Background
The use of frozen embryo transfer (FET) cycles has been increasing in recent years, because of the safety and comfort of patients, the decreased risk of ovarian hyperstimulation syndrome (OHSS) and the similar cumulative pregnancy rates compared with fresh embryo transfer cycles [
1‐
5]. The artificial hormone replacement therapy (HRT) cycle for FET is widely used for its stable clinical pregnancy rate and more convenient time schedule [
6,
7]. However, several studies suggested that HRT-FET could increase the risk of pregnancy-related complications (hypertension disorders, placenta accrete and intrahepatic cholestasis of pregnancy), as well as the risk of low birth weight and small for gestational age, which was speculated to be related to the abnormal oestrogen (E
2) level of HRT-FET cycles [
8,
9]. Therefore, the control of serum E
2 level in HRT-FET cycles needs further discussing.
Serum E
2 is essential for endometrial receptivity, myometrial spiral artery remodelling, and placental development [
10]. Studies have shown that higher serum E
2 levels lead to impaired endometrial receptivity and reduced clinical pregnancy rates [
11]. In addition, high serum E
2 levels before embryo transfer in fresh in vitro fertilization (IVF) cycles are strongly associated with decreased embryo implantation rates [
12]. Furthermore, high serum E
2 levels in IVF cycles increase the incidences of preeclampsia, foetal growth restriction and low-birth-weight infants [
13,
14]. Therefore, serum E
2 levels during the HRT cycle may be closely related to pregnancy outcomes.
Hormone replacement cycles are commonly employed in our reproductive medicine centre, and some patients have serum E
2 levels that are obviously higher than the natural physiologic levels before endometrial transformation. However, the number of relevant studies on the relationship between serum E
2 levels before endometrial transformation and pregnancy outcomes of FET cycles is limited to date [
11,
15,
16], and these studies have failed to reach a consensus. Therefore, we conducted a retrospective review of recent hormonal replacement FET cycles at the reproductive medicine centre of Nanjing Drum Tower Hospital to explore the relationships between serum E
2 levels before endometrial transformation and clinical pregnancy outcomes.
Discussion
In our study, the clinical pregnancy rate maintained a high level when the serum E2 level before endometrial transformation was less than 300 pg/mL. However, the serum E2 level couldn’t affect the clinical pregnancy rate of HRT-FET when the serum E2 level is more than 1400 pg/mL.
In 1983, Trounson A first reported a successful pregnancy by FET. FET is currently widely used in clinical human-assisted reproductive technology (ART) [
17], mainly for patients who have a previously failed fresh embryo transfer, whose fresh embryo transfer was cancelled due to the risk of OHSS, or who require embryo storage for other reasons [
18]. Artificial hormone replacement cycles have been more widely employed because of their convenience. Exogenous oestrogen and progestogen are orally administered to change the endometrium to achieve synchronization with the embryos, which is necessary for embryo implantation in HRT-FET cycles. Embryo implantation is one of the central factors in ART and depends mainly on the ability of the endometrium to receive the embryo for implantation and on the quality of the embryo. Endometrial growth in HRT-FET cycles relies on exogenous oestrogens taken by the patients, and there are two regimens: fixed or escalating doses. Considering patient compliance, the HRT-FET cycles we included in this study were all performed with a fixed oestrogen dose. The serum E
2 and P levels and endometrial thickness of the patients were monitored regularly, and progesterone was administered to transform the endometrium when the endometrium reached the expected thickness. However, there are differences in the absorption and metabolism of exogenous oestrogens among different individuals. Therefore, there are also differences in serum E
2 levels among different individuals. Of interest is whether the difference in serum E
2 levels influenced the clinical outcomes of HRT-FET cycles. Some studies have explored the association between serum E
2 levels before endometrial transformation and the clinical outcomes of HRT-FET cycles, but they remain inconclusive [
11,
15,
16,
19]. In our retrospective study, the results suggested that the levels of serum E
2 before endometrial transformation were closely related to the clinical pregnancy rates of patients with HRT-FET cycles when they were less than 1400 pg/mL. Higher clinical pregnancy and embryo implantation rates were achieved when serum E
2 levels were less than 300 pg/mL.
Although sustained elevations in oestrogen in the follicular phase are indispensable for endometrial growth, previous studies have suggested that excessive oestrogen may have adverse effects. In vitro studies reported that oestrogen overexpression in first-trimester human trophoblast cells and the first-trimester placenta is able to inhibit trophoblast invasion by inducing apoptosis, potentially leading to abnormal pregnancy outcomes [
20]. In addition, a number of in vivo studies have explored the effects of serum E
2 levels on endometrial function. In mouse models, oestrogen should be maintained in a certain range to enable the uterus to be receptive, and properly increased serum E
2 levels are closely associated with altered expression of genes involved in embryo implantation [
21]. At the same time, in a baboon model, higher serum E
2 levels during early pregnancy (first 60 days) allowed extra villous trophoblast invasion and uterine artery-related functions affecting the development of pregnancy [
21]. It follows that appropriate serum E
2 levels have important effects on both embryo implantation and ongoing pregnancy but that serum E
2 levels that are too high may adversely affect it. Finding a reasonable range of serum E
2 levels before endometrial transformation in HRT-FET cycles is essential to improve the pregnancy rate of ART cycles [
22]. Clinically, excessive serum E
2 levels after controlled ovarian hyperstimulation (COH) during IVF cycles may lead to a reduced clinical pregnancy rate [
23] or adverse pregnancy outcomes [
13,
24]. In a previous study, when serum E
2 levels reached 3560 ± 1233 pg/mL or even higher [
25] on the hCG trigger day in fresh transplant cycles, the clinical pregnancy rate was significantly lower. However, serum E
2 levels in fresh transplant cycles are much higher than those in HRT-FET cycles in most situations. Therefore, this serum E
2 limit is of limited significance for artificial hormone replacement cycle guidance. There are many studies focusing on HRT-FET cycles that have set cut-off values for serum E
2 levels before endometrial transformation: 299, 400, 600 or 689 pg/mL [
11,
16,
26]. Our study was not a direct equivalent of previous studies. In our retrospective study, a smooth curve fitting model was innovatively used to analyse HRT-FET data from our reproductive medicine centre over a period of nearly 3 years, suggesting that the peak clinical pregnancy rate occurs when the serum E
2 level before endometrial transformation is less than 300 pg/mL. Regarding this, we grouped serum E
2 [
12,
27,
28] levels before endometrial transformation, and further statistical analysis suggested that we could achieve better clinical pregnancy outcomes when serum E
2 levels were lower than 300 pg/mL in HRT-FET cycles. It is generally accepted that a high serum E
2 level in an FET cycle refers to more than a peak (284.5 ± 77.9 pg/mL) value in the physiological state of the natural cycle [
21]. In addition, the serum E
2 level of patients in group 1 (E
2 < 300 pg/mL) of our enrolled cycle was 194.11 ± 56.84 pg/mL, which was closer to the natural cycle situation, which had less of an effect on endometrial receptivity.
According to further observation of the initial fitting curve result, we found that when the serum E
2 level gradually increased to a certain extent, the clinical pregnancy rate had a slight fluctuation. There was a slightly positive correlation between increasing serum E
2 levels and the clinical pregnancy rate when the serum E
2 level was less than 300 pg/mL and within a certain range (approximately 1400–2400 pg/mL). Conversely, when the serum E
2 level was in a certain range (300–1400 pg/mL) and more than 2400 pg/mL, the expected clinical pregnancy rate decreased progressively with increasing serum E
2 levels. The serum E
2 increase in the HRT-FET cycles of this study far exceeded the normal physiological category, which was due mainly to 2 mg of oestradiol for the vaginal plug added in patients with unexpected endometrial thickness. It has been reported that in vaginal plugs with oestradiol 4 mg/day, the serum E
2 concentration can reach a maximum of 4800 pg/mL; in addition, the combination of oral and vaginal oestradiol can achieve better endometrial thickness and improve endometrial receptivity [
29]. Therefore, most of the patients in the HRT-FET cycles with higher serum E
2 levels (> 1400 pg/mL) had oestradiol medication vaginally because of an unexpected endometrial thickness. Higher serum E
2 levels may be better for endometrial proliferation, such that clinical pregnancy rates improve with higher serum E
2 levels when E
2 levels are between approximately 1400–2400 pg/mL. When serum E
2 levels reach a certain range, they will have a limited effect on the improvement of endometrial thickness. When serum E
2 levels are greater than approximately 2400 pg/mL, the clinical pregnancy rate will decrease with higher serum E
2 levels. These results are valuable for the regulation of oestrogen dosage during HRT-FET cycles in our reproductive medicine centre.
In IVF-ET cycles, endometrial thickness can reflect the functional status of the endometrium to a certain degree [
30]. Appropriate endometrial thickness is an essential condition for embryo implantation. Endometrial thickness is a routine detection index to evaluate the ability to accept embryo implantation because of the convenience and maturity of the measuring procedure. At present, most studies believe that endometrial thickness less than 6–8 mm may lead to adverse clinical outcomes [
31]. In our study, there was a significant difference in endometrial thickness between the two groups. Patients with high serum E
2 levels were mainly caused by vaginal medication, and the main reason for vaginal medication is the thin endometrial thickness. Therefore, the endometrial thickness of group 2 was slightly lower than that of group 1. To exclude the influence of endometrial thickness, we further used the threshold prediction model to calculate the cut-off value (8.4 mm) that might affect the clinical outcomes. On this basis, we divide the research data into two subgroups for smoothing curve fitting again. The results showed that the clinical pregnancy rate decreased with increasing serum E
2 levels before endometrial transformation, regardless of endometrial thickness. In addition, we also found the difference of serum P levels between these two groups. There are few studies on the impact of serum P level on the clinical pregnancy outcome of HRT-FET cycles, which mainly discuss the impact of serum P level after endometrial transformation or on the day of embryo transfer. Furthermore, we conducted a univariate analysis of the serum P level before transformation, and the results showed that it has no significant effect on the clinical pregnancy rate (Table. S
1).
The results of our study are not identical to those of some previous studies. Niu et al. [
16] retrospectively reviewed 274 FET cycles. Patients with different serum E
2 levels on the start day of progesterone had similar pregnancy rates. However, the higher E
2 level in their study (299 ± 48.9 pg/mL) was much lower than the higher serum E
2 level in our study. Moreover, the previous study included only the outcomes of cleavage-stage embryo transfer. Celik et al. conducted a retrospective study [
28] of 468 patients in 2019: Serum E
2 monitoring prior to progesterone administration could not predict patient live birth rates. A novel retrospective study [
32] suggested no significant difference in FET clinical outcomes when serum E
2 levels were between 100 and 500 pg/mL before endometrial transformation but that the spontaneous abortion rate was significantly increased when the serum E
2 level was below 100 pg/mL or over 500 pg/mL. However, the highest serum E
2 cut-off value of this study was only 500 pg/mL, which was much lower than the high serum E
2 cut-off value of our study, and this study included only the outcomes of blastocyst transfer. The retrospective analysis in our centre has a larger sample size than previous studies and incorporates different numbers and types of embryos transferred. To control for the influence of embryonic factors on the clinical outcomes of HRT-FET cycles, we further stratified the statistics for the different types of embryos transferred. A smooth curve fit was employed between the level of serum E
2 before endometrial transformation and the clinical pregnancy rate of patients with different numbers and types of embryos transferred, and we found that the clinical pregnancy rates of the different numbers and types of embryos transferred decreased gradually as the level of serum E
2 increased (Fig. S
3). The results suggested that the increased serum E
2 level before endometrial transformation impaired the clinical pregnancy rate regardless of the number and types of embryos transferred.
Our study was limited to HRT-FET cycles without pretreatment. However, a larger number of patients were pretreated with GnRHa before oral exogenous oestrogen and progesterone at our centre, and the relationship between pretransformation serum E2 levels and clinical outcomes in such patients needs further exploration. In addition, we didn’t include the duration of oestrogen used of patients in HRT-FET cycles, which might has influence on the clinical pregnancy outcomes. The main drawback of this study is its retrospective design. To further clarify the effect of serum E2 levels before endometrial transformation on clinical outcomes in HRT-FET cycles, higher-quality and large-scale randomized controlled trials are needed. We can further design clinical randomized controlled studies to clarify the impact of serum E2 level on clinical outcome under different method and dose of exogenous oestrogen administration. Therefore, the dosage of exogenous oestrogen can be reduced to avoid drug abuse and drug-related risks while maintaining a high clinical pregnancy rate.