Risk Factor Analysis of Residual Low Back Pain After Unilateral Biportal Endoscopic Discectomy in Patients With Lumbar Disc Herniation

Introduction

Lumbar disc herniation (LDH) is a common cause of low back pain (LBP) and sciatica, affecting approximately 1% to 5% of the general population annually [1]. Its pathophysiology involves rupture of the annulus fibrosus of the intervertebral disc, with subsequent herniation of the nucleus pulposus. This process leads to compression of the spinal nerves and cauda equina, triggers an inflammatory response, and results in LBP and lower extremity neurological dysfunction [2]. Generally, for patients with LDH-related clinical symptoms and signs that are consistent with imaging findings and who show no clinically significant improvement after 6 weeks of conservative treatment, surgical intervention is recommended [1]. Open fenestration discectomy has long been considered the standard surgical procedure for LDH, and its efficacy has been validated through extensive clinical practice [3]. However, open fenestration discectomy requires extensive dissection of the paravertebral muscles and large-scale resection of the lamina and facet joints, which substantially increases the incidence of adverse outcomes, including postoperative chronic back pain and lumbar spine degeneration [4]. Unilateral biportal endoscopy (UBE) is a novel surgical treatment for LDH, and recent studies have shown satisfactory outcomes with this approach [5–8]. The advantages of UBE include maximal preservation of the lamina and facet joints, minimal muscle injury, and considerable improvement in neurological symptoms immediately after surgery [9], making it the preferred surgical approach for many spine surgeons involved in LDH management. Despite its capacity to minimize damage to lumbar anatomical structures, residual low back pain (rLBP) may occur postoperatively. A previous study indicated that the proportions of patients with rLBP at 6–24 months and beyond 24 months after lumbar discectomy range from 3% to 34% and 5% to 36%, respectively [10]. Although substantial research has addressed the application of UBE in LDH, studies focusing on postoperative rLBP remain limited. The present study aimed to systematically analyze preoperative clinical indicators and imaging characteristics in patients with and without rLBP after UBE, identify risk factors associated with rLBP, and characterize high-risk populations. The findings may provide a reference for the development of individualized treatment strategies and selection of optimal surgical approaches.

Material and Methods

ETHICAL APPROVAL AND CONSENT TO PARTICIPATE:

This study was conducted in accordance with the ethical standards of the Declaration of Helsinki (1975, revised in 2008). Ethical approval (Approval No. 2025-002) was granted by the Ethics Committee of Jiangxi Province Hospital of Integrated Chinese and Western Medicine. All patients provided written informed consent prior to participation in the study.

PARTICIPANTS:

This study used a single-center, nonrandomized, retrospective case–control design. Clinical data from 203 consecutive patients with LDH who underwent UBE at Jiangxi Hospital of Integrated Traditional Chinese and Western Medicine between January 2020 and August 2024 were retrospectively collected. The analysis encompassed demographic characteristics, clinical outcomes, and imaging findings. During the evaluation of demographic characteristics, clinical outcomes, and radiological parameters, complete or partial blinding was utilized to reduce potential bias.

INCLUSION CRITERIA:

Patients were included if they (1) had a confirmed diagnosis of LDH; (2) received standard conservative treatment for more than 6 weeks without symptom improvement; (3) displayed single-level LDH that was treated by UBE; (4) had complete clinical data available; and (5) had at least 1 year of postoperative follow-up data.

EXCLUSION CRITERIA:

Patients were excluded if they (1) displayed lumbar spondylolisthesis, lumbar tumors, lumbar infections, or ankylosing spondylitis; (2) had a history of lumbar surgery; (3) were readmitted for recurrent LDH; and/or (4) did not complete at least 1 year of follow-up.

COLLECTION OF DEMOGRAPHIC CHARACTERISTICS AND CLINICAL OUTCOMES:

Demographic and clinical variables included age (years), sex (male/female), disease duration (months), body mass index (BMI, kg/m²), smoking status (yes/no), and alcohol consumption (yes/no). These data were collected and documented by a qualified clinical researcher (ZFC). Another qualified clinical researcher (TT) assessed and recorded LBP and leg pain (LP) visual analog scale (VAS) scores, as well as Oswestry Disability Index (ODI) scores, preoperatively and at 1 year postoperatively. MacNab scale outcomes were recorded at the final follow-up. Severe preoperative LBP was defined as a preoperative LBP VAS score of 8 or higher. Disease duration was defined as the interval from the onset of LDH-related symptoms to the date of surgery. Smoking was defined as consumption of at least 1 cigarette per day, and alcohol consumption was defined as alcohol intake at least 4 times per month.

ASSESSMENT OF RADIOLOGICAL PARAMETERS:

To ensure accuracy, radiological parameters were independently evaluated and recorded by 2 experienced spine surgeons (JFH and HNL). Assessed parameters were the disc height index (DHI) (Figure 1), lumbar lordosis (LL), sacral slope (SS), segmental range of motion (sROM) (Figure 2), degenerative scoliosis (coronal Cobb angle ≥10 degrees), intradiscal vacuum phenomenon (IVP), disc degeneration grade (Pfirrmann classification), lumbar facet joint osteoarthritis (FJOA), evaluated using the Framingham scale [11] (Figure 3), and Modic changes (MC) (Figure 4). In this study, FJOA was categorized as low grade (Grades I and II) or high grade (Grades III and IV). Pfirrmann Grades I to III were defined as low-level disc degeneration; Grades IV and V were defined as high-level disc degeneration. Prior to formal radiological assessment, 25% of the cases were randomly selected for evaluation by both surgeons to determine interobserver reliability using intraclass correlation coefficients and kappa statistics. Intraclass correlation coefficients for all continuous variables were 0.86 or greater, and kappa values for all categorical variables were 0.76 or greater, indicating excellent reliability. All radiological parameters were assessed in strict accordance with the criteria illustrated in Figures 1–4, ensuring methodological consistency for subsequent analyses.

SURGICAL PROCEDURE:

All patients underwent UBE performed by an experienced spine surgeon (BH). Patients were placed in the prone position on a radiolucent operating table, and fluoroscopy with a C-arm was used to confirm the target level. The working and viewing portals were identified; 2 longitudinal skin incisions (approximately 0.8–1.5 cm in length) were then created. After incision of the skin, subcutaneous tissue, and deep fascia, surgical instruments and the endoscope were introduced through the respective portals. Radiofrequency ablation was used to cauterize soft tissue and establish the working space. After the junction between the base of the spinous process and the inferior margin of the lamina of the upper vertebra had been reached, fluoroscopy was repeated to verify the surgical field. Subsequently, partial laminectomy was performed. Rongeurs and osteotomes were used to resect portions of the inferior articular process and lamina of the upper lumbar vertebra, as well as the superior articular process of the lower lumbar vertebra, until the cranial and caudal ligamenta flava had been exposed. The ligamenta flava were removed to expose the dural sac; the traversing and exiting nerve roots were fully decompressed. The herniated nucleus pulposus was removed using nucleus pulposus forceps, and annulus fibrosus suturing was performed when necessary. Adequate nerve root decompression and dural pulsation were confirmed, after which the incisions were closed (Figure 5).

COMPARISON OF CHARACTERISTICS BETWEEN GROUPS:

Patients with an LBP VAS score of 3 or higher at 1 year after UBE were assigned to the rLBP group, whereas those with a score below 3 were assigned to the non-rLBP group. Demographic variables, including age, sex, disease duration, BMI, alcohol consumption, and smoking status, were compared between groups. Clinical outcomes were also compared, including LBP VAS, LP VAS, and ODI scores obtained preoperatively and at 1 year postoperatively, as well as MacNab scale results at the final follow-up. Finally, radiological parameters were compared between groups, including DHI, LL, SS, sROM, degenerative scoliosis, IVP, lumbar disc degeneration grade, FJOA grade, and MC.

STATISTICAL ANALYSIS:

Statistical analyses were conducted using SPSS version 26.0 (SPSS Inc., Chicago, IL, USA). Post hoc power analysis was performed using G*Power version 3.1.9.7 to verify sample size adequacy. Categorical variables are presented as counts and percentages; they were analyzed using the chi-square test or Fisher’s exact test, as appropriate. For continuous variables, the Kolmogorov-Smirnov test was utilized to assess data distribution. The independent-samples t-test or the Mann-Whitney U test was then used, as appropriate. Results are expressed as mean±standard deviation or as median with interquartile range. To identify independent predictors of rLBP after UBE, univariate analyses and backward stepwise multivariate logistic regression analyses were performed, with rLBP as the dependent variable. Results are reported as odds ratios (ORs) with 95% confidence intervals (95% CIs). All statistical tests were 2-tailed, and P-values <0.05 were considered statistically significant. To control for type I error associated with multiple comparisons, Bonferroni correction was applied.

Results

DEMOGRAPHIC CHARACTERISTICS:

During the study period, 269 patients met the inclusion criteria. After exclusion of 14 patients with recurrent LDH, 6 patients with old lumbar fractures, 11 patients with follow-up durations of less than 12 months, 25 patients lost to follow-up, and 10 patients with incomplete radiological data, 203 patients were included in the analysis. Of the included patients, 44 and 159 were in the rLBP and non-rLBP groups, respectively. A statistically significant difference in disease duration was observed between the groups (P=0.048). No significant differences were identified between groups with respect to age, sex, BMI, alcohol consumption, or smoking status (Table 1).

CLINICAL OUTCOMES:

Clinical outcomes are summarized in Table 1. Compared with preoperative values, LBP VAS, LP VAS, and ODI scores in both groups were significantly improved at the 1-year postoperative follow-up. Preoperatively, LBP VAS scores were significantly higher in the rLBP group than in the non-rLBP group (P<0.001). At the 1-year follow-up, both LBP VAS and LP VAS scores remained higher in the rLBP group than in the non-rLBP group. The proportions of excellent and good outcomes according to the MacNab scale at the final follow-up were 70.5% (31/44) in the rLBP group and 96.9% (154/159) in the non-rLBP group; the difference was statistically significant. No statistically significant differences were observed between the 2 groups in preoperative LP VAS scores, preoperative ODI scores, or ODI scores at the 1-year follow-up.

RADIOLOGICAL OUTCOMES:

Radiological outcomes are presented in Table 2. The proportion of patients with high-grade FJOA was significantly higher in the rLBP group (43.2%) than in the non-rLBP group (22.6%) (P=0.007). Modic type 1 changes were also more frequent in the rLBP group (22.7%) than in the non-rLBP group (6.9%) (P=0.006). No statistically significant differences were identified between groups for DHI, LL, SS, sROM, degenerative scoliosis, IVP, or lumbar disc degeneration.

UNIVARIATE AND MULTIVARIATE LOGISTIC REGRESSION ANALYSIS:

Univariate logistic regression analysis showed that severe preoperative LBP, high-grade FJOA, and Modic type 1 changes were associated with postoperative rLBP (Table 3). Disease duration was evaluated but did not show a significant association. Variables with P-values <0.05 in univariate analysis were subsequently included in the multivariate logistic regression model. Multivariate analysis identified severe preoperative LBP (P<0.001; OR, 4.882; 95% CI, 1.920–12.412), high-grade FJOA (P=0.005; OR, 2.910; 95% CI, 1.374–6.161), and Modic type 1 changes (P=0.04; OR, 2.864; 95% CI, 1.047–7.834) as independent risk factors for postoperative rLBP among patients with LDH who underwent UBE (Table 4).

Discussion

LDH is a common cause of LBP and disability worldwide, affecting a substantial proportion of the population [12]. Surgical intervention is generally recommended when conservative treatment fails or when severe neurological deficits develop [13]. UBE is a minimally invasive surgical technique characterized by separate working and viewing portals, which provide enhanced intraoperative visualization. This feature enables thorough neural decompression while minimizing soft tissue injury [14]. Previous studies have indicated that, compared with percutaneous endoscopic lumbar discectomy (PELD), UBE is associated with a lower recurrence rate in the treatment of LDH [15], and it has thus become an increasingly preferred surgical option for patients with LDH [16]. However, rLBP after UBE for LDH remains relatively common [17], and its underlying mechanisms require further investigation. The occurrence of rLBP is multifactorial and may involve lumbar musculature, osseous and articular structures, intervertebral discs, neural elements, psychological factors, and various spine-related disorders [18,19]. The present findings indicate that UBE, as a minimally invasive procedure, is effective in alleviating lower extremity radicular pain caused by LDH [20,21]. Nevertheless, some patients continue to experience postoperative rLBP; 44 cases were identified in this cohort. Whereas previous studies have focused on rLBP after PELD or microdiscectomy, the present study specifically evaluated predictors of rLBP after UBE. The aim was to facilitate reduction of rLBP incidence through improved preoperative assessment and optimized intraoperative and postoperative management strategies.

FJOA is among the most prevalent degenerative spinal disorders and constitutes a major cause of LBP among older adults [22]. To date, most literature has focused on intervertebral disc pathology as the primary source of LBP, with relative underestimation of pain attributable to FJOA [23]. Previous research has shown FJOA prevalences of 59.6% in men and 66.7% in women [24]. Facet joints play a critical role in maintaining spinal stability, and degeneration of these joints results in reduced spinal stability [25], which may lead to LBP [26,27]. According to published reports, the proportion of LBP attributed to FJOA ranges from approximately 15% to 45% [22,28]. Suri et al [29] demonstrated that the proportion of patients who exhibit severe FJOA is higher among those with rLBP than among individuals without rLBP, highlighting the important contribution of FJOA to LBP. In the present study, high-grade FJOA was similarly identified as an independent risk factor for postoperative rLBP. For instance, FJOA-associated inflammation may irritate adjacent neural structures, resulting in localized LBP [22]. Advanced FJOA may cause segmental lumbar instability, which can also provoke LBP [30]. Additionally, during UBE, partial resection of the facet joints and laminae is often required to achieve adequate nerve root decompression. Such resection may further reduce lumbar spinal stability [31], and long-term progressive instability may contribute to LBP onset. Furthermore, degeneration of the lumbar facet joints has been associated with multifidus muscle atrophy [32], which may exacerbate lumbar instability and further accelerate facet joint degeneration. Izeki et al [33] reported that patients with lumbar FJOA may develop spontaneous facet joint fusion after percutaneous pedicle screw fixation, which substantially improves local spinal stability. Therefore, in patients with severe lumbar facet joint degeneration, fusion surgery may represent a more appropriate therapeutic option.

The present study indicates that severe preoperative LBP is an independent risk factor for postoperative rLBP. Kleinstück et al [34] reported that greater preoperative LBP severity is associated with worse clinical outcomes after decompression surgery. Taiji et al [35] demonstrated that, among patients undergoing surgery for lumbar spinal stenosis, a preoperative LBP VAS score greater than 7 could predict postoperative rLBP. Additionally, Zhu et al [17] reported that each 1-point increase in preoperative LBP VAS score was associated with an approximately 28.9% increase in postoperative rLBP incidence. Collectively, these findings support a clear association between elevated preoperative LBP VAS scores and postoperative rLBP. In the present study, 44 patients experienced persistent rLBP after UBE, and approximately one-third of these patients had a preoperative LBP VAS score exceeding 8. This observation reinforces the relationship between higher preoperative LBP VAS scores and postoperative rLBP. Complex pathophysiological mechanisms contribute to LBP. Higher preoperative LBP VAS scores have been associated with factors such as FJOA [36,37], degeneration of the cartilaginous endplates [38], and fatty infiltration of the paraspinal muscles [39,40]. Notably, these pathological factors often cannot be adequately addressed during surgery. In the early post-discharge period, oral nonsteroidal anti-inflammatory drugs are commonly prescribed to alleviate surgery-related pain; however, such treatment may mask the early manifestation of postoperative rLBP. After discontinuation of these medications, insufficient control of local inflammation may result in gradual rLBP onset. Therefore, among patients with high preoperative LBP VAS scores, comprehensive preoperative evaluation should be performed to identify potential sources of pain, and timely postoperative interventions should be implemented. Such measures are essential to improve postoperative outcomes and patient satisfaction.

MC, commonly observed on magnetic resonance imaging in degenerative spinal diseases, are classified into 3 types: Modic type 1, Modic type 2, and Modic type 3 changes [41]. Among these, Modic type 1 changes are believed to reflect bone marrow edema and inflammatory processes [42]. Increasing evidence indicates that Modic type 1 changes in adjacent vertebral bodies play a prominent role in LBP, and primary nociceptive innervation is localized to the vertebral endplate region [43,44]. In the present study, we identified an association between preoperative Modic type 1 changes and postoperative rLBP. Previous studies have shown that MC are highly prevalent among patients with LBP [45], that LBP accompanied by MC is characterized by more frequent and prolonged episodes [45–48], and that pain severity is correlated with the extent of MC lesions [49]. Zhang et al [50] reported that basivertebral nerve ablation may exert a beneficial therapeutic effect on rLBP associated with MC. Based on these findings, basivertebral nerve ablation may represent a potential treatment for MC-related LBP, warranting further investigation. Additionally, MC are regarded as indicators of cartilaginous endplate (CEP) injury. Persistent mechanical stimulation, cumulative damage, and other contributing factors may collectively lead to the development of symptomatic MC [41]. Feng et al [51] suggested that CEP degeneration represents the underlying pathological basis of MC. During CEP degeneration, various inflammatory mediators are produced [52–54], which may irritate adjacent neural structures and result in persistent postoperative LBP [55]. Zhao et al [56] further demonstrated that CEP degeneration strongly affects early postoperative pain and functional recovery in patients with LDH. Therefore, in future diagnostic and therapeutic strategies, greater attention should be directed to CEP degeneration, given its potential associations with MC and postoperative rLBP.

This study has some limitations. First, the follow-up duration was relatively short; longer follow-up periods are required in future studies to validate the present findings. Second, the single-center retrospective design is subject to the inherent limitations associated with such analyses. Furthermore, variations in surgical techniques and postoperative management compared with other institutions may have influenced the results. The findings should be confirmed through large-scale, multicenter prospective studies.

Conclusions

Residual symptoms after UBE are common. High-grade FJOA, severe preoperative LBP, and Modic type 1 changes were identified as risk factors for rLBP. Surgical intervention should be undertaken only after comprehensive informed consent has been obtained, with the goal of avoiding misunderstandings or unrealistic expectations regarding postoperative outcomes.