GnRH triggering may improve euploidy and live birth rate in hyper-responders: a retrospective cohort study

Abstract

Purpose

Despite the increasing use of GnRHa to trigger final oocyte maturation in segmented IVF cycles, the effects of trigger modality on chromosomal competence and embryo quality remain controversial. Hence, the purpose of this study was to compare euploidy rates and pregnancy outcomes among hyper-responding women using hCG versus GnRHa trigger.

Methods

This retrospective study included 333 hyper-responders, defined as >15 oocytes retrieved, who underwent preimplantation genetic testing (PGT-A) in segmented IVF cycles using either GnRHa or urinary hCG trigger. Live birth rate (LBR) was the primary outcome of interest. Implantation rate (IR), clinical pregnancy rate (CPR), and euploidy rate were secondary outcomes.

Results

GnRH triggering was associated with improved IR (70.5 vs. 53.2%, p = 0.0475), LBR (51.3 vs. 33.8%, p = 0.0170) compared to hCG. A greater number of oocytes were retrieved (21.9 vs 18.4%, p < 0.001) and euploid embryos produced (2.8 vs. 2.1, p = 0.0109) after GnRHa triggering, while higher euploidy rates were only observed among women <35-years-old (62.0 vs. 51.7%, p = 0.0307) using GnRHa trigger. Higher OHSS rates were observed after hCG triggering (10.6 vs. 2.1%, p = 0.0009).

Conclusion

Hyper-responders who received GnRHa trigger experienced improved pregnancy outcomes and lower rates of OHSS compared to hCG triggering. The higher number of oocytes retrieved and euploid embryos produced may reflect an improved developmental competence using GnRHa triggering due to physiologic induction of both LH and FSH surge or other undefined mechanisms that improve embryo development. However, higher overall euploid rates were only observed among women <35-years-old using the GnRHa trigger. Further prospective studies are required to validate this observation and evaluate the specific influence of different ovulation triggers on gamete developmental competence among hyper-responder women.

Electronic supplementary material

The online version of this article (10.1007/s10815-020-01842-2) contains supplementary material, which is available to authorized users.

Introduction

The physiological luteinizing hormone (LH) surge initiates a complex signaling cascade that mediates the process through which an oocyte resumes meiosis and attains competence for fertilization. For assisted reproduction techniques, human chorionic gonadotrophin (hCG) can be administered as a surrogate for the LH surge to catalyze the process of final oocyte maturation, and to promote subsequent implantation with appropriate luteal phase support in fresh transfer cycles [1, 2]. However, the long half-life of hCG that makes it clinically useful also increases the risk of ovarian hyperstimulation syndrome (OHSS) secondary to prolonged stimulatory effects on the corpus luteum [3]. More recently, the gonadotropin receptor hormone agonist (GnRHa) trigger, which induces quick and reversible luteolysis, has been favored, particularly among hyper-responders who are at particularly high risk of OHSS [4, 5]. Despite these advantages, the effect of the GnRHa trigger on oocyte quality, embryo development, and overall pregnancy outcomes remain a topic of ongoing debate.

Since its introduction, GnRHa triggering has been associated with a higher number of mature oocytes retrieved and a lower incidence of OHSS [6, 7], yet the effect on overall pregnancy and live birth rates are more controversial. A recent Cochrane review suggests that GnRHa triggering may reduce pregnancy rates and increase miscarriage rates in fresh autologous cycles compared to hCG trigger, owing to a luteal-phase deficiency as a result of the shorter duration and smaller amplitude of LH and FSH surge induced by GnRHa [1]. However, after appropriate luteal phase support, clinical pregnancy outcomes appear to be similar in fresh transfer cycles [8]. Furthermore, GnRHa has shown to differentially affect intracellular signaling and embryo developmental kinetics compared to hCG, thereby suggesting that the type of trigger used may not only have an effect on the luteal phase in fresh transfer cycles, but also a more direct effect on oocyte quality and subsequent embryo competence [9–11]. In frozen embryo transfer (FET) cycles where the endometrial preparation is controlled by exogenous hormonal support, the influence of each trigger on oocyte quality, embryo competence, and pregnancy outcomes can be further evaluated independent of the endometrial variable [12].

A recent retrospective study found similar euploidy rates after GnRHa trigger compared to hCG trigger, yet comparison groups were significantly different and subsequent pregnancy outcomes not reported [13]. Furthermore, Kaye al. [14] compared outcomes after immediate and delayed FET and demonstrated similar clinical and ongoing pregnancy rates after GnRHa and hCG triggers in a normal responder population. Conversely, Tannus et al. [15] found that the hCG trigger led to higher embryo implantation rates and significantly reduced time to live birth among hyper-responder women, thereby suggesting a suboptimal effect of GnRHa triggering on oocyte and embryo competence compared to hCG. However, markers of embryo quality such as embryo grading or aneuploidy testing were not reported, which may provide further insight into the specific effects of each trigger on embryo viability. Furthermore, the incidence of adverse outcomes was omitted in prior studies, which is a particularly important clinical outcome among hyper-responders who are at high risk of OHSS. Notwithstanding, it remains uncertain whether the limited LH surge seen in GnRHa-triggered cycles has any beneficial or deleterious effects on the important determinants of ART success after frozen embryo transfer, such as embryo quality, aneuploidy, and subsequent live birth rates compared to triggering with hCG. Hence, the aim of this study was to compare the effects of GnRHa and hCG triggering among hyper-responder women undergoing preimplantation genetic testing for aneuploidy (PGT-A) to determine whether different triggers affect the rate of aneuploidy and subsequent FET outcomes.

Materials & methods

Study design

A retrospective study was conducted at an academic fertility center between January 2014 and January 2017. All stimulated IVF cycles were reviewed for possible inclusion and all patients underwent a standard infertility evaluation within 12 months of starting IVF treatment, including transvaginal sonography and assessment of the uterine cavity via HSG (hysterosalpingogram) or hysteroscopy. Baseline FSH and antral follicle count (AFC) were recorded for all patients, but AMH was only performed for a minority of cases and hence was not reported. TSH and thyroid peroxidase antibody screening was also performed and thyroid hormone supplementation provided when indicated.

Inclusion criteria included stimulation with the GnRH antagonist protocol with final oocyte maturation induced by either 5000/10000 IU hCG or GnRH agonist resulting in hyper-response (defined as >15 oocytes retrieved [16, 17]), trophoectoderm biopsy for PGT-A, and subsequent frozen embryo transfer cycle with a euploid single embryo transfer. Exclusion criteria included GnRH agonist cycles, dual trigger protocols, cycles where <15 oocytes were retrieved, severe male factor infertility requiring surgical retrieval of sperm, non-PGT-A cycles, and cycles in which subsequent euploid blastocyst frozen-embryo transfer (FET) was not performed. The decision to undergo PGT-A was a joint discussion between patient and physician and choice of trigger was made according to physician discretion. The study was approved by the University of British Columbia Institutional Review Board (IRB).

Ovarian stimulation and oocyte retrieval

Ovarian stimulation was performed using a GnRH antagonist protocol with a combination of recombinant FSH (Puregon, Merck Canada) and highly purified urinary gonadotropins (Menopur, Ferring). Dosing regimens were chosen per physician discretion, considering patient specific factors such as age, BMI, and ovarian reserve tests (FSH, AMH and/or AFC, depending on which were available). Gonadotropin was started on day 2–3 of the menstrual cycle. Fixed start of antagonist (Orgalutran, Merck Canada) was initiated on day 6 of stimulation and continued until the day of trigger. Based on discretion of the treating physician, final oocyte maturation was achieved by administering either hCG (Pregnyl, Merck Canada) 5000 IU or 10,000 IU) or a single bolus of GnRH-a (Decapeptyl 0.2 mg, Ferring) when at least three follicles were > 17 mm and oocyte retrieval was performed 35 h after triggering.

Embryo culture, preimplantation genetic testing (PGT-A), and vitrification

Intracytoplasmic sperm injection (ICSI) was performed 4 h after retrieval after denuding of cumulus-oocyte-complexes (COCs). Embryos were cultured in Global-Plus media (Life Global) and assisted hatching of viable cleavage stage embryos performed on day 3 with the Zilos-tk Zona Infrared Laser Optical System (Hamilton-Thorne). Culture was continued in individual droplets until day 5 or 6 and trophectoderm biopsy performed on blastocysts with a viable grade, generally considered to be 3CC or greater according to SART grading criteria [18]. Micromanipulation and biopsy were performed in HEPES supplemented with 20% serum with the use of laser pulses to release five to ten Trophectoderm (TE) cells for analysis, which were sent for either aCGH (array comparative genomic hybridization) or NGS (next generation sequencing) testing, prior to and after April 2014, respectively, at commercial testing laboratories (Reprogenetics & iGenomix). Vitrification and eventual warming was performed with Vitrolife Rapid-I Vitrification System (Vitrolife) according to manufacturer’s protocols.

Endometrial preparation

Oral estradiol (Estrace 2 mg) was administered from day 2 of menses and escalated every 5 days at 2 mg intervals to a maximum of 6 mg daily. Transvaginal sonography (TVS) was used to assess the pattern and thickness of the endometrium approximately 14 days after menses, and progesterone (Endometrin 200 mg TID) was administered when a trilaminar pattern was achieved with a thickness between 8 and 14 mm. In all cases, single embryo transfer was performed after 5 days of progesterone treatment and subsequent luteal support continued until 12 weeks gestation.

Main outcomes and statistical analysis

Live birth rate (LBR) was the primary outcomes of interest. Secondary outcomes included implantation rate (IR), clinical pregnancy rate (CPR), miscarriage rate (MR), euploidy rate, and incidence of OHSS. IRs were calculated based on a positive pregnancy test after embryo transfer, which included all biochemical pregnancies, while CPR was defined by the presence of a viable fetal heart rate between 7 and 9 weeks’ gestation. LBR including all pregnancies beyond 24 weeks’ gestation resulting in a live birth. MR included both biochemical and clinical losses.

Statistical analysis was performed with the use of R v. 3.3.2. Statistical comparisons of outcomes were performed using chi-squared goodness-of-fit test. The Fisher’s exact test was used when the expected frequency of outcomes was less than 5 due to smaller sample size and comparisons of patient characteristics were performed using Mann-Whitney-Wilcoxon testing for non-parametric data. In all cases, statistical significance was considered to be at P < 0.05.

RESULTSs

Patients and cycle characteristics

Three hundred and thirty-three unique patients, with a mean age of 36.1 years, were included in this study. As shown in Table 1, no significant differences were observed in average age, BMI, duration of infertility, or prior pregnancy history between GnRHa and hCG trigger groups. With regard to baseline hormone status, the GnRHa group had a higher average baseline E2 level (164.2 vs. 137.9 pmol/L, p = 0.0116), but similar baseline FSH and AFC values (7.7 ± 2.6 vs. 8.1 ± 1.2, p = 0.17) were observed. Furthermore, no significant differences in the causes of infertility were appreciated between groups.

Table 1.

Baseline characteristics of study groups

Characteristic	GnRHa (n = 216)	hCG (n = 117)	P value
Age (y)	36.0 (29–46)	36.2 (28–44)	0.6124
BMI (kg/m2)	23.4 ± 4.31	23.9 ± 4.48	0.2022
No. of Previous Gravida	0 (0–1)	0 (0–1)	0.9323
No. of Previous Para	0 (0–0)	0 (0–0)	0.1470
No. of Previous Abortus	0 (0–1)	0 (0–1)	0.4710
Infertility Duration (months)	29 (15–42.5)	30.5 (18.8–49.3)	0.774
Baseline FSH (IU/L)	6.2 ± 1.67	6.4 ± 2.17	0.3
Baseline AFC	7.7 ± 2.6	8.1 ± 1.2	0.17
Baseline E2 (pmol/L)	164.2 ± 102.19	137.9 ± 62.31	0.0116
Infertility Diagnosis (%)
PCOS	9.7	6.0	0.3335
Male Factor	14.4	18.2	0.0953
Unexplained Infertility	28.2	18.8	0.0771
Other Ovulatory Disorders	12.0	19.7	0.0869
Tubal/Uterine Factor	7.9	5.7	0.0988
Mixed factors	20.8	25.6	0.3869
Other	6.9	6.0	0.9155

Continuous variables are presented as median (interquartile range) or as mean ± SD. Statistical significance was defined as P ≤ 0.05. BMI = body mass index

10,000 Vs. 5000 IU hCG trigger

A comparison of ovarian stimulation and subsequent embryo development outcomes between 5000 and 10,000 IU hCG trigger groups demonstrated a slightly higher number of oocytes retrieved in the 5000 IU hCG group (19.6 vs. 17.4, p = 0.0107, Supplemental Table 1). However, there was a higher percentage of metaphase II oocytes retrieved in the 10,000 IU hCG trigger group (82.8 vs. 77.0%, p = 0.0157), while the average number of MII oocytes were similar between groups. Fertilization, blastulation, and day 5 biopsy rates were comparable regardless of dose of hCG trigger used. Moreover, the number of euploid blastocysts produced (2.3 vs. 2.0, p = 0.84) and overall euploidy rate (45.4 vs. 42.9, p = 0.603) were similar between the 5000 IU and 10,000 IU hCG trigger groups.

Effects of GnRH vs. hCG on embryo development

As shown in Table 2, patients in the GnRHa trigger group received a similar average gonadotropin dose (2164 vs. 2208 IU, p = 0.153) and obtained similar peak E2 levels (13,101 vs. 12,616 pmol/L, p = 0.113). However, a significantly higher rate of OHSS (10.8 vs. 2.1%, p = 0.0009) and greater proportion of moderate severity cases (6.9 vs. 0.8%, p = 0.038; Supplemental Table 2) was noted in the hCG trigger group. With regard to oocyte retrieval and embryo development, a higher average number of total oocytes (21.9 vs. 18.4, p = 3.48e-09) and MII oocytes (17.6 vs 14.6, p = 1.5e-06) were retrieved in the GnRHa trigger group compared to hCG trigger group, while the fertilization and blastulation (per 2PN) rates were similar between the two trigger groups. Overall, a higher average number of euploid blastocysts were produced in the GnRHa group (2.8 vs. 2.1, p = 0.0109), while the euploid rate was comparable (47.9 vs. 43.8%, p = 0.1842).

Table 2.

Characteristics of ovarian stimulation cycles

Characteristic	GnRHa (n = 237)	hCG (n = 130)	P value
Gonadotropin Dose (IU)	2154 ± 437	2208 ± 757	0.353
Peak Stimulated E2 (pmol/L)	13,101 ± 3218	12,616 ± 3968	0.114
OHSS (%)	2.1 (5/237)	10.9 (14/130)	0.0009
Oocytes Retrieved	21.9 (15–53)	18.4 (15–41)	<0.001
MII Oocytes	17.6 (3–44)	14.6 (6–35)	<0.001
% MII Oocytes	80.2	79.8	0.903
Fertilized (2PN) Oocytes	13.2 (5.3, 1–39)	10.7 (3.8, 3–28)	<0.001
Fertilization Rate (%)	73.9	71.5	0.1001
Blastocysts (#)	5.5 (3.5, 1–23)	4.6 (2.5, 1–13)	0.0231
Blastulation Rate (% 2PN)	42.1 (19.5, 5.6–100)	43.4 (20.2, 8.3–100)	0.5925
Euploid Blastocysts	2.8 (0–14)	2.1 (0–9)	0.0109
Euploid Rate (% Blastocysts)	47.9	43.8	0.1842

Continuous variables are presented as median (interquartile range) or as mean ± SD. Statistical significance was defined as P ≤ 0.05

Age-stratified effect of GnRH vs. hCG on embryo development / euploidy rate

As shown in Table 3, subgroup analysis by maternal age demonstrated a consistent increase in both the total and mature number of oocytes retrieved in the GnRHa compared to hCG trigger group. Fertilization and blastulation rates were also consistently similar between study groups; however, the average euploidy rate was significantly higher with the GnRHa compared to hCG trigger (62.0 vs. 51.7%, p = 0.0307) among women <35-years-old (Fig. 1).

Table 3.

Characteristics of ovarian stimulation cycles stratified by age

Age	Trigger	Average Oocyte Retrieved	Average Matured Oocyte**	Average Fertilization Rate (%)	Average Blastocyst Rate (%) (/2PN)	Average Euploid Rate (%) (/Blast)
<35	GnRHa (n = 79)	23.2	18.9	73.7	43.8	62.0
	hCG (n = 46)	19.9	15.7	70.5	46.5	51.7
	P value	0.0015	0.0019	0.1692	0.5270	0.0307
35–39	GnRHa (n = 106)	21.3	17.2	75.0	44.7	48.4
	hCG (n = 50)	17.9	14.1	73.6	46.2	48.6
	P value	0.0004	0.0014	0.6802	0.7757	0.8247
>39	GnRHa (n = 52)	21.4	16.6	72.1	34.4	25.4
	hCG (n = 34)	17.2	14.0	69.7	35.3	26.1
	P value	0.0001	0.0329	0.2971	0.6843	0.5356

Statistical significance was defined as P < 0.05

Figure 1.

Euploid rates by trigger modality among hyper-responders in GnRH antagonist cycles. *denotes stastically significant difference. Stastical significance defined as P ≤ 0.05

Clinical outcomes after euploid FET

As shown in Table 4, the average number of FET cycles and embryos transferred per cycle were similar between study trigger groups. With regard to reproductive outcomes, the IR per transfer was significantly higher in the GnRHa group (70.5 vs. 53.2%, p = 0.0143). CPR per transfer was also higher in the GnRHa group, although not statistically significant (59.5 vs. 46.9%, p = 0.0842). Most notably, both LBR per transfer and cumulative LBR were higher in the GnRHa group (51.3 vs. 33.8%, p = 0.0170, and 72.1 vs. 51.0, p = 0.0145, respectively, Fig. 2), while the MR was similar between GnRHa and hCG trigger groups (19.23 vs. 19.48%, p = 0.988, respectively).

Table 4.

Clinical outcomes of frozen embryo transfer (FET) cycles

Variable	GnRHa (n = 156)	hCG (n = 77)	P value
No. of FET cycles per patient	1.4 ± 0.77	1.5 ± 0.92	0.5013
Implantation Rate (%)	70.5 (110/156)	53.2 (41/77)	0.0143
Clinical Pregnancy Rate (%)	62.18 (97/156)	49.35 (38/77)	0.0842
Miscarriage Rate (%)	19.23 (30/156)	19.48 (15/77)	0.988
Live Birth Rate (%)	51.3 (80/156)	33.8 (26/77)	0.0170
Cumulative Live Birth Rate (%)	72.1 (80/111)	51.0 (26/51)	0.0145

Statistical significance was defined as P < 0.05

Fig. 2.

Pregnancy outcomes by trigger modality among hyper-responders in GnRH antagonist cycles. *denotes stastically significant difference. Stastical significance defined as P ≤ 0.05

Discussion

In this retrospective study, we examine differences in embryo development and clinical outcomes using two different triggers of final oocyte maturation among hyper-responder women undergoing segmented IVF cycles for PGT-A. Overall, our results demonstrate that compared to women who received the hCG trigger, those who were triggered with GnRHa experienced a significantly higher implantation and LBR per ET as well as cumulative LBR. Despite similar baseline characteristics, women in the GnRHa group demonstrated a significantly higher number of oocytes retrieved, lower rate of OHSS, and greater number of euploid embryos compared to those in the hCG trigger group. To the best of our knowledge, this is the first study to evaluate the effect of different triggers on outcomes of PGT-A in a hyper-responder population.

Since its introduction, the primary value proposition of GnRHa triggering has been to reduce rates of OHSS, owing to the shorter duration and smaller amplitude LH and FSH surge, compared to the hCG trigger [19–21]. However, given the irreversible luteolysis generated by the GnRHa trigger, initial studies noted significantly worse pregnancy rates after fresh embryo transfer in patients who received a GnRHa compared to hCG trigger without adequate luteal phase support [22–24]. Besides their influence on the luteal phase and endometrial receptivity in fresh transfer cycles, several prior studies have attempted to evaluate the effect of different triggers on oocyte quality and embryo developmental competence using oocyte donor and fertility preservation models. For instance, Erb et al. [6] compared the effects of hCG and GnRHa triggers in an oocyte donor program and noted a significant increase in number of oocytes retrieved (23 vs. 15, p = 0.002), MII oocytes retrieved (22 vs. 13, p = 0.001), and high grade embryos (14 vs. 8, p = 0.01) per treatment cycle in the GnRHa compared to hCG trigger group. Although no differences in clinical pregnancy rates were observed, the authors postulated that GnRHa may produce higher-quality embryos due to a more physiologic LH surge compared to the hCG trigger, which may adversely affect oocyte quality by over-luteinizing follicles due to its long half-life. Similarly, our results are concordant with Reddy et al. [25] who noted improved cycle outcomes as evidenced by a significant increase in the number of mature oocytes and cryopreserved embryos obtained after GnRHa triggering, despite similar baseline AMH levels in a cohort of breast cancer patients undergoing fertility preservation. Conversely, Acevedo et al. [26] found no significant differences in the total number of retrieved oocytes or fertilization rates (80% vs. 65%, p > 0.05) between donors who ovulated with GnRHa compared to hCG triggers, although they did note a similarly significant reduction in the incidence of OHSS in the GnRHa trigger group (0 vs. 4, p < 0.05).

With respect to frozen transfer cycles, the adverse effects of rapid luteolysis from GnRHa triggering is no longer of concern since subsequent embryo transfer occurs in a different month after adequate endometrial preparation. In a prospective observational study, Griesinger et al. [27] noted similar quality embryos by morphology between the GnRHa and hCG trigger groups, yet both live birth rates per ET (30.0% vs. 18.5%) and cumulative live birth (21.6% vs. 16.1%) were higher in the GnRHa compared to hCG groups, respectively; however, these differences were not statistically significant, likely owing to the low number of patients who returned for frozen-embryo replacement cycles. In the only other reported study evaluating the effect of different triggers among hyper-responders undergoing a freeze-all protocol, Tannus et al. [15] found a similar cumulative live birth rate between trigger groups (48.08% vs. 48.15%, P = 0.9), but patients in the GnRHa group required more frozen-blastocyst transfer cycles (2.12 vs. 1.32, p < 0.001) and a higher proportion of embryos thawed and transferred (57% vs. 33%, P < 0.001) until a live birth was achieved. Interestingly, the authors noted a higher number of oocytes retrieved and mature oocytes available in GnRHa group, but no difference in embryo formation rate, thereby inferring that the GnRHa trigger may have negative effects on embryo and oocyte competence in this hyper-responder population. Despite this conclusion, no differences in embryo survival rate or embryo quality was observed. Thorne et al. [13] compared GnRHa and hCG triggering among patients undergoing PGT-A and found a greater number of oocytes retrieved and higher rate of euploid embryos produced (33.9% ± 2.2 vs 28.0% ± 1.9, P = .04) using the GnRHa trigger; however, average maternal age and baseline AMH were significantly different between groups and euploid rates were similar after multivariate linear regression. Furthermore, pregnancy outcomes were also similar between groups (63.2% vs. 59.7%, p = 0.58). In contrast, we found a higher number of oocytes retrieved and MII oocytes available for fertilization as well as a significantly higher implantation and live birth rate per euploid blastocyst transfer in the GnRHa trigger group despite similar baseline characteristics between groups. After stratifying for age, we also noted a significantly higher average euploidy rate in the <35-year-old age group (62% vs. 51.7%, p = 0.0307). As demonstrated by Ata et al. [28], euploidy rate is unrelated to the number of embryos available for biopsy, hence the improved euploidy rate observed in the GnRHa trigger group may not be a mere consequence of an increased oocyte retrieval and blastocyst rate and may instead reflect an intrinsic difference in the effect of each trigger on embryo competence. Furthermore, given the higher implantation and live birth rates after euploid embryo transfer, this may also suggest that there are other mechanisms through which the GnRHa trigger influences oocyte and embryo competence beyond simply having a normal complement of chromosomes.

Although controversy exists regarding the benefits and indications for a freeze-all approach in IVF [29–31], there appears to be significant consensus in support of a selective freeze- all approach among high responders who are at higher baseline risk of OHSS [32]. In addition, a recent review of registry data by Acharya et al. [33] further suggests that hyper-responders, defined as >15 oocytes retrieved, may also benefit from improved clinical pregnancy and live birth rates, as well as a lower incidence of low birth rate infants with a freeze-all approach. In addition to simulating an LH surge to induce final oocyte maturation, emerging evidence suggests that the underlying mechanism between different triggers differ substantially with respect to effect size and downstream hormone signaling that ultimately have significant effects both in terms of luteal phase endometrial receptivity as well as the intrinsic oocyte quality and development competence of a developing embryo. While hCG has been shown to elicit a supraphysiologic response of intracellular cAMP accumulation upon binding to the LH receptor, GnRHa binds to pituitary GnRH receptors and promotes endogenous secretion of both LH and FSH, and has been shown to preferentially stimulate AKT and extracellular signal regulated protein kinase (ERK1/2) phosphorylation which is responsible for granulosa cells proliferation, differentiation and survival [10]. Although mid-cycle FSH is not critical for oocyte maturation to occur, emerging evidence suggests that it has implications on pre-antral follicle recruitment, dominant follicle selection, and proper resumption of the meiotic processes of the oocyte, thereby influencing subsequent embryo competence [34–37]. In addition, FSH has been shown to increase LH receptor expression in granulosa cells and additionally may directly influence the expansion of cumulus–oocyte complexes and oocyte maturation through intracellular calcium mobilization, C-kinase activity, protein kinase C, and lipoxygenase [34, 35, 38]. Clinically, Lamb et al. [39] demonstrated improved oocyte developmental competence upon FSH supplementation as demonstrated by improved fertilization rates in a randomized placebo-controlled trial, further supporting the potential benefit of FSH at the time of final oocyte maturation. This also supports the notion of a potentially beneficial role of a GnRHa-induced FSH surge for promoting oocyte maturation and fertilization [40, 41].

In addition to the effects on FSH, GnRHa trigger leads to significantly lower levels of epidermal growth factor-like peptide amphiregulin protein in the follicular fluid [42] and higher levels of mRNA amphiregulin expression in granulosa cells [9] compared to the hCG trigger, both of which have been positively associated with improved markers of embryo quality and fertilization rates [43, 44]. With respect to post-fertilization embryo development, Gurbuz et al. [45] utilized time-lapse imaging to demonstrate different morphokinetic properties among embryos formed after GnRHa and hCG triggers in antagonist cycles. Specifically, the study showed that oocytes retrieved after GnRHa use were fertilized earlier, and the embryo cleavage until the six-cell stage was shorter in the GnRHa compared to hCG trigger group. However, later development was similar between the two groups and the clinical relevance of such differences were not evaluated. Nevertheless, these studies suggest that the choice of trigger influences both the cellular and clinical milieu which influence maturing oocytes and developing embryos and may explain the differences in euploidy rates and clinical outcomes in this study.

To the best of our knowledge, this is the first study to examine how triggering with hCG compared to GnRHa affects euploidy rates as determined by PGT-A, and the pregnancy outcomes in subsequent frozen embryo transfers of euploid embryos in these 2 groups. Given current advances in cryopreservation and increasing trends towards a freeze-all approach to IVF [29], the results presented in this study are clinically relevant. The major strengths of this study are the large number of patients undergoing PGT-A and subsequent FET, thereby controlling for differences in luteal phase support and endometrial receptivity associated with each trigger. Furthermore, inclusion of all women <43-years-old as well as reporting of various markers of embryo quality, such as euploidy testing and morphologic grading, provide further insight into possible mechanisms through which each trigger improves developmental competence and pregnancy outcomes compared to previous studies on this topic [15]. Finally, it is also important to emphasize that hyper-responders to ovarian stimulation are among those at highest risk of OHSS [46] and recent evidence suggests that a freeze-all policy may improve pregnancy outcomes in this subgroup [33].

The main limitation of our study is its retrospective design. Although all patients attended a single fertility clinic, the decision to choose GnRHa or hCG may have been influenced by each physicians’ intrinsic bias and subjective suspicion of the risk for OHSS, which increases heterogeneity and likelihood of confounding, particularly with the subgroup analysis. However, it is important to note that the baseline characteristics of patients in the different trigger groups were similar, which supports the validity of such a comparison. It is also important to note that more oocytes may have been recovered from the GnRHa group due the presence of a greater follicular mass at the time of triggering, but because it is not routine in our practice to measure all small follicles (<14 mm), this does not reflect a difference in patient characteristics. Furthermore, recruited patients underwent an elective freeze-all approach for PGT-A, as opposed to freeze-all due to hyper-response to ovarian stimulation, hence these results may not be representative of all hyper-responders, who may have other baseline characteristics that may affect reproductive outcomes. Those biases notwithstanding, our data still demonstrate GnRHa triggering may result in superior outcomes compared to hCG, and at very least, appears to be non-inferior as suggested by other investigators (Tannus et al., 2017). Finally, it is important to recognize that although differences in morphology and euploidy were evaluated, the exact mechanisms by which different triggers affect oocyte function and embryo competence requires further investigation.

Conclusion

Overall, the results of our study suggest that hyper-responders undergoing a freeze-all IVF protocol for PGT-A demonstrate improved outcomes after GnRHa compared to hCG triggering. In concordance with previous studies, GnRHa triggering resulted in a higher number of oocytes retrieved and significantly lower risk of OHSS. Our study is unique in that it is the first to examine the differences in euploidy rates subsequent to hCG vs GnRHa triggering. Furthermore, subsequent outcomes of frozen embryo transfers control for inclusion of only single euploid embryo transfers. The results demonstrate that women in the GnRHa trigger group have an overall higher number of euploid embryos, which may also translate to improved cumulative pregnancy rates, and women <35-years-old also benefited from a higher overall euploidy rate. Perhaps more interestingly, when transfer of euploid embryos are compared, pregnancy outcomes were better in blastocysts that resulted from GnRHa triggering, suggesting developmental factors other than chromosome copy number may be significantly influenced by the method used to induce the resumption of oocyte meiosis. These differences may reflect an improved developmental competence after GnRHa triggering due to physiologic induction of both an LH and FSH surge, promotion of granulosa cell-mediated function, or other undefined mechanisms that improve oocyte quality and subsequent healthy embryo development. Nevertheless, the effects of GnRHa on oocyte and embryo development need to be further evaluated in the context of prospective studies that will shed more light on the specific influence of different ovulation triggers on oocyte development, embryo competence, and overall pregnancy outcomes.

Compliance with ethical standards

Conflict of interest

The authors report no conflict of interest.