Caesarean Scar Endometrial Defects Contribute to Post-Caesarean Abnormal Uterine Bleeding and Chronic Endometritis: A Retrospective Case-Control Study.

Y.W., J.S. and X.L. designed and coordinated the study. Y.W., J.S., H.W. and L.G. recruited patients and performed the hysteroscopy. Y.W., Y.H. and X.G. evaluated the hysteroscopic features. Y.W., Y.H., Q.W. and Y.X. performed the pathological staining and analysis. Y.H. and X.G. performed the statistical analyses. Y.H. and X.G. wrote the initial draft of the manuscript. Y.W., J.S. and X.L. revised the manuscript.

The study protocol was approved by the Ethics Committee of Zhejiang Provincial People's Hospital (QT2022365).

In this case–control study, we enrolled patients with hysteroscopic video records obtained between January 2019 and December 2023 at two tertiary centres (Zhejiang Provincial People's Hospital and The First Affiliated Hospital, Zhejiang University School of Medicine). The study protocol was approved by the ethics committee of Zhejiang Provincial People's Hospital (QT2022365). All participants provided their informed consent. The inclusion criteria were patients with CSD diagnosed using transvaginal ultrasound who underwent hysteroscopy and had a complete video record. The exclusion criteria were miscarriage at the time of hysteroscopy, postmenopausal status, lack of satisfactory video records and presence of other uterine pathologies of AUB, including haemorrhagic polyps in the uterine cavity, endometrial hyperplasia, malignancy, submucosal myomas and adenomyosis. Patients were grouped into the AUB and non‐AUB groups according to their symptoms of pcAUB, which was defined as brownish discharge immediately after a menstrual period with a total duration of menstrual bleeding exceeding 7 days [ 10 ]. Detailed clinical information, including age, body mass index (BMI), reproductive history, menstrual history and ultrasound findings, was collected. Thirty‐eight women in the AUB group and six women in the non‐AUB group underwent laparoscopic surgery owing to AUB, intrauterine effusion and repeated implantation failure. Non‐contact hysteroscopy without dilation was performed by experienced gynaecologists who were not blinded to the patient's obstetric history using a 30° rigid 5–6 mm Bettocchi hysteroscope (Karl Storz, Tuttlingen, Germany) with a 5 Fr manipulating channel under a distension pressure of 80–100 mmHg. Two experienced hysteroscopists independently reviewed the hysteroscopy video records. The conditions of the cervix, diverticulum and uterine cavity were carefully reviewed to evaluate certain features, including bloody mucus, flap‐like fibrotic edges, size and position of the diverticulum, epithelial deficiency, isolated endometrial remnants, blood vessels and haemorrhagic lesions. Thirteen patients in each group who underwent biopsies of both the CSD surface and uterine endometrium underwent pathological comparisons. Haematoxylin–eosin (HE) and immunohistochemical (IHC) staining of cluster of differentiation (CD) 31 (CD31) and CD138 were used to evaluate endometrial thickness at the diverticulum and vascular patterns of CSD as well as CE, respectively. Histopathology was analysed using ImageJ software (National Institutes of Health, Bethesda, USA). For the HE staining, tissues were fixed with 4% paraformaldehyde and embedded in paraffin. Sections, 5‐μm thick, were prepared and routinely dewaxed, followed by HE staining using the standard protocol. The thickness of the endometrium, including both the epithelium and stroma, was measured as the average thickness at five random positions on each slide. For the IHC staining, paraffin‐embedded tissue sections were deparaffinised and rehydrated. After antigen retrieval and endogenous peroxidase activity blocking, the slides were incubated with a specific primary antibody at 4°C overnight, with a secondary antibody at 37°C for 25 min, and then with haematoxylin for 30 s. The anti‐CD138 antibody (1:100, ZA0584) used in this study was purchased from Zhongshan Jinqiao (Beijing, China), and the anti‐CD31 (1:1000, ab182981) antibody and the secondary antibody (1:2000, ab205718) were purchased from Abcam (Cambridge, UK). The area of blood vessel density (CD31‐positive tubules composed of endothelial cells) was calculated across five random high‐magnification fields as a proportion of the positive area. The diagnostic criterion for CE is the presence of one CD138‐positive plasma cell in the endometrial stroma per high‐power field (HPF) [ 11 ]. As the AUB group had a significantly wider and deeper niche with more previous caesarean sections than the non‐AUB group, propensity score matching (PSM) was employed to adjust for these confounding factors. Patients from the AUB group were matched 1:1 to non‐AUB patients using nearest neighbour matching without replacement, with a caliper of 0.2. The variables considered in the PSM were the width and depth of the niche and the number of previous caesarean sections. Measurement data were presented as mean (standard deviation). Between‐group differences in normally distributed continuous data were compared using a two‐sided t‐ test. For non‐normally distributed data, as detected using the Shapiro–Wilk test, the non‐parametric Mann–Whitney U test was used. Categorical data were presented as frequency and proportion [ N (%)], with the Fisher's exact test used for comparisons. The prevalence of hysteroscopic features was compared between the AUB and non‐AUB groups, and odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. The significance level, or two‐sided type 1 error rate (α), was set at 0.05. The power of the test (1 −  β , where β represents the type 2 error rate) was calculated using the observed effect size and actual sample size. Power ≥ 0.8 was considered satisfactory. Statistical analyses were conducted using R version 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria).

In total, 155 patients with CSD and satisfactory hysteroscopy video records were included in our study (69 with AUB and 86 without AUB) (Figure  S1 ). The AUB group exhibited a significantly wider and deeper diverticulum, a history of more caesarean sections and a higher prevalence of retroverted uterus (Table  1 ). These results indicate that in patients with CSD, those with AUB had more severe myometrial defects than those without AUB, which is consistent with the fact that larger CSDs cause more severe symptoms in general. Characteristics of patients with caesarean scar diverticulum before and after PSM. Abbreviations: AUB: abnormal uterine bleeding; BMI: body mass index; C/S: caesarean section; CSD: caesarean scar defect; n : number; PSM: propensity score matching. In order to balance the severity of myometrial defects, 30 patients with AUB were successfully matched with 30 patients without AUB after PSM. No statistically significant differences were observed in patient characteristics between the two groups after matching, including age, BMI, niche width and depth, number of previous caesarean sections and uterine position (Table  1 ). After PSM, the AUB group had a significantly longer menstrual period than the non‐AUB group (14.3 ± 3.7 vs. 5.5 ± 1.4 days, p  < 0.001). Hysteroscopy revealed several common features in the AUB group, including bloody mucus production, flap‐like fibrotic edges, spots of isolated endometrial remnants, exposed vessels, hyperplastic vessels and in situ haemorrhagic lesions (Table  2 and Figure  1 ). The following five features were significantly different between the AUB and non‐AUB groups (Figure  1 ): bloody mucus (OR 7.7 [2.2–31.1], p  = 0.001), epithelial deficiency (OR 5.5 [1.5–23.7], p  = 0.007), exposed vessels (OR 5.5 [1.6–20.6], p  = 0.004), hyperplastic vessels (OR 6.3 [1.9–24.2], p  = 0.002) and haemorrhagic endometrium (OR 40.7 [5.3–1852.0], p  < 0.001). Features of caesarean scar defect under hysteroscopy in patients with or without AUB. Abbreviations: α : type 1 error rate; β : type 2 error rate; AUB: abnormal uterine bleeding; n : number; PSM: propensity score matching. Hysteroscopic findings of caesarean scar endometrial defects. (A) Typical hysteroscopic findings. (B) Odds ratio of the presence of characteristic hysteroscopic features in the abnormal uterine bleeding (AUB) group compared with that in the non‐AUB group after propensity score matching. Consistent with hysteroscopic features, the pathological evaluation also revealed epithelial and vascular defects at the CSD site in the AUB group (Figure  2 ). The endometrial thickness on the CSD surface in the AUB group (0.23 ± 0.17 mm) was significantly lower than that in the non‐AUB group (1.02 ± 0.16 mm; p  < 0.005) (Figure  2I ). Niche resection showed epithelial deficiency and isolated endometrial residuals. A large number of small and congested capillaries were frequently observed near the surface of blood vessels. Upon counting the stained area of CD31‐positive cells in five different HPFs of view (HPF 400×), we found that the percentage of positive cell area was significantly higher in the AUB group than that in the non‐AUB group (5.77% ± 0.18% vs. 1.12% ± 0.13%, p  < 0.001) (Figure  2J ). Signs of fibrosis were observed in the perivascular stroma in the AUB group. The total number of CD138‐positive cells at CSD was significantly higher in the AUB group than that in the non‐AUB group (11.31 ± 5.30 vs. 4.53 ± 2.43, p  = 0.026, Figure  2K ). Eleven CD138‐positive patients (nine in the AUB group and two in the non‐AUB group) were diagnosed with CE and received antibiotic treatment. All nine patients with AUB and CE showed improved AUB symptoms after surgery and antibiotic treatment. Typical pathological findings in patients with caesarean scar defects with or without AUB. (A, B) Haematoxylin–eosin staining of CSD. (C, D) CD31 staining of CSD. (E, F) CD138 staining of CSD. (G, H) CD138 staining of the uterine endometrium. (I–K) Quantitative results of pathological findings. *: epithelial deficiency; ∆: isolated endometrial remnant; red arrow: exposed and hyperplastic vessels; back arrow: CD138+ plasma cell. AUB: abnormal uterine bleeding; CD: cluster of differentiation; CSD: caesarean scar defect; HPF: high‐power field. Menstrual recovery and postmenstrual bleeding were reviewed during the postoperative follow‐up of 23 patients in the AUB group who underwent laparoscopic surgery. Among them, 18 (78.3%) reported improved symptoms, whereas 5 (21.7%) reported no significant change in pcAUB. There were 24 patients (15 in the AUB group and 9 in the non‐AUB group) who desired pregnancy. In the AUB group, six patients experienced full‐term live births after caesarean section, with an uneventful obstetrical course; one was in the third trimester at the time of this report; two had miscarriages; two were in the 1‐year contraception interval and the remaining four did not experience conception. In the non‐AUB group, six women had full‐term live births, one was in the third trimester at the time of this report and the other two did not experience conception. No dehiscence was observed in the lower part of the uterus during subsequent caesarean sections in either group.

To our knowledge, this is the first study to explore the association between endometrial defects (demonstrated by hysteroscopic and pathological features) in CSD and pcAUB symptoms. In line with earlier research evaluated with sonography, we found that pcAUB is linked to wider and deeper niches and an increased number of previous caesarean sections [ 12 , 13 ]. To adjust for these factors, we performed PSM between the AUB and non‐AUB groups. Our findings revealed that even after adjusting for niche size and the number of previous caesarean sections, the AUB group still had a significantly higher prevalence of epithelial deficiency, exposed hyperplastic vessels and haemorrhagic endometrium under hysteroscopy. The retention of menstrual blood due to fibrotic tissue at the edges of the niche, which impairs normal myometrial contractions, has long been believed to be the aetiology of niche‐related postmenstrual spotting [ 14 , 15 ]. From an anatomical perspective, CSD stems from poor healing of the uterine scar, encompassing both CSMD and CSED. CSMD, which is associated with niche formation and fluid accumulation, has long been considered a cause of pcAUB and an indicator of symptom severity. This conclusion is supported by our results, which showed that the AUB group had larger niche sizes before PSM. However, CSED, which is associated with local recurrent spontaneous bleeding and exudation, has long been overlooked [ 9 ]. Our study found that CSED features conferred a significantly higher risk of pcAUB after PSM for CSMD. Thus, CSMD (represented by niche size) and CSED (represented by epithelial thinning and exposed vessels) may jointly influence symptoms and pathological features. After uterine incision healing, the CSMD structure stabilises. However, owing to cyclical changes in sex hormones, the endometrium covering the scar continues to undergo periodic proliferation, maturation and shedding. The caesarean incision typically lies on the isthmus, where the lining transitions from the endometrium to a less regenerative endocervical mucosa. Because of epithelial defects, the endometrium cannot completely cover the myometrium, making CSED a chronically unhealed wound that undergoes periodic bleeding, which is likely the real cause of pcAUB. Endometrial defects in CSD might increase tissue exudation and bleeding, disrupting the microbiota composition in the cervix and uterine cavity. This disruption can lead to abnormal metabolism and disordered immune responses, which can induce apoptosis in epithelial and vascular endothelial cells and lead to CE in CSD [ 16 , 17 ]. The epithelial defect and endometritis at the CSD site may be mutually causative, creating a vicious cycle. Our study further confirmed that the AUB group manifested more typical aberrations of the epithelial lining, in situ bleeding from superficial vessels and infiltrating CD138+ plasma cells, as reported in other studies [ 18 , 19 ]. CE is a chronic inflammatory state of the endometrium caused by abnormal microbiomes, which leads to repeated implantation failure and recurrent miscarriage [ 20 ]. Empirical medication is recommended when endometritis features are present. The current first‐line antibiotic treatment is bacteria specific or empiric: doxycycline 200 mg/day for 14 days or ciprofloxacin 1 g/day and metronidazole 1 g/day for 14 days [ 21 , 22 ]. Treatment with surgery and antibiotics was effective in relieving AUB in this study. However, whether the improvement comes from the surgery, antibiotics or both requires further investigation. The limited regenerative capacity of the damaged endometrium, chronic incomplete healing and periodic shedding induce inadequate vascular regeneration and compromise self‐haemostasis mechanisms, further exacerbating in situ bleeding. Long‐term in situ bleeding leads to blood accumulation in diverticula, fluid in the uterine cavity, CE, missed ovulatory intercourse, sperm transport issues and impaired endometrial receptivity, contributing to sub‐infertility [ 23 , 24 , 25 ]. Imaging techniques such as transvaginal sonography, hysterosalpingography and magnetic resonance imaging focus on evaluating CSMD and fluid accumulation. However, hysteroscopy, combined with pathological examination, can directly assess the diverticulum; determine the nature of uterine cavity fluid, endometrial defects and vascular patterns; and evaluate endometritis, which are directly related to reproductive outcomes. This makes hysteroscopy particularly valuable for precise assessments in infertile patients with a history of caesarean section, especially those with combined pcAUB or uterine cavity fluid. On the basis of our experience, we recommend performing hysteroscopy on Days 9–12 of the menstrual cycle to evaluate epithelial repair and bleeding conditions. A non‐contact hysteroscope with a distension pressure of 80–100 mmHg is preferable to avoid disturbing the natural state of the epithelium. After observing CSD, the distension pressure should be reduced to help identify the vascular pattern, and if necessary, an endometrial biopsy should be performed at the CSD site. Finally, evaluation and biopsy of the uterine cavity should be performed. Hysteroscopic evaluation can also guide treatment approaches for CSD. Some authors have proposed two key points for the hysteroscopic management of CSD. One is the resection on both edges of the scar diverticulum to eliminate the valve effect and the second is to destroy the epithelium by electrocoagulation [ 26 ]. However, when severe CSED is discovered during hysteroscopy, complete excision and repair of the CSD are recommended because hysteroscopy cannot strengthen the uterine wall, whereas laparoscopy and transvaginal repair allow better anatomical recovery [ 27 , 28 ]. Our arguments regarding the relationship between CSED and pcAUB also suggest the need for tissue healing strategies as alternatives, such as stem cell and regenerative medicine therapies. A limitation of this study was that the gynaecologists performing the hysteroscopies were not blinded to the patient's obstetric history. However, we attempted to mitigate this bias by involving two independent gynaecologists in the evaluations. The conclusions of the study were also constrained by the relatively small sample size. Nonetheless, the study was sufficiently powered, given that most comparisons of the hysteroscopic features had power > 0.8 (Table  2 ). The differences in endometrial remnants and flap‐like fibrotic edges between the AUB and non‐AUB groups did not reach significance or the desired power, probably due to the limited sample size, which calls for further research. After PSM, the width of the CSD and the time interval from hysteroscopy to the latest caesarean section showed no significant differences between the two groups. However, with a larger sample size, these differences may reach statistical significance.

Utilising hysteroscopic and pathological methodologies, we found that CSED exhibited a strong association with pcAUB severity after adjusting for CSMD. These findings are pivotal in advancing our understanding of this phenomenon and could lead to more targeted and effective treatment strategies for this common post‐caesarean complication. Additionally, we provided expertise regarding performing hysteroscopy in the context of CSD, offering significant practical guidance for clinicians in the field.

Caesarean section, the most prevalent surgical procedure in obstetrics, has recorded a global increase in its application, resulting in increased long‐term postoperative complications [ 1 ]. Caesarean scar defect (CSD), also known as caesarean scar diverticulum, uterine niche or isthmocele, is a common complication after caesarean section that results from suboptimal incisional healing, with an incidence as high as 56%–84% worldwide [ 2 ]. Patients with CSD may experience caesarean scar syndrome, including chronic pelvic pain, secondary infertility and post‐caesarean abnormal uterine bleeding (pcAUB) or postmenstrual spotting [ 3 ], which is the predominant symptom found in 33.6% of women with CSD [ 4 ]. Diverticulum formation due to the dehiscence of isthmic myometrium has long been proposed as the mechanism underlying pcAUB, which could lead to blood retention and delayed emptying within the defective cavity [ 5 ]. In patients with pcAUB, ultrasonography shows myometrial weakness and hysteroscopy usually shows characteristic fibrotic tissue, abnormal vascular morphology and bleeding [ 6 ]. Pathological evaluation shows signs of myometrial distortion and chronic inflammation with fibrosis [ 7 , 8 ]. Nevertheless, as only one‐third of all niches have menstrual disturbances and the size of CSD does not necessarily correlate with the symptoms, there remains a gap in the literature regarding the formation of CSD and pcAUB. Previously, we proposed that endometrial defects in CSD were an under‐recognised cause of pcAUB [ 9 ], in addition to myometrial defects. Although caesarean scar endometrial defects (CSEDs) and caesarean scar myometrial defects (CSMDs) are usually positively correlated, distinguishing between these two pathological mechanisms would help clarify how structural defects cause symptoms. A systematic evaluation of hysteroscopy and pathology could explain how endometrial defects and chronic endometritis (CE) contribute to pcAUB and infertility. Moreover, in addition to resculpting or resecting myometrial defects, attention must be paid to the recovery of endometrial defects during hysteroscopic isthmoplasty and laparoscopic or vaginal repair of CSD. We hypothesised that CSED features observed by hysteroscopy or pathology might be associated with clinical manifestations of CSD and could be used to determine the contribution of CSED to pcAUB after adjustment for CSMD. Thus, in this study, we aimed to explore the role of endometrial defects in the pathogenesis of AUB.

The authors declare no conflicts of interest.

Figure S1. Study flow‐chart.