Establishing a predictive model for ectopic pregnancy risk following assisted reproductive technology.

A total of 6872 cycles were screened for clinical data who met the inclusion criteria of this study, The dataset was randomly split into a training set (5496 cycles) and a validation set (1376 cycles) at a rate of approximately four to one. The initial clinical characteristics of training and validation groups are presented in Table  1 . Table 1 Baseline characteristics of the patients Validation sets( n  = 1376) Training sets( n  = 5496) t/χ 2 value p value Female age 31.67 ± 3.97 31.56 ± 4.03 0.853 0.245 Year of treatment 2010 69.00 (5.00%) 259.00 (4.70%) 2011 156.00 (11.30%) 649.00 (11.80%) 2012 214.00 (15.60%) 797.00 (14.50%) 2013 200.00 (14.50%) 873.00 (15.90%) 2014 240.00 (17.40%) 941.00 (17.10%) 2015 234.00 (17.00%) 957.00 (17.40%) 2016 240.00 (17.40%) 919.00 (16.70%) 2017 23.00 (1.70%) 101.00 (1.80%) EMT (mm) 9.69 ± 1.74 9.59 ± 1.67 1.984 0.011 BMI(kg/m 2 ) 21.75 ± 3.40 21.79 ± 2.95 -0.414 0.179 Underweight 0.759 0.384 NO 1251.00 (90.90%) 4954.00 (90.10%) YES 125.00 (9.10%) 542.00 (9.90%) Infertility type 0.712 0.399 primary 780.00 (56.70%) 3046.00 (55.40%) secondary 596.00 (43.30%) 2450.00 (44.60%) Endometriosis 0.697 0.404 NO 1304.00 (94.80%) 5238.00 (95.30%) YES 72.00 (5.20%) 258.00 (4.70%) PCOS 0.427 0.514 NO 1307.00 (95.00%) 5196.00 (94.50%) YES 69.00 (5.00%) 300.00 (5.50%) LH(mIU/ml) 4.973 ± 3.91 5.227 ± 5.74 -1.557 0.140 FSH(mIU/ml) 8.58 ± 8.21 8.51 ± 11.92 0.192 0.912 T(ng/ml) 0.58 0.29 0.61 ± 0.35 0.827 0.410 History of pelvic surgery 0.075 0.784 NO 859.00 (62.40%) 3409.00 (62.00%) YES 517.00 (37.60%) 2087.00 (38.00%) Hydrotubation 0.399 0.527 NO 1170.00 (85.00%) 4710.00 (85.70%) YES 206.00 (15.00%) 786.00 (14.30%) Caesarean section 0.003 0.953 NO 1358.00 (98.70%) 5423.00 (98.70%) YES 18.00 (1.30%) 73.00 (1.30%) History of induced abortion 1.723 0.189 NO 1065.00 (77.40%) 4161.00 (75.70%) YES 311.00 (22.60%) 1335.00 (24.30%) Type of infertility 3.324 0.068 Non-tubal factor 451.00 (32.80%) 1662.00 (30.20%) Tubal factor 925.00 (67.20%) 3834.00 (69.80%) Fresh or frozen embryo transfer 0.006 0.939 Fresh 632.00 (45.90%) 2518.00 (45.80%) Frozen 744.00 (54.10%) 2978.00 (54.20%) Male infertility 4.330 0.037 NO 828.00 (60.20%) 3474.00 (63.20%) YES 548.00 (39.80%) 2022.00 (36.80%) Number of embryos transferred 0.367 0.545 1 744.00 (54.10%) 2978.00 (54.20%) > 1 632.00 (45.90%) 2518.00 (45.80%) IVF/ICSI 3.707 0.054 IVF 999.00 (72.60%) 4129.00 (75.10%) ICSI 377.00 (27.40%) 1367.00 (24.90%) Outcome 2.210 0.137 IUP 1308.00 (95.10%) 5276.00 (96.00%) EP 68.00 (4.90%) 220.00 (4.00%) Type of embryo transferred 0.957 cleavage 1254 (91.10%) 4989 (91.10%) 0.170 blastocyst 122 (8.90%) 507 (9.20%) EP surgery history 0.024 0.807 NO 1168 (84.90%) 4656 (84.70%) YES 208 (15.10%) 840 (15.30%) EP history 0.060 0.877 NO 1138 (82.70%) 4530 (82.40%) YES 238 (17.30%) 966 (17.60%) Ectopic pregnancy, EP; polycystic ovary syndrome, PCOS; body mass index, BMI; endometrial thickness, EMT; in vitro fertilization, IVF; intracytoplasmic sperm injection, ICSI Baseline characteristics of the patients Ectopic pregnancy, EP; polycystic ovary syndrome, PCOS; body mass index, BMI; endometrial thickness, EMT; in vitro fertilization, IVF; intracytoplasmic sperm injection, ICSI Table 2 Univariate analysis of the risk factors for EP in ART population IUP( n  = 5276) EP( n  = 220) EP rate OR 95%CI P -value Female age 31.58 ± 4.02 31.26 ± 4.23 0.981 0.948–1.014 0.259 Year of treatment 0.353 2010 246 (4.70%) 13 (5.90%) 5.28% 2011 613 (11.60%) 36 (16.40%) 5.87% 2012 762 (14.40%) 35 (15.90%) 4.59% 2013 842 (16.00%) 31 (14.10%) 3.68% 2014 908 (17.20%) 33 (15.00%) 3.63% 2015 937 (17.80%) 20 (9.10%) 2.13% 2016 885 (16.80%) 34 (15.50%) 3.84% 2017 83 (1.60%) 18 (8.20%) 21.69% EMT (mm) 9.60 ± 1.69 9.34 ± 1.32 0.904 0.828–0.986 0.023 BMI(kg/m 2 ) 21.70 ± 3.33 21.67 ± 3.07 0.998 0.958–1.039 0.913 Abnormal BMI 1.333 1.005–1.768 0.046 NO 3736 (70.80%) 142 (64.50%) 3.80% YES 1540 (29.20%) 78 (35.50%) 5.06% Infertility type 0.828 0.632–1.084 0.17 primary 2934 (55.60%) 108 (49.10%) 3.68% secondary 2342 (44.40%) 112 (50.90%) 4.78% Endometriosis 0.977 0.573–1.665 0.827 NO 5029 (95.30%) 209 (95.00%) 4.16% YES 247 (4.70%) 11 (5.00%) 4.45% PCOS 1.784 1.109–2.870 0.017 NO 4996 (94.70%) 200 (90.90%) 4.00% YES 280 (5.30%) 20 (9.10%) 7.14% LH(mIU/ml) 5.22 ± 5.76 5.32 ± 5.13 FSH(mIU/ml) 8.50 ± 12.09 8.89 ± 6.80 LH/FSH 0.69 ± 0.83 0.67 ± 0.58 T(ng/ml) 0.61 ± 0.35 ± 0.62 0.38 1.100 0.757–1.601 0.617 History of pelvic surgery 0.814 0.619–1.069 0.139 NO 3283 (62.20%) 126 (57.30%) 3.84% YES 1993 (37.80%) 94 (42.70%) 4.72% Hydrotubation 0.765 0.537-1.091 0.14 NO 4529 (85.80%) 181 (82.30%) 4.00% YES 747 (14.20%) 39 (17.70%) 5.22% Caesarean section 0.33 0.046-2.386 0.272 NO 5204 (98.60%) 219 (99.50%) 4.21% YES 72 (1.40%) 1 (0.50%) 1.39% History of induced abortion 0.938 0.682–1.291 0.696 NO 3992 (75.70%) 169 (76.80%) 4.23% YES 1284 (24.30%) 51 (23.20%) 3.97% Type of infertility 1.059 0.787–1.425 0.705 Non-tubal factor 1598 (30.30%) 64 (29.10%) 4.01% Tubal factor 3678 (69.70%) 156 (70.90%) 4.24% Fresh or frozen embryo transfer 0.561 0.427–0.739 <0.001 Fresh 2387 (45.20%) 131 (59.50%) 5.49% Frozen 2889 (54.80%) 89 (40.50%) 3.08% Male infertility 1.271 0.967–1.670 0.086 NO 3347 (63.40%) 127 (57.70%) 3.79% YES 1929 (36.60%) 93 (42.30%) 4.82% Number of embryos transferred 1.779 0.936–3.381 0.079 1 412 (7.80%) 10 (4.50%) 2.43% > 1 4864 (92.20%) 210 (95.50%) 4.32% IVF/ICSI 1.112 0.820–1.508 0.496 IVF 3968 (75.20%) 161 (73.20%) 4.06% ICSI 1308 (24.80%) 59 (26.80%) 4.51% Type of embryo transferred 0.927 0.574–1.498 0.758 cleavage 4788 (90.80%) 201 (91.40%) 4.20% blastocyst 488 (9.20%) 19 (8.60%) 3.90% EP surgery history 1.283 0.907–1.817 0.159 NO 4477 (84.90%) 179 (81.40%) 4.40% YES 799 (15.10%) 41 (18.60%) 5.13% EP history 1.326 0.984–1.886 0.063 NO 4359 (82.60%) 171 (77.70%) 3.92% YES 917 (17.40%) 49 (22.30%) 5.34% IUP, Intrauterine pregnancy; EMT, endometrial thickness Univariate analysis of the risk factors for EP in ART population IUP, Intrauterine pregnancy; EMT, endometrial thickness Univariate analysis revealed that EMT(OR = 0.90, 95% CI:0.83–0.99, P  = 0.023), abnormal BMI(23.9) (OR = 1.33, 95% CI: 1.01–1.77, P  = 0.046), PCOS (OR = 1.78, 95% CI: 1.11–2.87, P  = 0.017), fresh or frozen embryo transfer (OR = 0.56, 95% CI: 0.43–0.74, P  < 0.001), male infertility (OR = 1.27, 95% CI: 0.97–1.67, P  = 0.086), number of embryos transferred(OR = 1.78, 95% CI: 0.94–3.38, P  = 0.079), EP history(OR = 1.33, 95% CI: 0.98–1.89, P  = 0.063) were associated with EP (Table  2 ). The multivariate analysis demonstrated that EMT(mm), PCOS, EP history, fresh or frozen embryo transfer and male infertility factors were independently associated with EP ( P  < 0.05). Specifically, after adjusting for female age, BMI, EMT in mm, PCOS, history of EP, ET/FET, and male infertility factors, ART patients with PCOS were found to have an over two-fold increased risk of EP (aOR = 2.07, 95% CI: 1.27–3.36, P  = 0.004). Furthermore, frozen embryo transfer can significantly reduce the risk of EP compared to fresh embryo transfer (aOR = 2.17, 95% CI: 1.62–2.91, P  < 0.001). Male infertility factor was associated with 1.4-fold increased risk of EP (aOR = 1.39, 95% CI: 1.05–1.85, P  = 0.021). Women with EP history was associated with 1.4-fold increased risk of EP (aOR = 1.41, 95% CI: 1.01–1.97, P  = 0.046) (Table  3 ). Table 3 Multivariate analysis of the risk factors for EP in ART population OR 95%CI P value EMT (mm) 0.86 0.77–0.93 < 0.001 PCOS NO 1.00 YES 2.07 1.27–3.36 0.004 EP History NO 1.00 YES 1.41 1.01–1.97 0.046 ET/FET Frozen ET 1.00 Fresh ET 2.17 1.62–2.91 < 0.001 Male infertile factors NO 1.00 YES 1.39 1.05–1.85 0.021 Multivariate analysis of the risk factors for EP in ART population An increase of 1 mm in EMT is correlated with a 15% decrease in the odds of developing EP (aOR = 0.86, 95% CI: 0.77–0.93, P  < 0.001) (Table  3 ). After adjusting for female age, BMI, methods of ART, tubal factor, PCOS and male infertility factors, we also found that there is a significant increase in the risk of EP when the EMT is less than 9.7 mm (aOR = 1.37, 95% CI: 1.03–1.82, P  = 0.03). Risk predictors were identified through the utilization of the multivariate logistic regression model. In addition, although there was no statistical significance between abnormal BMI and the occurrence of EP, considering its high proportion (29.20% in IUP, 35.50% in EP), we included it as one of the risk factors in the prediction model. A predictive model in the form of a nomogram was developed by integrating the significant prognostic factors. The EMT (mm) prior to ET, PCOS, EP history, fresh or frozen embryo transfer were set as independent variables to predict the incidence of EP. The risk prediction model of EP was logit P =-2.282792 + 0.778825× (PCOS = 1) + 0.338919× (EP history = 1) + 0.337529 ×(male infertile factors = 1) + 0.754782×(frozen embryo transfer = 1)- 0.163352 × (endometrial thickness) (Fig.  2 ). Fig. 2 The nomogram presented in the modeling group provides a visual representation of the clinical prediction model. Each variable value for a patient can be determined by locating the corresponding point on the nomogram. By drawing a vertical line from the variable score to the total score axis, the total score can be determined. From the total score, a vertical line can be drawn down to the predicted EP rate line, providing the estimated probability of ectopic pregnancy (EP) for a patient. For example, a total score of 140 on the nomogram indicates a probability of more than 10% for EP The nomogram presented in the modeling group provides a visual representation of the clinical prediction model. Each variable value for a patient can be determined by locating the corresponding point on the nomogram. By drawing a vertical line from the variable score to the total score axis, the total score can be determined. From the total score, a vertical line can be drawn down to the predicted EP rate line, providing the estimated probability of ectopic pregnancy (EP) for a patient. For example, a total score of 140 on the nomogram indicates a probability of more than 10% for EP The model’s predictive performance was evaluated using Area Under the Curve (AUC), a commonly used metric for assessing the performance of binary classification models. AUC values range from 0.5 to 1, with higher values indicating better model performance. After validating the training and validation (Fig.  3 ) sets separately, it was observed that the AUC remained largely unchanged, indicating that the model is relatively stable. The calibration plot indicates that the predicted values and observed values in both the training and validation (Fig.  3 ) datasets are in good agreement. Fig. 3 ROC curves and calibration curves for the EP rate nomogram. The accuracy of our prediction model was validated on both the ( A )training and ( B )validation datasets. Calibration curve for ( C ) training set and ( D )validation EP rate nomogram. On the calibration curve, X axis is nomogram predicted probability of EP and Y axis is actual probability of EP ROC curves and calibration curves for the EP rate nomogram. The accuracy of our prediction model was validated on both the ( A )training and ( B )validation datasets. Calibration curve for ( C ) training set and ( D )validation EP rate nomogram. On the calibration curve, X axis is nomogram predicted probability of EP and Y axis is actual probability of EP

This retrospective cohort study was conducted at the International Peace Maternal and Child Health Hospital. Prior to the study, approval and consent were obtained from the Ethics Committee of the International Peace Maternity &Child Health Hospital. The approval number is (GKLW)2013-51. A total of 7622 cycles performed between 2010 and 2017 were enrolled in our study. Cycles with incomplete clinical data, preimplantation genetic testing (PGT) were excluded. At last, 6872 cycles were included in the study. The included patients did not have diabetes, thyroid dysfunction, pituitary dysfunction and other endocrinological diseases. Total of 6872 cycles was randomly split into a training set (80%, 5496 cycles) and a validation set (20%, 1376 cycles) using the random sampling (Fig.  1 ). The training set was used for risk factor screening. During this process, univariate and multivariate analysis were used to determine the factors associated with the predicted outcome. Fig. 1 Flow chart of study selection Flow chart of study selection All women underwent COH with either a standard long agonist protocol or antagonist protocol. The dosage of gonadotropins was adjusted based on the ovarian response, as monitored by ultrasound and serum sex hormone levels (including FSH, LH, E 2 , and progesterone). When at least two follicles reached a size of 18 mm or larger, human chorionic gonadotropin (HCG) was administered. Oocyte retrieval was conducted within 34–36h, guided by transvaginal ultrasound. Depending on sperm quality, either in vitro fertilization (IVF) or intracytoplasmic sperm injection(ICSI) was performed. Embryos were transferred on any day from day 2 to day 6 following oocyte retrieval, based on the patient’s condition, including embryo quality, abdominal distention, and endocrine examination results. Cycle cancellation was defined as the absence of fresh embryo transfer after oocyte retrieval, excluding cycles cancelled before HCG triggering. Luteal phase support was given for women planning to transfer fresh embryos as routine. For FET, endometrium was prepared using several protocols, including hormone replacement with or without down-regulation, ovarian stimulation, and natural cycles. Cleavage stage embryos were transferred on the second or third day after ovulation or progesterone administration, whereas blastocysts were transferred on the fifth or sixth day. One or two embryos were transferred according to the clinicians’ advice and patients’ preference. Serum HCG levels were measured 14 days after embryo transfer, and if conception occurred, the luteal phase support was continued. Transvaginal ultrasonography was performed 28 days after embryo transfer. In this study, the primary outcome measured was EP, while the secondary outcome measured was Intrauterine pregnancy (IUP). EP was defined as the implantation of a developing blastocyst outside of the endometrial cavity. IUP was defined as the successful implantation of at least one embryo in the uterus. According to the Chinese Adult BMI Grading Standard, the BMI for Chinese adults is classified into the following four levels: Underweight: BMI < 18.5; Normal weight: 18.5 ≤ BMI < 24; Overweight: 24 ≤ BMI < 28; Obesity: BMI ≥ 28. We collected underweight, overweight and obesity as abnormal BMI. Obstructive azoospermia, oligoasthenospermia, teratospermia, poor semen quality and other factors were identified as male infertility factors. Additionally, tubal factors have been incorporated into our risk assessment, which includes hydrosalpinx, tubal inflammation, and congenital or acquired tubal obstruction. R and SPSS were used for statistical analysis. The data with normal distribution were expressed as mean ± standard deviation (x ± s). Count data were expressed as rates or constituent ratios (%). Continuous variables were tested using t-tests, while categorical variables were assessed using chi-square tests. Both univariate and multivariate logistic regression analyses were conducted to examine the impact of different factors on EP. The prediction performance of the logistic regression model is evaluated using the area under the receiver operating characteristic curve (ROC) and calibration curve. The cut off value was derived from the ROC curve. P -value less than 0.05 was considered significant.

The extrauterine implantation of an embryo, known as ectopic pregnancy (EP), is a serious complication of assisted reproductive technology (ART). EP is responsible for approximately 1% of maternal deaths in developing countries, while the incidence is even higher in developed countries, accounting for up to 5% of maternal deaths [ 1 ]. As ART has become increasingly common and advanced, the occurrence of ectopic pregnancy following ART has been found to range from 2.1–8.6%[ 2 , 3 ], which is more frequent than 1–2% [ 3 ] observed in the natural pregnancy group. Therefore, it is crucial to better predict the occurrence of EP following ART to prevent the serious adverse outcomes it can lead to. Previous studies suggested that the risk of EP may be related to tubal factor [ 4 ], abnormal body mass index (BMI) [ 5 ], fresh or frozen-thawed embryo transfer(FET) [ 6 ], endometrial thickness (EMT) [ 7 ]and PCOS [ 8 ]. Thus, having a more accurate understanding of the association between these risk factors and EP secondary to ART would be very helpful in EP prediction. There are currently prediction models available for adverse outcomes following ART, but the accuracy of their predictions may be limited due to the lack of risk factors such as PCOS, BMI and fresh or frozen-thawed embryos that are associated with ART-related EP [ 7 ]. Considering the high proportion of PCOS patients in the ART population, including PCOS as a predictive factor may improve the accuracy of the prediction model. PCOS is a prevalent endocrine disorder in women of reproductive age and is the primary cause of anovulatory infertility[ 9 , 10 ]. Besides, changes in oocyte competence (OC) are considered a potential cause of reduced fertility in women with PCOS. The impact of OC on the reproductive potential of women with PCOS varies significantly, largely depending on the phenotype of PCOS and the associated comorbidities [ 11 ]. According to the Rotterdam criteria, at least two conditions must be met for the diagnosis of PCOS: irregular ovulation, clinical or biochemical hyperandrogen, and polycystic ovary morphology [ 12 ]. Women with PCOS face an increased risk of both early and late pregnancy complications. They are more likely to experience late complications, including pregnancy-induced hypertension, pre-eclampsia, gestational diabetes mellitus, and preterm delivery [ 13 ], as well as early complications like miscarriage compared to matched controls without PCOS [ 14 ]. ART have provided a glimmer of hope for PCOS patients who struggle with infertility, but it has also brought about various negative effects to the patients, including ovarian hyperstimulation syndrome, miscarriage, gestational diabetes, preeclampsia, and very preterm birth[ 15 , 16 ]. Wang.et al. found that the overall incidence of adverse pregnancy outcomes (including EP, miscarriage, preterm birth) was significantly higher in PCOS phenotype A and D groups than in control group after IVF/ICSI treatment [ 17 ]. Liu.et al. found PCOS was associated with an increased risk of EP after controlled ovarian hyperstimulation (COH) in fresh ET cycles [ 8 ]. Although PCOS has been considered as a risk factor for EP following ART [ 18 ], the exact relationship between them is still inconclusive. This study identified PCOS as an important factor by evaluating multiple risk factors affecting EP in ART population, and used these factors to establish a potential risk assessment model to reduce the incidence of EP in ART pregnancy, aiming at early prevention or identification of EP, ensuring the quality of early intervention and treatment, and avoiding serious consequences.

This large retrospective cohort study of 6872 embryo transfer cycles has found that PCOS is an independent risk factor for EP after ART. Additionally, we also found EMT (mm) prior to ET, underweight, fresh or frozen embryo transfer were significantly correlated with the incidence of EP. In addition, this study established a predictive model for the incidence of EP to provide advice for each specific patient to evaluate the incidence of EP before embryo transfer. Though we found significant difference of EP risk between PCOS and non-PCOS group, the lack of data on certain factors related to PCOS, such as AMH level, blood glucose level, hypertension [ 19 ], makes it difficult to confirm whether PCOS is directly or indirectly related to EP. Therefore, more comprehensive research is needed to understand the mechanism behind this relationship. Moreover, since this study is based on a retrospective design from a single medical center, there may be many unknown confounding factors. Therefore, larger-scale multicenter prospective studies are needed to better address this issue. PCOS is one of the leading causes of infertility in women of child-bearing age [ 20 ]. Our study demonstrated a significant increase of EP risk (aOR = 2.07, 95% CI: 1.27–3.36, P  = 0.004) in PCOS patients following ART treatment. Our findings are consistent with the results reported by Wang et al. Their study showed a significantly higher incidence of adverse pregnancy outcomes in PCOS phenotype A and D groups compared to the control group after IVF/ICSI treatment. However, when ectopic pregnancy was analyzed separately, there were no significant differences between the PCOS group and the control group. This could be attributed to the limited sample size of ectopic pregnancy cases in both the PCOS and control groups (7/346 and 13/453) [ 17 ]. According to our study, the findings are also consistent with the research conducted by Liu et al. They found that PCOS was associated with an increased risk of EP after COH in fresh embryo transfer cycles ( n  = 3303, 7.0% vs. 2.4%; aOR, 3.06; 95% CI, 1.34–6.96). However, in cryo-thawed ET cycles, there was no significant association between PCOS and EP ( n  = 2036, 2.2% vs. 2.0%; aOR, 0.94; 95% CI, 0.22–4.07) [ 8 ]. Additionally, the impaired oocyte competence associated with PCOS may lead to a decline in embryo quality [ 11 ], which can affect implantation and consequently increase the risk of EP. Although the mechanism between PCOS and EP is unclear, clinical characteristics of PCOS can lead to dysregulation of endometrial sex hormone receptors, increased endometrial insulin resistance, chronic low-grade inflammation, immune dysfunction, altered uterine vascularity, abnormal endometrial gene expression, and cellular abnormalities and maybe the potential causes of the elevated risk of EP [ 21 ]. Women with PCOS were reported with higher levels of anti-Müllerian hormone (AMH) in their serum. Granulosa cells in these women have up to a 7.5-fold increase in AMH production [ 22 ], and anovulatory women with PCOS have AMH serum levels that are 18 times higher than the control group [ 23 ]. Considering previous research has shown that high levels of AMH are also associated with an increased risk of EP in fresh embryo transfer cycles [ 24 ], the underlying mechanism may correlate with elevated AMH level in PCOS patients. Additionally, this also highlights the importance of monitoring AMH level during ART cycles. Many studies have shown that the incidence of EP is lower in FET cycles compared to fresh embryo transfer cycles, which is consistent with our results[ 6 , 25 , 26 ]. The reduced risk of EP in FET cycles compared to fresh ET cycles was first reported by Ishihara et al. through the analysis of the Japanese ART registry [ 27 ]. A recent meta-analysis systematically evaluated the risks of ectopic pregnancy (EP) associated with frozen-thawed versus fresh blastocyst ET by synthesizing data from fourteen retrospective studies ( n  = 251,762 cycles). The analysis demonstrated that the incidence of EP in fresh single blastocyst ET cycles (1.2%) was significantly higher compared to that observed in frozen-thawed blastocyst ET cycles (0.80%) [ 28 ]. Similarly, according to a systematic review conducted by Wang et al. in 2021, FET was found to yield better outcomes in IVF compared to fresh ET. This improvement could be attributed to the achievement of better synchronization between the embryo and the endometrium through FET cycles [ 29 ]. One notable distinction between fresh and frozen cycles lies in the ovarian stimulation protocols employed. A previous study investigated the impact of these protocols on the risk of EP and found that the EP rate for natural (unstimulated) IVF cycles was 0.47%, which was significantly lower than the rates observed in stimulated cycles (ranging from 1.47 to 2.18%). Notably, the EP rate for unstimulated fresh cycles (0.47%) was comparable to the EP rate for all frozen-thawed cycles (0.52%) analyzed in the current investigation. These findings suggest that ovarian stimulation may play a mediating role in the increased risk of EP associated with fresh cycles [ 30 ]. Although FET may lower EP rate, there are also reports suggesting a higher incidence of biochemical pregnancy and pregnancy loss following FET [ 31 ]. Thus, the advisability of freeze-all cycles remains a topic of controversy, and further research is needed to develop individualized treatment plans for patients. The results of our study suggest that thinner EMT prior to embryo transfer was associated with higher EP rate, which is consistent with previous multiple studies[ 7 , 32 , 33 ]. Additionally, our results demonstrate that women with EMT less than 9.7 mm had 1.37-fold increased risk comparing with the women with thicker EMT. A recent systematic review finds that patients undergoing IVF/ICSI and having an EMT less than 8 mm are found to have an elevated risk of developing an EP [ 32 ]. According to another study, women who underwent ART and had an EMT more than 7.6 mm had a significantly lower risk of EP compared to women with an EMT of less than 7.6 mm [ 7 ]. Although there are various standards of thin endometrial thickness, the general trend of research results is consistent, which shows that the thinner the EMT, the higher the risk of EP. Besides, experimental and clinical data indicate that the endometrium of women with PCOS differs from that of healthy controls. The clinical features associated with this syndrome, either individually or in combination, may lead to dysregulation of sex hormone receptors and coreceptors in the endometrium of women with PCOS. This dysregulation can increase endometrial insulin resistance, accompanied by impaired glucose transport and utilization, resulting in chronic low-grade inflammation, immune dysfunction, altered uterine vasculature, abnormal endometrial gene expression, and cellular abnormalities [ 21 ]. While our study provides evidence, larger and more targeted multi-center clinical studies are needed to confirm the more accurate range of values for this relationship. The results of this study suggest that male infertility factor is associated with a 1.4-fold increase in EP risk, which is consistent with previous studies [ 34 ]. Many authors have suggested that semen quality related with embryo quality [ 35 , 36 ], which has also been considered as a significant aetiology for EP [ 37 ]. However, there is still limited research on the mechanism between male infertile factor and EP. While our study provided some evidence, further targeted research based on larger cohorts is still required to better understand the relationship between male infertility factors and EP.

In conclusion, our study demonstrated a strong correlation between PCOS and EP. In alternative ART cases, frozen embryo transfer may be a better choice for decreasing the incidence of EP in the treatment. The nomogram presented in this study and the included associated risk factors may provide new advice for the preparation phase of clinical ART cycles as well as early prediction and identification of EP.