Premature progesterone elevation (PE) on the day of human chorionic gonadotropin (hCG) administration has garnered significant attention due to its association with a substantial decrease—approximately 10%—in clinical pregnancy rates following fresh embryo transfer [1]. The pathophysiology of this effect remains controversial, with debate centered on whether PE primarily compromises the endometrium or the embryo. A prevailing view, supported by recent microarray studies, is that PE’s detrimental impact is mainly on endometrial receptivity rather than embryo quality, as evidenced by PE-induced changes in endometrial gene expression [2, 3, 4].

Given the significant clinical impact of PE, several preventive strategies have been proposed. The most prevalent approach is a “freeze-all” policy, with embryos cryopreserved for subsequent frozen-thawed transfer cycles [5, 6]. However, fresh embryo transfer remains the preferred option for many clinicians and patients, as it shortens the time to pregnancy and live birth. Consequently, an attractive alternative is to prevent PE from occurring during the ovulation induction process itself. To this end, some investigators have sought to identify reliable predictors of PE. While multivariable models have linked factors such as gonadotropin (Gn) dose, estradiol (E2) on hCG day, patient age and certain other factors to PE risk [7, 8], and one study identified basal progesterone as a predictor in Gonadotropin-releasing hormone (GnRH) antagonist cycles [9], a controversy remains. To date, no study has comprehensively investigated whether different ovulation induction regimens themselves are independently associated with the occurrence of PE.

Our preliminary, unpublished data revealed that P levels on the day of hCG administration differed significantly in the same patients when they underwent different ovulation induction regimens for controlled ovarian stimulation (COS). This led us to hypothesize that the choice of regimen is an independent risk factor for PE. Furthermore, the definition of PE itself is evolving; early studies have adopted a lower threshold of 0.9 ng/mL [10, 11, 12], observing detrimental effects on pregnancy rates at this level, in contrast to the traditional 1.5 ng/mL cut-off [13]. Therefore, this study aimed to develop a predictive nomogram for PE occurrence on the hCG trigger day, using this contemporary 0.9 ng/mL threshold. The model incorporates baseline indexes, including the ovulation induction regimen, to identify at-risk patients prior to in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) treatment.

We developed and validated a predictive model for PE on the day of hCG administration, incorporating three key variables: the ovulation induction regimen, BMI, and basal P level. The model demonstrated strong performance in both calibration and discrimination. To facilitate its clinical application, we translated the model into a user-friendly nomogram, which was confirmed to have good clinical validity. This tool enables clinicians to easily assess a patient’s risk of PE based on basic characteristics, allowing for proactive intervention. For example, if a patient preparing for a fresh embryo transfer with a low BMI (22 kg/m²) and high basal P ($\ge$1.26 ng/mL), the model suggests avoiding the antagonist regimen to mitigate PE risk. If the antagonist protocol is clinically necessary, the Gn dose should be carefully controlled to achieve an appropriate number of follicles, thereby preventing the occurrence of PE.

The reported incidence of PE on the day of hCG administration in controlled ovarian hyperstimulation (COS) cycles ranges from 6% to 30% [16]. The high incidence observed in our study (30%) may be attributed to the low progesterone cut-off value used (0.9 ng/mL). To our knowledge, this is the first retrospective cohort study to investigate the association between the ovulation induction regimen and PE occurrence. We found that PE rates varied significantly between regimens, with the antagonist protocol exhibiting a markedly higher rate than the agonist protocol (51.4% vs. 24.4%). This finding is consistent with numerous studies [17, 18, 19] that have reported higher progesterone levels in antagonist protocols compared to GnRHa long protocols. One potential explanation relates to patient selection: GnRHa protocols are typically reserved for patients with high or poor ovarian response (e.g., Polycystic ovarian syndrome (PCOS)), whereas antagonist protocols are primarily used for normal responders. It has been suggested that high responders develop more dominant follicles and higher E2 levels on the hCG trigger day, which aligns with previous observations [19]. Another compelling mechanism involves endogenous LH activity. During COS, the extended downregulation of pituitary GnRH receptors by GnRHa suppresses endogenous LH. In contrast, the antagonist protocol, which omits this downregulation, allows for higher endogenous LH levels. This uninhibited LH environment promotes progesterone production [19] and is concurrent with the growth of numerous small follicles, leading to a significant rise in serum estrogen and progesterone levels.

We identified a significant association between basal P and the occurrence of premature ovulation PE on the day of hCG administration. While this predictive capacity of basal P was previously established in a study limited to the GnRH antagonist protocol (n = 1702) [9], our research confirms this finding across three different GnRH regimens in a larger dataset (n = 2689 cycles), thereby providing substantially greater statistical power. The underlying reason for this association, however, remains unclear. One initial hypothesis is that basal P primarily reflects adrenal production. In this scenario, a high baseline level, when combined with the P secreted by follicular granulosa cells during ovulation induction, would lead to a higher total serum P concentration. However, research by Fanchin et al. [20] challenges this simple additive model. They demonstrated that while adrenal suppression lowered baseline P, the net increase in P during COS remained unchanged. This finding confirms that the adrenal gland contributes to baseline levels but suggests that the subsequent rise in P and thus the occurrence of PE is driven almost exclusively by the ovarian response to exogenous gonadotropins [2]. In line with this conclusion, our study also confirmed that the total Gn dose administered and the number of mature follicles achieved were significantly associated with PE (see Supplementary Fig. 2).

BMI was another significant factor influencing the occurrence of PE. Our data revealed that patients with a low BMI were more susceptible to PE than those with a high BMI, a finding likely linked to differences in ovarian response. To understand this, it is important to consider the effects of BMI across the spectrum. Previous report indicate that patients with obesity typically require higher doses of Gn but ultimately achieve higher serum E2 levels and yield more retrieved oocytes compared to normal-BMI patients [21]. Since elevated E2 and a greater number of oocytes are associated with higher progesterone concentrations, this suggests a complex relationship. The reason why obese patients require more Gn remains unclear, but one hypothesis involves leptin. Higher BMI is associated with increased adipose tissue and, consequently, higher leptin levels. While leptin stimulates the GnRH pulse, chronically high levels may induce a state of leptin resistance, which could have an anti-gonadotropic effect, reducing the efficacy of exogenous Gn and thereby increasing the dose required for optimal follicular development [22]. Considering the sensitivity of the ovaries to Gn, patients with a lower BMI may have higher ovarian sensitivity and stronger reactivity, resulting in a greater number of follicles developing, ultimately leading to an increase in serum P levels.

It is well-established that P levels exceeding a certain threshold can reduce the clinical pregnancy rate [1]. However, the precise target of this negative effect, whether the embryo or the endometriumremains controversial. Our findings help to clarify this issue: we observed that PE on the day of hCG administration had no detrimental impact on embryo quality, but it was significantly associated with reduced endometrial thickness (12.02 $\pm$ 2.42 mm vs. 11.58 $\pm$ 2.49 mm, p 0.001). This points to a primary effect on the endometrium. Indeed, other studies have provided direct evidence supporting this conclusion. For instance, endometrial gene expression analyses have demonstrated that PE directly alters the endometrial molecular environment [4, 5]. Furthermore, some reports suggest that elevated plasma P levels during the late follicular phase can induce premature secretory transformations in the endometrium, thereby modifying the window of implantation [23, 24]. This premature advancement of the endometrium could lead to asynchrony between the developing embryo and the uterine lining, ultimately compromising implantation. Currently, there is no consensus on the cut-off value for PE. While most studies define this cut-off as 1.5 ng/mL, our research and others indicate that a PE value exceeding 0.9 ng/mL already negatively impacts the clinical outcomes of fresh embryo transfer cycles. We hope the results of this study will provide theoretical support for establishing a stricter PE cut-off and offer guidance for clinical practice.

In this study, we developed an objective and accurate nomogram to predict the risk of PE on the day of hCG administration. This tool offers a novel and user-friendly method for clinicians to assess patient risk by evaluating key factors such as basal P and BMI. By identifying high-risk patients, the nomogram can guide the selection of appropriate ovulation induction regimens and inform adjustments to the Gn dose, with the ultimate goal of preventing PE. However, large-scale, multicenter clinical trials are warranted to provide more robust evidence to support our findings.

hCG, human chorionic gonadotrophin; PE, progesterone elevation; CI, confidence interval; Gn, gonadotropin; E2, estradiol; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; COS, controlled ovulation stimulation; BMI, body mass index; AMH, anti-müllerian hormone; AFC, antral follicle count; LH, luteinizing hormone; T, testosterone; DCA, decision curve analysis; CIC, clinical impact curve.

The datasets generated and analyzed during the current study are not publicly available due to institutional and patient confidentiality policies, but are available from the corresponding author upon reasonable request.

KX, L-RZ and C-MY contributed in the conception and design; J-MZ, Y-HS and H-YY contributed in statistical analysis. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study and the protocols used were approved by the Ethics Committee of the Changzhou Maternal and Child Health Care Hospital (2025-k161-01). All patients sighed informed consent forms. All data used in this study were collected from patients that underwent standard IVF treatments. The study was carried out in accordance with the guidelines of the Declaration of Helsinki.

This study was supported by the Affiliated Changzhou Maternal and Child Health Care Hospital. We thank the members of the IVF Unit of the Affiliated Changzhou Maternal and Child Health Care Hospital.

This work was supported by the Basic & Applied Research Project of Changzhou, China (grant number CJ20179049) and the Special Fund for Clinical Research of Chinese Medical Association (grant number CMCC202206).

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/CEOG46145.