Early MRI-Detected Interspinous Ligament Region Edema Is Associated With Delayed Pain Recovery and Persistent Low Back Pain After Unilateral Biportal Endoscopic Surgery

Abstract

Introduction

Persistent postoperative low back pain remains a concern after unilateral biportal endoscopic (UBE) surgery, but the impact of surgery-related interspinous ligament edema is unclear. This study investigated whether magnetic resonance imaging (MRI)-detected interspinous ligament region edema (ISLRE) was associated with delayed pain recovery and persistent postoperative low back pain after UBE surgery.

Methods

This retrospective cohort study included patients who underwent single-level UBE surgery between October 2021 and October 2024. On the basis of MRI within 3 days postoperatively, ISLRE was graded as grade 0 (no/minimal edema), grade 1 (localized edema), or grade 2 (diffuse or confluent edema involving most of the operated interspinous region). Clinical outcomes included visual analogue scale scores for low back pain (VAS-LBP), Oswestry disability index (ODI), and persistent postoperative low back pain, defined as 12-month VAS-LBP > 3. Multivariable linear and logistic regression analyses were performed to assess independent associations, and linear mixed-effects modeling was used to evaluate longitudinal pain recovery trajectories.

Results

A total of 245 patients were included (grade 0: n = 117; grade 1: n = 74; grade 2: n = 54). Interobserver agreement for MRI grading was substantial (κ = 0.802). At 12 months, VAS-LBP scores increased progressively across edema grades (1.66 ± 1.20, 2.34 ± 1.45, and 3.54 ± 1.65 for grades 0–2, respectively; P < 0.001), while persistent postoperative low back pain occurred in 21.4%, 41.9%, and 68.5% of patients, respectively (P < 0.001). Compared with grade 0 edema, grade 2 edema was independently associated with higher 12-month pain scores (β = 1.78, 95% confidence interval [CI] 1.34–2.22; P < 0.001) and more than eightfold increased odds of persistent postoperative low back pain (odds ratio [OR] 8.90, 95% CI 4.01–19.76; P < 0.001). Linear mixed-effects modeling demonstrated significantly delayed pain recovery trajectories in higher-grade edema groups. Although statistically significant, the magnitude of ODI differences at 12 months was below the minimal clinically important difference (MCID) threshold, suggesting limited clinical relevance.

Conclusions

The extent of postoperative ISLRE, particularly grade 2 edema, may be useful for postoperative risk stratification and targeted patient management after UBE surgery.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40122-026-00872-5.

Key Summary Points

Why carry out this study?

Unilateral biportal endoscopic (UBE) surgery is widely used for lumbar disc herniation, but postoperative pain outcomes remain variable and not fully understood.

Early postoperative magnetic resonance imaging (MRI) findings such as interspinous ligament region edema (ISLRE) may reflect tissue response but their clinical significance remains unclear.

This study aimed to investigate whether ISLRE is associated with postoperative pain trajectories and long-term outcomes.

What was learned from this study?

Higher-grade ISLRE was significantly associated with worse postoperative pain trajectories and increased risk of persistent low back pain at 12 months.

These findings suggest that ISLRE may serve as a postoperative imaging biomarker reflecting tissue response and recovery heterogeneity after UBE surgery.

However, the association is observational, and the clinical utility of early MRI-based risk stratification requires further validation.

Introduction

Unilateral biportal endoscopic (UBE) surgery has been increasingly adopted for the treatment of lumbar disc herniation (LDH) because of its minimally invasive nature and favorable clinical outcomes [1, 2]. Despite these advantages, a subset of patients still experience persistent postoperative low back pain, which may adversely affect functional recovery and patient satisfaction [3, 4].

The mechanisms underlying postoperative low back pain after UBE remain incompletely understood. Previous studies have mainly focused on paraspinal muscle injury, facet joint violation, and the extent of decompression [5–7], whereas the role of the posterior ligamentous complex has received limited attention. As an important component of the posterior tension band, the interspinous ligament contributes to segmental stability [8, 9], and surgical irritation or soft-tissue trauma in this region may contribute to postoperative inflammatory responses and subsequent mechanical low back pain [10].

With the routine use of early postoperative magnetic resonance imaging (MRI), edema-like signal changes are commonly observed in the interspinous ligament region, indicative of the postoperative soft tissue reaction [11]. However, the clinical relevance of such imaging findings has not been well established, particularly in a relatively homogeneous population such as patients with single-level LDH.

Therefore, this study aimed to investigate the association between MRI-detected interspinous ligament region edema (ISLRE) and postoperative low back pain and disability in patients undergoing single-level UBE surgery for LDH, with a follow-up period of 1 year.

Methods

Study Design and Population

This retrospective cohort study included consecutive patients who underwent single-level UBE surgery for LDH at a single center between October 2021 and October 2024.

Patients were eligible if they were aged 18 years or older, were diagnosed with single-level LDH and treated with UBE surgery, had preoperative lumbar MRI within 1 month before surgery as well as early postoperative MRI within 3 days after surgery, and completed clinical follow-up for at least 12 months, with available pain and functional outcome data. All included patients underwent a standardized UBE procedure for single-level LDH. The surgical technique was limited to UBE-assisted lumbar discectomy with decompression as required. No patients underwent additional procedures such as spinal fusion, instrumentation, deformity correction, or other adjunctive surgical interventions. Therefore, all cases represented a homogeneous surgical cohort. All surgical procedures were performed by three experienced spine surgeons, each of whom had completed more than 50 UBE procedures prior to the study period, ensuring a stable level of surgical proficiency beyond the initial learning curve. Patients were excluded if they had previous surgery at the index level, spinal infection, tumor, fracture, or deformity, underwent fusion surgery, experienced severe perioperative complications affecting postoperative pain assessment, or had poor-quality MRI precluding reliable evaluation of the interspinous ligament. Patients showing preoperative interspinous ligament region high signal intensity on MRI were also excluded to avoid confounding from preexisting interspinous ligament abnormalities. No imputation for missing data was performed. A complete-case analysis was conducted, and only patients with complete follow-up data were included in the final analysis.

Ethical Approval

The study protocol was approved by the Ethics Committee of Quzhou People’s Hospital (approval no. 2025-064) and was conducted in accordance with the principles of the Declaration of Helsinki. The requirement for informed consent was waived owing to the retrospective nature of the study.

MRI Acquisition and Assessment of ISLRE

All patients underwent standard lumbar MRI examinations preoperatively and within 3 days postoperatively. The imaging protocol included sagittal T1-weighted, T2-weighted, and fat-suppressed T2-weighted or short tau inversion recovery sequences, as well as axial T2-weighted images. ISLRE at the operated level was evaluated primarily on early postoperative sagittal T2-weighted and fat-suppressed sequences, with axial images used as supplementary references when necessary.

The severity of ISLRE was classified into three grades. The grading system was based on the extent and distribution pattern of hyperintense edema within the operated interspinous region. Grade 0 was defined as absent edema or only minimal punctate or linear hyperintensity without a definite edematous area. Grade 1 was defined as focal or segmental hyperintensity involving a limited portion of the operated interspinous region. Grade 2 was defined as diffuse or confluent hyperintensity involving most of the operated interspinous region (Fig. 1).

Fig. 1.

Magnetic resonance imaging (MRI)-based grading of postoperative interspinous ligament region edema (ISLRE). Representative early postoperative sagittal fat-suppressed T2-weighted MRI images illustrating the semiquantitative grading system at the operated level. Grade 0 shows absent or minimal punctate/linear hyperintensity without a definite edematous area. Grade 1 shows focal or segmental hyperintensity involving a limited portion of the operated interspinous region. Grade 2 shows diffuse or confluent hyperintensity involving most of the operated interspinous region. Arrows indicate representative postoperative edema

All MRI scans were independently reviewed by two experienced spine surgeons who were blinded to the clinical outcomes. Interobserver agreement was assessed using Cohen’s kappa coefficient, and discrepancies were resolved by consensus.

Clinical Outcome Assessment

Clinical outcomes were evaluated using the visual analogue scale for low back pain (VAS-LBP) and the Oswestry disability index (ODI). Assessments were performed preoperatively and at 3 days, 1 month, 6 months, and 12 months after surgery. Higher VAS-LBP scores indicated more severe low back pain, and higher ODI scores indicated greater functional disability.

Persistent postoperative low back pain was defined as a VAS-LBP score of 3 or higher at 12 months.

Covariates

A comprehensive set of clinically relevant covariates was included in the multivariable regression models to minimize potential confounding, including age, sex, BMI, smoking status, heavy physical labor, Modic changes, operative level, baseline VAS-LBP, baseline ODI, and operative time. Covariate selection was guided by clinical relevance and previously published literature regarding risk factors for postoperative low back pain.

Heavy physical labor was defined as sustained occupational exposure involving one or more of the following [12]: (1) frequent lifting or upward pulling of heavy objects, totaling several tonnes per day over multiple years; (2) recurrent handling of extremely heavy or awkward loads under similar conditions; (3) daily manual transfer of adults or older children with disabilities; or (4) prolonged daily exposure to whole-body vibration from operating heavy machinery.

Continuous covariates were entered into regression models as continuous variables, and their effects were quantified as regression coefficients in linear models or odds ratios (ORs) in logistic models. Categorical covariates were entered as indicator variables with predefined reference categories. Standardized regression coefficients were additionally calculated to compare the relative contribution of continuous covariates.

Statistical Analysis

Continuous variables were expressed as mean ± standard deviation or median with interquartile range, and categorical variables were expressed as frequencies and percentages. Baseline characteristics among ISLRE groups were compared using one-way analysis of variance or the Kruskal–Wallis test for continuous variables and the chi-square or Fisher’s exact test for categorical variables.

Differences in VAS-LBP and ODI scores among ISLRE groups at each postoperative time point were initially assessed using appropriate parametric or nonparametric tests. Multivariable linear regression analyses were performed to evaluate the independent association between ISLRE severity and VAS-LBP and ODI at 1 month, 6 months, and 12 months after surgery. Results were reported as regression coefficients with 95% confidence intervals. Multivariable logistic regression was performed to determine whether ISLRE was an independent risk factor for persistent low back pain at 12 months. Variance inflation factors (VIF) were calculated for all multivariable regression models to assess multicollinearity, and all values were below 2.0, indicating no significant collinearity among variables.

A linear mixed-effects model was used to evaluate longitudinal changes in postoperative VAS scores. Patient-specific random intercepts were included to account for within-subject correlation across repeated measurements. Time was treated as a categorical fixed effect (postoperative day 3, 1 month, 6 months, and 12 months, with day 3 as reference). Interactions between time and ISLRE grade were included. The model was adjusted for sex, age, BMI, smoking status, heavy manual labor, surgical level, preoperative VAS score, operative time, and Modic changes. Restricted maximum likelihood estimation was used.

The inverse probability of treatment weighting (IPTW) method was used to reclassify the patients into low-grade ISLRE (grade 0–1) and high-grade ISLRE (grade 2) groups for propensity score weighting, to reduce potential confounding factors. Covariate balance before and after weighting was assessed using standardized mean differences (SMD), with values < 0.10 indicating adequate balance. In addition, a sensitivity analysis was performed using a stricter definition of persistent pain (VAS > 4 at 12 months) to evaluate the robustness of the primary findings.

All statistical analyses were performed using IBM SPSS Statistics software (version 28.0), and a two-sided P-value < 0.05 was considered statistically significant.

Results

Patient Characteristics

A total of 245 patients were included in the final analysis, comprising 117 patients with grade 0 ISLRE, 74 with grade 1, and 54 with grade 2 (Fig. 2). Baseline demographic and clinical characteristics are summarized in Table 1. Baseline demographic, clinical, and radiological characteristics were generally comparable among the three groups.

Fig. 2.

Flowchart of patient selection. Of 336 consecutive patients undergoing unilateral biportal endoscopic (UBE) surgery, 91 were excluded, including patients with prior surgery at the index level (n = 15), fusion surgery (n = 16), incomplete follow-up (n = 27), poor-quality MRI precluding reliable assessment (n = 24) and others (n = 9). A total of 245 patients were included in the final cohort and classified into grade 0 (n = 117), grade 1 (n = 74), and grade 2 (n = 54) interspinous ligament region edema (ISLRE) groups

Table 1.

Baseline characteristics according to interspinous ligament region edema grade

Variable	Grade 0 (n = 117)	Grade 1 (n = 74)	Grade 2 (n = 54)	P-value
Age, years	57.34 ± 11.87	58.97 ± 13.19	59.98 ± 14.60	0.420
Male sex, n (%)	77 (65.8)	49 (66.2)	29 (53.7)	0.256
BMI, kg/m2	23.79 ± 3.66	24.59 ± 3.37	25.03 ± 3.61	0.079
Smoking, n (%)	30 (25.6)	22 (29.7)	12 (22.2)	0.625
Heavy labor, n (%)	48 (41.0)	35 (47.3)	27 (50.0)	0.484
L4/5, n (%)	52 (44.4)	46 (62.2)	32 (59.3)	0.068
L5/S1, n (%)	58 (49.6)	22 (29.7)	18 (33.3)
Other level, n (%)	7 (6.0)	6 (8.1)	4 (7.4)
Symptom duration, months	18.95 ± 22.92	14.12 ± 19.25	17.85 ± 21.66	0.315
Hypertension, n (%)	24 (20.5)	18 (24.3)	16 (29.6)	0.422
Diabetes, n (%)	9 (7.7)	5 (6.8)	7 (13.0)	0.416
Bone-density abnormality, n (%)	22 (18.8)	16 (21.6)	16 (29.6)	0.282
Baseline VAS-LBP	6.43 ± 0.98	6.55 ± 1.01	6.41 ± 1.06	0.632
Baseline ODI	56.53 ± 9.91	59.50 ± 9.24	58.37 ± 11.73	0.132
Operative time, min	114.98 ± 32.43	112.24 ± 30.05	116.52 ± 31.15	0.730
Modic changes, n (%)	30 (25.6)	17 (23.0)	18 (33.3)	0.405

BMI, body mass index; VAS-LBP, visual analogue scale for low back pain; ODI, Oswestry disability index

Interobserver agreement for MRI-based grading of ISLRE was substantial, with a Cohen’s kappa coefficient of 0.802.

Association Between ISLRE Severity and Postoperative Clinical Outcomes

Postoperative low back pain outcomes differed significantly according to ISLRE severity. Although a modest unadjusted difference was observed at postoperative day 3 (P = 0.030), this early difference was small and clinically less pronounced than differences emerging from 1 month onward (all P < 0.001).

At 12 months, mean VAS-LBP scores were 1.66 ± 1.20 in grade 0, 2.34 ± 1.45 in grade 1, and 3.54 ± 1.65 in grade 2 patients (P < 0.001). Persistent postoperative low back pain (VAS-LBP > 3) occurred in 21.4% of grade 0 patients, 41.9% of grade 1 patients, and 68.5% of grade 2 patients (P < 0.001). The between-group differences in VAS-LBP exceeded the minimal clinically important difference (MCID) threshold for postoperative low back pain (1.2 points) [13], indicating clinically meaningful pain disparities between ISLRE grades.

ODI differences were not statistically significant at 1 month (P = 0.053) but became significant at 6 and 12 months (P = 0.001 and P < 0.001, respectively) (Table 2). However, although statistically significant, the magnitude of ODI differences at 12 months was below the MCID threshold (12.8 points) [13], suggesting limited clinical relevance.

Table 2.

Postoperative outcomes according to edema grade

Outcome	Grade 0	Grade 1	Grade 2	P-value
POD3 VAS-LBP	3.27 ± 0.88	3.24 ± 0.93	3.63 ± 0.90	0.030
1-month VAS-LBP	2.20 ± 1.13	2.68 ± 1.23	4.24 ± 1.32	< 0.001
6-month VAS-LBP	1.89 ± 1.09	2.41 ± 1.46	4.00 ± 1.39	< 0.001
12-month VAS-LBP	1.66 ± 1.20	2.34 ± 1.45	3.54 ± 1.65	< 0.001
1-month ODI	30.01 ± 2.95	30.91 ± 3.02	31.06 ± 3.49	0.053
6-month ODI	19.57 ± 2.52	20.58 ± 2.61	21.00 ± 2.82	0.001
12-month ODI	12.03 ± 2.39	12.62 ± 2.66	13.74 ± 2.69	< 0.001
Persistent LBP, n (%)	25 (21.4)	31 (41.9)	37 (68.5)	< 0.001

VAS-LBP, visual analogue scale for low back pain; ODI, Oswestry disability index; POD3, postoperative day 3

Multivariable Linear Regression Analyses

After adjustment for potential confounders, including age, BMI, heavy labor, Modic changes, baseline VAS-LBP, baseline ODI, and operative time, ISLRE severity remained independently associated with worse postoperative pain outcomes.

Compared with grade 0 ISLRE, grade 1 ISLRE was associated with higher 12-month VAS-LBP scores (β = 0.67, 95% confidence interval [CI] 0.27–1.07; P = 0.001), whereas grade 2 ISLRE showed a stronger independent association (β = 1.78, 95% CI 1.34–2.22; P < 0.001) (Table 3). Heavy labor (β = 0.65, P < 0.001) and Modic changes (β = 0.38, P = 0.040) were also independently associated with worse pain outcomes.

Table 3.

Multivariable linear regression for 12-month VAS-LBP

Variable	β	95% CI	P-value
Grade 1 ISLRE (versus grade 0)	0.67	0.27–1.07	0.001
Grade 2 ISLRE (versus grade 0)	1.78	1.34–2.22	< 0.001
Male sex	−0.30	−0.69 to 0.10	0.138
Age (years)	0.01	−0.001 to 0.03	0.060
BMI (kg/m2)	−0.02	−0.07 to 0.03	0.482
Smoking	0.03	−0.39 to 0.46	0.878
Heavy manual labor	0.65	0.30–1.00	< 0.001
Operated level (L5/S1 versus L4/5)	0.18	−0.21 to 0.56	0.372
Operated level (other versus L4/5)	−0.26	−0.95 to 0.43	0.455
Baseline VAS-LBP	0.03	−0.14 to 0.20	0.714
Operative time (min)	−0.002	−0.008 to 0.003	0.401
Modic changes	0.38	−0.02 to 0.77	0.040

R² = 0.32, adjusted R² = 0.28, F = 9.06, P < 0.001

VAS-LBP, visual analogue scale for low back pain; CI, confidence interval; ISLRE, interspinous ligament region edema; BMI, body mass index

Bold values indicate statistically significant results (P < 0.05)

For ODI, grade 2 ISLRE remained independently associated with worse disability at 12 months (β = 1.42, 95% CI 0.60–2.23; P = 0.001), whereas grade 1 ISLRE was not significantly associated with ODI (Table 4). The explanatory power of the VAS model (R² = 0.28) exceeded that of the ODI model (R² = 0.12), suggesting a stronger association between ISLRE and postoperative pain than functional disability.

Table 4.

Multivariable linear regression for 12-month ODI

Variable	β	95% CI	P-value
Grade 1 ISLRE (versus grade 0)	0.36	−0.38 to 1.10	0.341
Grade 2 ISLRE (versus grade 0)	1.42	0.60–2.23	0.001
Male sex	−0.48	−1.21 to 0.25	0.198
Age (years)	0.006	−0.02 to 0.03	0.685
BMI (kg/m2)	0.03	−0.06 to 0.12	0.551
Smoking	0.17	−0.62 to 0.96	0.674
Heavy manual labor	0.77	0.13–1.42	0.020
Operated level (L5/S1 versus L4/5)	−0.26	−0.97 to 0.45	0.474
Operated level (other versus L4/5)	−0.93	−2.20 to 0.34	0.150
Baseline ODI	0.04	0.005 to 0.07	0.022
Operative time (min)	−0.01	−0.02 to −0.001	0.032
Modic changes	0.67	−0.06 to 1.40	0.073

R² = 0.17, adjusted R² = 0.12, F = 3.85, P < 0.001

ODI, Oswestry disability index; CI, confidence interval; ISLRE, interspinous ligament region edema; BMI, body mass index

Bold values indicate statistically significant results (P < 0.05)

VIFs for all covariates included in the multivariable regression models were below 2.0, indicating no evidence of significant multicollinearity among variables.

Multivariable Logistic Regression for Persistent Postoperative Low Back Pain

In multivariable logistic regression analysis, ISLRE severity was independently associated with persistent postoperative low back pain at 12 months.

Compared with grade 0 ISLRE, grade 1 ISLRE was associated with a 3.12-fold increased risk of persistent postoperative low back pain (OR 3.12, 95% CI 1.51–6.44; P = 0.002), whereas grade 2 ISLRE was associated with more than eightfold higher odds (OR 8.90, 95% CI 4.01–19.76; P < 0.001) (Fig. 3, Table 5).

Fig. 3.

Multivariable logistic regression analysis for persistent postoperative low back pain at 12 months. Forest plot showing adjusted odds ratios (ORs) with 95% confidence intervals (CIs) derived from the multivariable logistic regression model. Persistent postoperative low back pain was defined as a VAS-LBP score > 3 at 12 months after surgery. The model demonstrated good discrimination (area under the receiver operating characteristic curve (AUC) 0.791, 95% CI 0.732–0.850), acceptable calibration (Hosmer–Lemeshow χ² = 2.93, degrees of freedom (df) 8, P = 0.939), and explained 32.6% of the outcome variance (Nagelkerke R² = 0.326). Model fit was significant (likelihood ratio χ² = 67.07, df 12, P < 0.001). OR, odds ratio; CI, confidence interval; VAS-LBP, visual analogue scale for low back pain; BMI, body mass index; AUC, area under the receiver operating characteristic curve

Table 5.

Multivariable logistic regression for persistent postoperative low back pain

Variable	Adjusted OR	95% CI	P-value
Grade 1 ISLRE (versus grade 0)	3.12	1.51–6.44	0.002
Grade 2 ISLRE (versus grade 0)	8.90	4.01–19.76	< 0.001
Male sex	0.58	0.29–1.16	0.122
Age (years)	0.98	0.96–1.01	0.163
BMI (kg/m2)	1.01	0.93–1.11	0.774
Smoking	0.87	0.39–1.93	0.729
Heavy manual labor	2.86	1.54–5.32	0.001
Operated level (L5/S1 versus L4/5)	1.55	0.77–3.11	0.221
Operated level (other versus L4/5)	0.84	0.23–3.13	0.798
Baseline VAS-LBP	1.03	0.76–1.40	0.841
Operative time (min)	1.00	0.99–1.01	0.490
Modic changes	2.18	1.10–4.31	0.025

Model performance: The multivariable logistic regression model demonstrated good overall performance (likelihood ratio χ² = 67.07, df 12, P < 0.001), with a Nagelkerke R² of 0.326, AUC of 0.791 (95% CI 0.732–0.850), and good calibration according to the Hosmer–Lemeshow goodness-of-fit test (χ² = 2.93, df 8, P = 0.939)

OR, odds ratio; CI, confidence interval; ISLRE, interspinous ligament region edema; BMI, body mass index; VAS-LBP, visual analogue scale for low back pain; AUC, area under the curve

Bold values indicate statistically significant results (P < 0.05)

Heavy labor (OR 2.86, 95% CI 1.54–5.32; P = 0.001) and Modic changes (OR 2.18, 95% CI 1.10–4.31; P = 0.025) were also identified as independent risk factors. The multivariable logistic regression model demonstrated good overall performance (likelihood ratio χ² = 67.07, df 12, P < 0.001), with a Nagelkerke R² of 0.33, AUC of 0.79 (95% CI 0.73–0.85), and good calibration according to the Hosmer–Lemeshow goodness-of-fit test (χ² = 2.93, df 8, P = 0.939).

Longitudinal Analysis of Recovery Trajectories

Linear mixed-effects modeling demonstrated a significant effect of time on VAS-LBP recovery (P < 0.001). Significant time-by-edema grade interactions were also observed, indicating differential recovery trajectories according to ISLRE severity (Table 6).

Table 6.

Adjusted linear mixed-effects model for postoperative VAS-LBP trajectory

Variable	β	95% CI	P
Time
1 month	−1.08	−1.31 to −0.85	< 0.001
6 months	−1.38	−1.62 to −1.15	< 0.001
1 year	−1.62	−1.85 to −1.38	< 0.001
Grade
Grade 1	−0.06	−0.40 to 0.29	0.750
Grade 2	0.27	−0.11 to 0.65	0.161
Time × grade
1 month × grade 1	0.51	0.14–0.88	0.007
6 months × grade 1	0.55	0.18–0.92	0.004
1 year × grade 1	0.71	0.34–1.08	< 0.001
1 month × grade 2	1.69	1.28–2.10	< 0.001
6 months × grade 2	1.75	1.34–2.17	< 0.001
1 year × grade 2	1.52	1.11–1.93	< 0.001
Adjusted covariates
Male sex	−0.28	−0.53 to −0.03	0.031
Age	0.006	−0.004 to 0.016	0.231
BMI	−0.004	−0.034 to 0.026	0.798
Smoking	0.11	−0.16 to 0.39	0.426
Heavy manual labor	0.33	0.10–0.56	0.004
Surgical level	−0.09	−0.27 to 0.09	0.332
Baseline VAS-LBP	0.005	−0.11 to 0.12	0.931
Operative time	−0.001	−0.004 to 0.002	0.488
Modic changes	0.22	−0.03 to 0.46	0.087

Data were analyzed using a linear mixed-effects model with patient-specific random intercepts. Time (postoperative day 3 as reference), interspinous ligament injury grade (grade 0 as reference), and their interaction were included as fixed effects. The model was additionally adjusted for age, sex, body mass index, smoking status, heavy manual labor, surgical level, preoperative VAS score, operative time, and Modic changes. A significant time × grade interaction was observed (χ² = 91.75, P < 0.001), indicating that postoperative pain trajectories differed significantly among injury grades

VAS-LBP, visual analogue scale for low back pain; CI, confidence interval; BMI, body mass index

Significant time-by-edema grade interactions were observed, indicating delayed recovery trajectories in grade 1 and more pronounced delay in grade 2 edema.

Patients with grade 0 or grade 1 edema demonstrated progressive postoperative improvement, whereas those with grade 2 ISLRE showed persistently impaired recovery throughout follow-up (Fig. 4).

Fig. 4.

Pain recovery trajectories by interspinous ligament region edema grade. Mean VAS-LBP scores with 95% confidence intervals at baseline, postoperative day 3, 1 month, 6 months, and 12 months. Overall between-group differences across follow-up were assessed using a linear mixed-effects model and were statistically significant (P < 0.001). Shaded areas indicate 95% confidence intervals. VAS-LBP, visual analogue scale for low back pain

Sensitivity Analysis

After IPTW adjustment (Supplementary Table S1), all baseline covariates achieved adequate balance (SMD < 0.10). The IPTW-weighted logistic regression confirmed the robustness of the primary analysis, showing that high-grade interspinous ligament region edema remained significantly associated with persistent postoperative low back pain (OR 4.23, 95% CI 2.24–8.01, P < 0.001). Sensitivity analysis using VAS > 4 demonstrated consistent results with the primary analysis, confirming the robustness of the association between ISLRE grade and persistent postoperative low back pain (Supplementary Table S2).

Discussion

In this retrospective cohort study, we found that postoperative ISLRE severity was independently associated with worse low back pain recovery after UBE surgery. Patients with grade 2 edema showed significantly higher postoperative pain scores, substantially delayed recovery trajectories, and more than eightfold increased odds of persistent postoperative low back pain at 12 months, even after adjustment for potential confounders. By contrast, the association between ligament edema severity and disability seemed to lack clinical significance. These findings suggest that iatrogenic irritation to the interspinous region may be associated with persistent postoperative mechanical low back pain.

Our findings extend prior studies focusing primarily on neural decompression and disc-related factors as determinants of postoperative outcomes after endoscopic lumbar surgery [2, 6]. While UBE offers several advantages over conventional open procedures, the potential effect of iatrogenic irritation to posterior stabilizing structures has received limited attention. The interspinous ligament is an important component of the posterior tension band and contributes to resisting flexion, maintaining segmental stability, and modulating load transfer [14]. More extensive postoperative interspinous edema may reflect greater soft-tissue trauma in the posterior tension-band region and could serve as an imaging marker of surgical impact on the posterior spinal elements, which may be associated with delayed pain recovery, even when neural decompression is adequate [8–10].

A notable finding of the present study was the graded relationship between ISLRE severity and postoperative pain outcomes. Although postoperative pain was comparable across groups immediately after surgery, differences became apparent from 1 month onward and persisted throughout follow-up. Mixed-effects modeling demonstrated divergent recovery trajectories, with patients with grade 2 edema showing sustained delayed pain recovery. This temporal pattern suggests that the extent of early postoperative ISLRE may reflect surgical tissue response, rather than directly indicating structural ligament disruption.

The magnitude of association observed for grade 2 edema was clinically meaningful. In multivariable analysis, grade 2 edema was associated with an average increase of 1.76 points in 12-month VAS-LBP. In addition, grade 2 edema conferred more than eightfold increased odds of persistent postoperative low back pain. Importantly, the magnitude of VAS-LBP differences between ISLRE grades exceeded the reported MCID for low back pain, suggesting that the observed associations are not only statistically significant but also clinically meaningful. Together, these findings indicate that high-grade ISLRE is not merely an imaging observation, but may be clinically relevant.

Interestingly, the association between ISLRE and pain outcomes appeared stronger than that for functional disability. While grade 2 edema was independently associated with ODI at 12 months, the effect size was smaller and the explanatory power of the ODI model was lower than that of the pain model. This discrepancy may suggest that posterior tension-band edema preferentially contributes to mechanical pain generation rather than broad functional impairment, which is likely influenced by multiple psychosocial and structural factors [15, 16].

Heavy physical labor and Modic changes were also identified as independent risk factors for persistent postoperative low back pain. These findings are biomechanically plausible, as repetitive loading and preexisting vertebral endplate degeneration may amplify stress concentration at surgically altered segments, potentially interacting with ligament edema in relation to postoperative pain [17, 18]. Importantly, however, the independent effect of grade 2 edema remained robust after adjustment for these factors, supporting an independent association with postoperative persistent low back pain.

Although postoperative ISLRE may reflect localized tissue response following UBE surgery, it is unlikely to represent a purely isolated structural ligament injury. Instead, ISLRE may serve as a composite imaging biomarker integrating multiple perioperative factors, including the extent of surgical manipulation, local inflammatory response, irrigation-related fluid diffusion, and patient-specific healing capacity.

Importantly, all patients in the present study underwent a standardized single-level UBE procedure; however, subtle differences in operative handling and tissue response may still contribute to variability in postoperative MRI findings. Therefore, ISLRE should be interpreted not as a direct causal pathological lesion, but as a postoperative radiological manifestation reflecting the individual tissue response to a standardized surgical environment.

Despite this interpretation, the observed association between higher-grade ISLRE and both delayed pain recovery and persistent postoperative low back pain remained robust after adjustment for multiple clinical and surgical covariates, as well as in propensity score-weighted sensitivity analyses. These findings suggest that ISLRE may have prognostic value as an imaging-based marker of postoperative pain trajectory, rather than merely representing surgical extent alone.

From a clinical perspective, these findings highlight the potential utility of early postoperative MRI for identifying patients with extensive interspinous edema, who may be at higher risk of delayed pain recovery. Careful surgical manipulation to minimize posterior soft tissue trauma and targeted postoperative monitoring or rehabilitation for patients with high-grade edema may help improve long-term outcomes [8, 9]. Patients with grade 2 edema may warrant closer follow-up and potentially targeted rehabilitation strategies.

This study has several strengths. It incorporated early postoperative MRI-based grading of ISLRE, longitudinal pain trajectory analysis, multivariable adjustment, and sensitivity analysis, providing convergent evidence for the robustness of the findings. In addition, substantial interobserver agreement supported the reliability of the imaging assessment.

Several limitations should also be acknowledged. First, this was a retrospective single-center study, which may introduce selection bias. Second, although all procedures were performed by three experienced spine surgeons, each having completed more than 50 UBE procedures prior to the study period, and all patients with incomplete follow-up were excluded with a complete-case analysis applied, potential bias related to the retrospective design and institutional learning curve cannot be fully eliminated. Third, although multiple confounders were adjusted for, residual confounding cannot be excluded. Fourth, postoperative interspinous edema, as assessed on early MRI, may not directly reflect structural ligament disruption, and the study identified only associations rather than causation. Future studies integrating biomechanical assessments or longitudinal imaging may further clarify the relationship between surgical tissue trauma and persistent pain [15]. Finally, external multicenter validation is needed before broader generalization.

Conclusions

The extent of postoperative ISLRE, particularly grade 2 edema, may be useful for postoperative risk stratification and targeted patient management after UBE surgery.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

This study was conceptualized and initiated by Yi Mao, who also contributed to the investigation and drafted the original manuscript alongside Jiangbao Xu. Jiangbao Xu performed the formal analysis and participated in the investigation and initial writing. Junchao Zhang and Lei Tong curated the data, provided essential resources, and contributed to manuscript review and editing. Xiangyang Wang played a comprehensive role in conceptualization, investigation, data curation, and resource provision, and significantly revised and refined the manuscript. All named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published.

Funding

No funding or sponsorship was received for this study or publication of this article. The Rapid Service Fee was funded by the authors.

Data Availability

The datasets generated and/or analyzed during the current study are not publicly available owing to patient privacy protection and institutional data sharing policies, but are available from the corresponding author on reasonable request and with appropriate ethical approval.

Declarations

Conflict of Interest

Yi Mao, Jiangbao Xu, Junchao Zhang, Lei Tong, and Xiangyang Wang declare no conflicts of interest.

Ethics Approval

The study protocol was approved by the Ethics Committee of Quzhou People’s Hospital (approval no. 2025-064) and was conducted in accordance with the principles of the Declaration of Helsinki. The requirement for informed consent was waived owing to the retrospective nature of the study.