Analysis of discordance between AMH and AFC in IVF patients: a retrospective cohort study

Both anti-Mullerian hormone (AMH) and antral follicle count (AFC) predict ovarian reserve independently and have been shown to correlate well. However, it is quite common to encounter discordance between the two clinical indices. This objective of this study is to speculate which condition is favorable to patients and to decide what best to be done in cases of such discordance. This retrospective cohort study analyzed the medical records of 714 women undergoing the rst IVF cycle at the between January 2017 and December 2020. All patients had their essential characteristics, baseline FSH, Estradiol, PRL, LH, Testosterone, CA125, AMH, and AFC recorded. Patients were classied into three groups according to whether their AMH and AFC values were in the lower quartile, upper quartile, or in between, namely Group A (Concordance, concordant in AMH and AFC), Group B (HTP, AMH concentration higher than predicted according to AFC) and Group C (LTP, AMH concentration lower than predicted according to AFC). SPSS 25.0 software package (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis.

AMH, also known as Mullerian-inhibiting substance, is a member of the transforming growth factor β family, which is secreted by granulosa from small, growing follicles (6)(7)(8)(9). AMH exerts its biological effects on the ovaries generally by inhibiting the initial follicle recruitment and the FSH-dependent growth of the follicles and selection of pre-antral and small antral follicles through a transmembrane serine/threonine kinase type II receptor (AMHR II) (10)(11)(12). The absence of AMH has been shown to enhance FSH-induced follicle growth in female mice (10,13). Currently, AMH has been widely used as a golden marker for evaluating ovarian reserve in females, particularly in the eld of assisted reproduction (1,(14)(15)(16).
Antral follicle count (AFC), which consists of all ultrasonographically identi ed and counted antral follicles measuring 2-10 mm in diameter in both ovaries, is another reliable indicator of ovarian reserve owing to its close relationship to age-related declines in follicle counts (17)(18)(19)(20)(21). Although AFC may be more machine and operator-dependent and could have higher inter-operator variability compared to AMH, it correlates more directly with the ovarian reserve of a single ovary than AMH (1,3,5,19,22). Systematic reviews showed that AMH and AFC had a very good correlation with each other. In 2003, Fanchin et al. demonstrated that AFC was closely related to serum AMH level on cycle day 3 in infertile women, more strongly than other hormonal markers, such as inhibin B, E2, and FSH (23). A strong positive correlation between AMH and AFC and paralleled uctuations of AMH and AFC have also been demonstrated throughout the regular menstrual cycle (24)(25)(26). Yet, it is not unusual for IVF specialists to encounter a serum AMH level that is higher or lower than expected according to AFC, even when the measurement of AMH and counting of antral follicles are performed during the early follicular phase of the same menstrual cycle in a single clinical center (18,27).
In this article, we conducted a retrospective analysis to evaluate the discordance between AMH and AFC in women undergoing their rst IVF treatment, the reasons hiding in this phenomenon, and maybe the clinical implications.

Patient sampling
The medical records of 714 women at the Center for Reproductive Medicine, A liated Hospital of Nantong University between January 2017 and December 2020 were evaluated in the analysis. Only the rst cycle was included for each patient to eliminate previous failure bias.
The exclusion criteria were as follows: patients with a history of ovarian or pelvic surgery, endometriosis, chemotherapy or radiation exposure, hormonal therapy in the past 6 months, fragile X premutation, and abnormal karyotype or other endocrine diseases, including diabetes, thyroid dysfunction, and hyperprolactinemia.
During the study period, the type of protocol was determined by the doctor according to speci c characteristics, such as ovarian reserve and follicle size.
All patients who had their essential characteristics, baseline FSH, Estradiol, PRL, LH, Testosterone, CA125, AMH, and AFC recorded during their evaluation were included in the study. The Ethics Committee of A liated Hospital of Nantong University has authorized us to use existing patient data anonymously, waiving the requirement for written consent from individual patients. The main clinical and pathologic variables of the study patients are shown in Table 2.

Classi cation of patients
In the precycle analysis, the 25th percentile and 75th percentile values of serum AMH concentration were 2.05 ng/mL and 5.58 ng/mL, whereas the values of AFC were 12 and 31, respectively. Patients were classi ed into three groups according to whether their AMH and AFC values were in the lower quartile, upper quartile, or in between (Table 1), namely Group A (Concordant, concordance in AMH and AFC), Group B (HTP, AMH concentration higher than predicted according to AFC) and Group C (LTP, AMH concentration lower than predicted according to AFC).

Clinical investigation
Ultrasonographic examinations were performed by reproductive medicine specialists, who had at least 5 years of experience in gynecological ultrasound and have undergone uniform training. During the early follicular phase of the menstrual cycle, follicles measuring 2-10 mm in diameter were counted. The number of follicles in each ovary was assessed by using the same ultrasound equipment with a 2D transvaginal probe 6.5 MHz (Toshiba, Nemio, Japan). The total number of follicles in both ovaries was used as the AFC.

Laboratory analyses
Serum FSH, Estradiol, PRL, LH, and Testosterone were measured on day 2 or 3 of the menstrual cycle with the use of Beckman Coulter Dxl 800 (Beckman Coulter, Inc., Brea, USA). Random serum CA125 and AMH levels, unrelated to cycle day, were measured by using enzyme-linked immunosorbent assay (ELISA) on ARCHITECH I2000SI (Abbott, Chicago, IL, USA) and iFlash 3000-H (YHLO Biotechnology Co., Ltd., Shenzhen, CHINA), respectively.
The functional sensitivity, the coe cients of variation for intra-assay precision (repeatability) and the Statistical analysis SPSS 25.0 software package (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis. Patient characteristics were compared between the groups by using Mann-Whitney test. Continuous variables were compared using Student's t-test when data were normally distributed and by Mann-Whitney U when the data were not normally distributed. Chi-squared tests were used to compare proportions. All normal distribution measurement data are expressed as the mean ± standard deviation (SD). The level of signi cance was set at P < 0.05.

Results
In this study, 714 women, who underwent the rst IVF treatment cycle, ful lled the inclusion criteria, were included. The clinical and demographic parameters of the included subjects were listed in Table 2.
The correlation between AMH and AFC was visualized in Figure 1. A signi cant positive association of AMH and AFC was depicted in the entire study population (regression equation: AMH = 0.4167 × AFC; R 2 = 0.462, P < 0.001).
To better elucidate the relationship between AMH and AFC, the two ovarian reserve markers were strati ed into the following categories. The interquartile range (IQR) value of serum AMH concentration was 2.05-5.58 ng/mL, whereas which of AFC was 12-31. Patients were grouped according to whether their AMH and AFC values were in the lower quartile, upper quartile, or in between (Table 1). The clinical and biochemical characteristics of subjects after grouping were presented in Table 2. As seen in Table 2, the three cohorts had similar age, infertile duration, FSH, Estradiol, PRL, LH, Testosterone, and CA125 levels. To get a better understanding of BMI-related AMH and AFC discordance, women were subdivided into the following groups on account of the above BMI classi cation criteria (Table 3, P=0.009). The upper numbers were the observed values, and the numbers in the brackets were the expected ones. As shown in Table 3, the observed value in Group A (concordant) was notably higher than the expected one when BMI≤24.9 kg/m 2 (BMI<18.5 kg/m 2 and BMI 18.5-24.9 kg/m 2 ), which means in this BMI region, concordant in AMH and AFC was more common. Concurrently, the corresponding situation in Group B (HTP) generally occurred when BMI<18.5 kg/m 2 and BMI 25.0-29.9 kg/m 2 , while in Group C (LTP), widely appeared when BMI≥25.0 kg/m 2 (BMI 25.0-29.9 kg/m 2 and BMI 30.0-39.9 kg/m 2 ).

Discussion
It is important for patients undergoing assisted reproduction treatment to evaluate ovarian reserve precisely. Both AMH and AFC predict ovarian reserve independently and have been shown to correlate well, although AFC may have a higher inter-operator variability compared to AMH. However, it is quite common for IVF specialists to encounter serum AMH level that is higher or lower than expected according to AFC. It impels us to speculate which condition is favorable to patients and to decide what best to be done in cases of such discordance.
Our initial results showed a strong correlation between AMH and AFC, and then, we classi ed the two measurements into three intervals based on their respective lower and upper quartiles. The cut-offs for AFC were 12 and 31, and for AMH were 2.05 and 5.58 ng/mL, respectively. There was a high level of discordance (33.61% of cases) between AMH and AFC, and the incidence of HTP and LTP seemed similar ( There have been widely differing versions of the reasons for the discordance between AMH and AFC. The AFC refers to the number of follicles with a diameter of 2-10 mm in both ovaries and is a direct marker of the recruitable follicular cohort (28). AMH is a dimeric glycoprotein produced by granulosa cells of preantral and small antral follicles, and it indirectly re ects the population of early growing follicles. These pre-antral and small antral follicles are not able to completely develop into maturity leading to oocyte being yielded(3). Thus, the AMH level is not exactly re ecting the population of antral follicles counted in our AFC measurement. Moreover, the intra-and inter-menstrual cycle AMH variations may contribute to discordance in some cases(29).
A negative correlation between AMH and BMI has been previously reported in late reproductive-age women(30), young women using oral contraceptive pills(31), and infertile women with (diminished ovarian reserve) DOR but not in women with normal ovarian reserve (NOR)(32). In our analysis, we have found a signi cant role of BMI in the discordance between AMH and AFC. Group A(Concordant) had a moderate BMI and a medium ratio of completely mature oocytes (COM). Group B(HTP) had the lowest BMI and meanwhile, the worst ratio of completely mature oocytes (COM) and Group C(LTP) was its opposite. But we could not jump to a conclusion that a patient with AMH concentration higher or lower than predicted according to AFC will have a worse or better ratio of completely mature oocytes owing to individual BMI.

Conclusions
In conclusion, approximately 33.61% of patients undergoing the rst IVF therapeutic regime have discordance between AMH and AFC. Patients having a concordance in AMH and AFC have an intermediate ratio of completely mature oocytes and a moderate BMI. Those whose AMH concentration lower than predicted according to AFC have an obviously higher ratio, and their BMI are also above the average. The BMI is not an independent element that would in uence the complete maturity of oocytes. This study is limited by its retrospective nature and the small number of patients, making it di cult to generalize results. Such theoretical consideration, however, is needed to be elucidated by future clinical and basic studies.

Declarations
Ethics approval and consent to participate This study has been approved by the Ethics Committee of A liated Hospital of Nantong University.

Consent for publication
Not applicable.

Availability of data and materials
Datasets during the current study are available from the corresponding author upon reasonable request.  (7):2364-8. Figure 1 The relationship between serum anti-Müllerian hormone (AMH) concentration and antral follicle count (AFC) in the entire study population.