Association Between Hysteroscopic Treatment of Cesarean Scar Disorder and Changes in the Endometrial Microbiome and Clinical Outcomes: A Prospective Observational Study

All procedures complied with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and its later amendments. The study was approved by Sakarya University, Faculty of Medicine Ethics Committee (approval number 71522473/050.01.04; date March 5, 2020).

This work was supported by Sakarya Üniversitesi (2021‐7‐25‐76).

A total of 40 women who met the eligibility criteria were enrolled in the study. Their demographic characteristics are summarized in Table  1 . Demographic characteristics of the women who participated in the study. Abbreviations: BMI, body mass index; CS, cesarean section. Time elapsed since the last CS before hysteroscopy. Among the participants, 36 (90%) presented with AUB, 18 (45%) with secondary infertility, and 36 (90%) with chronic pelvic pain. The average age of women diagnosed with secondary infertility due to isthmocele was 30.7 years, and the mean duration of secondary infertility was 3 years. None of the women were diagnosed with infertility related to male factor, tubal factor, low ovarian reserve, recurrent implantation failure, or recurrent pregnancy loss. Among the 18 women diagnosed with secondary infertility, 12 (66.7%) achieved spontaneous pregnancy within the first year after isthmocele surgery, whereas 6 (33.3%) did not. H/S findings indicated symptoms consistent with chronic endometritis in all participants. Additionally, endometrial polyps were detected in six women (15%). Histopathological analysis of endometrial samples obtained after H/S revealed a proliferative endometrium in 16 cases (40%), polyps in eight (20%), fibrotic tissue in five (12.5%), early secretory endometrium in four (10%), endometrial tissue in four (10%), and secretory endometrium in two (5%). The most common complaints reported by women with isthmocele during outpatient visits were postmenstrual spotting, groin pain, and dyspareunia. Outpatient clinic follow‐ups after hysteroscopic isthmoplasty showed a significant decrease or resolution of all complaints. To objectively evaluate preoperative and postoperative changes in pelvic pain and bleeding irregularities, patients completed the Visual Analog Scale (VAS) and the Pictorial Blood Loss Assessment Chart (PBAC) scores during follow‐ups. Patients are asked to rate the pain felt during VAS on a scale of 1 to 10. The average preoperative pain level in women with isthmocele was 5.33, whereas this value decreased to 2.44 at 3 months postoperatively and to 1 at 6 months ( p  < 0.001) (Table  2 ). Changes in visual analog scale (VAS) and PBAC scores over time. 5.33 ± 2.54 b 6.00 (0.00–9.00) 2.44 ± 2.39 b 2.00 (0.00–7.00) 1.08 ± 1.97 b 0.00 (0.00–6.00) 151.45 ± 59.78 b 145.50 (67.00–264.00) 85.52 ± 22.96 b 81.00 (59.00–161.00) 68.93 ± 11.65 b 68.50 (51.00–111.00) Abbreviations: PBAC, Pictorial Blood Loss Assessment Chart; VAS, visual analog scale. Friedman analysis was performed. Group from which the difference originates. In the PBAC application, menstrual bleeding is considered pathologically excessive when the total score, calculated from daily pad usage during menstruation, exceeds 100. The mean preoperative PBAC score among the women in the study was 151, which decreased to 85 at 3 months and 69 at 6 months ( p  < 0.001) (Table  2 ). To evaluate the impact of hysteroscopic isthmoplasty on the quality of life of women with CSDi, the research form included questions addressing menstrual pathology parameters commonly associated with CSDi, such as prolonged menstruation, postmenstrual spotting, and dysmenorrhea, and the symptoms of dyspareunia. Significant reductions in menstrual bleeding and postmenstrual brown spotting duration, dysmenorrhea, and dyspareunia were observed over time at the 3‐month and 6‐month follow‐ups compared with the preoperative period ( p  < 0.001) (Table  3 ). Changes in bleeding‐related characteristics over time in 40 women. Abbreviation: MB, menstrual bleeding. Cochran Q test was performed. Indicates the group from which the difference originates. A significant reduction ( p  < 0.001) was observed over time in RMT, CSD depth, CSD width, CSD depth and width in the transverse section, CSD‐cervical ext. os distance, and CSD area measured by TV‐USG at the preoperative, 3‐month, and 6‐month assessments. Specifically, the reduction in RMT was attributable to differences between the preoperative, 3‐month, and 6‐month values, whereas all other parameters showed significant changes across all time points (Table  4 ). Change in USG measurements of isthmocele parameters over time. 5.68 ± 2.84 b 4.90 (1.40–13.90) 5.82 ± 2.29 b 5.55 (2.40–10.40) 8.68 ± 2.95 b 8.45 (4.00–17.00) 6.92 ± 2.65 b 6.40 (3.40–18.00) 3.98 ± 1.54 b 3.80 (0.00–7.10) 1.85 ± 1.95 b 2.10 (0.00–6.20) 8.43 ± 3.84 b 7.55 (2.80–16.70) 4.87 ± 2.32 b 4.50 (0.00–10.00) 2.28 ± 2.43 b 2.50 (0.00–7.80) 11.30 ± 6.18 b 10.00 (0.00–33.00) 6.29 ± 5.11 b 5.20 (0.00–31.00) 2.69 ± 3.47 b 0.00 (0.00–13.30) 5.25 ± 3.06 b 4.10 (0.00–15.80) 3.28 ± 1.91 b 3.00 (0.00–9.00) 1.51 ± 1.84 b 0.00 (0.00–6.10) 23.20 ± 3.89 b 22.00 (17.00–34.00) 29.28 ± 4.59 b 28.50 (22.00–39.00) 33.05 ± 6.92 b 34.00 (0.00–45.00) 29.76 ± 21.86 b 21.59 (8.50–123.30) 10.72 ± 7.93 b 8.83 (0.00–33.50) 4.08 ± 4.57 b 3.80 (0.00–15.81) Note: A. Residual myometrial thickness (RMT): Distance from the CSD (cesarean scar defect) ceiling to the CVB (cervico vesical border), representing the distance between the CSD and the uterine serosa. B. CSD depth: Distance from the CSD ceiling to the endocervical canal, measuring the distance between the CSD ceiling and line C. C. CSD base width: Distance between the lower lips of the CSD facing the endocervical canal. D. Isthmocele width (transverse section). E. Isthmocele depth (transverse section): Depth of the isthmocele measured at the midpoint of measurement D in the transverse section. F. Distance between the lower edge of the isthmocele and the external cervical os. Tx: Transverse section of USG. Abbreviation: CSD, cesarean scar defect. Friedman analysis was performed. Indicates the group from which the difference originates. The surface area of the triangular‐like isthmocele was calculated using the formula: Surface area = Height × Base/2, where height refers to the vertical depth (measured in mm) in the longitudinal section on TV‐USG (transvaginal ultrasonography), and base refers to the width of the isthmocele in the longitudinal section on TV‐USG. Conventional microbiological cultures were performed on endometrial lavage solutions obtained from the 40 participants. Growth in conventional cultures of first samples was detected in six patients: Lactobacillus acidophilus in two, Candida kefyr in two, Lactobacillus gasseri in one, and Escherichia coli in one. The patient with Escherichia coli received targeted antibiotic therapy based on the results of antibiotic susceptibility testing. In the second sampling performed at the 3rd postoperative month, Lactobecillus gasseri growth was detected in one case. The relative abundance of bacterial genera before and after surgery is presented in Figure  2 , which illustrates a decrease in microbial diversity following treatment. Figure  2 shows the percentages of bacterial genera detected in each sample within the total microbial community. Relative abundance reflects the proportion of a given taxon within the total microbiota. Colors represent different taxonomic groups. The bar chart allows comparisons of microorganisms detected in lavages taken from the cavity before and after hysteroscopy. The relative abundance graph at the genus level before and after surgery. From left to right, the first two columns with the same number represent samples from the same patient. The column labeled B (Before) indicates the pretreatment sample, and the adjacent column labeled A (After) indicates the post‐treatment samples. For example, S01B represents the bacterial diversity density of the same patient before surgery, while S01A represents the bacterial diversity of the same patient 3 months after hysteroscopy. Before treatment, the predominant genera included Clostridia , Bacteroidia , Lachnospiraceae , Prevotella , Ruminococcaceae , Staphylococcus , Coprococcus , and Corynebacterium . After surgery, the predominant taxa shifted to Lactobacillales , Bacilli , and Firmicutes ( p  < 0.001). From left to right, the first two columns with the same number represent samples from the same patient. The column labeled “B” (Before) indicates the pretreatment sample, and the adjacent column labeled “A” (After) indicates the posttreatment sample. For example, “S01B” represents the bacterial diversity density of the same patient before surgery, while “S01A” represents the bacterial diversity density of the same patient 3 months after hysteroscopy. Before treatment, the predominant taxa included Clostridia, Bacteroidia, Lachnospiraceae, Prevotella, Ruminococcaceae, Staphylococcus, Coprococcus , and Corynebacterium . After surgery, the predominant taxa shifted to Lactobacillales, Bacilli , and Firmicutes ( p  < 0.001), Figure  2 . A comparative metagenomic analysis was conducted on 44 samples, including 22 collected preoperatively and 22 collected 3 months postoperatively, to evaluate bacterial composition. The Shannon index, which accounts for both species richness and the balance of species distribution, was used to assess microbial diversity. A higher Shannon index reflects greater diversity and a more balanced distribution. Analysis revealed that the Shannon index of endometrial lavage samples collected 3 months after hysteroscopy was statistically significantly lower than that of preoperative samples ( p  = 0.0214) (Figure  3 ). The p ‐values of the Simpson and Chao1 indices were 0.0257 and 0.0316, respectively. Beta diversity analysis revealed no statistically significant difference between the pretreatment and posttreatment groups (Bray–Curtis PCoA PERMANOVA, p  = 0.362; Jaccard PERMANOVA, p  = 0.859). Alpha diversity analysis results. The Shannon index evaluates both species richness and evenness within a microbial community, with higher values indicating greater diversity and a more balanced distribution. Microbial diversity in endometrial lavages collected 3 months after hysteroscopy was significantly reduced compared with preoperative samples ( p  = 0.0214). LEfSe is a statistical method used to identify microbial taxa that differ significantly between groups. This analysis proceeds in three stages: (i) taxa with significant differences between groups are first identified using the Kruskal–Wallis test, (ii) pairwise comparisons are then performed using the two‐tailed Wilcoxon test, and (iii) the effect size of these differences is calculated using LDA. This method is widely applied to identify potential microbial biomarkers. The taxonomic distribution before and after treatment is presented in Figure  4 . In the pretreatment group, Corynebacteriales , Bacteroidia , Bacteroidetes , Clostridia , Bacteroidales , and Clostridiales were enriched at the order level, with Corynebacteriales the most abundant at the family level. Corynebacteriaceae , Rikenellaceae , Staphylococcaceae , Bacteroidales , Bacteroidia , Bacteroidetes , Ruminococcaceae , Clostridiales , Clostridia , Prevotellacea , and Lachnospiraceae were also prevalent in the pretreatment group. LDA/LEfSe analysis graph at different taxonomic levels comparing pretreatment and posttreatment groups. In the pretreatment group, the taxa enriched at the order level included Corynebacteriales , Bacteroidia , Bacteroidetes , Clostridia , Bacteroidales , and Clostridiales , with Corynebacteriales also abundant at the family level. Additional taxa enriched in this group were Corynebacteriaceae , Rikenellaceae , Staphylococcaceae , Bacteroidales , Bacteroidia , Bacteroidetes , Ruminococcaceae , Clostridiales , Clostridia , Prevotellacea , and Lachnospiraceae . In the post‐treatment group, Bacilli and Firmicutes ( Bacillota ) were highly abundant at the class level; Lactobacillales , Bacilli , and Firmicutes (novel name: Bacillota ) were enriched at the order level; and Lactobacillaceae , Lactobacillales , Bacilli , and Firmicutes were abundant at the family level ( p   2). At the species level, Lactobacillus gallinarum was significantly more abundant in the post‐treatment group than in the pretreatment group. In the post‐treatment group, Bacilli and Firmicutes ( Bacillota ) were highly abundant at the class level; Lactobacillales , Bacilli , and Firmicutes ( Bacillota ) were enriched at the order level; and Lactobacillaceae , Lactobacillales , Bacilli , and Firmicutes were abundant at the family level ( p   2). At the species level, Lactobacillus gallinarum was significantly more abundant in the post‐treatment group than in the pretreatment group (Figure  4 ).

CS rates are steadily increasing worldwide, leading to a higher incidence of cesarean‐related complications [ 22 ]. With growing clinical experience and expanding research, clinicians now have a better understanding of CSDi than in the past. In this study, we aimed to define the clinical success criteria for women undergoing surgical treatment for CSDi and to explore their association with the endometrial cavity microbiome. Although no universally accepted gold‐standard imaging method exists for diagnosing isthmocele, TV‐USG is commonly preferred. In a retrospective study involving 92 women, Fabres et al. reported that identifying a triangular anechoic shape on TV‐USG offers a cost‐effective and straightforward diagnostic approach, demonstrating 100% correlation with hysteroscopy findings [ 3 ]. Similarly, our study showed a 100% correlation between hysteroscopy and TV‐USG, mirroring the findings of Fabres et al. and Raimondo et al.'s [ 3 , 23 ] studies that recommended conducting the examination between the 7th and 12th day of the menstrual cycle. Consistent with these recommendations, we conducted our measurements during the follicular phase of the menstrual cycle. Several methods have been proposed for isthmocele measurement. Park et al. [ 10 ], in a study of 404 women, developed a schematic diagram illustrating isthmocele measurements in the longitudinal uterine plane. We adopted a similar approach, labeling the RMT above the isthmic isthmocele sac on the longitudinal TV‐USG section as “A,” the depth of the isthmocele sac as “B,” and the base width as “C.” Unlike Park et al. [ 10 ], we also measured the isthmocele diameter in the transverse TV‐USG section, labeling the length of the isthmocele sac as “D” and height as “E.” These measurements were inspired by the “modified Delphi procedure” proposed by Jordans et al. [ 11 ] Additionally, we measured the distance between the lower edge of the isthmocele sac and the external cervical os as “F” in the longitudinal section (Figure  1a,b ). Postoperative TV‐USG measurements at 3 and 6 months demonstrated a statistically significant reduction in the size of the isthmocele sac at all measurement locations (Table  4 ). Specifically, the A value increased at both time points, while the B, C, D, and E values decreased. The F value also increased. Additionally, on TV‐USG in the longitudinal section, we detected a significant reduction in the isthmocele surface area in measurements performed at the 3rd and 6th months after surgery compared with preoperative values: 29.7, 10.7, and 4.1 mm 2 , respectively (Table  4 ). Several surgical techniques have been proposed for treating isthmocele. In a prospective study involving 26 patients, Gubbini et al. [ 24 ] suggested that simultaneous isthmocele resection of the proximal and distal edges could be more effective in preserving the continuity of the cervical canal and minimizing the accumulation of menstrual blood. They also proposed that coagulating the roof might stop blood and mucus production [ 24 ]. By contrast, Raimondo et al. [ 23 ] argued that coagulation of the isthmocele roof was unnecessary. Other studies have suggested that resecting only the distal margin is sufficient [ 3 , 25 ]. We performed an arc‐opening resection of the distal part of the isthmocele sac. The arc was shaved at the distal end of the isthmocele sac toward the external os as extensively as possible to prevent menstrual blood accumulation and ensure the continuity of the cervical canal. Superficial veins and erythematous areas at the base of the isthmocele sac were coagulated with the same electrode tip using spray coagulation mode to accelerate healing (Video  S1 ). Although we resected only the distal end of the isthmocele sac, our postoperative sac closure rates were notably high. This outcome may be attributable to the distinct histological differences between the proximal and distal ends of the isthmocele sac. At the proximal end, the myometrium consists of three layers of smooth muscle fibers. As the thick layer of the uterine muscle thins toward the cervix, it transitions into collagen‐rich tissue at the uterocervical junction [ 26 ]. Following hysteroscopic isthmoplasty, in which abnormal scar tissue from a CS is excised, we believe the reappearance of collagen within the cervical stroma and lower uterine segment plays a biological role in healing. However, the specific collagen subtype present after hysteroscopic isthmoplasty and the mechanisms of extracellular matrix remodeling that fill the excised scar space remain unclear. To date, most studies have focused primarily on anatomical restoration and clinical outcomes. Studies evaluating histological, microbiological, molecular, and structural changes in the area before and after isthmoplasty are still needed. Our study addresses this gap by presenting findings on alterations in the local microbial flora. Romero et al. [ 27 ] considered the endometrial cavity as a sterile environment but reported that the presence of Enterobacteriaceae spp., Streptococcus spp., Staphylococcus spp., Escherichia coli , and other Gram‐negative bacteria adversely affected the success of in vitro fertilization (IVF). By contrast, other studies using classic culture methods have isolated Lactobacillus spp., Mycoplasma hominis, Gardnerella vaginalis , and Enterobacter spp. from the endometrial cavity samples of women who underwent hysterectomy [ 28 ]. In recent years, alternative methods have also been applied to the treatment of chronic endometritis. Kuroda et al. [ 29 ] diagnosed chronic endometritis when ≥ 5 CD138‐positive cells were detected in biopsies obtained during hysteroscopy performed for conditions such as endometrial polyps, intrauterine adhesions, and submucosal myomas. Among these women, those who declined antibiotic therapy underwent repeat pipelle endometrial biopsies during the first two cycles after surgery to reassess CD138 levels. In approximately 83% of these cases, the number of CD138‐positive cells fell below the cutoff value (≤ 4), indicating resolution of endometritis. Kuroda et al. [ 29 ] considered that the resolution of endometritis, despite the absence of antibiotic therapy, may have resulted from the combined effects of simultaneous irrigation with uromatic distension fluid and hysteroscopic surgery. Nobuta et al. [ 30 ] investigated the prevalence of chronic endometritis in women affected by infertility with and without CSDi using hysteroscopy and measured tumor necrosis factor‐α (TNF‐α) and interleukin‐1β (IL‐1β) levels in endometrial washings by enzyme‐linked immunosorbent assay. TNF‐α and IL‐1β levels were found to be statistically significantly higher in infertile cases of CSDi than in their counterparts [ 30 ]. The fact that CD‐138–positive plasma cell ratios were not evaluated in endometrial biopsy samples to support the endometritis findings observed during hysteroscopy is an important limitation of our study. A 2015 study investigating 12 bacterial species using quantitative PCR in endometrial swab samples obtained from patients undergoing hysterectomy detected at least one bacterial species in 95% of cases, with Lactobacillus iners , Prevotella spp., and Lactobacillus crispatus being the most common. Prior to microbiome studies employing array analysis, molecular methods similar to those employed in this study confirmed the presence of microorganisms in endometrial samples [ 31 ]. However, with advances in NGS, such research is increasingly contributing to the literature. Given the capabilities of NGS 16S rRNA gene technology, microbiome analysis without its use would now be considered insufficient. Although research remains limited, evidence suggests that the endometrial cavity is not sterile. These hypotheses are supported by microbiome findings obtained from NGS analysis of samples we obtained from the endometrial cavity before and after hysteroscopic isthmoplasty. While Clostridia, Bacteroidia, Lachnospiraceae, Prevotella, Ruminococcaceae, Staphylococcus, Coprococcus , and Corynebacterium dominated in pretreatment samples, Lactobacillales, Bacilli , and Firmicutes were dominant after surgery ( p  < 0.001) (Figures  2 and 4 ). Moreno et al. [ 32 ] employed NGS of the 16S rRNA gene to analyze endometrial and vaginal samples from 13 women with proven fertility and endometrial samples from an additional 22 women with proven fertility. In addition to identifying distinct microbial profiles between the endometrium and the vagina, they reported that the composition of the endometrial microbiota correlated with IVF outcomes. Specifically, while Lactobacillus species predominated, other bacteria, including Gardnerella, Prevotella, Atopobium, and Sneathia, were also present. Furthermore, samples with higher Lactobacillus abundance were associated with higher live birth rates [ 32 ]. Similarly, Nandagopal et al. [ 33 ] compared cultures and PCR‐based diagnoses of chronic endometritis in 500 patients with various gynecological conditions, detecting pathogens in 23% of cases by culture and 63.6% via PCR [ 33 ]. In our study, NGS‐based analysis of endometrial wash fluids revealed the presence of pathological microorganisms in nearly all samples, in addition to lactobacilli. In contrast, conventional culture identified pathogens in only 15% of cases. Beyond pathogen detection, NGS also characterized microbial diversity and the balance of species distribution. In patients with CSDi, alterations in anatomical structure and the accumulation of fluid and menstrual blood within the isthmocele area disrupt the physiological environment. These changes lead to variations in the nutritional properties and pH of the cavity, creating conditions that are conducive to the proliferation of diverse bacterial species. Consistently, bacterial diversity was significantly higher in samples collected before isthmocele repair than in those obtained 3 months after treatment. Molecular analysis in our laboratory confirmed this shift, showing a marked reduction in bacterial diversity posttreatment, with lactobacilli comprising more than 90% of the microbiome (Figures  2 , 3 , 4 ). However, the fact that bacterial change toward the return to normal flora bacteria was demonstrated using relative abundance; in other words, the inability to provide clear quantitative values limits the value of our findings. Furthermore, since all patients received antibiotic treatment after surgery, the observed microbiome changes cannot be attributed solely to the hysteroscopic procedure. In this context, we believe that the changes we observed in the endometrial microbiome may be due to the combined effect of medical and surgical treatments. Our results further indicate that the effectiveness of surgical and medical treatment may be evaluated through microbiome profiling using NGS 16S rRNA sequencing of endometrial samples obtained after treatment. For instance, when a Lactobacillus ‐dominated flora is detected in the sample taken before treatment, the indication for treatment may be reconsidered. Conversely, if postoperative microbiome analysis shows no increase in Lactobacillus abundance and clinical symptoms persist, treatment may be considered insufficient, warranting alternative therapeutic strategies. In our study, we did not perform repeat hysteroscopy to confirm whether endometritis had resolved during the postoperative follow‐up. Instead, we relied on microbiome analysis of endometrial lavage fluid to determine whether endometritis had resolved (Figures  2 , 3 , 4 ). In the 2‐year follow‐up after combined hysteroscopic and medical treatment, 12 (67%) of the 18 infertile patients in our study group conceived spontaneously. These findings are consistent with those in a review by Harjee et al., which summarized reproductive outcomes following isthmocele surgery [ 34 ]. Similarly, Tsuji et al. [ 35 ] reported a pregnancy rate of 71% among 38 patients with CSDi who underwent hysteroscopic isthmoplasty using a technique comparable to ours, with a mean follow‐up of 40 months. Changes in the endometrial cavity microbiome in favor of lactobacilli may have prepared the underlying basis for the occurrence of spontaneous pregnancies. However, considering that our study is observational in nature and the absence of multivariable adjustment for potential confounders, we cannot definitively state whether the microbiome changes we observed are independently associated with pregnancy outcomes. Our study has some limitations. While the clinical outcomes were assessed in the entire cohort, microbiome analysis was conducted in a predefined subgroup of 22 women due to technical and logistical constraints. This approach allowed us to explore potential microbiological correlates of clinical improvement without overextending the interpretive scope of the findings. Thus, this study does not claim a causal relationship between hysteroscopic treatment and microbiome normalization. Rather, the results highlight an association that warrants further investigation. The microbiome findings should be interpreted as exploratory, given the limited sample size and absence of a nonsurgical or non‐medicated control group. As noted by Kuroda et al. [ 29 ], irrigation with the uromatic distension fluid used during hysteroscopic isthmoplasty may have also contributed to the resolution of endometritis. However, we believe that the therapeutic effect of medical treatment on endometritis secondary to CSDi may be transient. By contrast, the surgical correction we performed can prevent recurrence in the long term. The basis for evaluating the endometrial cavity microbiome in the 3rd month after hysteroscopy lies in our clinical experience of observing significant improvement, both in USG findings and in symptom assessments, during postoperative 3rd‐month follow‐ups of patients in whom we performed hysteroscopic isthmoplasty due to CSDi. Future studies evaluating the endometrial cavity microbiome 6 months and beyond after hysteroscopy will substantially contribute to the data obtained. In conclusion, the parallel improvement in clinical symptoms and microbial composition after hysteroscopic isthmoplasty provides a rationale for future prospective, controlled studies integrating microbiological, immunological, histologic, and clinical outcomes.

Informed consent was obtained from all patients included in the study.

Cesarean section (CS) rates have been steadily increasing worldwide in recent years [ 1 , 2 ]. Consequently, early and late complications following CS have become more prevalent. One important late complication is cesarean scar disorder (CSDi), characterized by thinning and disruption of the myometrium, leading to an indentation of the myometrial opening toward the visceral peritoneum [ 3 ]. A cesarean scar defect (CSD) is also known as a niche, uterine diverticulum, or uteroperitoneal fistula. According to the developmental hypothesis, several patient‐related factors that impair wound healing and angiogenesis contribute to the development of CSDi. Additionally, specific surgical techniques, such as a lower uterine incision level and inadequate closure of the uterine incision, are believed to increase the risk of CSDi formation [ 4 ]. Depending on the severity of the defect, affected individuals may experience abnormal uterine bleeding (AUB), infertility, pelvic pain, or cesarean scar pregnancy [ 5 ]. Diagnostic imaging modalities include ultrasonography (USG), hysterosalpingography (HSG), sonohysterography, hysteroscopy (H/S), and magnetic resonance imaging. Among these, transvaginal (TV) USG is the preferred technique because it is noninvasive and cost‐effective [ 3 ]. A complete correlation has been reported between TV USG and H/S findings [ 3 ]. For treatment, management decisions depend on symptom severity; asymptomatic women should not undergo unnecessary interventions [ 6 ]. Surgical options include hysteroscopic, vaginal, laparotomic, laparoscopic, robotic, or combined approaches [ 6 ]. Recently, scientific interest in the role of the intestinal microbiota during pregnancy has increased. Alterations in intestinal microbiota occur as early as the first trimester. Until recently, the placenta was thought to be sterile [ 7 ]. The human microbiota begins to develop at birth, differs among individuals, and typically stabilizes by 2–3 years of age. However, composition at the genus and species levels remains dynamic throughout life. The detection of similar bacterial groups in endometrial and vaginal samples supports the hypothesis that the endometrial microbiome originates from bacteria ascending from the vagina. Nevertheless, differences in physicochemical and biological conditions between the uterus and vagina may result in distinct bacterial profiles. Therefore, endometrial sampling is required to accurately assess the uterine microbiome rather than relying on samples from the lower genital tract [ 8 ]. In women with endometritis‐associated infertility, Lactobacilli , the predominant bacteria in a healthy endometrial microbiome, are reduced. In contrast, pathogenic species such as Atopobium, Gardnerella, Streptococcus, Bifidobacterium, Chryseobacterium , and Klebsiella are increased [ 8 ]. However, the microbiome status within the endometrial cavity and the effect of hysteroscopic isthmoplasty on the endometrial microbiome in patients with isthmoceles remain unclear. This study aimed to evaluate the association between hysteroscopic treatment of cesarean scar disorder (CSDi), clinical improvement, and changes in the endometrial microbiome.

The authors declare no conflicts of interest.

This study was approved by the Sakarya University Faculty of Medicine Ethics Committee (approval number 71522473/050.01.04, March 5, 2020). Additional approval has also been obtained from the General Directorate of Pharmaceuticals and Pharmacy of the Ministry of Health of the Republic of Turkey. The study was conducted as a prospective observational study at our clinic between April 2020 and May 2022. Written informed consent was obtained from all participants. According to the standard protocol, 40 patients presenting with AUB, pelvic pain, or infertility, together with compatible imaging findings on USG and/or HSG, were enrolled in the study. The inclusion criteria were symptomatic isthmocele presenting with AUB, pelvic pain, or infertility; age between 18 and 45 years, a history of CS, regular menstrual cycles; completion of study questionnaires with uninterrupted follow‐up; and a residual myometrial thickness (RMT) > 2 mm on TV‐USG over the isthmocele sac. Exclusion criteria included antibiotic, probiotic, or vitamin use within the past 3 months; oral contraceptive use within the past 3 months; childbirth within the past year; a history of cervical or uterine gynecological procedures within the past 12 months; intrauterine device use; coexisting myometrial pathology such as adenomyosis or cavity‐associated myoma; and intrauterine synechiae detected on hysteroscopy. Patients with infertility due to causes other than CSDi were excluded. An indentation ≥ 2 mm in the cesarean scar area was defined as an isthmocele [ 9 ]. TV‐USG was performed using a Philips Affinity 70 ultrasound system (10 MHz endocavitary probe, Philips Inc., Bothell, WA, USA). All measurements were conducted by a single physician (ASC) during the first 14 days of the menstrual cycle, either during or after menstrual bleeding or spotting. Isthmocele size was recorded in millimeters. Sample measurements are illustrated in Figure  1a,b . Isthmocele size measurements. (a) Longitudinal section of the uterus. (A) Vertical distance from the cesarean scar defect (CSD) ceiling to the cervicovesical border (CVB). It is the distance between the CSD and the uterine serosa, as well as the residual myometrial thickness. (B) Distance from the CSD ceiling to the endocervical canal; depth of the CSD. B is the vertical distance between the ceiling of the CSD and line C. (C) The distance between the lower lips of the CSD facing the endocervical canal: CSD base width. (F) Longitudinal distance between the lower inferior edge of the isthmocele pouch and the external cervical os. (b) Transverse section of the uterus at the level of the endocervical canal. (D) Isthmocele width in transverse section. (E) Isthmocele depth in transverse section; depth in the middle of the D measurement taken in the transverse section. Isthmocele dimensions were assessed using a TV‐USG probe with gentle contact at the cervix. A longitudinal section of the uterus was obtained with the cervical canal in the center. Measurements were performed in the plane where the isthmocele sac borders were most clearly visualized. During the procedure, the bladder was maintained in a moderately filled state to optimize imaging. Measurement techniques for isthmocele sac diameter described by Park et al. and Jordans et al. [ 10 , 11 ] were modified and applied. Measurements were taken as shown in Figure  1a,b , capturing the full length of the cervical canal in the longitudinal section. For branched or complex isthmoceles, the primary isthmocele dimensions were recorded first. Any branching structures were drawn on a measurement form, and their dimensions and distances to the cervico‐vesical border (CVB) were documented. The following parameters were measured: distance A (RMT of the isthmocele, defined as the distance (in millimeters) from the top of the isthmocele to the CVB); distance B (depth of the isthmocele, measured from the ceiling of the isthmocele to the endocervical canal); and distance C (width of the isthmocele, measured as the distance of the plane between the lower edges of the isthmocele adjacent to the endocervical canal). Before transitioning from the longitudinal to the transverse plane, an additional measurement was taken for distance F, defined as the distance from where the lower edge of the CSD adjoins the internal cervical os to the external cervical os. After recording the distance from the CSD lower edge and the cervical external os, the TV probe was rotated to the transverse plane, maintaining alignment with the CSD measurement plane. The following transverse measurements were then recorded: distance D (width of the transverse section of the CSD) and distance E (depth of the transverse section of the CSD). A structured data collection form was used to assess changes in the quality of life among patients with isthmocele. Evaluated parameters included age, height, weight, menstrual patterns, pelvic pain, infertility, dyspareunia, and dysmenorrhea. Symptom assessment and isthmocele measurement were repeated at 3 and 6 months postoperatively. Types of AUB, including prolonged bleeding, postmenstrual spotting or bleeding, hypermenorrhea, and intermenstrual bleeding, were evaluated. Symptom duration and severity were recorded preoperatively and at 3 and 6 months postoperatively. The Pictorial Blood Loss Assessment Chart (PBAC), translated into Turkish, was used to quantify menstrual bleeding [ 12 ]. Patients who used sanitary pads were evaluated on the basis of the extent of bleeding, with a cutoff value of 100 points. This threshold demonstrated 86% sensitivity and 81% specificity for diagnosing menorrhagia. Pain intensity was assessed preoperatively and postoperatively using the Wong‐Baker Facial Pain Rating Scale and the Numerical Pain Scale. Evaluations were conducted before, immediately after surgery, and again at 3 and 6 months postoperatively [ 13 ]. Samples were collected from 40 patients. Endometrial lavage samples were cultured on Eosin‐methylene Blue and Sabouraud dextrose agar and incubated overnight for 48 h. Cultures exhibiting growth were identified using matrix‐assisted laser desorption/ionization‐time of flight mass spectrometry (BioMérieux, France). Antibiograms were performed with the VITEK 2 system (Biomérieux, France). The primary endpoint was next‐generation sequencing (NGS) results obtained by polymerase chain reaction (PCR). As no previous studies had employed identical methods, the sample size was estimated assuming an effect size of 0.8, type I error ( α ) = 0.05, and power (1 −  β ) = 0.90. Power analysis with G*Power version 3.1.9.7 (Franz Faul, Universitat Kiel, Germany) indicated a minimum of 19 patients. Due to budget constraints, microbiome analysis was performed for 22 patients before and 3 months postoperatively, yielding 44 endometrial lavage samples. Patients were randomly chosen using a simple lottery method. Other variables were analyzed using available patient data. All hysteroscopic procedures were performed after the end of active menstrual bleeding; between Days 6 and 12 of the cycle. The first endometrial samples were taken in the operating room prior to hysteroscopic surgery. Under general anesthesia, patients were positioned in the lithotomy position. After speculum insertion, the cervix and vagina were irrigated with normal saline. Once the irrigation fluid and cervical mucus were removed, a sterile intrauterine insemination (IUI) catheter with a guided wire (Medbar Medical Devices, Gaziemir, Türkiye) was inserted into the endometrial cavity. Upon removing the guided wire, the cavity was irrigated with 3 mL of sterile saline via the IUI catheter, and the aspirated fluid was placed in a sterile tube for microbiological analysis. The samples were delivered to the microbiology laboratory within 1 h. Approximately one‐third of the endometrial aspirate was cultured on blood agar and Sabouraud medium, while the remainder was promptly frozen and stored at −80°C. All procedures were performed by the same physician (ASC) using sterile, single‐use guides and disposable instruments, with no reusable sampling equipment. These measures were implemented specifically to minimize sampling differences that may depend on the person at the pre‐analytical stage. Hysteroscopic isthmoplasty was performed by the same physician (ASC). The cervix was dilated to a No. 8.5 Hegar plug size. An 8 mm unipolar resectoscope (Karl‐Storz, Germany) with a telescope and cable connections was inserted into the endometrial cavity. After endometrial cavity exploration, the resectoscope and camera were rotated 180°, and the resectoscope tip was pulled toward the cervical canal so that it rested on the isthmocele sac in the anterior isthmus. The lower edge of the isthmocele sac was resected toward the cervical external os with a loop electrode until the isthmus ceiling was reached. Superficial veins and erythematous areas at the base of the isthmocele sac were coagulated with the electrode tip using spray coagulation mode. In the electrosurgery generator (Petas, Petkod600, Türkiye), 70 W was generally used for cutting and 35 W for coagulation (Video). Hysteroscopic examination of all study patients revealed endometritis findings such as endometrial micropolyposis, strawberry spots, and diffuse hyperemia. Patients were discharged with an antibiotic regimen of 100‐mg doxycycline twice daily (Tetradox capsules, Actavis Pharmaceuticals, Istanbul, Türkiye) and 500‐mg metronidazole twice daily for 14 days (Flagyl 500‐mg film‐coated tablets, Sanofi Pharmaceuticals Inc., Kirklareli, Türkiye). Postoperative endometrial microbiota samples were obtained at 3 months in an outpatient setting using the same procedure as preoperative sampling (see Section  2.5 ). Endometrial sampling was performed after the end of active menstrual bleeding between Days 6 and 12 of the cycle. Under transabdominal USG guidance, an IUI catheter was inserted into the endometrial cavity. If necessary, the cervix was stabilized with a single thread. After removal of the guide wire, the cavity was irrigated with 3 mL of sterile saline through the IUI catheter. The aspirated fluid was collected in a sterile tube and transported to the microbiology laboratory within 1 h for evaluation. Genomic DNA was isolated from endometrial culture samples using the QIAamp Fast DNA Stool Mini kit (QIAGEN, Germany) following the manufacturer's protocol. Library preparation for sequencing involved PCR amplification using Proofreading DNA Polymerase 2× Reaction Mix and 200 nM primers. The thermal cycling program included an initial denaturation at 95°C for 3 min, followed by 25 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 1.5 min, with a final extension at 72°C for 5 min. Amplified nucleic acids were quantified fluorometrically (Qubit 3, ThermoFisher, Waltham, MA, USA), analyzed via agarose gel electrophoresis (approximately 1450 bp), and purified using the PCR Product Purification Kit. Metagenomic analysis followed previously described workflows [ 14 ]. Amplicon libraries targeted a region of approximately 1400 bp within the 16S rRNA gene [ 15 , 16 ]. Oxford Nanopore Technologies barcode DNA sequences of the resulting library were added to the 5′ end of the target‐specific primer pairs. The 16S rRNA‐specific primer‐connector sequences were: 5′‐TTTCTGTTGGTGCTGATATATTGC‐AGRGTTTGATYHTGGCTCAG‐3′ for the forward primer and 5′‐ACTTGCCTGTCGCTCGCTCTATCTTC‐TACCTTGTTAYGACTT‐3′ for the reverse primer. The ligation sequencing kit 1D (SQK‐LSK108; Oxford Nanopore Technologies) and amplicon library were loaded onto a MinION instrument (Oxford Nanopore Technologies, UK) for amplicon library construction. The device was loaded with a 45 μL mixture of barcoded DNA, containing 1–1.5 μg of DNA, along with 5 μL lambda phage DNA as a positive control. DNA end repair and dA addition were performed using a NEBNext End Repair/dA‐Tailing Module Kit (New England Biolabs). Purification was carried out with an Agencourt AMPure XP bead kit (Beckman Coulter). For adapter ligation, 0.2 pmol of ligated DNA was mixed with 50 μL of Blunt/TA ligase master mix (New England Biolabs) and 20 μL of adapter mix and incubated for 10 min at room temperature (22°C–23°C). Final purification for DNA library formation was achieved using Adapter Bead Binding buffer (SQK‐LSK108 kit) and 0.5× Agencourt AMPure XP beads (Beckman Coulter). The sequencing mixture, consisting of 14 μL of DNA library, was combined with 25.5 μL of loading beads and 35.5 μL of running buffer mix. The R9.4 flow cell designated for use was primed accordingly, and the sequencing mixture was transferred to the sample loading area of the flow cell. A 2‐day (R9.4) sequencing protocol was implemented using the MinION control software, MinKNOW version 0.46.1.9 (R9.4). Read data were processed using workflow 1.2.2 rev 1.5 and Metrichor agent (version 0.16.37960) software. Negative controls were incorporated throughout the molecular workflow, including extraction blanks and PCR blanks. These controls did not demonstrate significant amplification and yielded either no reads or read counts below analytical thresholds; therefore, they were excluded from the downstream analysis. Statistical analyses were performed using Statistical Package for the Social Sciences (version 22; IBM Corp., Armonk, NY, USA). Descriptive statistics are presented as counts and percentages for categorical data, and as mean ± standard deviation or median (range) for continuous data. Normality was assessed using the Kolmogorov–Smirnov test. Repeated measures of continuous variables were analyzed using the Friedman test, with Bonferroni‐adjusted Wilcoxon test for pairwise comparisons. Repeated categorical measures were analyzed with the Cochran Q test, with Bonferroni‐adjusted McNemar tests for pairwise comparisons. Statistical significance was set at p  < 0.05. Following sequencing, raw data in FAST5 format were converted to FASTQ format using Guppy v3.1.5 software for base‐calling and demultiplexing. Barcode and adapter sequences were removed using Porechop v0.2.3. Universal primers and tags were trimmed by removing 45 bases from both ends of the sequence. Following the cleaning process, reads with lengths between 1350 and 1550 bp were filtered, while any remaining reads were excluded from the analysis. To ensure data quality, chimeric sequences were removed, followed by sequence alignment and similarity matrix calculations. Reads with > 99% similarity were clustered into operational taxonomic units (OTUs) and taxonomically annotated by comparisons with the RDP 16S rRNA database. Statistical analyses were conducted at the genus level based on OTU associations. Graphs were generated using Minitab and R programs, incorporating the quantitative values of matched OTUs and sample metadata [ 17 ]. For further analysis, cleaned reads were processed through a Python‐based workflow. Taxonomic classification was performed with a locally adapted sequencing analysis tailored for 16S rRNA, leveraging the National Center for Biotechnology Information database to identify high‐probability organisms [ 18 ]. The similarity rates and quality controls of these matches, along with assignments demonstrating significant similarity, were recorded as taxon information. The dataset was systematically screened for the taxa commonly reported as reagent‐associated contaminants in low‐biomass 16S rRNA sequencing studies (e.g., Ralstonia, Pseudomonas, Sphingomonas, and Methylobacterium). These genera were not detected at biologically meaningful relative abundances in the studied samples. After finalizing the taxonomic classification, post‐analysis relative abundance values were computed for the remaining data. The report presented genome and relative abundance information of the organisms, along with phylogenetic data analyses performed using QIIME and BIOM data file formats. Additional outputs included alpha and beta diversity data analyses, linear discriminant analysis effect size (LEfSe), Krona plots, similarity matrix heatmaps, and statistical test data results ( p ‐values). Alpha diversity was assessed using Shannon, Simpson, and Chao1 indices, while beta diversity was evaluated through weighted UniFrac and Bray–Curtis distance, complemented by principal coordinate analysis (PCoA). Further analyses incorporated LDA‐LEfSe and cladograms [ 19 , 20 , 21 ]. Alpha diversity was analyzed using the Kruskal–Wallis test, whereas beta diversity was evaluated using Bray–Curtis PCoA, Jaccard, and PERMANOVA analyses.

Video S1: Hysteroscopic isthmoplasty.