Very Low Levels of Serum Anti-Müllerian Hormone as a Possible Marker for Follicle Growth in Patients with Primary Ovarian Insufficiency Under Hormone Replacement Therapy

Patients with primary ovarian insufficiency (POI) occasionally present with follicle growth; however, accurately predicting cycles accompanied by follicle growth is challenging. Early-stage follicles produce serum anti-Müllerian hormone (AMH), a useful marker of ovarian reserve. Therefore, serum AMH levels indicate growth of small follicles (which are difficult to detect ultrasonographically) and may predict follicle growth in patients with POI. Using an ultrasensitive enzyme-linked immunosorbent assay (ELISA) kit, we observed very low serum AMH levels in patients with POI. We further evaluated follicle growth in each patient during each cycle to determine the usefulness of measuring serum AMH levels as a predictor of follicle growth in patients with POI who receive hormone replacement therapy (HRT). We investigated 19 patients with POI in whom we analyzed 91 cycles; 14 cycles showed positive and 77 cycles showed negative results on serum AMH testing. The rate of cycles showing follicle growth in AMH-positive cycles was higher than that in AMH-negative cycles (64.3% vs. 6.5%, p = 0.0001). The median serum AMH level (7.7 pg/mL [25th and 75th percentiles 4.6 pg/mL and 22.3 pg/mL, respectively]) in AMH-positive cycles was lower than the lower limit of detection of conventional AMH ELISA kits. The positive predictive value of positive serum AMH levels for follicle growth was higher than that of follicle-stimulating hormone (< 10 mIU/mL). These results indicate that a very low level of serum AMH detected using picoAMH assays is a useful predictor of follicle growth in patients with POI receiving HRT.


Introduction
Primary ovarian insufficiency (POI) is defined as the development of hypergonadotropic hypogonadism before the age of 40 years [1,2]. POI is one of the causes of intractable infertility. Due to the current trend of women delaying having their first child until later in life, the number of patients with POI who desire children is increasing. Some patients with POI have follicle growth occasionally [3], although it is difficult to predict which cycles will show follicle growth. Currently, serum follicle-stimulating hormone (FSH) or estradiol (E2) levels are measured in the clinical setting to help predict follicle growth in patients with POI. Serum E2 levels were reported to be a useful predictive factor for ovarian function in patients with POI [4]; however, an elevation of serum E2 levels is a result of follicle growth, and therefore, follicles are already growing in these conditions. Likewise, serum FSH levels are affected by hormonal therapy, and consequently, some cycles with low serum FSH levels did not show follicle growth even with ovarian stimulation due to prior human gonadotropin injections [5,6].
Thus, markers with higher predictive value are required even in these conditions, in order to avoid ineffective ovarian stimulations which result in not only economic but also physical and psychosocial burdens for the patients.
Among several markers for ovarian reserve, the serum level of anti-Müllerian hormone (AMH), which is produced by the granulosa cells of early-stage follicles, is considered to be one of the most informative markers for ovarian reserve [7][8][9]. Measurement of AMH levels in the clinical setting has thus been used extensively in, for example, the prediction of ovarian response to controlled ovarian stimulation by gonadotropins, optimization of protocols in assisted reproductive technology, and assessment of ovarian toxicity in medical and surgical conditions [9][10][11][12]. In recent years, a new AMH ultrasensitive enzyme-linked immunosorbent assay (ELISA) kit, the picoAMH ELISA (MenoCheck picoAMH, Ansh Labs, Webster, TX USA) for the detection of very low serum AMH levels, has become available. The lower limit of detection (LoD) of the picoAMH ELISA is 1.2 pg/mL. This is significantly lower than that of the Gen II AMH ELISA (Beckman Coulter Inc., Brea, CA USA), one of the most prevalent AMH assay kits, which has a LoD of 80 pg/mL. We reported that measuring very low serum AMH, which is not detected by conventional ELISA kits, in patients with POI was useful for predicting true ovarian reserve [13].
In the current study, we investigated the usefulness of measuring very low levels of serum AMH using an ultrasensitive AMH assay kit, picoAMH, as the marker for follicle growth of a cycle in patients with POI under HRT conditions.

Patients
Patients with POI who underwent cyclic hormonal therapy between April 2013 and March 2019 were recruited in the study. Patients with POI are younger than 40 years old, exhibit serum FSH levels > 40 mIU/mL on at least two different tests, and present with amenorrhea for more than 3 months in the absence of hormonal therapy. The cyclic hormonal therapies administered included conjugated equine estrogen 0.625 mg (Premarin®; Pfizer Inc., New York, NY USA), 2 tablets/day for 10 days followed by an estrogen (0.05 mg ethinylestradiol) and progestin (0.5 mg norgestrel) combination (Planovar®; Takeda Pharmaceutical Company Limited, Chuo-ku, Tokyo, Japan), 1 tablet/day for 11 days or Premarin® 2 tablets/day for 21 days, and chlormadinone acetate (Lutoral®; Fuji Pharma Co. Ltd., Chiyoda-ku, Tokyo, Japan), 2 tablets/day, for the last 11 days. Following a 21-day treatment with a given cyclic hormonal therapy protocol, patients were subjected to a 7day withdrawal period before the next cyclic hormonal therapy was started. If serum E2 levels exceeded 30 pg/mL or if serum FSH levels dropped to < 15 mIU/mL, during the withdrawal period, the next cyclic hormonal therapy was suspended, and the follicles were monitored once or twice a week by ultrasonography. Follicle growth was defined as the presence of follicles with a mean diameter > 16 mm. In some cases, human menopausal gonadotropins were administered for controlled ovarian stimulation. This study was approved by the ethical committee of Nagoya University Graduate School of Medicine, and informed consent was obtained from all patients.

Hormonal Measurements
Blood samples were obtained during the drug withdrawal period from each patient. Serum was isolated from whole blood samples using a centrifuge, transferred to sterile polypropylene tubes, and stored at − 80°C until the assays were performed. The serum picoAMH concentrations were measured by the same operator using the picoAMH ELISA kit, according to the manufacturer's instructions [14]. The dynamic range of the standard curve of the picoAMH was from 3.0 to 746 pg/mL, and the LoD (limit of detection; the lowest amount of AMH in a sample that can be detected with a 95% probability) and limit of quantitation (LoQ) were 1.2 pg/mL and 3.9 pg/mL, respectively, in accordance with the manufacturer's specifications. In this study, when an AMH concentration of 0.00 pg/mL was obtained using the picoAMH, AMH was considered to be undetectable (negative). The absorbance measurements were performed using a microplate reader (ELx808™ Absorbance Microplate Reader; BioTek Instruments, Inc., Winooski, VT USA). Concentrations of serum antithyroperoxidase antibody (TPOAb), antithyroglobulin antibody (TgAb), E2, and FSH were measured using enzyme-linked immunosorbent assay kits (Elecsys Anti-Tg, Anti-TPO; Roche Diagnostics, Rotkreuz, Switzerland; ARCHITECT® Estradiol, ARCHITECT® FSH; Abbott Laboratories, Abbott Park, IL USA). Serum TPOAb levels > 16 IU/mL and TgAb levels > 28 IU/mL were considered positive.

Statistical Analyses
The data were analyzed using the Prism software (Prism 8; GraphPad Software, San Diego, CA USA) and EZR software (Version 1.40; Saitama Medical Center, Jichi Medical University, Saitama, Japan), the latter of which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). More precisely, EZR is a modified version of R Commander designed to add statistical functions that are frequently used in biostatistics [15]. Mann-Whitney U tests were used to compare patient characteristics, including hormone levels, while Fisher's exact tests were applied to compare the proportions of groups. A p value < 0.05 was considered to indicate statistical significance.

Results
A total of 91 cycles of 19 patients with POI were analyzed in this study. Of the 19 patients, 4 had histories of chemotherapy, 3 had documented chromosomal abnormalities, and 12 patients lacked definitive causes for the POI. Of these 12 patients, seven were positive for thyroid autoantibodies (TPOAb and/or TgAb) ( Table 3). Table 1 shows patient characteristics categorized by positive or negative serum AMH measured using the picoAMH kits. The median age of onset of POI did not differ between the negative and positive AMH groups. Follicle growth was observed in all patients with one or more AMH-positive cycles and only 14.3% in patients with AMH-negative cycles (Table 1).
To evaluate AMH as a predictive factor for follicle growth during a cycle, we compared the serum levels of AMH and FSH and the amount of follicle growth in each cycle. Of the total 91 cycles, 14 were positive for AMH and showed follicle growth. The levels of serum FSH and AMH were significantly lower and significantly higher, respectively, in cycles with follicle growth than those without (Fig. 1). The median serum FSH and AMH levels in cycles with follicle growth were 15.44 mIU/mL (25th, 75th percentile 5.03, 26.85) and 2.77 pg/mL (25th, 75th percentile 0.0, 9.64), respectively.
Of the 14 AMH-positive cycles, nine presented with follicle growth, and five did not show follicle growth. However, of these five cycles, two were positive for AMH during the previous cycles. There were more cycles showing follicle growth in AMH-positive cycles than in AMH-negative cycles (64.3% vs. 6.5%, p = 0.0001). Serum FSH levels in the AMH-positive cycles were significantly lower than in the AMH-negative cycles (Tables 2, 3).
The positive predictive value (PPV) and negative predictive value (NPV) of AMH-positive serum for follicle growth of the cycle were 0.643 and 0.935, respectively. At the same time, we assessed the usefulness of serum FSH levels in the prediction of follicle growth. When we set FSH levels < 10 mIU/mL as a predictor of follicle growth, the PPV of FSH was 0.250 and the NPV was 0.873. These results indicate that AMH-positive serum during the withdrawal period is superior to low levels of serum FSH as a predictor of follicle growth in the cycle.

Discussion
In this study, we demonstrated that detection of very low serum AMH levels in patients with POI can be used as a  [3]; however, it is difficult to predict the exact cycle that will present with follicle growth. Gonadotropin therapy with exogenous estrogen is one of the most common strategies for the induction of follicle growth in patients with POI [4,6,16].
Ovulation was reported to only occur in patients with lower serum FSH (< 15 IU/mL) after estrogen therapy [6], resulting in an ovulation rate of 32% with this therapy, which is higher than the spontaneous ovulation rate of patients with POI.
However, 70% of the patients who underwent ovarian stimulation did not achieve ovulation.
In the current study, serum FSH levels in cycles with follicle growth were significantly lower than those in cycles without follicle growth. However, the PPV and NPV of serum FSH < 10 mIU/mL as a prediction marker for follicle growth were 0.25 and 0.837, respectively. This resulted in 75% of the cycles with lower FSH levels not presenting with follicle growth. In contrast, the PPV and NPV of AMH-positive serum were 0.643 and 0.935, respectively. These values were significantly higher than those of low levels of serum FSH (<  10 mIU/mL). Moreover, the median serum AMH level for cycles with follicle growth was 2.77 pg/mL (25th, 75th percentile 0.0, 9.64). This level of AMH is undetectable using the Gen II AMH ELISA kit (LoD 80 pg/mL) or automated assay systems such as the Elecsys AMH (LoD 10 pg/mL, Roche Diagnostics) and Access AMH (LoD 20 pg/mL, Beckman Coulter Inc.) [14,17]. There were five cycles with AMH-negative serum that presented with follicle growth. Of these five cycles, two were AMH-positive in the previous cycle. Additionally, patients who had positive AMH levels, at least once, presented with follicle growth during the observational periods. These results indicate that the measurement of very low serum AMH levels was useful for predicting follicle growth in patients with POI.
We acknowledge that there were several limitations to this study. Firstly, the number of patients and cycles included in this study were small. However, the frequency of ovulation in the patients and the frequency of cycles were 3.7% and 1.5%, respectively, and these rates were similar to those reported in a previous study [3]. Secondly, an automated assay system for picoAMH is not presently available, and consequently, it is difficult to obtain same-day results for picoAMH tests. Since serum AMH levels are not affected by hormonal status [18,19], clinicians may measure AMH serum levels during hormonal therapy or several days before the withdrawal periods, in order to predict follicle growth. Thirdly, the results of this study are limited to the prediction of follicle growth, and all seven patients who exhibited follicle growth in this study failed to achieve pregnancy. Previous reports indicated that the median serum AMH level of patients with successful IVF was 837 (range 141.6-2750) pg/mL [20], and the live birth rate was significantly lower in women with AMH ≤ 0.1 ng/mL [21]. Consequently, it is impossible to conclude that the detection of AMH in POI patients using the picoAMH kit raises the probability of successful IVF or pregnancy. Fourthly, donor oocytes are available for patients with POI in some countries. However, accessing oocyte donations in some countries is still challenging. Moreover, in vitro fertilization (IVF) with donor oocytes has some issues, such as the high expense and concerns regarding genetics. Therefore, the availability and feasibility of IVF with autologous oocyte could be significant for patients with POI.
In 2018, the US Food and Drug Administration permitted the marketing of the picoAMH ELISA diagnostic test as an aid in the determination of a patient's menopausal status [22]. In addition to predicting menopausal status, the picoAMH ELISA kit may help patients with POI who desire to have children.
In conclusion, we demonstrated the usefulness of very low levels of serum AMH as a marker for follicle growth in patients with POI. Further studies with larger numbers of patients are needed to demonstrate the benefits of measuring AMH levels in patients with POI.