ullerian hormone (AMH) levels of various polycystic ovary syndrome–related phenotypes. Error bars represent 95%

confidence intervals. ANOV ¼ oligoanovulation; HA ¼ hyperandrogenism; PCOM ¼ polycystic ovary morphology.

Lin. Antim€ullerian hormone and PCOS. Fertil Steril 2011.

100/SHBG (nmol). Hyperandrogenism was defined as a total serum T level

R2.98 mmol/L.

PCOS was diagnosed according to the 2003 Rotterdam criteria, which re-

quire a minimum of two of the three following traits: ANOV, HA, and

PCOM. The definitions of ANOV, HA, and PCOM have been previously de-

scribed (7). A single ultrasound machine was used for the evaluation of

PCOM (Medison SA-8000 Live).

Insulin Sensitivity, Metabolic Syndrome, and GlucoseIntoleranceThe insulin sensitivity index was evaluated by HOMA-IR, and the following

formula was applied: HOMA-IR ¼ (fasting insulin [mU/mL] � fasting glu-

cose [mg/dL])/405. World Health Organization 2006 diagnostic criteria for

diabetes were used (fasting plasma glucose [FPG] R126 mg/dL or 2-hour

plasma glucose R200 mg/dL). Impaired glucose tolerance (IGT) was de-

fined as 2-hour glucose levels of 140–199 mg/dL in the 75-g oral glucose tol-

erance test. In women with IGT, the FPG level should be <126 mg/dL (8).

Metabolic syndrome was defined (2005 National Cholesterol Education

Program—Adult Treatment Panel III [ATP III]) as the presence of at least

three of the following criteria: abdominal obesity (waist circumference

>80 cm in women), serum TG R150 mg/dL, serum HDL <50, blood pres-

sure (BP) R130/85 mm Hg, and FPG R100 mg/dL.

Statistical AnalysisStatistical analysis was performed using SPSS 13.0 for Windows. Figure 1

presents the mean, error bars, and 95% confidence intervals of the AMH

levels of the eight subgroups. In Table 1, the data are presented as mean �SD. To compare the means of more than two groups, one-way analysis of var-

iance and post hoc Dunnett tests were performed, and equal variances were

not assumed. Receiver operating characteristic (ROC) curves of the AMH

levels were constructed to examine the diagnostic test performance for

PCOS. To determine the optimal thresholds for the AMH levels, the point

on the ROC curve with the maximumYouden index (sensitivity� [1� spec-

Fertility and Sterility�

ificity]) was calculated (9). We evaluated the correlation between the AMH

level and various parameters by determining the two-tailed Pearson correla-

tion coefficient. Differences among groups were considered to be significant

if the P value was < .05.

RESULTSIn total, 290 women were recruited, including 126 with PCOS and164 women without PCOS. The subjects were categorized into eightgroups based on their diagnostic characteristics. The 126 subjectswith PCOS were classified into four subgroups: ANOV þ HA,HA þ PCOM, ANOV þ PCOM, and HA þ ANOV þ PCOM.The 164 subjects without PCOS also were categorized into four sub-groups: HA only, ANOV only, PCOM only, and normal control.Figure 1 presents the mean serum AMH level of the eight subgroupsof PCOS-related phenotypes. The serum AMH levels of the fourPCOM subgroups were significantly greater than that of the controland ANOV-only subgroup. Women in the HA þ ANOV þ PCOMsubgroup had the highest levels of serum AMH, whereas womenin the normal control subgroup had the lowest levels.

The overall distribution of AMH in the entire population (n ¼290) was analyzed, and the subjects were divided into threegroups: high AMH (>11 ng/mL), moderate AMH (4–11 ng/mL),and low AMH (<4 ng/mL). The clinical and biochemical charac-teristics of subjects with low, moderate, and high AMH levels arepresented in Table 1. As the AMH level increased, the prevalenceof PCOS increased significantly from 21% in the low-AMH groupto 37% in the moderate-AMH group and 80% in the high-AMHgroup. Compared with the moderate AMH group, the rates ofHA, ANOV, and PCOM were also greater in the high-AMH group.Body weight, waist circumference, and BMI were significantlyhigher in the low-AMH group compared with the moderate- and

231

TABLE 1Biochemical and clinical characteristics of women in their reproductive age with low, moderate, and high levels of AMH.

Total AMH <4 (L) AMH 4–11 (M) AMH >11 (H)

P value

L vs. M L vs. H M vs. H

n 290 72 148 70

Age (y) 27.7 � 5.8 29.5 � 6.4 27.3 � 5.5 26.6 � 5.3 .040* .010* .727AMH (ng/mL) 7.9 � 5.3 2.4 � 1.0 7.0 � 2.0 15.4 � 4.2 < .001*** < .001*** < .001***

Menarche (y) 12.7 � 1.4 12.7 � 1.3 12.6 � 1.2 12.8 � 1.7 .983 .970 .843

No. of cycles per year 8.8 � 3.7 1.0 � 3.3 9.1 � 3.6 6.8 � 3.8 .191 < .001*** < .001***

Oligoanovulation 44% 31% 36% 73% .762 < .001*** < .001***Polycystic ovary morphology 43% 11% 43% 79% < .001*** < .001*** < .001***

Hyperandrogenemia 39% 25% 36% 57% .218 < .001*** .014*

Polycystic ovary syndrome 43% 21% 37% 80% .029* < .001*** < .001***

Obesity 45% 54% 45% 34% .521 .050 .319Anthropometric components

Weight (kg) 65.8 � 17.9 71.8 � 2.0 64.4 � 16.3 62.6 � 17.8 .023* .013* .839

Height (cm) 161.0 � 5.1 162.0 � 5.4 161.0 � 5.0 16.3 � 4.7 .488 .177 .756

Body mass index (kg/m2) 25.3 � 6.6 27.3 � 7.2 24.9 � 6.2 24.1 � 6.5 .047* .019* .790Waist (cm) 83.9 � 15.6 89.0 � 18.1 82.7 � 13.8 81.3 � 15.4 .033* .023* .890

Hip (cm) 10.2 � 11.4 104.1 � 12.1 99.6 � 1.6 97.5 � 11.4 .023* .003** .489

Waist-to-hip ratio .83 � 0.09 .85 � 0.09 .83 � 0.09 .83 � 0.09 .470 .603 1.000No. of follicles 8.7 � 5.2 3.9 � 2.9 8.6 � 4.3 13.7 � 4.2 < .001*** < .001*** < .001***

Androgens

Total T (nmol/L) 2.3 � 1.1 1.8 � 1.0 2.2 � 1.1 3.0 � 1.2 .046 < .001*** < .001***

Androstenedione (ng/dL) 2.7 � 1.3 2.2 � 1.1 2.7 � 1.1 3.5 � 1.3 .027* < .001*** < .001***Free androgen index 9.2 � 9.6 7.5 � 7.7 8.7 � 9.4 12.0 � 11.3 .681 .021* .111

DHEAS (ng/dL) 194.3 � 107.9 18.3 � 107.4 193.7 � 109.2 21.1 � 105.1 .770 .262 .640

17OH-P (ng/dL) 1.2 � 0.9 1.1 � 0.9 1.2 � 0.9 1.2 � 0.8 .772 .723 .998

Insulin sensitivity and glucose toleranceFasting insulin (uIU/mL) 1.8 � 13.5 13.1 � 16.4 9.3 � 9.1 11.6 � 17.2 .194 .929 .661

Fasting glucose (mg/dL) 92.7 � 16.8 95.7 � 18.3 92.5 � 17.1 9.0 � 13.9 .514 .110 .582

2-hour glucose (mg/dL) 114.2 � 45.3 12.8 � 51.1 112.1 � 41.5 112.0 � 46.6 .515 .638 1.000HOMA-IR 2.6 � 3.6 3.4 � 4.9 2.2 � 2.6 2.7 � 4.0 .159 .680 .789

Impaired glucose tolerance 21% 29% 20% 16% .608 .374 .893

Diabetes 5% 7% 4% 6% .783 .987 .939

Hormonal componentsLH (mIU/mL) 9.8 � 1.9 6.6 � 8.9 9.6 � 1.7 13.6 � 12.2 .096 .001** .062

FSH (mIU/mL) 6.2 � 2.0 6.4 � 2.6 6.1 � 1.9 6.1 � 1.7 .791 1.000 .840

TSH (mIU/mL) 1.8 � 1.0 1.6 � 1.0 1.9 � 1.1 1.8 � 0.8 .355 .775 .870

PRL (ng/mL) 14.5 � 5.3 15.4 � 5.9 14.6 � 4.9 13.4 � 5.2 .748 .109 .270SHBG (nmol/L) 42.4 � 26.5 43.1 � 29.6 42.9 � 26.0 4.6 � 24.5 1.000 .924 .894

Inflammatory components

hsCRP (mg/dL) 0.26 � 0.33 0.35 � 0.42 0.23 � 0.26 0.24 � 0.33 .091 .229 .998

Lin. Antim€ullerian hormone and PCOS. Fertil Steril 2011.

232Lin

etal.

Antim

€ullerianhorm

oneand

PCOS

Vol.

96,No.

1,July

2011

TABLE 1Continued.

Total AMH <4 (L) AMH 4–11 (M) AMH >11 (H)

P value

L vs. M L vs. H M vs. H

Liver function

GOT (IU/L) 23.5 � 14.9 26.3 � 2.9 22.6 � 12.4 22.5 � 12.0 .435 .479 1.000

GPT (IU/L) 23.2 � 22.9 26.2 � 28.6 22.5 � 22.8 21.5 � 15.6 .723 .549 .976Lipid profiles and blood pressure

Cholesterol (mg/dL) 182.4 � 33.5 175.5 � 35.2 181.9 � 31.3 19.6 � 35.0 .474 .034* .217

Triglycerides (mg/dL) 87.8 � 7.7 102.3 � 85.9 85.1 � 7.6 78.5 � 48.9 .370 .126 .806

HDL (mg/dL) 52.2 � 16.2 48.5 � 15.9 52.8 � 15.6 54.9 � 17.2 .165 .068 .778LDL (mg/dL) 108.6 � 3.2 104.6 � 28.6 107.5 � 29.1 115.1 � 33.4 .864 .130 .275

Systolic pressure (mm Hg) 112.9 � 17.5 117.3 � 17.1 111.9 � 16.7 11.3 � 19.0 .097 .079 .920

Diastolic pressure (mm Hg) 78.6 � 13.1 81.1 � 14.2 77.6 � 12.8 78.4 � 12.3 .236 .548 .963

Metabolic syndromeMetabolic syndrome 26% 34% 25% 20% .468 .183 .789

Hypertension 30% 36% 29% 24% .659 .332 .838

HDL <50 mg/dL 50% 64% 47% 44% .045* .056 .984Triglycerides >150 mg/dL 11% 15% 10% 9% .665 .523 .975

Waist >80 cm 52% 65% 50% 45% .108 .053 .865

Note: Unless indicated as percentage, values are mean � SD. GOT ¼ glutamic oxaloacetic transaminase; GPT ¼ glutamic pyruvic transaminase; HDL ¼ high-density lipoprotein; HOMA-IR ¼ homeostasis model

assessment insulin resistance index; hsCRP ¼ high-sensitivity C-reactive protein; LDL ¼ low-density lipoprotein.

* P< .05.

** P< .01.

*** P< .001.

Lin. Antim€ullerian hormone and PCOS. Fertil Steril 2011.

Fertilityand

Sterility

�233

FIGURE 2

Receiver-operating characteristic (ROC) curves of antim€ullerian

hormone (AMH) levels for the evaluation of polycystic ovarysyndrome (PCOS).

Lin. Antim€ullerian hormone and PCOS. Fertil Steril 2011.

high-AMH groups. The high-AMH group presented significantlyhigher total T and A levels than the low- and moderate-AMHgroups. Interestingly, significant differences in glucose tolerance, in-sulin resistance, liver enzyme concentration, lipid profile, or risk ofmetabolic syndrome were not observed among the low-, moderate-,and high-AMH groups. However, significant differences in the totalcholesterol level and the prevalence of HDL <50 mg/dL were ob-served among the groups.

Using the ROC curves, the diagnostic potency of the AMH assaywas tested. As shown in Figure 2, the AUC of AMH for the predictionof PCOS reached 0.774 (95% confidence interval [CI] 0.720–0.829).The largest Youden index and optimal specificity (76%) and sensitiv-ity (70%) for the prediction of PCOS were obtained when the thresh-old AMH concentration was set to 7.3 ng/mL.

The correlation analyses revealed that AMH levels were positivelycorrelated with LH (g ¼ 0.229; P<.001), total T (g ¼ 0.389;P<.001), A (g ¼ 0.396; P<.001), total cholesterol (g ¼ 0.174;P¼.003), HDL (g ¼ 0.119; P¼.044), LDL (g ¼ 0.136; P¼.020),and average ovarian volume (g ¼ 0.459; P<.001); however, theAMH concentration was negatively correlated with age (g ¼�0.205; P<.001), BMI (g¼�0.166; P¼.004), number of menstrualcycles per year (g ¼ �0.330; P<.001), and fasting glucose levels(g ¼ �0.130; P¼.027). AMH levels were not correlated with insulinresistance parameters (HOMA-IR), liver function (GOTand GPT), orhs-CRP. Logistic regression analysis was applied to PCOS-relatedparameters (AMH, LH, LH:FSH ratio, WHR, HOMA-IR, SHBG,HDL, GPT, and hsCRP), and the results revealed that AMH andSHBG were significantly associated with PCOS. Furthermore, a par-tial correlation analysis adjusted by age and BMI was conducted, andthe results revealed that the relationship between AMH levels andinsulin resistance parameters was independent.

234 Lin et al. Antim€ullerian hormone and PCOS

DISCUSSIONTo quantify the serum AMH levels and various factors related toPCOS, the subjects were categorized into eight subgroups accord-ing to the three hallmarks of PCOS. The results showed that theserum AMH levels of the four PCOM subgroups were signifi-cantly higher than that of the normal control subgroup. Similarlyto earlier reports, the results of the correlation analyses demon-strated that AMH was positively correlated with the total T andA levels. Furthermore, increased AMH levels were correlatedwith a reduction in the number of menstrual cycles per year (4,5). These results are consistent with those obtained by Pioukaet al., who demonstrated that high AMH levels reflect theseverity of PCOS (6).

Because AMH levels are positively correlated with the three com-ponents of PCOS, the concentration of AMH should be consideredto be a predictor of PCOS. In a study conducted on a cohort of 73PCOS patients and 96 control subjects, Pigny et al. (10) conductedan ROC analysis and found that the area under the curve of an AMHassay for the prediction of PCOS was equal to 0.851 (95% CI 0.796–0.905).When the threshold was set to 60 pmol/L (8.4 ng/mL), a com-promise between specificity (92%) and sensitivity (67%) wasobtained. In the present study, the area under the ROC curve forAMH reached 0.774 (95% CI 0.720–0.829) for the prediction ofPCOS. The primary difference between the present investigationand the study conducted by Pigny et al. is that 68 PCOS-risk caseswere included in the present study. These patients were diagnosedwith one of three PCOS components (HA, PCOM, or ANOV) andwere included in the non-PCOS group. PCOS is not a rare disease;therefore, to prevent selection bias, women with HA, ANOV, orPCOM that have been diagnosed as non-PCOS should not be ex-cluded (11). Our analysis demonstrated that the largest Youden in-dex and the optimal combined specificity (76%) and sensitivity(70%) for the prediction of PCOS were obtained when the AMHthreshold was set to 7.3 ng/mL. However, our results, and those ofPigny et al., suggest that the evaluation of serum AMH levels isan important breakthrough in the diagnosis and understanding ofPCOS (3).

In the present study, the subjects were classified into low-,moderate-, and high-AMH subgroups, and the results suggestedthat the prevalence of PCOS increased with an increase in AMHconcentration (from 21% in the low-AMH group to 37% in themoderate-AMH group and 80% in the high-AMH group). Alterna-tively, glucose intolerance, insulin resistance, lipid profile, andrisk of metabolic syndrome (except for total cholesterol level andHDL content) were not significantly different among the low-, mod-erate-, and high-AMH subgroups. In fact, the relationship betweeninsulin resistance and AMH has been a source of controversy (12,13). Chen et al. (12) discovered that AMHwas negatively associatedwith BMI and HOMA-IR. In contrast, Park et al. (13) observed anindependent relationship between HOMA-IR and AMH in womenwithout PCOS. The results of the present study suggest that an in-crease in the serum AMH level may increase the risk of PCOS inwomen of reproductive age. However, the AMH level did not affectthe risk of insulin resistance or metabolic syndrome. In the presentstudy and in earlier reports, serum AMH levels were negatively cor-related with age and BMI (2, 13). Although the low AMH group hadthe highest average age and BMI, an independent relationshipbetween AMH and insulin resistance and/or metabolic syndromewas observed when a partial correlation analysis adjusted by ageand BMI was conducted.

The subjects evaluated in the present study were recruited fromthe outpatient clinic of a tertiary care center and do not reflect the

Vol. 96, No. 1, July 2011

true distribution of the general population. As described in the inclu-sion criteria, women diagnosed with conditions that should be ex-cluded in the diagnosis of PCOS, such as hyperprolactinemia,premature ovarian failure, or even premenopause, were not includedin the present study. Therefore, the results should be applied to thegeneral population with caution.

Fertility and Sterility�

CONCLUSIONIn women with PCOS, AMH levels may reflect specific ovariananomalies but are not indicative of metabolic anomalies. Elevatedserum AMH levels are associated with an increased risk of PCOS;however, an increase in AMH levels was not correlated with an in-creased risk of insulin resistance and metabolic syndrome.

REFERENCES

1. Kwee J, Schats R, McDonnell J, Themmen A, de

Jong F, Lambalk C. Evaluation of anti-m€ullerian hor-

mone as a test for the prediction of ovarian reserve.

Fertil Steril 2008;90:737–43.

2. la Marca A, Sighinolfi G, Radi D, Argento C,

Baraldi E, Artenisio AC, et al. Anti-mullerian hor-

mone (AMH) as a predictive marker in assisted repro-

ductive technology (ART). Hum Reprod Update

2010;16:113–30.

3. Pigny P, Merlen E, Robert Y, Cortet-Rudelli C,

Decanter C, Jonard S, et al. Elevated serum level of

antimullerian hormone in patients with polycystic

ovary syndrome: relationship to the ovarian follicle

excess and to the follicular arrest. J Clin Endocrinol

Metab 2003;88:5957–62.

4. Carlsen SM, Vanky E, Fleming R. Anti-m€ullerian

hormone concentrations in androgen-suppressed

women with polycystic ovary syndrome. Hum Re-

prod 2009;24:1732–8.

5. Pellatt L, Rice S, Mason H. Anti-mullerian hormone

and polycystic ovary syndrome: a mountain too high?

Reproduction 2010;139:825–33.

6. Piouka A, Farmakiotis D, Katsikis I, Macut D,

Gerou S, Panidis D. Anti-mullerian hormone levels

reflect severity of PCOS but are negatively influenced

by obesity: relationship with increased luteinizing

hormone levels. Am J Physiol Endocrinol Metab

2009;296:E238–43.

7. HsuMI, Liou TH, Liang SJ, Su HW,WuCH, Hsu CS.

Inappropriate gonadotropin secretion in polycystic

ovary syndrome. Fertil Steril 2009;91:1168–74.

8. WorldHealth Organization. Definition and diagnosis of

diabetes mellitus and intermediate hyperglycemia; re-

port of aWHO/IDF consultation.WHODocument Pro-

duction. Geneva: World Health Organization; 2006.

9. Pepe M. The statistical evaluation of medical tests for

classification and prediction. New York: Oxford Uni-

versity Press; 2003.

10. Pigny P, Jonard S, Robert Y, Dewailly D. Serum anti-

mullerian hormone as a surrogate for antral follicle

count for definition of the polycystic ovary syndrome.

J Clin Endocrinol Metab 2006;91:941–5.

11. Bloom MS, Schisterman EF, Hediger ML. Select-

ing controls is not selecting ‘‘normals’’: design

and analysis issues for studying the etiology of

polycystic ovary syndrome. Fertil Steril

2006;86:1–12.

12. Chen MJ, Yang WS, Chen CL, Wu MY, Yang YS,

Ho HN. The relationship between anti-mullerian hor-

mone, androgen and insulin resistance on the number

of antral follicles in women with polycystic ovary

syndrome. Hum Reprod 2008;23:952–7.

13. Park HT, Cho GJ, Ahn KH, Shin JH, Kim YT, Hur JY,

et al. Association of insulin resistance with anti-

mullerian hormone levels in women without polycys-

tic ovary syndrome (PCOS). Clin Endocrinol (Oxf)

2010;72:26–31.

235