Threshold value of anti-Mullerian hormone for the diagnosis of polycystic ovary syndrome in Chinese women.

Polycystic ovary syndrome (PCOS) is a common reproductive endocrine disease that is prevalent in 6% of women of childbearing age [ 1 ]. Polycystic ovary syndrome (PCOS) is a common endocrine disease that is clinically manifested as menstrual thinning, hemorrhoids, hairiness, obesity and infertility. It is also characterized by abnormal levels of reproductive hormones, which can lead to anovulatory, infertile and menstrual disorders [ 2 ]. Anti-Müllerian hormone (AMH) is secreted by the granulosa cells of the pre-antral and antral follicles. AMH has drawn much attention in the field of female reproduction as a hormone that has important regulatory functions during follicular maturation. The level of AMH can reflect the number of ovarian antrum follicles, ovarian reserve, and ovarian function [ 3 ]. AMH levels in the blood are reliable biomarkers of actively growing primary to non-dominant antral follicles. In female primates, and most probably in women too, AMH does not regulate primordial follicle growth. It stimulates primary and secondary follicle growth, in contrast to its inhibitory role in female rodents. AMH, however, does inhibit FSH action on antral follicle growth, as it does in rodents [ 4 ]. Further, AMH is involved in follicular dysfunction and therefore reflects the degree of disturbance in follicular development [ 5 – 6 ]. Serum AMH is not affected by menstrual cycle and oral contraceptive use, so it has potential as a marker for the diagnosis of PCOS [ 7 ]. Studies have shown that the level of circulating AMH is two- to three-fold higher in women with PCOS than in healthy women of childbearing age, probably due to increased follicular mass associated with PCOS [ 8 – 9 ]. However, women without PCOS who have high AMH levels are commonly very fertile. Although increase in AMH levels has been reported to be associated with PCOS, there is little consensus on the appropriate reference range. Our aim was to explore the diagnostic value of AMH and other hormones or glucose metabolism markers for detecting PCOS in Chinese women. Our hypothesis is that AMH can be used as a potential predictor of PCOS, and that it can be used in combination with other indicators to improve detection.

This is a retrospective study that was conducted on women who consulted our endocrinology clinic at the Obstetrics and Gynecology Hospital of Fudan University between January 2016 and October 2016. Of the 3,400 women who consulted, we selected 653 women diagnosed with PCOS from among the outpatients for our retrospective analysis. In addition, 118 other participants were selected as healthy controls after exclusion of other endocrine and gynecological diseases. PCOS was diagnosed in the 653 participants based on the 2003 Rotterdam ESHRE/ASRM criteria: oligo and/or anovulation (OA); clinical and/or biochemical signs of hyperandrogenism (HA); and presence of polycystic ovaries (PCOs), defined as the presence of ≥12 follicles measuring 2–9 mm in diameter in each ovary and/or increased ovarian volume (>10 ml) [ 6 ]. OA was defined as a cycle length of more than 35 days or amenorrhea. Biochemical HA was defined as circulating total testosterone levels above the 95th percentile (0.51 ng/mL). Clinical HA was defined as the presence of hirsutism, acne, and/or alopecia. Hirsutism was defined as the presence of hirsutism, as evidenced by an m-FG score ≥ 8 [ 10 ]. The exclusion criteria were as follows: Cushing’s syndrome, dysfunctional uterine bleeding, primary amenorrhea, hypothalamic amenorrhea, pituitary amenorrhea, uterine amenorrhea, hyperprolactinemia, premature ovarian failure, ovarian functional tumors, theca cell proliferation, adrenal cortical hyperplasia or tumor, thyroid dysfunction, auto-immune disease, malignancy, central nervous system disease, current or previous use of oral contraceptives within 6 months of enrolment or the use of medications affecting the hypothalamic-pituitary-ovarian axis (e.g., anti-androgens, ovulation induction agents, antidiabetic medications, anti-obesity medications or glucocorticoids), age <20 or ≥40 years, and absence of complete records of hormone testing or ultrasonography. None of the participants were taking any form of steroid drugs before their samples were collected. The present study was conducted in accordance with the tenets of the Declaration of Helsinki and received the approval of the Research Ethics Committee of the Obstetrics & Gynecology Hospital of Fudan University (26 Feb 2016, Kyy2016-22). All data were fully anonymized before they were accessed, and the ethics committee waived the requirement for informed consent. Peripheral blood samples (5 ml) were obtained on days 2–5 of the menstrual cycle. The sample was collected in a serum separator tube and allowed to clot for 30 min before centrifugation at 1000 × g for 15 min. All peripheral blood samples were processed within 2 h of collection. Serum AMH was detected with the UNION immune analyzer’s AMH detection kit (single test strip) (YHLO, ShenZhen, China). Serum follicle-stimulating hormone (FSH), luteinizing hormone (LH), progesterone (P), prolactin (PRL), estradiol (E2), testosterone (T) and dehydroepiandrosterone sulfate (DHEA-S) were measured with Beckman Coulter DxI800. Sex hormone-binding globulin (SHBG) was measured on the Roche Cobas E601 automatic immunoassay apparatus. 17-α-OH progesterone (17α-OHP) was detected with the 17α-OH Progesterone ELISA Kit (DRG International, Germany). Fasting plasma glucose (FPG) was measured with the Hitachi 7180 automatic biochemical analyzer. Fasting insulin (INS) was measured with the Architect i2000 chemiluminescence immunoassay analyzer. Both intra-assay and inter-assay coefficients of variation were less than 10% for all the assays. Glucose tolerance was evaluated using the criteria of the American Diabetes Association [ 11 ]. Based on fasting serum glucose and insulin, the homeostasis model assessment for insulin resistance (HOMA-IR) index was calculated with a standard formula: HOMA-IR = fasting INS (mIU/ml) × fasting glucose (mmol/l)/22.5. The free androgen index (FAI) was calculated using the following formula: total testosterone/SHBG ×100 [ 12 ]. Ultrasonography was performed for all of the participants on the same day as blood sample collection. A transvaginal or transrectal ultrasound examination was performed by senior doctors with a 4–9 MHz probe using Voluson 730 Pro (General Electric, Milwaukee, WI, USA). Measurements were obtained in real time at the early follicular phase for determination of ovarian morphology. Each ovary was scanned in both longitudinal and transverse cross-sections from the inner to the outer margins to count the total number of follicles. The volume of each ovary was determined, and all follicles with a diameter of 2–9 mm were counted in each ovary. Ovarian volume was calculated for each ovary using the formula for a prolate ellipsoid: π/6 × (D1 × D2 × D3), where D represents the maximum diameter along the transverse, anteroposterior, and longitudinal axes [ 13 ]. For data with normal distribution and homogeneity of variance, an independent-sample t test was used to compare differences between two groups, and one-way ANOVA with Tukey’s post-hoc test was performed if there were three or more means. For non-normally distributed data, differences between groups were evaluated with the non-parametric Mann-Whitney U -test. In order to evaluate the diagnostic effectiveness of serum AMH and metabolic parameters in PCOS patients, we used receiver operating characteristic (ROC) curves and logistic regression analysis of AMH, LH/FSH, T and INS. Two different approaches were used to determine the optimal cutoff value of AMH to classify PCOS: Youden index, which is the maximum sensitivity-(1-specificity), and the shortest distance on the ROC between the optimal sensitivity and specificity ([1 − sensitivity] 2 + [1 − specificity] 2 ) [ 14 ]. Sensitivity, specificity, positive likelihood ratio (PLR: sensitivity/[1 − specificity]), and negative likelihood ratio (NLR: [1 − sensitivity]/specificity) of the cutoffs were also calculated [ 15 ]. All statistical analyses were performed by means of GraphPad Prism version 5.0 for Windows (GraphPad Software, USA), and p < 0.05 was considered to indicate significant differences.

The 653 patients were divided into four groups: OA+PCO (Type 1), OA+HA (Type 2), HA+PCO (Type 3), OA+HA+PCO (Type 4). Type 4 (characterized by all the Rotterdam criteria) is the most common type, while type 1 is the least common type. No statistical difference in age and BMI was observed between the PCOS groups and the control group (p > 0.05) ( Table 1 ). The hormonal and metabolic characteristics of the study patients and controls are shown in Tables 2 – 5 . FSH was significantly lower in the type 4 patients than in the controls (p < 0.05), but no significant differences were observed between the other types and the controls. The women with PCOS (all four types) had significantly higher LH, LH/FSH, AMH, FAI%, 17α-OHP, and fasting INS than women in the control group (p < 0.05) (Tables 2 – 5 ). Polycystic ovary patients aged 20–40 years are mainly concentrated in the 20–29 years of age, accounting for 73%. The AMH level was significantly higher in PCOS patients in the age groups 20–29 and 30–39 years than in the normal control group ( Table 3 ). Serum T, SHBG, DHEA-S and HOMA-IR levels were significantly higher in the PCOS patients than in the control group (Tables 4 and 5 ). No significant differences in PRL, P, E2, and fasting glucose were observed between the groups (Tables 2 – 5 ). ** P < 0.01; *** P < 0.001 *** P < 0.001 * P < 0.05; *** P < 0.001 * P < 0.05; ** P < 0.01; *** P < 0.001 For the ROC curves for AMH as a predictor of PCOS, we grouped the subjects into two age groups: 20–29 years and 30–39 years. The area under the curve (AUC) was 0.846 (95% confidence interval: 0.815–0.864) for AMH as a predictor of PCOS in the 20–29 years age group. The cut-off level of AMH was 8.16 ng/ml, which had a sensitivity of 78.4% and specificity of 80.9%. In the age group 30–39 years, the AUC was 0.865 (95% confidence interval: 0.791–0.887), and the cut-off level value of AMH was 5.89 ng/ml (sensitivity, 82.6%; specificity, 79.8%). For the ROC curves of the other metabolic parameters, the AUC, 95% confidence intervals, cut-off values, sensitivity, specificity, PLR, and NLR are summarized in Table 6 . With AMH, LH/FSH, T and INS as independent variables, a logistic regression model was established, in which the ROC curve for all the variables together was fitted with the probability value of the model. The area under the ROC curve was 0.951 (95% confidence interval: 0.922–0.954). The area under ROC curve for all the variables combined was significantly greater than that for each of the variables separately. The concentration of AMH was found to be significantly associated with FSH, LH, LH/FSH, T, and ovarian volume changes in PCOS patients. However, the AMH concentration was not significantly correlated with FAI%, INS or HOMA-IR ( Table 7 ).

Our results show that AMH may have potential as a marker of PCOS. The cut-off level of AMH for diagnosing PCOS was 8.16 ng/ml in the 20–29 years age group and 5.89 ng/ml in the 30–39 years group in the current study population from China. The results also indicate that serum AMH, LH/FSH, T and INS can be used for the diagnosis of PCOS; in particular, combined detection with these markers could improve the diagnostic specificity and sensitivity of PCOS.

(XLSX) Click here for additional data file. (XLSX) Click here for additional data file.