Neural Mobilization for Reducing Pain and Disability in Patients with Lumbar Radiculopathy: A Systematic Review and Meta-Analysis

Lumbar radiculopathy causes lower back and lower extremity pain that may be managed with neural mobilization (NM) techniques. This meta-analysis aims to evaluate the effectiveness of NM in alleviating pain and reducing disability in patients with lumbar radiculopathy. We hypothesized that NM would reduce pain and improve disability in the lumbar radiculopathy population, leveraging the statistical power of multiple studies. Electronic databases from their inception up to October 2023 were searched for randomized controlled trials (RCTs) that explored the impact of NM on lumbar radiculopathy. Our primary outcome measure was the alteration in pain intensity, while the secondary one was the improvement of disability, standardized using Hedges’ g. To combine the data, we employed a random-effects model. A total of 20 RCTs comprising 877 participants were included. NM yielded a significant reduction in pain intensity (Hedges’ g = −1.097, 95% CI = −1.482 to −0.712, p < 0.001, I2 = 85.338%). Subgroup analyses indicated that NM effectively reduced pain, whether employed alone or in conjunction with other treatments. Furthermore, NM significantly alleviated disability, with a notable effect size (Hedges’ g = −0.964, 95% CI = −1.475 to −0.453, p < 0.001, I2 = 88.550%), particularly in chronic cases. The findings provide valuable insights for clinicians seeking evidence-based interventions for this patient population. This study has limitations, including heterogeneity, potential publication bias, varied causal factors in lumbar radiculopathy, overall study quality, and the inability to explore the impact of neural pathology on NM treatment effectiveness, suggesting opportunities for future research improvements.


Introduction
Lumbar radiculopathy is a debilitating condition characterized by the compression of lumbar nerve roots [1].Causes of lumbar radiculopathy encompass bulging disc, herniation, hypertrophic facet or adjacent ligaments, spondylolisthesis, and, in rare instances, neoplastic and infectious conditions [1].Lumbar radiculopathy is often referred to as sciatica, a term derived from its hallmark symptom [2].It manifests as radiating leg pain over the distribution of the sciatic nerve [3].Treatment options for lumbar radiculopathy encompass conservative approaches, pharmaceutical interventions, and surgical procedures.Conservative care options are diverse, encompassing various exercise protocols, Life 2023, 13, 2255 2 of 17 the use of electrical modalities like transcutaneous electrical nerve stimulation, and techniques focused on mobilizing the affected tissue, including spinal mobilization and neural mobilization (NM) [4].
NM is a manual therapy technique designed to simultaneously lengthen a nerve at one joint while shortening it at an adjacent joint, with the primary objective of enhancing the smooth movement of neural structures within the surrounding tissues [5].This approach typically comprises two techniques, i.e., "slider" and "tensioner" [6].The former focuses on facilitating the sliding motion of neural tissue in relation to neighboring structures without generating significant tension [7].Conversely, the latter technique aims to create neural tension by increasing the distance between each end of the nerve tract, always within the tissue's elastic limits, ultimately enhancing the nerve's viscoelastic properties [7].The mechanisms underlying NM interventions involve restoring the equilibrium between the nerve and nearby connective tissues and reducing intraneural edema by dispersing fluid within the nerve axon [8].Furthermore, studies have indicated that NM has a hypoalgesic effect and can reduce the mechanical sensitivity of nerves [9,10].While there have previously been systematic reviews and meta-analyses addressing related topics [11,12], none have specifically focused on the lumbar radiculopathy population, nor have they delved into the further exploration of NM types, regimens, and lumbar radiculopathy symptom stages.Therefore, the primary aim of our meta-analysis is to investigate the benefits of NM on pain intensity and disability in individuals with lumbar radiculopathy.Furthermore, we aim to examine whether there are differential benefits based on NM type, regimen, and lumbar radiculopathy symptom stage.

Search Strategy
We performed an extensive literature review in accordance with the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, spanning from the inception of the databases to October 2023 [13].The PRISMA checklist is provided in Table S1.Our research protocol was duly registered on Inplasy.com with the registration number INPLASY2023100039.The screening process involved two reviewers, L.-H.L. and T.-Y.L., who meticulously examined various databases, including PubMed, ClinicalTrials.gov,Cochrane Library, and the Physiotherapy Evidence Database (PEDro).Our search strategy incorporated the following keywords: ("neural mobilization techniques" OR "neurodynamic mobilization techniques" OR "nerve mobilization techniques") AND ("lumbar radiculopathy" OR "sciatica").To ensure a comprehensive search, we employed different variations of these terms.Detailed information regarding the literature search is given in Table S2.

Inclusion and Exclusion Criteria
The PICO (population, intervention, comparison, outcome) setting of the current meta-analysis was as follows: P: human participants with lumbar radiculopathy; I: the NM technique; C: controls that did not employ NM; and O: changes in pain intensity and disability.
The inclusion criteria were as follows: (1) RCTs investigating pain intensity and disability before/after NM; (2) enrolling adults diagnosed with lumbar radiculopathy and/or sciatica based on radiography, reproducing radiated symptoms in the leg with a passive straight leg raise test or slump test; (3) intervention groups being treated with NM alone or NM plus other treatments; (4) at least one reference group using treatments other than NM (inclusion of both passive control, such as no intervention, and active control, such as conventional treatments).
The exclusion criteria were as follows: (1) non-RCTs; (2) studies that enrolled patients with recent associated neurological symptoms (e.g., foot drop and cauda equina), previous surgery in the lumbar spine, spine fractures, and lower extremity injuries that induced leg pain; (3) case reports, case series, and trials using quasi-experimental, single-arm, or Life 2023, 13, 2255 3 of 17 longitudinal follow-up designs; (4) studies lacking the desired outcome; and (5) those that enrolled participants duplicated with a previously published trial.

Primary Outcome Measurements
The primary focus of the study centered on evaluating alterations in pain intensity, quantified using the Numeric Rating Scale (NRS) and the Visual Analog Scale (VAS) before and after the intervention [14].Detailed descriptions can be found in Table S3 for clarity.

Data Extraction
The data extraction process was conducted by two independent reviewers (L.-H.L. and T.-Y.L.).This process involved gathering information encompassing demographics, study design, intervention details, outcome measures, and assessment timeframes.We utilized Excel to create the data collection form.In cases of disagreement between the two reviewers, consensus was achieved through discussion or consultation with the corresponding author for resolution.For studies with multiple arms, we organized similar eligible sets to facilitate straightforward pairwise comparisons, as previously reported [20].In situations where we encountered missing data in the published articles, we reached out their corresponding authors to obtain the original data.If the need arose to convert medians and interquartile ranges into means and standard deviations, we followed the guidelines provided in the Cochrane Handbook for Systematic Reviews of Interventions [21].

Assessment and Quality Classification
The evaluation of the quality of the Randomized Controlled Trials (RCTs) included was performed using the Cochrane risk-of-bias tool, known as RoB 2. RoB 2 assesses six primary criteria, namely the randomization process, intervention adherence, missing outcome data, outcome measurement, selective reporting, and the overall risk of bias [22].In the assessment of intervention adherence within the RoB 2 tool, two options were available: the intention-to-treat approach, which is based on intervention assignment, and the per-protocol approach, which is based on intervention adherence.For our meta-analysis, we used the per-protocol approach, which was in alignment with the study design used in the majority of the included studies [22].

Statistical Analysis
Due to the diverse measurement tools in the included studies, we deemed the mean difference inappropriate and opted for Hedges' g as our standardized effect size.This choice ensures outcome comparability across studies with varied measurement scales.Our results, expressed with Hedges' g and corresponding 95% CIs, quantify and convey the study findings.Effect sizes of 0.2, 0.5, and 0.8 were construed as representative of small, moderate, and large effects, respectively [23].This study employed a randomeffects model.This choice was based on the assumption that there was a distribution of intervention effects across studies, accommodating variability.The model recognizes that observed differences stem from both chance and genuine variation in intervention effects, resulting in broader confidence intervals when heterogeneity is present, as opposed to a fixed-effect model [24].Additionally, we assessed the extent of heterogeneity among the studies by employing I 2 and Cochran's Q statistics, with I 2 values exceeding 50% signifying significant heterogeneity [25].To appraise the influence of individual trials on the overall effect size, we conducted sensitivity analyses using the one-study removal method.Subgroup analyses were also carried out based on the type of NM techniques (slider or tensioner, slump or straight leg raise), symptom stage (chronic stage > 3 months, non-chronic stage ≤ 3 months), and NM regimen (NM only and NM plus).Furthermore, we performed meta-regression analyses to investigate potential associations between the pain-alleviating effects of NM and the duration of treatment per day, as well as the frequency of sessions per week.In order to scrutinize the potential presence of publication bias, we visually assessed the funnel plots and examined the statistics derived from Egger's regression test [26].Statistical analysis was performed using Comprehensive Meta-Analysis software (version 3, Biostat, Englewood, NJ, USA), with statistical significance defined as a two-tailed p-value below 0.05.

Study Identification and Selection
Figure 1 displays the PRISMA flowchart illustrating our literature search.Initially, we identified 1202 non-duplicated citations through our research efforts.Subsequently, we subjected 37 articles to further scrutiny for eligibility.After conducting a thorough full-text assessment, we excluded seventeen articles for various reasons: six were not RCTs, two did not provide both pre-and post-intervention data, one did not report the desired outcome measurements, five lacked a non-NM control group, two did not include patients with lumbar radiculopathy, and one had participants who overlapped with another publication by the same author.The specific rationales for these exclusions are given in Table S4.We ultimately incorporated 20 RCTs into our analysis, involving a total of 877 participants whose ages ranged from 20 to 60 years.The intervention durations spanned from two to six weeks.Among the trials, one was a three-arm study, while the remaining nineteen were two-arm RCTs.
of intervention effects across studies, accommodating variability.The model recognizes that observed differences stem from both chance and genuine variation in intervention effects, resulting in broader confidence intervals when heterogeneity is present, as opposed to a fixed-effect model [24].Additionally, we assessed the extent of heterogeneity among the studies by employing I 2 and Cochran's Q statistics, with I 2 values exceeding 50% signifying significant heterogeneity [25].To appraise the influence of individual trials on the overall effect size, we conducted sensitivity analyses using the one-study removal method.Subgroup analyses were also carried out based on the type of NM techniques (slider or tensioner, slump or straight leg raise), symptom stage (chronic stage > 3 months, non-chronic stage ≤ 3 months), and NM regimen (NM only and NM plus).Furthermore, we performed meta-regression analyses to investigate potential associations between the pain-alleviating effects of NM and the duration of treatment per day, as well as the frequency of sessions per week.In order to scrutinize the potential presence of publication bias, we visually assessed the funnel plots and examined the statistics derived from Egger's regression test [26].Statistical analysis was performed using Comprehensive Meta-Analysis software (version 3, Biostat, Englewood, NJ, USA), with statistical significance defined as a two-tailed p-value below 0.05.

Study Identification and Selection
Figure 1 displays the PRISMA flowchart illustrating our literature search.Initially, we identified 1202 non-duplicated citations through our research efforts.Subsequently, we subjected 37 articles to further scrutiny for eligibility.After conducting a thorough full-text assessment, we excluded seventeen articles for various reasons: six were not RCTs, two did not provide both pre-and post-intervention data, one did not report the desired outcome measurements, five lacked a non-NM control group, two did not include patients with lumbar radiculopathy, and one had participants who overlapped with another publication by the same author.The specific rationales for these exclusions are given in Table S4.We ultimately incorporated 20 RCTs into our analysis, involving a total of 877 participants whose ages ranged from 20 to 60 years.The intervention durations spanned from two to six weeks.Among the trials, one was a three-arm study, while the remaining nineteen were two-arm RCTs.The aforementioned three-arm study compared three interventions: NM slider technique, NM tensioner technique, and stretch exercises [27].Within the group of two-arm studies, six compared NM to conventional treatments, which included physical agents combined with lumbopelvic region exercises such as lumbopelvic muscle strengthening, stretching, or range-of-motion exercises [28][29][30][31][32][33].Additionally, one study compared NM to physical agents (e.g., transcutaneous electrical nerve stimulation) along with lumbar region massage [34]; another study examined NM versus lumbar range-of-motion exercises combined with lumbar stabilization exercises [35]; one study involved education about daily-life activities as the control [36]; and one study combined conventional treatments with lumbar region trigger point release and hamstring stretching [37].Similarly, another study paired conventional treatments with hamstring stretching [38], one study explored NM versus kinesio taping [39], three studies focused on NM versus lumbar stabilization exercises [40][41][42], three studies compared NM to physical agents [43][44][45], and one study combined conventional treatments with massage [46].

Methodological Quality of the Included Studies
With respect to the overall methodological quality of the included studies, we found that 15% had a low risk of bias (n = 3), 20% (n = 4) had some risk of bias, and 65% had a high risk of bias (n = 13) (Figure S1).The item that was mostly rated as some risk of bias was missing outcome data, followed by an absence of explanation for baseline differences and a lack of clear description of allocation concealment.The item that was mostly rated as high risk of bias was no information about the extent of missing outcome data.The details of the risk of bias assessment are summarized in Table 3.

Publication Bias
The funnel plot for pain intensity and disability revealed an asymmetry of effect size distribution, with both p-values < 0.001 according to Egger's regression test (Figures S8 and S9).

Discussion
Our meta-analysis revealed that the use of NM significantly reduced both pain and disability in lumbar radiculopathy.Regardless of the specific regimen and the type of technique employed, NM consistently produced positive results.Longer treatment duration and more frequent sessions further enhanced the effectiveness of NM in reducing pain and improving disability.Regarding disability, while the effect size of NM for participants in the chronic stage group was significant, it lacked significance for participants in the nonchronic stage group.Moreover, the concept of clinical significance is closely intertwined with effect size estimation, which, in turn, leads to power analysis estimation.Generally, effect sizes of 0.2, 0.5, and 0.8 are considered indicative of small, moderate, and large effects, respectively [23].In the current meta-analysis, the effect size of pain intensity reduction (Hedges' g = −1.097)and disability improvement (Hedges' g = −0.964)was large.This suggests their clinical significance.To the best of our knowledge, this article represents the first meta-analysis investigating the effectiveness of NM in reducing pain and disability specifically in individuals with lumbar radiculopathy.Moreover, this study includes subgroup analyses, addressing a significant gap in the current literature.
In a previous systematic review and meta-analysis, Neto et al. [11] included 10 RCTs wherein the effects of NM targeting the lower body quadrant in both healthy individuals and those with low back pain were examined.The results indicated that NM had moderate effects on flexibility in healthy adults.Notably, it led to a significant pain reduction and disability improvement in individuals with low back pain, highlighting the potential benefits of NM.Pourahmadi et al. [12] conducted a systematic review and meta-analysis of seven RCTs.Examining the effectiveness of NM for patients with low back pain, they demonstrated a significant reduction in pain and disability.However, it is worth noting that only two RCTs in Neto et al.'s study and three RCTs in Pourahmadi et al.'s study encompassed participants with lumbar radiculopathy, making our meta-analysis unique in its specific focus on this particular population.

Mechanisms of NM
Our meta-analysis uncovered that NM was more effective in reducing pain and disability when compared to control treatments.To understand how these techniques impact neural tissues, it is crucial to dive into the sequence of events that transpire when nerves encounter mechanical or chemical stimuli surpassing their tolerance threshold.Nerves possess the ability to adapt to various mechanical stresses during daily movements, but excessive stress can lead to ischemia and impaired function.Compressive stressors on nerve roots, such as disc herniation, osteophytes, or spinal stenosis, were observed to impede blood flow and cause sensory/motor dysfunction.They often result in pain due to microvascular alterations and inflammation [5].The latter could incite nerve mechanosensitivity, which arises from inflammatory mediators and sensitizing C fibers [5].Given the above-mentioned mechanisms of nerve injury, NM's effectiveness can reasonably be explained in several ways.
First, NM has been investigated for its ability to induce hypoalgesia.Beneciuk et al. [9] demonstrated an immediate hypoalgesic effect on C-fiber-mediated pain following specific tensioning techniques on the median nerve, as observed in thermal quantitative sensory testing.This effect might be attributed to a decrease in glia fibrillary acid proteins in the dorsal root ganglion and lumbar spinal cord after NM, associated with reduced allodynia and hyperalgesia [47].Second, NM has been investigated for its ability to reduce mechanosensitivity.Zhu et al. [10] reported lower concentrations of interleukin 1β (IL-1β) and tumor necrosis factor α (TNF-α) in the gluteal and trunk nerve branches following NM, correlating with reduced mechanical sensitivity.Third, there are reports showing that a promotion of nerve repair occurs after undergoing NM.In one study, increased neural growth factor and myelin protein zero levels, which play a crucial role in axonal regrowth and remyelination after injury, were demonstrated after NM [48].

Effectiveness of NM on Pain Reduction and Disability Improvement
NM has demonstrated benefits in reducing pain intensity in the participants analyzed in this review, whether they are in the chronic or non-chronic stage.However, when it comes to disability reduction, NM appears to be more effective in participants who are in the chronic stage.This finding can be explained by several factors.First, participants in the non-chronic stage group typically underwent shorter treatment durations (ranging from 2-3 weeks) and received fewer (i.e., three) treatment sessions per week, compared to those involved in the chronic stage group (with certain treatment durations lasting up to six weeks with up to five sessions per week).Second, nerve injuries need enough time to recover.In the early stages, nerve compression may be associated with a breakdown of the blood-nerve barrier.It could result in subperineurial edema with fibrosis and localized segmental demyelination [49].Based on Sunderland's classification, nerve injuries caused by mechanical stress can be categorized into multiple grades, depending on the status of continuity of the axon and myelin sheath [50].For instance, in a first-degree injury (according to Sunderland's classification), there is segmental demyelination, and sensory/motor functions are impaired until remyelination occurs.Full function can typically be expected without intervention within approximately 12 weeks [51].As such, we could suppose that reparative mechanisms following nerve injury necessitate a specific recuperation period, and in the (sub)acute phases of injury, nerve repair may not have fully transpired.Lastly, because gradual damage in chronic compression does not trigger an inflammatory response, macrophages arrive slowly, often after significant Schwann cell proliferation has already occurred.This situation is associated with an increase in Schmidt-Lanterman incisures, which are cytoplasmic components of Schwann cells believed to regulate myelin sheath metabolism.Therefore, their elevated levels suggest that Schwann cells boost their metabolic activity to facilitate remyelination in response to demyelination [49].These physiological mechanisms could have contributed to the enhanced therapeutic efficacy observed in the chronic phase.

Adverse Events
In the included 20 RCTs, none had reported any adverse events, which is consistent with previous systematic reviews and meta-analysis on NM [43].However, there were some contraindications for NM, including cauda equina lesions, cord signs, and other pathologies that would affect the nervous system, such as Guillain-Barre syndrome and multiple sclerosis [52].Therefore, it is crucial for healthcare professionals to identify these aforementioned issues before carrying out NM.

Limitations
It is essential to acknowledge several limitations of this study.First, we observed significant variability in the overall impact of pain and disability, indicating heterogeneity.To address this issue, subgroup analyses were conducted, focusing on different NM regimens to identify potential factors contributing to this heterogeneity.Second, our metaanalysis uncovered evidence of publication bias affecting the effect size of pain intensity and disability.To ensure the accuracy of NM's actual effects, a subsequent meta-analysis may be warranted to confirm whether the current publication bias continues to influence NM's true efficacy.Thirdly, lumbar radiculopathy has multifaceted origins, and among the studies included, causal factors are not uniform.Future investigations may explore the inclusion of more specific articles for a nuanced analysis.Fourthly, we acknowledge that the overall quality of the included studies is a limitation, with only seven out of the twenty studies assessed using RoB2 having a risk level lower than "high".Future studies need to enhance the understanding of lumbar radiculopathy treatment with NM by including a greater number of high-quality RCTs.Fourthly, factors such as muscle strength, sensory impairments, or walking ability are crucial aspects of lumbar radiculopathy.Future research should further analyze and explore these factors.Lastly, due to a lack of available studies on this aspect, it is crucial to highlight that this meta-analysis could not explore whether the extent of neural pathology played role in influencing the effectiveness of NM treatment.Investigating this relationship should be considered a valuable avenue for future research exploration.

Conclusions
In summary, our analysis confirms the effectiveness of NM in reducing pain and disability in individuals with lumbar radiculopathy.Regardless of the type, NM techniques demonstrated consistent positive results for both pain reduction and disability improvement.Whether used alone or in combination with other therapies, NM was beneficial.Moreover, NM effectively reduced pain in both the chronic and non-chronic stages, with greater disability reduction observed in the chronic stage.Longer treatment duration and more frequent sessions were associated with greater improvement in pain and disability.Future studies need to focus on follow-up duration and its effects on different neural pathologies.

Figure 1 .
Figure 1.PRISMA flow diagram describing the screening and review process for the current meta-analysis.

Figure 1 .
Figure 1.PRISMA flow diagram describing the screening and review process for the current meta-analysis.

Figure 3 .
Figure 3. Forest plot of subgroup analysis for pain intensity based on the type/regimen of neural mobilization (NM) and symptom stage of lumbar radiculopathy [27-39,41-46].

Table 1 .
Characteristics of the included studies.

Table 2 .
Summary of the intervention details of the included trials.

Table 3 .
Detailed quality assessment of included studies using Cochrane risk-of-bias 2 tool.