We retrospectively reviewed the records of all women treated at the Reproductive Medicine Centre, West China Second University Hospital, Sichuan University for ART treatment between January 2018 and December 2022. Inclusion criteria for participation were patients with a history of tubal ectopic pregnancy, whether treated surgically or conservatively, who underwent their first cycle of fresh or frozen embryo transfer. The exclusion criteria included: (1) a previous ectopic pregnancy following ART, (2) the cycles involving preimplantation genetic diagnosis and screening; and (3) any abnormal ultrasound findings before embryo transfer that could impact the assessment of the uterine cavity. The rationale for these exclusions was to concentrate the study on the effects of hysteroscopy within a more homogeneous group, thereby reducing variability from factors unrelated to the condition of the uterine cavity. The study flow is presented in Fig. 1.

Fig. 1.

Flowchart of the included population. ET, embryo transfer; EP, endometrial polyps; CE, chronic endometritis.

To account for potential confounding factors, we applied PSM using a multivariable logistic regression model based on age, body mass index (BMI), anti-Müllerian hormone (AMH) level, number of embryos transferred, and stage of embryos (whether they were cleavage-stage embryos or blastocysts). We also considered the type of treatment for ectopic pregnancy, which included salpingectomy, salpingostomy, or conservative treatment. This method enabled us to create pairs of patients with previous ectopic pregnancy, ensuring that those who underwent hysteroscopy were matched to those who did not, based on similar characteristics. The PSM was performed using 1:1 greedy nearest neighbor matching within a propensity score (PS) score range of 0.008 to balance the baseline characteristics and improve the comparability between groups.

In our study design, the patients in the study group were assigned to four subgroups based on hysteroscopic findings and pathological diagnosis: disease-free group, EP group, CE group, and CE plus EP (CE + EP) group.

Controlled ovarian hyperstimulation (COH) was performed using various protocols, including the gonadotropin-releasing hormone (GnRH) agonist long protocol, GnRH agonist short protocol, and GnRH antagonist protocol. These protocols were tailored to the patient’s age, ovarian reserve, and previous responses to ovarian stimulation. Human chorionic gonadotropin (hCG) (Lizhu Pharmaceutical Trading, Zhuhai, Guangdong, China) was administered when three or more follicles reached a diameter of 16–18 mm or greater. For patients at high risk for ovarian hyperstimulation syndrome (OHSS), hCG was used in combination with leuprolide acetate to trigger ovulation. The oocyte retrieval procedure was conducted 36 to 38 hours post-hCG administration, using transvaginal ultrasonography. Fertilization was achieved through conventional IVF or ICSI, and embryos were cultured until transfer on day 3, 5, or 6 following oocyte retrieval. The number of embryos transferred was determined according to the guidelines of the Chinese Society of Reproductive Medicine, with a maximum of two embryos per transfer [19]. All patients were evaluated postoperatively through a comprehensive follow-up program. A quantitative serum hCG level was determined two weeks post-embryo transfer. A subsequent transvaginal ultrasound examination was scheduled for five weeks post-transfer.

Office hysteroscopy was performed during the proliferative phase to diagnose and localize intracavitary lesions. All hysteroscopy treatments were conducted in the operating room by the same two physicians (Tianji Liao and Lijun Lin). Hysteroscopies employed a vaginoscopy approach under sedation, utilizing a 2.9-mm, 30-degree-angle hysteroscope with an external sheath of 4.4 mm diameter and offering inflow, outflow, and 5F working channels (Karl Storz, Tuttlingen, Germany). Saline solution (9% concentration) was employed to inflate the uterine cavity, with an expansion pressure approximating 100~120 mmHg. The procedure was performed with a 300-w light source with a high-definition digital camera/xenon bulb (Karl Storz™, Tuttlingen, Germany). All EP and CE were confirmed histologically. Syndecan-1 (CD138) immunohistochemistry (IHC) was conducted. CE diagnostic criteria consisted of $\ge$5 CD138+ cells identified within each high-magnification field (CD138+/high-power field (HPF) at $\times$400 magnification) within the endometrial stroma. As antibiotic therapy was recommended for CE patients [20], doxycycline (100 mg orally twice a day for 14 days) was administered. A diagnosis of normal endometrium was made when 5 CD138+/HPF or no plasma cell morphology was observed within the endometrium [21, 22].

The clinical pregnancy rate was defined as the number of intrauterine gestations with fetal cardiac activity per IVF-ET cycle. A biochemical pregnancy was defined as a positive hCG level without a gestational sac. Any pregnancy loss occurring after the visualization of intrauterine gestation was considered a spontaneous miscarriage, and any birth after 28 weeks of gestation was considered a live birth. An ectopic pregnancy was defined as a pregnancy in which the fertilized ovum implants outside the uterine cavity.

Statistical analysis was performed using SPSS software (version 25.0 for Windows; SPSS Inc., Chicago, IL, USA). Non-normally distributed data was expressed as the median with interquartile range (Q1, Q3). Qualitative variables were presented as a case quantity (n) or percentage (%). Categorical data are described by the number of cases, including the numerator/denominator and percentages. A value of p 0.05 was considered significant. Continuous variables were calculated via dependent-sample t tests or Mann-Whitney U tests as appropriate. Inter-cohort comparisons for nonparametric conditions were used by the Kruskal-Walli’s test, and the comparisons of distributed categorical variables were made using the Chi-square test when all expected cell frequencies were $\ge$5, or Fisher’s exact test when any expected cell frequency was 5. PSM was utilized for sampling at up to 1:1 nearest-neighbor matching with a caliper (0.008) to balance the baseline and improve the comparability between groups. The PSM allowed for the pairing of those who underwent hysteroscopy to those who did not, after hysteroscopy, with similar characteristics, which included age, BMI, number of transferred embryos, and stage of transferred embryos (cleavage embryos or blastocysts). Based on a previous study that reported clinical pregnancy rates of 63% versus 41% for women undergoing hysteroscopy in IVF/ICSI-ET without polyps [23], a sample size of 714 participants (post-PSM) was determined to provide 95% power, assuming a standard deviation of 2 and an alpha of 0.0008. A binary logistic regression was employed for multivariate analysis.

The clinical characteristics of the study and control groups before and following PSM are shown in Table 1. Both before and after PSM, baseline characteristics including age, BMI, gravidity, and treatment parameters were well-balanced between the two groups, confirming the effectiveness of the matching process.

As detailed in Table 2, pregnancy outcomes showed no statistically significant differences between the hysteroscopy and control groups after PSM. Clinical pregnancy rates were comparable between groups (58.26% vs. 53.22%, p = 0.397), as were biochemical pregnancy rates (p = 0.120) and live birth rates (p = 0.880). Notably, the spontaneous miscarriage rate remained significantly higher in the hysteroscopy group both before and after PSM (p = 0.031 and p = 0.042, respectively). Ectopic pregnancy rates were similar between groups.

Subgroup analysis revealed important differences in reproductive outcomes based on hysteroscopic findings (Tables 3,4,5). While baseline parameters were comparable across subgroups, the cured CE group demonstrated significantly worse outcomes. The spontaneous miscarriage rate was markedly elevated in the cured CE group (46.90%) compared to all other subgroups (disease-free: 18.00%, EP: 10.00%, CE + EP: 25.00%). Consequently, live birth rates were lowest in the cured CE group (25.00%). This association remained statistically significant after adjusting for confounding variables (adjusted odds ratio (OR) 0.348, 95% confidence interval (CI) 0.171–0.708).

Our study initially demonstrated that the utilization of hysteroscopy had an impact on pregnancy outcomes among patients with a history of ectopic pregnancies. The administration of hysteroscopy benefits patients without CE. The overall ectopic pregnancy rate in our cohort closely aligns with previously published data on similar populations, further supporting the generalizability of our findings [24]. Interestingly, although the hysteroscopy group showed a non-significant increase in biochemical and clinical pregnancy rates relative to the control group, there was a significantly higher rate of spontaneous miscarriage in the hysteroscopy group, ultimately resulting in comparable live birth rates between groups. A likely explanation for this observation is that CE serves as an important confounding factor. Our subgroup analysis demonstrated that CE was associated with the high rate of pregnancy loss in the hysteroscopy group, a finding that is consistent with a previous study [25]. Overall, our results indicate that routinely performing hysteroscopy in all patients with a history of ectopic pregnancy offers limited clinical benefit. An individualized strategy, emphasizing the detection and treatment of CE, may be more effective in enhancing reproductive outcomes for these patients.

Our data shows that patients with confirmed endometritis, diagnosed through hysteroscopy, experience the highest rate of pregnancy loss. Despite empirical antibiotic treatment aimed at reducing or eliminating plasmacyte infiltration within the endometrial stroma, the pregnancy loss rate in this subgroup remains the highest. This indicated that the current antibiotic approach might be insufficient for addressing endometritis. This finding is consistent with previous studies that have established CE as a significant risk factor for adverse reproductive outcomes [25, 26].

Several factors may explain why CE patients experienced persistently high pregnancy loss rates despite treatment. First, antibiotic resistance is a serious global medical problem in treating infectious diseases. A comprehensive survey of 3449 infertile women with RIF and three or more failed IVF-ET cycles found resistance to first-line 14-day oral doxycycline treatment in 21.2% of CE cases [27]. Our findings suggest that this resistance problem may be even more pronounced in patients with prior ectopic pregnancy, potentially due to more complex underlying pathophysiology or previous antibiotic exposures. As a result, it has been suggested that, alongside doxycycline, a 14-day course of a combination of levofloxacin lactate and metronidazole may be beneficial [20]. However, multidrug-resistant chronic endometritis CE (MDR-CE) is becoming an emerging concern in clinical management [12]. CE was resistant to two courses of combined oral antibiotic treatments (levofloxacin lactate and metronidazole) in 11.0% of cases [20]. This resistance pattern may partially explain why our hysteroscopy group, despite CE detection and antibiotic treatment, continued to demonstrate elevated pregnancy loss rates. On the other hand, a growing number of studies demonstrated that antibiotic treatments could enhance the clinical outcomes; this is only the case if follow-up biopsy confirms the successful eradication of CE [28]. Second-look hysteroscopy may be considered a potential option, despite the lack of conclusive evidence supporting its effectiveness in improving reproductive outcomes. An urgent need exists to establish universal diagnostic criteria that integrate histopathology, hysteroscopy, and microbiome analysis to address unanswered questions surrounding CE. Based on the results of our study, the pregnancy loss rate in patients with chronic endometritis is extremely high. Based on our results and the substantial gap between treatment intention and pregnancy outcomes, we strongly recommend that CE patients undergo second-look hysteroscopy to verify complete inflammation resolution before attempting conception. This approach, while requiring further validation through randomized controlled trials, represents a logical next step given the consistently poor pregnancy outcomes observed in our treated CE patients.

The natural progression of EP is still not well understood, but they are commonly found in infertile women. It has been hypothesized that EP may impair endometrial receptivity and subsequently influence ART outcomes [29]. Hysteroscopic polypectomy was associated with a higher C-reactive protein (CRP) rate and an increased total number of pregnancies [30]. In our study, EPs were identified in 18.21% of 357 patients with a previous history of ectopic pregnancy in this study. While our study found that patients with EP had higher live birth rates compared to the CE group, the clinical implications of this finding remain uncertain due to the absence of an untreated EP control group. This limitation prevents us from definitively attributing the improved outcomes to polypectomy versus the inherently better prognosis of EP compared to CE. The exact impact of EPs on fertility remains unclear, with most studies suggesting that pregnant women diagnosed with them should be managed expectantly [31]. However, clinical data on this topic are contradictory.

Regarding clinical decision-making, whether EP identified on hysteroscopy should always be surgically removed or can be managed conservatively remains unclear. In practice, polypectomy is generally recommended if EP is found before FET or ovarian stimulation for IVF, while management during ovarian stimulation is individualized based on polyp characteristics, embryo number, reproductive history, and clinic outcomes [32]. Our study found that hysteroscopy may be valuable for identifying EP in women with a history of ectopic pregnancy, but we cannot definitively conclude that polypectomy always improves outcomes without a randomized controlled trial comparing surgical removal versus conservative management. The evidence supporting the removal of EP is limited, and the best course of treatment must be determined by a carefully planned randomized controlled trial.

The primary strength of our study lies in the status as the first investigation into the administration of hysteroscopy in women with a previous history of ectopic pregnancy conducted at a single center. However, as with all clinical studies, our research has several limitations. The most notable drawback is its retrospective design, which limits the strength of our conclusions. While we matched multiple factors and found that the groups were generally similar in terms of age, BMI, number of transferred embryos, and types of embryos, unidentified confounding variables cannot be ruled out. Furthermore, the relatively small sample size may limit the generalizability of our results. Based on our findings, hysteroscopy appears most valuable for detecting specific uterine pathology (such as CE and EP) rather than universally improving reproductive outcomes in all patients with previous ectopic pregnancy. The clinical pregnancy rates did not differ significantly between groups, suggesting that routine hysteroscopy should not be universally recommended at this time. Regarding the CE subgroup, the underlying causes of the increased spontaneous miscarriage rate and decreased live birth rate remain unclear, including factors such as pre-existing microbial invasion, inadequate antibiotic administration, or potential direct effects of antibiotic treatment itself. Therefore, while our results suggest potential diagnostic value of hysteroscopy in identifying intrauterine abnormalities, we emphasize that further prospective, randomized controlled trials with larger sample sizes are essential before implementing this practice universally. Future studies should specifically compare outcomes in patients with detected pathology who receive treatment versus those managed conservatively, and establish clear criteria for when hysteroscopy would be most beneficial in this patient population.