Abstract
Purpose
Hydrosalpinx reduces the pregnancy outcomes of in vitro fertilization and embryo transfer (IVF-ET). Tubal occlusion has been shown to improve IVF-ET success; however, the optimal timing of oocyte retrieval and the molecular changes following tubal occlusion remain unclear.
Methods
Retrospective and prospective cohorts of patients with hydrosalpinx were analyzed. Logistic regression, piecewise regression, and mediation effect analyses were used to identify the optimal timing of oocyte retrieval and the impact of tubal occlusion on pregnancy outcomes. Bulk and single-cell RNA sequencing (RNAseq) were applied to investigate potential molecular changes underlying these effects.
Results
In total, 976 patients with hydrosalpinx were included. Logistic regression showed that delayed oocyte retrieval after occlusion significantly reduced clinical pregnancy (multivariate-adjusted: odds ratio [OR] = 0.904, P = 0.001) and live birth (multivariate-adjusted: OR = 0.926, P = 0.010). Curve estimation and piecewise regression indicated improved pregnancy outcomes within 7 months after occlusion. These findings were consistent regardless of baseline hormone levels or ovarian reserve. Endometrial RNAseq revealed activation of immune-related pathways before occlusion—including human T-cell leukemia virus 1 infection, natural killer cell–mediated cytotoxicity, cellular senescence, antigen processing and presentation, and complement and coagulation cascades—that were inactivated after occlusion, alongside upregulation of CXCL14 expression. CIBERSORTx identified a higher proportion of T follicular helper cells before occlusion (P = 0.02) and increased M2 macrophage infiltration after occlusion (P = 0.029), which was confirmed by CD163 immunohistochemistry. Finally, single-cell RNAseq suggested that CXCL14 expression in endometrial epithelial cells may interact with macrophages to promote M2 polarization, fostering a low-inflammatory microenvironment favorable for IVF-ET outcomes.
Conclusion
This study demonstrates that oocyte retrieval within 7 months after tubal occlusion is optimal for patients with hydrosalpinx, potentially due to CXCL14-mediated modulation of the endometrial immune microenvironment.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13048-025-01803-2.
Keywords: Hydrosalpinx, IVF-ET, CXCL14, M2-macrophage, Tubal occlusion
Introduction
Female infertility is a serious problem with far-reaching consequences for humanity. Approximately 25% of female infertility cases have been attributed to tubal factors [1]; and 30% of these tubal factors are classified as hydrosalpinx [1], which may be associated with pelvic inflammation. Hydrosalpinx has been linked to a dramatic decrease in pregnancy rates following in vitro fertilization-embryo transfer (IVF-ET) [2], possibly due the presence of inflammation in the endometrium [3]. It is now widely accepted that hydrosalpinx significantly reduces pregnancy rates for patients undergoing assisted reproductive technology (ART) treatment. For example, the success rate of IVF-assisted pregnancy for patients with hydrosalpinx was found to be approximately 30–50% lower than for those without hydrosalpinx [4]. In 2019, Harb et al. [5] conducted a large-sample meta-analysis of 14 studies to explore the impact of hydrosalpinx on pregnancy outcomes. The results of the meta-analysis indicate that hydrosalpinx is associated with a significant increase in miscarriage rate. Moreover, successful hydrosalpinx treatment led to significant improvements in IVF-ET pregnancy outcomes, coupled with decreased miscarriage rates [5]. Moreover, Peng et al. found that hydrosalpinx could result in a marked increase in the occurrence of chronic endometritis [6]; while promoting infiltration of inflammatory cells in the endometrium and impairing reproductive outcomes [7]. Thus, successful treatment of hydrosalpinx may effectively prevent excessive inflammation in the endometrium, and have positive effects on IVF-ET outcomes.
To date, several surgical procedures, including ultrasound-guided aspiration, laparoscopic salpingectomy, and proximal tubal occlusion + distal salpingostomy (hereafter tubal occlusion) have been successful in improving pregnancy outcomes when conducted before IVF-ET [8]. Ultrasound-guided aspiration of hydrosalpinx is a simple procedure that is associated with less damage and relatively lower cost than other methods. However, increased hydrosalpinx recurrence rates and lower efficacy with respect to improving pregnancy outcomes has restricted its application [9, 10]. Before the advent of IVF, laparoscopic salpingectomy was a commonly used pre-treatment method for hydrosalpinx patients. However, the secondary damage to surrounding tissues associated with this surgical procedure could markedly reduce ovarian reserves and ovulation induction [9, 10]. Laparoscopic bilateral tubal occlusion is a comparatively less invasive surgical procedure which could prevent the retrograde flow of hydrosalpinx fluid into the uterine cavity, improving endometrial receptivity, and achieving the same effect as fallopian tube resection. A 2020 meta-analysis reported that tubal occlusion could also significantly improve pregnancy outcomes compared with salpingectomy [11]. Although, prior surgical intervention was beneficial, the exact timing of oocyte retrieval following surgery was not reported [12]. Yilei et al. [13] recommended that oocyte retrieval should be performed within 4–12 months after salpingectomy, in order to improve pregnancy outcomes and live birth rates. As with other routine operations, tubal occlusion has also proved to be beneficial prior to IVF-ET, similar to other methods applied before IVF [14]. Moreover, tubal occlusion may result in significantly less damage to the ovarian reserve compared to salpingectomy [15]. To date, few studies have reported oocyte retrieval times following tubal occlusion, or its impact on the endometrial microenvironment.
Methods and materials
Study participants
Patients diagnosed with hydrosalpinx who visited the Reproductive Hospital of the Zhuang Autonomous Region between January 2016 and December 2022 for IVF or intracytoplasmic sperm injection (ICSI) were considered for inclusion. All patients underwent either proximal tubal occlusion or distal salpingostomy prior to controlled ovarian hyperstimulation (COH) and embryo transfer (ET). The exact day of oocyte retrieval after tubal occlusion was recorded, with 30 days defined as 1 month. The eligibility criteria were as follows: age of 20–45 years; regular menstruation; participation in a long gonadotropin-releasing hormone (GnRH) agonist protocol; hysterosalpingography (HSG) or vaginal ultrasound showing hydrosalpinx in one or both fallopian tubes, for which laparoscopic proximal occlusion and distal stoma of both fallopian tubes had been performed; IVF/ICSI performed within 365 days; and negative test results for chlamydia/gonorrhea infection. The exclusion criteria were endometriosis, uterine adenomyosis, prior ovarian surgery, uterine malformations (e.g., mediastinum, unicornis, or bicornis), endometrial lesions, severe intrauterine adhesions, other endocrine disorders, or impaired ovarian function due to radiotherapy, chemotherapy, or autoimmune disease. For patients with hydrosalpinx who first underwent occlusion and then IVF, fresh embryo transfer was carried out when conditions were suitable. If fresh transfer failed, frozen–thawed embryo transfer was performed. Single blastocyst transfer was generally preferred to maximize safety and outcomes. However, for older patients or those with prior failed implantation, two embryos could be considered. Regardless of whether cleavage-stage embryos or blastocysts were used, no more than two embryos were transferred. The primary endpoint was defined as either a live birth during the IVF cycle or exhaustion of all embryos available for transfer. All participants provided informed consent before laparoscopic bilateral proximal tubal occlusion. The study was approved by the Ethics Committee of the Reproductive Hospital of Guangxi Zhuang Autonomous Region in accordance with the Declaration of Helsinki (Approval No.: KY-LL-2022-13).
Clinical information
Clinical information specific to this study was collected, including age; body mass index (BMI); antral follicle count (AFC); infertility duration; basal follicle-stimulating hormone (FSH), luteinizing hormone (LH), and estradiol (E2) levels; and the proportion of primary versus secondary infertility. Ovarian response was assessed by monitoring LH, progesterone (P), and E2 levels, along with other factors such as the initial gonadotropin (Gn) dose, total Gn dose for ovulation stimulation, duration of Gn administration (days), endometrial thickness on the human chorionic gonadotropin (hCG) trigger day, and the number of oocytes retrieved. Embryo and oocyte data were also recorded, including oocyte maturation rate, normal fertilization rate, total cleavage rate, number of available cleavage-stage (D3) embryos, good-quality D3 embryo rate, blastocyst (D5/D6) formation rate, and high-quality blastocyst formation rate. When one principal follicle reached ≥ 18 mm in diameter, or two principal follicles reached ≥ 17 mm, or three principal follicles reached ≥ 16 mm, and hormone levels were optimal, 6000–8000 IU urinary-derived hCG [recombinant 250 µg hCG (Merck Serono S.p.A) combined with 2000 IU hCG (Livzon Pharmaceutical Group Inc.)] was administered as a trigger for ovulation. Transvaginal ultrasound-guided oocyte retrieval was then performed 34–36 h after hCG administration. Embryos classified as Grade II with ≥ 6 cells and derived from 2PN were rated as high-quality D3 embryos, while Grade III embryos with ≥ 4 cells were rated as usable embryos. Blastocysts graded 3BB or above were defined as high-quality blastocysts, whereas those graded 3BC or above were considered usable blastocysts [16].
Subgroup analysis
Subgroup analysis was based on the interval between tubal occlusion and oocyte retrieval, which was divided into 13 months using 30 days as 1 month. For further analyses, patients were grouped into ≤ 3 months, 3–7 months, and > 7 months. Patients aged ≥ 35 years were classified as advanced age, and low ovarian reserve was defined as an antral follicle count (AFC) of < 5–7 or anti-Müllerian hormone (AMH) level of < 1.1 ng/mL.
Statistical analysis
Statistical analysis was performed according to the interval between tubal occlusion and oocyte retrieval. One-way analysis of variance or the Welch test was used for continuous variables, and the chi-square test was applied for categorical variables. Logistic regression analyses were conducted to evaluate the association between the timing of oocyte retrieval after tubal occlusion and pregnancy outcomes, including clinical pregnancy, abortion, multiple pregnancy, ectopic pregnancy, and live birth. Binary logistic regression was performed using unadjusted, age-adjusted, and multivariate-adjusted models. In the multivariate-adjusted models, the following covariates were included: age, BMI, AFC, infertility duration, basal FSH, basal LH, and basal E2 levels. Additionally, the time to oocyte retrieval after tubal occlusion was divided into three groups (≤ 3 months, 3–7 months, and > 7 months), with the ≤ 3 month group used as the reference to further assess the potential influence of occlusion timing on pregnancy outcomes. All analyses were carried out using SPSS version 24.0 (IBM Corp., Armon, NY, USA). All statistical tests were two-tailed, and P < 0.05 was considered statistically significant.
Curve estimation
Curve estimation was conducted to identify the optimal timing of oocyte retrieval after tubal occlusion using binary logistic regression analysis. Dependent factors included clinical pregnancy, abortion, multiple pregnancy, ectopic pregnancy, and live birth, with adjustments made for age, BMI, AFC, infertility years, and basal FSH, LH, and E2 levels. The probability of pregnancy outcomes was applied in curve estimation using linear, logarithmic, inverse, quadratic, cubic, compound, power, S, growth, exponential, and logistic models. The R square value, adjusted R square value, and P-value were used to evaluate model fit, with the size of R square indicating the rationality of model construction. P < 0.05 was considered statistically significant.
Piecewise regression
Piecewise regression analysis was performed to identify the time point of oocyte retrieval after occlusion at which adverse assisted pregnancy outcomes occurred. In this method, the independent variables follow a linear relationship within a certain range, while another linear equation with a different slope is applied in other ranges. The segmented package in R software (version 4.4.1) was used. Logistic regression distribution probabilities of clinical pregnancy, live birth, and IVF-ET pregnancy time points were used to construct a linear regression model for estimating the segmented node. In addition, with clinical pregnancy outcome and live birth as dependent variables and the oocyte retrieval time after occlusion as the independent variable, generalized linear models were constructed for piecewise regression analysis, and the segmented node was again obtained.
Mediation effect
Mediation effect analysis was conducted to explore the role of ovarian reserve and hormone levels in influencing IVF-ET pregnancy outcomes because previous analyses had identified significant differences in these factors after tubal occlusion. The principle is as follows: when the independent variable X influences the dependent variable Y, and this influence occurs through another variable M, then M is considered a mediating variable. The effect of X on Y through M is referred to as the mediation effect (ME). In this causal pathway, M lies between X and Y, meaning that X leads to Y indirectly via M. Based on this framework, AFC, LH, E2, and FSH were used as mediating variables to compare the effects of IVF-assisted pregnancy timing (≤ 3 months, 3–7 months, and > 7 months after occlusion) on outcomes such as clinical pregnancy, live birth, miscarriage, and multiple pregnancy. The mediation package in R 4.4.1 was applied for the analysis. Both linear regression and generalized linear regression models were used, with multivariable adjustments for age, BMI, infertility duration, and infertility type. The final evaluation indices included average causal mediation effect (ACME, indirect effect), average direct effect (ADE), and total effect (TE), which together explain the association between X and Y through the mediating variables.
Ovarian reserve and hormone levels after tubal occlusion
Ovarian reserve and hormone levels were prospectively followed in patients scheduled for laparoscopic proximal fallopian tube disconnection combined with distal salpingostomy due to hydrosalpinx at Guangxi Reproductive Hospital between October 2021 and December 2023. Changes in ovarian function and hormone levels were recorded before occlusion and at 1, 2, 3, and 7 months after occlusion to evaluate the potential influence of ovarian reserve and hormone levels on assisted pregnancy outcomes.
Samples for RNA sequencing (RNAseq)
In total, 10 endometrial tissue samples were collected at the Reproductive Hospital of Guangxi Zhuang Autonomous Region between January 2024 and July 2024, including 5 samples obtained before tubal occlusion and 5 samples collected within 7 months after occlusion, under hysteroscopic guidance. Endometrial sampling was performed during the follicular phase, within 2–5 days after the end of menstruation. These samples were used to explore possible molecular mechanisms underlying the improved pregnancy outcomes observed within 7 months of tubal occlusion. All patients provided written informed consent prior to surgery. RNA was extracted using a commercial RNA extraction kit (Qiagen) according to the manufacturer’s instructions and subsequently used for RNAseq.
Differential expression gene analysis of RNAseq
Differential expression gene analysis of RNAseq was performed as follows. TrimGalore (https://github.com/FelixKrueger/TrimGalore/) was first used for data quality control, with the Phred quality score threshold set to 20 and read length > 20, along with other default parameters. Alignment of clean reads was then carried out using HISAT2 (version 2.1.0) with the NCBI GRCh38 reference genome (https://genome-idx.s3.amazonaws.com/hisat/grch38_genome.tar.gz). Gene expression levels were quantified with FeatureCounts (http://bioinf.wehi.edu.au/FeatureCounts/). Differentially expressed genes (DEGs) were identified using the DESeq2 package, comparing samples before and after occlusion. Visualization of results was performed with the ggplot and pheatmap packages. All analyses were conducted in R version 4.3.1.
KEGG pathway enrichment and GO annotation
The clusterProfiler package was used for KEGG enrichment and GO annotation of DEGs. Significant genes (P < 0.05 and |logFC| >1) were used to interpret the corresponding pathways and functional changes before and after tubal occlusion.
Gene set enrichment analysis (GSEA)
GSEA (http://software.broadinstitute.org/gsea/index.jsp) was performed to identify pathway activation before and after occlusion using DEGs (P < 0.05 and |logFC| >0.6). The Molecular Signatures Database (MSigDB, c2.cp.kegg.v7.4.symbols.gmt) was used as the reference set. KEGG signaling pathways were considered significantly enriched with P < 0.05 and |NES| >2.0.
CIBERSORTx
CIBERSORTx was used to estimate the abundance of cell types within the mixed cell population. The input signature matrix file consisted of 22 immune cell gene expression profiles provided by CIBERSORTx, while the mixture file was the RNAseq expression matrix before and after tubal occlusion. Analysis was performed using 1000 permutation tests.
Single-cell sequencing analysis
Single-cell data (GSE183837) from normal endometrium were obtained from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The expression matrix was processed using the Seurat R package (version 4.0.3), with strict quality control to ensure reliable results. Genes expressed in fewer than 10 cells and cells expressing fewer than 200 genes were excluded. To avoid low-quality or double cells, only cells with fewer than twice the median number of genes were retained, and the proportion of mitochondrial genes was limited to < 40%. To minimize potential batch effects, the functions SelectIntegrationFeatures, FindIntegrationAnchors, and IntegrateData in Seurat were applied to merge the single-cell data. Based on ElbowPlot results, 20 principal components were selected for dimensionality reduction using uniform manifold approximation and projection. The resolution parameter was set to 0.2 to identify distinct cell clusters. DEGs were calculated using the likelihood ratio test implemented in FindAllMarkers and FindMarkers.
Immunohistochemistry and immunofluorescence
Endometrial tissues were fixed in 4% paraformaldehyde for 24 h, then dehydrated and embedded in paraffin blocks. The blocks were sectioned into 4-µm pathological slices and stained using immunohistochemistry and immunofluorescence kits. For immunohistochemistry, the M2 macrophage marker CD163 (ab182422, diluted 1:500) was used. For immunofluorescence, primary antibodies included CD163 (ab182422, 1:500) and KRT8/18 (MA5-14088, 1:250). Alexa Fluor 488–conjugated donkey anti-mouse (1:200) and Alexa Fluor 555–conjugated donkey anti-rabbit (1:200) were used as secondary antibodies. Images were captured with an Olympus microscope.
Results
Time of oocyte retrieval after occlusion
In this study, 976 patients with hydrosalpinx were recruited. To identify the optimal timing of oocyte retrieval after occlusion, logistic regression analysis was performed. The results indicated that delaying oocyte retrieval after occlusion significantly reduced clinical pregnancy rates (unadjusted: odds ratio [OR] = 0.882, 95% confidence interval [CI] = 0.836–0.931, P < 0.001; multivariate-adjusted: OR = 0.904, 95% CI = 0.851–0.959, P = 0.001) and live birth rates (unadjusted: OR = 0.908, 95% CI = 0.863–0.956, P < 0.001; multivariate-adjusted: OR = 0.926, 95% CI = 0.874–0.982, P = 0.010) (Table 1). To further define the time points for oocyte retrieval, curve estimation was performed for the probability of clinical pregnancy (R2 = 0.264) and live birth (R2 = 0.184), with the cubic curve showing the best fit. Both curves suggested higher rates of clinical pregnancy and live birth within 3 months (Fig. 1). In addition, linear regression and piecewise logistic regression analyses indicated that 7 months might represent a critical time point, after which the probability of clinical pregnancy and live birth declined (Fig. 2).
Table 1.
Logistic regression analysis of pregnancy outcomes and IVF-ET assisted pregnancy interval after tubal occlusion
| Unadjusted | Multi-adjusted | ||||||
|---|---|---|---|---|---|---|---|
| OR | 95%CI | P | OR | 95%CI | P | ||
| Clinical pregnancy | 0.882 | 0.836–0.931 | < 0.001* | 0.904 | 0.851–0.959 | 0.001* | |
| Abortion | 0.982 | 0.897–1.075 | 0.696 | 1.035 | 0.932–1.149 | 0.519 | |
| Multiple pregnancies | 1.021 | 0.963–1.083 | 0.484 | 1.045 | 0.981–1.112 | 0.175 | |
| Live birth | 0.908 | 0.863–0.956 | < 0.001* | 0.926 | 0.874–0.982 | 0.010* |
* Multivariate correction factors included age, BMI, AFC, infertility years, basal FSH level, basal LH level, basal E2 level, total number of embryos transferred, Initial dose of Gn, total Gn, number of eggs harvested and number of high quality blastocyst. * indicated a statistical difference
Fig. 1.
Curve estimation of clinical pregnancy (a) and live birth (b) probability from Logistic regression with the time of oocyte retrieval after tubal occlusion. * the left is all the curve modes, and the right is the most significant curve mode
Fig. 2.
Piecewise logistic regression predicted the time node of oocyte retrieval after tubal occlusion according to the clinical pregnancy and live birth rate. (a) Linear model; (b) Generalized Linear model
Pregnancy outcomes relative to timing of oocyte retrieval after occlusion
According to the curve and piecewise analyses, 976 patients were divided into three groups (≤ 3 months, 3–7 months, and > 7 months after occlusion). Significant differences were observed in age (≤ 3 months: 33.11 ± 4.36; 3–7 months: 32.65 ± 4.33; >7 months: 34.18 ± 3.85; P < 0.001), infertility duration (≤ 3 months: 5.01 ± 3.50; 3–7 months: 5.10 ± 3.90; >7 months: 4.36 ± 3.91; P = 0.046), and total number of embryos transferred (≤ 3 months: 2.42 ± 1.34; 3–7 months: 2.69 ± 1.38; >7 months: 2.21 ± 1.17; P < 0.001). In addition, a decline in AFC (≤ 3 months: 23.03 ± 9.40; 3–7 months: 22.61 ± 8.85; >7 months: 20.70 ± 9.44; P = 0.017) and an increase in basal FSH levels (≤ 3 months: 7.32 ± 2.00; 3–7 months: 7.34 ± 1.70; >7 months: 7.76 ± 1.83; P = 0.008) were observed with longer intervals before oocyte retrieval. The number of oocytes retrieved (≤ 3 months: 13.94 ± 6.91; 3–7 months: 13.68 ± 6.01; >7 months: 12.49 ± 6.90; P = 0.048), available D3 embryos (≤ 3 months: 10.34 ± 5.40; 3–7 months: 10.24 ± 4.78; >7 months: 8.86 ± 5.69; P = 0.001), and high-quality blastocyst formation rate (≤ 3 months: 39.75%; 3–7 months: 39.16%; >7 months: 35.14%; P = 0.014) all declined with later oocyte retrieval. Conversely, the rate of biochemical abortion increased (≤ 3 months: 1.52%; 3–7 months: 5.24%; >7 months: 6.71%; P = 0.011), while cumulative live birth rates decreased markedly (≤ 3 months: 76.34%; 3–7 months: 71.53%; >7 months: 58.77%; P < 0.001) (Table 2).
Table 2.
Comparisons of basic information after tubal occlusion
| ≤ 3 months (N = 186) | 3–7 months (N = 562) | > 7 months (N = 228) | P | |
|---|---|---|---|---|
| Age (years) | 33.11 ± 4.36 | 32.65 ± 4.33 | 34.18 ± 3.85 | < 0.001 |
| Infertility age (years) | 5.01 ± 3.50 | 5.10 ± 3.90 | 4.36 ± 3.91 | 0.046 |
| BMI (kg/m2) | 21.75 ± 2.56 | 21.93 ± 2.88 | 22.15 ± 2.96 | 0.363 |
| AFC (NO.) | 23.03 ± 9.40 | 22.61 ± 8.85 | 20.70 ± 9.44 | 0.017 |
| Basel FSH level (mIU/ml) | 7.32 ± 2.00 | 7.34 ± 1.70 | 7.76 ± 1.83 | 0.008 |
| Basel E₂ level (pg/ml) | 46.07 ± 26.82 | 44.57 ± 27.86 | 42.86 ± 26.31 | 0.487 |
| Basel LH level (mIU/ml) | 5.40 ± 2.11 | 5.50 ± 2.39 | 5.51 ± 2.29 | 0.897 |
| AMH (ng/ml) | 3.06 ± 1.66 | 2.67 ± 1.40 | 2.53 ± 1.60 | 0.286 |
| Embryo transfer conditions | ||||
| Fresh | 116 (43.94%) | 598 (39.60%) | 192 (37.35%) | |
| Frozen | 148 (56.06%) | 912 (60.40%) | 322 (62.65%) | 0.206 |
| Total number of embryos transferred | 2.42 ± 1.34 | 2.69 ± 1.38 | 2.21 ± 1.17 | < 0.001 |
| Infertility type | ||||
| Primary infertility (%) | 31.18(58/186) | 33.63(189/562) | 28.07(64/228) | |
| Secondary infertility (%) | 68.82 (128/186) | 66.37 (373/562) | 71.93 (164/228) | 0.308 |
| Initial dose of Gn (IU) | 202.55 ± 54.88 | 198.13 ± 56.53 | 216.78 ± 62.83 | < 0.001 |
| Total Gn (IU) | 2434.95 ± 946.42 | 2345.73 ± 975.14 | 2610.96 ± 1062.81 | 0.003 |
| Gn days | 10.61 ± 1.64 | 10.74 ± 1.73 | 10.65 ± 1.74 | 0.593 |
| endometrial thickness at the time of hCG injection (mm) | 10.57 ± 1.92 | 10.71 ± 1.94 | 10.86 ± 1.83 | 0.311 |
| E2 level at the time of hCG injection (pg/ml) | 2377.47 ± 987.59 | 2366.93 ± 1142.75 | 2246.53 ± 1201.18 | 0.353 |
| LH level at the time of hCG injection (mIU/ml) | 1.89 ± 1.33 | 2.41 ± 10.92 | 1.92 ± 1.04 | 0.649 |
| Prog level at the time of hCG injection (ng/ml) | 0.67 ± 0.31 | 0.70 ± 0.36 | 0.68 ± 0.34 | 0.494 |
| Number of eggs harvested | 13.94 ± 6.91 | 13.68 ± 6.01 | 12.49 ± 6.90 | 0.048 |
| MII rate (%) | 81.48 (2112/2592) | 81.92 (6298/7688) | 80.76 (2300/2848) | 0.389 |
| 2PN rate (%) | 73.20 (1546/2112) | 71.39 (4496/6298) | 69.91 (1608/2300) | 0.054 |
| Number of available D3 embryos | 10.34 ± 5.40 | 10.24 ± 4.78 | 8.86 ± 5.69 | 0.001 |
| High quality D3 embryo formation rate (%) | 47.30 (910/1924) | 47.01 (2706/5756) | 46.04 (930/2020) | 0.688 |
| Available blastocyst formation rate (%) | 62.80 (888/1414) | 63.57 (2610/4106) | 58.75 (846/1440) | 0.005 |
| High quality blastocyst formation rate (%) | 39.75 (562/1414) | 39.16 (1608/4106) | 35.14 (506/1440) | 0.014 |
| Average frequency of embryos transferred | 1.70 ± 0.35 | 1.72 ± 0.31 | 1.72 ± 0.36 | 0.212 |
| Frequency of transplantation (time/period) | 1.42 ± 0.66 | 1.56 ± 0.73 | 1.31 ± 0.64 | < 0.001 |
| Unusable embryo rate (%) | 3.76(7/186) | 1.07(6/562) | 3.95(9/228) | 0.014 |
| Blood HCG positive rate (%) | 62.88(166/264) | 60.59 (532/878) | 59.73 (178/298) | 0.663 |
| Biochemical abortion (%) | 1.52(4/264) | 5.24 (46/878) | 6.71 (20/298) | 0.011 |
| Clinical pregnancy rate (%) | 61.36(162/264) | 54.90 (482/878) | 53.02 (158/298) | 0.104 |
| Early abortion rate (%) | 6.17(10/162) | 11.20 (54/482) | 12.66 (20/158) | 0.118 |
| Late abortion rate (%) | 6.17(10/162) | 5.39 (26/482) | 2.53 (4/158) | 0.265 |
| Ectopic pregnancy rate (%) | 0(0/264) | 0.46 (4/878) | 0 (0/298) | 0.277 |
| Multiple pregnancy rate (%) | 25.93(42/162) | 22.82 (110/482) | 36.71 (58/158) | 0.003 |
| Live birth rate (%) | 53.79(142/264) | 45.79 (402/878) | 44.97 (134/298) | 0.053 |
| Cumulative live birth rate (%) | 76.34 (142/186) | 71.53 (402/562) | 58.77 (134/228) | < 0.001 |
* Data given as mean ± SD, AFC: Antral follicle count, AMH: anti-mullerian hormone, BMI: Body mass index, FSH: Follicle-stimulating hormone, SD: Standard deviation, LH: Luteinizing hormone.
Further logistic regression confirmed that oocyte retrieval > 7 months after occlusion significantly reduced both clinical pregnancy (unadjusted: OR = 0.447, 95% CI = 0.290–0.690, P < 0.001; age-adjusted: OR = 0.478, 95% CI = 0.308–0.742, P = 0.001; multivariate-adjusted: OR = 0.557, 95% CI = 0.344–0.901, P = 0.017) and live birth rates (unadjusted: OR = 0.436, 95% CI = 0.285–0.666, P < 0.001; age-adjusted: OR = 0.469, 95% CI = 0.305–0.724, P = 0.001; multivariate-adjusted: OR = 0.524, 95% CI = 0.327–0.839, P = 0.007), even after full adjustment (Table 3).
Table 3.
Logistic regression analysis of pregnancy outcomes for three groups
| Unadjusted | Age-adjusted | Multi-adjusted | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | ||||
| Clinical pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.879 | 0.593–1.303 | 0.521 | 0.843 | 0.566–1.255 | 0.400 | 0.885 | 0.575–1.364 | 0.580 | |||
| > 7 months | 0.447 | 0.290–0.690 | < 0.001 | 0.478 | 0.308–0.742 | 0.001 | 0.557 | 0.344–0.901 | 0.017 | |||
| Early abortion | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.795 | 0.893–3.607 | 0.101 | 0.785 | 0.523–1.179 | 0.244 | 1.696 | 0.793–3.623 | 0.173 | |||
| > 7 months | 1.328 | 0.588–3.001 | 0.495 | 1.281 | 0.807–2.033 | 0.638 | 1.928 | 0.795–4.674 | 0.146 | |||
| Multiple pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.816 | 0.545–1.220 | 0.321 | 0.785 | 0.523–1.179 | 0.244 | 0.813 | 0.531–1.243 | 0.338 | |||
| > 7 months | 1.170 | 0.742–1.844 | 0.499 | 1.281 | 0.807–2.033 | 0.293 | 1.483 | 0.913–2.409 | 0.111 | |||
| Live birth | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.721 | 0.494–1.051 | 0.089 | 0.676 | 0.459–0.995 | 0.047 | 0.714 | 0.470–1.084 | 0.114 | |||
| > 7 months | 0.436 | 0.285–0.666 | < 0.001 | 0.469 | 0.305–0.724 | 0.001 | 0.524 | 0.327–0.839 | 0.007 |
* OR: odds ratios, 95% CI: confidence interval, Multiple logistic regression models adjusting for age, BMI, AFC, infertility years, basal FSH level, basal LH level, basal E2 level, total number of embryos transferred, Initial dose of Gn, total Gn, number of eggs harvested and number of high quality blastocyst
Influence of age on pregnancy outcomes
We further analyzed the influence of age on IVF-ET outcomes after occlusion. Consistent with the overall results, in younger women (< 35 years), delaying oocyte retrieval significantly reduced both clinical pregnancy (unadjusted: OR = 0.405, 95% CI = 0.212–0.775, P = 0.006; age-adjusted: OR = 0.404, 95% CI = 0.211–0.772, P = 0.006; multivariate-adjusted: OR = 0.387, 95% CI = 0.194–0.775, P = 0.007) and live birth rates (unadjusted: OR = 0.389, 95% CI = 0.212–0.714, P = 0.002; age-adjusted: OR = 0.387, 95% CI = 0.211–0.711, P = 0.002; multivariate-adjusted: OR = 0.376, 95% CI = 0.195–0.724, P = 0.003) at 3–7 months, with an even stronger effect observed beyond 7 months (Clinical pregnancy: unadjusted: OR = 0.246, 95% CI = 0.121–0.503, P < 0.001; age-adjusted: OR = 0.250, 95% CI = 0.122–0.513, P < 0.001; multivariate-adjusted: OR = 0.265, 95% CI = 0.123–0.572, P = 0.001. Live birth: unadjusted: OR = 0.293, 95% CI = 0.148–0.579, P < 0.001; age-adjusted: OR = 0.300, 95% CI = 0.151–0.595, P = 0.001; multivariate-adjusted: OR = 0.317, 95% CI = 0.152–0.664, P = 0.002). In addition, oocyte retrieval beyond 7 months increased the risk of multiple pregnancy in younger women (unadjusted: OR = 1.786, 95% CI = 0.985–3.239, P = 0.056; age-adjusted: OR = 1.939, 95% CI = 1.062–3.541, P = 0.031; multivariate-adjusted: OR = 2.476, 95% CI = 1.289–4.754, P = 0.006). By contrast, in older women (≥ 35 years), retrieval at 3–7 months was associated with an increased chance of clinical pregnancy (multivariate-adjusted: OR = 2.178, 95% CI = 1.099–4.318, P = 0.026) and a reduced risk of multiple pregnancy (multivariate-adjusted: OR = 0.411, 95% CI = 0.195–0.864, P = 0.019) (Table 4).
Table 4.
Logistic regression analysis of pregnancy outcomes for young or elderly patients
| Unadjusted | Age-adjusted | Multi-adjusted | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | ||||
| < 35 years | ||||||||||||
| Clinical pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.405 | 0.212–0.775 | 0.006 | 0.404 | 0.211–0.772 | 0.006 | 0.387 | 0.194–0.775 | 0.007 | |||
| > 7 months | 0.246 | 0.121–0.503 | < 0.001 | 0.250 | 0.122–0.513 | < 0.001 | 0.265 | 0.123–0.572 | 0.001 | |||
| Early abortion | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.811 | 0.611–5.372 | 0.284 | 1.836 | 0.618–5.454 | 0.274 | 1.693 | 0.510–5.624 | 0.390 | |||
| > 7 months | 2.000 | 0.585–6.839 | 0.269 | 1.803 | 0.522–6.224 | 0.351 | 2.190 | 0.576–8.328 | 0.250 | |||
| Multiple pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.051 | 0.627–1.763 | 0.849 | 1.032 | 0.614–1.735 | 0.904 | 1.181 | 0.677–2.060 | 0.558 | |||
| > 7 months | 1.786 | 0.985–3.239 | 0.056 | 1.939 | 1.062–3.541 | 0.031 | 2.476 | 1.289–4.754 | 0.006 | |||
| Live birth | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.389 | 0.212–0.714 | 0.002 | 0.387 | 0.211–0.711 | 0.002 | 0.376 | 0.195–0.724 | 0.003 | |||
| > 7 months | 0.293 | 0.148–0.579 | < 0.001 | 0.300 | 0.151–0.595 | 0.001 | 0.317 | 0.152–0.664 | 0.002 | |||
| ≥ 35 years | ||||||||||||
| Clinical pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.727 | 0.995-3.000 | 0.052 | 1.601 | 0.911–2.814 | 0.102 | 2.178 | 1.099–4.318 | 0.026 | |||
| > 7 months | 0.787 | 0.434–1.425 | 0.429 | 0.728 | 0.396–1.340 | 0.728 | 1.290 | 0.612–2.718 | 0.503 | |||
| Early abortion | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.868 | 0.745–4.686 | 0.183 | 1.967 | 0.779–4.966 | 0.152 | 2.155 | 0.761–6.102 | 0.148 | |||
| > 7 months | 0.872 | 0.290–2.624 | 0.807 | 0.898 | 0.298–2.712 | 0.849 | 1.640 | 0.455–5.917 | 0.450 | |||
| Multiple pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.513 | 0.266–0.989 | 0.046 | 0.470 | 0.241–0.915 | 0.026 | 0.411 | 0.195–0.864 | 0.019 | |||
| > 7 months | 0.676 | 0.330–1.387 | 0.286 | 0.633 | 0.306–1.311 | 0.218 | 0.754 | 0.328–1.733 | 0.506 | |||
| Live birth | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.208 | 0.706–2.066 | 0.490 | 1.097 | 0.631–1.906 | 0.742 | 1.236 | 0.665–2.295 | 0.503 | |||
| > 7 months | 0.660 | 0.366–1.193 | 0.169 | 0.599 | 0.325–1.102 | 0.100 | 0.767 | 0.383–1.539 | 0.456 |
* OR: odds ratios, 95% CI: confidence interval, Multiple logistic regression models adjusting for age, BMI, AFC, infertility years, basal FSH level, basal LH level, basal E2 level, total number of embryos transferred, Initial dose of Gn, total Gn, number of eggs harvested and number of high quality blastocyst
Ovarian reserve and pregnancy outcomes
Patients were divided into normal and low ovarian reserve groups. In women with normal ovarian reserve, delaying oocyte retrieval after occlusion—particularly beyond 7 months—significantly reduced the likelihood of clinical pregnancy (unadjusted: OR = 0.456, 95% CI = 0.288–0.721, P = 0.001; age-adjusted: OR = 0.485, 95% CI = 0.305–0.772, P = 0.002; multivariate-adjusted: OR = 0.572, 95% CI = 0.344–0.954, P = 0.032) and live birth (unadjusted: OR = 0.439, 95% CI = 0.281–0.686, P < 0.001; age-adjusted: OR = 0.473, 95% CI = 0.300–0.746, P = 0.001; multivariate-adjusted: OR = 0.546, 95% CI = 0.322–0.899, P = 0.017). By contrast, in women with low ovarian reserve, the timing of oocyte retrieval showed no significant influence on pregnancy outcomes (Table 5).
Table 5.
Logistic regression analysis of pregnancy outcomes for patients with different ovarian reserve
| Unadjusted | Age-adjusted | Multi-adjusted | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | ||||
| Normal ovarian reserve | ||||||||||||
| Clinical pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.931 | 0.616–1.409 | 0.736 | 0.888 | 0.584–1.349 | 0.557 | 0.966 | 0.614–1.518 | 0.879 | |||
| > 7 months | 0.456 | 0.288–0.721 | 0.001 | 0.485 | 0.305–0.772 | 0.002 | 0.572 | 0.344–0.954 | 0.032 | |||
| Early abortion | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 1.630 | 0.805-3.300 | 0.175 | 1.771 | 0.868–3.610 | 0.116 | 1.652 | 0.758–3.599 | 0.207 | |||
| > 7 months | 1.425 | 0.629–3.229 | 0.396 | 1.295 | 0.568–2.953 | 0.538 | 2.035 | 0.827–5.003 | 0.122 | |||
| Multiple pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.837 | 0.554–1.266 | 0.400 | 0.799 | 0.526–1.214 | 0.293 | 0.817 | 0.528–1.265 | 0.365 | |||
| > 7 months | 1.176 | 0.732–1.890 | 0.502 | 1.286 | 0.795–2.080 | 0.306 | 1.414 | 0.853–2.341 | 0.179 | |||
| Live birth | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.763 | 0.514–1.133 | 0.180 | 0.709 | 0.474–1.061 | 0.095 | 0.772 | 0.499–1.194 | 0.244 | |||
| > 7 months | 0.439 | 0.281–0.686 | < 0.001 | 0.473 | 0.300-0.746 | 0.001 | 0.546 | 0.332–0.899 | 0.017 | |||
| Low ovarian reserve | ||||||||||||
| Clinical pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.556 | 0.149–2.072 | 0.381 | 0.683 | 0.174–2.674 | 0.584 | 1.186 | 0.169–8.317 | 0.864 | |||
| > 7 months | 0.476 | 0.119–1.902 | 0.294 | 0.444 | 0.107–1.850 | 0.265 | 0.393 | 0.052–2.994 | 0.367 | |||
| Early abortion | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | NA | NA | NA | NA | NA | NA | NA | NA | NA | |||
| > 7 months | NA | NA | NA | NA | NA | NA | NA | NA | NA | |||
| Multiple pregnancy | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.537 | 0.087–3.322 | 0.503 | 0.565 | 0.090–3.560 | 0.543 | 0.905 | 0.050-16.449 | 0.946 | |||
| > 7 months | 1.435 | 0.248–8.291 | 0.687 | 1.430 | 0.247–8.277 | 0.690 | 5.136 | 0.375–70.448 | 0.221 | |||
| Live birth | ≤ 3 months | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. | ||
| 3–7 months | 0.425 | 0.114–1.583 | 0.202 | 0.514 | 0.131–2.010 | 0.339 | 0.794 | 0.113–5.580 | 0.816 | |||
| > 7 months | 0.476 | 0.119–1.902 | 0.294 | 0.441 | 0.106–1.846 | 0.263 | 0.391 | 0.054–2.857 | 0.355 |
* OR: odds ratios, 95% CI: confidence interval, Multiple logistic regression models adjusting for age, BMI, AFC, infertility years, basal FSH level, basal LH level, basal E2 level, total number of embryos transferred, Initial dose of Gn, total Gn, number of eggs harvested and number of high quality blastocyst
Mediation effects of hormones and ovarian reserve
Overall, the results identified the negative impact of delayed oocyte retrieval on pregnancy outcomes after occlusion. Both FSH levels and AFC varied significantly across different retrieval intervals, suggesting a potential role of hormones and ovarian reserve in influencing outcomes. Mediation effect analysis was then performed comparing retrieval times (≤ 3 months vs. 3–7 months; ≤3 months vs. >7 months). In this analysis, the ADE confirmed an association between retrieval time and pregnancy outcomes. However, hormones and ovarian reserve showed no significant ACME, indicating no mediating effect of these factors on the relationship between retrieval time and outcomes (Fig. 3).
Fig. 3.
Mediation effect analysis of ovarian reserve and hormone levels for the pregnancy outcomes in different time after tubal occlusion. (a) AFC; (b) E2; (c) FSH; (d) LH
To further validate this, 117 new patients were prospectively recruited (Table S1). In this cohort, an increase in FSH levels and a decrease in AFC were observed within 3 months, but both indices gradually returned to preoperative levels after 3 months. Similar patterns were noted across different ages and ovarian reserve groups (Table S2). These temporary changes further explain the absence of mediation effects of hormones and ovarian reserve on pregnancy outcomes.
Active immune microenvironment in endometrium with hydrosalpinx
To investigate potential changes in the endometrial microenvironment before and after occlusion in patients with hydrosalpinx, bulk RNAseq was performed on endometrial samples. Differential expression analysis identified 203 genes (e.g., LINC01391, GNLY, ITGAD, IL2RB) and 407 genes (e.g., CXCL14, CLIC6, FCGBP, PLAG1, ITM2A) significantly expressed before and after occlusion, respectively (Fig. 4b). GO annotation revealed prominent immune-related molecular functions before occlusion, such as MHC class Ib receptor activity (Fig. 4c). KEGG pathway analysis also suggested activation of immune signaling prior to occlusion, including human T-cell leukemia virus 1 infection, natural killer cell–mediated cytotoxicity, cellular senescence, antigen processing and presentation, and complement and coagulation cascades. By contrast, after occlusion, cAMP signaling and cGMP-PKG signaling pathways were activated (Fig. 4d). GSEA further confirmed enrichment of immune-related pathways in the endometrium before occlusion, compared with the postoperative state (Fig. 4e).
Fig. 4.
RNAseq of the endometrium before and after tubal occlusion of hydrosalpinx. (a) differentially expressed genes (DEGs) of RNAseq; (b) the GO annotations of DEGs before and after occlusion; (c) KEGG enrichment; (d) GSEA analysis of pathways
M2 macrophage recruitment in the endometrium after occlusion
The immune cell composition of the endometrium before and after occlusion was analyzed, revealing notable changes in the immune microenvironment (Fig. 5a). T follicular helper cells were more abundant in the endometrium of patients with hydrosalpinx (P = 0.02), whereas M2 macrophage infiltration increased after occlusion (P = 0.029) (Fig. 5b). Top DEGs identified after occlusion, such as CXCL14, were also significantly correlated with M2 macrophages (Fig. 5c). Furthermore, immunohistochemistry confirmed greater infiltration of M2 macrophages in the endometrium following occlusion (Fig. 5d).
Fig. 5.
M2 macrophage polarization after tubal occlusion of hydrosalpinx in the endometrium. (a) immune cell components in the endometrium before and after tubal occlusion; (b) the proportion of T follicular helper and M2 macrophage; (c) correlations of Top DEGs before and after occlusion with T and M2 macrophage; (d) CD163 expression in the endometrium before and after tubal occlusion identified with immunohistochemistry
CXCL14 in endometrial epithelial cells might be associated with M2 macrophage polarization
After occlusion of hydrosalpinx, pregnancy outcomes improved. To explore the formation of an immunosuppressive microenvironment in the endometrium, single-cell RNAseq data from normal endometrium were analyzed, identifying nine distinct cell types (Fig. 6a). CXCL14, the most highly expressed gene after occlusion, was localized primarily to epithelial cells of the endometrium, a finding further validated by immunofluorescence (Fig. 6b). Although no specific receptor for CXCL14 has been identified, it has been reported to bind with high affinity to CXCR4. In the single-cell data, CXCR4 expression was detected in macrophages, suggesting that CXCL14 may act through CXCR4 on macrophages to promote M2 macrophage polarization (Fig. 6c).
Fig. 6.
Single-cell RNAseq of endometrium identified highly expressed CXCL14 gene in the epithelium. (a) single-cell RNAseq defined 9 cell types; (b) the CXCL14 and CXCR4 expressed in the endometrium; (c) located the CXCL14 in the endometrial epithelium cells with immunofluorescence
Discussion
Although hydrosalpinx is a common factor that can affect female infertility and significantly influence assisted reproduction technology (ART) outcomes, the mechanisms behind these effects remain unclear [2]. Hydrosalpinx is associated with a range of conditions, including pelvic inflammation, endometriosis, appendicitis, or previous pelvic and abdominal surgery [17]. Treatment of hydrosalpinx with salpingostomy, salpingectomy or tubal occlusion has been linked to beneficial improvements in pregnancy outcomes [18]. In 2017, Xu et al. [19] compared the effects of various methods of hydrosalpinx treatment before IVF, and found that laparoscopic salpingectomy and proximal tubal occlusion had similar positive effects on pregnancy outcomes in women with hydrosalpinx undergoing IVF. In support of this, Bi et al. [20] also reported that prior treatment of hydrosalpinx by salpingectomy and proximal tubal occlusion was superior to other surgical procedures with respect to IVF pregnancy outcomes. Similarly, in 2023, Yilei et al. [13] reported higher clinical pregnancy rates when hydrosalpinx was treated by salpingectomy before IVF. At the same time, oocyte retrieval conducted 4–6 and 7–12 months after salpingectomy was associated with even higher cumulative pregnancy and live birth rates. Although our previous analyses support the effectiveness of tubal occlusion before IVF for hydrosalpinx treatment, no exact oocyte retrieval time was defined [12]. Based on the present study, and analysis of logistic regression and mediation effects, we recommend that oocyte retrieval should be performed within 7 months after tubal occlusion for hydrosalpinx treatment, with an optimal oocyte retrieval time of 3 months. In addition, it should also be mentioned that younger female patients with normal ovarian reserves are significantly more likely to benefit from tubal occlusion for hydrosalpinx treatment before IVF. If other critical factors that influence fertility are suboptimal, simply treating the hydrosalpinx will not likely improve IVF pregnancy outcomes. In particular, age has been identified as a determining factor that strongly influences female reproductive function; with the prevalence of infertility directly correlating with increasing age [21]. In addition to hydrosalpinx, other complex factors such as hyperprolactinemia, pelvic adhesions, endometriosis, and lower ovulatory dysfunction, etc., can also have additive or synergistic effects on reducing female fertility [22]. Similarly, ovarian reserves are another crucial factor directly affecting pregnancy outcomes. Low ovarian reserves have been linked to reduced pregnant rates and poorer ART prognosis [23, 24]. Tubal occlusion for the treatment of hydrosalpinx cannot reverse ovarian reserve function. Even in the early stages of tubal occlusion, ovarian reserves could already be suboptimal. Thus, our results also highlight the importance of careful pre-selection of patients for tubal occlusion in the context of IVF-ET treatment programs.
Based on our results we recommend an optimal oocyte retrieval time of ≤ 7 months after tubal occlusion; and speculate that hydrosalpinx recurrence and endometrial deterioration could negatively affect IVF outcomes for oocytes retrieved later than 7 months after tubal occlusion. As shown in another study, hydrosalpinx was associated with a high recurrence rate; and it is plausible that hydrosalpinx recurrence rates may increase in correlation with post-operation time. Liu et al. reported treatment of 562 hydrosalpinx cases in patients between January 1, 2012 and December 31, 2016. Of these, hydrosalpinx was found to have reoccurred in 146 patients after treatment [25]. As shown for other surgical hydrosalpinx treatment methods, such as hydrosalpinx aspiration with or without sclerotherapy, overall hydrosalpinx recurrence rates can reach as high as 30% [1]. Hydrosalpinx recurrence can cause toxic effects that directly reduce sperm motility, embryo development and endometrial receptivity [9]. Although hormone levels and ovarian reserves fluctuated after tubal occlusion, they returned to pre-operation levels after 3 months. However, our analysis did not identify any potential mediator effects by hormone and ovarian reserve levels on the relationship between pregnancy outcomes and tubal occlusion. Thus, the common influences of hydrosalpinx, or hydrosalpinx recurrence, may be related to deterioration of the endometrial environment and detrimental molecular changes in the endometrium. Studies have shown that hydrosalpinx can interfere with embryo implantation in the endometrium [26]. In addition, changes in the expression levels of specific candidate genes and proteins in hydrosalpinx patients can also affect the embryo implantation process [27]. While hydrosalpinx is a result of pelvic inflammation, local inflammation of the uterus may also occur. In 2022, Zou et al. reported that both bilateral and simple hydrosalpinx were closely related to chronic endometritis [28]. Using electron microscopy, Ajonuma et al. [29] found serious loss of cilia and microvilli on the epithelial surface of the fallopian tube in the hydrosalpinx. Moreover, inflammatory cells could also be seen in the hydrosalpinx, in areas with flattened or no epithelial cells in the lumen, and in dilated blood vessels and/or lymphatic vessels, which also could have adverse effects on the reproductive process and IVF outcomes [29].
Our previous results show that IVF outcomes in patients with hydrosalpinx are significantly improved after tubal occlusion. Further analysis of gene expression changes in the endometrial tissues of patients before and after hydrosalpinx, indicate that immuno-inflammatory pathways in the endometrial are significantly activated before tubal occlusion: including of T cell, antigen presenting and natural killer cell signaling pathways. These results strongly suggest a relationship between activated inflammatory responses in the endometrial tissues of hydrosalpinx patients and the high failure rate of IVF-ET before occlusion treatment. In addition, we also detected activated cAMP signaling pathways in the endometrium after occlusion, which could effectively inhibit tubal fluid stimulation by isoproterenol [30]. Thus, inactivation of cAMP signaling during hydrosalpinx recurrence may also aggravate tubal fluid and the negative effects of effusion on IVF. Due to the activation of a large number of immune signaling pathways in the endometrium by hydrosalpinx, further analysis indicates that M2 macrophage levels increase after occlusion, resulting in inhibition of the local immune inflammatory response in the surrounding tissues. As an important component of the immune system, macrophages are thought to be important factors in female infertility. Studies have shown that macrophages can accumulate in large numbers in patients with ectopic tubal pregnancy, salpingitis, hydrosalpinx and endometriosis [31]. In addition, in a mycoplasma-induced hydrosalpinx model, activation of macrophages and pro-inflammatory pathways was shown to promote the persistence of hydrosalpinx [32]. Polarization of M2 macrophages in the endometrial may effectively reduce endometrial inflammation and the risk of infertility [33]. Combining single-cell and total RNA sequencing analyses, we found high expression of the CXCL14 gene in the endometrium after tubal occlusion, which is thought to be closely related to M2 macrophages polarization. In 2020, Lv et al. found in both in vivo and in vitro experiments that CXCL14 effectively inhibits polarization of M1-type macrophages, while promoting production of M2-type macrophages [34]. In 2023, He et al. also found in mouse models that CXCL14 effectively induces polarization of M2-type macrophages, thereby improving atherosclerosis [35]. At the same time, in tumors, increased CXCL14 also promoted polarization of M2-type macrophages [36, 37]. In the present study, CXCL14 was mainly found to be highly expressed in the epithelial cells of normal endometrial tissues, which was also verified by immunofluorescence. Although the receptor for CXCL14 is not known at present, some studies have reported that the epithelial chemokine CXCL14 acts synergistically with CXCL12 through allosteric regulation of CXCR4 [38]. Interestingly, our data also indicate that CXCR4 is highly expressed in the macrophages of normal endometrial, which suggest that CXCL14 in the epithelial cells of the endometrium after occlusion might interact with CXCR4 in macrophages, helping to maintain a low inflammatory microenvironment that could ensure better IVF pregnancy outcomes.
Conclusions
Hydrosalpinx is a tubal factor strongly associated with female infertility and a high inflammatory microenvironment in the endometrium. Evidence suggests that tubal occlusion before IVF can result in improved IVF pregnancy outcomes. Increased CXCL14 gene expression in epithelial cells of the endometrium after tubal occlusion may promote polarization of M2-type macrophages. After tubal occlusion, we recommended that IVF should be conducted within 7 months for optimal pregnancy outcomes.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
This section thanks those other than the authors who have made substantive contributions to the manuscript, including participants in collaborative trials and persons providing only data collection or assistance with preparing the paper for submission or publication. Name only those who have made substantive contributions to the study.
Author contributions
JL, HL, JLL and LZ analyzed the data, performed the experiments, and wrote the manuscript. ZL, QLS and YC designed and supervised the study, revised the paper and provided the funding. SEM and BZ collected the samples. All the authors approved the final paper.
Funding
This study was funded by the National Natural Science Foundation of China (grant numbers 82460596 and 82060145), self-funded research projects by Guangxi Zhuang Autonomous Region Health Commission (grant numbers Z-A20220363), Guangxi Natural Science Foundation (grant number 2022JJB140197, 2024JJA141067, 2024JJB140509), and First-class discipline innovation-driven talent program of Guangxi Medical University (DC2500001200, DC2500001220), the National innovation and entrepreneurship training program for college students (202410598052, S202510598220).
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval
The study was approved by the Ethics Committee of the Reproductive Hospital of Guangxi Zhuang Autonomous Region (Number: KY-LL-2022-13).
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Jie Li, Hui Li, Jinling Liao and Li Zhou contributed equally to this work.
Contributor Information
Zhong Lin, Email: linzhong91316@126.com.
Qiuling Shi, Email: qshi@cqmu.edu.cn.
Yang Chen, Email: chenyang91316@126.com.
References
- 1.Cohen A, Almog B, Tulandi T. Hydrosalpinx sclerotherapy before in vitro fertilization: systematic review and meta-analysis. J Minim Invasive Gynecol. 2018;25:600–7. [DOI] [PubMed] [Google Scholar]
- 2.Camus E, Poncelet C, Goffinet F, Wainer B, Merlet F, Nisand I, Philippe HJ. Pregnancy rates after in-vitro fertilization in cases of tubal infertility with and without hydrosalpinx: a meta-analysis of published comparative studies. Hum Reprod. 1999;14:1243–9. [DOI] [PubMed] [Google Scholar]
- 3.Holzer I, Ott J, Kurz C, Hofstetter G, Hager M, Kuessel L, Parry JP. Is chronic endometritis associated with tubal infertility?? A prospective cohort study. J Minim Invasive Gynecol. 2021;28(11):1876–81. [DOI] [PubMed] [Google Scholar]
- 4.Bao H, Qu Q, Huang X, Wang M, Wang X, Hao C. Impact of hydrosalpinx fluid on early human embryos. Syst Biol Reprod Med. 2017;63:279–84. [DOI] [PubMed] [Google Scholar]
- 5.Harb H, Al-Rshoud F, Karunakaran B, Gallos ID, Coomarasamy A. Hydrosalpinx and pregnancy loss: a systematic review and meta-analysis. Reprod Biomed Online. 2019;38:427–41. [DOI] [PubMed] [Google Scholar]
- 6.Peng J, Guo F, Liu H, Gu Z, Zeng Z, Li M, Zeng H, Liang X. Correlation between hysteroscopy findings in patients with hydrosalpinx and chronic endometritis. Int J Gynaecol Obstet. 2022;157(2):471–5. [DOI] [PubMed] [Google Scholar]
- 7.Copperman AB, Wells V, Luna M, Kalir T, Sandler B, Mukherjee T. Presence of hydrosalpinx correlated to endometrial inflammatory response in vivo. Fertil Steril. 2006;86(4):972–6. [DOI] [PubMed] [Google Scholar]
- 8.Tjahyadi D, Udayana IS, Nisa AS, Rachmawati A, Djuwantono T. Comparison of salpingectomy and tubal occlusion for hydrosalpinx in in-vitro fertilization outcome. Ann Med Surg (Lond). 2024;86:886–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Palagiano A, Cozzolino M, Ubaldi FM, Palagiano C, Coccia ME. Effects of hydrosalpinx on endometrial implantation failures: evaluating salpingectomy in women undergoing in vitro fertilization. Rev Bras Ginecol Obstet. 2021;43:304–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chen T, Zhao F, Wang Q, Liu C, Lan Y, Wang S, Xin Z, Yang X. Salpingectomy may decrease antral follicle count but not live birth rate for IVF-ET patients aged 35–39 years: a retrospective study. J Ovarian Res. 2020;13:80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Melo P, Georgiou EX, Johnson N, van Voorst SF, Strandell A, Mol BWJ, Becker C, Granne IE. Surgical treatment for tubal disease in women due to undergo in vitro fertilisation. Cochrane Database Syst Rev. 2020;10:CD002125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li J, Mo S, Lin Z, Mo F, Shi Q. Proximal tubal occlusion first or oocyte retrieval first for patients with hydrosalpinx? Arch Gynecol Obstet. 2024;309:1597–608. [DOI] [PubMed] [Google Scholar]
- 13.Yilei H, Shuo Y, Caihong M, Yan Y, Xueling S, Jiajia Z, Ping L, Rong L, Jie Q. The influence of timing of oocytes retrieval and embryo transfer on the IVF-ET outcomes in patients having bilateral salpingectomy due to bilateral hydrosalpinx. Front Surg. 2023;9:1076889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Strandell A. Treatment of hydrosalpinx in the patient undergoing assisted reproduction. Curr Opin Obstet Gynecol. 2007;19:360–5. [DOI] [PubMed] [Google Scholar]
- 15.Wu S, Zhang Q, Li Y. Effect comparison of salpingectomy versus proximal tubal occlusion on ovarian reserve: a meta-analysis. Med (Baltim). 2020;99:e20601. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Gardner DK, Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: towards a single blastocyst transfer. Fertil Steril. 2000;73:1155–8. [DOI] [PubMed] [Google Scholar]
- 17.Puttemans PJ, Brosens IA. Salpingectomy improves in-vitro fertilization outcome in patients with a hydrosalpinx: blind victimization of the fallopian tube? Hum Reprod. 1996;11:2079–81. [DOI] [PubMed] [Google Scholar]
- 18.Ng KYB, Cheong Y. Hydrosalpinx - salpingostomy, salpingectomy or tubal occlusion. Best Pract Res Clin Obstet Gynaecol. 2019;59:41–7. [DOI] [PubMed] [Google Scholar]
- 19.Xu B, Zhang Q, Zhao J, Wang Y, Xu D, Li Y. Pregnancy outcome of in vitro fertilization after essure and laparoscopic management of hydrosalpinx: a systematic review and meta-analysis. Fertil Steril. 2017;108:84–95. [DOI] [PubMed] [Google Scholar]
- 20.Bi B, Han X, Dai W, Fang L, Shi H, Hu L. Comparisons of different treatment outcomes in IVF/ET patients with hydrosalpinx: a retrospective study. Gynecol Endocrinol. 2023;39:2249999. [DOI] [PubMed] [Google Scholar]
- 21.Dunson DB, Baird DD, Colombo B. Increased infertility with age in men and women. Obstet Gynecol. 2004;103:51–6. [DOI] [PubMed] [Google Scholar]
- 22.Awonuga AO, Camp OG, Biernat MM, Abu-Soud HM. Overview of infertility. Syst Biol Reprod Med. 2025;71:116–42. [DOI] [PubMed] [Google Scholar]
- 23.Domingues TS, Rocha AM, Serafini PC. Tests for ovarian reserve: reliability and utility. Curr Opin Obstet Gynecol. 2010;22:271–6. [DOI] [PubMed] [Google Scholar]
- 24.La Marca A, Sunkara SK. Individualization of controlled ovarian stimulation in IVF using ovarian reserve markers: from theory to practice. Hum Reprod Update. 2014;20:124–40. [DOI] [PubMed] [Google Scholar]
- 25.Liu H, Yao Z, Zhang R, Wu S, Lv S, Yan L. Effect of recurrence of hydrosalpinx after tubal ligation on the outcome of in vitro fertilization treatment: a retrospective cohort study. Gynecol Minim Invasive Ther. 2020;9(3):118–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Aljassim F, Georgopoulou N, Rigby CH, Powell SG, Wyatt JNR, Hapangama DK, Hill CJ. Exploring the presence of markers of decidualization in the fallopian tubes: a systematic review. Biol Reprod. 2023;109:125–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Changaei M, Javidan M, Ramezani Tehrani F, Mosaffa N, Noroozzadeh M, Hosseinzadeh R, Rajaei S. Reduced expression of Il10, Stat3, Hoxa10, and Itgb3 in the embryo implantation site of rat model with prenatal androgen-induced polycystic ovary syndrome. Am J Reprod Immunol. 2023;90:e13702. [DOI] [PubMed] [Google Scholar]
- 28.Zou Y, Li S, Ming L, Yang Y, Ye P, Zou J. The correlation between chronic endometritis and tubal-factor infertility. J Clin Med. 2022;12:285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ajonuma LC, Ng EH, Chan LN, Chow PH, Kung LS, Cheung AN, Ho LS, Briton-Jones C, Lok IH, Haines CJ, Chan HC. Ultrastructural characterization of whole hydrosalpinx from infertile Chinese women. Cell Biol Int. 2005;29:849–56. [DOI] [PubMed] [Google Scholar]
- 30.Li Y, Liu C, Guo N, Cai L, Wang M, Zhu L, Li F, Jin L, Sui C. Extracellular vesicles from human fallopian tubal fluid benefit embryo development in vitro. Hum Reprod Open. 2023;2023:hoad006. [DOI] [PMC free article] [PubMed]
- 31.Li P, Zhu WJ, Ma ZL, Wang G, Peng H, Chen Y, Lee KK, Yang X. Enhanced beta-catenin expression and inflammation are associated with human ectopic tubal pregnancy. Hum Reprod. 2013;28:2363–71. [DOI] [PubMed] [Google Scholar]
- 32.Harvie MC, Carey AJ, Armitage CW, O’Meara CP, Peet J, Phillips ZN, Timms P, Beagley KW. Chlamydia-infected macrophages are resistant to azithromycin treatment and are associated with chronic oviduct inflammation and hydrosalpinx development. Immunol Cell Biol. 2019;97:865–76. [DOI] [PubMed] [Google Scholar]
- 33.Xin L, Lin X, Zhou F, Li C, Wang X, Yu H, Pan Y, Fei H, Ma L, Zhang S. A scaffold laden with mesenchymal stem cell-derived exosomes for promoting endometrium regeneration and fertility restoration through macrophage immunomodulation. Acta Biomater. 2020;113:252–66. [DOI] [PubMed] [Google Scholar]
- 34.Lv J, Wu ZL, Gan Z, Gui P, Yao SL. CXCL14 overexpression attenuates sepsis-associated acute kidney injury by inhibiting proinflammatory cytokine production. Mediators Inflamm. 2020;2020:2431705. [DOI] [PMC free article] [PubMed]
- 35.He P, Wang H, Cheng S, Hu F, Zhang L, Chen W, Xu Y, Zhang Y, Gu Y, Li Z, Jin Y, Liu X, Jia Y. Geniposide ameliorates atherosclerosis by regulating macrophage polarization via perivascular adipocyte-derived CXCL14. J Ethnopharmacol. 2023;314:116532. [DOI] [PubMed] [Google Scholar]
- 36.Tian HY, Liang Q, Shi Z, Zhao H. Exosomal CXCL14 contributes to M2 macrophage polarization through NF-κB signaling in prostate cancer. Oxid Med Cell Longev. 2022;2022:7616696. [DOI] [PMC free article] [PubMed]
- 37.Gowhari Shabgah A, Haleem Al-Qaim Z, Markov A, Valerievich Yumashev A, Ezzatifar F, Ahmadi M, Mohammad Gheibihayat S, Gholizadeh Navashenaq J. Chemokine CXCL14; a double-edged sword in cancer development. Int Immunopharmacol. 2021;97:107681. [DOI] [PubMed] [Google Scholar]
- 38.Collins PJ, McCully ML, Martínez-Muñoz L, Santiago C, Wheeldon J, Caucheteux S, Thelen S, Cecchinato V, Laufer JM, Purvanov V, Monneau YR, Lortat-Jacob H, Legler DF, Uguccioni M, Thelen M, Piguet V, Mellado M, Moser B. Epithelial chemokine CXCL14 synergizes with CXCL12 via allosteric modulation of CXCR4. FASEB J. 2017;31:3084–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
No datasets were generated or analysed during the current study.