Evolution of serum progesterone levels in the very early luteal phase of stimulated IVF/ICSI cycles post hCG trigger: a proof of concept study

Abstract

Background

Studies have suggested that controlled ovarian hyperstimulation adversely affects endometrial receptivity due to advanced endometrial maturation. This adverse effect is mainly attributed to supraphysiological levels of both estrogen and progesterone identified in stimulated cycles. There is a paucity of published data investigating the very early luteal steroid profile following hCG trigger.

Aim of the study

This prospective, observational study was undertaken to determine the increase in serum progesterone levels after human chorionic gonadotrophin (hCG) trigger in stimulated IVF/ICSI cycles.

Materials and methods

This proof-of-concept study included 11 patients requiring ovarian stimulation for IVF/ICSI and who planned to avail of pre-implantation genetic screening with embryo vitrification of their biopsied embryos at blastocyst stage. For each study participant, five additional blood samples were drawn at the following specific times in the stimulation cycle, on the morning (10.00–12.00) of the assigned day to induce final oocyte maturation with hCG trigger, immediately prior to administration of hCG for final oocyte maturation, 1 h, 2 h, and 36 h post hCG trigger. A prediction model, the Gompertz curve, was used to determine serum progesterone levels at intervals between the 2 h post hCG trigger sample and the day of oocyte retrieval.

Results

Statistically significant increases in serum progesterone levels were identified following hCG administration as early as 1 h following trigger (P4 0.57 ng/ml, p < 0.05), 2 h following trigger (P4 0.88 ng/ml, p < 0.001) and on the day of oocyte retrieval (P4 9.68 ng/ml, p < 0.001). According to our prediction model, the Gompertz curve, the projected serum progesterone level at 4 h post trigger would have achieved a level of 1.45 ng/ml, 8 h post trigger of 3.04 ng/ml, and 12 h post trigger of 4.8 ng/ml. The very early and significant increases in serum progesterone following hCG trigger are clearly demonstrated in this study.

Conclusion

The endometrium is undoubtedly exposed to rapidly increasing serum progesterone levels post hCG trigger that would not be identified until much later in natural menstrual cycles.

Trial registration number: This study is registered with clinicaltrials.gov under the identifier NCT04417569.

Introduction

The establishment of a successful pregnancy in natural ovulatory cycles depends to a degree on endometrial development that is induced by a sequential and combined exposure to estrogen and progesterone, which are produced by the developing follicle prior to ovulation and by the corpus luteum after ovulation [1]. In assisted reproductive technology, a large bolus of hCG (5000–10,000 IU) is usually administered to mimic the midcycle surge of LH activity seen in a natural cycle to induce final oocyte maturation in stimulated cycles. In addition to the induction of follicular maturation, hCG trigger also directly stimulates endogenous progesterone production by the multiple corpora lutea during the early luteal phase [2–4]. Progesterone significantly reduces the production of luteinizing hormone (LH) via negative feedback mechanisms on the hypothalamus and pituitary gland [5]. As corpus luteum function is dependent on the frequency and amplitude of LH activity [1], the significant reduction in this gonadotropin following hCG trigger will result in corpus luteal dysfunction necessitating luteal phase support in stimulated IVF/ICSI cycles with fresh embryo transfer (ET), until pregnancy is well established [6, 7].

Despite the various advances and increasing success rates of assisted conception treatment in recent years, implantation continues to be a rate limiting step [8]. A contributory factor to this may be the early luteal steroid profile after hCG trigger [9, 10]. It is well recognized that the early luteal phase progesterone profile in IVF differs markedly from the progesterone profile of a natural cycle in which the peak progesterone level is attained approximately 6–8 days after ovulation which is the expected time of implantation [10, 11].

A previously published study confirmed that the early luteal phase endocrine profile is affected by the mode of final oocyte maturation in stimulated cycles [12]. Blood sampling in this study was performed on the day of final oocyte maturation, ovulation trigger + 1 day and ovum pick-up plus 5 days [12]. This study identified that in those who received hCG for final oocyte maturation, progesterone levels were significantly higher when compared to GnRH agonist trigger with and without standard luteal phase support [12]. In addition, serum FSH and serum LH levels were significantly lower following hCG trigger when compared to the levels seen following GnRH agonist trigger with and without standard luteal phase support [12]. The authors of this study concluded that induction of final oocyte maturation with a bolus of GnRH agonist in patients undergoing ovarian stimulation for IVF could be considered to be more physiological as opposed to the hCG trigger which resulted in significantly higher progesterone levels [12].

Studies suggest that the premature late follicular phase and early luteal phase rise in progesterone appearing after ovarian stimulation with exogenous gonadotropins and hCG trigger advances the window of implantation, causing asynchrony between the embryo and the endometrium, which may result in implantation failure [13–15]. Experimental evidence lends further credence with confirmation that prolonged exposure to hCG is detrimental to endometrial receptivity [16].

However, there are very little published data on the very early luteal steroid profile following hCG trigger and the limited data available are constrained due to limited serial blood sampling post hCG trigger ([9, 10, 17, 18]. The current study will advance our understanding of the very early luteal steroid profile following hCG trigger which is the cornerstone of follicular maturation in IVF/ICSI cycles. This study will take a step further in determining if hCG for final oocyte maturation may be the significant contributory factor to implantation rates remaining the rate-limiting step in fresh embryo transfer cycles [8].

Material and methods

This prospective observational study was conducted at ART Fertility Clinic Abu Dhabi, UAE, from February 2021 to May 2021. This proof-of-concept study included 11 patients presenting with primary/secondary infertility requiring ovarian stimulation for IVF/ICSI and who planned to avail of pre-implantation genetic screening with embryo vitrification of their biopsied embryos at blastocyst stage. Patients aged 18 to 42 years old, with a history of regular menstrual cycles with a cycle length of 26 to 35 days and a body mass index (BMI) of 18 to 30 kg/m² were included.

Patients with a history of a previous poor response to ovarian stimulation or who were expected to be poor responders in accordance with Bologna criteria and patients with a diagnosis of PCOS according to the Rotterdam criteria or a history of ovarian hyperstimulation syndrome in a previous ovarian stimulation were excluded [19, 20]. Patient information leaflets detailing the study were provided to all study participants and informed written consent obtained prior to study participation.

Details such as embryo development and pregnancy outcomes were not included in the study data as it has no bearing on the primary aim of this study which is to evaluate the early changes in serum progesterone following hCG trigger.

Ovarian stimulation protocol

Ovarian stimulation commenced on day 2 or day 3 of the menstrual cycle with recombinant FSH in a fixed GnRH antagonist protocol. The dose of recombinant FSH was individualized based on the patients’ ovarian reserve as assessed by a combination of parameters including female age, antral follicle count (AFC), and antimullerian hormone levels (AMH) [21]. Subcutaneous injections of recombinant FSH were administered daily at 8 p.m. In accordance with clinic protocol, the patients’ response to stimulation was monitored by a combination of serial ultrasound scans recording follicle size and number in conjunction with monitoring of serum levels of estrogen (E2), progesterone (P4), and follicle stimulating hormone (FSH). The dose of recombinant FSH was adjusted if deemed necessary based on the patient’s response to stimulation as assessed by a combination of ultrasound scans and serial endocrine monitoring. To achieve pituitary suppression, GnRH antagonist was commenced on day 5 of stimulation and administered daily at 08.00 until the day of final oocyte maturation. When three or more follicles attained the minimum required size of 17 mm, final oocyte maturation was induced by subcutaneous administration of 5.000 IU human chorionic gonadotropin (hCG) at 20.30 (12¹/2 h post antagonist injection at 08.00). Oocyte retrieval was performed 36 h following hCG administration in accordance with standard clinic protocol.

Blood sample and hormone assays

For each study participant, five additional blood samples of 10 ml venous blood were drawn at the following specific times in the stimulation cycle, on the morning (10.00–12.00) of the assigned day to induce final oocyte maturation with hCG trigger, immediately prior to administration of hCG for final oocyte maturation, 1 h, 2 h, and 36 h post hCG trigger. Blood samples were centrifuged and the obtained serum was divided into 3 vials/ aliquots prior to freezing at minus 20 °C. The samples were labeled with a unique code facilitating patient identification and ensuring patient anonymity. For the hormonal analysis, all frozen serum samples were placed in racks that allowed room-temperature air to circulate and were thawed at room temperature requiring 22 to 23 min to thaw. The samples were inverted at least 8 times to ensure proper mixing with no inadvertent leaks. All samples were measured and assayed at ART Clinical Laboratory, Abu Dhabi, in the same run.

Hormonal measurements

Progesterone analysis

ELECSYS® progesterone generation III assay is an electrochemiluminescence immunoassay (ECLIA) which uses sheep monoclonal antibodies due to their higher specificity towards progesterone. The measuring range is 0.47–56.4 ng/ml (ART clinical laboratory in-house validation). For the detection of analytical specificity, cross-reactivities towards other hormones were used with a maximum cross-reactivity of 3.93% towards 11-deoxycorticosterone and the minimum cross-reactivity of 0.001% towards Danazol (Elecsys Progesterone III Cobas Method Sheet, v.3).

Estradiol analysis

The Elecsys Estradiol III assay employs a competitive test principle using two monoclonal antibodies specifically directed against 17β estradiol. Endogenous estradiol released from the sample by mesterolone competes with the added estradiol derivative labeled with a ruthenium complex^a) for the binding sites on the biotinylated antibody. The measuring range is 21.2–2593 pg/ml (ART clinical laboratory in-house validation). Results are determined via a calibration curve which is an instrument specifically generated by 2-point calibration and a master curve provided via the reagent barcode or ebarcode (Elecsys Estradiol III Cobas Method Sheet, v.6).

LH analysis

The Elecsys LH assay employs two monoclonal antibodies specifically directed against human LH. The two specific antibodies used recognize particular conformations, with the biotinylated antibodies detecting an epitope constructed from both subunits whereas the antibody with the ruthenium complexa) label detects an epitope from the β subunit. As a result, the Elecsys LH assay shows negligible cross reactivity with FSH, TSH, hCG, hGH, and hPL. The measuring range is 0.99–197 mIU/ml (ART clinical laboratory in-house validation). Results are determined via a calibration curve which is an instrument specifically generated by 2-point calibration and a master curve provided via the reagent barcode or ebarcode (Elecsys LH Cobas Method Sheet, v. 22).

FSH analysis

The Elecsys FSH assay employs two different monoclonal antibodies specifically directed against human FSH. Cross reactivity with LH, TSH, hCG, hGH, and hPL is negligible. The measuring range is 1.817–99.060 mIU/ml (ART clinical laboratory in-house validation). Results are determined via a calibration curve which is an instrument specifically generated by 2-point calibration and a master curve provided via the reagent barcode or ebarcode (Elecsys FSH Cobas Method Sheet v.21).

Ethical approval

The study was approved by ART Fertility Clinical Ethical Committee under the identifying study code REFA052/REFA052a and registered with clinicaltrials.gov under the identifier NCT04417569.

Statistical analysis

Observational analysis

No formal sample size calculation was performed due to the scarcity of existing data on early luteal phase hormone levels. Prior to performing the main analysis, normality of the distribution was measured by employing the Jarque Bera and skewness test. Data including patient age, BMI, AMH, and parameters for the hormonal study (P4, LH, FSH, and E2) were recorded on a continuous scale. Data are summarized and presented as median, mean, and standard deviation values (SD). Mean values are further presented as minimum (min) and maximum (max) with confidence intervals at 95% (95% CI). Analysis of the variable data utilized the independent t-test, F-test for mean values and Kruskal–Wallis test for median values between compared groups to assess differences, depending on the normality of the variables under study. The mean differences for each patient’s serum progesterone values at different sample times were analyzed with a paired t-test. Analyses were performed using the STATA version 13.0 software.

Progesterone (P4) projection

We created a prediction model for serum progesterone levels at intervals between the 2 h post hCG trigger sample and the day of oocyte retrieval. Using statistical means, curve fitting was done using various equations to match our data points. The best fit equation was chosen for prediction purposes comparing adjusted R² values and matched with actual values available for post trigger times. P4 values (just before trigger, 1 h post trigger, 2 h post trigger) are used to project P4 level for every certain interval between trigger to day of oocyte retrieval. Every hour post trigger, a change in hormone level is assumed to determine a rate in change to facilitate an ongoing prediction for P4 levels. The curve fitting method of estimation used regression analysis with corresponding plots between P4 and sample time. The optimal curve fit equation of nonlinear regression function by least squares was chosen to fit the data trend over 4 time points. The Gompertz curve best suited the actual measured value (adj R² = 0.95). This is a type of mathematical model for a time series. Gompertz model was chosen comparing adjusted R² value (> 95%) and matched with actual values available for post hCG trigger times.

Results

A total of 11 patients were enrolled in this proof-of-concept study. The baseline characteristics of our study group are shown in Table 1. The mean age of the women was 33.3 years with a mean BMI of 23.8 kg/m². The mean AMH of the study group was 3.1 ng/ml with a range of 1.3–3.8 ng/ml. Outcomes following ovarian stimulation including the number of follicles measuring ≥ 11 mm recorded on the day of final oocyte maturation and the number of mature oocytes retrieved were recorded. Table 2 summarizes serum FSH, LH, E2, and P4 levels on the morning of trigger day, just prior to trigger, 1 h after trigger, 2 h after trigger, and on the day of oocyte retrieval. A mere 36 h following hCG trigger, the mean serum progesterone level recorded is 9.68 ng/ml.

Table 1.

Baseline characteristics and stimulation parameters (n = 11)

BMI, body mass index; AMH, anti Müllerian hormone; AFC, antral follicle count; E2, estrogen; P4, progesterone; FSH, follicle-stimulating hormone; LH, luteinizing hormone; COC, cumulus oocyte complex at oocyte retrieval; hCG, human chorionic gonadotrophin

Table 2.

Hormone levels’ from day of trigger to final oocyte retrieval

Min, minimum; Max, maximum; SD, standard deviation; CI, confidence interval; P4, progesterone; LH, luteinizing hormone; FSH, follicle stimulating hormone; E2, estrogen

Using a paired t-test to compare the serum progesterone levels at the five specified study sample times, a significant decrease was noted in serum progesterone levels recorded on the morning of trigger to immediately prior to administration of hCG trigger (p < 0.05) (Table 3). For the comparisons of the serum progesterone levels taken just prior to trigger to 1 h (p < 0.016), 2 h post trigger (p < 0.001), and on the day of oocyte retrieval (p < 0.001), a statistically significant increase in serum progesterone levels was seen for all mentioned comparisons (Table 3).

Table 3.

Test of difference in P4 level at different sample times

Footnote: Mean difference in P4 level at different sample times was computed using paired t-test, difference in mean value and corresponding p-values ( in parenthesis) are shown in each cell. − and + represent the decrease or increase respectively from level at time given in column to level at time given in row e.g. − 0.14(0.028) indicates a 0.14-point decrease in P4 level from day of trigger morning to just before trigger with p = 0.028

According to our prediction model, the Gompertz curve, the projected serum progesterone level at 4 h post trigger would achieve a level of 1.45 ng/ml, 8 h post trigger of 3.04 ng/ml, and 12 h post trigger of 4.8 ng/ml (Fig. 1).

Fig. 1.

Prediction of progesterone levels post hCG trigger

Discussion

Human chorionic gonadotropin (hCG) is commonly used in assisted reproduction technology in place of luteinizing hormone (LH) to trigger ovulation in controlled ovarian hyperstimulation (COH) [22]. This molecule is administered in the majority of controlled ovarian hyperstimulation regimens as natural preovulatory luteinizing hormone (LH) surges fail to occur and may exert direct effects on the endometrium as a consequence of stimulation of endometrial hCG/LH receptors [23].

For implantation to occur, a blastocyst must attach to and invade the endometrium under the influence of both estrogen and progesterone. Endometrial receptivity is driven by the secretory transformation of the endometrium under the influence of progesterone following estrogen exposure. It is well established that controlled ovarian hyperstimulation adversely affects endometrial receptivity due to supraphysiological levels of ovarian steroids and or untimely modifications in E2, progesterone, and androgen production which may explain the suboptimal implantation rates often associated with COH and fresh embryo transfer [22–26]. Further studies attribute both premature endometrial exposure to progesterone prior to and after hCG administration to adversely affecting endometrial receptivity [26, 27]. Despite the endometrium and implantation failure being the focus of investigation and discussion for many years, there is an obvious gap in our knowledge pertaining to the very early luteal steroid profile following hCG administration.

It is well accepted that endometrial morphology changes under the influence of ovarian sex steroids during the menstrual cycle. As far back as 1937, the effects of progesterone and estrogen on the endometrium were assessed. Rock and Bartlett reported that it may be possible to determine both the duration of progesterone secretion and the estrogen-progesterone balance from the presence or absence of sequential changes in the endometrium [28]. Subsequently, a paper was published identifying endometrial histology as the most accurate method of diagnosing corpus luteal insufficiency and established the criterion of a lag in endometrial histology of 2 or more days between expected and observed findings in at least two cycles [29]. A short time later in 1950, Noyes et al. documented the daily histological changes occurring under the influence of progesterone and concluded that endometrial dating gives an estimate of quantitative progesterone effect which reflected both duration and amount of progesterone secretion [30]. Previously published studies suggest that the early luteal phase rise in progesterone appearing after ovarian stimulation with exogenous gonadotropins and hCG trigger advances the window of implantation causing asynchrony between the embryo and the endometrium which may result in implantation failure [13–15].

This study clearly demonstrates the very early and significant increase in serum progesterone levels following hCG administration. Statistically significant increases in serum progesterone levels were identified following hCG administration as early as 1 h following trigger (P4 0.57 ng/ml, p < 0.05), 2 h following trigger (P4 0.88 ng/ml, p < 0.001), and on the day of oocyte retrieval (P4 9.68 ng/ml, p < 0.001). This data proves that the endometrium post hCG trigger is exposed to significantly increasing progesterone levels as early as 1 h following trigger inducing endometrial morphological and biochemical alterations that would not be identified until much later in natural menstrual cycles. This early progesterone exposure immediately posts hCG administration for final oocyte maturation and the possible early rise in serum progesterone in the late follicular phase of stimulation would clearly explain the earlier described endometrial advancement identified in stimulated cycles [13–15].

Kolibianakis et al. identified endometrial advancement on the day of oocyte retrieval in all cycles stimulated with recombinant FSH, GnRH antagonists, and hCG and no pregnancies were achieved when the endometrium was advanced for more than 3 days [15]. The take-home message of this study is clear in that the chance of pregnancy is significantly decreased in the presence of extreme endometrial advancement at the time of oocyte retrieval. The authors of this study attributed this finding to ovarian stimulation but admitted that a modifying effect of hCG on endometrial advancement observed at oocyte retrieval could not be excluded [15].

Van Vaerenbergh et al. conducted a micro-array-based study that investigated the impact of high serum progesterone levels on endometrial gene expression profiles [31]. The endometrial gene expression profile on the day of OPU was related to the progesterone concentration on the day of hCG administration. This was the first study to demonstrate a distinct difference in endometrial gene expression profile between patients with a serum progesterone level above and below the threshold of 1.5 ng/ml on the day of hCG administration [31]. This study clearly demonstrated accelerated endometrial maturation during the pre-receptive early secretory phase in those patients with high progesterone concentrations [31].

There is now substantial evidence available to suggest that in stimulated cycles, endometrial maturation is advanced on the day of oocyte retrieval and patients with a serum progesterone level of 1.5 ng/ml or above on the day of final oocyte maturation have different endometrial gene expression profiles [13, 15, 31, 32].

This progesterone level seems to represent the critical threshold above which a negative effect on the ongoing pregnancy rate in fresh IVF cycles can be observed [33]. The adverse effect of elevated peripheral progesterone levels in the late follicular phase and the rapidly increasing progesterone levels post hCG administration appear to be exerted on the endometrium and the window of implantation, which may lead to asynchrony between the endometrium and the embryo [33].

The obvious question now would be whether an elevated serum progesterone level on the day of final oocyte maturation will exacerbate the effect of hCG trigger on early luteal phase progesterone levels.

A recently published study examined the early luteal hormonal profile in IVF patients triggered with hCG [18]. It was interesting to see that the mean serum progesterone level prior to hCG in the study group was 1.46 ng/ml (SD ± 1.07 ng/ml) [18]. Of the 160 patients recruited into the study, 8 patients (5%) had a pre-hCG serum progesterone level ≥ 3 ng/ml [18]. The authors demonstrated that late follicular phase progesterone concentrations were significantly associated with progesterone concentrations measured at all time points until day OPU + 1 [18]. The mean serum progesterone level at hCG + 36 h was 14.06 ng/ml (SD ± 7.02 ng/ml). In contrast, the pre-hCG mean serum progesterone level is significantly lower in the herein presented study with a level of 0.61 ng/ml (SD ± 0.18 ng/ml). Premature early progesterone rise described in the late follicular phase was not seen as we adopt the practice of reducing the FSH drive to the granulosa cells during the follicular phase [34, 35]. It is interesting to note that studies have shown that stimulation with hCG/LH activity does not prevent a premature progesterone rise during the late follicular phase of stimulation [36–38]. In addition, there is a significant difference in the mean serum progesterone level at hCG + 36 h in our study 9.68 ng/ml ± 2.29 ng/ml as opposed to 14.06 ng/ml ± 7.02 ng/ml. This confirms that follicular LH sensitivity is reflected by late follicular phase progesterone concentrations, which then determine the early luteal phase response [18, 39]. Exogenous FSH administration results in supraphysiological levels of FSH which induce an abundance of LH receptors on granulosa cells causing the follicles to become hypersensitive to LH-like activity (hCG trigger) [37, 39].

Despite the limited number of study participants, which could be seen as a limitation of the study, the statistically significant increase in progesterone levels identified as early as 1 h following hCG trigger was clearly demonstrated and proved beyond doubt the concept of significantly increased serum progesterone levels post hCG trigger (Table 3). Using the Gompertz equation, we can statistically reliably predict serum progesterone levels to be as high as 4.8 ng/ml as early as 12 h following hCG trigger (Fig. 1). On the day of oocyte retrieval, the mean serum progesterone level was 9.68 ng/ml with a range of 7.56–15.54 ng/ml. A previously published paper examined the dynamics of ovarian and pituitary hormone changes during the midcycle of a spontaneous natural cycle [40]. A mean P₄ level of 0.55 ng/ml was documented at the time of the initiation of the LH surge [40]. A rapid rise in P4 levels was identified 36 h after the onset of the LH surge with a doubling time of 11.7 h [40]. In contrast, according to our prediction model, the Gompertz curve, the projected serum progesterone level at 4 h post trigger was 1.45 ng/ml and a mere 4 h later had more than doubled to a value of 3.05 ng/ml and at 12 h post trigger 4.8 ng/ml. The P4 level had more than tripled in value from 4 to 12 h post trigger as compared to the documented doubling time of 11.7 h in P4 level seen 36 h after the onset of LH surge in the natural cycle. The rapidity of the serum progesterone rise post hCG trigger is notable. The endometrium is undoubtedly exposed to rapidly increasing serum progesterone levels post hCG trigger that would not be identified until much later in natural menstrual cycles.

Conclusions

To the best of our knowledge, this is the first study to examine the very early luteal phase progesterone levels post hCG trigger and therefore contributes significantly to our understanding of the significance of hCG in ovarian stimulation. This study has clearly demonstrated the early and significant increases in serum progesterone following hCG trigger resulting in prolonged exposure of the endometrium to high progesterone levels. It can be assumed that the increase in progesterone levels after hCG trigger is even more pronounced in cases with high progesterone levels on the day of final oocyte maturation. Therefore, further studies are warranted to differentiate in detail the effect on the endometrial receptivity attributed to either the premature progesterone rise during ovarian stimulation, the administration of hCG for final oocyte maturation, or a combination of both.

Author contribution

Carol Coughlan: drafting of study protocol and manuscript

Raquel Vitorino: patient recruitment and manuscript drafting

Laura Melado: patient recruitment

Shieryl Digma: collating and presentation of study data

Junard Sibal: processing of study blood samples/drafting of manuscript

Rachana Patel: statistical analysis of data/drafting of manuscript

Barbara Lawrenz: drafting of study protocol and manuscript

Human Fatemi: conceptualization of study/recruitment/manuscript drafting

Declarations

Conflict of interest

The authors declare no competing interests.