Comparison of two different anti-mullerian hormone measurement methods and evaluation of anti-mullerian hormone in polycystic ovary syndrome

Introduction

Gynecological health issues include diseases of the female reproductive system or estrogen-controlled organs, such as polycystic ovarian syndrome (PCOS), an endocrine status that influences 15–20 % of women of fertile age [1], [2], [3]. PCOS, a heterogeneous condition, is a multifactorial, polygenic, inflammatory, dysregulated steroid state that can lead to obesity, insulin resistance, cardiovascular problems, and infertility [2], 4]. Endocrine, reproductive, metabolic, and psychosocial symptoms in women with PCOS might fluctuate throughout a lifetime [5]. According to the 2003 Rotterdam consensus, at least two of the following criteria must present for PCOS diagnosis: oligomenorrhea (OA), hyperandrogenism (HA), and polycystic ovary morphology on ultrasound [6].

Anti-mullerian hormone (AMH), also defined as the mullerian-inhibiting substance, belongs to the transforming growth factor-beta family. AMH contains homodimer disulfide-linked glycoprotein [7]. Male testicular Sertoli cells and female ovarian granulosa cells both produce it. One of PCOS’ characteristic features is the increment of ovarian follicles at their growing stage. The increment can be seen in both pre-antral follicles and small antral follicles that produce AMH, which is crucial for regulating folliculogenesis [8], 9]. AMH prevents primordial follicle reinforcement and modulates follicle-stimulating hormone (FSH) activity in early follicular development [10], 11].

AMH measurement is useful in determining gender, gonadal function, and fertility, as well as a gonadal tumor marker, due to the variations in AMH levels between genders, differences in circulating levels together with sexual feature development, and AMH’s specificity for Sertoli cells and granulosa cells. In women, AMH is known as an ovarian reserve marker, and there is a relationship between AMH and the primordial follicles and an inverse relationship with the chronological age, predicting the ovarian response to in vitro fertilization. It may even be able to predict when menopause will start. Serum AMH concentrations are generally steady during the menstrual cycle, unlike other ovarian reserve indicators that exhibit considerable changes. Serum AMH concentrations, which are correlated with HA and ovarian antral follicle count, are 2–4 times higher in women with PCOS due to increased stock of pre-antral follicles and small antral follicles of polycystic ovaries, and increased levels of AMH appear to be related to predicting anovulatory and irregular cycles and the severity of PCOS [12], [13], [14]. Some researchers have suggested that measuring serum AMH levels may have a role in the diagnosis of PCOS [15], 16]. Since there is no standardization for AMH immunoassays, values obtained using various assay techniques or kits may vary and cannot be used interchangeably. AMH levels and cut-off values depend on the method and, there is no international standard for AMH measurement [17], 18].

The aims of the present study are as follows: firstly, to measure AMH levels in two different devices that measure AMH levels by immunoassay method, secondly, to detect the measurement difference in these measurements, and thirdly, to evaluate the clinical meaning of this difference in PCOS diagnosed patient group and the control group. For the present research, the measurements of serum AMH and evaluation of the results for PCOS diagnosis were performed using two different AMH measurement methods: the Vidas® (bioMérieux SA, Marcy l’Etoile, France) and the Atellica® (Siemens Healthineers, Erlangen, Germany) methods.

To our knowledge, no research has been conducted to compare these AMH techniques. We also looked at the hormonal characteristics of women with PCOS and controls.

Materials and methods

Results

The PCOS group’s age (21.3 ± 1.89 years) was less than the control group (31.7 ± 7.83 years) (p<0.001). Serum prolactin concentrations were statistically higher in the control group. (p=0.002) In contrast, serum E2 levels, serum AMH levels measured by Vidas®, and Atellica® were statistically higher in the PCOS group (p=0.001, <0.001, p=0.001, respectively). Statistically, no difference was observed in serum FSH, LH, TSH, fT3, fT4, and progesterone levels between the groups. Hormonal and metabolic parameters of the groups are presented in Table 1.

The positive prediction values for Vidas® and Atellica® methods were 0.66 and 0.65, respectively. The negative prediction values for Vidas® and Atellica® methods were 0.34 and 0.35, respectively. The positive and negative likelihood ratios of the Vidas® and Atellica® methods were 1, 1.03 and 1, 0.95, respectively.

Serum AMH concentrations in PCOS and control groups

Serum AMH concentrations of PCOS group were significantly higher than the control group measured by Vidas® (4.48 ± 2.36 μg/L vs. 2.78 ± 2.20 μg/L, p<0.001) and by Atellica® (4.88 ± 2.65 μg/L vs. 3.17 ± 2.61 μg/L, p=0.001).

According to the ROC curve analysis; we got 3.23 μg/L as the best cut-off AMH level with a sensitivity of 65.5 % and, a specificity of 34.5 % by using the Vidas® method and we got 3.66 μg/L as the best cut off AMH level with sensitivity 65.4 % and specificity 36.4 % by using Atellica® method. The area under the curve (AUC) and 95 % confidence interval (CI) of the Vidas® and the Atellica® method were 0.702 (0.605–0.799) and 0.689 (0.59–0.789), respectively (Figure 1). Passing Bablock regression analysis was presented in Figure 2. Bland Altman plot presents the scatter of difference values ​​against the mean of the measurements of Atellica® and Vidas® methods. Bias distribution was given in Bland-Altman plots (Figure 3).

Figure 1:

Receiver operating characteristic-curve shows the best cut-off values of serum AMH levels between the PCOS and the control groups according to the Vidas® (3.23 μg/L as the best cut-off AMH level with a sensitivity of 65.5 % and a specificity of 34.5 %) and the Atellica® (3.23 μg/L as the best cut-off AMH level with a sensitivity of 65.5 % and a specificity of 34.5 %).

Figure 2:

Passing-Bablok regression analysis (y=0.0413 + 1.073x, n=110) (The 95 % CI for the intercept: −0.049, 0.799) (The 95 % CI for the slope: 0.908–1.102) (The area between dashed lines: the confidence interval) (r: 0.891).

Figure 3:

Bland-Altman plot graph.

Discussion

In the present study, a regression analysis was performed to compare AMH levels among Atellica® and the Vidas® devices, and a fairly strong positive relationship was observed (r=0.891) (Figure 2). The scatter of the difference values against the mean of the measurements of Atellica® and Vidas® methods was presented with the Bland Altman plot (Figure 3). Higher serum AMH concentrations were observed in PCOS women, and serum AMH cut-off values were presented with both the Atellica® and the Vidas® devices.

Additionally, we performed correlation analyses for variables with both Atellica® and Vidas® methods. We report that, serum AMH levels of women with PCOS, according to the Atellica® method, were positively correlated with both serum fasting glucose and LH levels and negatively correlated with serum progesterone levels and age. We also report that serum AMH levels of women with PCOS, according to the Vidas® method, were negatively correlated with serum progesterone levels and age. A statistically significant difference was observed in serum E2 and serum prolactin levels between the groups, but there was no statistically significant difference in serum FSH, TSH, fT3, and fT4 levels between the groups.

In this study, the mean age of the groups was not similar, but all the participants were of reproductive age. In the study conducted by La Marka et al., it was shown that AMH values decrease with increasing age and fall to undetectable levels after spontaneous menopause [20]. De Vet et al. reported that serum AMH levels in young normo-ovulatory women decreased with time [21]. In our study, there was a negative correlation between age and serum AMH levels, consistent with previous studies [20], 21].

In this study, we obtained results using the Atellica® and the Vidas® methods, and both represent higher mean serum AMH levels in women with PCOS compared to controls, in agreement with former studies. As mentioned in previous studies; this situation shows that serum AMH concentrations may be useful [22], [23], [24]. However, AMH was not included in the 2003 Rotterdam consensus for the diagnosis of PCOS.

In this study, we state two different best cut-off level values for serum AMH levels, but still, AMH alone is not sufficient diagnostic value to distinguish PCOS patients from controls. The Vidas® method presents 3.23 μg/L (sensitivity 65.5 % and specificity 34.5 %) and the Atellica® method presents 3.66 μg/L (sensitivity 65.4 % and specificity 36.4 %) and our results are nearly similar to the study by Sahmay et al. [24].

However, in some studies, cut-off values of serum AMH concentrations are higher to distinguish women with PCOS from controls [12], 25], 26]. In this academic work, mean serum AMH concentrations show a strong correlation between the Vidas® and the Atellica® methods (r=0.899).

In the study conducted by Halder et al., the AUC for total PCOS cases was 0.93 (95 % CI: 0.88–0.96) [27]. But in the present study, AUC of the Vidas® method was 0.702 (95 % CI: 0.605–0.799) and the AUC of the Atellica® method was 0.689 (95 % CI: 0.59–0.789).

AMH has important roles in gonadal functions and sexual differentiation. AMH has a potent effect on GnRH neuron firing and can increase GnRH-dependent LH pulse and secretion. LH pulsatility increment role is crucial for the PCOS’s pathophysiological process [28]. In this study, there was no statistically significant difference between the groups, but a positive correlation was observed between serum AMH and LH levels of women with PCOS, consistent with previous studies [29], 30].

PCOS is a common endocrine ailment, including progesterone secretion disorder [31]. Women with PCOS have low progesterone levels; because of oligo/anovulation [32]. In this academic work, serum AMH levels of women with PCOS were negatively correlated with serum progesterone levels, in line with previous studies [31], 32].

Zarei et al. stated higher E2 levels in women with PCOS than in controls. Our study was in line with the study by Zarei et al., having higher serum E2 levels in PCOS group and showing a significant difference between the groups [33]. In this study, serum prolactin levels showed statistical differences and were lower in women with PCOS according to the literature [34].

Women with PCOS are known to have elevated insulin and HOMA-IR levels and are expected to have an increased risk of type-2 diabetes mellitus [17]. Approximately, 40 % of PCOS patients have prediabetes [35]. Li et al. stated higher serum fasting glucose levels in women with PCOS compared with controls [36].

In our study, the control group’s serum fasting glucose levels, insulin levels, and HOMA-IR levels could not be obtained, but a positive correlation was observed between serum AMH and fasting glucose levels in women with PCOS. Zhao et al. stated a weak correlation between serum AMH and plasma fasting glucose levels in the classical PCOS phenotype [3]. Our result was in line with their study but not with some other studies [37], 38]. Further studies including more participants of similar ages with different serum AMH measurement methods may be useful to investigate whether serum AMH levels can determine the PCOS patients from healthy controls and to evaluate its relationship with hormonal parameters of PCOS patients.

A limitation of the this study is the inclusion of individuals of the same ethnicity and, given the study limitations, there were not many participants in any group. Another limitation of the present study is that serum fasting glucose, total testosterone, insulin, Homa-IR, DHEA-S, SHBG, and androstenedione levels were only available for women with PCOS. There was a discrepancy in age between the PCOS and control groups.

We stated the cut-off points for each method and performed correlation analyses with both the Vidas® and the Atellica® methods. Future investigations on this topic may find great value in all of these data points. To our knowledge, this is the first study to compare serum AMH levels using the Atelllica® and Vidas® methods and evaluate serum AMH levels in PCOS and control groups with both devices.

Conclusions

In conclusion, serum AMH levels analyzed with the Atellica® showed good correlation with serum AMH levels analyzed with the Vidas®. Our study showed that serum AMH levels measured with both the Atellica® and the Vidas® present significant differences between PCOS patients and the controls but still AMH alone was not diagnostic enough to distinguish PCOS patients from controls.

Furthermore, AMH were positively correlated with fasting glucose and LH level and negatively correlated with age and serum progesterone level. Further research with a larger sample of similar ages would be beneficial.