Effectiveness and safety of electroacupuncture as an adjunct to guideline-based usual care for acute lumbar disc herniation: a multicentre, randomised, sham-controlled trial protocol

Abstract

Introduction

Acute lumbar disc herniation (LDH) is a major cause of low back pain and lumbosacral radiculopathy. Although guideline-based conservative management (eg, non-steroidal anti-inflammatory drugs and physical therapy) is recommended as first-line care for most patients, pain relief and functional recovery are often incomplete and medication-related adverse effects can limit adherence. Electroacupuncture (EA) is widely used as an adjunctive therapy for LDH and may provide analgesic and functional benefits through neuromodulation. However, high-quality sham-controlled evidence in acute/subacute LDH remains insufficient. This trial aims to evaluate the effectiveness and safety of EA added to usual care in acute/subacute LDH.

Methods and analysis

This multicentre, parallel-group, randomised, sham-controlled trial will recruit 180 participants aged 18–65 years with acute/subacute LDH confirmed by lumbar CT or MRI (symptom duration 2–6 weeks) and baseline pain Visual Analogue Scale (VAS) score ≥4/10 from four tertiary hospitals in China. Participants will be randomised 1:1 using stratified variable block randomisation (strata: centre, baseline VAS ≤6 vs >6, age <45 vs ≥45) with concealed allocation. Both groups will receive usual care (celecoxib 100 mg twice daily plus physical therapy three times weekly). Rescue tramadol hydrochloride 50 mg will be permitted when VAS is ≥7, subject to a minimum 8-hour dosing interval, a maximum daily dose of 300 mg and a maximum continuous duration of 48 hours. The EA group will additionally receive EA at bilateral BL23, BL25 and EX-B2 with 2/100 Hz alternating stimulation (0.5–2 mA), 30 min per session, three sessions per week for 6 weeks (18 sessions). The sham group will receive non-penetrating Park sham needles with identical procedures but no insertion and no electrical output. Outcome assessors and statisticians will be blinded. The primary outcome is change in average low back/leg pain VAS from baseline to week 6. Secondary outcomes include Oswestry Disability Index, SF-36, MRI-based herniation volume and T2 signal intensity at week 6, Patient Global Impression of Change, rescue analgesic use and adverse events. Follow-up will continue to 6 months. Primary analysis will follow the intention-to-treat principle using linear mixed-effects models with multiple imputation for missing data.

Ethics and dissemination

Ethical approval has been obtained from the ethics committees of all four participating centres: the Research Ethics Committee of Suzhou Hospital of Traditional Chinese Medicine (2025-040), the Ethics Review Committee of Changshu Hospital of Traditional Chinese Medicine (CZYLS-2026005), the Ethics Committee of Sichuan Provincial Orthopaedic Hospital (KY2026-006-01) and the Medical Ethics Committee of Suqian Hospital of Traditional Chinese Medicine (2025120009).

Trial registration number

ITMCTR2025002522.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The multicentre, parallel-group, randomised, sham-controlled design across four tertiary hospitals includes blinded outcome assessment and statistical analysis.

• A validated non-penetrating Park sham procedure with an identical treatment setting is used to control for contextual and procedural effects.

• The electroacupuncture intervention is standardised across sites through prespecified stimulation parameters, practitioner training, fidelity checklists and formal assessment of treatment credibility and participant blinding.

• The trial combines a patient-centred pain outcome with disability, quality-of-life, rescue-analgesic, safety and exploratory MRI outcomes over 6 months of follow-up.

• Treating acupuncturists cannot be blinded, the sham procedure may not be physiologically inert and some between-centre variation in usual-care physical therapy may remain despite standardisation and statistical adjustment.

Background

Low back pain remains a leading cause of years lived with disability worldwide, and its absolute burden is projected to increase in coming decades.¹ Lumbar disc herniation (LDH) is a common structural cause of low back pain and sciatica (lumbosacral radicular pain), typically affecting working-age adults.^{2 3} Because degenerative imaging findings are also frequent in asymptomatic individuals, careful clinicoradiological correlation is essential when defining trial eligibility.^{4 5} Current guidance generally recommends an initial period of conservative management—such as education, activity modification, analgesics and exercise-based rehabilitation—while reserving surgery for selected patients with severe or persistent symptoms or neurological deficits.^{6 7} However, pharmacological options may be limited by adverse effects, contraindications or patient preference, and many patients seek non-pharmacological approaches to relieve pain and improve function. Consequently, there is a pressing clinical need for adjunctive interventions that can enhance the magnitude and durability of symptom relief beyond standard care in the acute phase of this condition.

Acupuncture is widely used for spinal pain and has shown clinically relevant benefits across chronic pain conditions.^8–10 For sciatica related to LDH, a recent multicentre, sham-controlled randomised clinical trial reported greater improvements in pain and function with acupuncture than with sham procedures, with benefits persisting through long-term follow-up.¹¹ Another sham-controlled trial also supported the efficacy and safety of acupuncture for chronic discogenic sciatica.¹² Systematic reviews suggest that acupuncture may be effective for sciatica, but the certainty of evidence is often limited by heterogeneity and methodological limitations.¹³ Electroacupuncture (EA) applies electrical stimulation through acupuncture needles and is commonly used in clinical practice. Prior controlled trials for discogenic sciatica indicate that EA may provide superior outcomes compared with electrotherapy modalities.¹⁴ Mechanistically, EA is posited to exert analgesic and anti-inflammatory effects through the modulation of descending pain inhibitory pathways, the release of endogenous opioids and the attenuation of neuroimmune signalling cascades implicated in radicular pain pathogenesis.^{15 16}

Preclinical evidence provides compelling biological plausibility for testing EA specifically within the acute/subacute LDH window. In rat models where autologous nucleus pulposus is applied to lumbar nerve roots or adjacent to the dorsal root ganglion (DRG), animals develop radiculopathy-like pain behaviours and neuroinflammatory changes, including nitric-oxide-related axonal dysfunction and early DRG apoptosis.^{17 18} In vitro, disc-related cytokines such as TNF and IL-1β can inhibit axonal outgrowth from cultured DRG neurons, supporting an inflammation-driven component of nerve root dysfunction beyond pure mechanical compression.¹⁹ In LDH rodent models, EA has been reported to reverse mechanical allodynia and attenuate neuroimmune signalling, including reducing HCN2 expression in DRG and microglial activation in the spinal dorsal horn and promoting a shift towards anti-inflammatory microglial phenotypes.^{20 21} Together, these findings support a mechanistically grounded expectation that EA may mitigate acute radicular pain and related disability by modulating peripheral and central sensitisation pathways.

When performed by appropriately trained practitioners, acupuncture appears to be relatively safe, with serious adverse events (AEs) reported rarely in large prospective datasets.²² Nevertheless, rigorous evidence for EA in acute/subacute LDH remains insufficient and high-quality sham-controlled trials. To address this gap, we designed a multicentre, randomised, sham-controlled trial to evaluate whether the addition of a standardised EA protocol to guideline-based conservative management yields clinically meaningful improvements in pain and function among patients with acute/subacute LDH. By employing a validated non-penetrating sham device with identical procedural rigour and patient–provider interaction, this trial aims to isolate the specific efficacy of EA while controlling for the contextual and non-specific effects inherent to procedural interventions.

Objectives

Primary objective

To determine whether EA plus usual care reduces pain intensity (Visual Analogue Scale, VAS) more than sham EA plus usual care from baseline to week 6.

Secondary objectives

To compare changes in disability (Oswestry Disability Index, ODI), quality of life (SF-36), imaging outcomes, global improvement (Patient Global Impression of Change, PGIC), rescue analgesic use and safety outcomes over the intervention and follow-up period.

Exploratory objective

To explore prespecified subgroup effects by radicular pain characteristics and age strata.

Hypothesis

EA plus usual care will yield a clinically meaningful greater reduction in pain than sham EA plus usual care at week 6 (between-group difference ≥1.5 VAS points).

Methods

Protocol reporting

This protocol was developed in accordance with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) 2025 statement.²³ The trial will be reported following the Consolidated Standards of Reporting Trials (CONSORT) 2010 statement and the CONSORT extension for nonpharmacologic treatments, as applicable.^{24 25} Intervention details are described using the STandards for Reporting Interventions in Clinical Trials of Acupuncture (STRICTA) 2010 checklist and the template for intervention description and replication (TIDieR) guide to facilitate transparency and reproducibility.^{26 27}

Study design and setting

This is a multicentre, parallel-group, randomised, sham-controlled trial with assessor-blinding and analyst-blinding. The intervention period is 6 weeks, followed by follow-up assessments at 3 and 6 months. The trial will be conducted at four tertiary hospitals in China: Suzhou Hospital of Traditional Chinese Medicine, Changshu Hospital Affiliated to Nanjing University of Chinese Medicine, Sichuan Provincial Orthopaedic Hospital and Suqian Hospital of Traditional Chinese Medicine. The trial will be coordinated by Suzhou Hospital of Traditional Chinese Medicine. The overall study flow is shown in figure 1.

Figure 1. Flow diagram of the study. AE, adverse event; EA, electroacupuncture; LDH, lumbar disc herniation; ODI, Oswestry Disability Index; PGIC, Patient Global Impression of Change; SAE, serious adverse event; VAS, Visual Analogue Scale.SF-36,36-item short-form health survey.

Participants

Inclusion criteria

Eligible participants must meet all of the following criteria: (1) age 18–65 years, any sex; (2) LDH confirmed by lumbar CT or MRI, classified as protrusion, extrusion or sequestration, and anatomically concordant with the participant’s symptoms and neurological findings; (3) symptom duration 2–6 weeks; (4) baseline pain VAS ≥4/10 and (5) provision of written informed consent.

Exclusion criteria

Exclusion criteria are as follows: (1) symptom duration outside 2–6 weeks or prior lumbar spine surgery; (2) severe complications or alternative diagnoses requiring urgent management (eg, cauda equina syndrome, tumour, infection); (3) pregnancy, breastfeeding or known bleeding disorders; (4) implanted cardiac pacemaker or other contraindications to acupuncture/electrical stimulation; (5) acupuncture within the previous 4 weeks or epidural steroid injection within the previous 6 months or (6) cognitive impairment precluding reliable outcome assessment.

Recruitment and consent

Participants will be recruited from outpatient clinics and through advertisements at the four participating centres. Screening will include confirmation of the LDH diagnosis and eligibility using lumbar CT or MRI. Either modality may be used to establish imaging eligibility, consistent with the registered eligibility criterion. After written informed consent, all enrolled participants will undergo the protocol-specified baseline lumbar MRI for imaging outcome assessment; this MRI is not an additional eligibility criterion. CT myelography will not be routinely performed as a study-mandated procedure and may be undertaken only when clinically indicated independently of the trial. The consent discussion will explicitly address the sham procedure, potential risks and participants’ right to withdraw at any time.

Randomisation and allocation concealment

Participants will be randomised 1:1 to EA plus usual care or sham EA plus usual care. The randomisation sequence will be generated by an independent statistician using stratified block randomisation with variable block sizes (4–6). Strata are trial centre, baseline VAS (≤6 vs >6) and age (<45 vs ≥45). Allocation concealment will be ensured using sequentially numbered, sealed, opaque envelopes opened only after enrolment and baseline assessment. Allocation concealment procedures are designed to minimise selection bias.²⁸ A trained research coordinator will enrol participants after confirming eligibility and obtaining consent. After enrolment, the next sequentially numbered envelope will be opened by a staff member not involved in outcome assessment, and the assigned intervention will be communicated only to the treating acupuncturist/therapist. The allocation sequence and block sizes will not be accessible to personnel responsible for screening, enrolment, or outcome assessment; envelope use will be logged (date/time, participant ID and signature) and audited periodically to detect any irregularities.

Blinding

Because the treating acupuncturists deliver different procedures, they cannot be blinded to group assignment. Participants will be masked to the extent feasible by using a non-penetrating sham device and an identical treatment setting, session frequency and duration, device appearance (including lights/sounds) and standardised practitioner–participant communication. Outcome assessors, data managers, and statisticians will remain blinded throughout the trial. Unblinding will be permissible only when knowledge of the allocated intervention is essential for urgent clinical management; requests for unblinding will be documented, and the decision will be made by the principal investigator in consultation with the Data and Safety Monitoring Board (DSMB), with unblinding restricted to the minimum personnel necessary.

Interventions

Usual care (both groups)

All participants will receive guideline-based usual care comprising celecoxib 100 mg orally twice daily and physical therapy three times per week for 6 weeks. Concomitant treatments will be recorded, and additional acupuncture outside the protocol will be discouraged during the intervention period.

Rescue analgesia: Participants with severe breakthrough pain (VAS ≥7) may receive tramadol hydrochloride tablets 50 mg (CSPC Ouyi Pharmaceutical, Shijiazhuang, China). The single dose will be 50 mg, with a minimum dosing interval of 8 hours, a maximum daily dose of 300 mg and a maximum continuous treatment duration of 48 hours. Rescue medication use will not be considered a protocol deviation and will not lead to discontinuation of scheduled outcome assessments. The following will be recorded in the medication diary and electronic case report form (eCRF): whether rescue medication was used, time to first use, VAS immediately before each dose, reason for use, dose per administration, cumulative dose, number of treatment days and related AEs.

EA group

Acupoints: bilateral BL23 (Shenshu), BL25 (Dachangshu) and EX-B2 (Yaoyan). Needles: disposable sterile Hwato acupuncture needles (0.25×40 mm; Suzhou Medical Supplies Factory, Suzhou, China). Needling technique: needles will be inserted at the specified points and manually manipulated by rotation to elicit deqi sensation. Electrical stimulation: an Hwato SDZ-V Electronic Acupuncture Treatment Instrument (Suzhou Medical Supplies Factory, Suzhou, China) will deliver alternating 2/100 Hz stimulation at 0.5–2 mA (adjusted to participant tolerance). Dose: 30-minute sessions, three times per week for 6 consecutive weeks (18 sessions). Practitioners: licensed TCM physicians/acupuncturists with ≥5 years of experience who have completed standardised protocol training. Acupoint locations follow the WHO Standard Acupuncture Point Locations.²⁹ A schematic diagram of the EA intervention is shown in figure 2.

Figure 2. Schematic diagram of the electroacupuncture (EA) intervention in the EA group.

Sham EA group

Sham device: a Park Sham Acupuncture Device with retractable non-penetrating stainless-steel blunt-tipped needles (0.25×40 mm; Suzhou Medical Supplies Factory, Suzhou, China) will be applied at the same point locations. Procedures will mimic those of the EA group, but without skin penetration and with no electrical current output. Standardised verbal interaction will be used to maintain procedural credibility. The number and duration of sessions will match the EA group. The Park sham device has demonstrated validity as a non-penetrating sham control,³⁰ and placebo-needle approaches have been used in acupuncture trials.³¹ The overall treatment protocol for both groups is summarised in figure 3.

Figure 3. Schematic flow chart of the treatment protocol. EA, electroacupuncture; ITMCTR, International Traditional Medicine Clinical Trial Registry; ODI, Oswestry Disability Index; PGIC, Patient Global Impression of Change; VAS, Visual Analogue Scale.SF-36,36-item short-form health survey.

Intervention fidelity and adherence

Intervention delivery will be standardised through video training, site initiation visits, and ongoing supervision. Each session will be documented using checklists covering point selection, stimulation parameters and session duration. Reasons for missed sessions will be recorded; participants will be encouraged to complete all scheduled sessions. Participant expectations and treatment credibility will be assessed using the Credibility/Expectancy Questionnaire after the first treatment session.³² Success of participant blinding will be evaluated using the Bang blinding index at week 6.³³

Outcomes and assessment schedule

Outcomes will be assessed at baseline, week 3, week 6 (end of intervention), month 3 and month 6 by blinded assessors, as summarised in table 1. To promote retention, participants will receive appointment reminders (telephone/text), flexible scheduling within predefined visit windows, and, where appropriate, remote completion of questionnaire-based outcomes. For participants who discontinue or deviate from the assigned intervention, the primary outcome (VAS) and key secondary outcomes (ODI, SF-36, PGIC) as well as safety data will still be collected whenever possible.

Table 1. Schedule of enrolment, interventions and assessments.

	Trial period
	Enrolment		Postrandomisation			Close-out
Time point	Screening	Baseline (week 0)	Week 3	Week 6	Month 3	Month 6
Enrolment
Eligibility criteria	X
Informed consent	X
Randomisation		X
Intervention
EA/sham EA sessions		X		X
Assessments
VAS		X	X	X	X	X
ODI		X	X	X	X	X
SF-36(36-Item Short Form Health Survey)		X		X	X	X
MRI (volume, T2 signal)		X		X
PGIC				X	X	X
Rescue medication use		X	X	X	X	X
Adverse events (AE/SAE)		X	X	X	X	X

AE, adverse event; EA, electroacupuncture; ODI, Oswestry Disability Index; PGIC, Patient Global Impression of Change; SAE, serious AE; VAS, Visual Analogue Scale.

Primary outcome

The primary outcome is the change in average low back/leg pain intensity over the preceding week measured using a 0–10 VAS from baseline to week 6. The VAS is a widely used and validated pain intensity measure.³⁴

Secondary outcomes

Secondary outcomes are as follows: (1) ODI (0–100, higher scores indicate worse disability; minimal clinically important difference (MCID) 12 points); (2) SF-36 physical and mental component summary scores; (3) MRI-derived herniation volume (mm³) and T2 signal intensity changes at week 6; (4) PGIC; (5) rescue analgesic use, including the proportion of participants requiring rescue medication, time to first use, number of treatment days and cumulative dose and (6) safety outcomes including AEs and serious AEs (SAEs), graded using Common Terminology Criteria for Adverse Event (CTCAE) v5.0. ODI and SF-36 are well-established patient-reported measures for spine-related disability and health-related quality of life, respectively.^{35 36} PGIC is commonly used to capture patients’ perceived overall improvement.³⁷ The outcome domains were selected with reference to IMMPACT guidance for pain clinical trials.³⁸

For imaging outcomes, all enrolled participants will undergo lumbar MRI at baseline and week 6 using a MAGNETOM Skyra V.3.0 T MRI system (Siemens Healthcare, Erlangen, Germany). MRI acquisition parameters will be standardised across centres. Image analyses will be performed by trained readers blinded to treatment allocation and assessment time point. Herniation segmentation and three-dimensional volumetric measurement will be performed using 3D Slicer version 5.12.3 (Brigham and Women’s Hospital and 3D Slicer contributors, Boston, Massachusetts, USA), and T2 signal intensity will be measured using ImageJ version 1.54u8 (US National Institutes of Health, Bethesda, Maryland, USA) with prespecified regions of interest. A prespecified measurement protocol (segmentation rules and definitions) will be used; a random sample of at least 10% of scans will be re-read to assess intrarater and inter-rater reliability, and discrepancies will be resolved by consensus. The equipment, devices, medication and software used in the trial are summarised in table 2.

Table 2. Equipment, devices, medication and software used in the trial.

Category	Commercial name/model/version	Manufacturer/developer	City and country	Use in the trial
Acupuncture needles	Hwato disposable sterile acupuncture needles, 0.25×40 mm	Suzhou Medical Supplies Factory	Suzhou, China	Needling at bilateral BL23, BL25 and EX-B2
Park sham device and needles	Park Sham Acupuncture Device with retractable non-penetrating stainless-steel blunt-tipped needles, 0.25×40 mm	Suzhou Medical Supplies Factory	Suzhou, China	Sham needling without skin penetration
Electroacupuncture device	Hwato SDZ-V Electronic Acupuncture Treatment Instrument	Suzhou Medical Supplies Factory	Suzhou, China	Alternating 2/100 Hz stimulation at 0.5–2 mA
MRI system	MAGNETOM Skyra 3.0 T MRI system	Siemens Healthcare	Erlangen, Germany	Lumbar MRI at baseline and week 6
Electronic data capture	Research Electronic Data Capture (REDCap), institutionally hosted deployment	Vanderbilt University	Nashville, Tennessee, USA	eCRFs, validation, audit trails and queries
Sample-size software	PASS, V.15.0.13	NCSS	Kaysville, Utah, USA	Primary sample-size calculation
Sample-size verification	G*Power, V.3.1.9.7	Heinrich Heine University Düsseldorf	Düsseldorf, Germany	Independent verification of sample size
3D image segmentation	3D Slicer, V.5.12.3	Brigham and Women’s Hospital and 3D Slicer contributors	Boston, Massachusetts, USA	Segmentation and volumetric measurement of LDH
T2 signal analysis	ImageJ, V.1.54u8	US National Institutes of Health	Bethesda, Maryland, USA	T2 signal measurement using prespecified regions of interest
Rescue medication	Tramadol hydrochloride tablets, 50 mg	CSPC Ouyi Pharmaceutical	Shijiazhuang, China	Rescue analgesia when VAS ≥7

eCRFs, electronic case report forms; LDH, lumbar disc herniation; VAS, Visual Analogue Scale.

Sample size

The sample size calculation was based on the primary outcome—change in average low back/leg pain intensity measured by the 0–10 VAS from baseline to week 6. Based on previously published sham-controlled acupuncture trials in sciatica¹¹ and systematic reviews of conservative management for acute LDH, the MCID for VAS in this population is generally accepted as 1.0–1.5 points. To ensure adequate power to detect a modest but clinically relevant effect, we adopted a conservative between-group difference (delta) of 1.2 points. The expected SD of the change score was set at 2.5 points, reflecting the greater variability typically observed in acute/subacute pain populations compared with chronic cohorts. Under these assumptions (Cohen’s d=1.2/2.5=0.48), a two-independent-sample t-test with a two-sided alpha of 0.05 and 80% power requires 69 participants per group (total N=138). Calculation was performed using PASS V.15.0.13 (NCSS) and verified with G*Power V.3.1.9.7 (Heinrich Heine University Düsseldorf, Düsseldorf, Germany).

To accommodate an anticipated 20% attrition rate by the primary endpoint at week 6, the required sample size is inflated to 69/(1−0.20)=87 participants per group (total N=174). Given the multicentre design involving four tertiary hospitals, a modest design effect (variance inflation factor) of approximately 1.03 was applied to account for minimal between-centre heterogeneity, yielding a final requirement of 90 participants per group (total N=180). This sample size provides >80% power to detect a between-group difference of 1.2 VAS points under the specified assumptions, and it also preserves sufficient precision for prespecified subgroup analyses and secondary outcomes without compromising the integrity of the primary comparison.³⁹

Data collection and management

Data will be captured in eCRFs using Research Electronic Data Capture (REDCap; institutionally hosted deployment; Vanderbilt University, Nashville, Tennessee, USA). Data entry will include validation rules, range checks and audit trails. Trained personnel will perform quality control and resolve queries. Participants will be assigned unique identifiers, and personal identifiers will be stored separately from analysis datasets. Access to data will be restricted to authorised study staff.

Personal identifiers will be stored separately from trial data using a unique study ID. Access to identifiable information will be restricted to authorised staff only, and all electronic data transfers will use encrypted channels. Source documents and consent forms will be stored in locked cabinets at each site; data will be retained for at least 10 years after study completion in accordance with institutional policy and applicable regulations.

Data monitoring and interim review

An independent DSMB will oversee safety and trial conduct. The DSMB will consist of at least three independent members with expertise in spine/orthopaedics, acupuncture or rehabilitation and biostatistics. All DSMB members will declare potential conflicts of interest prior to appointment. A written DSMB charter will define responsibilities, meeting frequency, safety review procedures and reporting lines; the charter will be available from the corresponding author on reasonable request.

A planned interim review will occur after approximately 50% of participants have been enrolled to evaluate safety and futility. Interim monitoring will follow an O’Brien-Fleming alpha-spending approach to preserve the overall type I error rate. On-site monitoring will be performed quarterly to verify consent documentation, source data and adherence to the protocol. Interim review boundaries will follow conservative spending functions (eg, O’Brien-Fleming type) to preserve the overall type I error rate.⁴⁰ Only the DSMB will have access to any interim, unblinded comparative results. The DSMB may recommend trial continuation, modification, temporary suspension or termination based on prespecified stopping guidelines and safety considerations; the final decision will be made by the principal investigator in consultation with the ethics committees, with the sponsor having no authority over this decision.

Statistical analysis

Analyses will be conducted using SAS V.9.4 and/or R (V.4.0 or later). Baseline characteristics will be summarised using appropriate descriptive statistics. The primary analysis will use linear mixed-effects modelling with fixed effects for group, time and group-by-time interaction, and random effects for centre and participant. Prespecified covariates are age, sex and baseline VAS. Results will be presented as between-group differences with 95% CIs. The primary analysis population will be intention-to-treat (ITT), including all randomised participants. Missing outcome data will be handled using multiple imputation. Per-protocol analyses will be performed as sensitivity analyses. Missing data will be handled using multiple imputation under plausible missing-at-random assumptions, with chained equations implemented where appropriate.⁴¹

Secondary outcomes will be analysed using mixed-effects models or generalised mixed models as appropriate. Multiple comparisons for secondary outcomes will be adjusted using Bonferroni correction. Prespecified subgroup analyses will explore heterogeneity by radicular pain characteristics and age strata. Effect sizes (eg, Cohen’s d) will be reported alongside p values to aid interpretation.

Rescue analgesic use will be analysed as a prespecified secondary outcome. The groups will be compared with respect to the proportion of participants requiring rescue medication, time to first use, number of treatment days and cumulative tramadol dose. All randomised participants will remain in the ITT analysis regardless of rescue medication use, and outcome assessments will continue as scheduled. Sensitivity analyses will evaluate the robustness of the primary outcome to rescue medication use.

Harms

AEs and SAEs will be collected at each assessment and throughout the intervention period. Expected EA-related events include transient needling pain, bruising, minor bleeding, dizziness and local infection; serious complications are expected to be rare. All events will be graded (CTCAE v5.0) and assessed for relatedness to the intervention. SAEs will be reported promptly to the principal investigator and ethics committees according to local requirements. AEs will be classified and graded using CTCAE v5.0, and the overall safety monitoring plan is informed by large-scale observational evidence regarding acupuncture-related AEs.²² Participants will receive appropriate medical care for any AEs during the trial. If a participant suffers harm related to trial participation, compensation and/or reimbursement will be provided in accordance with institutional policies and applicable regulations; details of coverage (eg, insurance, reimbursement procedures) are available from the coordinating centre on request.

Patient and public involvement

Patients and the public were not involved in the design of this protocol. Participants will be informed of the study results on completion through public summaries and academic dissemination.

Ethics and dissemination

The trial will be conducted in accordance with the Declaration of Helsinki and Good Clinical Practice. Ethical approval has been obtained from the ethics committees of all four participating centres: the Research Ethics Committee of Suzhou Hospital of Traditional Chinese Medicine (approval no. 2025-040; 31 October 2025), the Ethics Review Committee of Changshu Hospital of Traditional Chinese Medicine (approval no. CZYLS-2026005; 6 March 2026), the Ethics Committee of Sichuan Provincial Orthopaedic Hospital (approval no. KY2026-006-01; 2 February 2026) and the Medical Ethics Committee of Suqian Hospital of Traditional Chinese Medicine (approval no. 2025120009; 29 December 2025). No participating centre will enrol participants without valid site-specific ethics approval. All participants will provide written informed consent. Data will be anonymised and protected in accordance with institutional policies and applicable regulations.

Results will be disseminated through peer-reviewed publication and conference presentations. Authorship will follow ICMJE criteria. After publication, de-identified data and statistical code may be shared on reasonable request, subject to ethics approval and data-sharing agreements.

Discussion

This protocol has several methodological strengths. The multicentre, randomised, sham-controlled design is intended to strengthen internal validity while improving generalisability across four tertiary hospitals. A validated non-penetrating Park sham device, standardised practitioner–participant communication, blinded outcome assessment and statistical analysis, intervention-fidelity procedures and formal evaluation of expectancy and participant blinding are incorporated to reduce bias. The prespecified EA protocol and mixed-effects analytical approach also support consistency across sites and account for centre-level heterogeneity.

Several limitations remain. Treating acupuncturists cannot be blinded because they deliver distinct procedures, creating a potential source of performance bias despite standardised scripts and training. The non-penetrating sham intervention may not be physiologically inert because tactile stimulation could produce nonspecific sensory effects, potentially attenuating between-group differences. In addition, the multicentre usual-care physical therapy component may vary to some extent across sites; fidelity monitoring and centre-level random effects are intended to mitigate but cannot fully eliminate this heterogeneity. MRI outcomes at week 6 are exploratory and should therefore be interpreted as mechanistic or phenotypic measures rather than definitive evidence of structural treatment effects.

Trial status

Protocol version 2.0, revised 29 August 2026. All four participating centres have obtained the required site-specific ethics approvals. Recruitment is planned to commence after completion of site training and study initiation procedures. The planned study duration is 24 months, including recruitment, the 6-week intervention, 6-month follow-up and final analyses.

Any important protocol modifications (eg, changes to eligibility criteria, outcomes, sample size or analysis methods) will be submitted to the ethics committees for approval prior to implementation and will be updated in the trial registry. Amendments will also be communicated to participating sites and, where relevant, to trial participants.

The funder/sponsor will provide financial support only and will have no role in study design, data collection, data analysis, interpretation, or publication decisions.