Analysis of related factors of biochemical pregnancy loss in single blastocyst transfer in the hormone replacement protocol frozen-thawed cycle.

Between January 2012 and November 2024, a total of 18,937 cycles of single blastocyst transfer using the HRT protocol were performed at the Center for Reproductive Medicine and Obstetrics and Gynecology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University. Of these, 10,274 cycles exhibited positive serum β-hCG levels (≥ 100 U/L) two weeks post-transfer. At 4 weeks post-transfer, 9,644 cycles (50.93%) had a clinical pregnancy confirmed by transvaginal ultrasound, with the presence of a gestational sac. BPL was observed in 630 cycles (3.33%), where no gestational sac was detected, and serum β-hCG levels did not return to baseline and/or showed stagnation. The incidence of BPL among patients with positive β-hCG was 6.13% (630/10,274). Baseline characteristics of the BPL group were compared with the clinical pregnancy (CP) group. Significant differences were observed between the two groups in terms of female age, male age, female BMI, number of transfer cycles, basal antral follicle count (AFC), estradiol (E2) and progesterone levels prior to endometrial transformation, the number of high-quality embryos transferred, blastocyst expansion, blastocyst morphological score, days after embryo transfer, and endometrial thickness (all P < 0.05). Specifically, the BPL group had significantly higher levels of female age, male age, female BMI, the number of transfer cycles, and pre-endometrial transformation E2 levels compared with the CP group. Conversely, the BPL group had lower basal AFC, fewer high-quality embryos transferred, poorer blastocyst expansion and quality scores, fewer days post-transfer, and thinner endometrial thickness compared to the CP group.However, no significant differences were found between the two groups in terms of male infertility factors, duration of infertility, infertility type (primary or secondary), ovulation disorders, diminished ovarian reserve, endometriosis, adenomyosis, progesterone levels before endometrial transformation, or basic sex hormone levels (including basal FSH, luteinizing hormone [LH], and anti-Müllerian hormone [AMH]). Additionally, no significant differences were found regarding endometrial morphology classification [ 10 ], uterine cavity space-occupying lesions, uterine cavity adhesions, the use of GnRH-a, or assisted hatching (all P >0.05) (Table 1 ). Table 1 Comparison of baseline characteristics between the HRT-FET single blastocyst cycle biochemical pregnancy loss group and the clinical pregnancy group (Mean ± SD) Pregnancy loss group (study group) Clinical pregnancy group (control group) Standardize diff. P value N 630 9644 Female age (year, x ˉ± s ) 31.87 ± 4.34 31.18 ± 4.17 0.161 (0.080, 0.241) <0.001 BMI(kg/m 2 , x ˉ± s ) 23.36 ± 3.46 22.90 ± 3.39 0.134 (0.053, 0.215) 0.002 Male age(year, x ˉ± s ) 33.05 ± 5.12 32.44 ± 4.88 0.122 (0.042, 0.203) 0.005 Infertility years(year, x ˉ± s ) 3.56 ± 2.61 3.38 ± 2.33 0.071 (−0.012, 0.154) 0.441 Infertility type(%) 0.035 (−0.046, 0.117) 0.396  Primary infertility 320(52.37%) 5142(54.14%)  Secondary infertility 291(47.63%) 4356(45.86%) Infertility factors (%)  Ovulation disorders 129(20.48%) 1971(20.44) 0.001 (−0.080, 0.082) 0.981  Diminished ovarian reserve 15(2.38%) 223(2.31%) 0.005 (−0.076, 0.085) 0.912  Endometriosis 31(4.92%) 551(5.71%) 0.035 (−0.045, 0.116) 0.404  Adenomyosis 27(4.29%) 302(3.13%) 0.061 (−0.019, 0.142) 0.111  Male factor 201(31.90%) 3153(32.69%) 0.017 (−0.064, 0.097) 0.682 Embryo transfer cycle number( x ˉ± s ) 2.90 ± 1.62 2.67 ± 1.19 0.157 (0.076, 0.237) 0.02 Basic FSH (IU/L, x ˉ± s ) 6.98 ± 2.53 7.01 ± 2.34 0.014 (−0.067, 0.096) 0.404 Basic LH (IU/L, x ˉ± s ) 6.35 ± 4.51 6.50 ± 4.57 0.034 (−0.048, 0.115) 0.323 AMH (ng/mL, x ˉ± s ) 4.77 ± 3.58 4.99 ± 3.80 0.060 (−0.048, 0.167) 0.283 Basic AFC ( x ˉ± s ) 20.16 ± 10.03 21.19 ± 10.61 0.100 (0.018, 0.182) 0.03 Use of GnRH-a (%) 136(21.59%) 1888(19.58%) 0.050 (−0.031, 0.130) 0.219 HRT-E 2 (pg/ml, x ˉ± s ) 430.63 ± 636.51 344.86 ± 498.61 0.150 (0.066, 0.234) 0.013 HRT-P (ng/ml, x ˉ± s ) 0.20 ± 0.19 0.21 ± 0.19 0.055 (−0.030, 0.139) 0.080 HRT-LH (IU/L, x ˉ± s ) 9.41 ± 8.63 10.81 ± 10.29 0.147 (0.027, 0.268) 0.055 High quality embryo transfer (%) 378(60.00%) 6993(72.51%) 0.267 (0.186, 0.348) <0.001 Assisted hatching (%) 499(79.21%) 7617(78.98%) 0.006 (−0.075, 0.086) 0.893 Degree of blastocyst expansion (%) 0.137 (0.056, 0.217) 0.006  Expansion 3 11(1.75%) 127(1.32%)  Expansion 4 488(77.46%) 7490(77.66%)  Expansion 5 85(13.49%) 1579(16.37%)  Expansion 6 46(7.30%) 448(4.65%) Morphological score (%) 0.314 (0.233, 0.395) <0.001  AA 35(5.56) 587(6.09%)  AB 81(12.86%) 2043(21.18%)  BA 30(4.76%) 476(4.94%)  BB 309(49.05%) 4819(49.97%)  AC 4(0.63%) 74(0.77%)  CA 0(0.00%) 5(0.05%)  BC 112(17.78%) 1215(12.60%)  CB 55(8.73%) 390(4.04%)  CC 4(0.63%) 35(0.36%) Days of blastocysts (day, %) 0.203 (0.122, 0.284) <0.001  Day 5 485(76.98%) 8189(84.91%)  Day 6 145(23.02%) 1455(15.09%) Endometrial thickness (mm, x ˉ± s ) 9.66 ± 1.59 9.89 ± 1.59 0.140 (0.059, 0.221) <0.001 Endometrial patterns by ultrasound (%) 0.054 (−0.028, 0.136) 0.658  Pattern A 207(33.77%) 3325(35.39%)  Pattern B 406(66.23%) 6062(64.52%)  Pattern C 0(0.00%) 8(0.09%) Endometrial adhesions (%) 0(0.00%) 6(0.06%) 0.035 (−0.045, 0.116) 0.531 Endometrial separation (%) 3(0.48%) 29(0.30%) 0.028 (−0.052, 0.109) 0.444 Occupation of uterine cavity (%) 1(0.16%) 20(0.21%) 0.028 (−0.052, 0.109) 0.793 Comparison of baseline characteristics between the HRT-FET single blastocyst cycle biochemical pregnancy loss group and the clinical pregnancy group (Mean ± SD) To investigate the potential associations between embryo- and endometrium-related factors and the incidence of BPL following single blastocyst transfer in frozen-thawed cycles, we performed univariate analyses on relevant patient characteristics and treatment variables. The analysis revealed that several factors were significantly associated with BPL, including female age, female BMI, male age, number of transfer cycles, basal AFC, serum E2 and LH levels prior to endometrial transformation, number of high-quality embryos transferred, degree of blastocyst expansion, blastocyst morphology score, day of embryo transfer, and endometrial thickness before transformation (all P  < 0.05). Notably, transfers involving blastocysts graded as CB or BC, as well as blastocysts transferred on Day 6, were more likely to result in biochemical pregnancy loss compared with those involving high-quality blastocysts or Day 5 transfers. Detailed statistical outcomes are presented in Table 2 . Table 2 Univariate analysis of relevant data in the HRT-FET cycle and the clinical outcome of biochemical pregnancy loss Statistics OR(95%CI) P value Female age (year, x ˉ± s ) 31.22 ± 4.18 1.039 (1.020, 1.059) < 0.001 BMI(kg/m 2 , x ˉ± s ) 22.93 ± 3.39 1.024 (1.009, 1.041) 0.001 Male age(year, x ˉ± s ) 32.48 ± 4.89 1.039 (1.016, 1.064) 0.002 Infertility years(year, x ˉ± s ) 3.39 ± 2.35 1.031 (0.997, 1.066) 0.076 Infertility type(%)  Primary infertility 5462 (54.03%) Reference value  Secondary infertility 4647 (45.97%) 1.073 (0.911, 1.265) 0.396 Infertility factors (%)  Ovulation disorders 2100(20.44%) 1.002 (0.821, 1.224) 0.981  Diminished ovarian reserve 238 (2.32%) 1.030 (0.607, 1.749) 0.912  Endometriosis 582 (5.66%) 0.854 (0.589, 1.238) 0.405  Adenomyosis 329 (3.20%) 1.385 (0.926, 2.071) 0.112  Male factor 3354 (32.65%) 0.965 (0.811, 1.147) 0.682 Embryo transfer cycle number( x ˉ± s ) 2.69 ± 1.22 1.132 (1.071, 1.196) < 0.001 Basic FSH (IU/L, x ˉ± s ) 7.01 ± 2.36 0.994 (0.959, 1.029) 0.721 Basic LH (IU/L, x ˉ± s ) 6.49 ± 4.56 0.992 (0.974, 1.011) 0.420 AMH (ng/mL, x ˉ± s ) 4.98 ± 3.79 0.984 (0.955, 1.014) 0.288 Basic AFC ( x ˉ± s ) 21.13 ± 10.58 0.992 (0.974, 1.011) 0.020 Use of GnRH-a (%) 2024 (19.70%) 1.131 (0.929, 1.376) 0.219 HRT-E 2 (pg/ml, x ˉ± s ) 349.958 ± 508.222 1.000 (1.000, 1.000) <0.001 HRT-P (ng/ml, x ˉ± s ) 0.21 ± 0.19 0.747 (0.473, 1.181) 0.212 HRT-LH (IU/L, x ˉ± s ) 10.72 ± 10.20 0.985 (0.972, 0.998) 0.025 High quality embryo transfer (%) 7371 (71.74%) 0.569 (0.482, 0.671) < 0.001 Assisted Hatching (%) 8116 (79.00%) 1.014 (0.831, 1.236) 0.8933 Degree of blastocyst expansion (%)  Expansion 3 138 (1.34%) Reference value  Expansion 4 7978 (77.65%) 0.752 (0.404, 1.402) 0.370  Expansion 5 1664 (16.20%) 0.622 (0.323, 1.195) 0.154  Expansion 6 494 (4.81%) 1.185 (0.597, 2.356) 0.627 Morphological score (%)  AA 622 (6.05%) Reference value  AB 2124 (20.67%) 0.665 (0.443, 0.999) 0.049  BA 506 (4.93%) 1.057 (0.640, 1.747) 0.829  BB 5128 (49.91%) 1.075 (0.750, 1.541) 0.692  AC 78 (0.76%) 0.907 (0.313, 2.623) 0.856  CA 5 (0.05%) 0.000 (0.000, inf.) 0.964  BC 1327 (12.92%) 1.546 (1.045, 2.288) 0.029  CB 445 (4.33%) 2.365 (1.519, 3.682) <0.001  CC 39 (0.38%) 1.917 (0.645, 5.697) 0.242 Days of blastocysts (day, %)  Day 5 8674 (84.43%) Reference value  Day 6 1600 (15.57%) 1.683 (1.386, 2.042) < 0.001 Endometrial thickness (mm, x ˉ± s ) 9.87 ± 1.59 0.910 (0.861, 0.961) <0.001 Endometrial patterns by ultrasound (%)  Pattern A 3532 (35.29%) Reference value  Pattern B 6468 (64.63%) 1.076 (0.905, 1.279) 0.407  Pattern C 8 (0.08%) 0.000 (0.000, inf.) 0.955 Endometrial adhesions (%) 6 (0.06%) 0.000 (0.000, inf.) 0.961 Endometrial separation (%) 32 (0.31%) 1.586 (0.482, 5.222) 0.448 Occupation of uterine cavity (%) 21 (0.20%) 0.765 (0.103, 5.709) 0.794 OR O dds ratio, CI C onfidence interval Univariate analysis of relevant data in the HRT-FET cycle and the clinical outcome of biochemical pregnancy loss OR O dds ratio, CI C onfidence interval To further explore the dynamic associations between clinical and embryological factors and the incidence of BPL, smooth curve fitting analyses were performed. The results revealed that increasing female age, BMI, male age, number of embryo transfer, and serum E2 levels prior to endometrial transformation were each positively associated with an elevated risk of BPL. Conversely, decreases in the basal AFC, LH levels before transformation, and endometrial thickness were also associated with a rising incidence of BPL. These findings are illustrated in Fig. 1 . In addition, transfer cycles involving non-high-quality blastocysts-specifically those graded as CB or BC-as well as transfers performed on Day 6, exhibited a significantly higher likelihood of BPL when compared with cycles using high-quality blastocysts or transfers conducted on Day 5. Among all morphological categories, CB-graded blastocysts were associated with the highest incidence of BPL, followed by BC-graded embryos. Fig. 1 Smooth curve fitting analysis of factors related to single blastocyst transfer in the HRT protocol of the freeze-thaw cycle and the incidence of biochemical pregnancy loss. A - E . Smooth fitting curve analysis of female age, BMI, male age, number of transplantation cycles, E 2 level before endometrial transformation and biochemical pregnancy loss rate. The curve shown shows the relationship between FET cycle-related factors and the biochemical pregnancy loss rate. The area between the two dotted lines is represented as the 95% CI. The rate of biochemical pregnancy loss in patients gradually increases with the increase of related factors. F - H . Smooth fitting curve analysis of the number of basal antral follicles, endometrial thickness before endometrial transformation, LH level before endometrial transformation and biochemical pregnancy loss rate. The curve shown shows the relationship between FET cycle-related factors and the biochemical pregnancy loss rate. The area between the two dotted lines is represented as the 95% CI. The rate of biochemical pregnancy loss in patients gradually decreases with the increase of related factors. Notes: CI: confidence interval; EMT: Endometrial thickness Smooth curve fitting analysis of factors related to single blastocyst transfer in the HRT protocol of the freeze-thaw cycle and the incidence of biochemical pregnancy loss. A - E . Smooth fitting curve analysis of female age, BMI, male age, number of transplantation cycles, E 2 level before endometrial transformation and biochemical pregnancy loss rate. The curve shown shows the relationship between FET cycle-related factors and the biochemical pregnancy loss rate. The area between the two dotted lines is represented as the 95% CI. The rate of biochemical pregnancy loss in patients gradually increases with the increase of related factors. F - H . Smooth fitting curve analysis of the number of basal antral follicles, endometrial thickness before endometrial transformation, LH level before endometrial transformation and biochemical pregnancy loss rate. The curve shown shows the relationship between FET cycle-related factors and the biochemical pregnancy loss rate. The area between the two dotted lines is represented as the 95% CI. The rate of biochemical pregnancy loss in patients gradually decreases with the increase of related factors. Notes: CI: confidence interval; EMT: Endometrial thickness To reduce potential confounding bias and better delineate independent predictors of BPL, a multivariate logistic regression analysis was conducted. Based on the univariate findings, the following variables were included in the model: female age, female BMI, male age, basal AFC, number of embryo transfer cycles, number of high-quality embryos transferred, blastocyst morphology score, day of embryo transfer, endometrial thickness prior to transformation, and pre-transformation serum levels of E2 and LH.Considering that female age, BMI, male age, AFC, and the number of transfer cycles could serve as potential effect modifiers, these variables were included in the model as covariates. Endometrial thickness, E2 and LH levels, blastocyst quality, and day of transfer were treated as exposure variables of interest.After adjusting for the confounding variables, only endometrial thickness prior to transformation remained significantly associated with the incidence of BPL (OR = 0.86, 95% CI: 0.79–0.93, P = 0.0002). No statistically significant associations were observed for pre-transformation E2 or LH levels, day of blastocyst transfer (Day 6 vs. Day 5), assisted hatching, or blastocyst quality in the multivariate model. Full results are shown in Table 3 . Table 3 The results of multiple regression analysis of the relevant data in the HRT-FET cycle and the clinical outcome of biochemical pregnancy loss Exposure variable OR(95%CI) P value Endometrial thickness(mm, x ˉ± s ) 0.855 (0.787, 0.929) <0.001 HRT-E 2 (pg/ml, x ˉ± s ) 1.000 (1.000, 1.000) 0.930 HRT-LH(IU/L, x ˉ± s ) 0.992 (0.978, 1.005) 0.214 Days of blastocysts  Day5 Reference value  Day6 1.391 (0.906, 2.135) 0.132 Assisted Hatching 1.084 (0.790, 1.488) 0.617 (%) High quality embryo transfer 0.772 (0.536, 1.112) 0.164 Adjustment variables: Female age, BMI, Male age, Basic AFC, Embryo transfer cycle number OR Odds ration, CI Confidence interval The results of multiple regression analysis of the relevant data in the HRT-FET cycle and the clinical outcome of biochemical pregnancy loss Adjustment variables: Female age, BMI, Male age, Basic AFC, Embryo transfer cycle number OR Odds ration, CI Confidence interval The smooth curve fitting analysis indicated a negative correlation between endometrial thickness prior to transformation and the incidence of BPL. Specifically, the risk of BPL decreased progressively with increasing endometrial thickness. This relationship is visualized in Fig.  2 . To further evaluate this relationship, a threshold effect analysis was conducted, which identified 10 mm as the optimal cutoff point. When endometrial thickness before endometrial transformation was <10 mm, a statistically significant inverse association with BPL was observed ( P = 0.0005). However, when the thickness was ≥10 mm, the correlation was no longer significant ( P = 0.629). Detailed results are presented in Table 4 . Fig. 2 Smooth curve fitting analysis of endometrial thickness before transformation and the incidence of biochemical pregnancy loss Table 4 Threshold effect analysis of endometrial thickness and clinical outcomes of biochemical pregnancy loss before FET cycle transformation Model OR (95%CI) P value Model I (linear) Standard linear effect 0.906 (0.857, 0.958) <0.001 Model II (polyline) Folding point (K) Inflection point of endometrial thickness 10 Thickness of the endometrium < 10 mm 0.818 (0.730, 0.916) <0.001 Thickness of the endometrium ≥ 10 mm 0.978 (0.894, 1.070) 0.629 Difference in effect 1.196 (1.005, 1.424) 0.044 Predicted value of the equation at the turning point -2.893 (-3.036, -2.750) Logarithmic Likelihood ratio test (LRT) 0.046 Exposure variable: Endometrial thickness; Adjustment variables: Female age, BMI, Male age, Basic AFC, Embryo transfer cycle number OR O dds rati, CI C onfidence interval Smooth curve fitting analysis of endometrial thickness before transformation and the incidence of biochemical pregnancy loss Threshold effect analysis of endometrial thickness and clinical outcomes of biochemical pregnancy loss before FET cycle transformation Folding point (K) Inflection point of endometrial thickness Exposure variable: Endometrial thickness; Adjustment variables: Female age, BMI, Male age, Basic AFC, Embryo transfer cycle number OR O dds rati, CI C onfidence interval To validate the threshold effect, patients were stratified into two subgroups: Group A: Endometrial thickness <10 mm; Group B: Endometrial thickness ≥10 mm. The incidence of BPL was significantly higher in Group A compared to Group B (6.76% vs. 5.20%, P < 0.05), confirming the clinical relevance of the 10 mm threshold. These results are summarized in Table 5 . Receiver operating characteristic (ROC) analysis was conducted using endometrial thickness as a risk assessment indicator for BPL, and the area under the curve (AUC) of endometrial thickness was 0.547 (Fig.  3 ). Table 5 Verification of the cutoff value of endometrial thickness of 10 mm before transformation Group A Group B P Value Embryo transfer cycle number 5958 4177 Biochemical pregnancy loss 403 (6.76%) 217 (5.20%) 0.001 Group A༚endometrial thickness before transformation <10 mm Group B: endometrial thickness before transformation ≥ 10 mm Fig. 3 The ROC of endometrial thickness as a risk assessment indicator for BPL Verification of the cutoff value of endometrial thickness of 10 mm before transformation Group A༚endometrial thickness before transformation <10 mm Group B: endometrial thickness before transformation ≥ 10 mm The ROC of endometrial thickness as a risk assessment indicator for BPL

This retrospective cohort study was conducted at the Center for Reproductive Medicine and Obstetrics and Gynecology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University, from January 2012 to November 2024. A total of 54,413 FET cycles were evaluated. The inclusion criteria comprised patients who underwent single blastocyst transfer with a HRT endometrial preparation protocol. Exclusion criteria were: (1) patients with clinically significant chromosomal or genetic abnormalities in either partner, for whom preimplantation genetic testing (PGT) was required; (2) egg donor recipients; and (3) patients lost to follow-up. After applying these criteria, 18,937 single blastocyst FET cycles were included in the analysis, comprising 630 BPL cycles and 9,644 clinical pregnancy (CP) cycles. The study adhered to the principles of the Declaration of Helsinki, and informed consent was obtained from all participants. The study protocol was approved by the ethics committee of Nanjing Drum Tower Hospital (2021-384-01). All patients underwent a comprehensive pre-pregnancy physical examination to rule out any contraindications to pregnancy, including internal or surgical conditions. In this study, the endometrial preparation, luteal support, embryo transfer and clinical follow-up of the HRT protocol were all carried out in accordance with the routine procedures of our center. HRT protocol: Endometrial preparation for all patients followed the standard HRT protocol. Briefly, patients were administered 6 mg of estradiol (Femoston, Abbott, USA) daily for 14 days starting from the second day of menstruation. Upon reaching the optimal endometrial thickness (EMT), patients were switched to a combination of 6 mg estradiol and 10 mg dydrogesterone (Femoston yellow tablets) and were also given intramuscular progesterone (Xianju, Zhejiang, 20 mg per vial) at 60 mg per day for luteal phase support. It lasted for 5 or 6 days to induce endometrial transformation. Down-regulation + HRT protocol: For patients with endometriosis or adenomyosis, a down-regulation phase with long-acting GnRHa (Triptorelin, Decapeptyl, Ferring, Switzerland) was performed prior to the HRT protocol. After 1 to 3 cycles of down-regulation with 3.75 mg, HRT protocol was carried out. Cleavage embryos are thawed and transplanted on the fifth day of endometrial transformation. The blastocyst is thawed and transplanted on the sixth day of endometrial transformation. The blastocyst scoring adopts the Garnder blastocyst grading method, which is commonly used internationally [ 9 ]. The evaluation criteria are mainly based on the developmental stage of the blastocyst, the Inner cell mass (ICM), and the Trophoblast cells. (1) The development stages of blastocysts are classified into grades 1 to 6 based on their expansion and hatching degrees: Grade 1, blastocyst development and stage status. Grade 2, blastocoel cavity more than half the volume of the embryo. Grade 3, full blastocyst, cavity completely filling the embryo. Grade 4, expanded blastocyst, cavity larger than the embryo, with thinning of the shell. Grade 5, hatching out of the shell. Grade 6, hatched out of the shell. Blastocysts of grades 3 to 6 need to be scored for the ICM and trophoblast ectodermal cells (TE). (2) ICM grade: Grade A, many cells, tightly packed. Grade B, several cells, loosely grouped. Grade C, very few cells. (3) TE grade: Grade A, many cells, forming a cohesive layer. Grade B, few cells, forming a loose epithelium. Grade C, very few large cells. After embryo transfer, all drugs were maintained at their original doses. Patients usually take Femoston yellow tablets (1 mg estradiol plus 10 mg dydrogesterone tablet) 6 mg/ day and progesterone sustained-release Vaginal Gel (Crinone, 90 mg/ vial, Merck, Germany) 90 mg/ day for luteal support. Pregnancy outcomes were determined by measuring serum β-hCG levels two weeks post-transfer, with values ≥ 100 U/L considered indicative of biochemical pregnancy. Transvaginal ultrasound at 4 weeks post-transfer was used to confirm clinical pregnancy, defined by the presence of a gestational sac. A pregnancy outcome where a gestational sac is observed by ultrasound is classified as clinical pregnancy. Conversely, a pregnancy outcome where no gestational sac is detected, accompanied by a decrease in serum β-hCG to the normal range or stagnation in β-hCG levels without returning to baseline, is classified as BPL. After the patient became pregnant, luteal support was continued until 2 months after embryo transfer, and the patient was followed up continuously until parturition. Statistical analysis was performed using the R software package ( http://www.R-project.org , version 3.6.0) and Empower Stats ( www.empowerstats.com , X&Y solutions, Inc. Boston MA). Continuous variables were analyzed for normal distribution using the Kolmogorov-Smirnov test, with comparisons made using t-tests for normally distributed data and Mann-Whitney U tests for non-normally distributed data. The variables were described as mean ± standard deviation (SD). Univariate logistic regression and smooth curve fitting were used to assess the potential associations between patient characteristics and biochemical pregnancy loss, followed by multivariate logistic regression to adjust for confounders, including age, body mass index (BMI), follicular-stimulating hormone (FSH), and the number of transfer cycles. The value of the evaluation results for the BPL risk assessment index was analyzed and evaluated by using the receiver operating characteristic (ROC) curve. The area under the curve (AUC) is an important indicator for evaluating the overall performance of the model. The closer the AUC value is to 1, the better its predictive value is. P values < 0.05 were considered statistically significant.

This study systematically analyzed the risk factors associated with BPL following single blastocyst transfer in FET cycles based on a large-scale retrospective cohort. Our findings underscore the critical role of endometrial thickness, along with blastocyst morphology and developmental timing, in predicting BPL. Strategies such as optimizing endometrial preparation and careful embryo selection may effectively reduce BPL incidence and improve ART outcomes.

With the rapid advancement of ART, the utilization of single embryo transfer (SET) and FET has increased significantly in recent years [ 11 ]. Due to their superior implantation potential, blastocyst-stage embryos are now preferred over cleavage-stage embryos, particularly in vitrified-thawed cycles, where embryo survival rates have markedly improved [ 12 ]. Consequently, single blastocyst transfer (SBT) has become a common strategy. However, the incidence of BPL following SBT in FET cycles remains a concern, with reported rates ranging from 10% to 13.8% [ 13 , 14 ]. BPL is now recognized as a key limiting factor affecting successful pregnancy outcomes in ART. BPL is characterized by the detection of hCG in serum or urine without the visualization of a gestational sac on ultrasound. This condition, also referred to as subclinical miscarriage or early embryonic loss, may indicate implantation failure or very early pregnancy attrition. Embryo implantation is a highly complex and regulated process that involves a sequence of steps, including embryo localization, adhesion to the endometrial lining, and subsequent invasion. These processes are mediated by intricate molecular signals exchanged between the embryo and the endometrial environment. Disruptions in these signaling pathways may contribute to the occurrence of BPL; however, the precise biological mechanisms underlying this phenomenon remain largely unexplored. Although both SA and BPL fall under the category of EPL, they are influenced by distinct factors and may arise from different mechanisms. SA is commonly associated with chromosomal abnormalities in the embryo, as well as various uterine and endocrine disorders [ 8 ]. In contrast, while current literature suggests that BPL may be linked to ART-specific factors such as ovarian stimulation protocols and embryo manipulation, definitive evidence remains insufficient [ 15 ]. he variability in ART protocols and the lack of large-scale randomized controlled trials further complicate our understanding of the etiology of BPL. Therefore, additional studies are imperative to better delineate the risk factors associated with BPL in the context of ART. To minimize potential confounding variables, the present study focused exclusively on SBTs within hormone replacement therapy (HRT)-based FET cycles at a single high-volume reproductive center. This design provided enhanced control over embryo developmental stages and endometrial preparation protocols, thereby enabling a more reliable analysis of the risk factors associated with biochemical pregnancy loss (BPL). The relationship between embryo quality and BPL remains a subject of ongoing debate. Musters et al. suggested that BPL is suggested by endometrial factors and is not significantly associated with embryo quality [ 16 ], developmental stage, or chromosomal euploidy [ 17 ]. However, a substantial body of evidence indicates that poor embryo quality and limited developmental potential are key factors contributing to BPL [ 6 ]. In the present study, we initially conducted univariate analyses to examine the associations between various embryo- and endometrium-related factors and the incidence of BPL. The results revealed that certain variables—particularly lower morphological scores (e.g., blastocysts graded as CB or BC) and blastocyst transfers performed on Day 6—were significantly associated with a higher risk of BPL. These observations are consistent with prior studies, which have demonstrated that suboptimal embryo morphology and delayed blastocyst development may compromise implantation potential and increase the likelihood of early pregnancy failure. Blastocyst morphology is a critical parameter in assessing embryonic developmental potential. Numerous studies have shown that higher-grade blastocysts possess stronger implantation competence. In line with these results, our univariate analyses findings confirm that blastocyst morphology is closely related to BPL risk following HRT-FET single blastocyst transfer [ 18 ]. The number of developmental days and the morphological quality of the blastocyst were both significantly correlated with BPL. Day 5 high-grade blastocysts demonstrated stronger developmental potential, whereas lower-grade blastocysts (CB/BC) were more likely to exhibit deficiencies in cell number, compaction, and trophectoderm development—factors that may contribute to implantation failure and BPL. Several studies, including those by Dai [ 19 ] and Abdala [ 20 ], have demonstrated a significant association between the developmental stage of embryos transferred during FET and BPL. In particular, Day 6 blastocysts exhibit a higher incidence of BPL compared to Day 5 blastocysts. This difference may be partially attributed to metabolic and genetic factors: mitochondrial DNA (mtDNA) content in Day 6 blastocysts is significantly lower than in Day 5 blastocysts. Given the critical role of mitochondria in cellular energy metabolism, reduced mtDNA content may impair metabolic processes, leading to developmental delays and an increased risk of chromosomal abnormalities, such as aneuploidy or mosaicism [ 21 ]. Our univariate analysis results are consistent with previous studies indicating that the transfer of blastocysts graded as BB, BC, or CB increases the risk of BPL by approximately 41% compared to embryos graded AA, AB, or BA. This increased risk may be attributed to prolonged in vitro culture, which can exacerbate abnormal gene expression, epigenetic instability, and mitochondrial dysfunction, ultimately compromising embryo viability [ 22 ]. Vaiarelli et al. [ 17 ] also reported that blastocysts at expansion stages 5/6 are associated with a higher risk of BPL, suggesting that asynchronous embryo–endometrium development may be another potential contributing factor. However, in our study, the number of blastocysts at stages 5/6 was relatively small, and no definitive association between blastocyst expansion stage and BPL was observed. Nevertheless, our findings emphasize the importance of selecting embryos with optimal morphology and appropriate developmental timing in ART practice. In our subsequent multivariate logistic regression analysis, which aimed to control for confounding factors, only endometrial thickness prior to transformation remained a significant independent predictor of BPL (OR = 0.86, 95% CI: 0.79–0.93, P = 0.0002). Notably, after adjusting for variables such as female age, BMI, male age, number of prior transfer cycles, and embryo quality, neither blastocyst morphology nor the timing of transfer maintained statistical significance in their association with BPL (P >0.05). While blastocyst morphology remains an important parameter in assessing embryonic developmental potential, with higher-grade blastocysts typically associated with better implantation competence, our study showed that these associations were influenced by other variables in the multivariate model. Although suboptimal blastocyst grades and Day 6 transfers were linked to increased BPL in univariate analyses, they lost their significance in the multivariate context. Thus, while careful embryo selection is crucial, our findings underscore that the most significant predictor of BPL in this study was endometrial thickness. In addition to embryo-related factors, our univariate analysis also revealed that advanced maternal and paternal age, as well as diminished ovarian reserve, were associated with an increased risk of BPL. In older women, the higher incidence of embryo chromosomal abnormalities may partially explain the increased risk of BPL. However, even after preimplantation genetic testing for aneuploidy (PGT-A), the BPL rate remains elevated at 18.3%, suggesting the involvement of non-chromosomal age-related factors [ 23 , 24 ]. Age-related metabolic decline in women, including reductions in muscle mass and basal metabolic rate, may contribute to systemic inflammation and coagulation abnormalities [ 25 , 26 ]. Additionally, endometrial aging can lead to defective stromal decidualization and impaired receptivity, both of which increase the risk of early pregnancy loss [ 27 , 28 ]. Similarly, advanced paternal age has been associated with decreased sperm motility, increased morphological defects, and impaired embryo quality, further elevating the risk of BPL [ 29 ]. Basal AFC, a key marker of ovarian reserve, also plays a crucial role in the risk of BPL. A reduced AFC reflects a diminished follicular pool and impaired oocyte quality, leading to fewer high-quality embryos and limited transfer options, thereby increasing the risk of BPL [ 30 ]. To further clarify the independent relationship between these factors and BPL and to control for potential confounding variables identified in the univariate analysis, we conducted multivariate logistic regression. Variables such as female age, BMI, male age, basal AFC, and number of transfer cycles—factors known to influence embryo quality—were included as covariates. Although these variables were significantly associated with BPL in the univariate analysis, they did not emerge as independent predictors after multivariate adjustment. This lack of significance may be attributed to confounding effects or the adjustments made within the statistical model. For instance, while advanced maternal age is associated with an increased incidence of chromosomal abnormalities, the multivariate analysis did not substantiate age as an independent contributor to BPL. The analysis revealed that endometrial thickness prior to transformation remained a significant independent predictor of BPL following HRT-mediated FET with single blastocyst transfer (P < 0.05). Several studies have examined the relationship between endometrial thickness and pregnancy outcomes in assisted reproductive technology. Yang’s studies suggest that, for normal pregnancies, there is no statistically significant difference in endometrial thickness between BPL, SA, and normal pregnancies [ 31 ]. In contrast, other studies consistently report that a thinner endometrium is associated with increased risks of BPL and lower pregnancy rates. Zhang's research [ 32 ] fa statistically significant difference in endometrial thickness on the day of oocyte retrieval between the BPL and the CP group. The endometrial thickness in the CP group was significantly higher than that in the BPL group. Regarding cutoff values, various studies have identified different thresholds. Dickey et al. [ 33 ] suggested that an endometrial thickness of less than 9 mm on the peak day of luteinizing hormone (LH) surge significantly increases the risk of BPL. Our analysis, using smooth curve fitting, further corroborated these findings, showing that an endometrial thickness of less than 10 mm is associated with a higher incidence of BPL, consistent with previous results reported by Zeng [ 34 ]. The endometrium plays a crucial role in implantation, and sufficient thickness is essential for proper embryo attachment and development. Previous studies have shown that an endometrial thickness of less than 7 mm is considered thin and is associated with reduced pregnancy, ongoing pregnancy, and live birth rates in ART cycles [ 35 ]. Thin endometrium is typically characterized by inadequate glandular growth, impaired vascular development, and increased resistance to blood flow. Additionally, reduced expression of estrogen receptors (ER), integrin β3, leukemia inhibitory factor (LIF), and vascular endothelial growth factor (VEGF) may disrupt implantation and early embryonic development, thereby contributing to BPL [ 36 ]. Traditionally, a threshold of 7–8 mm has been used in clinical practice as the benchmark for endometrial transformation in FET cycles.However, our results suggest that in HRT or down-regulated artificial cycles, clinicians should consider delaying endometrial transformation until the endometrial thickness exceeds 10 mm, when feasible. This provides a more precise and evidence-based reference for reducing the risk of BPL in HRT-FET single blastocyst transfers. Nevertheless, when endometrial thickness was used as a risk assessment parameter for BPL in ROC analysis, the AUC was 0.547. This result highlights the limitations of our study's retrospective design and the variability in clinical practices. We therefore recommend exercising caution when applying this threshold universally in clinical decision-making. Despite the valuable insights gained from this study, several limitations must be acknowledged. First, the retrospective nature and single-center design of this study may limit the external validity of our findings. The results may not be generalizable to other settings or populations, as medical practices and patient demographics can vary significantly across institutions. Second, there may be variability in laboratory conditions, particularly concerning embryo freezing and thawing techniques. Differences in cryoprotectant usage, thawing rates, embryo handling protocols, and embryo placement could lead to inconsistencies in outcomes and introduce bias. These factors can greatly influence embryo viability and implantation success and should be carefully considered when interpreting our results. Furthermore, the cycles included in this study encompassed repeated cycles from the same patients, which introduces a potential risk of bias due to uncorrected repeated measurements. This may slightly overestimate the significance of the correlation analysis, resulting in a marginally smaller p-value. Due to autocorrelation within the data from the same patient, the effect size (odds ratio) of the unadjusted primary outcome measure may be slightly inflated. Given the limitations of our data system, we were unable to retain only independent cycles for each patient, and there was no clear method for selecting which cycle to include. However, considering the large volume of cycles analyzed, we believe that this issue does not significantly affect the overall directional conclusions. Additionally, while we made efforts to control for common confounding factors, there may still be residual confounding variables that were not measured or included in our analysis, such as genetic or immunological factors that could affect embryo development and implantation rates. These unaccounted variables could potentially influence our conclusions about the associations between embryo quality, transfer timing, and BPL risk. Finally, as the retrospective design of this study precludes the establishment of causality, we emphasize the need for prospective, multicenter studies. Such research would not only help validate our findings but also contribute to a more comprehensive understanding of the factors influencing pregnancy outcomes in FET protocols.

Supplementary Material 1 Supplementary Material 1