The primary aim of this prospective cohort study is to evaluate the association between perioperative inflammatory markers (CRP, IL-6, and TNF-$\alpha$) and the incidence and severity of IUAs after hysteroscopic submucosal myomectomy. We further seek to explore the influence of key clinical variables, including myoma classification and surgical parameters, on biomarker profiles and adhesion outcomes. By integrating biomarker surveillance with clinical and hysteroscopic data, our study seeks to advance risk stratification and inform future preventive strategies in gynecologic surgery.

Women of reproductive age scheduled for hysteroscopic submucosal myomectomy were assessed for eligibility. Participants were recruited through the gynecologic outpatient clinic and underwent preoperative evaluation including ultrasonography, hormonal profile testing, and infection screening. A total of 200 participants were enrolled with consideration of a 10% –15% attrition rate to maintain adequate statistical power for final analysis.

The required sample size was determined a priori based on published post-hysteroscopic recurrence rates of IUAs. In a randomized controlled trial of 98 women undergoing hysteroscopic transcervical resection of adhesions, Qu and Zhou reported a 25%–40% re-adhesion rate at three months despite the use of barrier therapy [18]. Based on this rate, we calculated that a minimum of 170 participants would provide 80% power at $\alpha$ = 0.05 to detect an odds ratio (OR) of at least 1.5 for biomarker-defined risk groups. To compensate for a projected 10%–15% loss to follow-up, we increased the enrollment target to 200 women to maintain adequate power for the final analysis. Post-hoc, the achieved sample (n = 180 with 45 mild-to-severity IUA events) supported an event-per-variable ratio $\ge$10 for multivariable logistic regression, ensuring adequate model stability.

The primary outcome was the incidence of IUAs, diagnosed by follow-up hysteroscopy at 3 months postoperatively. Secondary outcomes included IUA severity, changes in inflammatory biomarker levels, and their associations with clinical and surgical variables.

$\bullet$ Women aged 18–45 years with submucosal myomas classified types 0–II (according to the International Federation of Gynecology and Obstetrics [FIGO] system) scheduled for hysteroscopic removal were eligible.

$\bullet$ Menstrual phase at surgery was determined primarily from the last menstrual period (LMP) and corroborated with transvaginal ultrasound measurement of endometrial thickness. Serum estradiol and progesterone assays were performed only in cases with irregular cycles or an indeterminate phase by LMP and ultrasound. When surgery was performed outside a clearly defined follicular or luteal window, the phase classification was based on the best available clinical and imaging data.

$\bullet$ Previous history of IUAs or uterine surgeries such as myomectomy, cesarean section, or curettage within the past year.

$\bullet$ Presence of endometrial tuberculosis or chronic endometritis.

$\bullet$ Active pelvic infection or untreated sexually transmitted disease.

$\bullet$ Known or clinically suspected endometriosis or chronic systemic inflammatory conditions (such as autoimmune disease, chronic inflammatory bowel disease, rheumatoid arthritis, or other conditions requiring long-term immunomodulatory therapy). All participants underwent preoperative evaluation, including medical history, ultrasound, and laboratory screening, to minimize unrecognized chronic inflammatory conditions. Routine histological confirmation was not performed.

$\bullet$ Known coagulation or bleeding disorders.

$\bullet$ Immunosuppressive medication use.

$\bullet$ Contraindications to hysteroscopic surgery or anesthesia.

All hysteroscopic submucosal myomectomies were performed using a standardized bipolar resectoscope system (Karl Storz, Tuttlingen, Germany) under general anesthesia induced with propofol (AstraZeneca, Shanghai, China). Intrauterine distension was achieved with isotonic saline, and careful resection of the myoma was performed to minimize injury to the surrounding endometrium. Operative videos were archived for documentation and teaching purposes.

To allow unbiased assessment of adhesion formation, no postoperative anti-adhesion barriers or estrogen therapy were administered. All patients received routine postoperative monitoring for pain, bleeding, and infection and were given standard discharge instructions. Follow-up hysteroscopy was scheduled 12 weeks postoperatively to evaluate adhesion formation, and any adverse events were recorded prospectively.

To standardize postoperative care and avoid confounding effects, no prophylactic anti-adhesion barrier or postoperative estrogen therapy was administered to any participant. In our institution, such measures are not routinely used following hysteroscopic myomectomy but are reserved for high-risk or recurrent cases.

Peripheral blood samples were collected at three time points: 24 hours preoperatively, 24–48 hours postoperatively, and at 3 months post-surgery. At the 3-month follow-up, participants completed a short clinical questionnaire and were screened for recent febrile illnesses, infections, or new inflammatory conditions. If an acute illness was present at the time of the scheduled blood draw, sampling was postponed until recovery, whenever feasible. Primary inflammatory markers assessed included CRP, IL-6, and TNF-$\alpha$, while secondary markers included IL-1$\beta$, erythrocyte sedimentation rate (ESR), and white blood cell (WBC) count (BC-6800Plus, Shenzhen Mindray Bio-Medical Electronics Co., Ltd., Shenzhen, Guangdong, China) in the hospital’s central clinical laboratory). Serum cytokines were measured using enzyme-linked immunosorbent assay (ELISA) (Cat. No. DCRP00; R&D Systems, Inc., Minneapolis, MN, USA), and complete blood counts were analyzed using an automated hematology analyzer. All laboratory personnel were blinded to the surgical and hysteroscopy findings.

Follow-up hysteroscopy was scheduled for all participants at 12 weeks following the index procedure. IUAs were evaluated and graded based on the European Society for Gynecological Endoscopy (ESGE) classification system, which stratifies adhesions into grades I–V based on their density, extent, and anatomic involvement. Evaluations were conducted by two independent reviewers blinded to inflammatory marker data. Discrepancies in scoring were resolved through consensus review.

Patient-level data on potential confounders were recorded, including age, body mass index (BMI), parity, smoking history, hormonal phase at the time of surgery (follicular vs. luteal), and presence of anemia. Menstrual phase was determined using the date of LMP and ultrasound endometrial thickness, and was corroborated by mid-cycle serum estradiol and progesterone measurments when needed.

Missing data were minimal: CRP (1.7%), IL-6 (2.2%), TNF-$\alpha$ (1.7%), and baseline covariates 3% each. No variable exceeded 5% missing data. Analyses were conducted using a complete-case approach. Sensitivity analyses using available-case data produced results virtually identical to those of the complete-case analysis, indicating that missing data did not materially bias the primary findings.

All statistical analyses were performed using SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to summarize baseline characteristics. Continuous variables were assessed for normality using the Shapiro-Wilk test and compared with Student’s t-test or the Mann-Whitney U test, as appropriate. Categorical variables were analyzed using Pearson’s chi-square test when all expected cell counts were $\ge$5. Yates’ continuity correction was applied to 2 $\times$ 2 tables with small but nonzero expected counts, and Fisher’s exact test was used when any expected cell count was 5. Spearman’s rank correlation examined relationships between inflammatory markers and adhesion severity.

Of 220 women screened, 20 were excluded (n = 12 did not meet inclusion criteria; n = 5 refused participation; n = 3 other reasons). 200 participants were enrolled. Postoperative attrition included 15 lost to follow-up and 5 withdrawals, leaving 180 for final analysis (Fig. 1).

Fig. 1.

STROBE flow diagram. STROBE, Strengthening the Reporting of Observational studies in Epidemiology.

Flowchart of participant enrollment, attrition, and final analysis cohort. Of 220 women screened, 200 met eligibility criteria and underwent hysteroscopic submucosal myomectomy. 20 participants were excluded due to failure to meet inclusion criteria (n = 12), refusal to participate (n = 5), or other reasons (n = 3). Postoperative attrition included 15 participants lost to follow-up and 5 withdrawals, resulting in 180 participants for final analysis.

Baseline demographic and clinical characteristics are summarized in Table 1a. The mean age of participants was 34.5 $\pm$ 5.2 years, and 62% were nulliparous. Most myomas were FIGO type I (68%), followed by type 0 (22%) and type II (10%). Table 1b shows comparison of two groups by IUA status. Baseline demographic and clinical characteristics were similar between groups, except for a higher proportion of FIGO type II myomas among patients who developed IUAs (17.7% vs. 7.4%, p = 0.08).

Out of 200 enrolled patients, 180 completed the full follow-up, including follow-up hysteroscopy at 3 months postoperatively. Based on ESGE grading, 83 patients (46.10%) exhibited no IUAs, while 46 (25.62%) had mild IUAs, 34 (18.92%) had moderate IUAs, and 17 (9.40%) had severe IUAs. The distribution of IUA severity is summarized in Table 2.

Inflammatory marker levels varied significantly across the different adhesion severity groups. The average levels and SDs for CRP, IL-6, and TNF-$\alpha$ across all groups are presented in Table 3a.

$\bullet$ CRP levels were elevated in patients with IUAs compared to those without, with a stepwise increase from mild to severe grades (Fig. 2).

$\bullet$ IL-6 showed the most pronounced gradient, especially between the moderate-to-severe IUA groups compared to those without (Fig. 3).

$\bullet$ TNF-$\alpha$ levels were significantly higher in the severe IUA group, distinguishing it from all others (Fig. 4).

$\bullet$ Mean CRP, IL-6, and TNF-$\alpha$ levels were significantly higher at both 48 hours and 3 months postoperatively in patients who developed IUAs compared to those who did not (Table 3b), whereas baseline levels did not differ significantly between groups.

Fig. 2.

Boxplot showing CRP levels increase progressively with IUA severity.

Fig. 3.

IL-6 levels were significantly elevated in moderate and severe IUA groups.

Fig. 4.

TNF-$\alpha$ levels are notably higher in the severe group, indicating fibrosis-related inflammation.

Spearman’s correlation analysis demonstrated a moderate positive correlation between inflammatory marker levels and IUA severity scores: IL-6 exhibited the strongest correlation ($\rho$ = 0.51), followed by CRP ($\rho$ = 0.43), and TNF-$\alpha$ ($\rho$ = 0.39), all with p 0.01.

These results are detailed in Table 4 and visually summarized via the heatmap in Fig. 5.

Fig. 5.

Heatmap displaying Spearman’s correlation between inflammatory biomarkers and IUA severity. Strongest correlation observed between IL-6 and adhesion grade.

To assess predictive power, a logistic regression model was developed to identify patients at risk of developing moderate-to-severe IUAs based on perioperative biomarker levels. IL-6 emerged as the strongest independent predictor:

$\bullet$ IL-6: Coefficient = 0.84, p 0.001

$\bullet$ CRP: Coefficient = 0.63, p = 0.004

$\bullet$ TNF-$\alpha$: Coefficient = 0.42, p = 0.019

These findings are presented in Table 5a, with the ROC curves for IL-6, CRP, and TNF-$\alpha$, shown in Fig. 6, indicating excellent discriminatory performance.

Fig. 6.

Receiver operating characteristic (ROC) curves for prediction of IUA using inflammatory markers. (A) CRP. (B) IL-6. (C) TNF-$\alpha$.

In multivariate analysis, both IL-6 and CRP remained independent predictors of IUA development after adjustment for surgical complexity, FIGO myoma type, BMI, parity, smoking history, menstrual phase, and anemia. The risk associated with elevated IL-6 levels was especially pronounced in women with FIGO type II myomas.

Table 5b presents both unadjusted and adjusted ORs for all predictors. In univariable models, IL-6 and CRP showed strong crude associations with IUA development, which remained significant after multivariable adjustment. In contrast, TNF-$\alpha$ showed a weaker crude association that was no longer significant after adjustment. FIGO type II myomas remained a significant predictor of adhesion formation in both unadjusted and adjusted models.

Histograms were created to visualize the overall distribution of inflammatory biomarker levels in the study population. CRP levels exhibited a mildly right-skewed distribution, with most values clustering between 5.5–6.0 mg/L and a longer tail toward higher concentrations (Fig. 7). The mean exceeding the median is consistent with the right-skewed distribution and aligns with the summary statistics. IL-6 levels showed a normal distribution with a higher tail in severe cases (Fig. 8). Histogram demonstrates a slightly right-skewed distribution of IL-6, with most values between 14 and 20 pg/mL, and a tail extending beyond 21 pg/mL. TNF-$\alpha$ showed a moderately skewed distribution with high variability (Fig. 9). Histogram reveals a moderately right-skewed distribution, with TNF-$\alpha$ values concentrated between 9.0 and 11.5 pg/mL and a few elevated cases above 12 pg/mL. These patterns support the findings that higher inflammatory states are present in more severe IUA cases.

Fig. 7.

Histogram of CRP levels (n = 180). The distribution is mildly right-skewed (mean $>$ median), with most values clustered around 5.5–6.0 mg/L. Shapiro-Wilk test p 0.05, indicating mild non-normality.

Fig. 8.

Histogram of IL-6 levels (n = 180). The distribution is slightly right-skewed with a tail extending beyond 21 pg/mL. Shapiro-Wilk test p 0.05, indicating mild non-normality.

Fig. 9.

Histogram of TNF-$\alpha$ levels (n = 180). The distribution is moderate right-skewed, with values concentrated between 9.0 and 11.5 pg/mL. Shapiro-Wilk test p $>$ 0.05, approximating normality.

Table 6a summarizes the AUC values and significance levels of inflammatory markers for predicting moderate-to-severe IUAs. IL-6 showed the highest predictive performance (AUC = 0.82), showing strong discrimination for moderate-to-severe IUAs. Both CRP and TNF-$\alpha$ also demonstrated acceptable discriminatory ability with statistically significant p-values. ROC analyses, including AUC (95% CI) and Youden-based operating characteristics (sensitivity, specificity, LR⁺, LR^–, PPV, NPV with 95% CIs) are summarized in Tables 6a,6b,6c. Table 6b shows the optimal cut-off values for IL-6, CRP, and TNF-$\alpha$ derived from the Youden index. IL-6 $\ge$18.0 pg/mL provided the highest accuracy for moderate-to-severe IUAs (sensitivity 82%, specificity 73%, Youden index 0.55), followed by CRP $\ge$6.3 mg/L, and TNF-$\alpha$ $\ge$10.5 pg/mL. ROC analyses yielded the following optimal cut-off values (Table 6b), with corresponding sensitivity, specificity, LR⁺, LR^–, PPV, and NPV (Table 6c). IL-6 $\ge$18.0 pg/mL demonstrated the highest discriminative performance (AUC = 0.82, 95% CI 0.74–0.90; sensitivity 82%, specificity 73%, LR⁺ 3.04, LR^– 0.25, PPV 68%, NPV 85%). CRP $\ge$6.3 mg/L and TNF-$\alpha$ $\ge$10.5 pg/mL also showed clinically relevant predictive ability with acceptable sensitivity and specificity.

Table 7 summarizes surgical outcomes, including operative time and intraoperative and postoperative complications, comparing the IUA group and non-IUA group. Operative time was significantly longer in patients who developed IUAs (p = 0.02), possibly due to more complex resection or endometrial trauma. Fluid deficit was also greater in the IUA group (p = 0.04), suggesting higher intrauterine irrigation volume needs. Complication rates (bleeding, uterine perforation, and postoperative pain) were higher in the IUA group, though not all reached statistical significance.

Fig. 10 shows a line graph comparing trends of CRP, IL-6, and TNF-$\alpha$ levels across three time points (preoperative, 48 hours, and 3 months) between adhesion and non-adhesion patient groups. All three markers peaked at 48 hours in both groups, with consistently higher values in the adhesion group across all time points.

Fig. 10.

Trends of CRP, IL-6, and TNF-$\alpha$ levels across three time points (preoperative, 48 h postoperative, 3 months postoperative) comparing adhesion versus non-adhesion groups. All three markers peaked at 48 hours with persistently higher values in the adhesion group.

Next, we developed a Venn diagram that shows the overlap of the three inflammatory markers (IL-6, CRP, and TNF-$\alpha$) by number of IUA patients. The largest overlaps are between IL-6 and TNF-$\alpha$ (6 individuals) and IL-6 and CRP (5 individuals), while only 4 individuals tested positive for all three markers (Fig. 11).

Fig. 11.

Overlap of elevated inflammatory biomarkers in IUA patients. Venn diagram with the number of patients with elevated IL-6, CRP, and TNF-$\alpha$, including all single, double, and triple marker combinations. Each region is labeled with the corresponding biomarker(s) and patient count.

Beyond their potential role in predicting IUA risk after hysteroscopic myomectomy, systemic inflammatory indices are increasingly being applied across gynecological oncology and reproductive medicine. For example, recent work has demonstrated that combining systemic inflammatory indices with tumor markers such as CA-125 (SIR-125 and SIRI-125) significantly improves the preoperative differentiation of borderline ovarian tumors from early-stage ovarian carcinoma, achieving AUCs above 0.80 [22]. Similarly, the SIR-En index, which integrates systemic inflammation measures with endometrial thickness, distinguished endometrial carcinoma from atypical hyperplasia in postmenopausal women presenting with abnormal uterine bleeding, demonstrating high specificity [23]. Collectively, these findings highlight the broader applicability of inflammation-based biomarkers in gynecologic settings, including ovarian and endometrial disease, and support our conclusion that perioperative monitoring of inflammatory markers in intrauterine surgery may represent a clinically relevant and cost-effective extension of this diagnostic paradigm.

This study has several limitations. First, although our study was prospective and included standardized biomarker assessment, detailed fibroid metrics beyond FIGO classification or operative energy settings were not captured, which may represent important modifiers of adhesion risk. Second, chronic endometritis and other baseline inflammatory conditions were not assessed histologically, limiting our ability to distinguish pre-existing from procedure-induced inflammation. Third, our follow-up was limited to 3 months after surgery, which may underestimate late or progressive adhesion formation; longer surveillance (e.g., at 6 or 12 months) could capture delayed cases and refine predictive cut-offs. Fourth, while we recalculated all p-values using Fisher’s exact test or Yates’ continuity correction for low-cell-count categorical variables, some residual risk of type-I error remains due to multiple comparisons. Finally, as an observational study, our findings establish an association but not causality; randomized controlled trials are needed to determine whether modifying perioperative inflammation can reduce adhesion rates. Our study focused on adhesion incidence and severity, without directly assessing fertility outcomes such as conception rates, pregnancy maintenance, or live birth. Because IUAs primarily affect reproductive health, longer-term follow-up capturing fertility and obstetric outcomes is essential to establish the clinical utility of perioperative biomarker monitoring. Another limitation is that we did not conduct histological examinations to definitively rule out endometriosis or chronic endometritis. Although women with known or clinically suspected conditions were excluded based on history, imaging, and laboratory testing, subclinical cases may have been present and could act as unmeasured confounders. Future studies incorporating histopathologic screening may help clarify the association between perioperative biomarkers and adhesion risk.

This study underscores the potential of inflammation-guided, personalized risk stratification for women undergoing intrauterine surgery. Serial perioperative measurement of IL-6 and CRP, combined with established clinical risk factors, can identify patients at increased risk of IUAs who may benefit from targeted preventive strategies such as early follow-up hysteroscopy, application of anti-adhesion barriers, or adjunctive hormonal therapy. Although the cost-effectiveness and operational feasibility of such protocols require further evaluation, our findings indicate that perioperative monitoring of IL-6, CRP, and TNF-$\alpha$ could be integrated into routine postoperative care pathways after hysteroscopic submucosal myomectomy. Early recognition of persistently elevated inflammatory markers may enable timely, individualized interventions to reduce adhesion formation and mitigate subsequent infertility risk. Given the relative affordability and widespread availability of these biomarkers, this approach has the potential to improve reproductive outcomes and optimize resource allocation, particularly in settings where fertility preservation is a primary concern. From a clinical perspective, the primary significance of IUAs lies in their impact on fertility and pregnancy outcomes. Although reproductive endpoints were not directly assessed in this cohort, elevated perioperative IL-6 and CRP levels may also indicate an increased risk of subfertility, miscarriage, or adverse obstetric outcomes. Future longitudinal studies incorporating fertility and pregnancy outcomes are warranted to validate these associations and establish the full clinical utility of perioperative biomarker surveillance.