Elevated serum estradiol levels on embryo transfer day predict higher pregnancy success in HRT-FET cycles.

A total of 175 eligible women undergoing HRT-FET cycles were analyzed based on demographic characteristics and reproductive outcomes. Patients were divided into two groups according to the decision tree analysis illustrated in Fig.  1 .A: Group 1 (estrogen level ≤ 201 pg/mL) and Group 2 (estrogen level > 201 pg/mL). Fig. 1 Decision tree models generated using the CART (Classification and Regression Tree) algorithm to predict clinical pregnancy based on serum estrogen levels. A Tree using estrogen levels on the first day of progesterone administration as the primary variable. B Tree using estrogen levels on the day of embryo transfer as the primary variable Decision tree models generated using the CART (Classification and Regression Tree) algorithm to predict clinical pregnancy based on serum estrogen levels. A Tree using estrogen levels on the first day of progesterone administration as the primary variable. B Tree using estrogen levels on the day of embryo transfer as the primary variable Our findings revealed that an estrogen level cut-off of 201 pg/mL, with a 95% CI for the mean of 459.597 ± 162.402 (ranging from 297.195 to 621.999), on the day of embryo transfer yielded favorable pregnancy outcomes. Furthermore, Fig.  1 B demonstrated that considering both estrogen and progesterone levels on the day of embryo transfer, if the estrogen level on the first day of progesterone administration is not less than 144.7 pg/mL (with a 95% CI for the mean of 270.207 ± 120.618), satisfactory clinical pregnancy rates will be achieved. Both models identified a threshold of 201 pg/mL, which emerged independently from the CART process. This consistency is due to mutual referencing of the two time-point variables in the respective models. To evaluate the discriminative ability of serum estrogen levels, two approaches were employed. First, a decision tree (CART) model identified a clinically relevant threshold of 201 pg/mL for estrogen levels on the day of embryo transfer, which demonstrated strong predictive performance. When this threshold was applied in ROC curve analysis as a binary classifier (≤ 201 vs. >201 pg/mL), the model achieved an AUC of 0.72 (95% CI: 0.64–0.80), with 90% sensitivity and 70% specificity, supporting its clinical utility (Fig.  2 ; Table  1 ). Fig. 2 Receiver operating characteristic (ROC) curve for prediction of clinical pregnancy using serum estrogen levels on the 1 st and 4th day of progesterone administration Receiver operating characteristic (ROC) curve for prediction of clinical pregnancy using serum estrogen levels on the 1 st and 4th day of progesterone administration Table 1 Summary of specificities related to the ROC curve Test Result Variable(s) Area Cut point sensitivity specificity 95% C.I Lower Bound Upper Bound E2 of the 1 st day of P administration 0.80 146 0.76 0.68 0.73 0.87 E2 of the day of transfer 0.72 202 0.90 0.70 0.64 0.80 Summary of specificities related to the ROC curve To further assess whether continuous estrogen levels alone (without thresholding) differed in predictive value between Day 1 and Transfer, we performed a DeLong test comparing their ROC curves. The AUCs were 0.52 and 0.50, respectively (Z = 0.394, p  = 0.694), indicating no statistically significant difference in predictive performance when using raw values. This finding emphasizes that the discriminative power lies in the thresholded decision rule, rather than the continuous hormone levels alone. Thus, the CART-derived cutoff enhances interpretability and clinical applicability over the raw values. The mean age of patients was 34.73 ± 5.85 years, and the mean BMI was 23.5 ± 4.2 kg/m². Among the patients, 43.42% had estrogen level ≤ 201 pg/mL, while 56.58% had estrogen level > 201 pg/mL Patients in the two estrogen groups were comparable regarding BMI, the number of embryos transferred, endometrial thickness, and serum levels of AMH and FSH. However, a significant difference in age was observed, suggesting an inverse relationship with increasing age, estrogen levels tended to decrease (Table  2 ). Table 2 Comparison of demographic and clinical outcomes between groups Estrogen (transfer day) Estrogen ≤ 201 Estrogen > 201 p -value Parameters (n= 76) (n= 99) Age (years: mean ± SD) 35.98 ± 6.01 33.76 ± 6.84 0.02* BMI 23.8 ± 3.2 23.2 ± 3.3 0.71 Duration of infertility (years: mean ± SD) 7.06 ± 5.17 6.84 ± 4.25 0.75** Endometrial thickness (mean ± SD) 7.47 ± 0.93 7.68 ± 0.94 0.12** FSH (mean ± SD) 4.37 ± 2.35 4.38 ± 2.17 0.81** AMH (mean ± SD) 3.34 ± 2.90 3.31 ± 2.96 0.92** Biochemical pregnancy (n (%)) 16(21.05%) 69(70%) < 0.001*** Clinical pregnancy (n (%)) 2(2.63%) 62(62.62%) < 0.001*** Early pregnancy loss (n (%)) 14(87.5%) 5(7.25%) < 0.001*** BMI  Body mass index, FSH  Follicle-stimulating hormone, AMH  Anti-Müllerian Hormone * Independent T Test ** Mann-Whitney Test *** Fisher’s Exact Test Comparison of demographic and clinical outcomes between groups Duration of infertility (years: mean ± SD) BMI  Body mass index, FSH  Follicle-stimulating hormone, AMH  Anti-Müllerian Hormone * Independent T Test ** Mann-Whitney Test *** Fisher’s Exact Test Overall, the biochemical pregnancy rate was 48.57% (85/175), the clinical pregnancy rate was 36.57% (64/175), and the early pregnancy loss was 22.35% (19/85). The biochemical pregnancy rate was significantly higher in the estrogen level > 201 pg/mL group compared to the estrogen level ≤ 201 pg/mL group (70.0% [69/99] vs. 21.1% [16/76]; p   201 pg/mL group compared to the estrogen level ≤ 201 pg/mL group (62.62% (62/99) versus 2.63% (2/76), ( p  < 0.001). Furthermore, the early pregnancy loss rate was significantly lower in the high estrogen group compared with the low estrogen group (7.25% [5/69] vs. 87.5% [14/16]; p  < 0.001; Table  2 ). The model’s resubstituting risk estimate was 0.229 (SE: 0.046), and the 10-fold cross-validated risk was also 0.229 (SE: 0.046), indicating stable model performance and minimal overfitting. This corresponds to an overall classification accuracy of 77.1%. Table  3 presents both unadjusted and adjusted binary logistic regression analyses assessing the association between various clinical and hormonal parameters and clinical pregnancy. It presents the odds ratios (ORs) with the corresponding 95% confidence intervals (CIs) and p-values for each parameter included in the regression model. Clinical pregnancy was the main outcome of the study. The parameters incorporated in the binary regression analysis were age, FSH level, AMH level, endometrial thickness, progesterone levels on the day of embryo transfer, estrogen levels on the day of embryo transfer, and estrogen levels on the first day of progesterone treatment. Table 3 Binary regression model (Unadjusted and Adjusted) Dependent Variable Independent Variable B (Unadj.) p -value (Unadj.) OR (Unadj.) 95% CI for OR (Unadj.) B (Adj.) p -value (Adj.) OR (Adj.) 95% CI for OR (Adj.) Clinical Pregnancy Age −0.125 0.018* 0.882 0.795–0.979 −0.180 0.190 0.840 0.650–1.090 FSH 0.146 0.322 1.157 0.867–1.545 −0.060 0.800 0.940 0.580–1.520 AMH −0.076 0.605 0.927 0.694–1.237 0.010 0.930 1.010 0.740–1.400 E2 (1st day of P treatment) 0.001 0.631 1.001 0.996–1.006 0.000 0.690 1.000 0.990–1.000 E2 (day of transfer) 0.000 0.849 1.000 0.999–1.001 0.010 0.040* 0.990 1.010–1.180 P (day of transfer) −0.032 0.558 0.969 0.872–1.077 0.450 0.050* 1.570 1.100–2.730 Endometrial thickness 0.018 0.942 1.018 0.632–1.639 −0.010 0.990 0.990 0.290–3.390 P -values < 0.05 are considered statistically significant OR Odds Ratio, CI Confidence Interval, Unadj . = Unadjusted, Adj . = Adjusted Binary regression model (Unadjusted and Adjusted) P -values < 0.05 are considered statistically significant OR Odds Ratio, CI Confidence Interval, Unadj . = Unadjusted, Adj . = Adjusted The regression analysis highlights that while age is a significant predictor of clinical pregnancy in the unadjusted model, its effect is no longer significant after adjusting for other relevant variables. A stratified analysis by age (< 35 and ≥ 35 years) was performed to assess whether the association between estrogen levels and clinical pregnancy differed across age groups. For women under 35 years ( n  = 41), the Pearson chi-square test did not show a significant association between estrogen level groups and clinical pregnancy ( p  = 0.516). Similarly, in women aged 35 years or older ( n  = 42), no significant association was observed ( p  = 0.501). Significant p-values were identified for estrogen levels ( p  = 0.04) and progesterone levels ( p  = 0.05) on the day of embryo transfer in patients with estrogen levels > 201 pg/mL (Table  3 ).

This retrospective cohort study was conducted among 210 infertile women who underwent frozen embryo transfer at the Infertility Center of Shahid Mohammadi Hospital. The study received approval from the Institutional Review Board and Ethics Committee of Hormozgan University of Medical Sciences. This was a non-interventional, retrospective observational study and was not registered as a clinical trial. Therefore, a clinical trial registration number is not applicable. Eligible participants were women aged ≤ 40 years, with an endometrial thickness of ≥ 7 mm, optimal serum P4 levels from the day of transfer, and the ones who received of two or three cleavage-stage embryos transfer. In this study, optimal P4 levels on embryo transfer day were defined as P4 levels within ranges of ≥ 11ng/mL and ≤ 25ng/mL. Exclusion criteria included the presence of uterine cavity lesions and malformations, uterine fibroids, endometriosis, hydrosalpinx, chromosome abnormalities, history of thrombosis, contraindications for estrogen use, and embryos subjected to preimplantation genetic testing for aneuploidy (PGT-A). A total of 175 cycles met these inclusion and exclusion criteria. All patients received hormone replacement therapy (HRT) for endometrial preparation with pituitary downregulation using a GnRH agonist. Specifically, oral estradiol valerate treatment began on the second or third day of the menstrual cycle at a dose of 2 mg every 8 h. After 10–12 days of estradiol administration, if an ultrasound showed optimal endometrial thickness (≥ 7 mm with a triple-line pattern), estradiol valerate was continued at the same dosage, and intramuscular administration of 50 mg progesterone was started. E2 and P4 levels were measured on the 1 st and 4th days of progesterone administration. Embryo transfer was scheduled on the 4th day of progesterone administration. If serum P4 levels were within optimal ranges on the day of cleavage embryo transfer, a standard luteal phase support (LPS) regimen consisting of 50 mg progesterone intramuscularly with estradiol 6 mg orally was continued until gestational week 12 in pregnant patients. A clinical pregnancy was defined as an intrauterine pregnancy with a heartbeat at week 7 of gestation. On the first day and 4th day of progesterone treatment, a blood sample was taken, and serum estrogen (E2, ng/mL) was analyzed. Estrogen was assessed by using chemiluminescence (ADVIA Centaur XP™ Automated Chemiluminescence System). The intra and inter-assay coefficients of variation for the E2 determinations were less than 5% and less than 10%, respectively. Moreover, progesterone (P) serum levels were measured on embryo transfer day as previously described [ 14 ]. Post-thaw embryos were transferred at the cleavage stage under the guidance of abdominal ultrasound. Depending on the patient’s age and the availability and quality of embryos, two to three embryos were transferred. Embryos were graded according to the criteria proposed by Depa-Martynow et al. This evaluation considered the number and quality of blastomeres and the percentage of fragmentation. The grading criteria were as follows: Grade A: Embryos with 7-9 blastomeres and ≤20% cytoplasmic fragmentation (highest quality). Grade B: Embryos with 7–9 blastomeres and > 20% cytoplasmic fragmentation. Grade C: Embryos with 4–6 blastomeres and ≤ 20% cytoplasmic fragmentation. Grade D: Embryos with 4–6 blastomeres and > 20% cytoplasmic fragmentation (lowest quality) [ 15 ]. Grade A: Embryos with 7-9 blastomeres and ≤20% cytoplasmic fragmentation (highest quality). Grade B: Embryos with 7–9 blastomeres and > 20% cytoplasmic fragmentation. Grade C: Embryos with 4–6 blastomeres and ≤ 20% cytoplasmic fragmentation. Grade D: Embryos with 4–6 blastomeres and > 20% cytoplasmic fragmentation (lowest quality) [ 15 ]. The primary outcome of this study was clinical pregnancy, defined as the presence of an intrauterine gestational sac with fetal heartbeat detected at 7 weeks of gestation. Secondary outcomes included biochemical pregnancy (positive β-hCG without a gestational sac), implantation rate, and miscarriage rate (loss of clinical pregnancy before 12 weeks of gestation). Normality was assessed using the Kolmogorov-Smirnov test. Measurement data were expressed as mean ± standard deviation (SD) for continuous variables and as numbers (percentage) for categorical variables. Fisher’s exact test, chi-square test, Mann-Whitney test, and independent t-test were used as appropriate. Decision tree analysis was performed to determine the optimal estrogen threshold on the day of embryo transfer, considering the estrogen levels on the first day of progesterone administration. The CART growing algorithm was chosen, and the splitting criterion was Gini index. The maximum tree depth was equal to 1 to find the cut-off point. In addition, the minimum cases in parents node and child node were 20 and 5 in order. To assess the generalizability of the model, 10-fold cross-validation was performed. The dataset was randomly partitioned into 10 subsets, and the model was iteratively trained on 9 folds and tested on the remaining fold. The average misclassification risk and standard error were calculated across all folds. The Receiver Operating Characteristic (ROC) curve was utilized to evaluate the sensitivity, specificity, and area under the curve (AUC) for the threshold obtained through the decision tree, based on the E2 serum levels on days 1 and 4 of progesterone treatment. To further assess whether continuous estrogen levels alone (without thresholding) differed in predictive value between Day 1 and the transfer day, a DeLong test was performed in order to compare their ROC curves using R software (version 4.5.0; R Foundation for Statistical Computing, Vienna, Austria). The pROC package was used to generate ROC curves and perform the DeLong test for comparing the areas under correlated ROC curves. A binary logistic regression model was used to assess the effects of various independent variables on clinical (positive/negative) pregnancy outcome in patients with estrogen levels above the threshold. Variables with p-value < 0.1 in univariate analysis and those deemed clinically relevant were included in the multivariable model. P-values less than 0.05 were considered statistically significant in the final analysis. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, NY, USA).

Our retrospective study demonstrated that E2 levels on the day of embryo transfer are a strong determinant of treatment success following the transfer of frozen-thawed cleavage-stage embryos. Patients with E2 > 201 pg/mL on the day of FET after an artificial endometrial preparation had significantly higher biochemical and clinical pregnancy rates, along with a notably lower incidence of early pregnancy loss. Importantly, this association persisted after controlling for embryo quality, endometrial thickness, and progesterone levels, suggesting that E2’s role extends beyond mere endometrial preparation. The association between low E2 levels (< 201 pg/mL) and poor reproductive outcomes may be due to the facts that estrogen is essential for endometrial proliferation, glandular secretion, and vascular endothelial growth factor (VEGF) upregulation, all of which are critical for embryo implantation. Below 201 pg/mL, the endometrium may lack sufficient decidual transformation, leading to impaired embryo attachment [ 16 , 17 ]. In addition, estrogen promotes immune tolerance by increasing regulatory T cells (Tregs) and reducing pro-inflammatory cytokines (e.g., TNF-α, IFN-γ). Suboptimal E2 levels may predispose to excessive inflammation, increasing early pregnancy loss risk [ 18 , 19 ]. The findings of this study also suggest that minimum serum E2 thresholds of 144.7 pg/mL on the first day of progesterone administration and 201 pg/mL on the day of transfer need to be achieved. These thresholds, determined by decision tree analysis and confirmed through ROC curve analysis, demonstrated sensitivity of 76% and 90% and specificity of 68% and 70%, respectively. Although the AUC for estrogen levels on the first day of progesterone administration was slightly higher than that observed on the day of embryo transfer, its sensitivity and specificity were comparatively lower. Given that predictive accuracy relies not only on AUC but also on practical classification performance, we prioritized the estrogen level measured on the embryo transfer day. This time point is also more closely aligned with the implantation window and may better reflect endometrial receptivity. Therefore, the cut-off value of 201 pg/mL derived from transfer-day measurements was selected for further analysis and clinical interpretation. Existing data regarding optimal E2 levels on embryo transfer day during artificial endometrial preparation remains limited and inconsistent. Previous studies have primarily examined estrogen concentrations prior to progesterone administration, reflecting the late follicular phase rather than the implantation-critical luteal phase [ 20 – 22 ]. Consequently, no consensus exists regarding ideal serum E2 thresholds [ 20 , 23 , 24 ]. Vyas et al. established a physiological range of 300–500 pg/mL for peak follicular-phase E2 preceding progesterone administration, with values outside this range correlating with significantly reduced live birth rates [ 25 ]. In a recent cohort study of 412 single euploid blastocyst transfers during FET cycles with optimal mid-luteal progesterone, Alsbjerg et al. demonstrated that mid-luteal E2 levels between 292 and 409 pg/mL (1070–1500 pmol/L) on progesterone day 6 were optimal for ongoing pregnancy and live birth, with deviations predicting poorer outcomes [ 26 ]. Notably, despite standardized 6 mg/day oral estradiol administration, not all patients achieved optimal transfer-day E₂ levels. Our findings corroborate this heterogeneity - only 57% of patients receiving identical estradiol dosages (with endometrial thickness > 7 mm at progesterone initiation) attained E₂ levels above the identified threshold by transfer day. These results suggest that endometrial thickness alone may be insufficient for optimal transfer timing, underscoring the need for combined hormonal and morphological assessment. The regression analysis highlights that while age is a significant predictor of clinical pregnancy in the unadjusted model, its effect is no longer significant after adjusting for other relevant variables. This suggests that age may confound the relationship between hormonal levels and pregnancy outcome but is not an independent predictor when other factors are considered. Our findings align with previous studies indicating that estradiol (E2) levels on the day of progesterone initiation can significantly impact pregnancy outcomes in frozen embryo transfer (FET) cycles. In particular, Li et al. demonstrated that elevated E2 levels were associated with lower implantation, clinical pregnancy, and live birth rates in cleavage-stage embryo transfers, but not in blastocyst-stage transfers. Their study, which analyzed 776 HRT-FET cycles, found that E2 levels in the highest decile (≥ 508.4 pg/mL) significantly reduced reproductive success compared to lower ranges [ 27 ]. These findings underscore the importance of optimal estrogen exposure, suggesting that supraphysiological E2 levels may negatively affect endometrial receptivity during the implantation window, especially in cleavage-stage transfers. Our study reinforces this concept by identifying a clinically relevant E2 threshold (201 pg/mL) that correlates with significantly improved clinical pregnancy rates. Our findings are further supported by the prospective study by Alyasin et al. which investigated the association between serum progesterone levels and pregnancy outcomes in hormone replacement therapy cycles. Unlike our study, which included only patients with progesterone levels within a predefined optimal range (11–25 ng/mL), Alyasin et al. included a broader population and identified that progesterone levels exceeding 32.5 ng/mL on the day of embryo transfer were significantly associated with reduced live birth rates. While their study included both cleavage-stage and blastocyst-stage embryo transfers and focused on live birth as the primary outcome, the consistency in findings reinforces the concept that excessive progesterone exposure may impair endometrial receptivity, even when endometrial thickness is optimal [ 14 ]. In contrast to our findings, Du et al., in a retrospective analysis of 708 HRT cycles with frozen-thawed blastocyst transfers, categorized the cycles into quartiles based on serum estrogen levels on the endometrium transformation day (A1: E2 < 157.5 pg/ml, A2: 157.5 pg/ml ≤ E2 < 206.4 pg/ml, A3: 206.4 pg/ml ≤ E2 < 302.3 pg/ml, and A4: E2 ≥ 302.3 pg/ml) and on the blastocyst transfer day (B1: E2 < 147 pg/ml, B2: 147 pg/ml ≤ E2 < 200.4 pg/ml, B3: 200.4 pg/ml ≤ E2 < 323 pg/ml, and B4: E2 ≥ 323 pg/mL). They found that higher serum estrogen levels (≥ 302.3 pg/mL) on the endometrium transformation day were associated with lower blastocyst implantation rates and reduced multiple pregnancy rates but did not significantly impact clinical pregnancy rates, abortion rates, or live birth rates. Overall, they reported that neither serum estrogen levels nor the duration of Progynova administration significantly influenced pregnancy outcomes in HRT cycles. Moreover, multivariate regression analysis of this study indicated the age and number of high-quality blastocysts transferred were independent factors affecting clinical pregnancy, while the estrogen levels before blastocyst transfer, days of taking Progynova and progesterone level on the blastocyst transfer day did not significantly influence clinical pregnancy rates [ 28 ]. Hung et al. also reported that the peri-implantation estradiol level did not significantly influence pregnancy outcomes [ 29 ]. This discrepancy in findings may be attributed to differences in the stage of transferred embryos (cleavage versus blastocyst transfer), patient population differences (such as age and BMI), E2 measurement techniques, and the timing of E2 measurements (the day or the timing of blood sampling relative to estradiol administration). This study has several limitations that must be acknowledged. First, its retrospective design inherently introduces bias. Second, we were unable to standardize the exact time interval between the last oral dose of estradiol valerate (E2) and the collection of blood samples for serum E2 measurement, potentially contributing to variability in E2 levels. Additionally, the poor clinical outcomes in the low-estrogen group (Group A) due to low embryo genetic quality cannot be fully excluded. Embryo morphology and the number transferred were similar, but the lack of PGT-A data limits our ability to rule out embryonic aneuploidy as a contributing factor. Investigating outcome-specific thresholds through independent analyses should be considered as well. Also, we acknowledge that our data-driven dichotomization may not fully capture the potentially nonlinear relationship between estradiol concentrations and clinical outcomes, as suggested by previous research. One strength of this study is the use of cross-validation to assess the robustness of the decision tree model. The cross-validated risk estimate was identical to the resubstitution error, suggesting that the model is unlikely to be overfitted and may generalize well to similar patient populations. However, as the model was developed and validated on the same dataset, future studies with external validation cohorts are necessary to confirm the predictive value of the proposed estradiol threshold. These factors underscore the need for prospective studies with larger datasets and nonlinear modeling approaches to more accurately assess the clinical relevance of serum E2 levels. Despite these limitations, the study has notable strengths. It focused on women under 40 years old, with two to three high-quality embryos transferred into a normal endometrium, allowing for the evaluation of E2 effects in cleavage-stage transfers while minimizing confounding variables. Furthermore, both E2 and progesterone (P4) levels were administered via a single consistent route, reducing variability from different administration methods and enhancing the reliability of the findings. In conclusion, this study identifies serum estradiol (E2) levels on the day of embryo transfer as a significant prognostic marker for clinical outcomes in frozen–thawed cleavage-stage embryo transfer cycles. E2 concentrations exceeding 201 pg/mL were independently associated with higher biochemical and clinical pregnancy rates, as well as a lower incidence of early pregnancy loss. These associations remained robust after adjusting for potential confounders such as embryo quality and endometrial thickness. Our findings suggest that monitoring serum E2 levels may provide additional predictive value beyond conventional morphological assessments and could contribute to more individualized and effective management of HRT-FET cycles.

Over the past decade, the global use of frozen–thawed embryo transfer (FET) has increased significantly [ 1 , 2 ]. This trend reflects major advancements in controlled ovarian hyperstimulation (COH) protocols, improvements in vitrification and warming techniques, the adoption of elective single embryo transfer (eSET) policies, and the expanded application of preimplantation genetic testing for aneuploidy (PGT-A) [ 3 – 5 ]. Moreover, the “freeze-all” strategy has demonstrated high pregnancy rates in both high responders and normo-responders [ 6 , 7 ]. Despite substantial progress in embryo culture and cryopreservation, achieving optimal endometrial preparation for FET remains a critical challenge [ 8 ]. Endometrial preparation can be performed via natural or artificial (hormonally substituted) cycles. Artificial cycles typically involve hormone replacement therapy (HRT) using exogenous estradiol and progesterone to simulate the natural cycle and support implantation. In recent years, numerous efforts have been made to optimize HRT protocols to enhance pregnancy outcomes. Studies have suggested that an endometrial thickness between 8.7 and 14.5 mm is associated with the highest live birth rates in HRT-FET cycles, while deviations from this range are linked to poorer outcomes [ 9 ]. Furthermore, serum progesterone levels on the day of embryo transfer have been identified as independent predictors of both clinical pregnancy and live birth [ 10 , 11 ]. However, the role of serum estradiol (E2) levels in this context remains unclear. In natural cycles, E2 levels progressively increase during the follicular phase, typically exceeding 200 pg/mL in the late follicular stage and maintaining that level for at least 50 h prior to ovulation [ 12 ]. Whether similar E2 dynamics are beneficial or necessary in artificial HRT cycles is still uncertain. Importantly, no consensus has been reached regarding a specific threshold or optimal range of serum E2 that correlates with successful pregnancy outcomes in HRT-FET cycles. Shuai et al. recently conducted a large cohort study involving 26,194 patients and reported that higher serum E2 levels prior to progesterone administration were associated with reduced clinical pregnancy and live birth rates [ 13 ]. Conversely, another retrospective study found no significant correlation between different ranges of late proliferative phase serum E2 levels (≤ 144 pg/mL, 145–438 pg/mL, > 439 pg/mL) and live birth rates in HRT-FET cycles. Given these inconsistencies and the limited data regarding the impact of serum E2 after progesterone initiation, we conducted a retrospective study of 175 HRT-FET cycles performed at our center between January and May 2024. This study aimed to evaluate the effect of serum estradiol levels on the day of cleavage-stage embryo transfer on clinical pregnancy outcomes in cycles with optimal progesterone levels, thereby contributing to the refinement of HRT-FET protocols.