Acupuncture for chronic neck pain

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To assess the effects of acupuncture on chronic neck pain, with a focus on pain, disability, quality of life, patient satisfaction, and global perceived effect.

Background

Description of the condition

Neck pain is commonly understood as pain in the neck region with proximal areas often involved, such as the upper back, shoulders, jaw, occiput and head (as in headache, or with pain radiating into the head, trunk, or arms, called radiculopathy). Neck pain has been anatomically defined as 'pain between the superior nuchal line and the spine of the scapula, across the upper back posteriorly, and laterally between the exterior occipital protuberance/superior nuchal line and the superior border of the clavicle and suprasternal notch' (Guzman 2009). Neck pain is considered chronic if it lasts three or more months, often with decreased range of motion.

A ‘pain in the neck’ is an English language idiom describing something or someone as ‘annoying, aggravating, or inconvenient’, and may reflect the common experience of neck pain. Neck pain is among the most common pain conditions in the USA, together with migraines, headaches, and low back pain (National Health Interview Survey 2016). The prevalence of neck pain in adults, over a 12‐month period, is 30% to 50%, and frequently recurs in 50% to 75% of people, one to five years after a prior incident (Carroll 2009). Neck pain can be episodic, recurring, or persistent. Episodic neck pain can occur over a lifetime with variable degrees of recovery between episodes (Guzman 2009). Of the estimated 33.8 million people with neck pain in the USA, almost half (49.3%) experience persistent pain (Kennedy 2014). Daily feelings of anxiety, depression, and fatigue are correlated with persistent pain. Even with persistence, pain varies in severity. About half (50.5%) of those with persistent neck pain report it can be 'unbearable and excruciating', and 67.2% report their pain is 'constantly present’ to some degree. Only 28.2% say ‘medication can take away my pain completely’ (Kennedy 2014).

Neck pain and low back pain were the leading global cause of disability from 2005 to 2015 in most countries, and the leading cause of disability in all high‐income countries in 2015 (GBD 2016). The US Social Security Administration reported that work disability attributed to musculoskeletal and connective tissue disorders, including neck pain, increased from 20.6% of beneficiaries in 1996 to 25.4% in 2005 (Deyo 2009), to 31.7% in 2015 (Social Security Administration 2016).

There is a substantial economic impact of pain conditions, including neck pain, for individuals (White 2005), insurers, and employers (Lerner 2015), which includes healthcare costs, and direct and indirect workforce financial costs, including both workplace absenteeism (Phillips 2017), and lost productivity from ill and injured workers while still working, called presenteeism (Goetzel 2004; Koopman 2002).

Description of the intervention

Acupuncture is the insertion and manipulation of fine, solid core needles at specified points, or combination of points, on the body. Acupuncture therapy is derived from the traditional East Asian paradigm the recognizes the inter‐relationship of organs, body points, and channels, and associated symptoms, diseases, and dysfunction. In the USA, depending on a state’s scope of practice, acupuncture therapy may include treatment by means of mechanical, thermal, or electrical stimulation. Acupuncture needling is the insertion and manipulation of fine, solid‐core needles at specified points, or combination of points, on the body (Tick 2018).

How the intervention might work

Needle stimulation of acupuncture points can trigger complex, interactive mechanical and chemical processes. Studies show these acupuncture biomechanisms interact at the local, proximal, regional, distal, and systemic level. An acupuncture point itself may be quiescent or activated, becoming sensitized and responsive to needling (Kim 2017; Quiroz‐Gonzalez 2017; Rong 2013; Yu 2015).

Selection of acupoints to be treated for neck pain may be informed by a traditional East Asian perspective or contemporary research, or combined with other approaches. The process may incorporate palpation to identify the exact location of known channels and points intended for treatment, and to assess sensitivity. Needle insertion can then be simple, and often involves rotation manipulation. Needle rotation leads to winding and deformation of collagen fibers, and mechanotransduction of signals within connective tissue, recognized as a body‐wide network, intimately associated with all other tissues, including the organ systems (Langevin 2001; Langevin 2002; Langevin 2006; Langevin 2006a). The channels are thought to reside in connective tissue planes, providing lower‐resistant passages for the chemotactic migration of cells, including mast cells, which can migrate longitudinally along meridians (Ding 2018; Langevin 2002a).

Interacting with collagen fibers, mast cells appear to be a central structure activated in acupuncture points by needling (Huang 2018). There is a higher density of mast cells at acupuncture points than at non‐acupoints that are in turn preferentially co‐localized with sensory neurons (Yin 2018; Zhong 1994). Acupuncture needling causes mast cell aggregation and degranulation, activated by a surface mechanosensitive protein, called transient‐receptor‐potential cation channel (TRPV2) (Huang 2018; Zhang 2012). The degranulation of mast cells increases adenosine and histamine concentrations, and activates adenosine and histamine receptors, which in turn act on nerve endings, relax vascular smooth muscle, increase blood vessel permeability, and transfer the acupuncture signal to the central nervous system (Huang 2018). The release of 5‐HT (serotonin), histamine, and tryptase during degranulation of mast cells may underlie the cross‐communication between the circulatory, nervous, and immune systems at acupoints (Ding 2018). These complex chemical and mechanical interactions create an analgesic effect (Ding 2013; Goldman 2010; Huang 2018; Wang 2014).

Needle manipulation at acupuncture points induces higher microcirculatory changes locally and proximally (Wang 2017), which along with mast cell degranulation, correlate with increased analgesic effects locally (Goldman 2010), and at relevant distal (Min 2015), and regional sites (Calamita 2018).

These mechanisms join the long‐held view that acupuncture promotes a peripheral and central analgesic response by releasing endorphins and endogenous opioid peptides (Han 2004; Huang 2018; Trinh 2016a). Recent research has highlighted the central role of endocannabinoids in pain modulation, and the endocannabinoid system’s participation in the analgesic effects of electro‐acupuncture (EA (Hu 2017; Zogopoulos 2013)).

Brain imaging studies (fMRI) show acupuncture alters brain activation patterns in areas associated with pain processing (Trinh 2016a,Huang 2012; Scheffold 2015; Cai 2018). Analgesic effects outlast the needling period; pain‐associated brain areas are modulated in direct response to a preceding acupuncture treatment (Theysohn 2014).

Although there is a lot of literature on the hypothesized mechanisms of acupuncture in treating pain, many of these theories are not supported by high‐quality human research; they are based on animal studies. The continued use of acupuncture is based on a long history of clinical experiences; the potential mechanisms remain uncertain.

Why it is important to do this review

The Global Spine Care Initiative has recommended non‐invasive procedures for chronic neck pain (Chou 2018), as they can produce outcomes similar to more invasive surgical procedures, with less risk of adverse effects (Acaroğlu 2018). These non‐invasive procedures include exercise, yoga, cognitive behavioral therapies, acupuncture, biofeedback, progressive relaxation, massage, manual therapy, and interdisciplinary rehabilitation (Chou 2018).

The 2006 Cochrane Review, Acupuncture for neck pain, focused on acupuncture treatment of both acute and chronic neck pain and was updated and then withdrawn by the authors in 2016 (Trinh 2006; Trinh 2016a; Trinh 2016b). Recent reviews have supported the use of acupuncture for chronic neck pain (Chou 2018; Seo 2017; Vickers 2018). Given the heavy public health and economic burden associated with chronic neck pain, and encouragement to base healthcare decisions on best evidence, an updated Cochrane systematic review focused on chronic neck pain is justified.

The purpose of this review is to summarize the most current scientific evidence on the effectiveness of acupuncture for chronic neck pain, by incorporating the information from relevant studies, and including long‐term outcomes, which were absent from previous studies.

Objectives

To assess the effects of acupuncture on chronic neck pain, with a focus on pain, disability, quality of life, patient satisfaction, and global perceived effect.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomized controlled trials (RCTs). We will exclude quasi‐randomized trials. We will not restrict study eligibility by language or publication status.

If the study report does not describe the method of randomization, we will contact the study authors for information on the randomization method to confirm that the trial was truly randomized.

Types of participants

We will include trials of adults (aged 18 years or greater) with chronic neck pain, defined as neck pain of at least three months duration, in the following categories (Misailidou 2010):

• Mechanical neck pain, with or without radicular findings, including whiplash‐associated disorders (Spitzer 1995), and pain due to degenerative changes (Borenstein 2008); and

• Headache attributed to neck disorders (Headache Classification Subcommittee 2004)

Neck pain with cervical radiculopathy includes sensory, motor, or reflex changes of the upper back, shoulders, or arms. Whiplash injury (a sudden acceleration and deceleration of the neck) can present as chronic neck pain. Headache attributed to neck disorders is defined as head pain that originates from a neck problem.

We will exclude studies with participants who have neurological disease (e.g. myelopathies), neck fractures, neoplasms, or other pathological conditions that may be causing the neck pain (Borenstein 2008).

If studies include participants with both eligible and ineligible types of neck pain, or participants with neck pain that is both chronic and non‐chronic, we will only include the studies if they separately report the outcome data for participants with eligible chronic neck pain.

Types of interventions

We will include RCTs that evaluate traditional East Asian needle acupuncture. In traditional East Asian medicine, acupuncture is embedded in a whole system context that engages needling of known acupuncture points, both locally and distally, to an identified problem area, using known points and channels with the option of additional points, based on palpation. We will only include studies in which acupuncture needles are used to puncture the skin.

We will exclude acupuncture‐type interventions that do not use the traditional East Asian concept of points and channels (e.g. trigger‐point needling), or do not use acupuncture needles to puncture the skin (e.g. acupressure, laser acupuncture). We will include trials that follow participants for at least four weeks after the start of treatment, since we are interested in effects on chronic neck pain, which is a long‐term condition.

We will include studies that compare acupuncture to sham acupuncture, to an inactive intervention, or to another active intervention. We will also include studies that compare acupuncture as an adjunct to other active therapies, versus those other therapies alone. We will compare acupuncture to all drug interventions in the primary analysis, and then compare acupuncture to the main types of drugs in a subgroup analysis (Subgroup analysis and investigation of heterogeneity).

We will not include studies that compare two different types, protocols, or doses of acupuncture to each other. We will exclude studies that assess acupuncture as part of a multi‐modal intervention package where it is not possible to isolate the effect of the acupuncture intervention.

The comparisons of interest will be:

1. acupuncture versus sham acupuncture;

2. acupuncture versus inactive intervention (wait list, or a minimal intervention, such as an educational pamphlet);

3. acupuncture versus other active intervention (in which each intervention is considered separately); and

4. acupuncture + other active intervention versus other active intervention alone (in which each intervention is considered separately)

For studies comparing acupuncture to an active intervention, we will group similar active interventions together, and use subgroup analyses to examine similar, but distinct, interventions separately. For example, if the active intervention is medication, we will carry out a main analysis of acupuncture versus medication, and within the analysis, we will group different medication classes into subgroups.

Concomitant treatments will be allowed, as long as they are given to both intervention groups.

Types of outcome measures

All outcomes will be assessed at short‐term (closest to four weeks), intermediate term (closest to six months), and long‐term (closest to one year) time points. We will consider the short‐term outcomes to be the outcomes of primary interest.

Primary outcomes

• Pain (e.g. numeric rating scale (NRS) or visual analogue scale (VAS));

• Disability (e.g. Neck Disability Index (NDI))

Secondary outcomes

• Physical quality of life (e.g. as measured on the SF‐36 PCS);

• Mental quality of life (e.g. as measured on the SF‐36 MCS)

• Patient satisfaction;

• Global perceived effect;

• Anxiety or depression (e.g. as measured on the Hospital Anxiety and Depression Scale (HADS));

• Work status (e.g. number of sick days);

• Adverse effects (e.g. participant withdrawal due to adverse effects, serious adverse events).

We will consider the minimum clinically important difference thresholds to be 1.3 for the numeric rating scale for pain (where 0 is no pain and 10 is the worst pain imaginable), and 19 percentage points for the NDI (where the total possible score is 100 percent (Cleland 2008)).

Search methods for identification of studies

Electronic searches

We will search the following databases from inception to the present with no language restrictions:

• Cochrane Central Register of Controlled Trials, includes Cochrane Back and Neck (CBN) Trials Register and Cochrane Complementary Medicine Field Specialized Register (CENTRAL; current issue via CRS web) in the Cochrane Library;

• MEDLINE Ovid (Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Daily and MEDLINE; 1946 to current);

• Embase OvidSP (1980 to current);

• Cumulative Index to Nursing and Allied Health Literature (CINAHL EBSCO; 1981 to current);

• AMED Ovid (Allied and Complementary Medicine; 1985 to current);

• Index to Chiropractic Literature (www.chiroindex.org/#results)

• U.S. National Institutes of Health ClinicalTrials.gov (www.clinicaltrials.gov)

• World Health Organization (WHO) International Clinical Trials Registry Platform (www.who.int/ictrp/search/en/)

The searches were developed using methods from Furlan 2015 and Higgins 2011. We have included a draft strategy for MEDLINE in Appendix 1.

We chose not to include Chinese databases in our search, as some studies have shown that the quality of reporting and conduct of Chinese trials (including traditional Chinese medicine studies) is poor, and we do not have the resources to contact Chinese authors to verify the methodological quality of their trials (He 2011; Wang 2007; Zhang 2008; Zhang 2011).

Searching other resources

We will screen the reference lists of included studies and any relevant systematic reviews retrieved by our searches or already known to our author team, and contact experts in the field for additional trials, including ongoing or unpublished studies.

Data collection and analysis

Selection of studies

Two authors (SH, THK) will independently screen the titles and abstracts of all retrieved citations, and assess the full text of all potentially eligible references. We will pilot test the inclusion criteria at the full‐text stage by having two authors (SH, THK) screen the first 10 full texts. Before continuing the full‐text selection, we will check the agreement rate and clarify anything that is unclear. For non‐English language papers, we will seek assistance from pairs of native speakers to assist in determining relevance. We will record reasons for exclusion at the full‐text stage in the 'Characteristics of excluded studies' table. We will resolve disagreements by consensus, or by involving a third author (MSL or AN).

Data extraction and management

One author (MSL, AN or THK) will extract the characteristics of each primary study using a standard, pre‐piloted data extraction form to collect the following details:

• Methods: study design, study setting, recruitment method, inclusion and exclusion criteria;

• Participants: number of participants randomized and analyzed, age, gender, diagnostic criteria, baseline pain and disability status;

• Interventions: description of acupuncture interventions, including the style of acupuncture, methods of point selection, number and names of selected points, total length of treatment period, the number, duration, and frequency of treatment sessions, whether or not de qi was sought, and method of needle stimulation;

• Comparators: description of control conditions and presence of cointerventions;

• Outcomes: listing of study outcomes; and

• Other: trial sponsorship, conflicts of interest, details of assessment of adverse effects (active or passive surveillance, whether and how authors defined serious adverse events, whether and how authors assessed adverse effects as related or unrelated to the acupuncture intervention).

A second author (LSW) will check the extraction of the characteristics of studies.

Two authors (MSL, AN, or THK) will independently extract the outcome data for our prespecified primary and secondary outcomes from each primary study, using a standard pre‐piloted extraction form. They will resolve any disagreements by consensus, or by consulting a third author if they cannot reach consensus. Should we identify non‐English language studies for inclusion, we will seek assistance from native speakers for single data extraction and checking of study characteristics, and dual extraction of outcome data from those studies.

Should any included trials involve authors of this review, we will ensure that the authors do not extract the data or assess the risk of bias for those studies. If necessary, we will solicit any necessary clarification of methods from those authors, and include this information in the notes of the relevant 'Characteristics of included studies' table. For all studies, we will contact authors to request any missing data on primary outcomes or key study characteristics.

When both endpoint and change data are available in the study report, we will use endpoint data in our primary analysis. In cases where study participants are lost to follow‐up, and both available case analyses and intention‐to‐treat analyses using imputation are presented, we will use the imputed data for our primary analysis. In cases where both unadjusted and adjusted data are presented, we will use the adjusted data for our primary analysis.

Assessment of risk of bias in included studies

Two authors (THK, LSW) will independently assess the risk of bias for each included study, using the criteria outlined in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions and adapted by the Cochrane Back and Neck (Furlan 2015; Higgins 2011; Table 1; Table 2). Information on risk of bias, including relevant quotations from the study and an explanation of the risk of bias decision, will be extracted into a standard pre‐piloted form based on Table 1 and added to the 'Risk of bias' tables in RevMan. One author will be a content expert (THK) and one author will be a methods expert (LSW). They will resolve disagreements by consensus, or by involving a third author.

Table 1.

Sources of risk of bias

Bias domain	Source of bias	Possible answers
Selection	(1) Was the method of randomization adequate?	Yes/no/unsure
Selection	(2) Was the treatment allocation concealed?	Yes/no/unsure
Performance	(3) Was the patient blinded to the intervention?	Yes/no/unsure
Performance	(4) Was the care provider blinded to the intervention?	Yes/no/unsure
Detection	(5) Was the outcome assessor blinded to the intervention?	Yes/no/unsure
Attrition	(6) Was the drop‐out rate described and acceptable?	Yes/no/unsure
Attrition	(7) Were all randomized participants analyzed in the group to which they were allocated?	Yes/no/unsure
Reporting	(8) Are reports of the study free of suggestion of selective outcome reporting?	Yes/no/unsure
Selection	(9) Were the groups similar at baseline regarding the most important prognostic indicators?	Yes/no/unsure
Performance	(10) Were cointerventions avoided or similar?	Yes/no/unsure
Performance	(11) Was the compliance acceptable in all groups?	Yes/no/unsure
Detection	(12) Was the timing of the outcome assessment similar in all groups?	Yes/no/unsure
Other	(13) Are other sources of potential bias unlikely?	Yes/no/unsure

Furlan 2015

Table 2.

Criteria for a judgment of ‘‘yes’’ for the sources of risk of bias

1	A random (unpredictable) assignment sequence. Examples of adequate methods are coin toss (for studies with 2 groups), rolling a dice (for studies with 2 or more groups), drawing balls of different colours, drawing ballots with the study group labels from a dark bag, computer‐generated random sequence, preordered sealed envelopes, sequentially‐ordered vials, telephone call to a central office, and preordered list of treatment assignments. Examples of inadequate methods are: alternation, birth date, social insurance or security number, date on which they are invited to participate in the study, and hospital registration number.
2	Assignment generated by an independent person not responsible for determining the eligibility of the patients. This person has no information about the persons included in the trial, and has no influence on the assignment sequence, or on the decision about eligibility of the patient.
3	Index and control groups are indistinguishable for the patients, or if the success of blinding was tested among the patients and it was successful.
4	Index and control groups are indistinguishable for the care providers, or if the success of blinding was tested among the care providers and it was successful.
5	Adequacy of blinding should be assessed for each primary outcome separately. This item should be scored yes if the success of blinding was tested among the outcome assessors and it was successful or: for patient‐reported outcomes in which the patient is the outcome assessor (e.g. pain, disability): the blinding procedure is adequate for outcome assessors if participant blinding is scored yes for outcome criteria assessed during scheduled visit, and that supposes a contact between participants and outcome assessors (e.g. clinical examination): the blinding procedure is adequate if patients are blinded, and the treatment or adverse effects of the treatment cannot be noticed during clinical examination for outcome criteria that do not suppose a contact with participants (e.g. radiography, magnetic resonance imaging): the blinding procedure is adequate if the treatment or adverse effects of the treatment cannot be noticed when assessing the main outcome for outcome criteria that are clinical or therapeutic events that will be determined by the interaction between patients and care providers (e.g. cointerventions, hospitalization length, treatment failure), in which the care provider is the outcome assessor: the blinding procedure is adequate for outcome assessors if item 4 (caregivers) is scored yes for outcome criteria that are assessed from data of the medical forms: the blinding procedure is adequate if the treatment or adverse effects of the treatment cannot be noticed on the extracted data
6	The number of participants who were included in the study but did not complete the observation period, or were not included in the analysis, must be described and reasons given. If the percentage of withdrawals and dropouts does not exceed 20% for short‐term follow‐up, and 30% for long‐term follow‐up, and does not lead to substantial bias, score a yes. (N.B. these percentages are arbitrary, not supported by literature).
7	All randomized patients are reported and analyzed in the group to which they were allocated by randomization for the most important moments of effect measurement (minus missing values), irrespective of noncompliance and cointerventions.
8	All the results from all prespecified outcomes have been adequately reported in the published report of the trial. This information is either obtained by comparing the protocol and the report, or in the absence of the protocol, assessing that the published report includes enough information to make this judgment.
9	Groups have to be similar at baseline regarding demographic factors, duration and severity of complaints, percentage of patients with neurological symptoms, and value of main outcome measure(s).
10	If there were no cointerventions, or they were similar between the index and control groups.
11	The reviewer determines if the compliance with the interventions is acceptable, based on the reported intensity, duration, number, and frequency of sessions for both the index intervention and control intervention(s). For example, physiotherapy treatment is usually administered for several sessions; therefore it is necessary to assess how many sessions each patient attended. For single‐session interventions (e.g. surgery), this item is irrelevant.
12	Timing of outcome assessment should be identical for all intervention groups and for all primary outcome measures.
13	Other types of biases. For example: When the outcome measures were not valid. There should be evidence from a previous or present scientific study that the primary outcome can be considered valid in the context of the present. Industry‐sponsored trials. The conflict of interest (COI) statement should explicitly state that the researchers have had full possession of the trial process from planning to reporting, without funders with potential COI having any possibility to interfere in the process. If, for example, the statistical analyses have been done by a funder with a potential COI, usually one should score unsure.

Furlan 2015

We will assess the risk of bias according to the following domains:

• Selection bias (random sequence generation,allocation concealment, and group similarities at baseline);

• Performance bias (blinding of participants, blinding of personnel or care providers, cointerventions, compliance);

• Detection bias (blinding of outcome assessors, timing of outcome assessments);

• Attrition bias (incomplete outcome data, intention‐to‐treat analysis);

• Reporting bias (selective reporting);

• Other bias (other sources of bias related to a particular trial design or specific circumstances, such as conflicts of interest (Furlan 2015)).

We will classify the risk of bias as low, high, or unclear in each domain (Higgins 2011a). We will consider trials that are at low risk of bias in each domain to be at overall low risk of bias and carry out sensitivity analyses by overall risk of bias (Sensitivity analysis). We will provide graphic summaries of risk of bias in the completed review.

Measures of treatment effect

We will calculate risk ratio (RR) and 95% confidence intervals (CI) for dichotomous data, except for adverse effects, for which we will calculate the risk difference (RD) and 95% CI. We will calculate the mean difference (MD) or the standardized mean difference (SMD) and 95% CI for continuous data.

Unit of analysis issues

We do not expect to find relevant cluster‐randomized trials. However, if we find such trials, we will follow the guidance in chapter 16.3 of the Cochrane Handbook for Systematic Reviews of Interventions to assess their suitability, and include them in the analysis if appropriate (Higgins 2011b). If we find eligible cross‐over trials, we will follow the guidance in chapter 16.4 of the Cochrane Handbook for Systematic Reviews of Interventions to assess their suitability, and include the first period of the trial in the analysis if appropriate (Higgins 2011b).

Dealing with missing data

If key study characteristics or primary outcome data are missing or incomplete, we will contact the first author or primary investigator to obtain the missing information.

Assessment of heterogeneity

We will assess clinical heterogeneity between studies qualitatively. For studies that are considered sufficiently clinically homogenous to combine in a meta‐analysis, we will assess the presence of statistical heterogeneity using the Chi² test. We will consider P < 0.10 as evidence of statistically significant heterogeneity.

The I² statistic describes the percentage of variability in the overall effect that is due to heterogeneity rather than to chance (Deeks 2011). The suggested thresholds for interpretation of I² percentages are as follows:

• 0% to 40%: might not be important;

• 30% to 60%: may represent moderate heterogeneity;

• 50% to 90%: may represent substantial heterogeneity;

• 75% to 100%: considerable heterogeneity.

We will assess the importance of the value of the I² statistic in light of the range and direction of effects as observed from the forest plots, and the strength of the evidence for heterogeneity based upon the Chi² test (Deeks 2011). We will also use subgroup analyses to explore clinical heterogeneity (see Subgroup analysis and investigation of heterogeneity).

Assessment of reporting biases

For meta‐analyses in which at least 10 studies are included, we will use the Egger's test and a funnel plot to assess the potential for small study bias (Egger 1997; Sterne 2011). We will also evaluate the possibility of selective outcome reporting for each study as part of the assessment of risk of bias.

Data synthesis

We will use meta‐analysis to combine the outcome data across trials when there are available data and sufficient clinical comparability of population, interventions, outcomes, and timing of assessment. In our main analysis, we will group all subtypes of chronic neck pain (e.g. neck pain with radicular symptoms, neck pain without radicular symptoms, whiplash). We will then carry out subgroup analyses by type of chronic neck pain, if the data permit (see Subgroup analysis and investigation of heterogeneity). We will use a random‐effects model for all analyses, because we expect some between‐study variation. When it is not possible to carry out a meta‐analysis, we will carry out a qualitative description of the results from clinically similar trials.

We will summarize dichotomous data using the RR and 95% CI for effectiveness outcomes, and the RD and 95% CI for adverse effects. We will summarize continuous data using the MD and 95% CI for measures using the same scale, or the SMD and 95% CI when different scales were used to evaluate the same outcome.

Regardless of whether sufficient data are available to use quantitative analyses to summarize the data, we will assess the overall certainty of the evidence for each outcome. We will use the GRADE approach, as recommended in chapter 12.2 of the Cochrane Handbook for Systematic Reviews of Interventions and adapted in the updated CBN Review Group method guidelines, to assess the certainty of the evidence (Furlan 2015; Schünemann 2011). Two review authors (MSL, THK) will independently apply the GRADE approach to each outcome; in cases of disagreement they will reach consensus through discussion or involvement of a third author (LSW). The levels of certainty of the evidence and the criteria for judgement are listed in Appendix 2.

'Summary of findings' table

We will create a 'Summary of findings' table (SoF) for our main comparison, acupuncture versus sham acupuncture. We will present the primary outcomes of pain and disability, and adverse effects. We will present the primary outcomes at the short‐term time point (closest to four weeks; see Table 3).

Table 3.

Acupuncture versus sham acupuncture

Acupuncture versus sham acupuncture for chronic neck pain
Patient or population: patients with chronic neck pain Settings: outpatient Intervention: acupuncture Comparison: sham acupuncture
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Sham acupuncture	Acupuncture
Pain (VAS; 0 to 10 scale; less is better) Follow‐up: short‐term (closest to four weeks)
Disability (Neck Disability Index (NDI); 0 to 50; less is better) Follow‐up: short‐term (closest to four weeks)
Withdrawals due to adverse events Follow‐up: at any time point
Serious adverse events Follow‐up: at any time point
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). [review] CI: Confidence interval; SMD: Standardised mean difference; RR: Risk ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Subgroup analysis and investigation of heterogeneity

If there are available data, we will carry out subgroup analyses based upon factors that may cause the effect of the intervention to vary. This will include the following factors:

• Type of neck pain (e.g. radicular pain, nonradicular pain, whiplash pain), as different types of pain may respond differently to acupuncture;

• Type of needle stimulation (e.g. manual stimulation, electrical stimulation), as different types of needle stimulation may produce different effects;

• Type of drug comparator (e.g. opioid, NSAID), as these drug classes have different general characteristics; and

• Number of acupuncture sessions (fewer than eight sessions versus eight or more sessions), as some recent research suggests that eight or more sessions may be associated with greater effects (McKee 2018).

For each factor with two or more subgroups, we will use the I² statistic to assess the percentage of the variability in effect estimates that is due to genuine subgroup differences rather than to chance (Deeks 2011), however we will be cautious about drawing conclusions based on between‐studies differences.

Sensitivity analysis

If there are available data, we will carry out sensitivity analyses to explore the impact of the following factors:

• Including all trials versus including only trials at low risk of bias (as defined in Assessment of risk of bias in included studies);

• Using endpoint outcome data versus using change outcome data;

• Using imputed outcome data versus using available case outcome data; and

• Using adjusted ocutome data versus using unadjusted outcome data.

Appendices

Appendix 1. MEDLINE search strategy

1. randomized controlled trial.pt.

2. controlled clinical trial.pt.

3. pragmatic clinical trial.pt.

4. random*.ti,ab.

5. placebo.ab,ti.

6. drug therapy.fs.

7. trial.ab,ti.

8. groups.ab.

9. or/1‐8

10. (animals not (humans and animals)).sh.

11. 9 not 10

12. neck pain/

13. (neck pain or neckache? or neck ache? or cervicodynia or cervicalgia).ti,ab.

14. neck muscles/

15. neck/

16. neck injuries/

17. cervical plexus/

18. neck.tw,kf.

19. neck disorder?.tw,kf.

20. whiplash injuries/

21. whiplash.tw,kf.

22. myofascial pain syndromes/

23. (myofascial adj3 (pain or trigger point?)).tw,kf.

24. radiculopathy/

25. (cervical adj3 (radiculopath* or pain)).tw,kf.

26. cervical spine.tw,kf.

27. cervical vertebrae/

28. trapezius.tw,kf.

29. (cervicobrachial* or cervico‐brachial*).tw,kf.

30. cervicogenic headache?.tw,kf.

31. or/12‐30

32. exp Acupuncture Therapy/

33. exp Acupuncture/

34. electroacupuncture.tw,kf.

35. moxibustion.tw,kf.

36. acupuncture.tw,kf.

37. exp Meridians/

38. meridian*.tw,kf.

39. trigger points/

40. trigger point*.tw,kf.

41. dry needl*.tw,kf.

42. or/32‐41

43. 11 and 31 and 42

Appendix 2. The GRADE approach to assessing the certainty of the evidence

The certainty of evidence will be categorized as follows:

• High (⊙⊙⊙⊙): further research is very unlikely to change the confidence in the estimate of effect.

• Moderate (⊙⊙⊙○): further research is likely to have an important impact in the confidence in the estimate of effect.

• Low (⊙⊙○○): further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

• Very Low (⊙○○○): any estimate of effect is very uncertain.

The evidence available to answer each sub‐question will be graded on the domains in the following manner:

1. Study design

Confidence in the estimate of the effect is higher if it is based upon randomised controlled trials rather than observational studies. All included studies will be randomised controlled trials, and therefore, we will begin with a high level of confidence in the estimates of effect, and downgrade certainty of the evidence if there are problems in the following domains.

2. Risk of bias

Confidence in the estimate of the effect decreases if studies have major limitations in design and conduct. We assessed five types of bias, described in detail in Table 1 and Table 2: • selection bias (random sequence generation, allocation concealment, group similarities at baseline); • performance bias (blinding of participants, blinding of personnel or care providers, cointerventions, compliance); • attrition bias (incomplete outcome data, intention‐to‐treat analysis); • detection bias (blinding of outcome assessors, timing of outcome assessments); • reporting bias (selective reporting).

We will downgrade the certainty of evidence one level for an estimate of effect that relies on studies with a high risk of bias in one of these domains. The certainty of evidence will be downgraded by two levels for an estimate of effect that relies heavily (i.e. approximately 50% or greater weight in the meta‐analysis) on studies with a high risk of bias in two or more of these domains.

3. Inconsistency

Inconsistency refers to heterogeneity between studies that does not have a plausible explanation. The I² statistic is an estimate of the percentage of the variability in effect estimates that is due to clinical or methodological heterogeneity rather than chance. An I² statistic of 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, and 75% to 100% may represent considerable heterogeneity (Deeks 2011). The design and conduct of acupuncture interventions for neck pain are highly variable, and therefore, heterogeneity is expected.

The certainty of evidence will be downgraded by one level when heterogeneity is substantial (i.e. when I² ≥ 50%), and by two levels when heterogeneity is considerable (i.e. when I² ≥ 75%), unless there is a plausible explanation for the heterogeneity.

4. Indirectness

Indirectness refers to a mismatch between the population, intervention, comparator, or outcomes for the studies included in the review and the population, intervention, comparator, or outcomes for the research question being posed by the systematic review.

The certainty of evidence will be downgraded by one level when there is indirectness for one element of the research question (e.g. outcome), and by two levels when there is indirectness for two or more elements of the research question.

5. Imprecision

Imprecision refers to uncertainty in the results due to few participants or to wide confidence intervals.

For continuous outcomes, we will downgrade the certainty of evidence by one level when: • the total number of participants is fewer than 400; or • the 95% confidence interval around the estimate of effect covers both no effect and a minimally important difference for that outcome; or if a minimally important difference is not prespecified, no effect and a standardized mean difference (SMD) of ± 0.5. An outcome will be downgraded two levels for imprecision if both points above are true.

For dichotomous outcomes, we will downgrade the certainty of the evidence by one level when: • the total number of events is less than 300; or • the 95% confidence interval around the estimate of effect includes both no effect and either appreciable benefit or appreciable harm. The threshold for ’appreciable benefit’ or ’appreciable harm’ will be a relative risk reduction (RRR) or relative risk increase (RRI) greater than 25%. An outcome will be downgraded two levels for imprecision if both points above are true.

6. Publication bias

Publication bias refers to the selective publication of studies, which may bias the estimate of effect that is based on available studies.

The certainty of evidence will be downgraded by one level if a funnel plot to assess the potential for small‐study bias suggests that publication bias is present, or there is any other reason to strongly suspect that publication bias is present.

Contributions of authors

MSL, AN, THK, IHH, SH and LSW drafted the protocol.

Sources of support

Internal sources

• Korea Institute of Oriental Medicine, Korea, South.

  Myeong Soo Lee was supported by KIOM (KSN1522120).

External sources

• National Institutes of Health (NIH), USA.

  Support was received from the National Center for Complementary and Integrative Health of the NIH under award number R24AT001293. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Declarations of interest

MSL has no known conflicts of interest.

AN has no known conflicts of interest.

THK has no known conflicts of interest.

IHH has no known conflicts of interest.

SH is Managing Editor of Cochrane Back and Neck review group. She was not involved in any of the Editorial decisions regarding this review and has no known conflicts of interest.

LSW has no known conflicts of interest.

New