Is it necessary to monitor the serum luteinizing hormone (LH) concentration on the human chorionic gonadotropin (HCG) day among young women during the follicular-phase long protocol? A retrospective cohort study

Background The normal physiological function of LH requires a certain concentration range, but because of pituitary desensitization, even on the day of HCG, endogenous levels of LH are low in the follicular-phase long protocol. Therefore, our study aimed to determine whether it is necessary to monitor serum LH concentrations on the day of HCG (LHHCG) and to determine whether there is an optimal LHHCG range to achieve the desired clinical outcome. Methods A retrospective cohort study included 4502 cycles of in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) from January 1, 2016, to June 30, 2019, in a single department. The main outcome measures included retrieved eggs, available embryos, and live birth rate. Results The LHHCG was divided into five groups: Group A (LH ≤ 0.5), Group B (0.5 IU/L < LH ≤ 1.2 IU/L), Group C (1.2 IU/L < LH ≤ 2.0 IU/L), Group D (2.0 IU/L < LH ≤ 5.0 IU/L), Group E (LH > 5 IU/L). In terms of the numbers of retrieved eggs (15.22 ± 5.66 vs. 13.54 ± 5.23 vs. 12.90 ± 5.05 vs. 12.30 ± 4.88 vs. 9.6 ± 4.09), diploid fertilized oocytes (9.85 ± 4.70 vs. 8.69 ± 4.41 vs. 8.39 ± 4.33 vs. 7.78 ± 3.96 vs. 5.92 ± 2.78), embryos (7.90 ± 4.48 vs. 6.83 ± 4.03 vs. 6.44 ± 3.88 vs. 6.22 ± 3.62 vs. 4.40 ± 2.55), and high-quality embryos (4.32 ± 3.71 vs. 3.97 ± 3.42 vs. 3.76 ± 3.19 vs. 3.71 ± 3.04 vs. 2.52 ± 2.27), an increase in the LHHCG level showed a trend of a gradual decrease. However, there was no significant difference in clinical outcomes among the groups (66.67% vs. 64.33% vs. 63.21% vs. 64.48% vs. 63.33%). By adjusting for confounding factors, with an increase in LHHCG, the number of retrieved eggs decreased (OR: -0.351 95%CI − 0.453-[− 0.249]). Conclusion In the follicular-phase long protocol among young women, monitoring LHHCG is recommended in the clinical guidelines to obtain the ideal number of eggs. Supplementary Information The online version contains supplementary material available at 10.1186/s12958-022-00888-4.


Introduction
Appropriate protocols of controlled ovarian stimulation (COS) are critical for assisted reproductive technology (ART) outcomes. The discovery of Page 2 of 7 Zhang et al. Reproductive Biology and Endocrinology (2022) 20:24 gonadotropin-releasing hormone (GnRH) analogs has offered multiple options in assisted reproduction and improved in vitro fertilization (IVF) success rates [1]. Effective control of the premature luteinizing hormone (LH) peak, reduction of the cycle cancellation rate and more mature oocytes make the GnRH agonist protocol popular in many reproductive centers. In China, the follicular-phase long protocol has gradually become mainstream because of its simple and convenient single administration, good follicular homogeneity, and excellent fresh cycle outcomes [2]. In the GnRH agonist protocol, because of pituitary desensitization, endogenous levels of LH are very low during the late stimulation phase [3]. Moreover, approximately 50% of patients undergoing IVF/intracytoplasmic sperm injection (ICSI) using a GnRH agonist are LH deficient [3]. Thus, it would seem logical that LH supplementation would be beneficial.
According to the "two-cell, two-gonadotrophin theory", both follicle-stimulating hormone (FSH) and LH are important for follicle development in humans. This model explains our understanding of folliculogenesis [4]. LH stimulates theca cells, promoting androgen production, and FSH regulates the proliferation of granulosa cells (GCs) and promotes E2 synthesis. Whereas FSH is the main regulator of antral follicular growth, LH plays key roles in promoting steroidogenesis and in the development of the leading follicle. Moreover, LH exerts different functions during the different stages of both natural and stimulated cycles [5]. During the menstrual cycle, LH not only promotes the growth of larger follicles but also increases granulosa cell FSH activity by increasing androgen synthesis, and then LH promotes the recruitment of follicles.
The normal physiological function of LH requires a certain concentration range, namely, the "LH window" [6]. When below the LH threshold, the eggs cannot fully mature, and there is not enough androgen and estrogen synthesis. Moreover, there is a lack of paracrine signals between granulosa cells and membrane cells. In contrast, when beyond the upper limit, granulosa cell proliferation is inhibited, which can lead to a series of problems in egg development. Therefore, we speculate that it is necessary to monitor the level of LH during the early follicular phase long protocol.
An increasing amount of evidence has demonstrated that downregulation with a GnRH agonist in some normogonadotrophic women may result in profound suppression of LH, which in turn impairs adequate estradiol synthesis during FSH stimulation for IVF/ICSI [4][5][6][7] and reduces the fertilization rate [8], the number of clinical pregnancies [8], and the pregnancy outcomes [9]. However, there is some controversy about this conclusion.
Consequently, this study aimed to determine whether it is necessary to monitor the serum LH concentration on the HCG day (LH HCG ), to identify whether the LH HCG has an impact on the clinical outcome and to determine whether there is an optimal LH range to achieve the expected clinical outcome.

Study design and participants
The subjects of this retrospective study underwent 4502 cycles of IVF or ICSI from January 1, 2016, to June 30, 2019, in our department. The following inclusion criteria were applied: 1) age < 40 years, 2) follicular-phase single-dose GnRH agonist protocol, and 3) fresh cycle transplants. Thin endometrium on HCG day, recurrent miscarriage and endometriosis were excluded. Although GnRH-a has a therapeutic effect on endometriosis, patients diagnosed with endometriosis were excluded in this study because endometriosis leads to infertility and affects the outcome of ART outcomes in many ways [10] ( Fig. 1).
The study was approved by the ethics committee of the Third Affiliated Hospital of Zhengzhou University (2021-105).
Data on patient age and infertility treatment-related characteristics were collected from the files. Basic sex hormone levels (FSH, LH, estradiol [E2], progesterone [P]), Start-up gonadotropin (Gn) does, Gn dose, number of oocytes retrieved, number of diploid fertilized oocytes, number of embryos, number of embryos transferred, and the LH level on day 1 of ovarian stimulation (HCG day 1)/ the HCG day were recorded. E2/P/on the day of HCG administration was recorded.

Stimulation protocols
Follicular-phase single-dose GnRH agonist protocol: A single dose of 3.75 mg long-acting GnRH agonist triptorelin (Diphereline, IPSEN, Paris, France) was administered on day 2 or 3 or 4 of the menstrual cycle. Twenty-eight days later, serum FSH, LH, E2, and P were examined. When FSH and LH were both < 5 IU/L, P < 3.18 nmol/L and E2 < 183.5 pmol/L, Gn was initiated daily until ovulation induction. Chosing the most appropriate FSH starting dose based on patients' age, body mass index (BMI), Ant-Mullerian hormone (AMH) and FSH can obtain a more optimized effect [11,12] ICSI was performed only for severe male factor infertility or previous fertilization failure. Luteal phase support was initiated on day 1 after oocyte retrieval. The cleavage stage embryo transfer time was 3 days after egg retrieval. The blastocyst transfer time was 5 days after egg retrieval. Morphologic criteria were used for embryo scoring [13].

Primary and secondary outcomes
Biochemical pregnancy was assessed by serum hCG detection 14 days after ET [14]. Clinical pregnancy was confirmed if a fetal heartbeat could be observed by transvaginal ultrasound [14]. Severe ovarian hyperstimulation sydrome (OHSS) was diagnosed in women who were hospitalized because they fulfilled one or more of the following criteria: clinical ascites; hydrothorax; and dyspnea (exertional or at rest) [15]. Early abortion was defined as a clinical pregnancy that failed to reach the 12th gestational week. Live birth was defined as the birth of a live child after 28 weeks of gestation per embryo transfer cycle. Ectopic pregnancy was a pregnancy outside the uterine cavity, diagnosed by ultrasound, surgical visualization or histopathology. The premature birth rate was defined as delivery before 37 gestational weeks but not earlier than 28 gestational weeks [16].

Statistical analysis
Differences in variables between the groups were statistically analyzed with Student's t-test and chi-squared tests when appropriate. A bilateral p value < 0.05 was considered to be significant. The results are presented as the mean ± standard deviation. If the variances were not uniform, we used the rank sum test for comparison. Linear regression analysis was performed to assess the association between the LH level on the HCG day and the number of eggs retrieved. The same set of potential confounders was introduced into the regression models for adjustment by the enter method, regardless of whether significant differences between groups were observed. Statistical analysis was performed with the Statistical Package for the Social Sciences (version 24.0; SPSS Inc., USA). All P < 0.05 on one-sided tests was considered to be statistically significant.

Results
When the serum LH level is lower than 0.5 IU/L, both the fertilization rate and the number of embryos decreased significantly. Some clinicians believe that the threshold of LH can be set to 1.2 IU/L; if the LH is lower than this level, follicular development and endometrial growth will be severely insufficient [17]. When the serum LH level is greater than 5.0 IU/L on the day of HCG in the follicular-phase long protocol, the pregnancy rate significantly decreased [18]. Thus, we divided the LH HCG into five groups: Group There were 9 cycles with LH HCG ≥ 10 IU/L in the follicular-phase long protocol, of which only 1 cycle was greater than 20 IU/L. There was no obvious follicle luteinization in these cycles.
According to the basic information and clinical characteristics of the patients, the age of group D was slightly lower than that of group C. The BMI of groups D and E was lower. The basal FSH and LH in groups C, D and E were higher than those in groups A and B ( Table 1).
The level of P in groups A and B was higher than that in the other groups. The AMH value of group C was significantly lower than that of group A. The start-up Gn dose in group D was significantly lower than that in groups B and C. The total Gn dose in group D was significantly lower than that in the other groups. For the hormone concentrations of HCG, the E 2 levels of group B and group C were lower, and the P level in group A was the highest. On the HCG day, the endometrium in group C was thicker than that in groups A and E (Table 1).
In terms of the numbers of retrieved eggs, embryos, high-quality embryos and diploid fertilized oocytes, an  Tables 2 and 3). Logistic regression analysis was used to determine whether the LH HCG level was related to the number of oocytes retrieved. The results showed that in the follicular-phase long protocol, with the increase in LH HCG , the number of retrieved eggs decreased (Table 2 and Fig. 2).

Discussion
The findings of our study suggest that the level of LH HCG can affect the IVF/ICSI outcome in the follicular-phase long protocol in a young population.
The "two cell-two gonadotropin" model has been the key to our understanding of folliculogenesis [4]. According to this model, LH stimulates theca cells, thereby advancing androgen production, and FSH governs the proliferation of granulosa cells (GCs) and promotes E2 synthesis. Both FSH and LH play an important role in folliculogenesis. The details of the specific role of LH in folliculogenesis have not been fully studied. LH, which is well known for its importance in the late follicular phase, supports theca cells in the ovary to provide androgens and hormonal precursors for estradiol production. LH not only promotes the growth of larger follicles but also increases FSH activity in granulosa cells by increasing androgen synthesis [19]. During  Table 3 Linear regression analysis between serum LH concentration of HCG day and eggs retrieved the development of a dominant follicle, its dependence on FSH gradually declines, whereas its dependence on LH gradually increases. FSH's partial role is replaced by LH in the late follicular phase. This facilitates the selection and maintenance of the dominant follicle. The follicular-phase long protocol uses a single dose of 3.75 mg GnRH-a in the early follicular phase for pituitary downregulation. Two weeks after a single injection of 3.75 mg GnRH-a, endogenous sex hormones were almost completely inhibited. The FSH levels gradually recover after 3-4 weeks, and the E 2 levels begin to rise after 7-8 weeks, whereas the inhibitory effect on LH lasts up to 8 weeks after administration [20]. After 30 days, as the FSH level gradually recovers, Gn is used to stimulate the development of multiple follicles in the ovary. In the follicular-phase long protocol, due to pituitary desensitization, endogenous LH levels are very low in the later stage of stimulation. Previous studies had shown that when the serum LH concentration after pituitary downregulation is between 0.5-2.5 U/L, we can obtain a normal ovarian response [21]. Too high or too low is not ideal.
Physiological levels of LH are clearly important for follicle development. If the LH level is abnormal, it will lead to abnormal follicular development. Therefore, it was speculated that there should be an LH threshold window. The Asia-Pacific consensus on the application of LH in assisted reproductive treatment published in 2011 concluded that the reasonable threshold window for LH ranges from 1.2-5 IU/L [22]. However, some studies found that in the process of controlled ovarian hyperstimulation (COH), when the LH level was lower than 1.2 IU/L on the HCG day, it did not affect the clinical outcome [23], whereas for the follicular-phase long protocol, LH was very rarely elevated due to suppression. In this study, the LH concentration was greater than 10 IU/L in only 10 cases, there was only 1 case with LH > 20 IU/L, and the P concentrations on the HCG day were all normal.
Our study found that in the follicular-phase long protocol, with an LH ≤ 0.5 IU/L on the hCG day, patients had a higher number of oocytes retrieved, a better embryo rate, and more embryos, but there was no significant difference in the live birth rate compared with the other groups. Thus, we speculate that when the LH ≤0.5 IU/L, although there may be pituitary hypersuppression leading to inadequate LH secretion in the late follicular phase, it does not affect the outcome. This is similar to the results of the study of Brjercke et al. [24], that is, a lower LH can satisfy the development of follicles. As long as less than 1% of LH receptors are occupied, they can play a normal physiological role [24]. However, when there are older women, a low response, a slow response, etc. in the COH cycle, adding LH may be beneficial. With a normal ovarian response, a low LH concentration does not affect the fertilization outcome [25].
However, as LH increased, the numbers of oocytes retrieved, embryos, and high-quality embryos and the number of diploid fertilized oocytes decreased gradually, which may be because excessive LH levels may cause the ovaries to secrete more androgens, resulting in follicular atresia and then fewer oocytes retrieved. However, the pregnancy rate did not decrease, which may be because the increase in LH level only affected the number of oocytes retrieved but did not affect the quality of the oocytes, the embryo development potential or endometrial receptivity. For Preimplantation Genetic Testing (PGT) patients, it may be more suitable to control the LH at a lower level on the HCG day to retrieve more oocytes.
The main limitations of this article are as follows: 1. The current measurement of LH levels may not accurately reflect the true level of internal LH; 2. Our study mainly enrolled young people, which is a specific group. This may have a certain influence on the outcome; 3. According to follicular development in the later stage, HMG or rLH may be added, which may cause the conclusion to be inaccurate. The cumulative pregnancy rate and cumulative live birth rate were not calculated. The data sample is still not large enough, and additional prospective largesample studies are needed to confirm this idea.
In conclusion, in the follicular-phase long protocol, the LH level is maintained at a low level throughout COH. From the data of our study, we should pay more attention to LH HCG . Elevated LH levels on the hCG day may lead to a reduction in oocyte retrieval. However, even when LH ≥ 5 IU/L, approximately 10 oocytes are retrieved, which, for clinical purposes, is sufficient. However, those who undergo PGT may benefit more when the LH level is controlled within a certain range. Therefore, in the follicular-phase long protocol among young women, monitoring LH HCG has clinical applications.