Introduction
Low back pain (LBP) with or without radicular symptoms caused by the herniated nucleus pulposus is common in patients aged 40–80 years [1]. The recommended treatments include bed rest, medications, physiotherapy, pain intervention, and surgery. However, in this condition, the pain pathophysiology is more likely from inflammation than from nerve root compression [2]. The high-severity subtypes of spinal pathology are associated with severe inflammation and macrophage activation, leading to severe pain. Recent evidence has focused on the role of macrophages and monocytes in the herniated disc in matrix remodeling and neovascularization, resulting in spontaneous disc regression [2]. This phenomenon highlights the role of inflammation in lumbar disc herniation.
A lumbar epidural steroid injection (ESI) involves injecting steroids into the epidural space close to the targeted inflamed nerve root or herniated disc to reduce nerve root inflammation and relieve pain. The effect of ESI ranges from 29% to 74% [3,4]. Currently, no standard clinical or radiographic parameters have been established to predict favorable responses to ESI in patients diagnosed with lumbosacral radiculopathy. Some patients with pain caused by nerve compression or spinal canal stenosis may benefit less from the injection [5]. According to the effectiveness level of ESI for the treatment of LBP with sciatica, the predictors of a good response to treatment become essential and reduce the risk of steroid exposure. Previous studies have evaluated the physical and biological factors related to the favorable response from ESI. These included pain characteristics, physical examination, magnetic resonance imaging (MRI), nerve conduction study, electromyography, and biochemical markers. Several studies have reported acute onset of symptoms, decreased posterior disc height, positive sharp waves, and fibrillation found in an electrodiagnostic study as predictors of the response to ESI [6,7].
This prospective analytical study aimed to investigate the clinical and radiographic predictive factors for favorable outcomes of lumbar ESI in patients with lumbosacral radiculopathy.
Materials and Methods
Study design and patients
This study was approved by the institutional review board (IRB) of the Faculty of Medicine of Chulalongkorn University, Bangkok, Thailand (IRB no., 321/63). Patients aged at least 20 years with a history of LBP and radicular pain and diagnosed with lumbosacral radiculopathy were recruited from August 2020 to December 2022. Transforaminal ESI (TFESI) was considered a treatment option for eligible patients whose conservative treatments failed, including medications, bed rest, activity modifications, and physical therapy. Failed conservative treatment was defined by reporting a pain score upon movement of at least 4 out of 10 assessed by a Numeric Rating Scale (NRS 0–10). Radicular pain was described as neuropathic pain, following a distinct dermatomal pattern associated with the involvement of one or two lumbar nerve roots and corresponding to MRI findings of disc herniation or degenerative spinal stenosis. The exclusion criteria were as follows: previous lumbar spine surgery, trauma, infection, tumor, progressive or profound motor weakness (motor power grade 0–II on physical examination), and cauda equina syndrome. Patients diagnosed with systematic inflammatory diseases (rheumatoid arthritis, inflammatory arthritis, ankylosing spondylitis, etc.), allergy to medications, depression, current use of oral steroid medications, and history of ESI in the last 3 months. Individuals experiencing depression, as determined using the hospital anxiety and depression scale questionnaire, utilizing a threshold of at least 11, were also excluded from the study.
After informed consent was obtained, all eligible patients received TFESI under fluoroscopic guidance by a pain specialist with 10 years of experience on an outpatient basis. All patients were positioned prone with standard monitoring and no sedation. The procedure was performed under local anesthesia using a 22G Quincke spinal needle. The needle was gently advanced to the targeted epidural space via the subpedicular approach in the oblique view, and the final position of the needle was checked in the lateral view. In the anteroposterior view, the contrast medium (Omnipaque; GE Healthcare, Chicago, IL, USA) was injected through the needle to confirm proper needle placement, which illustrated the spreading of the contrast medium into the epidural space with no vascular or intrathecal pattern. The 5.0 mg preservative-free dexamethasone mixture in 1.5 mL of 2.0% lidocaine and 0.5 mL of 0.5% bupivacaine was injected into the indexed epidural level. A single TFESI was administered, and its effect was observed until it relieved pain by <50.0%. All patients were asked to rate their pain score using the NRS 0–10 and to complete the Oswestry Disability Index (ODI) questionnaire at baseline (before TFESI) and at 2 weeks, 1 month, 2 months, and 3 months following the procedure at the pain clinic. At the 2-week follow-up, the patients were divided into two groups based on their responses. The favorable response to TFESI was defined as a pain (upon movement) reduction >50.0% calculated by the difference between the pre- and post-TFESI scores divided by the pre-TFESI score. The others were defined as unfavorable responders.
After a 3-month follow-up at the pain clinic, all patients were interviewed by phone each month and visited the clinic every 3 months, continuing until 12 months postprocedure. Data collected included baseline patient characteristics, pain duration, body mass index (BMI), straight leg raising test (SLRT), motor deficit, and neurogenic claudication. Two certified radiologists evaluated the lumbosacral MRI and reached an agreement on various parameters, including the extent and location of disc herniation according to the Pfirrmann classification (ranging from bulging and protrusion to extrusion and sequestration) [8], degree of central canal stenosis (measured by the cross-sectional area and classified as mild, moderate, or severe) [9], grading of nerve root compromise [10] and foraminal stenosis [11], and spondylolisthesis. The primary outcomes were the differences in clinical presentations and MRI findings between the favorable and unfavorable responders. The secondary outcomes were pain duration reduction >50% following TFESI and spinal surgery following TFESI between favorable and unfavorable responders at 6 and 12 months.
Statistical analysis
The sample size was calculated using the 10 events per covariate rule formula, a common practice in logistic regression modeling to ensure reliable parameter estimates. According to this formula, having a minimum of 10 occurrences of the outcome variable for each predictor variable (covariate) in the logistic regression model is recommended. The primary study outcome has seven variables as clinical predictors, requiring a total of 70 events. According to a previous study [12], the proportion of favorable responses to TFESI was 61.0%, resulting in a requirement of 114 participants. Considering a dropout rate of 5.0%, the total sample size required was 120 participants.
An unpaired t-test and chi-square test were performed to compare continuous and categorical data between the two responder groups. The descriptive data were reported as means and standard deviations, whereas categorical data were reported as percentages. Univariate logistic regression analysis was performed to assess the primary outcome. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to determine the reliability of the estimates. Subsequently, multiple logistic regression analysis was performed to identify the individual predictors of a substantial response highlighted in the univariate analysis. The pain and ODI scores over time were evaluated between the two responder groups using a repeated-measure analysis of variance test. The correlations between the pain score, ODI score, and MRI findings were analyzed using Spearman’s rank correlation. The chi-square test was performed to compare the likelihood of undergoing surgery between the two groups. The sensitivity and specificity of the mean difference in the pain score to predict a favorable response at various time points were shown by a receiver operating characteristic (ROC) curve with an area under the curve. Data analysis was performed using SPSS ver.29 and Stata ver. 16.0 (Stata Corp., College Station, TX, USA). A p-value of <0.05 was considered a significant difference.
Results
The study initially enrolled 168 participants, of which 48 were excluded based on predefined criteria. Among the excluded participants, 23 had a history of lumbar spine surgery, 12 had recently undergone TFESI, five displayed depressive symptoms, five had systemic inflammatory disease, two were using oral prednisolone, and one was diagnosed with a spinal tumor involving epidural invasion. Demographic data, pain characteristics, clinical presentations, and MRI findings were collected from 120 patients, who were followed up for 12 months (Fig. 1). After the procedure, every patient expressed a pain reduction of >90% during postprocedure observation in the postanesthetic care unit, confirming that the injection effectively treated the underlying spinal pain pathology. Two weeks after TFESI, a favorable response was observed in 60.8% of the patients (73/120), whereas 39.2% (47/120) exhibited unfavorable responses. No significant differences in age, BMI, duration of pain symptoms, presence of neurogenic claudication, SLRT, and motor deficit were found between individuals categorized as favorable and unfavorable responders (Table 1).
The MRI findings of both groups were comparable, except for the extent of central canal stenosis, which varied. Patients classified as unfavorable responders exhibited severe central canal stenosis, whereas favorable responders showed a significant presence of mild central canal stenosis (p<0.01). Furthermore, MRI did not demonstrate a significant correlation between the extent of spinal pathology and the baseline pain intensity level (p>0.05), except for the presence of spondylolisthesis. The coexistence of spondylolisthesis exhibited a modest association with high baseline pain intensity (with a coefficient of 0.19 and a p-value of 0.04). The results of the univariate and multivariate logistic regression analyses are presented in Table 2. In the multivariate analysis, patients with mild central canal stenosis had 4.8 times higher odds of a positive response to TFESI than those with severe central canal stenosis (adjusted OR, 4.8; 95% CI, 1.2–18.8; p=0.023).
Significant differences in pain scores (NRS) and functional scores (ODI) were found between favorable and unfavorable responders following TFESI, and these differences persisted for a minimum of 3 months (Fig. 2). Favorable responders experienced a notable 29-week (95% CI, 10.3–47.8) period of pain reduction after a single TFESI, along with a significantly lower incidence of subsequent surgery at 6 and 12 months (p<0.01) (Table 3). Unfavorable responders experienced only limited pain relief, typically lasting up to 1 week, and one-third to one-half ultimately underwent surgery. ROC curve analysis (Fig. 3) showed that the mean pain score differences of 3.5 out of 10.0, noted at 2 weeks and 1 month post-TFESI, represented optimal values for the greatest sensitivity, specificity, positive-predictive value, negative-predictive value, and accuracy in forecasting favorable responder status (p<0.01) (Table 4).
Discussion
In this prospective analytical study, 60.8% of the patients diagnosed with lumbosacral radiculopathy due to lumbar disc herniation or spinal canal stenosis showed a favorable response to TFESI. Significantly, severe central canal stenosis identified by MRI was a robust predictive factor for a negative outcome following TFESI. The mean pain score differences of 3.5 out of 10.0, noted at 2 weeks and 1 month post-TFESI, indicated the optimal estimation for forecasting a long-term positive outcome. The effect of a single TFESI on individuals with a favorable response lasted for approximately 29 weeks, resulting in a significant reduction in surgical rates to 6.8% at 6 months and 13.7% at 12 months after the injection.
These results were consistent with those of a previous study highlighting that severe central canal stenosis is a significant negative predictive factor, with an OR of 3.6 [13]. Severe central stenosis may exert direct pressure on the neural elements, particularly in the upper lumbar region, compared with other spinal pathologies such as foraminal stenosis and posterolateral disc herniation. Inflammation is not the only primary mechanism, but mechanical compression also contributes to unfavorable outcomes following TFESI in cases of severe central canal stenosis. Current studies did not establish a connection between a longer duration of pain symptoms and a poor response, which contradicts the findings of previous studies [6,13]. A study indicated that being relatively young and being female was linked to a more positive response [14]. However, the present study demonstrated that factors such as age, sex, BMI, pain location, neurogenic claudication, positive SLRT, and motor deficit should not be regarded as predictors of TFESI outcomes. These results were consistent with those of previous studies [6,13,15]. Moreover, a history of spine surgery was recognized as a negative prognostic factor for ESI in patients with lumbosacral radiculopathy [13]. Nevertheless, the proposed pain mechanism was hypothesized to be related to epidural scar formation rather than inflammation. Consequently, TFESI may provide limited pain relief, which is why these patients were excluded from the present study.
Regarding MRI findings, previous studies have illustrated that low-grade nerve root compression and disc height reduction were associated with a favorable response to TFESI [7,15]. In this study, nerve root compression, foraminal stenosis, and coexisting spondylolisthesis were not predictors of a favorable response. These findings aligned with those of earlier research [16,17]. Although the presence of spondylolisthesis demonstrated a modest association with higher baseline pain intensity, it did not serve as a predictor for a positive outcome following TFESI. Furthermore, the severity of the spinal pathology observed in MRI did not correlate with the pain intensity level. Therefore, physicians should not solely rely on MRI results to determine whether a patient should undergo TFESI or proceed with surgery.
In this study, the proportion of patients demonstrating a favorable response (60.8%) at 2 weeks postprocedure was remarkably similar to previous reports, that is, 61.0% at 2–4 weeks [6] and 62.0% at 3 months [13], despite the variations in timeframes. Although previous studies have shown the short-term effect of ESI, a single TFESI in the present study exhibited a positive response up to approximately 29 weeks. Despite the average age of the patients being 58 years and not appearing to have disc herniation alone and the likelihood of having spinal stenosis was high, the results demonstrated that the primary source of their pain symptoms was likely associated with inflammation. Our findings also demonstrated an unfavorable response after TFESI in some patients, indicating that not every patient receiving TFESI will experience pain reduction. The present study revealed a substantial decrease in surgical rates, dropping from 51.1% in unfavorable responders to 13.7% in favorable responders. This finding indicates that a positive response to TFESI can reduce the likelihood of requiring surgery in certain patients within 12 months. In a recent meta-analysis, both favorable and unfavorable responders to TFESI were included and compared with a placebo. The analysis revealed a tendency to reduce the requirement for surgery in the short-term (<1 year); however, this effect was not observed in the long-term (≥1 year) [18]. Another meta-analysis indicated that ESI was probably slightly more effective than placebo in reducing leg pain and disability during short-term follow-ups [19]. However, these improvements may not be deemed clinically significant by both patients and clinicians. Moreover, the long-term efficacy of ESI is constrained by either the progression of spinal conditions or the limited duration of action of steroids. The risks, benefits, and efficacy of both ESI and surgery should be carefully evaluated and discussed with patients to achieve a balanced decision.
The novel aspect of this study is the establishment of specific pain score thresholds at different periods following TFESI, which aids physicians in providing patients with more accurate prognostic information. The mean pain score differences of 3.5 out of 10.0, noted at 2 weeks and 1 month following TFESI, serve as benchmarks for predicting a long-term positive outcome. This promising discovery has practical implications in clinical practice, offering vital information for prognostic outcomes and aiding surgical decision-making. The study’s strength lies in its prospective design, which contrasts with previous retrospective studies. This design presents a more precise correlation between clinical presentations and MRI findings, enhancing the ability to predict responses to lumbar TFESI. However, the first limitation of this study is the absence of data regarding outcomes following TFESI beyond 1 year. Second, whether repeated TFESI yield a response similar to the initial injection remains uncertain. Third, no comparison of pain medications used was made between the two groups before and after the TFESI; however, the patients had already reported the failure of conservative treatments since recruitment. Fourth, other potential variables such as nerve conduction studies, quantitative sensory testing, or rehabilitation modalities were not included. Lastly, this study assessed the favorable response to TFESI as a predictor of surgical outcomes rather than compared the effect of receiving TFESI versus not receiving TFESI on the surgical outcomes. Potential future research can include scoring the clinical presentations and MRI findings to predict the response rates following lumbar TFESI in patients with degenerative lumbar spine diseases.
Conclusions
The mean pain score differences of 3.5 out of 10.0 noted at 2 weeks and 1 month post-TFESI represented the optimal sensitivity and specificity in forecasting favorable response. The presence of severe central canal stenosis likely indicates an unfavorable response to TFESI.
Key Points
A pain reduction of 3.5 out of 10 within 1 month following transforaminal epidural steroid injection (TFESI) is considered a clinical benchmark for predicting long-term positive outcomes.
The presence of severe central canal stenosis is significantly associated with an unfavorable response to TFESI.
The effect of a single TFESI on individuals with a favorable response lasted for approximately 29 weeks, resulting in a significant reduction in surgical rates to 13.7% at 12 months.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
