Clinical and perinatal outcomes of fresh single-blastocyst-transfer cycles under an early follicular phase prolonged protocol according to day of trigger estradiol levels

Introduction

Assisted reproductive technology (ART), which largely utilizes controlled hyperstimulation (COH), has made the dream of parenthood a reality for many infertile patients (Hilbert & Gunderson, 2019). However, the use of exogenous gonadotropin (Gn) during the COH cycle can often cause serum estradiol (E₂) levels to be higher than during the natural cycle (Li et al., 2019). High E₂ concentration in the follicular phase causes early onset of the luteinizing hormone (LH) peak, early endometrial luteinization, and a reduced embryo implantation rate (Fatemi & Popovic-Todorovic, 2013). A previous study found that live birth rates were not affected by E₂ levels on the day of trigger during the pituitary down-regulation cycle (Huang et al., 2014).

Gonadotropin releasing hormone agonist (GnRH-a) remains the most commonly-used peptide hormone in COH during ART treatments because it can obtain the most oocytes, prevent early-onset LH peaks, reduce luteinization, and improve cycle completion rates (Haydardedeoğlu & Kılıçdağ, 2016). Patients that have been treated with long-acting GnRH-a during COH have shown good compliance and clinical pregnancy rates (Gao et al., 2014; Liao et al., 2015). Our previous study also showed encouraging clinical outcomes for long-acting GnRH-a, regardless of the patient’s menstrual cycle time (Ying et al., 2019).

The ultimate goal of ART is to deliver a single, healthy, living baby as quickly and economically as possible. Single-blastocyst-transfer (SBT) can prevent high incidences of complications such as multiple pregnancies and ovarian hyperstimulation syndrome (OHSS) which results in the continuous emergence of maternal and fetal related complications such as gestational hypertension, premature delivery and low-weight newborns (Zeng & Li, 2019).

SBT clinical outcomes are better during the fresh cycle than clinical outcomes during the fresh cleavage stage transplantation (Glujovsky et al., 2016). Moreover, a recent meta-analysis suggested that in patients undergoing in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) cycles, cryopreserved SBT may not be the best choice when compared to fresh SBT (Zeng & Li, 2019).

Until now, the application of fresh SBT under an early follicular phase prolonged regimen has not been thoroughly examined. Do different E₂ levels on the day of trigger influence the live birth rate or increase maternal and fetal complications? In this study, we compared the clinical, obstetrical, and perinatal outcomes of using fresh D5 SBT under an early follicular phase long-acting GnRH-a protocol based on our analysis of E₂ levels on the day of trigger.

Materials & methods

Participants

This was a retrospective study. We collected data from 1,897 cycles of prolonged protocol between June 2016 and December 2018 at a single reproductive center. We received written informed consent from all participants. Blastocysts from 759 cycles had undergone cryopreservation, three patients canceled their cycles for personal reasons, and 364 cycles showed β-hCG values lower than 10 IU/L on the 12th day following blastocyst transfer. Ultimately, we used data from 771 prolonged regimen cycles that used fresh SBT and had β-hCG values of at least 10 IU/L (Fig. 1).

The inclusion criteria were as follows: (1) all candidates were between the ages of 20 and 42 years old and had infertility due to tubal obstruction, endometriosis (EMs), polycystic ovary syndrome (PCOS), or ovarian polycystic changes confirmed by transvaginal ultrasound, male factors, or idiopathic reasons; (2) the endometrial double layer thickness on the day of trigger was greater than 6 mm; (3) the serum progesterone (P₄) level did not exceed 1.98 ng/ml on the day of trigger; and (4) they had one or more D5 3BC blastocyst.

The exclusion criteria were as follows: (1) patients required all blastocysts to be cryopreserved on the day of fresh SBT treatment; (2) endometrial adhesion, submucosal myoma, or uterine diameters longer than 65 mm; (3) obvious infection after oocyte retrieval; and (4) complications with malignant tumors or other systemic diseases, such as an active stage of systemic lupus erythematosus, that were not suitable for pregnancy.

The research was approved by the Research Ethical Committee of the Second Hospital Affiliated to Wenzhou Medical University (approval number: L-2020-09).

Ovarian stimulation and embryo culture

We injected a single full dose of 3.75 mg GnRH-a (Triptorelin, Kilferin, Germany) during the D1-D3 menstrual cycle. The ovarian inhibition situation was evaluated 32 and 38 days after pituitary down-regulation. If there were no follicles greater than 8 mm, the E₂ level did not exceed 50 pg/ml, and the LH level was less than 5 IU/L, we determined that the pituitary desensitization was successful. Additionally, we used 87.5-375 IU of recombinant human FSH (rhFSH; Gonal-F, Merck Serono, Geneva, Switzerland) for COH according to the patient’s age, BMI, antral follicle count (AFC), and levels of LH, P₄, and E₂ on the starting day.

We used timely ultrasonic and blood examination to monitor the patients’ follicular growth, ovarian response, peripheral serum E₂ levels, and LH and P₄ values. If the LH level did not exceed 0.5 IU/L and the follicular growth speed was not ideal, we added 75 IU of recombinant LH (r-LH; Merck Serono, Aubonne, Switzerland). When the average diameter of the two follicles reached 18 mm, we administered a single dose of 4,000–10,000 IU of human chorionic gonadotropin (hCG, Guangdong Lizhu, China). Egg harvesting was performed after 34–36 h.

The embryo culture program has been previously described and the embryo scoring standard referred to the 2011 Istanbul Consensus (Embryology ASiRMaESIGo, 2011; Ying et al., 2019). A blastomere with seven to nine cells of an A or B grade was considered a good quality D3 blastomere. Blastocysts above 4BB of D5 or D6 were considered good quality, and D5 blastocysts that reached at least 3BC could be freshly transferred (Fig. 2).

Results

In this study, we analyzed 771 fresh SBT cycles under an early follicular phase prolonged protocol with β-hCG values no less than 10 IU/L. We compared the patients’ clinical outcomes and perinatal and obstetrical complications according to the serum E₂ levels on the day of trigger, baseline information, COH, and laboratory parameters. Baseline information was similar across the four groups, except with regards to BMI; BMI decreased as E₂ levels increased on the day of trigger (Table 1).

When performing COH under the prolonged protocol, the >75^th percentile group had the highest LH level and AFC on the commencing day (0.48 ± 0.21 IU/L and 19.10 ± 6.26, respectively; all P < 0.001). The 25^th–50^th percentile group had the highest E₂ level on the commencing day and the days of stimulation were highest (28.19 ± 10.22 pg/ml, P = 0.003 and 11.40 ± 2.48 days, P < 0.001). The <25^th percentile group had the highest P₄ level on the commencing day (0.52 ± 0.32 ng/ml P < 0.001). LH levels on the day of trigger were highest in the 51^st–75^th percentile group (0.81 ± 0.62 IU/L, P < 0.001). The >75^th percentile group had the highest levels of total E₂, E₂ per follicle, P₄ and endometrial thickness on the day of trigger (3,992.8 ± 724.4 pg/ml, 259.68 ± 101.14 pg/ml, and 0.85 ± 0.34 ng/ml, 11.42 ± 2.06 mm, respectively; all P < 0.001). However, initial and total Gn doses were highest in the <25^th percentile group (201.66 ± 59.03 IU and 2,434 ± 929.88 IU, respectively; all P < 0.001) (Table 2).

The laboratory results showed that the >75^th percentile group had the highest number of retrieved oocytes, double pronuclear (2PNs), cleavage embryos, D3 embryos, good quality D3 embryos, culture D3 embryos into blastocysts number, blastocysts, and good quality blastocysts per cycle (17.79 ± 4.86, 13.47 ± 4.11, 13.26 ± 4.17, 13.73 ± 4.23, 6.91 ± 3.19, 12.20 ± 4.15, 7.37 ± 2.85, and 3.84 ± 2.61, respectively; all P < 0.001) (Table 3). The rate of good quality D3 embryos was highest in the <25^th percentile group (0.53 ± 0.21, P < 0.001). Conversely, the rates of blastocyst formation and good quality blastocyst formation were highest in the 25^th–50^th percentile group (0.60 ± 0.19, P = 0.017; 0.34 ± 0.19, P = 0.037) (Table 3). Our comparisons of the grades of blastocyst cavity expansion, inner cell mass, and trophoblastic layer showed no significant differences. Blastocyst cavity expansion was almost the 4^th degree of all the groups. The rate of Grade A or B inner cell masses was similar across the four groups, while the trophoblastic layer was mainly composed of Grade B cells with no statistical difference across the groups (Table 3).

Although the >75^th percentile group had the lowest rates of clinical gravidity, embryo transplantation, persistent pregnancy, and living birth, there were no significant differences (85.57% (166/194), 86.60% (168/194), 71.13% (138/194), and 71.13% (138/194), respectively; all P > 0.05). Additionally, rates of severe OHSS, early miscarriage, and ectopic pregnancy showed no significant differences (Table 4).

In order to observe the influence of variables with statistical differences among groups on the live birth percentage, we performed a logistic regression analysis. We found that there was no correlation between BMI, LH level, E₂ level, P₄ level and AFC on the first day of treatment; LH level, total E₂ level, E₂ level per follicle, P₄ level and endometrial thickness on the day of trigger; or the initial and total FSH doses, days of stimulation and number of obtained oocytes, 2PNs, cleavage stage embryos, D3 embryos, good quality D3 embryos, D3 embryos cultured into blastocysts, blastocysts, and good quality blastocysts, rate of good quality D3 embryos, rate of blastocyst formation, rate of good quality blastocyst formation (all P > 0.05, Fig. 3).

We used ROC curves to further analyze the correlation between the E₂ levels on the day of hCG administration and the prediction of blastocyst plantation, clinical gravidity, sustained gestation, early abortion, living birth, and early-onset severe OHSS. We found that E₂ levels on the day of trigger could not be used to predict the rates of fresh single SBT, clinical pregnancy, sustained pregnancy, early abortion, or live birth, and that the AUC was from 51.33% to 55.82% with P values greater than 0.05. We noted that the E₂ level and early onset OHSS ROC curves showed a moderate intensity correlation with an AUC of 72.39%, P value of 0.0292, and a Youden index (YI) of 44.59%; the cut off value for E₂ was 2,893 pg/ml with the 75% sensitivity and 70% specificity (Fig. S1).

The >75th percentile group had the highest male-female neonate rate (1.68 (89/53)), but we did not find any significant differences when compared to the other three groups. The 51^st–75^th percentile group had the highest percentages of low birth weight infants, premature deliveries, hospitalizations in the neonatal intensive care unit (NICU), twin pregnancies, and monochorionic diamniotic pregnancies (11.73% (19/162), P = 0.0408; 11.43% (20/175), P = 0.0269; 10.49% (17/162), P = 0.0029; 8.57% (15/175), P = 0.0047; and 8.57% (15/175); P = 0.001, respectively) (Table 5). Additionally, we found no differences in birth defects across the four groups. Obstetrical complications including GDM, PROM, gestational hypertension, placenta previa, and PPROM also did not show significant differences (Table 5).

Discussion

Few studies have examined the influence of fresh SBT E₂ levels on clinical and maternal-fetal perinatal outcomes under a full long acting GnRH-a regimen on the day of trigger. Our results indicated that E₂ levels on the day of trigger are not effective when used as a predictor for the clinical and perinatal outcomes of mothers and fetuses. However, we found that E₂ levels could still be used to predict early-onset severe OHSS in IVF/ICSI patients with fresh SBT cycles.

There is growing interest in the safety of IVF/ICSI for mothers and newborns, and a great number of doctors and patients in reproductive centers around the world are using SBT because fresh blastocyst stage transfer is a better choice compared to fresh cleavage stage embryo transfer (Glujovsky et al., 2016). No consensus has been reached on the greater success of fresh SBT or vitrified-warmed SBT. Shavit et al. found that vitrified-warmed SBT was associated with poor clinical outcomes and a higher incidence of preeclampsia compared to fresh SBT (Shavit et al., 2017). However, Wei et al. found that frozen SBT resulted in more successful live births in ovulatory women than fresh SBT, but that frozen SBT resulted in an increased risk of pre-eclampsia (Wei et al., 2019). A different study found that there was an equal number of clinical pregnancies and live births using frozen-thawed and fresh blastocyst transfer (Feng et al., 2012). Nonetheless, the majority of reproductive clinicians select the most suitable transplant plan after considering the patient’s safety and the optimal clinical benefits.

Prolonged pituitary down-regulation before starting Gn resulted in higher live-birth rates during fresh embryo transfer due to the greater endometrial receptivity during embryo implantation (Ren et al., 2014; Tian et al., 2019). In our previous study, we found one GnRH-a depot across different fresh embryo transfer menstrual cycles, which was consistent with the findings of Ying et al. (2019). However, few studies have further explored how E₂ levels influence fresh SBT’s clinical and maternal-fetal perinatal outcomes under this protocol.

During the fresh embryo transfer scheme, the hyper-physiological levels of E₂ elevated the P₄ levels, which reduced endometrial receptivity (Wu et al., 2012). Li et al. confirmed that in IVF/ICSI patients, the supra-physiological E₂ concentration in COH affected endometrial transcriptome profiles, resulting in a shift of the embryo transfer window via endometrial mRNA and lncRNA sequencing (Li et al., 2020). Other researchers found that there was not sufficient evidence to confirm or deny the effect of E₂ levels on clinical outcomes during the IVF/ICSI cycle on the day of hCG administration (Kosmas, Kolibianakis & Devroey, 2004). Although the basic BMI information was biased across the four groups, this information did not affect our logistic regression analysis that calculated the patients’ live birth rates.

In this study, the >75^th percentile group had the highest LH level and AFC on the commencing day, total E₂, E₂ per follicle, P₄ levels and endometrial thickness on the day of trigger. Table 1 shows that, although we found no statistical differences in the basic endocrinology across the four groups, the group with the highest E₂ level (the >75^th percentile group) had the lowest average age, basal FSH level, and FSH/LH ratio, indicating that this group may have better ovarian reserve function (Lensen et al., 2018). Therefore, patients with the highest E₂ levels had more eggs with the minimum initial and total Gn amount. Additionally, the >75^th percentile group had the highest number of retrieved oocytes, 2PNs, cleavage embryos, D3 embryos, good quality D3 embryos, D3 embryos cultured into blastocysts, blastocysts, good quality blastocysts per cycle. While, the 25^th–50^th percentile group had the highest rates of blastocyst formation and good quality blastocyst formation. Though we saw no statistical differences in the clinical outcomes and P values ranged from 0.06 to 0.08, the cluster with the highest E₂ levels had the lowest rates of blastocyst embeddedness, clinical and ongoing pregnancy, and live birth. It is likely that statistical differences will be observed if the sample size increases.

In this study, we mapped the relationship between E₂ levels and six different clinical outcomes. Although the ROC results could not predict the clinical outcomes, the E₂ level on the day of trigger was still a relatively reliable indicator of early-onset severe OHSS, an E₂ cut-off value of 2,893 pg/ml with the 75% sensitivity and 70% specificity (Magnusson et al., 2018). The mean E₂ level for all patients in the study was 2,483 pg/ml. The span between this value and the E₂ cut-off value for early-onset severe OHSS was 410 pg/ml. Even though the incidence of early-onset severe OHSS was low (ranging from 0–1.6%) in regards to this prolonged regimen, clinicians should strictly measure Gn dosages during the final stage of COH for patients with good ovarian function, and should be cautious when increasing the Gn dosage in order to better control the risk of early-onset severe OHSS.

Previous studies have confirmed that retrieving 11–15 oocytes produces satisfactory live birth rates and cumulative live births with the lowest rate of OHSS, while retrieving more than 15 oocytes can critically increase the risk of OHSS without increasing the live birth rate (Malchau et al., 2019; Steward et al., 2014). In this study, the average number of eggs retrieved in the <25^th, 25^th–50^th, and 51^st–75^th groups was between 11.04 and 15.44, and the clinical outcomes were encouraging. This result was in accordance with the previous studies, suggesting that a prolonged protocol combined with fresh SBT would be beneficial for most suitable patients.

The results of this study suggested that the twin conception rate from SBT was much less than that of two blastomeres transfer, which was consistent with the results of a previous study (De Croo et al., 2019). Although SBT reduces the rate of twin pregnancy, it increased the incidence of monozygotic twin pregnancies (Busnelli et al., 2019; Liu et al., 2018; Mateizel et al., 2016), which we observed in our study, especially in the 51^st–75^th percentile group (8.57% (15/175), P = 0.0047). The incidence of monozygotic twins is 0–13.2% during blastocyst transplantation (Hviid et al., 2018), and another study showed that the incidence of multiple pregnancies with zygotic division after single embryo transfer is 1.36% (Ikemoto et al., 2018). Nevertheless, there is no consensus on why the monozygotic twinning rate increases during blastocyst transfer, but the possible explanations are the impact of maternal age, extended culture in fresh and frozen cycles, fresh transfer, in vitro maturation, assisted hatching, and the treatment year (Busnelli et al., 2019; Ikemoto et al., 2018; Knopman et al., 2014; Liu et al., 2018; Mateizel et al., 2016; Ota et al., 2020). Other studies have shown that the monozygotic twin rate is unrelated to maternal age, the use of diverse zona pellucida operation techniques, the type of culture medium, ovarian stimulation, or fertilization methods (Ikemoto et al., 2018; Mateizel et al., 2016).

The mechanism of monozygotic twinning is not fully understood. The accepted theory is that the embryo splits after fertilization to produce monozygotic twins (Sundaram, Ribeiro & Noel, 2018). Corner’s model suggested that herniation and subsequent splitting of the inner cell mass during the blastocyst stage could produce monozygotic twins (Knopman et al., 2010; Nakasuji et al., 2014). Some scholars believe that prolonging embryo culture time could reduce intercellular adhesion and lead to the division of the inner cell mass (Behr et al., 2000). Embryo splitting occurs after the blastocyst stage (Kyono, 2013). It has also been suggested that the occurrence of monozygotic twins could be reduced by using the time-lapse technique to exclude blastocysts that contain loose inner cell mass (Otsuki et al., 2016). We hypothesize that the incidence of monozygotic twins was so high (between 2,407 and 3,028 pg/ml) on the day of trigger in our experiment because the concentration of E₂ may have destroyed the adhesion of the inner cell mass. However, this needs to be further explored.

Conclusion

By analyzing the effect of E₂ levels in fresh SBT under a prolonged protocol on clinical, maternal, and fetal perinatal outcomes on the day of trigger, we found that E₂ concentration may not be effective at predicting patients’ clinical and perinatal outcomes. However, E₂ levels can still be used to predict the incidence of early-onset severe OHSS, and the cut-off value was 2,893 pg/ml with the 75% sensitivity and 70% specificity under a prolonged protocol. When estrogen levels were between 2,407 and 3,028 pg/ml, we found the highest proportion of single chorionic twins transferred with fresh SBT, leading to an increase in premature births, NICU hospitalizations, and low birth weight. Overall, using a prolonged protocol with fresh SBT offers patients efficiency and security. It may be worthwhile to design a prospective study that can further investigate the effect of trigger day E₂ levels on the clinical and perinatal effects of fresh SBT under an early follicular phase prolonged protocol. It would also be valuable to explore the mechanism of E₂ concentration on the day of trigger in the incidence of monozygotic twins.

Supplemental Information