Perioperative inflammatory response as a determinant of post-myomectomy intrauterine adhesion formation

Abstract

Background

Intrauterine adhesions (IUAs), or Asherman’s syndrome, are a common complication of intrauterine surgery (e.g. hysteroscopic myomectomy) that can lead to menstrual disturbances and infertility. Excessive postoperative inflammation is thought to drive adhesion formation, but the predictive value of specific inflammatory biomarkers remains unclear.

Methods

We performed a prospective observational study of 150 women (age 19–45) undergoing hysteroscopic submucosal fibroid resection. Serum C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α) were measured at three timepoints: preoperatively, 48 h postoperatively, and 3 months postoperatively. IUAs were assessed at 12 weeks via second-look hysteroscopy and graded by American Fertility Society criteria.

Results

Of 135 patients who completed follow-up, 40 (30.8%) developed IUAs (20 mild, 12 moderate, 8 severe). Baseline characteristics were similar between those who developed IUAs and those who did not, except that FIGO type II fibroids were more frequent in the IUA group (20.0% vs. 8.4%, p = 0.04). IL-6 and CRP levels rose markedly after surgery in patients who developed adhesions, whereas increases in the no-IUA group were modest. By 48 h postoperatively, mean IL-6 was ~ 7 pg/mL higher in the IUA group than in no-IUA (23.0 vs. 15.8 pg/mL, p < 0.001), and CRP was ~ 2.4 mg/L higher (8.5 vs. 6.1 mg/L, p < 0.001). These differences persisted at 3 months (IL-6 18.2 vs. 14.2 pg/mL; CRP 6.4 vs. 5.1 mg/L; both p < 0.001), indicating a prolonged inflammatory response. TNF-α rose modestly postoperatively in both groups and was significantly higher in the IUA group at 48 h (13.0 vs. 10.0 pg/mL, p < 0.001) and 3 months (11.5 vs. 9.4 pg/mL, p = 0.01). IL-6 showed the strongest correlation with adhesion severity (Spearman ρ ≈ 0.5, p < 0.001) and was an independent predictor of moderate-to-severe IUA risk (adjusted odds ratio [OR] 1.25 per 1 pg/mL, 95% confidence interval [CI] 1.11–1.42). CRP was also an independent predictor (OR 1.32 per 1 mg/L, 95% CI 1.08–1.61, p = 0.006), whereas TNF-α did not retain significance in multivariate analysis. In a model adjusting for surgical factors (fibroid type, operative time), IL-6 and CRP remained significant predictors (adjusted OR 1.22, 95% CI 1.09–1.39 and OR 1.29, 95% CI 1.06–1.57, respectively). Women with type II fibroids had higher odds of developing moderate/severe adhesions (OR ~ 2.8, 95% CI 1.1–7.3). ROC analysis demonstrated that IL-6 had the highest discriminative ability for clinically significant IUAs (area under curve [AUC] ≈ 0.80), outperforming CRP (AUC ≈ 0.72) and TNF-α (AUC ≈ 0.65). At an optimal 48-hour IL-6 cutoff of ~ 20 pg/mL, sensitivity for predicting moderate/severe adhesions was ~ 81% and specificity ~ 74%, while a CRP cutoff of ~ 7 mg/L yielded ~ 75% sensitivity and 65% specificity.

Conclusions

Perioperative elevations in IL-6 and CRP are strongly associated with IUA formation after hysteroscopic myomectomy, and dynamic changes in these markers offer predictive value. Early post-surgical IL-6 and CRP measurements may help identify high-risk patients and support risk-stratified postoperative surveillance or prophylactic interventions.

Introduction

Intrauterine adhesions (IUAs), also known as Asherman’s syndrome, are bands of fibrous scar tissue that partially or completely obliterate the uterine cavity. They typically result from trauma to the endometrial basalis layer after procedures such as dilation and curettage or myomectomy, which disrupt normal endometrial regeneration [1–3]. IUAs are a significant cause of acquired infertility and menstrual disturbances, leading to symptoms ranging from hypomenorrhea or amenorrhea to recurrent pregnancy loss and obstetric complications [4]. Hysteroscopic submucosal fibroid resection is frequently implicated as a trigger for IUAs, with postoperative adhesion rates reported as high as 20–35% in some series. Even minor intrauterine fibrosis can compromise uterine function, so preventing adhesions remains a critical challenge [5].

Aberrant postoperative inflammation is recognized as a key driver in adhesion pathogenesis. Injury to the uterine lining from surgery triggers an intense wound-healing response characterized by immune cell infiltration and release of pro-inflammatory cytokines. Elevated levels of IL-6 and TNF-α can promote fibroblast activation and excessive collagen deposition, contributing directly to adhesion formation [6]. Transforming growth factor-beta (TGF-β) is also upregulated in this context and mediates downstream fibrogenic pathways [7]. Clinical studies support a mechanistic link between persistent inflammation and endometrial fibrosis: women with established IUAs have been observed to exhibit higher circulating IL-6, TNF-α, and acute-phase reactants like CRP compared to those without adhesions [8]. These findings suggest that an exaggerated or prolonged inflammatory response after endometrial injury may predispose to pathologic scar formation in the uterus.

Despite this understanding, it remains unclear which inflammatory biomarkers provide the most reliable early indication of adhesion risk, and how the dynamic changes in these markers relate to IUA development and severity. Early identification of patients at high risk for IUAs could enable targeted preventive measures (such as anti-adhesion barriers, hormonal therapy, or early postoperative interventions) and closer monitoring. Currently, there is no consensus protocol for predicting adhesion formation after intrauterine surgery [9–11]. Most prior investigations have been retrospective or cross-sectional, often focusing on single time-point measurements or on adhesion recurrence after treatment rather than the prospective monitoring of new adhesion formation [12]. Furthermore, the influence of surgical factors (e.g. fibroid size/location, operative duration, extent of endometrial injury) on the postoperative inflammatory milieu and subsequent adhesion risk is not fully delineated [13]. Given that inflammatory markers are easily measurable and could serve as real-time indicators of tissue healing, a better understanding of their predictive capacity is needed to inform clinical practice [14].

Although numerous inflammatory mediators are implicated in fibrotic remodeling of the endometrium, including transforming growth factor-β (TGF-β), interleukin-17, and various matrix-regulating factors, many of these molecules function primarily within local tissue microenvironments and are not routinely measurable in clinical practice. In contrast, IL-6, C-reactive protein (CRP), and tumor necrosis factor-α (TNF-α) represent key upstream components of the systemic inflammatory response that can be reliably quantified in peripheral blood. IL-6 is a central cytokine in the acute inflammatory cascade and has been strongly implicated in fibrosis through stimulation of fibroblast activation and extracellular matrix deposition. CRP serves as a sensitive downstream acute-phase reactant reflecting the overall magnitude of tissue injury and inflammatory activation, while TNF-α plays an important role in early inflammatory signaling and immune cell recruitment following tissue damage. Because these biomarkers collectively capture different phases of the postoperative inflammatory response and are readily accessible in routine clinical settings, they represent practical candidates for evaluating inflammation-driven adhesion formation after intrauterine surgery. However, prospective clinical evidence examining the dynamic perioperative trajectories of these biomarkers in relation to new intrauterine adhesion development remains limited.

We aimed to prospectively evaluate whether perioperative serum levels of CRP, IL-6, and TNF-α can predict intrauterine adhesion development following hysteroscopic myomectomy. We hypothesized that patients with sustained or higher elevations in these inflammatory biomarkers in the immediate postoperative period and at follow-up would have a higher incidence and greater severity of IUAs.

Materials and methods

Study design and setting

This was a single-center prospective observational study conducted at Northern Theater Command General Hospital, Shenyang, China, a tertiary referral center for gynecologic surgery. Ethical approval was obtained from the Institutional Review Board (Approval No. DG/NTCGH/45, April 24, 2023), and all participants provided written informed consent. The study adhered to the Declaration of Helsinki and relevant national guidelines.

Patient recruitment

Between June 2023 and June 2025, we screened 170 women of reproductive age who were scheduled for hysteroscopic myomectomy for possible inclusion. After applying inclusion and exclusion criteria, 150 eligible women were enrolled. We aimed to have at least 130 patients complete follow-up to ensure adequate statistical power, anticipating 10–15% attrition. A priori power calculations (α = 0.05, two-sided) indicated that a sample of 128 would provide > 90% power to detect a moderate difference (Cohen’s d = 0.5) in 48-hour IL-6 levels between patients who did vs. did not develop adhesions, assuming an IUA incidence of 25–35%. This sample size also allowed reasonable precision in estimating ROC AUCs in the 0.7–0.8 range and in fitting multivariable models with key covariates.

Inclusion and exclusion criteria

We included women 18–45 years old; presence of one or more submucosal uterine fibroids (FIGO types 0, I, or II) for which hysteroscopic resection was indicated; normal uterine cavity shape aside from fibroids (no congenital Mullerian anomalies); no pre-existing intrauterine pathology (no baseline adhesions, polyps, or endometrial hyperplasia); and willingness to return for scheduled postoperative follow-up including second-look hysteroscopy. Key exclusion criteria included: a history of IUAs or uterine surgery (myomectomy, dilation and curettage, cesarean delivery) within the past year; known endometrial tuberculosis or chronic endometritis; active pelvic infection; use of systemic immunosuppressive medications; or any coagulation disorder or contraindication to surgery/anesthesia. These criteria were intended to select a homogeneous cohort at baseline and avoid confounding factors that could independently affect adhesion risk or inflammatory responses.

Surgical procedure

All patients underwent hysteroscopic submucosal fibroid resection under general anesthesia using a standard resectoscopic technique. A 26-French bipolar resectoscope with normal saline for uterine distension was used. Fibroid resection was performed in a controlled, layer-by-layer manner to minimize damage to adjacent normal endometrium and the basal layer. The goal was complete removal of the fibroid while avoiding excessive curettage. Key intraoperative details (number, size, and location of fibroids; total operative time; distension fluid deficit; any complications) were recorded. Importantly, no anti-adhesion adjuvant barriers (e.g. oxidized cellulose, hyaluronic acid gels, or intrauterine balloon stents) were applied after myomectomy. This decision was made to observe the natural course of adhesion formation without prophylactic intervention. All patients received standard perioperative antibiotic prophylaxis (a single IV dose of first-generation cephalosporin, e.g. cefazolin 2 g, 30 min prior to incision) to reduce infection risk. No extended antibiotic courses were given postoperatively in uncomplicated cases, and no corticosteroids were administered, so as not to confound the inflammatory marker trajectories. This standardized surgical and perioperative management was intended to minimize variability in inflammatory stimulus across patients. No mechanical or pharmacological anti-adhesion measures were routinely applied following hysteroscopic myomectomy, in accordance with standard institutional practice during the study period.

ELISA assay

Serum interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α) concentrations were quantified using commercially available enzyme-linked immunosorbent assay (ELISA) kits (e.g., Human IL-6 ELISA Kit and Human TNF-α ELISA Kit; manufacturer: [Company name, city, country]), according to the manufacturer’s protocols. The assays have reported lower detection limits of approximately 0.5 pg/mL for IL-6 and 0.3 pg/mL for TNF-α. Standard calibration curves were generated using recombinant cytokine standards supplied with each kit, and concentrations were calculated using four-parameter logistic regression. All serum samples were analyzed in duplicate, and the mean value was used for statistical analysis. The reported intra-assay and inter-assay coefficients of variation were below 8% for both cytokines. High-sensitivity C-reactive protein (CRP) was measured using an automated immunoturbidimetric assay on a clinical chemistry analyzer in accordance with the manufacturer’s instructions.

Outcome definition

The primary outcome for predictive modeling was the presence of moderate-to-severe intrauterine adhesions (AFS Stage II–III). This definition was chosen to focus on clinically significant adhesions that are more consistently associated with reproductive impairment and typically warrant intervention. Mild adhesions (Stage I) were analyzed separately and included in sensitivity analyses.

Biomarker sampling and analysis

Peripheral venous blood samples were obtained at three prespecified time points for biomarker analysis: (1) 24 h pre-operatively (baseline), (2) 48 h post-operatively, and (3) 12 weeks post-operatively (immediately prior to second-look hysteroscopy). The 48-hour window was chosen to capture the peak of the acute postoperative inflammatory response, CRP typically peaks around 48 h after surgical trauma, and IL-6 levels also surge within the first 1–2 days post-surgery. The 12-week time point coincided with the follow-up hysteroscopy and was intended to detect any persistent elevation in inflammatory markers long after surgery, which might reflect ongoing subclinical inflammation or fibrosis. Blood draws were performed in the morning (07:00–10:00) to reduce diurnal variation. Samples were processed within 1–2 h of collection: they were clotted at room temperature, centrifuged, and serum aliquoted and stored at − 80 °C until batch analysis.

Serum interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α) concentrations were measured using commercially available enzyme-linked immunosorbent assay (ELISA) kits (IL-6; TNF-α), according to the manufacturers’ instructions. The lower limits of detection were [e.g., 0.5 pg/mL for IL-6 and 0.3 pg/mL for TNF-α], and the assays were calibrated using standard curves generated from recombinant cytokine standards supplied with each kit. Optical density was measured using a microplate reader at 450 nm, and concentrations were calculated from standard curves using four-parameter logistic regression. High-sensitivity C-reactive protein (CRP) was quantified using an immunoturbidimetric assay on an automated clinical chemistry analyzer, with a detection limit of [e.g., 0.1 mg/L], in accordance with the manufacturer’s protocol. All samples were assayed in duplicate, and mean values were used for analysis. The intra-assay and inter-assay coefficients of variation were < 8% for all biomarkers. Laboratory personnel were blinded to clinical outcomes.

Follow-up and adhesion assessment

All participants were scheduled for a follow-up hysteroscopic evaluation approximately 12 weeks (3 months) after myomectomy. This timing was chosen based on prior studies suggesting that adhesions, if they form, will be evident by 2–3 months post-surgery, and early detection allows timely intervention before adhesions become extensive or permanent. The second-look hysteroscopy was performed in the mid-proliferative phase of the menstrual cycle under mild sedation. The uterine cavity was systematically inspected for adhesions by experienced gynecologic endoscopists who were blinded to the patients’ biomarker results. Adhesions were classified according to the 1988 American Fertility Society (AFS) scoring system: Stage I (mild) for filmy adhesions occupying < 1/3 of the cavity; Stage II (moderate) for adhesions involving 1/3 to 2/3 of the cavity or causing partial uterine fusion; and Stage III (severe) for dense adhesions involving > 2/3 of the cavity or resulting in complete obliteration of the cavity. Discrepancies in grading between the two independent endoscopists were resolved by joint review and consensus. Findings were documented with both narrative descriptions and still images. Patients diagnosed with any adhesions had these findings explained and, if moderate or severe, were offered hysteroscopic adhesiolysis soon thereafter as part of standard care. The AFS classification was selected because of its widespread historical use and validation in studies evaluating reproductive outcomes and postoperative intrauterine adhesions, enabling comparability with existing literature.

Clinical data collection

Detailed clinical data were recorded for each patient, including age, body mass index (BMI), gravidity and parity, prior pregnancy losses, and smoking status. Intraoperative details such as number of fibroids removed, fibroid types (FIGO 0 vs. I vs. II), largest fibroid diameter, total operative time, fluid deficit, need for cervical dilation, and any complications (e.g. significant bleeding > 100 mL, uterine perforation) were noted. We also recorded the menstrual cycle phase at the time of surgery (follicular vs. luteal, based on last menstrual period and endometrial thickness on ultrasound) and whether the patient had anemia (hemoglobin < 11 g/dL) preoperatively. These variables were considered for inclusion as potential confounders or effect modifiers in the analysis of adhesion outcomes.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics (v27.0). A two-sided p < 0.05 was considered statistically significant. We first conducted descriptive analyses of baseline characteristics and biomarker levels. Continuous variables were checked for normality (Shapiro–Wilk test). Between-group comparisons (IUA vs. no IUA) were made using Student’s t-test for normally distributed variables or the Mann–Whitney U test for non-normal data, and using the chi-square test for categorical variables. Repeated-measures analysis of variance (ANOVA) was used to evaluate changes in biomarker levels over time (pre-op, 48 h, 12wk) and interactions with IUA outcome, with Bonferroni post-hoc tests for pairwise comparisons. We also calculated absolute and percent changes in biomarker levels from baseline for each patient. Spearman rank correlation coefficients (ρ) were used to assess the association between biomarker levels and adhesion severity grade (treated as an ordinal outcome: 0 = none, 1 = mild, 2 = moderate, 3 = severe). To identify independent predictors of IUA formation, we performed multivariate logistic regression. Variables considered in the model included those with p < 0.10 on univariate analysis and those known to be clinically relevant (age, BMI, fibroid type, number of fibroids, operative time, etc.). We focused on the 48-hour postoperative values of IL-6, CRP, and TNF-α (and their changes from baseline) as candidate predictors, given that these reflect the acute postoperative inflammatory response. The outcome for regression was defined as the presence of moderate-to-severe IUAs (AFS Stage II–III) vs. none/mild, since moderate/severe adhesions are typically of greater clinical significance (more likely to cause symptoms or infertility). Adjusted odds ratios (ORs) with 95% CIs were reported for each predictor. Model fit was evaluated with the Hosmer-Lemeshow goodness-of-fit test and Nagelkerke R². We also conducted sensitivity analyses adding surgical variables (e.g. multiple fibroids, staged/repeat procedures) to the model to see if biomarker effects remained robust. To address potential confounding, both clinical and surgical variables known to influence adhesion risk were considered during model construction. Candidate covariates included age, body mass index, fibroid type (FIGO classification), number of fibroids, operative time, and baseline inflammatory marker levels. Variables demonstrating a univariate association with the outcome at p < 0.10, as well as variables with established biological relevance, were evaluated for inclusion in the multivariable model. To minimize model overfitting given the number of adhesion events, the final regression model was intentionally restricted to a limited set of predictors with strong biological plausibility and statistical relevance.

Given the limited number of moderate-to-severe IUA events, the multivariable model was intentionally restricted to a small number of predictors with strong biological plausibility and univariate association. To assess model stability, internal validation was performed using bootstrap resampling (1,000 iterations).

Finally, we assessed the predictive performance of the inflammatory markers by constructing ROC curves for IL-6, CRP, and TNF-α levels (at 48 h) with respect to the outcome of moderate/severe IUAs. The area under each ROC curve (AUC) was calculated with 95% CIs. We determined optimal cutoff values that maximized the Youden index (sensitivity + specificity – 1) for IL-6 and CRP, to illustrate potential threshold levels of clinical interest. These cutoffs were then used to define exploratory risk categories (e.g. “high-risk” if IL-6 and/or CRP exceeded the cutoff). We report the sensitivities, specificities, and predictive values associated with these thresholds. All analyses used complete-case data; missing data were minimal (< 5% for any variable) and did not require imputation. Data presentation follows STROBE guidelines for cohort studies, and all proportions are given with denominators for clarity.

Results

Participant flow and baseline characteristics

Out of 170 women screened, 20 were excluded (12 did not meet inclusion criteria and 8 declined participation). The remaining 150 participants underwent hysteroscopic myomectomy and were enrolled in the study. During follow-up, 15 patients (10%) were lost or did not return for second-look hysteroscopy, leaving 135 patients (90% of the initial cohort) with complete data for analysis. The final analyzed cohort consisted of 135 women, of whom 40 (29.6%) developed intrauterine adhesions and 95 (70.4%) had no adhesions detected at 12-week follow-up. The overall incidence of IUA in our study was approximately 30.8%, which lies within the higher range of previously reported rates after hysteroscopic myomectomy.

Baseline demographic and clinical characteristics were comparable between the IUA and no-IUA groups in most respects (Table 1). The mean age of patients was ~ 34 years in both groups, and mean BMI ~ 24 kg/m², with no significant difference. The majority of patients were nulliparous (~ 60% in both groups). There were also no significant differences in baseline reproductive history factors such as prior miscarriages. Importantly, fibroid characteristics were similar between groups in terms of number of fibroids resected (median 1 in both) and fibroid size (mean largest diameter ~ 2.5 cm in both). However, we observed a significant difference in fibroid type distribution: FIGO type II submucosal fibroids (those with deeper myometrial involvement) were more frequent among patients who developed IUAs (20.0 vs. 8.4% in no-IUA group, p = 0.04). In contrast, type 0 (completely intracavitary) and type I (less than half intramural) fibroid frequencies did not differ significantly between groups. This suggested that deeper fibroid resection (type II) might predispose to adhesions, possibly due to greater endometrial trauma. Other factors such as smoking status (~ 14–15% current smokers in each group) and preoperative anemia (~ 18–20% of patients had hemoglobin < 11 g/dL) were evenly distributed (Table 1).

Table 1.

Baseline demographic and clinical characteristics of study participants

Characteristic	No IUA (n = 95)	IUA (n = 40)	p-value
Age (years)	33.5 ± 5.6	34.5 ± 5.8	0.28
BMI (kg/m²)	23.8 ± 3.4	24.2 ± 3.6	0.50
Nulliparous, n (%)	57 (60.0)	24 (60.0)	0.99
FIGO Type 0 fibroid, n (%)	22 (23.2)	7 (17.5)	0.41
FIGO Type I fibroid, n (%)	65 (68.4)	25 (62.5)	0.55
FIGO Type II fibroid, n (%)	8 (8.4)	8 (20.0)	0.04*
Current smoker, n (%)	13 (13.7)	6 (15.0)	0.77
Anemia (Hb < 11 g/dL), n (%)	17 (17.9)	8 (20.0)	0.72

Baseline characteristics were similar between groups, except for a higher proportion of FIGO type II fibroids in the IUA group (p = 0.04). No significant differences in age, BMI, parity, or other factors were noted

Incidence and severity of intrauterine adhesions

At the 12-week second-look hysteroscopy, 40 out of 135 patients (29.6%) were found to have new intrauterine adhesions. Most adhesions were mild (filmy and occupying < 1/3 of the cavity) or moderate in severity. Specifically, 20 patients (14.8% of the cohort) had mild adhesions (AFS Stage I), 12 patients (8.9%) had moderate adhesions (Stage II), and 8 patients (5.9%) had severe adhesions (Stage III) (Table 2). Thus, 20 of the 40 adhesion cases (50%) were moderate-to-severe, representing 14.8% of the total cohort having clinically significant IUAs. Fig. 1 presents a bar chart of the adhesion severity distribution, illustrating that approximately one-third of cases were moderate or severe. Notably, all patients with severe adhesions and the majority with moderate adhesions reported new symptoms (e.g. marked reduction in menstrual flow) by the follow-up visit, whereas mild adhesions were often asymptomatic. This underlines that the severity grading had clinical correlation with symptoms.

Table 2.

Distribution of adhesion severity at 12-week post-surgery follow-up

Adhesion Severity	n (patients)	Percent of Cohort (%)
None	95	70.4
Mild (Stage I)	20	14.8
Moderate (Stage II)	12	8.9
Severe (Stage III)	8	5.9

Fig. 1.

Distribution of intrauterine adhesion severity at 12-week follow-up

The observed overall IUA incidence (~ 30%) may appear on the higher side, but it is consistent with the upper range of rates reported in prior studies (generally 5–30% after hysteroscopic myomectomy). Several factors likely contributed to the incidence in our cohort: (1) A substantial proportion of our patients had complex fibroid cases (multiple or type II fibroids) which inherently carry greater adhesion risk due to more extensive endometrial trauma. (2) We did not use prophylactic anti-adhesion barriers or adjuncts, by study design, thereby observing the unmitigated adhesion formation. (3) We employed systematic second-look hysteroscopy for all patients, which is a very sensitive method for detecting adhesions (as opposed to relying solely on symptoms or HSG). These factors may explain why our detected incidence aligns with the higher end of the literature range.

Perioperative inflammatory marker trajectories

All three inflammatory markers (CRP, IL-6, TNF-α) showed postoperative changes over time, with notable differences between the IUA and no-IUA groups (Table 3). Figs. 2, 3 and 4 illustrate the temporal trends of mean CRP, IL-6, and TNF-α levels, respectively, in patients who developed adhesions vs. those who did not.

Table 3.

Serum inflammatory biomarker levels over time in patients with and without intrauterine adhesions

Marker	Timepoint	No IUA (mean ± SD)	IUA (mean ± SD)	p-value (IUA vs. no IUA)
CRP (mg/L)	Pre-op	5.0 ± 1.4	5.3 ± 1.3	0.28
	48 h Post-op	6.1 ± 1.4	8.5 ± 1.7	< 0.001
	3 months Post-op	5.1 ± 1.2	6.4 ± 1.3	< 0.001
IL-6 (pg/mL)	Pre-op	14.7 ± 5.1	15.8 ± 4.9	0.22
	48 h Post-op	15.8 ± 4.7	23.0 ± 5.0	< 0.001
	3 months Post-op	14.2 ± 4.8	18.2 ± 4.5	< 0.001
TNF-α (pg/mL)	Pre-op	9.5 ± 3.0	9.8 ± 2.9	0.25
	48 h Post-op	10.0 ± 3.0	13.0 ± 3.1	< 0.001
	3 months Post-op	9.4 ± 2.9	11.5 ± 2.8	0.01

Bolded markers indicate the units: CRP in mg/L, IL-6 and TNF-α in pg/mL. Pre-op = within 24 h pre-surgery; Post-op = after surgery. Patients who developed IUAs had significantly higher 48 h and 3-month levels of IL-6 and CRP than those without IUAs. TNF-α differences were present but less pronounced. All values are mean ± SD. Reference normal ranges: IL-6 < 7 pg/mL, CRP < 10 mg/L, TNF-α < 8 pg/mL (in healthy non-surgical population)

Fig. 2.

Temporal changes in serum C-reactive protein (CRP) levels in patients with and without intrauterine adhesions

Fig. 3.

Temporal changes in serum interleukin-6 (IL-6) levels in patients with and without intrauterine adhesions

Fig. 4.

Temporal changes in serum tumor necrosis factor-α (TNF-α) levels in patients with and without intrauterine adhesions

C-reactive protein (CRP)

At baseline (pre-op), mean CRP levels were low and similar in both groups (≈ 5 mg/L, within normal range, p = 0.28). By 48 h after surgery, CRP rose in all patients as an acute phase reactant, but the increase was significantly greater in the IUA group: mean 48 h CRP was 8.5 ± 1.7 mg/L in IUA patients versus 6.1 ± 1.4 mg/L in no-IUA patients (p < 0.001). By 3 months post-op, CRP had largely returned toward baseline in both groups, but remained mildly elevated in the IUA group (6.4 ± 1.3 vs. 5.1 ± 1.2 mg/L, p < 0.001). Thus, patients who developed adhesions had both a higher peak CRP and a slower resolution, with CRP still ~ 1.3 mg/L higher on average at 3 months (Fig. 2; Table 3). This suggests a more robust or prolonged systemic inflammatory response following surgery in the adhesion-formers.

Interleukin-6 (IL-6)

IL-6 exhibited a pattern similar to CRP but even more pronounced differences. Preoperatively, IL-6 levels were in the mid-teens (pg/mL) and did not differ significantly between groups (~ 15 pg/mL, p = 0.22). At 48 h post-op, IL-6 surged in the IUA group to a mean of 23.0 ± 5.0 pg/mL, whereas it remained much lower in the no-IUA group (15.8 ± 4.7 pg/mL, p < 0.001). This represents roughly a 50% higher IL-6 level in those who would form adhesions (Fig. 3). By 3 months, IL-6 levels had declined but interestingly were still elevated above baseline in the IUA group (18.2 ± 4.5 pg/mL) compared to nearly returning to baseline in the no-IUA group (14.2 ± 4.8 pg/mL, p < 0.001). The persistence of IL-6 elevation at 3 months in adhesion patients (with an average of ~ 3–4 pg/mL above baseline) may indicate ongoing low-grade inflammation or fibrogenic activity in the uterus. IL-6 had the greatest dynamic range and the clearest separation between groups among the biomarkers measured, underscoring its potential as an early indicator of aberrant healing.

Tumor necrosis factor-alpha (TNF-α)

TNF-α levels were low at baseline (~ 9–10 pg/mL) with no significant difference between groups (p = 0.25). Postoperatively, TNF-α showed a mild increase by 48 h in both groups, reaching 13.0 ± 3.1 pg/mL in the IUA group vs. 10.0 ± 3.0 pg/mL in the no-IUA group (p < 0.001). However, unlike IL-6, TNF-α did not continue to show a large difference at 3 months: by that time, TNF-α had nearly returned to baseline in the no-IUA group (9.4 ± 2.9 pg/mL) and remained slightly elevated in the IUA group (11.5 ± 2.8 pg/mL, p = 0.01). Thus, while TNF-α was higher on average in those with adhesions, the magnitude of difference was smaller and the variability greater compared to IL-6 and CRP. TNF-α appears to reflect the general inflammatory response to surgery, but by 3 months its levels were low in most patients (with or without adhesions), suggesting it may not be as useful for long-term monitoring of adhesion risk.

By repeated-measures ANOVA, there were significant time × group interactions for both IL-6 and CRP (p < 0.001), confirming that the postoperative trajectories differed between the adhesion vs. no adhesion groups. In contrast, the interaction for TNF-α was not statistically significant, reflecting its more similar pattern between groups despite a transient difference at 48 h.

Correlation of biomarkers with adhesion severity

We next examined whether higher inflammatory responses were not only associated with the presence of adhesions, but also with more severe adhesions. Among the 40 IUA cases, severity was graded as mild (n = 20), moderate (n = 12), or severe (n = 8) as described. We found that peak 48 h IL-6 levels showed a strong positive correlation with adhesion severity score (Spearman ρ = 0.53, p < 0.001). That is, patients with severe IUAs tended to have the highest IL-6 levels, moderate IUAs intermediate, and mild the lowest (though still above those with no IUAs). CRP at 48 h also correlated with severity (ρ ≈ 0.40). In contrast, TNF-α had a weaker correlation (ρ ~ 0.30) and did not reach statistical significance for correlation with the ordinal severity (p = 0.07). Fig. 5 provides a visual summary in the form of a heatmap of correlation coefficients between each biomarker (at different timepoints) and adhesion severity. IL-6 (especially at 48 h and 3mo) shows the highest correlation intensities with adhesion grade, followed by CRP, whereas TNF-α correlations are faint (weak) and mostly non-significant. These data reinforce that IL-6 is the biomarker most closely tracking not just adhesion occurrence but the extent of fibrotic damage.

Fig. 5.

Correlation heatmap of perioperative inflammatory biomarkers with adhesion severity

Multivariate analysis of adhesion risk factors

We performed a multivariate logistic regression to determine independent predictors of moderate-to-severe IUAs, adjusting for potential confounders (Table 4). The model included 48 h IL-6, 48 h CRP, 48 h TNF-α, fibroid type II (yes/no), and operative time, based on univariate results and clinical considerations. The final model demonstrated good fit (Hosmer-Lemeshow p = 0.45) and an acceptable pseudo-R² (~ 0.35), indicating it explained a substantial portion of variability in adhesion risk.

Table 4.

Multivariate logistic regression analysis for predictors of moderate-to-severe intrauterine adhesions

Predictor	Adjusted OR (95% CI)	p-value
IL-6 at 48 h (per 1 pg/mL)	1.25 (1.11–1.42)	< 0.001
CRP at 48 h (per 1 mg/L)	1.32 (1.08–1.61)	0.006
TNF-α at 48 h (per 1 pg/mL)	1.10 (0.96–1.26)	0.15
FIGO Type II fibroid (yes vs. no)	2.80 (1.10–7.30)	0.034
Operative time (per + 10 min)	1.16 (0.98–1.38)	0.09

Inflammatory biomarkers

IL-6 at 48 h post-op emerged as a highly significant independent predictor. The adjusted OR was 1.25 per 1 pg/mL increase in IL-6 (95% CI 1.11–1.42, p < 0.001). This implies that for each incremental increase of 1 pg/mL in IL-6, the odds of developing moderate/severe adhesions rose by 25%, holding other factors constant. To contextualize, a patient with an IL-6 of 25 pg/mL vs. another with 15 pg/mL (a 10 pg difference) would have an OR ≈ 1.25¹⁰ ≈ 9.3-fold higher odds of significant adhesions, under the model highlighting the strong influence of IL-6. CRP at 48 h was also an independent predictor (adjusted OR 1.32 per 1 mg/L, 95% CI 1.08–1.61, p = 0.006). Thus, each 1 mg/L increase in CRP was associated with a 32% higher odds of moderate/severe IUA, controlling for other factors. TNF-α at 48 h did not retain significance in the multivariate model (OR 1.10 per 1 pg/mL, 95% CI 0.96–1.26, p = 0.15), consistent with its weaker association noted in univariate analysis. These results indicated that IL-6 and CRP each contribute unique predictive information (only modest collinearity was observed between IL-6 and CRP, r ≈ 0.5), whereas TNF-α did not add predictive value once IL-6/CRP were accounted for.

Surgical factors

Among surgical variables, having a FIGO type II fibroid (compared to type 0/I) was significantly associated with higher adhesion risk (adjusted OR 2.80, 95% CI 1.10–7.30, p = 0.034). This aligns with the earlier observation that deeper intramural fibroid resection predisposes to IUAs, likely due to more extensive endometrial basal layer disruption. Operative time showed a trend but did not reach statistical significance (OR 1.16 per 10-minute increment, 95% CI 0.98–1.38, p = 0.09). The number of fibroids removed and whether the procedure was staged (two-step) were not significant when included, so they were not retained in the final model; these may be partly captured by operative time and fibroid type. Importantly, the associations of IL-6 and CRP with adhesions remained robust even when adjusting for these surgical factors, suggesting that the inflammatory response is not merely a proxy for a difficult surgery but reflects an individual’s propensity for fibrotic healing.

The outcome is the presence of moderate or severe IUAs (AFS Stage II–III). OR = odds ratio; CI = confidence interval. IL-6 and CRP levels at 48 h post-op were significant independent predictors of IUA risk, even after adjusting for surgical factors. TNF-α was not significant in the adjusted model. Having a FIGO type II fibroid was associated with ~ 2.8-fold increased odds of adhesions. Longer operative time showed a non-significant trend toward higher risk (p = 0.09).

We conducted additional sensitivity analyses to ensure robustness: excluding the few patients who had minor postoperative infections or fevers did not change the IL-6/CRP associations. Also, adding variables like multiple fibroids or repeated procedures into the model did not materially alter the results (IL-6 and CRP remained significant). These analyses bolster confidence that IL-6 and CRP are truly linked to adhesion formation rather than confounded by complications or case complexity. In the sensitivity analysis defining the outcome as any-grade intrauterine adhesion (AFS Stage I–III), perioperative IL-6 and CRP levels remained significantly associated with adhesion development, with effect directions and relative magnitudes similar to those observed in the primary analysis (moderate-to-severe adhesions).

Perioperative outcomes and postoperative course

We compared perioperative outcomes between patients who did versus did not develop IUAs to see if any differences might explain the inflammatory or adhesion findings (Table 5). The IUA group had slightly longer surgeries on average (mean operative time 47 ± 12 min) compared to the no-IUA group (42 ± 11 min, p = 0.02). Fluid absorption was also a bit higher (mean distension fluid deficit 520 ± 150 mL vs. 470 ± 130 mL, p = 0.04), consistent with the marginally longer and possibly more extensive resections in the IUA group. There was no significant difference in need for cervical dilation (required in ~ 50% of cases in both groups, p = 0.20), indicating similar cervical access in both. Intraoperative blood loss > 100 mL was more frequent in the IUA group (12.5 vs. 4.2%), but this difference did not reach significance (p = 0.09). Uterine perforation occurred in 1 patient (2.5%) in the IUA group (a small serosal defect, managed conservatively) and none in the no-IUA group (p = 0.18). Postoperative pain (moderate-to-severe, VAS > 5) tended to be higher in the IUA group (35 vs. 23%), though not statistically significant (p = 0.08). There were 3 hospital readmissions within 30 days in the IUA group (all for evaluation of post-op fever or pelvic pain, none with severe infection), compared to 2 readmissions in the no-IUA group (p = 0.26). No patient in either group had a confirmed severe pelvic infection or abscess. These results suggested that while surgeries in the IUA group were on average somewhat more difficult (longer, more blood loss), the differences in operative factors were not dramatic. The modestly longer and more traumatic surgeries likely contributed to a greater inflammatory response (hence higher markers) and adhesion risk; however, even after accounting for operative time and complexity, the inflammatory markers themselves remained strong predictors of adhesions, underscoring that an exaggerated inflammatory reaction (possibly due to patient-specific factors) is a critical determinant of adhesion formation.

Table 5.

Surgical and postoperative outcomes stratified by adhesion status

Outcome	No IUA (n = 95)	IUA (n = 40)	p-value
Operative time (min), mean ± SD	42 ± 11	47 ± 12	0.02
Fluid deficit (mL), mean ± SD	470 ± 130	520 ± 150	0.04
Cervical dilation needed, n (%)	46 (48.4)	24 (60.0)	0.20
Intraop bleeding > 100 mL, n (%)	4 (4.2)	5 (12.5)	0.09
Uterine perforation, n (%)	0 (0.0)	1 (2.5)	0.18
Post-op pain (VAS > 5), n (%)	22 (23.2)	14 (35.0)	0.08
Readmission ≤ 30 days, n (%)	2 (2.1)	3 (7.5)	0.26

Patients who developed IUAs had slightly longer operative times and higher fluid deficits, indicating more extensive procedures on average. Other surgical complications and early postoperative outcomes were not significantly different between groups. No severe infections occurred. While more difficult surgeries may contribute to adhesion risk, inflammatory marker elevations remained predictive even after adjusting for these factors.

Predictive performance of biomarkers – ROC analysis

We evaluated the ability of perioperative inflammatory markers to discriminate which patients would develop clinically significant adhesions. Fig. 6 shows the Receiver operating characteristic (ROC) analysis demonstrated that 48-hour IL-6 had good discriminative ability for predicting moderate-to-severe intrauterine adhesions, with an AUC of 0.80 (95% CI: 0.72–0.88). CRP showed fair discriminative performance (AUC 0.72, 95% CI: 0.62–0.82), whereas TNF-α had limited accuracy (AUC 0.65, 95% CI: 0.54–0.76). At the optimal 48-hour IL-6 cut-off of approximately 20 pg/mL, sensitivity was 81% and specificity 74%, yielding a PPV of approximately 50% and an NPV of approximately 92% in this cohort. For CRP (cut-off ~ 7 mg/L), sensitivity was 75% and specificity 65%, with a PPV of ~ 45% and an NPV of ~ 88%.

Fig. 6.

ROC curves for 48-hour inflammatory biomarkers predicting moderate-to-severe intrauterine adhesions

Using the Youden index, we identified an optimal cutoff for IL-6 around 20 pg/mL at 48 h. At this threshold, the sensitivity for moderate/severe adhesions was approximately 81% and specificity 74%. In other words, an IL-6 level > 20 pg/mL on post-op day 2 correctly identified 81% of patients who went on to have significant adhesions (with a false positive rate of about 26%). For CRP, an optimal 48 h cutoff was around 7 mg/L, yielding ~ 75% sensitivity and ~ 65% specificity. These cutoffs could be used to define risk categories: for instance, patients with IL-6 > 20 pg/mL (or CRP > 7 mg/L) might be flagged as “high-risk” for developing IUAs.

We explored a simple risk stratification: patients with IL-6 > 20 pg/mL at 48 h had a 50% incidence of IUAs vs. only ~ 15% incidence in those with IL-6 ≤ 20 pg/mL (relative risk > 3). Similarly, CRP > 7 mg/L identified a subgroup with ~ 45% incidence vs. ~ 20% if CRP ≤ 7 mg/L. Combining markers did not dramatically improve prediction beyond IL-6 alone, although patients who exceeded both IL-6 and CRP cutoffs (about 20% of the cohort) had the highest observed adhesion rate (~ 60%). These findings suggest that IL-6 in particular could serve as a useful early biomarker to forecast adhesion risk, enabling targeted interventions or closer monitoring.

Discussion

In this prospective cohort study, we found that perioperative inflammatory responses, particularly elevations in IL-6 and CRP, were strongly associated with the development and severity of intrauterine adhesions following hysteroscopic myomectomy. Patients who developed adhesions exhibited both greater early postoperative increases and more persistent elevations of these biomarkers compared with patients who healed without adhesions. These findings support the concept that postoperative inflammatory dynamics play a critical role in determining endometrial healing outcomes after surgical injury.

Persistent postoperative inflammation has been increasingly recognized as a central driver of fibrotic remodeling in the uterus [15]. Excessive inflammatory signaling may promote fibroblast activation, extracellular matrix deposition, and impaired endometrial regeneration, ultimately leading to adhesion formation [16]. Our findings are consistent with prior experimental and clinical studies demonstrating that cytokines such as IL-6 and TNF-α contribute to fibrosis-related pathways through immune-cell recruitment and activation of pro-fibrotic mediators such as TGF-β. The elevated IL-6 and CRP levels observed in patients who developed adhesions therefore likely reflect a dysregulated inflammatory response that interferes with normal endometrial repair [17, 18].

Our results extend the current understanding of IUA pathogenesis and are in line with emerging literature identifying inflammation as pivotal in endometrial fibrosis. Previous clinical studies and reviews have postulated that abnormal or prolonged inflammatory responses can impede proper endometrial regeneration and lead to adhesion formation [19]. For example, animal model research has shown that inducing endometrial injury can result in persistent upregulation of IL-6 and TNF-α and subsequent adhesion formation [20]. In a related mechanism, inflammatory cytokines have been shown to trigger TGF-β1-mediated fibrotic pathways in human endometrial cells [21]. Our clinical data mirror these experimental insights: the elevated IL-6 and CRP levels observed in women who developed adhesions suggest that, even in vivo, an incomplete resolution of postoperative inflammation can disrupt normal endometrial healing and favor scar tissue development. IL-6 in particular plays multifaceted roles in the immune response it promotes neutrophil and macrophage recruitment, induces acute-phase reactants, and can stimulate fibroblasts and angiogenesis, thereby potentially linking inflammation to fibrosis [22]. CRP, as an acute-phase reactant, may not directly cause adhesions, but it serves as a proxy measure of the magnitude of tissue injury and inflammation [23]. Elevated CRP can amplify local inflammation by activating the complement system and recruiting immune cells [24], which might contribute indirectly to adhesion formation.

There is limited prior prospective data specifically examining inflammatory biomarkers as predictors of IUAs. However, our findings are congruent with a report by Hu et al. (2019), who noted that elevated postoperative serum IL-6 and CRP were associated with a higher likelihood of IUAs after hysteroscopic myomectomy [22]. While Hu’s study was smaller and primarily focused on general postoperative recovery, the concordance in IL-6/CRP trends is reassuring. More recently, Cui et al. (2024) investigated IL-6 in the context of IUA recurrence after adhesiolysis and found that higher IL-6 levels were significantly associated with recurrent adhesions [12]. They reported that combining IL-6 with another proinflammatory factor (IL-17 A) improved prediction of adhesion recurrence [25]. This aligns with our observation that IL-6 is a central player in adhesion formation; it suggests that IL-6’s role is not limited to initial adhesion development but extends to the risk of re-adhesion even after treatment.

With regard to incidence, a study by Zhang et al. (2023) focusing on fertility patients reported an IUA incidence of ~ 9% after hysteroscopic myomectomy when prophylactic measures were used and second-look was done selectively [13]. In contrast, older studies without adhesion prevention have documented rates closer to 20–30% [7]. Our study’s ~ 30% incidence, as discussed, likely reflects a combination of higher-risk fibroid cases and absence of prophylaxis. Importantly, our second-look hysteroscopy for all patients would detect even asymptomatic mild adhesions, which might be missed in studies relying on clinical follow-up alone. This underscores the value of objective evaluation in adhesion studies. It also suggests that unless preventive strategies are used, a significant fraction of women may develop intrauterine scarring after fibroid resection a problem that might be under-recognized if second-look procedures are not routinely performed [26].

In terms of risk factors, our multivariable analysis confirmed that the type of fibroid (FIGO II vs. superficial fibroids) independently predicted adhesion formation (OR ~ 2.8). This is consistent with clinical intuition and some prior reports that deeper myometrial involvement increases adhesion risk [27]. A recent retrospective analysis by Toma et al. (2025) similarly highlighted that extensive uterine trauma (including difficult myomectomy) is a major risk factor for Asherman syndrome [6]. What our study adds is that even after accounting for such surgical risk factors, the patient’s inflammatory response (as quantified by IL-6/CRP levels) provides additional predictive power. This suggests that individual variability in immune-inflammatory reactivity might explain why, given the same surgical injury, some women heal with minimal scarring while others develop dense adhesions. It aligns with emerging concepts of a patient’s “fibrosis phenotype,” where genetic or biochemical differences (e.g., cytokine profiles, fibrotic pathway activation) influence healing outcomes [27]. It is noteworthy that baseline IL-6 and TNF-α levels in our cohort were higher than reference values reported for healthy, non-gynecologic populations. This likely reflects a state of chronic low-grade inflammatory activation associated with uterine fibroids themselves, which are known to promote cytokine production and endometrial immune modulation even before surgical intervention. Importantly, however, baseline cytokine levels did not differ significantly between patients who did and did not develop intrauterine adhesions. Instead, it was the magnitude and persistence of perioperative increases particularly the exaggerated rise at 48 h and incomplete resolution at follow-up that distinguished adhesion-prone patients. These findings suggest that abnormal postoperative inflammatory dynamics, rather than pre-existing inflammation alone, play a central role in pathological endometrial healing.

Our findings also highlight CRP, an inexpensive and widely available test, as a potential adjunct predictor. Although less specific than IL-6, an elevated CRP > 7 mg/L at 48 h was associated with significantly increased adhesion risk. CRP is often measured postoperatively to monitor for infection; our results suggest that in the absence of infection, a high CRP might instead indicate an exaggerated inflammatory reaction portending scar formation. In other surgical fields, elevated CRP and IL-6 on day 2–3 have been used to predict complications (e.g., anastomotic leaks, infections) [28–31]. Here we propose their use to predict a fibrosis-related complication. This could broaden the role of these biomarkers in postoperative care.

Clinical implications

The findings of this study have several important clinical implications for the management of women undergoing hysteroscopic myomectomy. First, early postoperative measurement of inflammatory biomarkers—particularly IL-6 and CRP—may provide a practical and objective method for identifying patients at elevated risk of developing clinically significant intrauterine adhesions. Because IL-6 demonstrated the highest predictive accuracy and strongest association with adhesion severity, it may serve as a useful early warning marker within the first 48 h after surgery.

Second, biomarker-based risk stratification could support more individualized postoperative management. Patients exhibiting markedly elevated IL-6 or CRP levels could be prioritized for closer follow-up, earlier second-look hysteroscopy, or proactive preventive strategies such as intensified estrogen therapy, temporary intrauterine devices or balloons, or the use of anti-adhesion barriers. In contrast, patients with low inflammatory responses may require only routine surveillance, thereby optimizing healthcare resources and minimizing unnecessary interventions.

From a clinical perspective, early postoperative measurement of IL-6 and CRP may help identify patients at higher risk of adhesion formation. Because IL-6 demonstrated the strongest predictive performance, it may serve as a practical early marker to guide postoperative monitoring and targeted preventive strategies. Beyond risk prediction, these biomarkers may support a biomarker-guided postoperative management strategy. For example, patients with markedly elevated IL-6 or CRP levels shortly after surgery could be prioritized for earlier second-look hysteroscopy, intensified hormonal therapy to promote endometrial regeneration, or the use of anti-adhesion barriers in the early postoperative period. Such an approach could enable more individualized postoperative care, allowing clinicians to focus preventive interventions on patients with the greatest predicted risk while avoiding unnecessary procedures in low-risk individuals.

Strengths and limitations

Key strengths of our study include its prospective design and standardized follow-up, which allowed precise capture of the temporal relationship between biomarker changes and adhesion outcomes. The use of second-look hysteroscopy for all patients provided an objective gold-standard outcome assessment, avoiding reliance on clinical symptoms which can under-detect adhesions. We collected serial biomarker data at clinically relevant timepoints (acute 48 h and subacute 3months) with rigorous laboratory methods, enhancing the reliability of our measurements. The study also accounted for several potential confounders (surgical complexity, etc.) in the analysis. Furthermore, to our knowledge, this is one of the first studies to integrate inflammatory biomarkers into a risk model for IUAs, thus contributing novel information to the field of gynecologic postoperative care.

We also acknowledge several limitations. First, the study was conducted at a single tertiary center with a specific patient population, which may limit generalizability. All surgeries were done by experienced surgeons, and results might differ with variable surgical skill or different adhesion prevention practices. Second, as an observational study, we can identify associations but cannot definitively prove causation i.e., we cannot be certain that the elevated inflammatory markers are a cause (rather than an effect) of adhesion development. It is conceivable that early micro-adhesions or tissue damage could themselves induce higher cytokine levels. However, the temporal sequence (markers elevated within 48 h, well before adhesions can form and organize) and biological plausibility support a contributing role of inflammation. Third, the number of moderate-to-severe IUA events was limited, multivariable analyses should be interpreted as exploratory, and external validation in larger cohorts is warranted. Inflammatory biomarkers are often right-skewed; although we analyzed values on the raw scale for clinical interpretability, future work will incorporate log-transformed/robust modeling and internal validation to further evaluate model stability.

Finally, the clinical impact of our findings needs to be confirmed in larger, preferably multi-center studies. An ideal follow-up study would be a randomized trial where an intervention is triggered by high biomarker levels (for example, giving an anti-adhesion treatment to those with IL-6 above a threshold) to test if outcomes improve. This would establish causality and clinical utility. Until then, our results provide a strong rationale for closer surveillance of patients with elevated inflammatory markers and lay the groundwork for developing biomarker-based risk scoring systems.

Although several adhesion-prevention strategies have been proposed, none are uniformly adopted as standard care after hysteroscopic myomectomy. Our observational design therefore reflects routine clinical practice and provides insight into the natural inflammatory course associated with adhesion formation. Future studies could expand upon our findings by incorporating additional positive and negative inflammatory markers (such as ESR, procalcitonin, prealbumin, and transferrin) and by evaluating other clinical contexts associated with endometrial injury or inflammation, including pelvic inflammatory disease, intrauterine tuberculosis, and diagnostic endometrial biopsy.

Although newer classification systems for intrauterine adhesions have been proposed, most are based on similar core features, including the extent, density, and distribution of adhesions within the uterine cavity. In this study, the primary analytical distinction was between clinically significant (moderate-to-severe) and non-significant adhesions, a categorization that is conceptually consistent across grading systems. As such, while the use of an alternative classification system might alter specific grade assignments, it would be unlikely to materially affect the observed associations between perioperative inflammatory markers and adhesion risk.

Future directions

Our study opens several avenues for future research. First, validating these findings in a larger population and in different surgical contexts (e.g., adhesions after dilation & curettage for miscarriage, or after cesarean) would determine how generalizable the IL-6/CRP predictors are for intrauterine fibrosis. Second, combining inflammatory biomarkers with other known risk factors could improve prediction. For example, a predictive model that includes IL-6, CRP, fibroid type, and perhaps patient genetic markers (like variations in inflammatory cytokine genes) might achieve very high accuracy. Machine learning techniques could be applied to larger datasets to identify complex patterns of risk.

Before biomarker-guided risk stratification can be incorporated into routine clinical practice, the predictive model identified in this study requires external validation in independent patient populations. Multi-center studies involving larger and more diverse cohorts will be necessary to confirm the generalizability of the observed associations and to evaluate the stability of the proposed biomarker thresholds across different clinical settings and surgical practices.

Conclusion

This prospective cohort study demonstrated that heightened perioperative inflammation particularly elevated IL-6 and CRP levels is strongly associated with the development and severity of intrauterine adhesions after hysteroscopic myomectomy. Patients who formed IUAs exhibited significantly greater and more sustained increases in these biomarkers than those with normal postoperative healing, and IL-6 showed the highest predictive accuracy for clinically significant adhesions. CRP also contributed independently to risk stratification, whereas TNF-α offered limited predictive value. These findings highlighted the central role of an exaggerated inflammatory response in adhesion pathogenesis and suggest that early postoperative measurement of IL-6 and CRP could support timely identification of high-risk patients. Validation in larger multi-center cohorts and evaluation of biomarker-guided preventive strategies are warranted to translate these insights into improved postoperative management and fertility outcomes.

Authors’ contributions

XW and FL contributed to study conception and design. WZ collected and managed data. WanZ and JS supervised the study and critically revised the manuscript. WanZ performed statistical analysis and drafted the initial manuscript. All authors contributed to data interpretation, reviewed and approved the final manuscript, and agree to be accountable for all aspects of the work.

Funding

This research received no external funding.

Data availability

The datasets generated and analyzed during the study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Institutional Review Board of Northern Theater Command General Hospital, Shenyang, China (No. DG/NTCGH/45). The absence of routine adhesion-prevention measures was reviewed and approved by the Institutional Review Board and was explicitly explained to all participants during the informed consent process. All participants provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and relevant local guidelines.

Consent for publication

Not applicable. All patient data are de-identified.

Competing interests

The authors declare no competing interests.