Needle knife therapy for cervical spondylotic radiculopathy: A systematic review and meta-analysis

Abstract

BACKGROUND

Cervical spondylotic radiculopathy (CSR), a prevalent form of cervical spondylopathy, significantly impacts patients’ quality of life. Current management primarily relies on conservative approaches. In recent years, needle knife therapy, as a new type of therapy, has been used in the treatment of CSR.

AIM

To evaluate the therapeutic efficacy of needle-knife therapy for CSR and compare its advantages over conventional treatments, including acupuncture, massage, and warm acupuncture.

METHODS

Chinese Biomedical Literature Database, China National Knowledge Infrastructure, Wanfang, PubMed, VIP Chinese Science and Technology Journal Database, China Biology Medicine, and other databases were searched for randomized controlled trials (RCTs) of needle-knife therapy for CSR from the establishment of the database to November 3, 2025. Two independent researchers performed literature screening based on predefined inclusion and exclusion criteria, followed by secondary screening and quality assessment. The Cochrane risk of bias tool was utilized for assessing the methodological quality of the included RCTs. Data synthesis and analysis were performed using RevMan 5.4 software, focusing on treatment effective rates, Visual Analog Scale (VAS) scores, Neck Disability Index (NDI) scores, and Yasushi Tanaka cervical spondylitis symptom scale 20 (YT20) scores.

RESULTS

A total of 15 RCTs, encompassing 1184 patients (592 in the experiment group and 592 in the control group), were included in this meta-analysis. Potential high risks of bias in randomization, allocation concealment, and blinding procedures may adversely affect the overall methodological quality of the included trials. The results of meta-analysis showed that the effective rate of the experiment group was significantly higher than that of the control group [mean differences (MD) = 0.10, 95% confidence interval (CI): 0.07-0.14, Z = 6.03, P ≤ 0.00001]; the safety and YT20 scores [safety: MD = 0.34, 95%CI: 0.14-0.83, Z = 2.36, P = 0.02, YT20 scores: standardized MD (SMD) = 2.05, 95%CI: 1.24-2.86, Z = 4.94, P ≤ 0.00001] in the experiment group were higher than those in the control group. The VAS score (SMD = -0.93, 95%CI: -1.59 to -0.27, P ≤ 0.0001), NDI score (SMD = -4.04, 95%CI: -5.01 to -3.07, P ≤ 0.0001) in the experiment group were lower than those in the control group. Significant heterogeneity was observed for VAS (I² = 99%) and NDI (I² = 84%) scores. Sensitivity analysis did not find significant reversal, indicating that the results of the included studies were reliable.

CONCLUSION

Needle-knife therapy for CSR presents a higher efficacy and improved clinical outcome scores when contrasted with conventional treatments. However, these findings are constrained by the generally low quality of the included trials and notable heterogeneity observed in some studies. Consequently, the results should be interpreted with caution and considered preliminary. Further validation through larger, high-quality RCTs is warranted, given the existing limitations in sample size and methodological rigor.

Key Words: Needle-knife therapy; Cervical spondylotic radiculopathy; Randomized controlled trial; Systematic review; Meta-analysis

Core Tip: This study aims to evaluate the therapeutic efficacy of needle-knife therapy for cervical spondylotic radiculopathy through a meta-analysis of 15 randomized controlled trials encompassing 1184 patients. The research addresses three core issues: comparative effectiveness versus conventional treatments (acupuncture, massage, warm acupuncture), comprehensive clinical outcome assessment (effective rates, Visual Analog Scale scores, Neck Disability Index scores, and Yasushi Tanaka cervical spondylitis symptom scale 20 scores), and methodological quality evaluation using Cochrane risk of bias tool. Results demonstrate superior effective rates and improved symptom scores. Findings provide preliminary support for needle-knife therapy while emphasizing the need for larger, high-quality randomized controlled trials to validate these results.

INTRODUCTION

Cervical spondylotic radiculopathy (CSR) is a common type of cervical spondylosis, accounting for approximately 55% of all cervical spondylosis patients in China[1]. The pathogenesis of CSR primarily involves thickening and calcification of the cervical ligaments, degeneration of the intervertebral discs, and osteophyte formation. These pathological changes result in intervertebral foramen stenosis, thereby compressing and irritating the spinal nerve roots[2]. The clinical symptoms include stiffness and pain in the neck and shoulders, accompanied by headache, dizziness, numbness in the upper limbs, and muscle weakness, which significantly affect the daily life and work of patients[3]. Currently, the treatment of CSR mainly relies on conservative methods, such as massage, acupuncture, and local block. While these approaches are effective to some extent, they have drawbacks, including a slow onset of action and prolonged treatment courses. Outcomes are often unsatisfactory for patients with poor treatment compliance[4]. Needle-knife therapy evolved from traditional acupuncture therapy, compensating for the limitation of conventional acupuncture in effectively releasing the tension of soft tissue surrounding the spine[5]. Based on the results of our team’s previous clinical trials[6], needle-knife therapy can effectively reduce the Visual Analog Scale (VAS) and Yasushi Tanaka cervical spondylitis symptom scale 20 (YT20) scores of CSR patients and increase the clinical effective rate. However, there is a lack of systematic research to prove the advantages of needle-knife therapy for CSR. Therefore, this study conducted a meta-analysis of randomized controlled trials (RCTs) to evaluate the efficacy of needle-knife therapy for CSR, aiming to provide evidence-based support for its clinical application. The research protocol has been registered on the PROSPERO platform (registration No. CRD42025633257).

MATERIALS AND METHODS

This meta-analysis was conducted in strict accordance with the PRISMA 2020 statement as displayed in the PRISMA checklist (Table 1)[7,8].

Table 1 Summary of findings for Grading of Recommendations Assessment, Development and Evaluation ratings.

Outcome measure	Quality of evidence assessment (GRADE)
	No of patients (effect size)	Risk of quality of bias1	Inconsistency2	Evidence6	Indirectness3	Imprecision4	Publication5
Effective rate	1184	Serious	Very low	Low	Not serious	Very low	Very low
VAS	1184	Low	Very serious	Low	Not serious	Very low	Very low
NDI	734	Low	Very serious	Low	Not serious	Very low	Very low
YT20	246	Serious	Very serious	Low	Not serious	Not serious	Very low

¹Risk of bias based on the Cochrane risk of bias tool. This tool assesses selection bias, performance bias, detection bias, attrition bias, and reporting bias.

²Downgraded if there was a substantial unexplained heterogeneity. Downgraded one level for serious inconsistency: 50% < I² ≤ 75%; downgraded two level for very serious inconsistency: 75% < I² < 100%.

³Downgraded if there were factors present relating to the participants, interventions, or outcomes that limited the generalizability of the results.

⁴Downgraded if the number of participants are very small (n < 200).

⁵Downgraded if there was an evidence of publication bias using Egger’s test.

⁶Given that all included studies were meta-analyses of randomized controlled trials, the evidence certainty was initially rated as high for all outcomes and sub- sequently downgraded according to predefined criteria. Quality was graded as high, moderate, low, and very low.

GRADE: Grading of Recommendations Assessment, Development and Evaluation; VAS: Visual Analog Scale; NDI: Neck Disability Index; YT20: Yasushi Tanaka cervical spondylitis symptom scale 20 scores.

A comprehensive literature search was performed from the inception of each database up to December 3, 2025, encompassing eight authoritative international and domestic databases: PubMed, EMBASE, Web of Science, the Cochrane Library, China National Knowledge Infrastructure, Wanfang Data Knowledge Service Platform, VIP Chinese Science and Technology Journal Database, and the Chinese Biomedical Literature Database. A combined search approach incorporating both Medical Subject Headings (MeSH) terms and free-text keywords was applied. The detailed search strategy is presented in Figure 1. Search term is: #1 (“Acupuncture Therapy” [Mesh] OR “Acupuncture” [Mesh] OR acupotom*[tiab] OR “needle-knife” [tiab] OR “needle knife” [tiab] OR “miniscalpel needle” [tiab]) #2 (“Cervical Vertebrae” [Mesh] OR “Radiculopathy”[Mesh] OR “cervical spondylotic radiculopathy” [tiab] OR CSR [tiab] OR “cervical radiculopathy” [tiab]) #3 (“Randomized Controlled Trial” [Publication Type] OR randomized [tiab] OR randomized [tiab] OR randomly [tiab] OR trial [tiab])#4 #1 AND #2 AND #3. We also manually searched the bibliographies of relevant systematic reviews and meta-analyses as a supplementary measure to identify any additional RCTs that could satisfy our eligibility criteria.

Figure 1 Literature search process. CNKI: China National Knowledge Infrastructure; VIP: VIP Chinese Science and Technology Periodical Database; RCT: Randomized controlled trial; VAS: Visual Analog Scale.

Literature screening criteria

The literature screening was performed by two investigators (Wang ZY and Xu HH) in accordance with predefined inclusion and exclusion criteria. Any discrepancies were resolved through consensus discussion.

Inclusion criteria: (1) The study type is a RCT; (2) The research subjects meet the diagnostic criteria for CSR as stipulated in the “Diagnosis and Therapeutic Standards for Traditional Chinese Medicine Diseases and Syndromes” in 2024[9]; (3) The treatment group consists of needle-knife therapy or needle-knife combined with moxibustion, massage, etc. while the control group consists of acupuncture or moxibustion, massage, etc.; (4) There is good balance and comparability among the groups in the original data; and (5) The outcome indicators include the effective rate of treatment, VAS, Neck Disability Index (NDI) score, YT20, etc.

Exclusion criteria: (1) Literature from which the full article cannot be obtained; (2) Republished literature; (3) Incomplete clinical data or insufficient statistical data; (4) Other diseases similar to CSR; (5) Literature whose outcome indicators do not meet the requirements; (6) Non-RCTs, such as reviews, clinical experience introductions, summary analyses, clinical case reports, literature involving massage genres, and animal experiments; and (7) Retracted articles.

Study selection process

The retrieved records were organized and deduplicated using EndNote X9 software. Following deduplication, the titles and abstracts of the remaining records were screened for eligibility based on the predefined inclusion and exclusion criteria. Two reviewers (Li XF and Zhang ZW) independently conducted the initial screening by assessing the titles, abstracts, and introductions against the inclusion criteria. Irrelevant studies were excluded with recorded reasons. In the event of disagreements, the reviewers re-evaluated the full texts and discussed to reach a consensus. The final screening results were cross-checked by a third reviewer (Xu HH) to validate the selection process and to establish a solid foundation for the ensuing meta-analysis.

Data extraction and management

Data were extracted independently by two researchers (Li XF and Zhang ZW) into a pre-designed Microsoft Excel 2023 spreadsheet and then cross-checked for consistency. The scope of extraction included: (1) Basic study information (first author, publication year, country); (2) Participant characteristics (group sample sizes, sex distribution); (3) Intervention details for both control and experimental groups, with precise description of the needle-knife technique; (4) Outcomes, focusing on pre- and post-treatment values for VAS, NDI, and YT20 scores; and (5) Safety indicators, encompassing adverse event rates and other relevant safety assessments.

Risk of bias assessment

The methodological quality of the eligible studies was evaluated by two independent reviewers using the Cochrane Risk of Bias tool. This tool assesses six domains: Random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting. Each domain was judged as having a low, high, or unclear risk of bias, following the criteria outlined in the Cochrane Handbook. The overall assessment was then visualized in a risk of bias summary figure.

Statistical analysis

Statistical analyses were performed using Review Manager 5.4 software, following Cochrane guidelines. For dichotomous variables, results were pooled as risk ratios; for continuous variables, mean differences (MD) or standardized MD (SMD) were used, all with 95% confidence intervals (CI). Heterogeneity was assessed using the I² statistic and χ² test. A fixed-effect model was applied when heterogeneity was low (I² ≤ 50%, P ≥ 0.1); otherwise, a random-effects model was used. Significant heterogeneity (I² > 50%) was explored through sensitivity or subgroup analysis.

RESULTS

A preliminary search yielded a total of 612 potentially relevant studies. After removing 312 duplicate records and excluding 237 studies based on title and abstract review, 73 studies remained for full-text screening. Ultimately, 15 RCTs[10-24] meeting the inclusion criteria were included, with the study selection process illustrated in Figure 1.

Characteristics of included studies

A total of 15 studies were included, comprising 1 English study and 14 Chinese studies. The average age range of patients in the included studies was 45-55 years. All studies employed needle-knife treatment in the experimental group. Ten studies utilized needle-knife alone, two studies combined needle-knife with tuina therapy, and three studies combined needle-knife with acupuncture treatment. All 15 studies assessed therapeutic effectiveness using the overall effective rate and evaluated pain improvement using the VAS score. Ten studies employed the NDI to assess the improvement in cervical function, while four studies used the YT20 to evaluate the improvement in symptoms and signs in patients with CSR. The basic characteristics of the included literature are shown in Table 2.

Table 2 Basic characteristics of the include literature.

Ref.	Nation	Sample size	Duration of treatment	Interventions	Therapeutic process	Outcome indicator	Safety evaluation
Deng[10], 2017	China	E: 30 C: 30	E: 3 weeks C: 3 weeks	E: N + C C: C	E: Once every 2 days	Effective rate	N
					C: Once a day	VAS
Yang[11], 2019	China	E: 30 C: 30	E: 3 weeks C: 3 weeks	E: N + C C: C	E: Once every 2 days	Effective rate	Y
					C: Once a day	VAS, NDI, YT20
Xu[12], 2018	China	E: 33 C: 33	E: 4 weeks C: 4 weeks	E: N C: C	E: Once every day	Effective rate	N
					C: Once every day	VAS, YT20
Wang and Huo[13], 2019	China	E: 43 C: 43	E: 2 weeks C: 2 weeks	E: N C: C	E: Once every day	Effective rate	N
					C: Once every day	VAS, NDI, YT20
Li et al[14], 2022	China	E: 30 C: 30	E: 6 weeks C: 6 weeks	E: N C: C	E: Once every day	Effective rate	N
					C: Once very day	VAS
Wu et al[15], 2022	China	E: 18 C: 18	E: 2 weeks C: 2 weeks	E: N + C C: C	E: Once every 3 days	Effective rate	Y
					C: Once every day	VAS, NDI
Ma[16], 2019	China	E: 102 C: 102	E: 2 weeks C: 2 weeks	E: N C: C	E: Once every 2 days	Effective rate, VAS	N
					C: Once every day
Hu[17], 2019	China	E: 30 C: 30	E: 4 weeks C: 4 weeks	E: N C: C	E: Once every day	Effective rate, VAS,	Y
					C: Once every day
Zhou et al[18], 2022	China	E: 50 C: 50	E: 4 weeks C: 4 weeks	E: N + C C: C	E: Once a week	Effective rate	Y
					C: Once every day	VAS, NDI
Yin et al[19], 2024	China	E: 50 C: 50	E: 2 weeks C: 2 weeks	T: N C: C	E: Once ten days	Effective rate	N
					C: Once every day	VAS, NDI
Li et al[20], 2023	China	E: 53 C: 53	E: 3 weeks C: 3 weeks	E: N + C C: C	E: Once every day	Effective rate	N
					C: Once every day	VAS, NDI
Liu et al[21], 2024	China	E: 30 C: 30	E: 2 weeks C: 2 weeks	E: N C: C	E: Once every day	Effective rate	N
					C: Once every day	VAS, NDI, YT20
Zhang and Yao[22], 2020	China	E: 30 C: 30	E: 3 weeks C: 3 weeks	E: N C: C	E: Once every day	Effective rate	N
					C: Once every day	VAS
Jin[23], 2021	China	E: 33 C: 33	E: 4 weeks C: 4 weeks	E: N C: C	E: Once every day	Effective rate, VAS, NDI	N
					C: Once every day	Effective rate, VAS, YT20
Li and Wang[24], 2021	China	E: 30 C: 30	E: 3 weeks C: 3 weeks	E: N C: C	E: Once every day		N
					C: Once every day

E: Experiment group; T: Control group; N: Needle knife; C: Conventional therapy; VAS: Visual Analog Scale; NDI: Neck Disability Index; YT20: Yasushi Tanaka cervical spondylitis symptom scale 20 scores.

Risk of bias assessment

The methodological quality of the 15 included RCTs was appraised using the Cochrane Risk of Bias tool. Seven studies (46.7%) explicitly described using SPSS for random sequence generation, while five (33.3%) mentioned randomization without detailing methods. Allocation concealment using opaque envelopes was reported in one study (6.7%). Participant and/or outcome assessor blinding was implemented in two studies (13.3%). One study (6.7%) noted protocol deviations and dropouts. No study had a pre-registered or published protocol, raising concerns regarding selective reporting bias; other bias sources remained unclear. Overall Judgment: The overall methodological rigor was limited, with a high risk of bias attributed primarily to insufficient reporting of randomization, allocation concealment, and blinding procedures. The detailed assessments are summarized in Figure 2.

Figure 2  Risk assessment of included literature.

Meta-analysis results

Effective rate of treatment: Fifteen studies (n = 1184) compared the clinical efficacy of needle-knife therapy vs traditional therapy. The heterogeneity among the studies was low (I² = 1%, P = 0.44), so the fixed-effect model was adopted. The detailed analysis results are shown in Figure 3A.

Figure 3 The difference in effective rates between the experimental group and the control group before and after treatment. A: Forest plot of treatment efficacy; B: Sensitivity analysis. CI: Confidence interval.

Analysis results indicate that the treatment efficacy rate in the experimental group was significantly superior to that in the control group, with a statistically significant difference (MD = 0.10, 95%CI: 0.07-0.14, P ≤ 0.00001). Sensitivity analysis revealed no significant reversal, indicating high reliability of the results. Detailed analysis results are presented in Figure 3B. Combined with the Egger test (P > 0.05), suggests no publication bias exists. However, it is important to note that “effective rate” is a composite and subjective outcome measure, which may limit its comparability across studies and its international clinical relevance.

Pain level (VAS score): Fifteen articles reported VAS scores as an outcome measure, involving 1184 patients. Significant heterogeneity existed among studies (I² = 99%, P < 0.0001), necessitating the use of a random-effects model. Detailed analysis results are presented in Figure 4A.

Figure 4 The difference in Visual Analog Scale scores before and after treatment. Visual Analog Scale score. A: Forest plot of Visual Analog Scale score; B: Sensitivity analysis; C-E: Subgroup analysis of Visual Analog Scale score. CI: Confidence interval.

The results of the meta-analysis indicated that the VAS scores in the experiment group were consistently lower than those in the control group (SMD = -0.93, 95%CI: -1.59 to -0.27, P = 0.006). The sensitivity analysis showed no significant reversal, suggesting the reliability of these findings is high. The specific analysis results are shown in Figure 4B. Furthermore, combined with Egger’s test analysis (P > 0.05), it indicated no significant publication bias. To investigate the source of heterogeneity, a subgroup analysis was performed on the included studies.

Subgroup analysis (VAS scores)

Intervention measures: The included studies were grouped according to the different intervention measures, categorized into the needle-knife group and the needle-knife combined with traditional treatment group. Subgroup analysis results showed that the post-treatment VAS scores in both the needle-knife group and the needle-knife combined with traditional treatment group were lower than that in the traditional treatment group, with statistically significant differences (needle-knife group: SMD = -0.12, 95%CI: 0.02-0.21, P < 0.0001; combined group: SMD = -1.40, 95%CI: -1.52 to -1.28, P < 0.0001). The detailed analysis results are shown in Figure 4C. Both groups exhibited high heterogeneity, which decreased after the exclusion of studies[11,17,18,20,23]. Potential reasons for the high initial heterogeneity include variations in needle-knife techniques, operator skill, treatment frequency/duration, and baseline patient characteristics across studies. The specific analysis results are presented in Figure 4D.

Treatment cycle: The included studies were categorized based on the duration of the treatment cycle. Studies with a treatment cycle of less than 4 weeks were assigned to group A, while those with a treatment cycle exceeding 4 weeks were designated as group B. Subgroup analysis showed that in populations receiving 2 weeks, 3 weeks, 4 weeks, or 6 weeks of treatment, the VAS scores in the experimental group were superior to those in the control group, with the differences being statistically significant (SMD = -0.45, 95%CI: -0.53 to -0.38, P < 0.0001). The detailed analysis results are presented in Figure 4E.

NDI score

Eleven studies reported the NDI score as an outcome measure, involving a total of 734 patients. Significant heterogeneity was observed among the studies (P < 0.0001, I² = 84%); therefore, a random effects model was applied. The meta-analysis results demonstrated that the experiment group was significantly superior to the control group in reducing the NDI score in patients with CSR (SMD = -4.04, 95%CI: -5.01 to -3.07, P < 0.0001), and the difference was statistically significant. The detailed analysis results are presented in Figure 5A. Sensitivity analysis indicated no significant reversal of the results, suggesting high reliability of the included studies. Specific results are shown in Figure 5B. Furthermore, Egger’s test analysis (P > 0.05) indicated no evidence of publication bias. To explore the source of heterogeneity, a subgroup analysis was performed based on different intervention measures.

Figure 5 Changes in Neck Disability Index scores before and after treatment. A: Neck Disability Index score; B: Sensitivity analysis; C and D: Subgroup analysis of Neck Disability Index score.

Subgroup analysis (NDI scores)

Intervention measures: The included studies were grouped according to the different intervention measures, categorized into the needle-knife group and the needle-knife combined with traditional treatment group. The subgroup analysis results showed that both the needle-knife group and the needle-knife combined with traditional treatment group had lower post-treatment NDI scores compared to the traditional treatment group, with statistically significant differences (needle-knife group: SMD = -3.90, 95%CI: -5.55 to -2.26, P < 0.0001; combined group: SMD = -4.19, 95%CI: -5.42 to -2.96, P < 0.0001). The detailed analysis results are presented in Figure 5C. Both subgroups exhibited high heterogeneity, which decreased after the exclusion of studies[19,21]. Potential reasons for the high initial heterogeneity may include variations in treatment protocols, considerable differences in baseline functional impairment, and substantial changes in functional scores from pre- to post-treatment. The large effect sizes observed for the NDI should be interpreted with caution, as it remains unclear whether these represent genuine treatment effects or are influenced by biases inherent in the study designs. This finding necessitates verification in more rigorously conducted trials. The detailed results of the analysis are presented in Figure 5D.

YT20 scores: Four studies reported the YT20 scores as an outcome measure, involving a total of 246 patients. Significant heterogeneity was observed among the studies (P = 0.01, I² = 72%); therefore, a random effects model was applied for the meta-analysis. The results demonstrated that the needle-knife treatment group was significantly superior to the traditional treatment group in increasing the YT20 Score in patients with CSR (SMD = 2.05, 95%CI: 1.24-2.86, Z = 4.94, P < 0.001), with the difference being statistically significant. The detailed analysis results are presented in Figure 6A. The results of the sensitivity analysis revealed no significant reversal of the analytical outcomes, indicating a high reliability of the included studies. Specific results are presented in Figure 6B. Furthermore, the Egger’s test analysis (P > 0.05) suggested no evidence of publication bias.

Figure 6 Changes in the Yasushi Tanaka cervical spondylitis symptom scale 20 scores score before and after treatment. A: Yasushi Tanaka cervical spondylitis symptom scale 20 scores; B: Sensitivity analysis.

Safety evaluation

Four studies compared the adverse reaction situation between the two groups of patients. The heterogeneity test indicated low heterogeneity among the studies (I² = 0, P = 0.92); therefore, a fixed effects model was used for analysis. The meta-analysis results showed that the incidence of adverse events in the needle-knife group was lower than that in the traditional treatment group, and the difference was statistically significant (MD = 0.34, 95%CI: 0.14-0.83, Z = 2.36, P = 0.02). The detailed analysis results are presented in Figure 7.

Figure 7 Safety evaluation. CI: Confidence interval.

Sensitivity analysis

A sensitivity analysis was performed by sequentially removing each individual study. The conclusions were considered robust if the pooled effect size did not change substantially and the direction of the forest plot remained consistent. The results indicated that sequentially omitting any single study had minimal impact on the pooled effect sizes for the NDI score, VAS score, YT20, and treatment efficacy rate. The Meta-analysis demonstrated low sensitivity, suggesting that the findings are relatively reliable.

Publication bias

The funnel plot was constructed using the effective rate as the analysis indicator. Data points were distributed on both sides of the central line. All points fell within the 95% confidence interval, with a relatively even distribution across the top and bottom of the plot, though fewer points were located on the left side. The distribution of data points was narrow, indicating low heterogeneity. The specific analysis results are shown in Figure 8.

Figure 8  Publication bias.

DISCUSSION

This study followed Cochrane guidelines to conduct a comprehensive analysis of the 15 included articles, evaluating the efficacy and safety of needle-knife treatment for CSR. Meta-analysis results indicated that needle-knife treatment for CSR was significantly superior to conventional treatments in terms of treatment effective rate, safety, and scale scores, with the differences being statistically significant (P < 0.05). The evidence has certain limitations. The primary limitation of this evidence is the high risk of bias in the included trials. Inadequate reporting and likely inadequate implementation of randomization, allocation concealment, and blinding significantly undermine the internal validity of the individual studies and, consequently, the reliability of our study. Further subgroup analysis explored the efficacy and heterogeneity of needle-knife treatment for CSR from the perspectives of intervention measures and treatment cycle. The results showed that needle-knife treatment maintained a consistent direction of therapeutic effect across all subgroups. This indicates that the efficacy of needle-knife treatment for CSR is relatively robust; different patient characteristics and treatment plans do not significantly diminish its therapeutic effect, providing a reliable basis for individualized clinical treatment. However, the high heterogeneity within subgroups tempers the strength of these conclusions.

The mechanisms of needle-knife treatment for CSR primarily involve the following aspects: (1) Relieving soft tissue adhesion and spasm: Through its cutting and separation effects, the needle-knife releases adhesions in cervical soft tissues such as muscles, ligaments, and fascia, improves local blood circulation, alleviates muscle tension and spasm, and thereby reduces compression and irritation of the nerve roots[24]; (2) Improving local microcirculation: The stimulation of local tissues by needle-knife promotes vasodilation, increases local blood flow, enhances local microcirculation, and accelerates the uptake of inflammatory substances and the elimination of metabolic products[25]; (3) Restoring cervical biomechanical balance: By releasing the soft tissues surrounding the spine, needle-knife helps correct mechanical imbalances in the cervical spine, restores the normal physiological curvature, and reduces mechanical compression on the nerve roots[26]; (4) Regulating nerve function: Needle-knife stimulation can activate afferent fibers (such as Abeta, Adelta, and C-fos fibers), triggering the release of various signaling molecules (e.g., neurotransmitters and endogenous opioids). This process modulates nerve function and alleviates local pain[27]; and (5) Reducing neurotransmitter levels: Needle-knife treatment can significantly downregulate the level of substance P. Substance P plays a crucial role in pain signal transmission, and the reduction in its level contributes to the alleviation of pain symptoms in patients[28]. The detailed mechanism diagram is shown in Figure 9.

Figure 9 Needle-knife mechanism of action. MMP: Matrix metalloproteinase; TIMP: Tissue Inhibitor of metalloproteinase; ECM: Extracellular matrix; TNF: Tumor necrosis factor; IL: Interleukin.

This study also has several limitations: (1) Although all included studies mentioned randomization, seven did not specify the specific randomization method; (2) Most studies did not address allocation concealment or the blinding method; (3) The sample sizes varied considerably, with a maximum of 100 and a minimum of 18 participants; (4) Five studies involved additional treatment modalities in the experimental group, rather than pure needle-knife treatment; (5) Most trials did not clearly report adverse reactions or long-term outcomes, making it impossible to assess the safety of needle-knife treatment; (6) Among the included studies, the needle-knife group’s treatment protocols, session durations, and needling sites differed, introducing significant subjectivity; (7) The high statistical heterogeneity for key outcomes (VAS, NDI) was not fully resolved by subgroup analyses; and (8) “Effective rate” is a composite and subjective outcome measure, which may limit its comparability across studies and its international clinical relevance.

CONCLUSION

The preliminary findings of this study indicate that needle-knife treatment holds considerable advantages for CSR, providing a theoretical basis for its clinical application in managing this condition.

However, the low quality and high heterogeneity of the included trials preclude definitive conclusions. The findings of this meta-analysis should be considered preliminary evaluation. Future research should involve more large-scale, high-quality RCT studies, with emphasis on proper implementation of randomization and blinding methods. Standardized improvements should be made to the needle-knife treatment plan and course of therapy, with particular attention paid to adverse reactions associated with needle-knife therapy. Future trials should prioritize the use of internationally recognized, patient-centered outcome measures (e.g., Medical Outcomes Study 36-Item Short-Form Health Survey) and clearly differentiate between needle-knife monotherapy and combination therapies. They should also include longer-term follow-up to assess the durability of effects. Conducting research in a more standardized and rigorous manner will yield higher-quality evidence for this field.