Introduction

Despite attempts at suppressing endogenous gonadotropins with gonadotropin-releasing hormone (GnRH) agonists and antagonists, progesterone elevation (PE) at the time of trigger is common during controlled ovarian stimulation. The incidence of PE on the day of trigger has been demonstrated to be as high as 13–46% in GnRH agonist cycles [1] and 23% in GnRH antagonist cycles [2]. These subtle increases in serum progesterone are concerning, as PE on the day of trigger has been associated with decreased implantation [3], clinical pregnancy [4, 5], and live birth rates [6] following fresh embryo transfer when compared to cycles which proceeded to fresh embryo transfer following low progesterone levels on the day of trigger.

The mechanism by which PE exerts these deleterious effects is unclear; previous studies have attributed these adverse pregnancy outcomes to asynchrony between the endometrium and the embryo [6,7,8] or impairments in oocyte quality [9, 10]. In natural menstrual cycles, the presence of late follicular phase progesterone is essential for follicular development, ovulation, and endometrial receptivity [6]. In contrast, the hyperstimulated endocrine milieu which results from controlled ovarian stimulation may cause premature PE, thereby maturing the endometrium beyond the optimal window of implantation at the time of embryo transfer.

Previous studies have demonstrated a decline in IVF outcomes and live birth rates when serum progesterone is elevated ≥ 1.5 ng/mL at the time of trigger [2, 11]. Although these data have prompted several IVF clinics to adopt a freeze-all approach if serum progesterone is elevated on the day of trigger, there is a paucity of data on the impact of progesterone elevation in the early and mid-follicular phases.

Due to previous research linking elevated serum progesterone levels at the beginning of ovarian stimulation with inferior ART outcomes [12], we believed that a single assessment of serum progesterone at the time of trigger may be an insufficient proxy of the endometrium’s exposure to progesterone. Consequently, our center began measuring serum progesterone levels from stimulation start in 2015. Taking advantage of this fact, the objective of this study is to characterize early and mid-follicular phase PE (defined as serum progesterone ≥ 1.5 ng/mL) during IVF stimulation and its effects on implantation, clinical pregnancy, and live birth rates in fresh IVF cycles. We hypothesize that PE ≥ 1.5 ng/mL at any point during stimulation prior to trigger will be associated with decreased implantation, clinical pregnancy, and live birth rates.

Methods

This study was approved by the institutional review board at Mass General Brigham (Protocol # 2016P001715). Approval and waiver of written informed consent to retrieve and analyze data were obtained from the institutional review board at Mass General Brigham.

Inclusion criteria

This retrospective cohort study included patients undergoing fresh IVF and IVF/ICSI cycles at the Brigham and Women’s Hospital from October 1, 2015 to June 30, 2021. Data were collected from a prospectively maintained departmental database. Cycles had estradiol and serum progesterone (PR) levels measured with each follicular monitoring ultrasound visit beginning on day 7 of stimulation until the day of trigger. Cycles proceeded to fresh transfer of day 3 or day 5 embryos if TPR < 2.0 ng/mL. Cycles utilizing letrozole, donor egg, gestational carriers, and PGT tested embryos were excluded.

Ovarian stimulation

Ovarian stimulation protocols included GnRH antagonists, GnRH agonists for downregulation, estradiol priming, low-dose GnRH agonist flare protocols, or minimal stimulation with clomiphene citrate [12, 13]. Cycles had serum estradiol and PR levels measured with each follicular monitoring ultrasound visit beginning on day 7 of stimulation until the day of trigger. When two or more follicles were ≥ 18 mm in diameter, 5000 IU hCG, 10,000 IU hCG, or dual trigger of 40 mg leuprolide with 1500 IU hCG was utilized for final oocyte maturation. Oocyte retrieval was performed 36 h after the trigger injection. Fertilization was achieved with either conventional IVF or intracytoplasmic sperm injection (ICSI) as clinically indicated. Patients triggered with hCG started Crinone 8% intravaginal gel daily 48 h after retrieval. Cycles with a dual trigger started oral Estrace 3 mg BID and Crinone 8% intravaginal gel 48 h after oocyte retrieval for luteal phase support.

Laboratory protocols and embryo grading

Gametes and embryos were cultured in a humidified incubator maintained at 37 °C under an atmosphere of 5–6% CO2, 5% O2, and the balance of N2. IVF or ICSI was performed 4–6 h or 3–5 h after egg retrieval. The fertilization check was performed 16–18 h after insemination. Zygotes with 2 pronuclei (2PN) were cultured in 25 uL microdrops of a single-step medium (Global Total, IVF OnLine, Guelph, Ontario, Canada) under mineral oil.

Embryos were evaluated on day 3 between 66- and 69-h post-insemination. Morphological grading was based on the number of blastomeres, the extent of fragmentation, and the degree of asymmetry, as previously described [14]. Blastocyst morphology was evaluated on day 5 and scored according to the stage of development and to the quality of the inner cell mass (ICM) and trophectoderm (TE) [15]. Each embryo was scored from 1 to 9 to indicate its development at the time of grading: 1 denoted an arrested embryo, 2 an embryo with < 50% compaction, 3 a full morula, 4 an early blastocyst, 5 an embryo in which > 50% of the volume is occupied by the blastocyst, 6 to 8 with increasing blastocoel formation and expansion, and 9 a fully hatched blastocyst. A grade of A–D and a–d (best to worst) indicated the quality of the ICM and TE, respectively.

Ultrasound-guided embryo transfer was performed on day 3 or 5 depending on the number and quality of embryos available. Serum hCG levels were assessed 14 days after embryo transfer, followed by ultrasound confirmation of an intrauterine pregnancy in all pregnant patients. Patients with a confirmed intrauterine pregnancy continued progesterone supplementation until 10 weeks of gestation.

Progesterone assay

Serum progesterone levels were measured with a solid-phase competitive chemiluminescent enzyme immunoassay. The lower limit of detection for the assay was 0.05 ng/mL, and the analytical sensitivity of the assay was 0.03 ng/mL. Intra-assay and inter-assay coefficients of variation were 6.5 and 6.9%, respectively.

Primary and secondary outcomes

The primary outcome was live birth, defined as a live-born infant after 22 weeks of gestation. Secondary outcomes included implantation rate and clinical pregnancy. Implantation rate was defined as the number of gestational sacs visualized on ultrasound divided by the number of embryos transferred. Clinical pregnancy was defined as the presence of fetal cardiac activity visualized on ultrasound.

Statistical analysis

Patients were categorized into two exposure groups based on PR levels at any point during stimulation prior to the day of trigger: PR < 1.5 ng/mL (low PR group) and PR ≥ 1.5 ng/mL (high PR group). Means and proportions were generated for continuous variables; frequencies and proportions were generated for categorical variables. Relative risks (RR) and 95% confidence intervals (CI) were produced using Poisson regression for counts, Poisson regression with an offset for rates, and log binomial regression for dichotomous outcomes. Regression models for cycle outcomes were adjusted for age, AMH, day 3 FSH, BMI, infertility diagnosis, stimulation protocol, and serum estradiol on the day of trigger. Regression models for transfer outcomes were adjusted for all previously stated covariates in addition to trigger type (hCG versus dual trigger), number of oocytes retrieved, embryo quality, and day of transfer. Generalized estimating equations were used to account for multiple cycles from the same patient. An alpha of 0.05 was considered statistically significant. All statistical analyses were performed with SAS® version 9.4 (Cary, NC, USA).

Results

A total of 6961 cycles met inclusion criteria: 5393 (77.4%) in the low PR group (PR < 1.5 ng/mL) and 1568 (22.6%) in the high PR group (PR ≥ 1.5 ng/mL). The two groups were similar with respect to age at cycle start, BMI, infertility diagnosis, use of HMG, and day 3 FSH (Table 1). The stimulation protocols employed in both groups were comparable: antagonist protocols comprised 76.6 and 80.74% of cycles in the low PR and high PR groups, respectively.

Table 1 Demographic characteristics of patients in the low and high PR groups
Full size table

A total of 3757 cycles proceeded to embryo transfer on day 3 or day 5 if TPR < 2.0 ng/mL: 3341 (88.9%) in the low PR group and 416 (11.1%) in the high PR group. The two groups were similar with respect to age at cycle start, BMI, infertility diagnosis, use of HMG, day 3 FSH, and stimulation protocol (Table 2).

Table 2 Demographic characteristics of patients in the low and high PR groups who underwent a fresh embryo transfer
Full size table

When assessing pregnancy outcomes by serum progesterone levels prior to the day of trigger, the high PR group had significantly lower IR (RR 0.75; 95% CI 0.64–0.88), CPR (aRR 0.74; 95% CI 0.64–0.87), and LBR (aRR 0.71; 95% CI 0.59–0.85) when compared to the low PR group (Table 3).

Table 3 Pregnancy outcomes by PR group
Full size table

When assessing pregnancy outcomes stratified by serum progesterone on the day of trigger (TPR), 3398 cycles had TPR < 1.5 ng/mL and 359 had TPR ≥ 1.5 ng/L (Table 4). Of the cycles with TPR < 1.5 ng/mL, 3341 (98.3%) were in the low PR group and 57 (1.7%) were in the high PR group. In cycles with TPR < 1.5 ng/mL, the high PR group had clinically lower IR (16.8% vs 23.3%), CPR (28.1% vs 36.0%), and LBR (22.8% vs 28.9%) when compared to the low PR group (Table 4).

Table 4 Pregnancy outcomes from fresh embryo transfers stratified by TPR
Full size table

Discussion

This study confirms our hypothesis that progesterone elevation ≥ 1.5 ng/mL at any point during stimulation confers a poor prognosis following fresh embryo transfer, as the high PR group experienced significantly lower IR, CPR, and LBR when compared to the low PR group.

Although our center began measuring serum progesterone levels during ovarian stimulation in 2015 as a function of logistical ease, the results of this study indicate that it may be advantageous to check serum progesterone levels for patients intending a fresh embryo transfer, as there was a clinically notable decline in IR, CPR, and LBR in the high PR group even if TPR < 1.5 ng/mL (Table 4). Due to the potential for asynchrony between the embryo and endometrium in cycles with TPR < 1.5 ng/mL but serum progesterone ≥ 1.5 ng/mL prior to trigger, these patients may benefit from a freeze-all approach.

In cycles with TPR < 1.5 ng/mL, the high PR group had clinically lower IR (16.8% vs 23.3%), CPR (28.1% vs 36.0%), and LBR (22.8% vs 28.9%) when compared to the low PR group (Table 4). Although we believe that these results are clinically notable and have the potential to impact medical practice, statistical significance between the low PR and high PR groups was not attained likely due the low number of transfers in the high PR group. The discrepancy in the number of transfers in the low PR group (N = 3341) and the high PR group (N = 57) in cycles with TPR < 1.5 ng/mL is largely due the fact that providers at our center began adopting a freeze-all approach if serum progesterone ≥ 1.5 ng/mL at any point during stimulation due to the observed decrease in live birth rates. It is essential to note that high PR during stimulation with TPR < 1.5 ng/mL was not a rare occurrence at our center, as 66.5% (135/203) of cycles with TPR < 1.5 ng/mL resulted in a freeze-all strategy due to elevated progesterone during stimulation despite intending a fresh transfer at cycle start.

Critics may argue that routine monitoring of serum progesterone levels prior to trigger is costly and, thus, a practice which should be discouraged. A serum progesterone assay at our center costs $30.00 USD [16] and ranges from $27.00 to $45.00 USD at private laboratories [17]. If patients have daily serum progesterone measurements from day 7 to day of trigger, this will cost an additional $210–420, on average, depending on the length of the cycle. Although this may marginally increase the total cost of the cycle, we believe this is negligible when compared to the financial and emotional toll of a failed fresh transfer and costs associated with a subsequent fresh stimulation.

Although many IVF centers have ameliorated the decline in pregnancy rates with a freeze-all approach when TPR ≥ 2 ng/mL [18], the progesterone threshold at which a freeze-all strategy is indicated remains controversial. Our study demonstrated a clinically notable decrease in LBR between the low PR and high PR groups when TPR < 1.5 ng/mL (28.9% vs. 22.8%, Table 4) and a statistically significant decrease in LBR (21.7%, aRR 0.69, 95% CI 0.57–0.85) when serum progesterone levels are persistently elevated ≥ 1.5 ng/mL during stimulation and on the day of trigger. The amplified decline in IR, CPR, and LBR seen in cycles with progesterone levels persistently elevated ≥ 1.5 ng/mL during stimulation and on the day of trigger is likely due to the duration of endometrial exposure to elevated progesterone, as has been previously published [5].

A primary strength of this study is the large sample size, allowing us to assess the impact of progesterone elevation in the early and mid-follicular phases on pregnancy outcomes. By conducting the study at a single large academic center utilizing a singular solid-phase competitive chemiluminescent enzyme progesterone immunoassay, we capitalized on low inter-assay and intra-assay variability, allowing for the appropriate comparison between exposure groups. One limitation of this study is its retrospective design; thus, future prospective studies assessing pregnancy outcomes at varying follicular phase progesterone thresholds and durations of progesterone elevation are warranted.

Conclusion

In fresh IVF cycles in which TPR < 1.5 ng/mL, progesterone elevation ≥ 1.5 ng/mL at any point in time prior to trigger negatively impacts IR, CPR, and LBR. These data support routine testing of serum progesterone, as these patients may benefit from a freeze-all approach.