Efficacy of Superficial Acupuncture Stimulation with Dynamic Mobilization in Chronic Neck Pain: A Multicenter Randomized Controlled Trial

Abstract

Introduction

This multicenter randomized controlled trial investigated the clinical efficacy and durability of shallow acupuncture (SA; superficial needling with concurrent active mobilization) for chronic neck pain (CNP).

Methods

A total of 252 participants were randomized into SA, celecoxib, or wait-list groups. The SA intervention involved standardized superficial stimulation at biomechanically sensitized points coupled with active neck mobilization. Clinical outcomes, including Overall Clinical Efficacy, Visual Analog Scale (VAS), Neck Disability Index (NDI), Self-Rating Anxiety Scale (SAS), 36-Item Short-Form Health Survey (SF-36) and Cervical Range of Motion (CROM), were evaluated at baseline (T0), post-treatment (T1), and 3-month follow-up (T2).

Results

238 participants were included in the primary analysis (14 dropouts). At T1, the SA group (n=81) demonstrated significant clinical superiority over both pharmacological (n=79) and wait-list controls (n=78). Beyond achieving profound analgesia, SA facilitated superior functional restoration and psychological recalibration. Crucially, these multidimensional therapeutic benefits were robustly sustained through the 3-month follow-up, whereas the effects of the pharmacological intervention diminished significantly after treatment cessation. The intervention was well-tolerated with no serious adverse events reported.

Conclusion

Superficial needling combined with active mobilization is a safe, clinically effective, and durable non-pharmacological intervention for chronic neck pain, providing superior and sustained pain relief and functional restoration compared to conventional pharmacological care. This trial was registered at the International Traditional Medicine Clinical Trial Registration Platform (ITMCTR, Registration Number: ITMCTR2024000219).

Introduction

Chronic neck pain (CNP) has become a major global public health concern and is a leading cause of long-term disability and reduced quality of life.¹ The widespread use of digital devices and sedentary habits have led to a surge in musculoskeletal imbalances, which result in persistent neck strain.^2,3 Beyond the physical sensation of pain, CNP significantly limits range of motion and is frequently linked to psychological distress, such as anxiety.⁴ This creates a heavy socioeconomic burden for both individuals and healthcare systems.⁵

Standard management for chronic pain relies heavily on medication, with selective Cyclooxygenase-2 (COX-2) inhibitors.^6,7 Although drugs effectively reduce inflammation and provide systemic pain relief, their long-term use is limited by potential gastrointestinal and cardiovascular risks.⁸ Furthermore, drug treatments often fail to address the physical aspects of the disease, such as high muscle tension and structural stiffness.⁹ This relief-relapse cycle highlights the urgent need for more targeted, non-drug alternatives.

Current clinical guidelines increasingly recommend non-drug interventions, with acupuncture being a key option.^10,11 However, it is important to distinguish between conventional and superficial dry acupuncture.¹² Conventional acupuncture typically involves deep needle insertion into muscle tissue to trigger De-qi sensations (sense of soreness, numbness, or heaviness). While effective, this high-intensity approach can cause needle fear, low patient tolerance, or a small risk of deep tissue injury. In contrast, shallow acupuncture (SA) is a minimally invasive technique that targets only the skin and shallow fascia layers, avoiding deep muscle penetration, coupled with active neck mobilization during needle retention.¹³ Despite its clinical promise and high safety profile, evidence for shallow acupuncture remains limited, with most studies being small or single-center pilots. This multicenter randomized controlled trial aims to fill this gap by evaluating the efficacy and safety of shallow acupuncture versus celecoxib and a wait-list control in improving pain intensity, neck function, and overall quality of life, while providing a clinical framework to explore potential superficial mechanotransduction mechanisms as prospective pathways underlying shallow acupuncture.

Methods

Study Design and Ethical Considerations

This study was a multicenter, randomized, parallel-group, controlled clinical trial. It was conducted across three independent clinical centers (Center 1: Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Center 2: First Teaching Hospital of Tianjin University of Traditional Chinese Medicine and Center 3: Shenzhen Bao’an Traditional Chinese Medicine Hospital) in accordance with the Declaration of Helsinki and approved by the centers’ Ethics Committees (Approval Numbers: BF2024-042-01, TYLL2024[K]No.021, KY-2025-007-01, respectively). The trial was registered at the International Traditional Medicine Clinical Trial Registration Platform (ITMCTR, Registration Number: ITMCTR2024000219). Prior to study screening and enrollment, the study objectives, procedures, potential risks, and benefits were thoroughly explained to all potential participants. Standard written informed consent was obtained from each participant prior to study commencement. Trial period: August 31, 2024 to November 30, 2025. The study was conducted in strict adherence to the Consolidated Standards of Reporting Trials (CONSORT) statement, and the Standards for Reporting Interventions in Clinical Trials of Acupuncture (STRICTA) standards. The detailed study protocol for this study has been previously published.¹⁴ The current report focuses on the clinical efficacy of shallow acupuncture for chronic neck pain, which was defined as the primary objectives of this multicenter trial. To maintain a concise focus on clinical outcomes and functional recovery, the neuroimaging data, intended to explore the underlying central mechanisms of the intervention, will be presented in a separate specialized manuscript currently under preparation.

Participants

A total of 252 patients with CNP were enrolled, with 84 participants recruited from each of the three clinical centers. The detailed parameters and calculation process for sample size estimation have been pre-specified and described in the study protocol.¹⁴ Inclusion criteria were as follows: (1) diagnosis of neck-type or radicular cervical spondylosis without prior surgical intervention; (2) primary complaint of neck and shoulder pain for >3 months; (3) age 18–75 years; (4) Visual Analog Scale (VAS) score between 4 and 6; and (5) conscious and capable of providing informed consent. Exclusion criteria included: (1) severe systemic diseases or history of significant psychiatric disorders; (2) use of antidepressants, analgesics, or corticosteroids within the past 30 days; (3) pregnancy or lactation; (4) contraindications to acupuncture or bleeding risks; and (5) known allergy to celecoxib.

Randomization and Blinding

Participants were randomly assigned to the shallow acupuncture group, the celecoxib control group, or the wait-list group in a 1:1:1 ratio within each of the three clinical centers. Randomization was performed using a block randomization method. An independent statistician, who was not involved in the clinical implementation or recruitment, generated the randomization sequence using a computer. To ensure rigorous allocation concealment, the sequence was placed in sequentially numbered, opaque, sealed envelopes. These envelopes were kept by a dedicated coordinator and were only opened after the participant had completed the baseline assessment and confirmed their eligibility.

Due to the distinct nature of the interventions (physical therapy versus pharmacological treatment), participant blinding was not feasible, and no sham control was included. To minimize ascertainment bias, a strict blinding protocol was maintained for all other key personnel throughout the study: outcome assessors, data collectors, and statisticians were blinded to group allocations and specific study hypotheses, while acupuncturists and clinical staff remained masked to the overall trial design and allocation across other centers. All blinded personnel remained masked until the statistical analysis was finalized and the blind was officially unmasked. Nevertheless, assessor and analyst blinding alone cannot eliminate patient-level expectancy effects or non-specific placebo responses inherent to patient-reported pain outcomes.

Interventions

To ensure methodological consistency, interventions at all three centers were performed by senior acupuncturists. Prior to the study, all practitioners underwent standardized training in the specific shallow acupuncture protocol and successfully completed rigorous proficiency assessments.

Shallow Acupuncture (SA) Group

Participants in the SA group received treatment according to a strictly standardized protocol. To ensure clinical precision and address the potential heterogeneity of chronic neck pain, a standardized palpation-based approach was employed instead of a fixed acupoint formula. Intervention sites were localized through systematic palpation of the posterior neck region by experienced practitioners. These “pathological reaction points” were identified based on a rigorous set of objective cutaneous and structural manifestations (details in protocol). The intervention zone for shallow acupuncture was precisely demarcated within the posterior cervical region, delineated by the following anatomical landmarks: the superior boundary consisted of the superior nuchal line and the external occipital protuberance; the lateral boundaries were defined by the medial borders of the sternocleidomastoid muscles; and the inferior boundary was established by a horizontal line passing through the spinous process of the seventh cervical vertebra (C7).

Operationally, SA utilized conventional acupuncture needles targeting reaction points, with needles inserted solely into the skin and superficial fascia without seeking De-qi or penetrating deep muscles. During needle retention, participants performed active neck movements across standard planes within their pain-free range. Specifically, disposable sterile filiform needles (0.25×25 mm, Beijing Hanyi Medical Instruments Co., Ltd., China) were utilized. The needle tip was directed toward the primary site of pain, followed by a shallow insertion at an angle of 10°–25° relative to the skin surface. The insertion depth and angle were adjusted in real-time based on the patient’s immediate feedback regarding symptomatic alleviation. Needles were retained for 15 minutes, during which patients performed gentle, active neck and shoulder movements. The treatment regimen comprised three sessions per week on alternate days for two consecutive weeks, totaling 6 sessions of shallow acupuncture treatment.

Celecoxib Control Group

Participants assigned to the pharmacological control group were prescribed oral celecoxib capsules (Shiyao Group Ouyi Pharmaceutical Co., Ltd., China, State Medical Permit No. H20203296) at a standardized dosage of 100 mg twice daily. The medication was administered for a continuous duration of 14 days. To monitor treatment adherence and safety, participants were required to maintain a daily medication log, and any unused capsules were collected and counted at the end of the 2-week intervention period.

Wait-List Control Group

Participants in the wait-list group underwent a 2-week observation period, during which they were strictly instructed not to receive any form of therapeutic intervention for their neck pain, including acupuncture, physical therapy, or analgesics. Following the completion of this waiting period and the associated post-treatment assessments, these participants were offered the identical standardized shallow acupuncture protocol as the Shallow Acupuncture group. This delayed-treatment design was implemented to ensure ethical parity and to maintain participant recruitment motivation while providing a true natural-history control for the primary endpoint.

Outcome Measurement

Assessments were performed at baseline (T0), post-treatment (T1, day 14), and 3-month follow-up (T2). Detailed descriptions of these measures were previously published in the study protocol.

Primary Outcomes

(1) Clinical Efficacy Rate: According to the Guidelines for Evaluation of traditional Chinese medicine (TCM) Diagnosis and Treatment (DB44/T 1425–2014) and General Rules for the Formulation and Revision of the Diagnosis and Efficacy Evaluation Standards of TCM Diseases and Syndromes (ZY/T 10–2024), participants were categorized into four levels based on symptomatic relief: Cured (complete resolution of symptoms), Marked improvement (significant symptom relief), Improvement (reduction in pain frequency and intensity), or No effect.

(2) Visual Analog Scale (VAS): Used to quantify pain intensity on a 0–10 cm scale, where 0 represented no pain and 10 represented the most severe pain imaginable.

(3) Neck Disability Index (NDI):¹⁵ A 10-item instrument employed to evaluate the impact of neck pain on activities of daily living.

Secondary Outcomes

(1) Self-Rating Anxiety Scale (SAS):¹⁶ Used to assess the presence and severity of anxiety-related symptoms.

(2) 36-Item Short-Form Health Survey (SF-36):¹⁷ A comprehensive tool used to evaluate multi-dimensional health-related quality of life.

(3) Cervical Range of Motion (CROM):¹⁸ Measured in six planes—flexion, extension, bilateral rotation, and bilateral lateral flexion—using EasyAngle (Meloq, Stockholm, Sweden), a digital goniometer.

Statistical Analysis

A unified Electronic Data Capture (EDC) system provided by National Clinical Research Center for Chinese Medicine Acupuncture and Moxibustion was employed across the three centers for data entry and verification. Statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY). Continuous variables were summarized as mean ± standard deviation (SD) with 95% confidence intervals (CIs). Categorical data were presented as frequencies and percentages. The primary and secondary outcomes were evaluated using a linear mixed-effects model (LMM) to account for the repeated-measures design and potential correlation within subjects. The model included Treatment group (SA, Celecoxib, and Wait-list), Time (T0, T1, and T2), and the Group × Time interaction as fixed effects. To address the multicenter nature of the trial and the observed inter-site variability, “Center” was incorporated into the LMM as a random effect. Baseline scores for each outcome were included as covariates to adjust for initial severity. For significant interaction effects, post-hoc pairwise comparisons were performed using Bonferroni correction to control for multiple testing. Between-group effect sizes were calculated as Cohen’s d = (Mean_intervention − Mean_control)/SD_pooled. Under this sign convention, negative d values represent lower scores in the intervention group relative to controls (eg, VAS), whereas positive d values represent higher scores in the intervention group (eg, CROM). Effect sizes were classified as small (|d| = 0.2), medium (|d| = 0.5), and large (|d| = 0.8). All statistical tests were two-sided, with the significance level set at P < 0.05. To control for multiplicity across endpoints, the three co-primary outcomes were evaluated under an intersection-union framework requiring all three to reach significance at α = 0.05. Multiplicity across multidimensional secondary outcomes was controlled using the Benjamini–Hochberg false discovery rate (FDR) procedure, with FDR-adjusted P < 0.05 defining statistical significance.

Results

Participant Flow

A total of 252 patients with CNP were initially enrolled and randomized from 3 centers. Fourteen participants (5.56%) were excluded from the final analysis due to dropout. The final analysis comprised 238 participants: 81 in the Shallow Acupuncture group, 79 in the Celecoxib group, and 78 in the Wait-list group. The CONSORT flow chart is presented in Figure 1.

Figure 1.

CONSORT flow diagram for the trial.

Baseline Characteristics

Baseline demographic and clinical characteristics are summarized in Table 1. Statistical analysis confirmed that there were no significant differences among the three groups regarding age (F = 0.19, P = 0.824), gender distribution (χ² = 0.44, P = 0.801), educational level (χ² = 6.12, P = 0.410) or disease duration (H = 0.86, P = 0.652).

Table 1.

Baseline Characteristics of Participants

Characteristic	SA (n=81)	Celecoxib (n=79)	Wait-List (n=78)	Statistical	P
Gender [n (%)]				χ2 = 0.44	0.801
- Female	61 (75.3%)	56 (70.9%)	58 (74.4%)
- Male	20 (24.7%)	23 (29.1%)	20 (25.6%)
Age (years) [95% CI]	37.70 ± 12.21 [35.00, 40.40]	38.70 ± 13.47[35.68, 41.72]	38.96 ± 14.70 [35.65, 42.27]	F = 0.19	0.824
Educational level [n (%)]				χ2 = 6.12	0.410
- Postgraduate or above	29 (35.8%)	25 (31.6%)	21 (26.9%)
- Undergraduate	39 (48.1%)	32 (40.5%)	38 (48.7%)
- Secondary school	11 (13.6%)	21 (26.6%)	16 (20.5%)
- Primary school or below	2 (2.5%)	1 (1.3%)	3 (3.8%)
Duration (months, M [IQR])	39.52 [16.35, 78.08]	29.60 [15.92, 70.60]	37.16 [10.00, 87.75]	H = 0.86	0.652

Abbreviations: SA, Shallow acupuncture; IQR, Interquartile Range.

Overall Clinical Efficacy

At T1, the distribution of clinical efficacy grades exhibited significant variation across the three groups (Kruskal–Wallis H = 112.38, P < 0.001). The total effective rate in the Shallow Acupuncture group was 90.12% (73/81), which was significantly higher than those in the Celecoxib group (77.22%, 61/79; χ² = 3.99, P = 0.046) and the Wait-list group (17.95%, 14/78; χ² = 80.65, P < 0.001). Specifically, 71.6% (58/81) of participants in the Shallow Acupuncture group achieved either Cured or Marked Improvement status, compared with 25.3% (20/79, P < 0.001) in the Celecoxib group. Additionally, a logistic regression analysis was conducted to test for the interaction between treatment group and clinical center. The interaction term was not statistically significant (P = 0.285). Results of clinical efficacy are presented in Table 2.

Table 2.

Comparison of Clinical Efficacy Among Three Groups

Group	N	Cured	Marked Improvement	Improvement	No Effect	Total Effective Rate
Shallow Acupuncture	81	19 (23.5%)	39 (48.1%)	15 (18.5%)	8 (9.9%)	90.12%
Celecoxib Control	79	6 (7.6%)	14 (17.7%)	41 (51.9%)	18 (22.8%)	77.22%
Wait-list Group	78	0 (0.0%)	1 (1.3%)	13 (16.7%)	64 (82.1%)	17.95%

Notes: Total Effective Rate = n (Cured + Marked Improvement + Improvement)/N.

VAS

At T0, pain intensity was comparable across groups (SA: 5.27 ± 0.97, 95% CI [5.05, 5.48]; Celecoxib: 5.30 ± 0.91, 95% CI [5.10, 5.50]; and Wait-list: 5.18 ± 0.85, 95% CI [4.99, 5.38]). At T1, the SA group achieved a significant reduction in pain (2.23 ± 1.30, 95% CI [1.94, 2.51]), demonstrating superior efficacy over celecoxib (3.83 ± 1.12, 95% CI [3.58, 4.09]; Mean Diff: −1.57, 95% CI [−1.88, −1.26], P < 0.001, Cohen’s d = −1.330) and the wait-list control (5.06 ± 1.12; Mean Diff: −2.89, 95% CI [−3.20, −2.58], P < 0.001, Cohen’s d = −2.712). Furthermore, at T1, the celecoxib group also showed significant improvement compared to the wait-list control (Mean Diff: −1.32, P < 0.001, Cohen’s d = −1.336). These gains were robustly sustained at T2, where SA (2.52 ± 1.35, 95% CI [2.22, 2.82]) remained markedly more effective than celecoxib (3.97 ± 1.54, 95% CI [3.62, 4.31]; Mean Diff: −1.40, 95% CI [−1.78, −1.01], P < 0.001, Cohen’s d = −0.918) and the wait-list control (3.66 ± 1.58; Mean Diff: −1.18, 95% CI [−1.56, −0.79], P < 0.001, Cohen’s d = −0.816). Notably, at T2, no significant difference was observed between the celecoxib and wait-list groups (Mean Diff: 0.22, P = 0.266, Cohen’s d = 0.118). While a significant center effect was observed regarding VAS scores at T1 (F [2, 234] = 4.54, P = 0.012) and T2 (F [2, 234] = 16.21, P < 0.001), the clinical trend of SA achieving lower pain intensity relative to controls was universally sustained across all settings. No significant Group × Center interactions were observed at T1 or T2. Results of VAS scores are presented in Table 3 and Figure 2. And Center-specific VAS statistics are presented in Table 4.

Table 3.

Summary of VAS Among Three Groups

Outcome & Comparison	Baseline (T0)	Post-Treatment (T1)	3-Month Follow-Up (T2)
	Mean ± SD [95% CI]	Mean ± SD [95% CI]	Mean ± SD [95% CI]
Shallow acupuncture (n=81)	5.27 ± 0.97 [5.05, 5.48]	2.23 ± 1.30 [1.94, 2.51]	2.52 ± 1.35 [2.22, 2.82]
Celecoxib (n=79)	5.30 ± 0.91 [5.10, 5.50]	3.83 ± 1.12 [3.58, 4.09]	3.97 ± 1.54 [3.62, 4.31]
Wait-list (n=78)	5.18 ± 0.85 [4.99, 5.38]	5.06 ± 1.12 [4.81, 5.31]	3.66 ± 1.58 [3.30, 4.01]
		Diff [95% CI] (P, Cohen’s d)	Diff [95% CI] (P, Cohen’s d)
SA vs Celecoxib		−1.57 [−1.88, −1.26] (<0.001, −1.330)	−1.40 [−1.78, −1.01] (<0.001, −0.918)
SA vs Wait-list		−2.89 [−3.20, −2.58] (<0.001, −2.712)	−1.18 [−1.56, −0.79] (<0.001, −0.816)
Celecoxib vs Wait-list		−1.32 [−1.63, −1.01] (<0.001, −1.336)	0.22 [−0.17, 0.61] (0.266, 0.118)
Center Effect (P)	—	0.012	< 0.001

Abbreviation: SA, Shallow acupuncture.

Figure 2.

Changes in Visual Analog Scale (VAS) scores across three groups. Data are presented as mean ± standard deviation (SD). * indicates statistically significant differences compared with the Shallow Acupuncture group at the same time point (P < 0.05). # indicates statistically significant differences compared to baseline within each group (P < 0.05).

Abbreviations: VAS, Visual Analog Scale; T0, baseline; T1, post-treatment; T2, 3-month follow-up.

Table 4.

Center-Specific VAS Statistics (N=238)

Clinical Center	Group	Post-Treatment (T1)	3-Month Follow-Up (T2)
		Mean ± SD [95% CI]	Mean ± SD [95% CI]
Center 1	SA (n=27)	1.77 ± 0.73 [1.48, 2.06]	1.92 ± 0.75 [1.62, 2.22]
	Celecoxib (n=28)	4.05 ± 0.73 [3.77, 4.33]	4.06 ± 0.68 [3.80, 4.32]
	Wait-list (n=24)	5.13 ± 0.44 [4.94, 5.32]	4.12 ± 0.62 [3.86, 4.38]
Center 2	SA (n=28)	2.63 ± 1.10 [2.20, 3.06]	3.16 ± 1.43 [2.61, 3.71]
	Celecoxib (n=28)	4.13 ± 1.14 [3.69, 4.57]	4.93 ± 1.21 [4.46, 5.40]
	Wait-list (n=28)	5.54 ± 0.91 [5.19, 5.89]	4.66 ± 1.23 [4.18, 5.14]
Center 3	SA (n=26)	2.27 ± 1.76 [1.56, 2.98]	2.44 ± 1.47 [1.85, 3.03]
	Celecoxib (n=23)	3.22 ± 1.28 [2.67, 3.77]	2.67 ± 1.77 [1.90, 3.44]
	Wait-list (n=26)	4.48 ± 1.48 [3.88, 5.08]	2.15 ± 1.43 [1.57, 2.73]
Summary	Center Effect (P)	0.012	< 0.001

Abbreviation: SA, Shallow acupuncture.

NDI

NDI scores were comparable across the three groups at T0 (SA: 45.60 ± 9.63, 95% CI [43.48, 47.73]; Celecoxib: 47.85 ± 11.33, 95% CI [45.31, 50.39]; and Wait-list: 44.28 ± 8.88, 95% CI [42.28, 46.29]). At T1, the SA group achieved a significant improvement in neck function (32.07 ± 7.46, 95% CI [30.42, 33.72]), demonstrating superior efficacy over celecoxib (41.34 ± 10.97; Mean Diff: −7.81, 95% CI [−10.02, −5.60], P < 0.001, Cohen’s d = −0.809) and the wait-list control (43.15 ± 10.02; Mean Diff: −11.93, 95% CI [−14.14, −9.72], P < 0.001, Cohen’s d = −1.592). The celecoxib group also significantly outperformed the wait-list group at T1 (Mean Diff: −4.12, P < 0.001). These functional gains were robustly sustained through the 3-month follow-up (T2), where SA (33.68 ± 8.08, 95% CI [31.89, 35.47]) remained markedly more effective than celecoxib (41.90 ± 10.42; Mean Diff: −7.05, P < 0.001, Cohen’s d = −0.649) and the wait-list control (39.00 ± 8.67; Mean Diff: −5.95, P < 0.001, Cohen’s d = −0.748). Notably, at T2, the difference between the celecoxib and wait-list groups was no longer statistically significant (P = 0.355). While the center effect was not significant at T1 (P = 0.345), it became significant at T2 (F [2, 234] = 4.54, P = 0.011). However, the overall trend favoring SA was consistently maintained across all participating sites. No significant Group × Center interactions were observed at T1 or T2. Results of NDI scores are presented in Table 5 and Figure 3. And Center-specific NDI statistics are presented in Table 6.

Table 5.

Summary of NDI Among Three Groups

Outcome & Comparison	Baseline (T0)	Post-Treatment (T1, 2-wk)	3-Month Follow-Up (T2)
	Mean ± SD [95% CI]	Mean ± SD [95% CI]	Mean ± SD [95% CI]
Shallow Acupuncture (n=81)	45.60 ± 9.63 [43.48, 47.73]	32.07 ± 7.46 [30.42, 33.72]	33.68 ± 8.08 [31.89, 35.47]
Celecoxib (n=79)	47.85 ± 11.33 [45.31, 50.39]	41.34 ± 10.97 [38.88, 43.80]	41.90 ± 10.42 [39.56, 44.23]
Wait-list (n=78)	44.28 ± 8.88 [42.28, 46.29]	43.15 ± 10.02 [40.89, 45.41]	39.00 ± 8.67 [37.05, 40.95]
		Diff [95% CI] (P, Cohen’s d)	Diff [95% CI] (P, Cohen’s d)
SA vs Celecoxib		−7.81 [−10.02, −5.60] (<0.001, −0.809)	−7.05 [−9.35, −4.76] (<0.001, −0.649)
SA vs Wait-list		−11.93 [−14.14, −9.72] (<0.001, −1.592)	−5.95 [−8.25, −3.65] (<0.001, −0.748)
Celecoxib vs Wait-list		−4.12 [−6.37, −1.88] (<0.001, −0.737)	1.10 [−1.23, 3.44] (0.355, 0.075)
Center Effect (P)	—	0.345	0.011

Abbreviation: SA, Shallow acupuncture.

Figure 3.

Changes in Neck Disability Index (NDI) scores across three groups. Data are presented as mean ± standard deviation (SD). * indicates statistically significant differences compared with the Shallow Acupuncture group at the same time point (P < 0.05). # indicates statistically significant improvements compared to baseline within each group (P < 0.05).

Abbreviations: NDI, Neck Disability Index; T0, baseline; T1, post-treatment; T2, 3-month follow-up.

Table 6.

Center-Specific NDI Statistics (N=238)

Clinical Center	Group	Post-Treatment (T1, 2-wk)	3-Month Follow-Up (T2)
		Mean ± SD [95% CI]	Mean ± SD [95% CI]
Center 1	SA (n=27)	29.26 ± 7.11 [26.45, 32.07]	28.96 ± 5.72 [26.70, 31.22]
	Celecoxib (n=28)	39.36 ± 8.33 [36.13, 42.59]	38.86 ± 7.93 [35.79, 41.93]
	Wait-list (n=24)	44.42 ± 7.69 [41.17, 47.67]	40.50 ± 7.60 [37.29, 43.71]
Center 2	SA (n=28)	33.00 ± 7.07 [30.26, 35.74]	35.21 ± 7.70 [32.22, 38.20]
	Celecoxib (n=28)	42.14 ± 10.57 [38.04, 46.24]	44.29 ± 10.51 [40.21, 48.37]
	Wait-list (n=28)	44.50 ± 12.14 [39.79, 49.21]	42.00 ± 9.86 [38.18, 45.82]
Center 3	SA (n=26)	34.00 ± 7.63 [30.92, 37.08]	36.92 ± 8.55 [33.47, 40.37]
	Celecoxib (n=23)	42.78 ± 14.02 [36.72, 48.84]	42.70 ± 12.35 [37.36, 48.04]
	Wait-list (n=26)	40.54 ± 9.22 [36.82, 44.26]	34.38 ± 6.22 [31.87, 36.89]

Abbreviation: SA, Shallow acupuncture.

SAS

SAS scores were comparable across the three groups at T0 (SA: 47.00 ± 8.88, 95% CI [45.04, 48.96]; Celecoxib: 45.24 ± 9.54, 95% CI [43.10, 47.38]; and Wait-list: 43.60 ± 9.05, 95% CI [41.56, 45.64]). At T1, the SA group showed a significant reduction in anxiety (39.17 ± 7.86, 95% CI [37.43, 40.91]), demonstrating superior efficacy over celecoxib (41.67 ± 8.57; Mean Diff: −3.72, 95% CI [−5.54, −1.89], P < 0.001, Cohen’s d = −0.602) and the wait-list control (42.55 ± 9.48; Mean Diff: −5.74, 95% CI [−7.59, −3.90], P < 0.001, Cohen’s d = −1.090). The celecoxib group also showed a mild but significant improvement compared to the wait-list group at T1 (Mean Diff: −2.03, P = 0.031). These improvements were sustained through the 3-month follow-up (T2), where the SA group (39.02 ± 7.38, 95% CI [37.39, 40.66]) remained more effective than celecoxib (41.15 ± 8.70; Mean Diff: −3.26, P < 0.001, Cohen’s d = −0.568) and the wait-list control (41.19 ± 8.25; Mean Diff: −4.35, P < 0.001, Cohen’s d = −0.920). Notably, at T2, the difference between the celecoxib and wait-list groups was no longer statistically significant (Mean Diff: −1.09, P = 0.220). The center effect was not significant at either T1 (P = 0.307) or T2 (P = 0.101). No significant Group × Center interactions were observed at T1 or T2. Results of SAS scores are presented in Table 7 and Figure 4.

Table 7.

Summary of SAS Among Three Groups

Outcome & Comparison	Baseline (T0)	Post-Treatment (T1, 2-wk)	3-Month Follow-Up (T2)
	Mean ± SD [95% CI]	Mean ± SD [95% CI]	Mean ± SD [95% CI]
Shallow Acupuncture (n=81)	47.00 ± 8.88 [45.04, 48.96]	39.17 ± 7.86 [37.43, 40.91]	39.02 ± 7.38 [37.39, 40.66]
Celecoxib (n=79)	45.24 ± 9.54 [43.10, 47.38]	41.67 ± 8.57 [39.75, 43.59]	41.15 ± 8.70 [39.20, 43.10]
Wait-list (n=78)	43.60 ± 9.05 [41.56, 45.64]	42.55 ± 9.48 [40.41, 44.69]	41.19 ± 8.25 [39.33, 43.05]
		Diff [95% CI] (P, Cohen’s d)	Diff [95% CI] (P, Cohen’s d)
SA vs Celecoxib		−3.72 [−5.54, −1.89] (<0.001, −0.602)	−3.26 [−4.98, −1.54] (<0.001, −0.568)
SA vs Wait-list		−5.74 [−7.59, −3.90] (<0.001, −1.090)	−4.35 [−6.09, −2.60] (<0.001, −0.920)
Celecoxib vs Wait-list		−2.03 [−3.88, −0.18] (0.031, −0.409)	−1.09 [−2.83, 0.65] (0.220, −0.262)
Center Effect (P)	—	0.307	0.101

Abbreviation: SA, Shallow acupuncture.

Figure 4.

Changes in Self-Rating Anxiety Scale (SAS) scores across three groups. Data are presented as mean ± standard deviation (SD). * indicates statistically significant differences compared with the Shallow Acupuncture group at the same time point (P < 0.05). # indicates statistically significant improvements compared to baseline within each group (P < 0.05).

Abbreviations: SAS, Self-Rating Anxiety Scale; T0, baseline; T1, post-treatment; T2, 3-month follow-up.

SF-36

Baseline SF-36 scores across all eight dimensions were comparable among the three groups at T0 (all P > 0.05). At T1, the SA group demonstrated significant improvements across multiple dimensions compared to the celecoxib group, including Physical Functioning (PF; SA: 90.14, 95% CI [87.77, 92.51] vs Celecoxib: 84.68, 95% CI [81.31, 88.05]; Mean Diff: 4.00, 95% CI [1.48, 6.52], P = 0.002, Cohen’s d = 0.386), Bodily Pain (BP; SA: 80.68, 95% CI [78.39, 82.96] vs Celecoxib: 74.30, 95% CI [71.82, 76.79]; Mean Diff: 5.79, 95% CI [2.85, 8.73], P < 0.001, Cohen’s d = 0.432), General Health (GH; SA: 66.48, 95% CI [61.46, 71.50] vs Celecoxib: 62.15, 95% CI [57.65, 66.66]; Mean Diff: 4.03, 95% CI [0.25, 7.80], P = 0.036, Cohen’s d = 0.300), and Vitality (VT; SA: 72.70, 95% CI [69.50, 75.89] vs Celecoxib: 68.16, 95% CI [64.41, 71.92]; Mean Diff: 5.72, 95% CI [2.37, 9.07], P = 0.001, Cohen’s d = 0.472). Compared to the wait-list control, the SA group achieved significantly higher scores in all eight dimensions at T1 (all P < 0.05).

At T2, the superiority of SA over celecoxib was sustained in the dimensions of PF (Mean Diff: 2.84, 95% CI [0.06, 5.63], P = 0.046), BP (Mean Diff: 6.40, 95% CI [3.80, 9.01], P < 0.001), GH (Mean Diff: 5.25, 95% CI [1.45, 9.05], P = 0.007), and VT (Mean Diff: 3.50, 95% CI [0.31, 6.69], P = 0.032). In contrast, the celecoxib group showed no significant difference compared to the wait-list group in most dimensions at T2, including PF (P = 0.478) and GH (P = 0.963). Significant center effects were observed primarily at the 3-month follow-up (T2) for dimensions including Physical Functioning (PF; F [2, 232] = 8.91, P < 0.001), Bodily Pain (BP; F [2, 232] = 6.43, P = 0.002), and General Health (GH; F [2, 232] = 12.53, P < 0.001). At T1, a significant center effect was only noted for the GH dimension (F [2, 232] = 5.42, P = 0.005). No significant Group × Center interactions were observed at T1 or T2. The therapeutic trend favoring SA was consistently maintained across all clinical sites. Results of SF-36 scores are presented in Table 8.

Table 8.

Summary of SF-36 Among Three Groups

Dimension	Time	Shallow Acupuncture (n=81)	Celecoxib (n=79)	Wait-List (n=78)	P
Physical Functioning (PF)	T0	84.20 ± 13.95 [81.11, 87.28]	81.77 ± 16.35 [78.11, 85.43]	85.63 ± 12.36 [82.85, 88.42]	0.234
	T1	90.14 ± 10.71 [87.77, 92.51]	84.68 ± 15.05 [81.31, 88.05]	85.29 ± 14.48 [82.03, 88.56]	< 0.001
	T2	89.75 ± 11.96 [87.11, 92.40]	85.27 ± 13.74 [82.19, 88.34]	88.72 ± 12.70 [85.85, 91.58]	0.010
Role Physical (RP)	T0	54.63 ± 41.48 [45.46, 63.80]	50.95 ± 39.52 [42.10, 59.80]	60.26 ± 40.58 [51.11, 69.41]	0.352
	T1	74.69 ± 36.12 [66.70, 82.68]	63.61 ± 39.99 [54.65, 72.57]	63.14 ± 41.62 [53.76, 72.53]	< 0.001
	T2	71.30 ± 37.73 [62.95, 79.64]	60.13 ± 41.51 [50.83, 69.42]	69.55 ± 39.40 [60.67, 78.44]	0.001
Bodily Pain (BP)	T0	68.77 ± 12.36 [66.03, 71.50]	67.72 ± 13.68 [64.66, 70.78]	67.40 ± 14.48 [64.14, 70.67]	0.801
	T1	80.68 ± 10.33 [78.39, 82.96]	74.30 ± 11.08 [71.82, 76.79]	67.05 ± 14.28 [63.83, 70.27]	< 0.001
	T2	80.28 ± 8.26 [78.45, 82.10]	73.48 ± 10.10 [71.22, 75.74]	76.28 ± 11.06 [73.79, 78.78]	< 0.001
General Health (GH)	T0	57.48 ± 19.50 [53.17, 61.79]	57.30 ± 20.36 [52.74, 61.86]	61.01 ± 17.91 [56.97, 65.04]	0.400
	T1	66.48 ± 22.71 [61.46, 71.50]	62.15 ± 20.12 [57.65, 66.66]	62.92 ± 18.93 [58.65, 67.19]	< 0.001
	T2	66.58 ± 21.21 [61.89, 71.27]	61.09 ± 20.77 [56.44, 65.74]	63.96 ± 18.59 [59.77, 68.15]	< 0.001
Vitality (VT)	T0	62.59 ± 17.80 [58.66, 66.53]	64.14 ± 19.20 [59.84, 68.43]	67.80 ± 15.53 [64.30, 71.30]	0.163
	T1	72.70 ± 14.45 [69.50, 75.89]	68.16 ± 16.78 [64.41, 71.92]	71.03 ± 15.40 [67.55, 74.50]	< 0.001
	T2	71.05 ± 13.53 [68.06, 74.04]	68.65 ± 15.89 [65.09, 72.21]	70.77 ± 13.70 [67.68, 73.86]	< 0.001
Social Functioning (SF)	T0	81.02 ± 15.83 [77.52, 84.52]	77.85 ± 20.11 [73.34, 82.35]	82.37 ± 19.26 [78.03, 86.72]	0.290
	T1	87.50 ± 12.50 [84.74, 90.26]	84.18 ± 15.07 [80.80, 87.55]	80.13 ± 19.58 [75.71, 84.54]	< 0.001
	T2	86.73 ± 12.55 [83.95, 89.50]	84.97 ± 13.78 [81.88, 88.05]	85.10 ± 14.95 [81.73, 88.47]	0.026
Role Emotional (RE)	T0	57.61 ± 41.84 [48.36, 66.86]	56.54 ± 42.82 [46.95, 66.13]	66.03 ± 40.42 [56.91, 75.14]	0.298
	T1	75.31 ± 35.27 [67.51, 83.11]	65.40 ± 40.80 [56.26, 74.54]	68.38 ± 40.52 [59.24, 77.51]	< 0.001
	T2	76.54 ± 34.34 [68.95, 84.14]	68.35 ± 38.08 [59.82, 76.88]	70.73 ± 40.04 [61.70, 79.75]	< 0.001
Mental Health (MH)	T0	64.84 ± 18.58 [60.73, 68.95]	65.37 ± 18.50 [61.22, 69.51]	69.96 ± 15.94 [66.37, 73.55]	0.139
	T1	72.35 ± 15.92 [68.83, 75.87]	69.37 ± 16.48 [65.68, 73.06]	70.87 ± 14.83 [67.53, 74.21]	< 0.001
	T2	71.95 ± 16.04 [68.40, 75.50]	69.77 ± 16.25 [66.13, 73.41]	70.97 ± 13.39 [67.96, 73.99]	0.003

Notes: T0, baseline; T1, post-treatment; T2, 3-month follow-up.

Abbreviation: SA, Shallow acupuncture.

CROM

Regarding cervical range of motion, baseline scores for all six dimensions were comparable among the three groups at T0 (all P > 0.05). At T1, the SA group demonstrated significant improvements in all directions compared to the celecoxib group, including Flexion (SA: 42.24, 95% CI [41.41, 43.08] vs Celecoxib: 40.02, 95% CI [38.89, 41.15]; Mean Diff: 2.42, 95% CI [1.39, 3.46], P < 0.001, Cohen’s d = 0.578), Extension (Mean Diff: 2.03, 95% CI [0.76, 3.30], P = 0.002, Cohen’s d = 0.470), Left Rotation (Mean Diff: 4.13, 95% CI [2.42, 5.83], P < 0.001, Cohen’s d = 0.592), and Right Rotation (Mean Diff: 4.60, 95% CI [2.91, 6.28], P < 0.001, Cohen’s d = 0.746). Compared to the wait-list control, the SA group achieved markedly higher mobility in all dimensions at T1 (all P < 0.001). At T2, the therapeutic advantages of SA over the wait-list control remained significant for Flexion (P = 0.029) and Right Rotation (P = 0.010). When compared to celecoxib at T2, SA maintained significantly higher scores in Right Rotation (Mean Diff: 9.63, 95% CI [3.41, 15.84], P = 0.002) and Left Flexion (Mean Diff: 2.21, 95% CI [0.91, 3.50], P = 0.001). While significant inter-center variability was observed in all dimensions at T1 and T2 (all P < 0.05), the overall trend favoring SA for functional recovery was consistently observed across all three clinical sites, particularly in rotational and flexional mobility. No significant Group × Center interactions were observed at T1 or T2. Results of CROM are presented in Table 9.

Table 9.

Summary of CROM Among Three Groups

Dimension	Time	SA (n=81)	Celecoxib (n=79)	Wait-List (n=78)		P
Flexion (°)	T0	38.64 ± 5.66 [37.39, 39.89]	39.08 ± 5.61 [37.83, 40.34]	37.94 ± 5.57 [36.69, 39.20]		0.385
	T1	42.24 ± 3.77 [41.41, 43.08]	40.02 ± 5.04 [38.89, 41.15]	37.79 ± 5.91 [36.46, 39.12]		< 0.001
	T2	41.41 ± 3.95 [40.53, 42.28]	40.54 ± 10.73 [38.13, 42.94]	38.78 ± 5.12 [37.63, 39.94]		0.015
Extension (°)	T0	39.48 ± 5.54 [38.25, 40.70]	39.54 ± 5.35 [38.34, 40.73]	41.67 ± 18.04 [37.60, 45.73]		0.407
	T1	42.91 ± 4.20 [41.98, 43.84]	40.87 ± 4.15 [39.94, 41.80]	39.06 ± 5.81 [37.75, 40.37]		< 0.001
	T2	43.46 ± 15.21 [40.09, 46.82]	40.00 ± 4.63 [38.97, 41.04]	41.93 ± 15.54 [38.42, 45.43]		0.010
Left Rotation (°)	T0	62.56 ± 9.07 [60.56, 64.57]	60.87 ± 9.32 [58.79, 62.96]	59.98 ± 9.36 [57.87, 62.09]		0.176
	T1	67.51 ± 9.12 [65.50, 69.53]	62.22 ± 9.57 [60.08, 64.36]	60.39 ± 10.33 [58.06, 62.72]		< 0.001
	T2	66.11 ± 8.46 [64.24, 67.98]	63.55 ± 15.62 [60.05, 67.05]	62.22 ± 9.93 [59.98, 64.46]		0.017
Right Rotation (°)	T0	61.22 ± 10.07 [59.00, 63.45]	61.27 ± 10.25 [58.97, 63.57]	60.85 ± 10.97 [58.38, 63.33]		0.959
	T1	68.12 ± 8.56 [66.23, 70.01]	63.51 ± 9.46 [61.39, 65.63]	62.56 ± 10.80 [60.13, 65.00]		< 0.001
	T2	72.07 ± 34.67 [64.40, 79.74]	62.48 ± 10.44 [60.14, 64.81]	63.44 ± 10.64 [61.04, 65.84]		< 0.001
Left Flexion (°)	T0	34.36 ± 6.97 [32.82, 35.90]	34.79 ± 6.12 [33.42, 36.16]	33.31 ± 7.79 [31.56, 35.07]		0.407
	T1	39.28 ± 5.51 [38.06, 40.49]	37.01 ± 6.24 [35.61, 38.41]	34.24 ± 7.62 [32.53, 35.96]		< 0.001
	T2	37.79 ± 6.06 [36.45, 39.13]	35.66 ± 6.94 [34.10, 37.21]	36.67 ± 6.92 [35.11, 38.23]		0.001
Right Flexion (°)	T0	34.73 ± 6.56 [33.28, 36.18]	35.77 ± 16.87 [31.99, 39.55]	34.19 ± 6.94 [32.63, 35.76]		0.612
	T1	40.41 ± 15.33 [37.02, 43.80]	36.22 ± 6.26 [34.81, 37.62]	34.32 ± 7.15 [32.70, 35.93]		< 0.001
	T2	37.48 ± 6.44 [36.06, 38.90]	35.62 ± 6.55 [34.15, 37.08]	36.30 ± 7.39 [34.64, 37.97]		0.010

Notes: T0, baseline; T1, post-treatment; T2, 3-month follow-up.

Abbreviation: SA, Shallow acupuncture.

Adverse Events

No serious adverse events were reported during the study, and there was no significant difference in the incidence of minor adverse events—primarily subcutaneous bruising in the shallow acupuncture group (2.47%) and gastrointestinal discomfort in the celecoxib group (5.06%)—between the two groups (P = 0.428), indicating both interventions were safe and well-tolerated.

Discussion

This multicenter randomized controlled trial establishes the clinical efficacy of SA in alleviating CNP and underscores the significance of the superficial sensory interface as a potent target for pain modulation. Our findings reveal the potential for functional reconstruction without the necessity of deep intramuscular penetration.

From a neuroanatomical perspective, evidence suggests that mechanical signals from acupuncture, when transmitted to subcutaneous tissues, induce mast cell recruitment and degranulation, thereby mediating therapeutic effects through local microenvironmental modulation.¹⁹ The dermis and superficial fascia function not merely as physical barriers but as highly sensitive somatosensory interfaces.²⁰ The analgesic mechanism of SA likely stems from the stimulation of this interface, triggering mechanotransduction.²¹ Specifically, the mechanical strain targets superficial low-threshold mechanoreceptors (LTMRs), such as Meissner corpuscles and Ruffini endings.^22,23 We hypothesize that such superficial mechanical stimulation may activate Piezo2 ion channels on the receptor membranes, converting mechanical forces into electrical signals that recruit A-β afferent fibers.^24,25 Afferent volleys mediated by A-β fibers recruit inhibitory interneurons in the spinal dorsal horn, facilitating presynaptic inhibition that effectively attenuates the ascending nociceptive transmission from C fibers.²⁶ This competitive regulatory mechanism might demonstrate superior efficacy in reversing peripheral sensitization associated with CNP and possibly serve as the fundamental mechanistic basis for subsequent systemic central analgesic modulation.^27,28 However, direct biological and tissue-level investigations are needed to confirm these mechanisms and determine the broader generalizability of this approach across other chronic pain populations.

Simultaneously, the emphasis on the superficial fascia in this study provides pivotal biomechanical evidence for pain alleviation. Recent evidence suggests that the fascia at acupoints is not merely inert connective tissue but a richly innervated, highly mechanosensitive nociceptive network.²⁹ Our observation of significant improvements in CROM across all dimensions suggests that superficial stimulation might trigger deep myofascial remodeling.³⁰ By releasing the physical constraints and mechanical tethering of superficial tissues, SA may directly attenuate movement-related nociceptive inputs, aligning with established mechanobiological mechanisms of acupuncture.³¹ Evidence supports the therapeutic benefits of myofascial release in managing chronic neck pain.³² The mechanical stimulation of the fascial layer via SA might induce fibroblast-mediated ATP release and modulate the viscoelasticity of the extracellular matrix (ECM). This ECM rearrangement facilitates a reduction in the mechanical stress state of surrounding tissues, thereby alleviating chronic pain and discomfort.³³ Furthermore, distinct from conventional approaches, our SA protocol incorporates moderate active neck-and-shoulder mobilization during needle retention, functioning as a form of kinetic acupuncture.³⁴ Studies indicate that sustained active mobilization can influence the intervertebral foramen area in neck pain patients.³⁵ The resulting dynamic shear stress acts effectively at the interface between the superficial fascia and deep tissues to release pathological adhesions. This process might constitute the fundamental mechanistic rationale for the robust superiority demonstrated by the SA group in restoring three-dimensional functional stability of the cervical spine.³⁶

On the other hand, the rehabilitative effects of SA on CNP patients are further manifested as a systemic recalibration of psychophysical states, evidenced by the synchronous improvement of anxiety levels and quality of life. However, baseline SAS scores (~43–47) were below the standard cutoff for clinical anxiety (50), showing mild, pain-related stress rather than an anxiety disorder. Even so, small improvements within this normal range are a meaningful side benefit of physical pain relief. Chronic pain is not merely an isolated physical symptom but a multidimensional construct characterized by maladaptive homeostasis, emerging from the intricate interplay between peripheral insult and affective distress.³⁷ SA likely modulates sympathetic nervous system activity and the hypothalamic–pituitary–adrenal (HPA) axis, thereby disrupting the self-perpetuating cycle of pain, stress, and muscular hypertonicity.³⁸ This restoration of autonomic balance may enhance localized tissue perfusion and oxygenation, ultimately facilitating an integrated body-mind synchronization therapeutic paradigm.³⁹

Furthermore, in this study, targeting positive reaction points rather than fixed anatomical coordinates aligns methodologically with the core tenets of precision medicine. Empirical evidence suggests that incorporating individualized protocols into acupuncture regimens optimizes therapeutic outcomes for patients with chronic pain.⁴⁰ The evolution of these reaction points is intrinsically linked to the phenomenon of acupoint sensitization—the transition of an acupoint from a quiescent to an active state. By utilizing standardized palpation protocols to identify biomechanical markers such as taut bands and nodules, we ground individualized treatment in procedural standardization.⁴¹ This pathology-driven intervention logic not only maximizes external validity within real-world clinical settings but also provides clinical evidence for integrating this efficient and safe technique.⁴²

Notably, the large effect sizes for pain relief in the SA group reflect both specific needle actions and strong patient expectations in this study. Because patients were aware of their group allocation, non-specific therapeutic elements—such as treatment preference, tender-point palpation, tactile stimulation, and guided active movements—naturally enhanced the therapeutic relationship and amplified patient-reported pain relief relative to oral pharmacotherapy.⁴³ Furthermore, unlike trials using sham controls to isolate placebo responses, our comparison with the wait-list arm captured the distinct contrast between an active physical intervention and a complete absence of care. Consequently, the superior clinical outcomes of SA reflect the comprehensive benefit of the clinical encounter rather than isolated needle stimulation alone. In addition, the wait-list design introduces a methodological limitation at the T2 follow-up. Because ethical considerations warranted providing active treatment to the wait-list group after T1, parallel three-arm comparisons were no longer feasible at T2.⁴⁴ Therefore, the observed long-term outcomes at T2 primarily reflect sustained clinical benefit relative to celecoxib, and in the absence of an untreated parallel control at T2, spontaneous symptom resolution or regression to the mean over time cannot be entirely ruled out.

Limitations

Although this multicenter study provides robust evidence for the efficacy of SA, several limitations persist. First, the inherent nature of acupuncture precluded practitioner blinding and the use of a sham acupuncture control, potentially introducing performance bias and complicating the isolation of non-specific effects. Second, the 3-month follow-up period may not fully capture the long-term trajectory of CNP, necessitating extended observation in future trials. Finally, the proposed mechanisms involving neurobiological modulation and myofascial remodeling remain hypothetical, as direct biological validation such as through biomarkers was not incorporated in the current study.

Funding Statement

This study is supported by Open Project of National Clinical Research Center for Chinese Medicine Acupuncture and Moxibustion (Grant No. NCRCOP20230010), the Traditional Chinese Medicine Inheritance Innovation Development Research Project of Zhongshan City (Grant No. 2024B3058), and Program of Guangdong Provincial Bureau of Traditional Chinese Medicine (Grant No.20243008).

Data Sharing Statement

All data are available upon reasonable request to the corresponding author, Zhenhua Xu (Email: xzh197011@163.com).

Disclosure

The authors report no conflicts of interest in this work.