Anti-Müllerian Hormone (AMH) New Cutoff as a Possible Tool for the Diagnosis of Polycystic Ovary Syndrome (PCOS)

Al-Jefout M, Alnawaiseh C, Saleh M and Warwar K

Published on: 2021-02-20

Abstract

Objective: to find AMH cutoff value for PCOS diagnosis.

Methods: A total of 105 women: 85 with PCOS and 20 without PCOS. The cutoff value of the AMH level was evaluated using the receiver operating characteristic (ROC) analysis.

Results: AMH levels were significantly higher in PCOS women than controls (p < .001). AMH levels did not differ between PCOS phenotypes. ROC of AMH shows that participants with the cutoff of 23.97 pmol/L had the area under the curve of 0.844 (95% CI, 0.756–0.932), p-value <0.001 76.5% sensitivity, and 85% specificity for the presence or absence of PCOS. AMH in the PCOS group shows a gradual and steady decrease around the age of 50, while among controls, there was a sharp decrease at 35. Of all hormones, AMH correlated positively only with the Free androgen Index (p-value = 0.016). Lipid indices and AMH among PCOS participants did correlate, p-value >0.05.

Conclusions: The serum AMH level cutoff of 23.9 has good sensitivity and high specificity and maybe a new and useful diagnostic tool in PCOS diagnosis.

Keywords

Polycystic ovary syndrome; Anti-Müllerian hormone; Diagnosis; PCOS phenotypes

Introduction

PCOS is a heterogeneous endocrine disease that can be presented by oligo- or anovulation (O), biochemical or clinical hyperandrogenism (H.A.), and polycystic ovaries (PCO). Patients with PCOS can be grouped according to their PCOS features in four possible phenotypes [1] as follows: Type A: hyperandrogenism, chronic anovulation, and polycystic ovaries (O+ HA+ PCO). Type B: hyperandrogenism and chronic anovulation (O+ H.A.). Type C: hyperandrogenism and polycystic ovaries (HA+ PCO) and Type D: chronic anovulation and polycystic ovaries (O+ PCO).

Several publications have supported the anti-Müllerian hormone (AMH) role in the pathophysiology of PCOS [2-4]. Anti-Müllerian hormone (AMH) is mainly secreted by the ovary's granulosa cells, the preantral and antral follicles, and belongs to the group of transforming growth factor-beta [5]. AMH prevents primordial follicle recruitment by inhibiting the transition of the primordial to the primary follicular step and affecting FSH action on early follicular development [6]. There is substantial evidence reporting serum AMH elevation in PCOS is mainly caused by an increase in the numbers of small antral follicles [7], in addition to the synergistic effect of hyperinsulinism secondary to insulin resistance on AMH concentrations in PCOS patients [8,9]. However, the methodology by which AMH was measured is behind the lack of proper AMH cutoff.

The quest for a better diagnostic test for PCOS is a matter of debate; At the same time, some researchers suggested several cutoff values of AMH; however, a recent paper by Teede et al. [10] claimed that lack of a well-defined PCOS population in addition to the lack of international standardization in AMH assays are a real obstacle for the establishment of clear cut off values.

In this study, we investigated serum levels of AMH in PCOS and different phenotypes and the association of AMH levels with hormonal and metabolic parameters and tried to find a cut off value of AMH to help in the diagnosis of PCOS.

Methods

This study was approved by the Tawam Hospital Human Research Ethics committee SA/AJ/628 2019 (T-HREC). Women, aged 14–50 years, who had a new diagnosis of PCOS and were seen at Tawam Hospital in the Department of Obstetrics & Gynecology (AL-ain city, UAE) from January 1, 2017, through December 31, 2020, were considered for the study. The Control group was considered of women without PCOS attending for other gynecological conditions. To be included in the study, patients had to be newly diagnosed with PCOS and had to have undergone clinical evaluation at Tawam Hospital in the Department of Obstetrics & Gynecology. The Rotterdam criteria were used for the PCOS diagnosis, that is, to have at least two of three criteria: (1) clinical or biochemical evidence of hyperandrogenism; (2) oligo-ovulation or anovulation; and (3) Polycystic Ovarian Morphology PCOM [11]. Polycystic ovaries (as defined on ultrasonography as the presence of increased ovarian volume > 10 ml, and or > 12 follicles in each ovary, measuring 2–9 mm in diameter). For the assessment of clinical hyperandrogenism, the scoring of hirsutism was performed using the modified Ferriman–Gallwey system [12] with hirsutism to be considered if score>6, and alopecia was evaluated with Ludwig's scoring system [13].

Hormonal Assessment

Blood samples were collected from all participants after having been instructed not to eat, drink, or smoke. All patients reported no consumption of alcohol ever. A 10-cc sample of blood was drawn during day 2–6 of the menstrual cycles (natural or bleeding after progestin withdrawal) in plain-top tubes for subsequent hormonal analysis. If amenorrhoea was present, this was performed after excluding a dominant follicle with ultrasound. Blood was analyzed for free and total Testosterone (F.T. and T.T.) (F.T.; nmol/L), sex hormone-binding globulin (SHBG; normal values- 40 - 120 nmol/L), L.H. (units per L), FSH (units per L), fasting blood glucose (mg/dL), fasting insulin (microunits per mL) was detected by the glucose oxidase method (AU640 automation biochemistry analyzer and their relevant reagent, Olympus Company, Hamburg, Germany), sex hormone-binding globulin (SHBG), luteinizing hormone (L.H.), and follicle-stimulating hormone (FSH). LH/FSH ratio was calculated for each subject, LH/FSH ratio > two, and was considered abnormal. The L.H. and FSH were measured by electrochemiluminescence technique in COBAS autoanalyzer using reagents manufactured by Roche Diagnostics, and quality control was done by precise-control 1 and 2, then mean LH/FSH ratio was obtained. Free androgen index (FAI) was calculated by [total testosterone] X 100 divided by [SHBG]. Normal insulin sensitivity was defined by fasting insulin levels <12 mU/mL. We also calculated indices of the homeostasis model assessment (HOMA-IR)[14]. HOMA−IR was calculated using the equation: HOMA−IR=Fasting insulin (μU/mL) x Fasting glucose (mg/dL) /405 [15] with HOMA-IR > 2.5 mol μU/mL as cut off abnormal value [16].

 In high fasting glucose (Fasting Glucose (F.G.), 110–125 mg/dL), subjects were referred to a diabetes specialist. Additionally, serum PRL (nomograms per mL), TSH (microunits per mL), dehydroepiandrosterone sulfate ug/dL (DHEAS), and 17α-hydroxyprogesterone (17-αOHP; nomograms per mL) levels were examined in blood samples of all participants. Normal values of 20-100 ng/dl were accepted to exclude other causes of menstrual disorders [17]. AMH serum concentrations were examined with an enzyme-linked immunosorbent assay (ROCHE, Platform: Cobas 8000) according to the manufacturer's instructions; the respective intra and inter-assay coefficient of variations was <12%. A receiver operating characteristic curve analysis was used to determine the best cutoff point for AMH to distinguish PCOS women from controls.

Results

Baseline Characteristics

The average age and age menarche of our study participants did not differ between cases and controls (age-25.52±7.18, age at menarche-12.98±1.56) as the expected period frequency was significantly higher among PCOS cases (38.99±16.05) compared to the control participants (29.85±3.41), with p-value = 0.013. mFG score was significantly higher among PCOS cases (10.84±3.70) compared to the control participants (4.00±3.48), with p-value = 0.001. The ovarian volume also was higher among PCOS cases (12.63±3.07), while it was (4.55±3.17) among participant controls, and this difference in the mean ovarian volume (cm3) was statistically significant with p-value = 0.001. PCOM was significantly higher among PCOS cases (n=65 (76.5%)) compared to the control participants (n=2 (10.0%)), with p-value = 0.001 (Fisher's Exact test). Waist circumference was significantly higher among cases (84.46±11.42) than controls (78.50±12.88), with p-value = 0.043. However, BMI was not significantly different between cases and controls. Moreover, Waist circumference as a category (obese [≥88.0 cm] was significantly associated with cases (n=40 (47.1%)) rather than controls (n=4 (20.0%)), with p-value = 0.017.

All patients with Acanthosis Nigricans were PCOS cases (n- 46 (54.1%)) compared to no cases of Acanthosis Nigricans among controls with p-value = 0.001 (Fisher's Exact test). All grades of Alopecia score (Ludwig) were significantly higher among PCOS cases (p=0.007). Hyperandrogenism was significantly more prevalent among PCOS cases (n=76 (89.4%)) in comparison to the controls (n=8 (40.0%)), with p-value = 0.001 (Pearson Chi-Square).

Figure 1: Serum AMH mean values between PCOS and controls.

Hormonal profile; FSH, L.H., Prolactin, AMH, Total Testosterone, FAI, 17alpha HOP, DHAES, Fasting Insulin, and HOMA-IR were all significantly having higher means among PCOS cases than controls, with p-value <0.001 (t-test). Conversely, SHBG was substantially higher among controls rather than PCOS cases with a p-value <0.001. Women with PCOS had higher values of AMH in comparison with controls (Figure 1) AMH in PCOS (40.65±22.02 pmol/L), in controls (21.71±9.98 pmol/L) with p <0.001.

Our results also showed that AMH (pmol/L) among PCOS participants (n=85) shows a gradual and steady decrease in the AMH (pmol/L) around the age of 50 years old, while among the controls (n=20), there was a sharp decrease in the AMH (pmol/L) at the age of 35 years old (Figure 2A & 2B).

Figure 2: A- AMH and age correlation in the PCOS group, B- control.

AMH Values Cutoff as a Diagnostic Tool

To find a cutoff that will be good enough to diagnose PCOS, we used the receiver operating characteristic curve (ROC) of AMH pmol/L for the presence or absence of PCOS, and we constructed the curve with data generated from study subjects with AMH cutoff values of 23.97 pmol/L. The sensitivity and specificity of the serum values of AMH can help differentiate PCOS women from controls were evaluated using cutoff values according to the ROC curve. The best-combined values of 76.5% sensitivity and 85% specificity for the presence of PCOS was obtained using an AMH cutoff value of 23.97 pmol/L (Figure 3).

Figure 3: Receiver operating characteristic curve of AMH pmol/L and presence or absence of PCOS.

In figure 3 the Area under the curve was 0.844 (95%confidence interval, 0.756–0.932), p-value <0.001. An AMH of >23.97 pmol/L had 76.5% sensitivity and 85% specificity for the presence of PCOS (Table 1).

Table 1: Area under the curve of AMH values.

Area Under the Curve

Test Result Variable(s): AMH (pmol/L)

 

 

Area

Std. Error

Asymptotic Sig. b

Asymptotic 95% Confidence Interval

Lower Bound

Upper Bound

0.844

0.045

0.000

0.756

0.932

Correlation of AMH with Different Hormonal Parameters

There was no statistically significant correlation between HOMA-IR and AMH (pmol/L) among all participants, p-value = 0.737, and among PCOS participants, p-value >0.05. There was also no significant correlation between HOMA-IR and AMH (pmol/L) among different PCOS phenotypes, and among controls, p-value >0.05 (Table 2).

Table 2: HOMA-IR and AMH correlation.

All participants (n-105)

 

 

AMH (pmol/L)

HOMA-IR

 

 

 

 

r

0.033

p-value

0.737

N

105

PCOS participants (n-85)

 

 

AMH (pmol/L)

HOMA-IR

r

-0.186

p-value

0.088

N

85

PCOS phenotypes & controls

Phenotype

AMH (pmol/L)

A

HOMA-IR

r

-0.083

p-value

0.644

N

33

B

HOMA-IR

r

-0.403

p-value

0.051

N

24

C

HOMA-IR

r

-0.132

p-value

0.547

N

23

D

HOMA-IR

r

-0.475

p-value

0.419

N

5

No PCOS

HOMA-IR

r

0.000

p-value

0.999

N

20

There was no significant correlation between Hormonal Profile (L.H., FSH, SHBG) and AMH (pmol/L) among PCOS participants, p-value >0.05. However, the Free androgen Index was positively correlated with the AMH, p-value = 0.016 (Table 3)

Table 3: AMH correlation with age and hormonal parameters in PCOS participants (n-85).


 

 

AMH (pmol/L)

L.H. (units per L)

r

0.136

p-value

0.214

Free Testosterone (ng/dL)

r

-0.028

p-value

0.796

Total Testosterone (nmol/L)

r

0.148

p-value

0.178

SHBG (nmol/L)

r

-0.181

p-value

0.098

BMI

r

-0.105

p-value

0.341

WC (cm)

r

-0.178

p-value

0.104

FSH (units per L)

r

0.079

p-value

0.475

FBS (mmol/L)

r

-0.011

p-value

0.923

Fasting insulin (pmol/L)

r

-0.176

p-value

0.107

HOMA-IR

r

-0.186

p-value

0.088

Ovarian volume in cm3

r

-0.132

p-value

0.229

Free androgen Index

r

0.260*

p-value

0.016

DHAES/FAI ratio

r

-0.123

p-value

0.263

E2 (ng/ml)

r

 0.030

p-value

0.782

Age (years)

r

-0.150

p-value

0.172

*Correlation is significant at the 0.05 level (2-tailed).

Discussion

In this study, we found two folds increase in AMH levels for PCOS women than controls; this is in concordance with other reports [7,18]. Moreover, our results show that an AMH value cutoff of >23.97 pmol/L had 76.5% sensitivity and 85% specificity for the presence of PCOS. In a European study, a cutoff value of 8.40 ng/mL was suggested for PCOS diagnosis, with a sensitivity of 67% and a specificity of 92% [19]. Moreover, one study from Australia suggested that an AMH value of 35.8 pmol/L (sensitivity 84%, specificity 82%) was the best determinant of PCOS presence [20]. Different methodologies used in different studies may be the reason for such a variation of AMH cutoff values. We propose this cutoff as it has good sensitivity and relatively high specificity, as some reports have good sensitivity yet poor specificity, or vice versa, where others had a higher cutoff than our proposed one (Table 4).

In women with PCOS, we found that serum AMH levels did not correlate with any of the hormones measured and AMH among PCOS participants, p-value >0.05. However, the Free androgen Index was positively correlated with the AMH, p-value = 0.016, indicating a correlation between AMH and androgen secretion. Such interaction may play a role in the pathogenesis of PCOS, this in concordance with the previous report [25] and in contrast to the report where age- and BMI- stratified deep analysis showed that AMH was more predictive of PCOS independently of free androgen index (FAI) [26]. Moreover, there was no significant correlation between HOMA-IR and AMH among all participants, among PCOS participants, and among different PCOS phenotypes, in contrast to the finding of other studies where the median value of AMH was highest in phenotype A [3,25]. These findings may indicate that AMH is an independent factor in PCOS.

Table 4: Different AMH levels cutoff in different studies.

Study

Country

AMH Cutoff

SI units (conventional units)

Sensitivity

specificity

Study sample size

Al-Jefout et al. (our study)

UAE

23.97 pmol/L

(3.36 ng/ml)

76.5%

85%

120

Pigny, P. [19]

France

60 pmol/L

(8.4 mg/ml)

92%

67%

169

Mahajan, N. [21]

India

35.91 pmol/L

(5.03 ng/mL)

70.68%

79.91%.

367

Song, D. K.[22]

Korea

 71.4 pmol/L (10.0 ng/mL)

71%

 93%

207

Casadei, L.[23]

Italy

37.13 pmol/L (5.20 ng/ml)

79%,

80%

140

Sova, H. [24]

Finland

42.1 pmol/L

(5.894 ng/ml)

67%

83%

418

Tremellen, K. [20]

Australia

36 pmol L

(5.04 ng/ml)

83.7%

82.3%

654

Serum AMH level has been reported to decline with age [27]. Our study showed that AMH levels are double those of controls and remained elevated around the age of 50, this inconsistency with a European study, where serum AMH levels in women with PCOS were 2- to 3-fold higher with a plateau until the age of 40 [28]. The steady and high levels of AMH with age in women with PCOS may be due to inherited high levels of AMH [29].

Strengths and Limitations

The strengths of this study are the homogenous study population, which included only women of Arabic ancestry, and both our groups are well defined. Another strength of our study is that both groups are almost similar in age and weight, so that we could exclude the confounding effect of age and weight on AMH levels. The limitation of our study is the relatively small number of participants. Because of the lack of universally accepted methods and assays for measuring serum AMH level, we may have some discrepancies with other reports, and it is difficult to generalize the results of our study. Furthermore, the study population was not population-based but consisted of women who had visited the gynecology clinic. These women probably had different PCOS manifestations, moreover, given that the control group did not always include women with all PCOS symptoms eligible for diagnosis, which could lead to higher serum levels of AMH in the PCOS group versus the controls.

Conclusion

In conclusion, we found that serum AMH level cutoff of 23.9 has good sensitivity and high specificity and maybe a new and useful diagnostic tool in PCOS diagnosis. This cutoff level of AMH may useful to use for either the confirmation or as an additional tool for the diagnosis of PCOS when imaging is not possible to perform, as in adolescent population with intact hymen and obesity.

Authors' Roles

M.A.: Concept/design, Data analysis/interpretation, Drafting article, Critical revision of the article, Approval of article, data collection.

N.A.: Statistics/Data analysis/interpretation, Drafting article, Critical revision of the article, Approval of article.

M.K.: Interpretation, Drafting article, Critical revision of the article, Approval of article.

W.K.: Interpretation, data collection, Drafting article, Critical revision of the article, Approval of article.

Disclosure Statement

The authors have no conflict of interest to declare.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

  1. Stojkovic M, Beleslin B, Ciric J, Savic S, Nisic T, Stojanovic M, et al. Potential discriminant factors for different PCOS phenotypes. 2013.
  2. La Marca A, Broekmans FJ, Volpe A, Fauser BC. Anti-Mullerian hormone (AMH): what do we still need to know? Hum Reprod. 2009; 24: 2264-2275.
  3. Gupta M, Ritu Y, Mahey R, Agrawal A, Upadhyay A, Malhotra N, et al. Correlation of body mass index (BMI), anti-mullerian hormone (AMH), and insulin resistance among different polycystic ovary syndrome (PCOS) phenotypes - a cross-sectional study. Gynecol Endocrinol. 2019; 35: 970-973.
  4. Kedem A, Hourvitz A, Yung Y, Shalev L, Yerushalmi GM, Kanety H, et al. Anti-Mullerian hormone (AMH) downregulation in late antral stages is impaired in PCOS patients. A study in normo-ovulatory and PCOS patients undergoing in vitro maturation (IVM) treatments. Gynecol Endocrinol. 2013; 29: 651-656.
  5. Hampl R, Snajderova M, Mardesic T. Antimüllerian hormone (AMH) not only a marker for prediction of ovarian reserve. Physiologi Res. 2011; 60: 217-223.
  6. Visser JA, Themmen AP. Anti-Müllerian hormone and folliculogenesis. Molecular and cellular endocrinology. 2005; 234: 81-86.
  7. Pigny P, Merlen E, Robert Y, Cortet-Rudelli C, Decanter C, Jonard S, et al. Elevated serum level of anti-mullerian hormone in patients with polycystic ovary syndrome: relationship to the ovarian follicle excess and the follicular arrest. J Clinical Endocrinology Metabolism. 2003; 88: 5957-5962.
  8. La Marca A, Malmusi S, Giulini S, Tamaro LF, Orvieto R, Levratti P, et al. Anti-Müllerian hormone plasma levels in the spontaneous menstrual cycle and during treatment with FSH to induce ovulation. Human Reproduction. 2004. 19: 2738-2741.
  9. Nardo LG, Yates AP, Roberts SA, Pemberton, Laing I. The relationships between AMH, androgens, insulin resistance, and basal ovarian follicular status in non-obese subfertile women with and without polycystic ovary syndrome. Human Reproduction. 2009; 24: 2917-2923.
  10. Teede H, Misso M, Tassone EC, Dewailly D, Hy Ng E, Azziz R, et al. Anti-Müllerian hormone in PCOS: a review informing international guidelines. Trends in Endocrinology Metabolism. 2019; 30: 467-478.
  11. ESHRE TR, A.-S.P.C.W. Group, Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertility and sterility. 2004; 81: 19-25.
  12. Ferriman D, Gallwey J. Clinical assessment of body hair growth in women. J Clinical Endocrinology Metabolism. 1961; 21: 1440-1447.
  13. Ludwig E. Classification of the types of androgenetic alopecia (common baldness) occurring in the female sex. British J Dermatol. 1977; 97: 247-254.
  14. 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 Metabol. 1998; 83: 2694-2698.
  15. Wallace TM, Matthews DR. The assessment of insulin resistance in man. Diabet Med. 2002; 19: 527-534.
  16. Ascaso JF, Pardo S, Real JT, Lorente RI, Priego A, Carmena R. Diagnosing Insulin Resistance by Simple Quantitative Methods in Subjects With Normal Glucose Metabolism. Diabetes Care. 2003; 26: 3320-3325.
  17. Azziz R, Zacur HA. 21-Hydroxylase deficiency in female hyperandrogenism: screening and diagnosis. J Clin Endocrinol Metab. 1989; 69: 577-584.
  18. Mulders AG, Laven JS, Eijkemans MJ, de Jong FH, Themmen AP, Fauser BC. Changes in anti-Müllerian hormone serum concentrations over time suggest delayed ovarian aging in normogonadotrophic anovulatory infertility. Human Reproduction. 2004; 19: 2036-2042.
  19. Pigny P, Jonard S, Robert Y, Dewailly D. Serum anti-Mullerian hormone as a surrogate for antral follicle count for a definition of the polycystic ovary syndrome. J Clinical Endocrinol Metab. 2006; 91: 941-945.
  20. Tremellen K, Zander?Fox D. Serum anti?Mullerian hormone assessment of ovarian reserve and polycystic ovary syndrome status over the reproductive lifespan. Australian and New Zealand J Obstetrics Gynaecol. 2015; 55: 384-389.
  21. Mahajan N, Kaur J. Establishing an Anti-Mullerian Hormone Cutoff for Diagnosis of Polycystic Ovarian Syndrome in Women of Reproductive Age-Bearing Indian Ethnicity Using the Automated Anti-Mullerian Hormone Assay. J Hum Reprod Sci. 2019; 12: 104-113.
  22. Song DK, Jee-Young Oh, Lee H, Yeon-Ah S. Differentiation between polycystic ovary syndrome and polycystic ovarian morphology by means of an anti-Mullerian hormone cutoff value. Korean J Intern Med. 2017; 32: 690-698.
  23. Casadei L, Fanisio F, Sorge RP, Collamarini M, Piccolo E, Piccione E. The diagnosis of PCOS in young infertile women according to different diagnostic criteria: the role of serum anti-Mullerian hormone. Arch Gynecol Obstet. 2018; 298: 207-215.
  24. Sova H, Unkila-Kallio L, Tiitinen A, Hippeläinen M, Perheentupa A, Tinkanenet H, et al. Hormone profiling, including anti-Mullerian hormone (AMH), for the diagnosis of polycystic ovary syndrome (PCOS) and characterization of PCOS phenotypes. Gynecol Endocrinol. 2019; 35: 595-600.
  25. Skalba P, Anna C, Pawe? M, Anna D, Jerzy S, Gayane M, et al. Is the plasma anti-Mullerian hormone (AMH) level associated with body weight and metabolic and hormonal disturbances in women with and without polycystic ovary syndrome? Eur J Obstet Gynecol Reprod Biol. 2011; 158: 254-259.
  26. Sathyapalan T, Al?Qaissi A, Kilpatrick ES, Dargham SR, Atkin SL. Anti-Mullerian hormone measurement for the diagnosis of polycystic ovary syndrome. Clin Endocrinol (Oxf). 2018; 88: 258-262.
  27. Durlinger A, Visser J, Themmen J. Regulation of ovarian function: the role of the anti-Mullerian hormone. Reproduction, 2002.
  28. Piltonen T, Morin-Papunen L, Koivunen R, Perheentupa A, Ruokonen A, Tapanainen JS. Serum anti-Müllerian hormone levels remain high until late reproductive age and decrease during metformin therapy in women with polycystic ovary syndrome. Human reproduction. 2005; 20: 1820-1826.
  29. Do Kyeong Song JYO, Lee H, Sung YA. Differentiation between polycystic ovary syndrome and polycystic ovarian morphology by means of an anti-Müllerian hormone cutoff value. Korean J Internal Med. 2017; 32: 690.