Massage for neck pain

Abstract

Background

Massage is widely used for neck pain, but its effectiveness remains unclear.

Objectives

To assess the benefits and harms of massage compared to placebo or sham, no treatment or exercise as an adjuvant to the same co‐intervention for acute to chronic persisting neck pain in adults with or without radiculopathy, including whiplash‐associated disorders and cervicogenic headache.

Search methods

We searched multiple databases (CENTRAL, MEDLINE, EMBASE, CINAHL, Index to Chiropractic Literature, trial registries) to 1 October 2023.

Selection criteria

We included randomised controlled trials (RCTs) comparing any type of massage with sham or placebo, no treatment or wait‐list, or massage as an adjuvant treatment, in adults with acute, subacute or chronic neck pain.

Data collection and analysis

We used the standard methodological procedures expected by Cochrane. We transformed outcomes to standardise the direction of the effect (a smaller score is better). We used a partially contextualised approach relative to identified thresholds to report the effect size as slight‐small, moderate or large‐substantive.

Main results

We included 33 studies (1994 participants analysed). Selection (82%) and detection bias (94%) were common; multiple trials had unclear allocation concealment, utilised a placebo that may not be credible and did not test whether blinding to the placebo was effective.

Massage was compared with placebo (n = 10) or no treatment (n = 8), or assessed as an adjuvant to the same co‐treatment (n = 15). The trials studied adults aged 18 to 70 years, 70% female, with mean pain severity of 51.8 (standard deviation (SD) 14.1) on a visual analogue scale (0 to 100). Neck pain was subacute‐chronic and classified as non‐specific neck pain (85%, including n = 1 whiplash), radiculopathy (6%) or cervicogenic headache (9%). Trials were conducted in outpatient settings in Asia (n = 11), America (n = 5), Africa (n = 1), Europe (n = 12) and the Middle East (n = 4). Trials received research funding (15%) from research institutes. We report the main results for the comparison of massage versus placebo.

Low‐certainty evidence indicates that massage probably results in little to no difference in pain, function‐disability and health‐related quality of life when compared against a placebo for subacute‐chronic neck pain at up to 12 weeks follow‐up. It may slightly improve participant‐reported treatment success. Subgroup analysis by dose showed a clinically important difference favouring a high dose (≥ 8 sessions over four weeks for ≥ 30 minutes duration). There is very low‐certainty evidence for total adverse events. Data on patient satisfaction and serious adverse events were not available.

Pain was a mean of 20.55 points with placebo and improved by 3.43 points with massage (95% confidence interval (CI) 8.16 better to 1.29 worse) on a 0 to 100 scale, where a lower score indicates less pain (8 studies, 403 participants; I² = 39%). We downgraded the evidence to low‐certainty due to indirectness; most trials in the placebo comparison used suboptimal massage doses (only single sessions). Selection, performance and detection bias were evident as multiple trials had unclear allocation concealment, utilised a placebo that may not be credible and did not test whether blinding was effective, respectively.

Function‐disability was a mean of 30.90 points with placebo and improved by 9.69 points with massage (95% CI 17.57 better to 1.81 better) on the Neck Disability Index 0 to 100, where a lower score indicates better function (2 studies, 68 participants; I² = 0%). We downgraded the evidence to low‐certainty due to imprecision (the wide CI represents slight to moderate benefit that does not rule in or rule out a clinically important change) and risk of selection, performance and detection biases.

Participant‐reported treatment success was a mean of 3.1 points with placebo and improved by 0.80 points with massage (95% CI 1.39 better to 0.21 better) on a Global Improvement 1 to 7 scale, where a lower score indicates very much improved (1 study, 54 participants). We downgraded the evidence to low‐certainty due to imprecision (single study with a wide CI that does not rule in or rule out a clinically important change) and risk of performance as well as detection bias.

Health‐related quality of life was a mean of 43.2 points with placebo and improved by 5.30 points with massage (95% CI 8.24 better to 2.36 better) on the SF‐12 (physical) 0 to 100 scale, where 0 indicates the lowest level of health (1 study, 54 participants). We downgraded the evidence once for imprecision (a single small study) and risk of performance and detection bias.

We are uncertain whether massage results in increased total adverse events, such as treatment soreness, sweating or low blood pressure (RR 0.99, 95% CI 0.08 to 11.55; 2 studies, 175 participants; I² = 77%). We downgraded the evidence to very low‐certainty due to unexplained inconsistency, risk of performance and detection bias, and imprecision (the CI was extremely wide and the total number of events was very small, i.e < 200 events).

Authors' conclusions

The contribution of massage to the management of neck pain remains uncertain given the predominance of low‐certainty evidence in this field. For subacute and chronic neck pain (closest to 12 weeks follow‐up), massage may result in a little or no difference in improving pain, function‐disability, health‐related quality of life and participant‐reported treatment success when compared to a placebo. Inadequate reporting on adverse events precluded analysis. Focused planning for larger, adequately dosed, well‐designed trials is needed.

Plain language summary

Massage for neck pain

Key message

This review shows that in subacute (medium‐term) or persisting neck pain, at closest to 12 weeks follow‐up, massage compared to a 'dummy' massage probably results in little to no difference in pain, function‐disability, quality of life and participant‐reported treatment success. Possible side effects may include treatment soreness.

Background

Neck pain is a common condition in adults leading to acute (< 4 weeks), subacute (4 to 12 weeks) or persisting (>12 weeks) neck pain, disability and substantial economic costs. It may lead to headaches stemming from the neck, referral of pain into the upper back and arms, and signs of weakness or numbness in the arms. It can be affected by bones, joints, muscles, ligaments, tendons and nerves, and be influenced by social, psychological and personal factors.

Massage therapy involves the movement of the soft tissues of the body through manual touch to reduce pain and muscle tension, and promote relaxation. Because massage therapy is commonly used, inexpensive and has minimal side effects, it is of great interest to people with neck pain.

What did we want to find out?

What are the benefits and risks of massage for treating acute, subacute and chronic persisting neck pain in adults with or without arm pain and headache or associated with a whiplash injury? Does the dose (frequency per week, total number of weeks, session duration) of massage influence findings?

What did we find?

We found 33 studies that involved 1994 people suffering subacute and persisting moderate to very strong neck pain including arm pain (6%), whiplash (3%) or headache stemming from the neck (9%). There were no studies addressing acute pain. Studies examined adults aged 18 to 70 years, who were mostly female (70%). Trials were conducted in outpatient settings in Asia, America, Africa, Europe and the Middle East. Funding, when reported (in 15% of studies), was from research centres or universities. We included studies evaluating many massage techniques (with ischaemic compression, a sustained pressure over the soft tissue, being the most common technique) that vary in the manner in which touch is applied, the amount of pressure and intensity that is applied, and session frequency. We did not include techniques that used a massage tool to deliver treatment or non‐touch energy techniques.

Compared with a 'dummy' massage treatment, massage probably results in little to no difference in the following:

• Pain (3% improvement): People who had a 'dummy' massage rated their pain at 20.55 points while people who had massage rated their post‐treatment pain as 17.12 points. Those who used massage improved by 3.43 points (8.16 better to 1.29 worse) on a 0‐ to 100‐point scale, where a lower score means less pain.

• Function‐disability (10% improvement): People who had a 'dummy' massage rated their function‐disability at 30.9 points. People who had massage rated their post‐treatment function‐disability as 21.21 points. Those who used massage improved by 9.69 points (17.57 better to 1.81 better) on a 0‐ to 100‐point scale, where a lower score means better function‐disability.

• Participant‐reported treatment success (11% improvement): People who had a 'dummy' massage rated their self‐reported treatment success at 3.1 points. People who had massage treatment rated their post‐treatment success as 2.3 points. Those who used massage improved by 0.80 points (1.39 better to 0.21 better) on a 1‐ to 7‐point scale, where a lower score means more self‐reported treatment success.

• Health‐related quality of life (5% improvement): People who had a 'dummy' massage rated their quality of life at 43.2 points. People who had massage rated their post‐treatment quality of life as 48.5 points. Those who used massage improved by 5.30 points (8.24 better to 2.36 better) on a 0‐ to 100‐point scale, where a higher score means better quality of life.

• Patient satisfaction and serious adverse events ‐ these outcomes were not reported.

• Minor adverse events ‐ minor adverse (unwanted or harmful) events were poorly reported, were the same as with the dummy massage and included minor temporary effects such as mild treatment soreness.

What are the limitations of the evidence?

We have little confidence in our findings that massage provides little to no benefit in terms of pain, function‐disability, quality of life and participant‐reported treatment success compared to a 'dummy' massage, because it is possible that people in the studies were aware of which treatment they were getting. We believe that these biases may result in either an over or underestimation of the size of the effect reported. The occurrence of adverse events was not commonly reported. The number of participants in most trials was small. Focused planning of studies with larger numbers of participants, adequate massage doses and control for study biases is needed.

How up‐to‐date is this evidence?

We searched databases up to 1 October 2023.

Summary of findings

Summary of findings 1. Summary of findings: Massage compared with placebo or sham for neck pain.

Massage compared with placebo or sham for neck pain
Patient or population: neck pain Setting: outpatient ambulatory care services Intervention: massage (various types) Comparison: placebo or sham
Outcome	No. of participants (studies)	Relative effect (95% CI)	Anticipated absolute effects (95% CI)			Certainty	What happens
			Placebo or sham	Massage	Difference
Pain at closest to 12 weeks follow‐up Assessed with: 0 to 100 scale, lower score indicates lower pain intensity	403 (8 RCTs)	‐	The mean pain was 20.55points	17.12	MD 3.43 points better (8.16 better to 1.29 worse)	⨁⨁◯◯ Lowa	Massage may result in little to no difference in pain in the short term.
Function‐disability at closest to 12 weeks follow‐up Assessed with: NDI 0 to 100, lower score indicates lower disability	68 (2 RCTs)	‐	The mean function‐disability was 30.90 points	21.21	MD 9.69 points better (17.57 better to 1.81 better)	⨁⨁◯◯ Lowb	Massage may result in a slight improvement to no difference in function‐disability in the short term.
Participant‐reported treatment success at closest to 12 weeks follow‐up Assessed with: Patient Impression of Global Improvement (PGI‐I) 1 to 7, lower score indicates very much improved	54 (1 RCT)	‐	The mean participant‐reported treatment success was 3.1points	2.3	MD 0.8 points better (1.39 better to 0.21 better)	⨁⨁◯◯ Lowb	Massage may slightly increase participant‐reported treatment success in the short term.
Health‐related quality of life at closest to 12 weeks follow‐up Assessed with: SF‐12 physical component (PCS) from 0 to 100, 0 indicating lowest level of health	54 (1 RCT)	‐	The mean health‐related quality of life was 43.2points	48.5	MD 5.3 points better (8.24 better to 2.36 better)	⨁⨁◯◯ Lowb	Massage may result in little to no difference in health‐related quality of life in the short term.
Patient satisfaction ‐ not reported	‐	‐	‐	‐	‐	‐	‐
Adverse events ‐ minor, such as slight post‐treatment soreness, sweating, low blood pressure, at closest to 12 weeks follow‐up	175 (2 RCTs)	RR 0.99 (0.08 to 11.55)	14.8%	14.6% (1.2 to 100)	0.1% fewer (13.6 fewer to 155.9 more)	⨁◯◯◯ Very lowc	We are uncertain whether massage results in increased adverse events.
Adverse events ‐ serious ‐ not reported	‐	‐	‐	‐	‐	‐	‐
*The risk in the intervention group (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). CI: confidence interval; MD: mean difference; NDI: Neck Disability Index; PGI‐I: Patient Impression of Global Improvement; RCT: randomised controlled trial; RR: risk ratio; SF‐12 (PCS): Short‐Form 12 (physical component scale); VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Explanations aDowngraded by two levels, once due to indirectness and once due to risk of selection, performance and detection bias. Most trials in the placebo comparison used suboptimal massage doses (only single sessions), which may not be applicable to clinical practice. Additionally, selection, performance and detection bias affected the certainty of the evidence. bDowngraded by two levels, once due to imprecision and once due to risk of bias (performance and detection bias). Studies included relatively few patients and there was a wide CI resulting in uncertainty about the results; the lower versus the upper boundary of the CI represented a slight to moderate improvement in function‐disability (where a minimally important change is defined as 10 points on a 0‐ to 100‐point scale), global perceived effect (where a minimal change is defined as 0.7 on a 1‐ to 7‐point scale) and quality of life (where a minimal change is defined as 5 to 10 SF‐12 points on a 0‐ to 100‐point scale). Performance and detection bias were noted. One trial for the function‐disability outcome (contributing few patients to the meta‐analysis, but at high risk of bias across most domains) affected the interpretation of results; when removed, the effect estimate shifted to a moderate effect (MD 10.80 better, 95% CI 19.47 better to 2.13 better). In this case, the overall limitations were not serious and the evidence was not downgraded for risk of selection or detection bias. However, performance bias, where the therapist is not blind to the massage intervention, was unavoidable. cDowngraded by three levels due to inconsistency (I2 = 77%), risk of bias (reporting bias) and imprecision as the 95% confidence intervals are extremely wide and the total number of events was very small (< 200).

Summary of findings 2. Summary of findings: Massage compared with no treatment for neck pain.

Massage compared with no treatment for neck pain
Patient or population: neck pain Setting: outpatient ambulatory care services Intervention: massage Comparison: no treatment (no intervention/wait‐list)
Outcome	№ of participants (studies)	Relative effect (95% CI)	Anticipated absolute effects (95% CI)			Certainty	What happens
			No treatment	Massage	Difference
Pain at closest to 12 weeks follow‐up Assessed with: 0 to 100 scale, lower score indicates lower pain intensity	299 (6 RCTs)	‐	The mean pain was 46.83 points	22.59	MD 24.24 points better (28.48 better to 19.99 better)	⨁⨁⨁◯ Moderatea	Massage for neck pain probably results in a large reduction in pain in the short term compared with no treatment.
Function‐disability at closest to 12 weeks follow‐up Assessed with: NDI 0 to 100, lower score indicates lower disability	300 (6 RCTs)	‐	The mean function‐disability was 31.94 points	22.43	MD 9.51 points better (13.68 better to 5.33 better)	⨁⨁⨁◯ Moderateb	Massage for neck pain probably result in a slight improvement in function‐disability in the short term compared with no treatment.
Health‐related quality of life at closest to 12 weeks follow‐up Assessed with: SF‐12 physical component (PCS) 0 to 100, 0 indicates the lowest level of health	89 (1 RCT)	‐	The mean health‐related quality of life was 42.40points	48.1	MD 5.7 points better (8.84 better to 2.56 better)	⨁⨁◯◯ Lowc	Massage for neck pain may result in little to no difference in health‐related quality of life in the short‐term compared with no treatment.
Participant‐reported treatment success ‐ not reported	‐	‐	‐	‐	‐	‐	‐
Patient satisfaction ‐ not reported	‐	‐	‐	‐	‐	‐	‐
Adverse events (minor or serious) ‐ not reported	‐	‐	‐	‐	‐	‐	‐
*The risk in the intervention group (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). CI: confidence interval; MD: mean difference; NDI: Neck Disability Index; RCT: randomised controlled trial; SF‐12 (PCS): Short‐Form 12 (physical component scale); VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Explanations aDowngraded once for high risk of selection and detection bias; there may be high risk of detection bias for self‐reported outcomes and an unclear risk of selection bias for allocation concealment. A score of lower than 10 points is shown to be of little or no difference for patients. Additional dichotomous data at short‐term follow‐up showed a RR of 0.31 (95% CI 0.17 to 0.57) for pain events (where an event includes those who continue to have substantive pain). This showed 51.3% fewer (61.7 fewer to 31.9 fewer) pain events. bDowngraded once for high risk of bias (selection, performance and detection bias). A change of 5 to 10 on the NDI (0 to 100 points) is shown to be a small difference for patients. Sensitivity analysis by allocation concealment may help to explain the inconsistency in the results: adequate allocation concealment (MD 9.48 better, 95% CI 12.87 better to 6.08 better, I2 = 0%) versus all studies (MD 9.51 better, 95% CI 13.68 better to 5.33 better). Caution should be applied as there is a paucity of studies. cDowngraded by two levels, once for risk of detection bias and once for imprecision, with data from a single small trial. A change of 5 to 10 SF‐12 PCS points is shown to be a small difference for patients.

Summary of findings 3. Summary of findings: Massage as an adjuvant to the same co‐treatment for neck pain.

Massage as an adjuvant to the same co‐treatment for neck pain
Patient or population: neck pain Setting: outpatient ambulatory care services Intervention: massage as an adjuvant to a co‐treatment Comparison: same co‐treatment
Outcome	№ of participants (studies)	Relative effect (95% CI)	Anticipated absolute effects (95% CI)			Certainty	What happens
			Co‐treatment	Massage as an adjuvant to the same co‐treatment	Difference
Pain at closest to 12 weeks follow‐up Assessed with: VAS 0 to 100, lower score indicates lower pain intensity	795 (12 RCTs)	‐	The mean pain was 34.41 points	27.12	MD 7.29 points better (11.43 lower to 3.16 lower)	⨁⨁◯◯ Lowa,b	Massage as an adjuvant to a co‐treatment for neck pain may result in little to no difference in pain in the short term.
Function‐disability at closest to 12 weeks follow‐up Assessed with: varied tools (NDI 0 to 100, NDI 0 to 50, NPQ 0 to 100, CGPQ 0 to 10), lower score indicates lower disability	532 (7 RCTs)	‐	The mean function‐disability was 42 pointsc	33.72	MD 8.28 better (16.56 lower to 0.18 lower)c	⨁◯◯◯ Very lowa, d	SMD 0.46 SD better (0.92 better to 0.01 better) Massage as an adjuvant to a co‐treatment for neck pain may have little to no effect on function‐disability in the short term, but the evidence is very uncertain.
Health‐related quality of life at closest to 12 weeks follow‐up Assessed with: varied scales (SF‐36 0 to 100; EQ5D 0 to 1), lower score indicates lower quality of life	352 (2 RCTs)	‐	The mean health‐related quality of life was 88.9 pointsc	89.04	MD 0.14 better (0.73 better to 0.43 worse)c	⨁⨁◯◯ Lowe,f	SMD 0.1 SD better (0.52 better to 0.31 worse) Massage as an adjuvant to a co‐treatment for neck pain may not improve health‐related quality of life in the short‐term.
Participant‐reported treatment success ‐ not reported	‐	‐	‐	‐	‐	‐
Patient satisfaction ‐ not reported	‐	‐	‐	‐	‐	‐
Adverse events (minor or serious) ‐ not reported	‐	‐	‐	‐	‐	‐
*The risk in the intervention group (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). CI: confidence interval; CPGQ: Chronic Pain Grade Questionnaire; EQ5D: European Quality of Life‐5 dimensions; MD: mean difference; NDI: Neck Disability Index; NPQ: Northwick Park Pain Questionnaire; SF‐36: short form‐36; SMD: standardised mean difference; VAS: visual analogue scale
Explanations aDowngraded one level due to high risk of selection and detection bias: selection bias due to unclear allocation concealment and detection bias due to self‐reported outcomes. A score of 5 to 10 points is shown to be a small improvement for patients. bDowngraded once for inconsistency (I2 = 69%): the effect measures of the summary estimates for two categories (disorder classification and co‐treatment type) had wide CIs, with overlap of the CIs and thus could be considered not statistically significant. cControl group risk was estimated from the placebo value at follow‐up for function‐disability and health‐related quality of life in Skillgate 2020. The SMD was back‐translated to a typical scale by multiplying the SMD by the among‐person standard deviation (e.g. the standard deviation of the control group at baseline from Skillgate 2020: function‐disability = SD 18.0 CPGQ (transformed to 0 to 100); health‐related quality of life SD = 1.4 EQ‐5D (transformed to 0 to 100)). dDowngraded once for inconsistency (I² = 79%): heterogeneity may in part be explained by the disorder classification as well as symptom duration; the non‐overlap of the CIs indicates statistical significance. However, the power to detect a difference is low and thus these may be spurious findings. eDowngraded once for high risk of bias: selection and detection bias due to self‐reported outcomes. fDowngraded once for inconsistency (I² = 58%). Sensitivity analysis was not feasible due to a paucity of studies (n = 2).

Background

Description of the condition

Neck pain with its associated disorders, including cervicogenic headache and pain radiating into the upper back and arms, is a highly prevalent condition leading to substantial pain, disability and economic costs (Haldeman 2008; Safiri 2020). This translates to a global point prevalence of 4.9%, with nearly half of these cases becoming chronic (Hoy 2014; Munoz‐Garcia 2016). Neck pain can be defined as acute (< 4 weeks), subacute or chronic (>12 weeks). Many non‐specific causal factors may contribute to the painful condition, such as impairments related to the neuromusculoskeletal system and other biopsychosocial factors. Neck pain can be episodic and it impacts people to various degrees over their lifetime (Haldeman 2008). Neck pain has a considerable impact on function, which leads to a limited ability to participate in activities required for daily living. As a result, neck pain is the fourth‐highest cause of global disability (Cieza 2020; Hoy 2014; Vos 2020). Neck pain and function‐disability are therefore considered to be primary outcomes in this review.

Neck pain is associated with large healthcare costs and loss of productivity attributable to sick leave taken from work. Three to 11 percent of claimants are off work each year due to neck pain (Côté 2008), with 14.2% of those required to miss work due to neck pain experiencing at least one reoccurrence resulting in subsequent absenteeism in the following two years (Van Eerd 2011). Consequently, the total cost of neck pain in 2013 was determined to be USD 87.6 billion (Dieleman 2016). Despite neck pain posing a large economic and physical burden, no evidence‐based consensus on its management has been drawn at this time. However, massage and soft tissue work are commonly used (79% of physiotherapists or chiropractors) for neck pain treatment by massage therapists, physiotherapists and chiropractors, and used 92% to 95% of the time for acute or chronic neck pain, respectively, based on an international survey (Carlesso 2014).

Description of the intervention

Massage therapy involves the manipulation of the soft tissues of the body through manual touch (Sherman 2006). Massage techniques include Swedish techniques, fascial or connective tissue release techniques, cross fibre friction and myofascial trigger point techniques (Kutner 2008). See Appendix 1 for a glossary of massage and bodywork terms. Techniques vary in the manner in which touch is applied, as well as the amount of pressure and intensity that is applied (Kutner 2008). There is a wide spectrum of interventions and techniques that fall under the umbrella term of massage therapy; specific definitions for the techniques are lacking and there is substantial overlap amongst them. As such, Sherman et al proposed a three‐level classification system for the different massage therapy types based on the goals of the treatment, the style and the technique (Sherman 2006). The proposed treatment goals were classified into the following: relaxation massage, clinical massage, movement re‐education and energy work. Relaxation massage proposes to move body fluids, nourish cells and relax muscles with the aim of reducing pain. Clinical massage involves focused soft tissue manipulation of the muscle and fascia, and may address the lymphatic, circulatory and nervous systems. Movement re‐education involves massage that is expected to enhance posture, movement and body awareness such as strain, counter strain and proprioceptive neuromuscular facilitation. Energy work uses massage that is hypothesised to balance and “assist the flow of energy in the body” with light touch or positioning hands just by the skin (for example: holding, rocking, traction); it may use massage traditions derived from Eastern cultures, such as acupressure, Amma, Shiatsu and Tuina (Sherman 2006). These goals can be addressed by a variety of different styles and each style is characterised by specific techniques to achieve the goal (Sherman 2006). See Appendix 1 for a list of massage styles and techniques commonly accepted in massage research.

To place our results in a larger context, our original protocol considered factors like plausibility and dose response. Supporting evidence considering the dose of manual therapy care in different conditions is emerging (Haas 2018; Sherman 2014). The dose of massage may have an important influence on the magnitude of effect and may include session duration, frequency per week, number of sessions and number of weeks (Sherman 2014). Evidence for single sessions of care is being challenged (Gross 2018). Based on these data, we defined the dose as high, medium or low as follows:

Table 4. Dose range definitions.

1. Sources of risk of bias.

Bias domain	Source of bias	Possible answers
Selection	(1) Was the method of randomisation 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 dropout rate described and acceptable?	Yes/no/unsure
Attrition	(7) Were all randomised participants analysed 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 co‐interventions 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

Dose	Session duration	Number of sessions	Number of weeks
High	> 30 minutes	≥ 8 sessions	4 weeks
Medium	≤ 30 minutes	< 8 sessions	< 4 weeks
Low	—	1 session	—

Placing massage in the context of a placebo control is complex and poses a methodological challenge (Giandomenico 2022). No guidelines exist on how to conduct appropriate sham therapy to ensure the robustness of the methodology and results of manual therapy randomised controlled trials (RCTs). Equality assumptions in manual therapy trials are detailed by Giandomenico and colleagues and should include:

• contextual factors – process of care, positive framing may include reputation, confidence and therapeutic alliance;

• patient factors – patient is naive to the procedure, credibility of the provided treatment, patients’ previous experiences and psychological features, and reimbursement of the patient; and

• practitioner factors − type, experience, training, gender, age and number of practitioners.

The success of blinding is essential to map the effect estimates in massage RCTs successfully.

The most feasible alternative is to use a wait‐list or no treatment control group (no intervention, wait‐list, massage as an adjunct where both arms of the trial receive the same co‐treatment). All of these carry the risk of overestimating the effect estimate.

How the intervention might work

Massage likely works through biomechanical, neurophysiological and psycho‐neuromuscular mechanisms. The mechanisms underpinning the clinical effectiveness of various types of massage are likely a complex interaction of factors associated with the patient, the care provider, and the context or environment in which the care occurs. Bialosky and colleagues note that traditional emphasis on solely biomechanical mechanisms of massage has a misguided focus (Bialosky 2018). They have created a comprehensive model of the possible mechanisms of manual therapy to advance research on biological mechanisms (Bialosky 2017; Bialosky 2018). The interaction of multiple mechanisms supposedly includes altering the processing of nociception to impact the pain experience (pain modulatory capacity), systemic neurophysiological responses from the peripheral and central nervous systems and psycho‐neuromuscular mechanisms. Other mechanisms of massage could be non‐specific responses, such as placebo or expectation responses, and psychological influences. Specific biomechanical mechanisms may assist in the breakdown of adhesions (Crane 2012; Moyer 2004). Theoretically, massage may increase blood and lymph flow, cause a shift from sympathetic to parasympathetic response, reduce fibrosis, reduce pain (De Domenico 2007; Kanazawa 2009; Moyer 2004; Zaleska 2020), increase oxygenation to the muscles (Munk 2012; Soares 2020), and enable a neuroimmunologic response (Schipholt 2021). Additionally, massage has been proposed to increase neural activity at the spinal cord level and also at the subcortical nuclei, which in turn can affect mood and pain perception (Sagar 2007). As such, it is inferred that massage could potentially reduce anxiety and depression as noted by biological markers of an increase of serotonin and endorphins, although this needs more research (Heidari 2021; Moyer 2004).

Why it is important to do this review

For neck pain, a readership survey of over 45,000 people (Consumer Reports 2011), and Cross‐sectional European Social Surveys of 4950 older adults with musculoskeletal pain (Morrissey 2022), showed that massage (deep tissue or Swedish) or manual body‐based therapies were highly rated by their users, receiving scores above certain medications (prescription and over‐the‐counter) or other complementary and alternative medicine treatments (osteopathy, homoeopathy, herbal use) in the proportion of users who found them helpful. In the former survey, more than half of the users rated massage as helping 'a lot' and in the latter it was often combined with, or replaced, more traditional pain management strategies (e.g. physiotherapy and pharmacology) amongst older adults. These survey data, although anecdotal, show that many people utilise massage for neck pain and, by self‐report, find it helpful.

Our previous review reported a very low to low quality of evidence supporting the immediate and short‐term benefits of massage therapy for neck pain and tenderness compared to no treatment or placebo (Patel 2012). Multiple systematic reviews have presented similar recommendations for the benefits of massage for immediate post‐treatment reduction of neck pain (Brosseau 2012; Cheng 2014; Furlan 2012; Kong 2013; Wong 2016). Most adverse events were mild, transient and affected a small percentage of patients. However, numerous studies evaluate the efficacy of only a single massage therapy session. Wong and colleagues suggested that there should be caution in including results from exploratory studies (efficacy trials) in clinical recommendations and concluded that more robust evaluation studies (comparative effectiveness trials) would lead to greater clarity about clinical benefit (Wong 2016). They also identified the need to contextualise the dose and duration of outcomes to reflect clinical practice. Thus, the effectiveness of massage therapy remains unclear. The Royal Dutch Society for Physical Therapy issued a clinical practice guideline in 2018 noting that clinicians may consider the use of massage for patients with grade I, II and III neck pain when the primary advised treatments are not sufficiently effective (Bier 2018). However, a further review on a specific type of massage treatment (trigger point manual therapy) for chronic musculoskeletal pain (including neck pain) noted that the evidence is not sufficient to warrant this treatment (Denneny 2019). In summary, there is still very limited information on this topic, new trials have been published and new methods have been used, indicating that updates of this review remain necessary. This review was conducted according to the guidelines recommended by the Cochrane Musculoskeletal Editorial Board (Ghogomu 2014).

Objectives

To assess the benefits and harms of massage compared to placebo or sham, no treatment or exercise as an adjuvant to the same co‐intervention for acute to chronic persisting neck pain in adults with or without radiculopathy, including whiplash‐associated disorders and cervicogenic headache.

Our secondary objective was to assess the effect of massage dose.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs). Parallel, cross‐over and cluster‐RCT designs were included. We excluded quasi‐RCTs and non‐RCTs.

We included studies reported as full text, those published as abstract only and unpublished data. There were no language restrictions.

Types of participants

We included adults (≥ 18 years old) with a diagnosis of acute (< 30 days), subacute (30 to 90 days) or chronic persisting (> 90 days) neck pain classified as:

1. neck pain, including non‐specific, mechanical or simple neck pain of unidentified aetiology (Spitzer 1987; Tsakitzidis 2009); neck pain associated with myofascial pain syndrome; neck pain associated with degenerative changes (Klippel 2008); and neck pain associated with aetiology related to trauma or injury such as whiplash‐associated disorders I or II and work‐related injury (Spitzer 1987; Spitzer 1995; State Insurance Regulatory Authority 2014);

2. neck pain with radiculopathy (Rubinstein 2007; Spitzer 1987; State Insurance Regulatory Authority 2014); this category includes whiplash‐associated disorders III; and

3. neck pain associated with cervicogenic headache (Arnold 2018).

For mixed population studies, we a) contacted authors for data on the eligible participants or b) if the trial had 80% eligible participants, we included the study as if all the participants were eligible.

We excluded participants with the following co‐morbidities/characteristics.

• Definite or possible long tract signs (i.e. myelopathies).

• Neck pain caused by other pathological entities (i.e. rheumatoid arthritis, ankylosing spondylitis, spasmodic torticollis, fractures and dislocations; Klippel 2008), including grade IV neck pain ‐ pain with signs or symptoms of disease such as infection, structural pathology, tumours, spinal cord disease such as myelopathy or systemic disease (Haldeman 2008).

• Headache not of cervical origin but associated with the neck.

• Co‐existing headache, when either neck pain was not dominant or the headache was not provoked by neck movements or sustained neck postures.

• 'Mixed' subtypes of headache (i.e. migraine and cervicogenic headache).

Types of interventions

We included trials comparing massage of any type (Appendix 1) with:

1. placebo or sham (defined in any way by study authors);

2. no treatment (Davey 2011):

   • no intervention or wait‐list control;

   • massage adjuvant to the same co‐treatment. Both intervention arms received the same co‐treatment; this co‐treatment could either be an active (including exercise) or passive (i.e. manual therapy or modality) treatment. In this latter group, the massage and control groups were offered, or received, the same co‐treatment, allowing the effect of the massage treatment to be isolated.

We included the following co‐interventions that were not part of the randomised treatment: concomitant medications (i.e. analgesics) and general practitioner care. Interventions and comparators were eligible regardless of the delivery of the massage technique, dose, duration or intensity; these treatment parameters are detailed in the Characteristics of included studies table.

We did not compare massage against 'usual care' in this review.

Studies were also excluded if they used:

• multimodal treatment approaches that included massage and the unique contribution of massage could not be determined;

• instrumentation for massage application;

• non‐touch massage approaches such as 'energy work', i.e. Reiki, Polarity, Therapeutic Touch;

• massage versus another standard non‐surgical treatment using either an active (including exercise) or passive (i.e. manual therapy, passive stretch, acupuncture) method;

• one type of massage versus another; or

• one dose of massage versus another dose (i.e. three weeks with nine sessions of Chinese massage versus three weeks with three sessions of Chinese massage).

Types of outcome measures

We had six outcomes of interest: pain, function‐disability, health‐related quality of life, participant‐reported treatment success, patient satisfaction and adverse events. We did not use the outcomes listed as major and minor outcomes as inclusion criteria for the review. Relevant trials that measured these outcomes but did not report the data or did not report the data in a usable format were included in the review as part of the narrative.

We did consider what should be in a core outcome set (Turk 2004). There were no core outcome sets specific to neck pain identified through the Initiative for Measurement and Methods in Pain Clinical Trials (IMMPACT‐Turk 2008) or Core Outcome Measures in Effectiveness Trials (COMET‐Kirkham 2017). There was one protocol noted (Maujean 2018) and one recently published set for patient self‐reported disability in those with whiplash‐associated disorders (Griffin 2021).

Major outcomes

The major outcomes of interest were pain intensity and function‐disability, with the related minimally important between‐group differences used for interpreting results. We used a similar approach to the American College of Physicians 2017 guidelines for low back pain (Chou 2017), since this is not well established for neck pain.

1. Pain intensity (i.e. visual analogue scale (VAS), numeric pain rating scale (NRS), pain intensity Likert scale) or pain event (the proportion of participants not reporting at least 50% or substantial pain reduction; count; dichotomous yes or no).

2. Function‐disability could be measured using:

   1. disease‐specific patient‐reported outcomes;

   2. patient‐specific patient‐reported outcomes;

   3. observer‐based physical performance tests;

   4. other workplace metrics (i.e. return to work, sick leave); or

   5. disability episode (Beattie 2001; Finch 2002). Each disability concept was analysed separately as listed when available and not considered as one construct. Function and disability are interchangable terms in this review, representing one category of measures. Patient‐reported outcome measures would include disease‐specific disability that measures a person's ability to carry out usual activities (i.e. Neck Disability Index (NDI), Northwick Park Neck Pain Questionnaire, Neck Bournemouth Questionnaire, Neck Pain‐10, Chronic Pain Questionnaire (pain‐related disability), Activities of Daily Living) or patient‐specific disability that is used to quantify activity limitation and function specific to a patient (i.e. Patient‐Specific Function Scale). Measures of physical performance had to test the participant's ability to execute a simple activity in a standardised environment using a standardised test and scoring procedure. They are concerned with the testing of a co‐ordinated set of functions. These formed a component of functional purposeful activity (i.e. reaching, walking, driving). We defined a minimally important between‐group difference to be five points on a 0 to 100 scale.

Minor outcomes

1. Participant‐reported treatment success: based on one systematic review on global rating of change for neck pain (GROC, Bobos 2019), the outcome may include: Patient's Impression of Global Improvement (PIG‐I) 1 to 7; Global Assessment ‐1 to +10 (dichotomised expressed as events).

2. Health‐related quality of life: i.e. Short Form‐12 (SF‐12) physical (0 optimal health to 100 poor health), Short Form‐36 (SF‐36) physical (0 to 100), European Quality of Life Five Dimensions (EQ‐5D) (0 death to 1 optimal health).

3. Health‐related patient satisfaction may be judged on both process (e.g. how confident the massage practitioner is viewed to be in performing skills) and outcome (e.g. perceived improvement in health status); i.e. Satisfaction and Recovery Index 0 to 100% satisfied (Modarresi 2021); Patient Satisfaction 1 to 7 (1 fully satisfied, 4 is neutral, 7 is completely dissatisfied); numeric rating scale 0 to 10 (0 not at all satisfied, 5 moderately satisfied, 10 completely satisfied).

4. Adverse events (serious or other). If adverse events were not mentioned in the protocol or if the trial did not mention them at all, it was assumed that adverse events were not measured.

Timing of outcome assessments

We considered the following time points for the extraction of outcomes:

• short‐term from immediate post‐treatment to closest to 12 weeks; and

• long‐term from closest to 24 weeks up to one year.

The primary time point that we included in the summary of findings tables was short‐term; however, long‐term data were considered when available.

Search methods for identification of studies

A research librarian searched the following computerised bibliographic databases of the medical, chiropractic and allied health literature from their inception in past reviews and this update from February 2012 to October 2023. There were no restrictions imposed on language of publication. Anatomical terms, disorder or syndrome terms, treatment terms and methodological terms were included in subject headings (MeSH) and key text words.

Electronic searches

We searched the following databases:

• Cochrane Central Register of Controlled Trials in the Cochrane Library (1 October 2023);

• MEDLINE Ovid (1946 to 1 October 2023);

• Embase Ovid (1974 to 1 October 2023);

• CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature; 1982 to 1 October 2023); and

• Index to Chiropractic Literature (1 October 2023).

We used an RCT filter (Lefebvre 2022). We have included the complete search strategies implemented for the primary databases in Appendix 2 with the limits and filters used.

We conducted a search of the clinical trials register ClinicalTrials.gov (searched 15 October 2022), and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP; searched 15 October 2022). Additionally, we carried out a monthly screen of PubMed, ClinicalTrials.gov and the WHO ICTRP search portal up to October 2022.

Searching other resources

We checked the reference lists of all primary studies and review articles for additional references.

Key conference proceedings were searched (AG) for relevant grey literature (i.e. World Federation of Physical Therapists (WCPT 2015; WCPT 2019), World Physiotherapy Congress 2021 (WC 2021), International Federation of Orthopaedic and Manipulative Therapists (IFOMPT 2012;IFOMPT 2016)).

We searched for errata or retractions from included studies published in full text on PubMed (pubmed/ncbi.nlm.nih.gov) on 5 October 2022.

We also screened references and searched our personal files to identify studies.

Data collection and analysis

We used pre‐piloted forms for all phases of data extraction. We noted key data in the Characteristics of included studies table. We listed the reason for excluding studies by broad categories: ineligible clinical condition, inappropriate study design and irrelevant intervention. We mapped included or excluded trials to one primary reference if multiple records referred to one clinical trial.

Selection of studies

Two review authors (NC, NG, TH, AG) each with one or more areas of expertise from medicine, physiotherapy, chiropractic, massage therapy, statistics or clinical epidemiology, independently screened titles and abstracts of all the potentially relevant studies we identified as a result of the search for inclusion and coded them as 'retrieve' (eligible or potentially eligible or unclear) or 'do not retrieve'. We retrieved the full‐text study reports or publications and two review authors (NC, NG, TH, AG) independently screened the full text and identified studies for inclusion, and identified and recorded reasons for exclusion of the ineligible studies. We resolved any disagreement through discussion or, if required, we consulted a third person (CHG).

We identified and excluded duplicates and collated multiple reports of the same study under a single reference ID so that each study, rather than each report, was the unit of interest in the review. References of retrieved articles were independently screened by two review authors. We recorded the selection process in sufficient detail to complete a Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) flow diagram (PRISMA Statement 2020) and Characteristics of excluded studies table. We managed the search process using Covidence. We assessed the level of agreement for study selection using the quadratic weighted Kappa statistic (Kw); Cicchetti weights (Cicchetti 1976).

Data extraction and management

We used a data collection form for study characteristics and outcome data, which had been piloted on at least five studies in the review. Two review authors (JE, CD, HL, AM) independently extracted study characteristics from the included studies. We extracted the following study characteristics.

1. Methods: study design with design characteristics, number analysed or randomised, intention‐to‐treat analysis, power analysis, funding source for trial and declaration of interest of trial authors.

2. Participants: disorder (diagnostic classification), radicular sign and symptoms (present or absent), sex (% females), age (mean; range or standard deviation (SD)), severity (of condition), duration of complaint, setting and country.

3. Interventions and comparison: intervention and comparison description by author; activity (context: description of the pre‐treatment phase, detailed description of the sham therapy protocol, description of the post‐treatment phase); patient factors (patient‐related equality assumptions such as patients’ expectations, questionnaire or interview assessing the success of blinding, credibility questionnaire or interview, patients’ previous experiences with the given therapy); success of blinding measured; provider factors (type, years of practitioners’ experience, pre‐trial training for practitioners); mode; dose (frequency: number sessions per week over number of weeks and duration of a session); monitoring (adherence measures); route (body region); treatment schedule (number of total sessions); duration of follow‐up (time points assessed); co‐intervention (inclusion or exclusion of concomitant medication or treatment); side effects and cost of care. The description of the interventions followed the TIDieR guidance (Hoffman 2014). We defined high‐, medium‐ and low‐dose massage as per Box 1 (Description of the intervention).

4. Outcomes: clinical outcome measure, scale, baseline data (mean, SD or risks), study reported results, major or minor outcomes with time points.

5. Characteristics of the design as outlined in the Assessment of risk of bias in included studies section.

6. Notes: additional contact of trial authors and details; errata.

7. Information needed to assess GRADE (e.g. baseline risk in the control group for key outcomes).

Two review authors (JE, CD, HL, AM) independently extracted outcome data from included studies using pre‐piloted standardised forms. We extracted the number of events and number of participants per treatment group for dichotomous outcomes, and means and standard deviations and number of participants per treatment group for continuous outcomes. We noted in the Characteristics of included studies table if outcome data were not reported in a useable way and gave the authors' reported results. If data were transformed or estimated from a graph, we noted this in both the analysis footnote and text. We resolved disagreements by consensus or by involving a third person (CHG). One review author (AG, HL) transferred data into RevMan Web (RevMan Web 2020). We double‐checked that data were entered correctly by comparing the data presented in the systematic review with the study reports.

We used a Digitiser software to extract data when presented in graphic form (www.digitizeit.de/). These data were extracted in duplicate. For studies with a dose gradient, we extracted the data for each dose level separately and noted these data in a dose gradient figure for primary outcomes. For data that were missing, we wrote to the study authors to request additional or clarify data and recorded this correspondence within the notes section. When detected, we also noted errata or retraction statements. When data could not be retrieved from the author, we reported the author's report of significance in tabular form (see Characteristics of included studies). All results reported were based on the sample size analysed using the 'intention‐to‐treat' principle; in other words, the sample entering the study as these data were consistently available. We used the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017) and MECIR Standards to guide this process.

Assessment of risk of bias in included studies

At least two review authors (CHG, GG, MF, HL) from a calibrated team of assessors independently critically appraised the risk of bias for each study. This team consisted of an expert panel including a senior statistician, methodologists and interdisciplinary members. We resolved any disagreements by team debate and discussion or by involving another author (SL). We used the quadratic weighted Kappa statistic to assess agreement on risk of bias assessment (Kw 0.21 to 1.00) to maximise inter‐rater reliability (Graham 2012). We did not exclude studies from further consideration in this review on the basis of the risk of bias assessment but did carry out sensitivity analysis based on risk of bias to inform our findings.

We assessed the risk of bias using the original Cochrane risk of bias tool and according to the following domains listed in Table 4 and defined in Table 5 (Furlan 2015; Higgins 2017). We assessed each potential source of bias as high, low or unclear risk, and provided a quote from the study report together with a justification for our judgement in the risk of bias table. We considered blinding separately for different key outcomes where necessary (e.g. self‐reported clinical outcome measures versus performance‐based outcome). Blinding for placebo‐controlled trials in manual therapy considered context‐related, patient‐related and practitioner‐related equality factors to determine the success of the placebo (Giandomenico 2022). In addition, we considered the impact of missing data by key outcomes. Where information on risk of bias related to unpublished data or we had correspondence with the trial author, we noted this in the risk of bias table. When considering treatment effects, we took into account the risk of bias for the studies that contributed to that outcome. We presented the figures generated by the risk of bias tool to provide summary assessments of the risk of bias.

2. Criteria for a judgement of low risk of bias or '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 die (for studies with 2 or more groups), drawing of balls of different colours, drawing of 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/security number, date in 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 a 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. co‐interventions, hospitalisation 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 on 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 a 'yes' is scored. (N.B. these percentages are arbitrary, not supported by literature).
7	All randomised patients are reported/analysed in the group they were allocated to by randomisation for the most important moments of effect measurement (minus missing values) irrespective of non‐compliance and co‐interventions.
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 judgement.
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 co‐interventions, 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 statement should explicitly state that the researchers have had full possession of the trial process from planning to reporting without funders with potential conflict of interest having any possibility to interfere in the process. If, for example, the statistical analyses have been done by a funder with a potential conflict of interest, usually 'unsure' is scored.

Assessment of bias in conducting the systematic review

We conducted the review according to our published protocol and the updated Cochrane MECIR standards and reported any deviations from our original protocol in the Differences between protocol and review section of this systematic review.

Measures of treatment effect

We analysed dichotomous data as risk ratios or Peto odds ratios when the outcome is a rare event (approximately less than 10%), and used 95% confidence intervals (CIs). We analysed continuous data as a mean difference (MD) or standardised mean difference (SMD), depending on whether similar scales were used to measure the same conceptual construct in an outcome, along with their respective 95% CIs. We transformed data to one metric when possible (i.e. pain intensity 0 to 100 or disability 0 to 100) and pooled the data. We entered data presented as a scale with a consistent direction of effect across studies. When different scales were used to measure the same conceptual outcome (e.g. disability), we calculated SMDs instead, with corresponding 95% CIs. We back‐translated SMDs to a typical scale (e.g. 0 to 10 for pain) by multiplying the SMD by a typical amongst‐person standard deviation (e.g. the standard deviation of the control group at baseline from the most representative trial) (Higgins 2022b). Cohen 1988 provides the following interpretation: slight or small 0.2 to ≤ 0.5; moderate 0.5 to ≤ 0.8; large > 0.8. Higgins 2017 identifies the RR of an event to be slight or small when 0.8 to ≤ 1 or 1 to ≤ 1.25; moderate 0.5 to ≤ 0.8 or 1.25 to ≤ 2.0 and large or substantial < 0.5 or > 2.0. We converted data not reported in a format that could be entered directly into a meta‐analysis to the required format using the information in Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022b); we noted this as an asterisk in the footnotes of the respective figure.

For dichotomous outcomes, we calculated the absolute percent change from the difference in the risks between the intervention and control group using GRADEpro (GRADEpro GDT) and expressed this as a percentage. The relative percent change was calculated as the risk ratio ‐ 1 and expressed as a percentage. Additionally, we calculated the number needed to treat for an additional beneficial outcome (NNTB), or the number needed to treat for an additional harmful outcome (NNTH), from the control group event rate and the relative risk using the Visual Rx NNT calculator (Cates 2008). For continuous outcomes, we calculated the absolute percent change by dividing the mean difference by the scale of the measure and expressed this as a percentage. We calculated the relative difference as the absolute benefit (mean difference) divided by the baseline mean of the control group, and expressed this as a percentage. In the Effects of interventions results section and the 'What happens' column of the summary of findings tables, we provided the absolute percent change, the relative percent change from baseline, and the NNTB or NNTH (the NNTB or NNTH was provided only when the outcome showed a clinically significant difference).

Minimally important differences

Rather than using minimally important differences (MIDs), which are not well‐defined for neck pain, we judged the magnitude of the effect based on mean between‐group differences for continuous measures, based upon a similar approach used in the American College of Physicians 2017 guidelines for low back pain (Chou 2017).

Table 5. The magnitude of effect was divided into a slight/small, moderate and large/substantial difference based on between‐group differences.

Outcome	Slight/small	Moderate	Large/substantial
Pain intensity (0 to 100 scale)	5 to 10 points	> 10 to 20 points	> 20 points
Disability‐function (0 to 100 scale)	5 to 10 points	> 10 to 20 points	> 20 points
Health‐related quality of life (0 to 100)	5 to 10 points	> 10 to 20 points	> 20 points

For participant‐reported treatment success and patient satisfaction, we defined a standardised mean difference (SMD) of 0.2 (small effect, Cohen 1988) as the minimally important difference.

Unit of analysis issues

For continuous outcomes reported as medians, we converted the interquartile range (IQR) to SD to represent calculated effect sizes, based on Wan 2014. We transformed a standard error (SE) into a SD using the RevMan calculator (Cates 2015 - UK Cochrane Centre) and marked the study in the footnote. For missing statistics such as SDs, we used the following methods: 1) we converted 95% CI or SE to SD; 2) we used Digitiser software to digitise a scanned graph or chart into (x, y)‐data, when data were presented in graphic form (www.digitizeit.de/). When studies were reverse scaled (i.e. higher values indicated better outcomes rather than lower values), we multiplied the mean in each group by ‐1 as recommended in the Cochrane Handbookfor Systematic Reviews of Interventions. To facilitate analysis, we used data imputation rules when necessary (Gross 2002; Appendix 3). When data were not 'useable', we qualitatively recorded the reported significance level and analysis variables in the Characteristics of included studies table. Where multiple trial arms were reported in a single trial, we included only the relevant arms within the main analysis and described the other arms in the text and Characteristics of included studies table. If two comparisons (e.g. massage A versus placebo and massage B versus placebo) were combined in the same meta‐analysis, we halved the control group to avoid double‐counting. For RCTs with repeated outcome measurement at different time points, we used the time point closest to 12 weeks for short‐term data and closest to one year for long‐term data. For metrics that were considered not sufficiently similar, we did not perform a meta‐analysis, but we presented the results in a narrative format.

For cross‐over trials, we used first‐period data only to avoid cross‐over effects. For trials with more than two arms, we used the treatment and control arms that fit the structure of the review (massage compared with no treatment or sham control; massage compared with active treatment; or massage as adjuvant to another standard treatment compared with standard treatment alone).

Dealing with missing data

We contacted investigators to verify key study characteristics and obtain missing numerical outcome data where possible (e.g. when a study is identified as abstract only or when data are not available for all participants). Where this was not possible, and the missing data were thought to introduce serious bias, we explored the impact of including such studies in the overall assessment of results by a sensitivity analysis. We described any assumptions and imputations to handle missing data under 'Notes' in Characteristics of included studies table and explored the effect of imputation by sensitivity analyses. In addition, to facilitate analysis we used data imputation rules when necessary, following prior decisions and statistical principles (Appendix 3).

For dichotomous outcomes, we calculated the withdrawal rate (number analysed or randomised) using the number of patients randomised in the group as the denominator. For continuous outcomes (e.g. mean change in pain score), we calculated the MD or SMD based on the number of patients analysed at that time point. If the number of patients analysed was not presented for each time point, the number of randomised patients in each group at baseline was used. Where possible, we calculated missing standard deviations from other statistics, such as standard errors, confidence intervals or P values, according to the methods recommended in Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022b). If standard deviations could not be calculated, they were imputed (e.g. from other studies in the meta‐analysis).

Assessment of heterogeneity

We assessed clinical and methodological diversity in terms of participants, interventions, outcomes and study characteristics (e.g. study design, outcome measurement tools, etc.) for the included studies to determine whether a meta‐analysis was appropriate, by observing the data from the Characteristics of included studies table. We assessed statistical heterogeneity by visual inspection of the forest plot to assess the direction and magnitude of effects and the degree of overlap between CIs. We used the I² statistic to quantify inconsistency amongst the trials in each analysis. We considered the P value from the Chi² test. The Chi² test was interpreted whereby a P value of 0.10 indicated evidence of statistical heterogeneity. If we identified substantial heterogeneity, we reported it and explored possible causes by prespecified subgroup analysis.

When there were few studies, we used caution in applying the thresholds below to interpret statistical heterogeneity. We used this approximate guide for the interpretation of I² values (Deeks 2022): 0% to 40% might 'not be important'; 30% to 60% may represent 'moderate' heterogeneity; 50% to 90% may represent 'substantial' heterogeneity; 75% to 100% represents 'considerable' heterogeneity. If we identify substantial heterogeneity, we reported it and investigated possible causes by following the recommendations in section 10.10 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022). We explored the impact of heterogeneity on the meta‐analysis using either subgroup analysis (i.e. dose, disorder classification, symptom duration) or sensitivity analysis (i.e. risk of bias ‐ selection and detection bias) prior to making the decision to perform the meta‐analysis. We performed a test for subgroup differences by dose, based on a random‐effects model as described in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Section 9.6.3.1), across all analyses.

Assessment of reporting biases

We created and examined a funnel plot to explore possible small‐study biases. In interpreting funnel plots of ≥ 10 studies, we had planned to examine the different possible reasons for funnel plot asymmetry as outlined in section 13 of the Cochrane Handbook for Systematic Reviews of Interventions and related these to the results of the review. We had planned to undertake formal statistical tests to investigate funnel plot asymmetry (Egger 1997), and follow the recommendations in section 13.3 of the Cochrane Handbook for Systematic Reviews of Interventions (Page 2022).

To assess outcome reporting bias, we checked trial protocols against published reports. For studies published after 1 July 2005, we screened the ClinicalTrials.gov and WHO ICTRP trial registers for an a priori trial protocol. We evaluated whether selective reporting of outcomes was present.

Data synthesis

Our primary comparison was massage versus placebo or sham, stratified by dose.

Our other main comparisons were grouped as follows:

1. massage versus no treatment (no intervention or wait‐list); and

2. massage versus no treatment (massage as an adjuvant to a co‐treatment versus the same co‐treatment).

Meta‐analysis of numerical data

We undertook meta‐analyses only where this was meaningful, i.e. if the participants, interventions and outcomes for the three comparisons were similar enough for pooling. The dose parameters were defined as low (single session), medium (< 8 sessions over < 4 weeks for ≤ 30 minutes) and high (≥ 8 sessions over ≥ 4 weeks for > 30 minutes) (Haas 2018; Sherman 2014). We presented outcomes stratified by dose for the three comparisons. We conducted separate analysis for short‐term (closest to 12 weeks) and long‐term (closest to 24 weeks up to 12 months) outcomes. We used a random‐effects model and performed a sensitivity analysis with a fixed‐effect model. We used the random‐effects model based on the assumption that clinical diversity is likely to exist, and that different studies are estimating different intervention effects.

Subgroup analysis and investigation of heterogeneity

We carried out the following subgroup analyses in pain and function for massage versus placebo and massage versus no treatment.

• Disorder classification (neck pain with whiplash‐associated disorders versus neck pain; radiculopathy versus neck pain; cervicogenic headache versus neck pain).

• Symptom duration (acute versus chronic; subacute versus chronic).

• Co‐treatment type for 'massage adjunct to co‐treatment' (active included exercise versus passive included other manual therapy, acupuncture or modalities such as ultrasound).

First, we acknowledge that debate exists as to the accepted classification system for neck pain (Haldeman 2008). We note that those with whiplash‐associated disorders, neck pain with radiculopathy and cervicogenic headache could be identified as a distinct subpopulation from non‐specific neck pain (Anstey 2016; Haldeman 2008). However, there are similarities in prognostic and psychological factors (i.e. negative expectations, catastrophising and cognitions indicative of passive coping style) predicting poor outcome in these groups. Since there is continued debate regarding the disorder classification, we performed subgroup analysis contrasting those with neck pain without radiculopathy versus whiplash or radiculopathy or cervicogenic headache.

Second, regarding symptom duration (acute, subacute, chronic), most patients with acute neck pain have self‐limited episodes that should resolve on their own and individuals may or may not seek medical care. Chronic neck pain can be persistent or episodic and may have no identifiable underlying serious or specific disorder.

Finally, post hoc analysis for 'massage adjuvant to co‐treatment' included massage added to an active co‐treatment including exercise versus massage added to a passive co‐treatment that may include other manual therapy, acupuncture and modalities such as ultrasound. We anticipated that a different interaction effect may be present when massage was added to different co‐treatments and that this may influence our findings.

We used the formal test for subgroup interactions in Review Manager (RevMan Web 2020) and used caution in the interpretation of subgroup analyses as advised in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022). The magnitude of the effects was compared between the subgroups by assessing the overlap of the CIs of the summary estimate. Non‐overlap of the CIs indicates statistical significance.

Sensitivity analysis

We planned to carry out the following sensitivity analyses to investigate the robustness of the treatment effect for pain and function outcomes.

• Selection and detection bias (adequate versus all studies).

• Characteristics of comparator: poorly described or questionably adequate comparison classifications for placebo or sham (adequate versus all studies).

Summary of findings and assessment of the certainty of the evidence

We followed the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions, Chapters 14 and 15 for interpreting results, and distinguished a lack of evidence of effect from a lack of effect (Schünemann 2022; Schünemann 2022a). We created summary of findings tables using the following outcomes: pain at short‐term, function‐disability at short‐term, health‐related quality of life at short‐term, participant‐reported treatment success at short‐term, patient satisfaction at short‐term and adverse events. The comparison in the first summary of findings table was massage versus placebo or sham, in the second massage versus no treatment (no intervention or wait‐list) and in the third massage adjuvant to a co‐treatment versus the same co‐treatment.

Two people (AG, HL, JE, AM) independently assessed the certainty of the evidence with disagreements resolved by discussion or involving a third review author (CHG). We used the five GRADE considerations (study limitations using overall risk of bias, consistency of effect, imprecision, indirectness and publication bias) to assess the certainty of the body of evidence as it relates to the studies that contribute data to the analyses for the prespecified outcomes, and reported the certainty of evidence as high, moderate, low or very low. We justified, documented and incorporated judgements into the reporting of results for each outcome. We used GRADEpro software to prepare the summary of finding tables (GRADEpro GDT). We used version three of the GRADEPro view to display our summary of finding tables. We justified all decisions to downgrade the certainty of evidence for each outcome using footnotes, and we made comments to aid the reader's understanding of the review where necessary.

Results

Description of studies

Results of the search

We identified 5595 references from our database searching and 77 references identified through other sources to October 2023. After the removal of duplicates, we screened 4147 references by title and abstract. Following screening, we assessed 101 references in full text for eligibility (93 studies). We included 33 studies (37 references) at this update, in addition to the five studies included in the previous Cochrane review update (Figure 1). Additionally, we identified three ongoing studies (four references) and five studies are awaiting classification. Those awaiting assessment are either controlled trials (Matthieu 2013; NCT03585283), or had data errors detected (authors contacted) (Buttagat 2021; Monteiro de Sobral 2010; Stuner 2016).

1.

PRISMA flow diagram of included studies

Included studies

We included 33 RCTs (37 references) with 2053 participants (1994 analysed/2053 randomised) in this review. We contacted seven authors to request additional data such as mean and standard deviation data for primary outcomes (Briem 2007; Iqbal 2016; Klein 2013; Matsubara 2011; Phadke 2016; Skillgate 2020; Zhang Jun 2013). We contacted one author to provide details of treatment technique (Irnich 2001).

Trials were small and sample sizes varied from 19 to 313 participants (mean 62.21, SD 59.12). Details of treatment characteristics, co‐interventions, baseline values, reported results, side effects and costs of care are shown in the Characteristics of included studies table. The intervention description and replication table helps readers to assess the descriptions, compare interventions and use the information (Table 6).

3. Intervention characteristics by comparison group using TiDier guidance.

Study	Dose and treatment schedule	Intervention	Control	Co‐intervention
Versus PLACEBO
Antolinos‐Campillo 2014	Low‐dose; 1 session 1 day	Suboccipital muscle inhibition: Activity context: the participant lying supine on a bed, therapist placed hands under participant’s head and placed fingers between the suboccipital condyles and C2. The pressure was applied upwards and towards the therapist; Setting: private practice; Patient‐factor: treatment expectation was not assessed; Provider: physical therapists with 6 years of experience in the field of manual therapy; Mode: manual application with hands; Dose:frequency: 1 session; duration: 4 minutes; Monitoring: not reported; Route: suboccipital muscles Duration of follow‐up: 1 day immediately post‐treatment	Mock lower extremity treatment: Activity context: active range of motion of the hip and knee on the opposite side was performed; Setting: private practice; Success of blinding measured: not reported; Patient‐factor: treatment expectation was not assessed; Provider: physical therapists with over 6 years of experience in the field of manual therapy; Mode: active range of motion; Dose:frequency: 1 session; duration: 4 minutes; Monitoring: not reported; Route: hip and knee Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Blikstad 2008	Low‐dose; 1 session 1 day	Myofascial band therapy: Activity context: firm thumb pressure was applied in a slow stroking motion from lateral shoulder to mastoid process along the upper trapezius muscle and through active trigger points; Setting: university; Patient‐factor: treatment expectation was not assessed; Provider: chiropractor; Mode: soft tissue massage with hands; Dose:frequency: 1 session; duration: 1 minute; Monitoring: not reported; Route: upper fibres of trapezius Duration of follow‐up: 1 day immediately post‐treatment	Sham ultrasound: Activity context: Medi Link Systems ultrasound machine was used. Patients were informed that a pulsed ultrasound was going to be applied, they were notified that they should not feel any heat or pain with the ultrasound; Setting: university; Success of blinding measured: not reported; Patient‐factor: treatment expectation were not assessed; Provider: chiropractor; Dose: frequency: 1 session; duration: 2 minutes; Monitoring: not reported; Route: upper fibres of trapezius Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Briem 2007	Low‐dose; 1 session 1 day	Inhibitive distraction: Activity context: patient in supine, therapist placed fingertips onto suboccipital musculotendinous structures just caudal to the superior nuchal line and induce a sustained force in a ventrocranial direction for exerting compressive forces and distraction to cervical and suboccipital structures. Pressure was applied slowly into perpendicular to longitudinal axis of muscles and tendons. Amount of pressure was at the patient's tolerance as reflected by muscle response; Setting: private physiotherapy practice; Patient‐factor: treatment expectation was not assessed; Provider: a physical therapist, 10 years of clinical experience with 6 years experience of technique; Mode: manual approach; Dose:frequency: 1 session;duration: 3 to 3.5 minutes; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Placebo ‐hands‐on positioning: Activitycontext: patient supine and rested their head in palms of clinician for the same duration to mimic treatment position. Participants received effects of touch, warmth and rest; Setting: private physiotherapy practice; Success of blinding measured: not reported; Patient‐factor: treatment expectation were not assessed; Provider: a physical therapist, 10 years clinical experience with 6 years experience of technique; Mode: manual application; Dose:frequency: 1 session; duration: 3 to 3.5 minutes; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	No other treatment was offered to either group
Capo‐Juan 2017	Low‐dose; 1 session 1 day	Pressure release: Activity context: pressure release on myofascial on the sternoclavicular mastoid muscle was applied; Setting: general hospital rehabilitation services; Patient‐factor: treatment expectation was not assessed; Provider: physiotherapist; Mode: soft tissue massage with hands; Dose:frequency: 1 session; duration: not reported; Monitoring: not reported; Route: sternoclavicular mastoid muscle Duration of follow‐up: 1 day immediately post‐treatment	Placebo ‐ light pressure using an algometer: Activity context: application of algometric bilateral pressure on sternocleidomastoid muscle for placebo; Setting: general hospital rehabilitation services; Success of blinding measured: not reported; Patient‐factor: treatment expectation were not assessed; Provider: physiotherapist; Dose:frequency: 1 session; duration: not reported; Monitoring: not reported; Route: sternocleidomastoid muscle Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Gemmell 2008	Low‐dose; 1 session 1 day	Soft tissue mobilisation: Activity context: ischaemic compression was applied with sustained deep pressure with thumb to trigger point. Pressure was released when there was a decrease in tension in the trigger point or when the trigger point was no longer tender or when 1 minute had passed; Setting: university; Patient‐factor: treatment expectation was not assessed; Provider: chiropractor; Mode: Soft tissue massage with hands; Dose:frequency: 1 session; duration: 30 to 60 seconds; Monitoring: not reported; Route: upper fibres of trapezius Duration of follow‐up: 1 day immediately post‐treatment	Sham ultrasound: Activity context: a detuned Medi‐Link Systems ultrasound machine was used. Patients were told pulsed ultrasound was going to be used; Setting: university; Success of blinding measured: not reported; Patient‐factor: treatment expectation was not assessed; Provider: chiropractor; Mode: ultrasound; Dose:frequency: 1 session; duration: 2 minutes; Monitoring: not reported; Route: upper fibres of trapezius Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Haller 2016	High‐dose; 8 sessions 8 weeks	Craniosacral therapy: Activity context: releasing restrictions of the cranium of the spine up to the pelvis and the sacrum using standardised application of gentle fascial traction, release, unwinding techniques; Setting: university ‐ Department of Internal and Integrative Medicine, Kliniken Essen‐Mitte, University of Duisburg‐Essen, Essen, Germany; Patient‐factor: patients’ expectancy, credibility and therapeutic alliance was measured, patients of both groups received initial structural craniosacral therapy examination; Provider: 4 licenced physiotherapists with advanced craniosacral therapy qualification, and on average 6 years of clinical practice; Mode: soft tissue craniosacral therapy with hands; Dose:frequency: 1 session per week; duration: 45 minutes; Monitoring: treatment steps were recorded by therapists using a structured log; Route: cranium of the spine up to the pelvis and sacrum Duration of follow‐up: immediately post‐treatment (2 days post treatment), short‐term (12 weeks post‐treatment)	Sham ‐ light touch treatment: Activity context: light touch was applied on standardised anatomic areas, equal to those treated with craniosacral therapy, for 2 minutes each time. Body awareness instructions were given to stimulate craniosacral therapy dialogue techniques; Setting: university ‐ Department of Internal and Integrative Medicine, Kliniken Essen‐Mitte, University of Duisburg‐Essen, Essen, Germany; Success of blinding measured: treatment expectancy was assessed as part of the Credibility/Expectancy Questionnaire on a 9‐point rating scale from 1 (not at all) to 9 (very much). Treatment credibility and quality of the therapeutic alliance, measured by the Helping Alliance Questionnaire, were analysed and reported separately. Patients’ expectancy, credibility and therapeutic alliance did not appear to affect study outcomes, blinding patients to group allocation was possible, and sham intervention was tolerable and safe; Patient‐factor: Credibility/Expectancy Questionnaire and quality of the therapeutic alliance, measured by the Helping Alliance Questionnaire; patients of both groups received initial structural craniosacral therapy examination; Provider: 4 licenced physiotherapists with advanced CST qualification, and on average 6 years of clinical practice; Mode: manual application with hands; Dose:frequency: 1 session per week; duration: 45 minutes; Monitoring: treatment steps were recorded by therapists using a structured log; Route: cranium of the spine up to the pelvis and sacrum Duration of follow‐up: 1 day immediately post‐treatment (2 days post treatment), short‐term (12 weeks)	Not reported
Irnich 2001	Medium‐dose; 5 sessions 3 weeks	Soft tissue mobilisation: Activity context: massage: conventional western massage(effleurage, pétrissage, friction, tapotement); Setting: university; Patient‐factor: most patients believed in the potential benefit of the treatment; Provider: experienced physiotherapists; Mode: soft tissue massage with hands; Dose:frequency: 5 sessions over 3 weeks; duration: 30 minutes; Monitoring: not reported; Route: cervical spine and shoulder Duration of follow‐up: immediately post‐treatment (1 week post‐treatment); short‐term (3 months post‐treatment)	Sham laser: Activity context: inactivated laser pen was placed on sensitive spots; Setting: university; Success of blinding measured: patients’ beliefs about the treatment was assessed by asking 4 questions; Patient‐factor: most patients believed in the potential benefit of the treatment; Provider: experi­enced, licenced medical acupuncturists; Mode: manual application with laser pen; Dose:frequency: 5 sessions over 3 weeks; duration: 30 minutes; Monitoring: not reported; Route: acupuncture points Duration of follow‐up: 1 day immediately post‐treatment (1 week), short‐term (3 months)	Not reported
Klein 2013	Low‐dose; 1 session 1 day	Strain counter‐strain: Activity context: the affected body parts were positioned to the free direction contrary to the restriction for 90 seconds. The tender point was monitored in the position with minimal tension by a therapist. Then, a slow reposition to basic position was carried out; Setting: private practice; Patient‐factor: treatment expectation were not assessed; Provider: general physician and Osteopathic Medicine practitioner, 8 years of experience in osteopathic treatments; Mode: soft tissue stretch with hands; Dose:frequency: 1 session; duration: 90 seconds; Monitoring: not reported; Route: cervical spine and shoulder Duration of follow‐up: immediately post‐treatment (2 to 3 days post treatment)	Sham ‐ hand‐on neck movements: Activity context: the finger of the therapist was placed at the height of C4 para vertebrally on the right‐hand side of the dorsal part and the head was rotated by 30 degrees to the left to basic position and was held for 90 seconds, then, slowly back to basic position; Setting: private practice; Success of blinding measured: not reported; Patient‐factor: treatment expectation were not assessed; Provider: general physician and osteopathic medicine practitioner, 8 years of experience in osteopathic treatments; Mode: manual application; Dose:frequency: 1 session; duration: 90 seconds; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment (2 to 3 days post treatment)	A full individualised osteopathic treatment
Segura‐Orti 2016	Medium‐dose; 3 to 6 sessions 3 weeks	Strain counter‐strain: Activity context: the selected upper trapezius muscle was placed in a shortened position for 90 seconds and the patient’s cervical spine was in lateralisation ipsilateral to and rotation contralateral to the symptomatic myofascial trigger point. Once the selected myofascial trigger point no longer elicited a pain response, the position was passively maintained for 90 seconds; Setting: not reported; Patient‐factor: treatment expectation were not assessed; Provider: physiotherapist; Mode: manual application; Dose:frequency: 1 to 2 sessions per week; duration: 180 seconds; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Sham ‐ light pressure strain counter‐strain: Activity context: participants were placed in a supine position and a therapist put digital pressure below the pain threshold over the selected myofascial trigger point. Gentle, slow lateral and circular digital motions were applied over the myofascial trigger point for 90 seconds.; Setting: not reported; Success of blinding measured: not reported; Patient‐factor: treatment expectation were not assessed; Provider: physiotherapist; Mode: soft tissue stretching with hands; Dose:frequency: 1 to 2 sessions per week; duration: 180 seconds; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Stieven 2021	Low‐dose; 1 session 1 day	Myofascial release: Activity context: gentle pressure was applied to the myofascial trigger point in a lengthened position of upper trapezius. Pressure was sustained until the pain reduced by 50% then increased pressure. The upper trapezius muscle was passively stretched for 30 seconds. The combination of myofascial release plus stretching was performed 3 times, with an average duration of 90 seconds. It was performed in a sitting; Setting: university; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist with 6 years experience in myofascial release and manual therapy; Mode: manual application; Dose:frequency: 1 session; duration: 270 seconds; Monitoring: not reported; Route: upper trapezius muscle Duration of follow‐up: 1 day immediately post‐treatment	Sham ‐ dry needling: Activity context: patients were in a prone position, the therapist manipulated the tissues adjacent to the myofascial trigger point with a light touch using thumb and index finger for 10 seconds. Then therapist put a needle guide tube on the myofascial trigger point with a gentle tap. No actual needle puncture occurred; Setting: university; Success of blinding measured: not reported; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist with 6 years experience in myofascial release and manual therapy; Mode: sham dry needling tube; Dose:frequency: 1 session; duration: 10 seconds; Monitoring: not reported; Route: upper trapezius muscle Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Versus NO TREATMENT: no intervention or wait‐list
Brück 2021	Medium‐dose; 2 sessions 2 weeks	Fascial treatment: Activity context: 8 different fascia of the neck and cervical region were gently stretched, and slight pressure applied in supine. Each of the fasciae was treated only once; Setting: private practice; Patient‐factor: treatment expectations were not assessed; Provider: 2 osteopathic practitioners with ten years of experience in physiotherapy, manual therapy and osteopathy; Mode: soft tissue massage with hands; Dose:frequency: 1 session per week; duration: 15 to 20 minutes; Monitoring: not reported; Route: neck, cervical region Duration of follow‐up: 1 day immediately post‐treatment	No treatment: Activity context: untreated during study period, outcome data were taken by visiting the osteopathic practice in the same time intervals as patients in the treatment group; Setting: private practice; Mode: none; Dose: none; Monitoring: not reported; Route: none	Not reported
Cen 2003	High‐dose; 18 sessions 6 weeks	Traditional Chinese therapeutic massage: Activity context: traditional Chinese therapeutic massage using the following 2 techniques: 1 finger mediation massage that uses tip and/or whole surface of thumb, rolling massage uses the fifth metacarpophalangeal joint and hypothenar eminence, both use swinging back and forth motion 120 times per minute; Setting: university; Patient‐factor: treatment expectations were not assessed; Provider: one acupuncturist trained in TCTM for 5 years; Mode: soft tissue massage with hands; Dose:frequency: 3 sessions per week; duration: 30 minutes; Monitoring: not reported; Route: cervical spine and upper shoulder Duration of follow‐up: 1 day immediately post‐treatment	No treatment: Activity context: no treatment given during Phase I (6 weeks); Setting: university; Mode: none; Dose: none; Monitoring: not reported; Route: cervical spine and upper shoulder	Not reported
Forman 2014	Medium‐dose; 6 sessions 6 weeks	Soft tissue mobilisation: Activity context: effleurage/gliding strokes using warming massage oil applied. Therapists used the flats of the knuckles, finger pads and/or thumb for the active muscle therapy massage strokes, cervical extensor muscles were passively shortened, and then patients slowly and actively contracted the neck flexor muscles to flex their necks as the cervical paraspinals were stripped with the finger pads from distal to proximal attachments; Setting: community in the Silicon Valley region of California; Patient‐factor: treatment expectations were not assessed; Provider: therapist; Mode: soft tissue massage with hands; Dose:frequency: 1 session per week; duration: 30 minutes; Monitoring: not reported; Route: cervical spine and shoulder muscles Duration of follow‐up: 1 day immediately post‐treatment	No treatment: Activity context: no treatment, instruction or intervention was given; Mode: none; Dose: none; Monitoring: not reported; Route: none	Not reported
Jafari 2017	Medium‐dose; 4 sessions 2 weeks	Ischaemic compression: Activity context: myofascial trigger point was grasped between the physiotherapist's thumb and index fingers. Then sustained and constant pressure for ischaemic compression was applied to maximum tolerable level before reproducing a headache pattern. When the patient reported a 50% reduction of pain, the pressure was increased to maximum tolerable level again; Setting: university outpatient headache clinic; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: manual approach; Dose:frequency: 1 session per 2 days; duration: 4 minutes; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	No treatment: Activity context: no treatment was given between outcome measure sessions; Mode: none; Dose: none; Monitoring: not reported; Route: none	Not reported
Kim 2021	Medium‐dose; 8 sessions 4 weeks	Acupressure: Activity context: in supine, the therapist applied gentle pressure using the middle finger on 3 unilateral acupoints in the cervical spine. Then, the gentle pressure was applied to 9 bilaterally symmetrical acupoints with the thumb; Setting: not reported; Patient‐factor: treatment expectations were not assessed; Provider: aesthetician with experience of manipulative therapy and 2‐year acupressure training; Mode: manual application; Dose:frequency: 2 sessions per week; duration: 30 minutes; Monitoring: all selected acupoints were evaluated and validated by a traditional Korean medical doctor.; Route: neck and shoulder Duration of follow‐up: 1 day immediately post‐treatment	No treatment: Activity context: were paid 18 USD (20,000 Won) for the study period; Provider: aesthetician with experience of manipulative therapy and 2‐year acupressure training; Mode: none; Dose: none; Monitoring: not reported; Route: none	Not reported
Matsubara 2011	Low‐dose; 1 session 1 day	Acupressure to local points: Activity context: acupressure was performed to 3 local points on both sides of the neck and shoulders. The therapist used the pulp of the right thumb and treated with 20 to 25 cycles per minute in each point; Setting: university; Patient‐factor: treatment expectations were not assessed; Provider: 1 investigator, department of rehabilitation; Mode: acupuncture; Dose: 3 sets for 30 seconds (1.5 minutes) and bilaterally applied, 3 total minutes; Monitoring: not reported; Route: cervical spine and shoulder Duration of follow‐up: 1 day immediately post‐treatment	No treatment: Activity context: none; Setting: university; Mode: none; Dose: none; Monitoring: not reported; Route: none	Not reported
Pach 2018	Medium‐dose; 6 sessions 3 weeks	Tuina therapy: Activity context: meridian point manipulation, soft tissue massage at neck region, local muscle stretching, mobilisation and traction of cervical spine, manipulation of local pain (trigger points); Setting: outpatient clinic, university; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist with 2‐years of Tuina therapy training; Mode: manual application with hands; Dose:frequency: 2 sessions per week; duration: 30 minutes; Monitoring: not reported; Route: meridian points, wrist, elbow, shoulder, neck (cervical spine), trigger points Duration of follow‐up: 1 week after the end of treatment; short‐term at 9 weeks post‐treatment	No treatment: Activity context: no intervention was given during the study period; Setting: outpatient clinic, university; Mode: none; Dose: none; Monitoring: not reported; Route: none	Avoided in trial design ‐ both groups were instructed not to receive other treatments
Sherman 2014	High‐dose; 12 sessions 4 weeks	Massage: Activity context: consisting of 3 minutes of cervical range of motion assessment, 1 minute of hands‐on check‐in/tissue warming, 2 to 4 minutes of lymph drainage, 12 to 20 minutes of neck work consisting of: friction, myofascial techniques, effleurage, proprioceptive neuromuscular facilitation, muscle energy technique, active assisted stretching, 15 to 20 minutes for compensatory patterns, 6 to 10 minutes for neck work part 2, 8 to 18 minutes for full body integration and 1 minute for completion; Setting: group health integrated healthcare system, Seattle; Patient‐factor: treatment expectations were not assessed; Provider: massage therapists, 5 years experience; Mode: soft tissue massage with hands; Dose:frequency: 3 sessions per week; duration: 60 minutes; Monitoring: not reported; Route: cervical spine and upper back and chest Duration of follow‐up: 1 week post‐treatment	Wait‐list: Activity context: 4‐week period on waiting list; Mode: none; Dose: none; Monitoring: not reported; Route: none	Not reported
Versus NO TREATMENT: massage adjunct to co‐treatment vs the same co‐treatment
Abaspour 2020	Medium‐dose; 3 sessions 2 weeks	Muscle energy technique + infrared radiation: Activity context: patient’s head was put in the elastic barrier and contracted muscles, then relaxed in each direction of movements including cervical rotations, side flexions, extension and flexion. The head was taken to new barrier. The technique was applied with 20% to 30% of maximal voluntary contraction in a supine position. Each technique was repeated 3 times for both sides; Setting: neurology or physiotherapy clinics; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: manual application; Dose:frequency: 3 sessions over 2 weeks; duration: 9 minutes; Monitoring: not reported; Route: cervical muscles Duration of follow‐up: 1 day immediately post‐treatment; 2 weeks post‐treatment	Same co‐treatment ‐ infrared radiation: Activity context: infrared radiation from a single flame light (a power of 250 Watts and voltage 230 to 250 volts) was adjusted perpendicular to the exposed upper cervical region in prone; Setting: neurology or physiotherapy clinics; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: modality; Dose:frequency: 3 sessions over 2 weeks; duration: 15 minutes; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment, 2 weeks post treatment	Not reported
Afzal 2019	Medium‐dose; 9 sessions 3 weeks	Cervical manual traction + mobilisation and hot pack: Activity context: a patient's head was cradled by a physiotherapist from chin and the occiput in a supine position, and traction force applied in 25 degrees of neck flexion; Setting: Fauji Foundation Hospital; Patient‐factor: treatment expectations were not assessed; Provider: an experienced orthopaedic manual physical therapist; Mode: manual traction; Dose:frequency: 3 sessions per week; duration: 10 minutes with 10 seconds pull and 5 seconds rest; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment — mobilisation and hot pack: Activity context: the intervertebral foramen technique with movement was applied with pulling over the restricted area of the neck by performing movement into opening; hot pack was applied to the cervical spine before all other treatment techniques; Setting: Fauji Foundation Hospital; Patient‐factor: treatment expectations were not assessed; Provider: an experienced orthopaedic manual physical therapist; Mode: manual application; Dose:frequency: 3 session per week; duration: 15 minutes for hot pack; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Celenay 2016	Medium‐dose; 12 sessions 4 weeks	Connective tissue massage + exercises: Activity context: massage was applied according to the vascular response of the tissue. The application of short strokes on the lumbosacral area was applied and once the condition of subcutaneous tissues in the massaged region returned to normal, treatment progressed to the thoracic (scapular and inter‐scapular area) and then cervical (cervical‐occipital area). After 2 to 4 sessions, the next region was treated. Each treatment was given while sitting on a stool in the upright position, with 90 degrees of flexion of the hips, knees and ankles. The thighs and feet were fully supported; Setting: university physiotherapy and rehabilitation clinic; Patient‐factor: treatment expectations were not assessed; Provider: an experienced physical therapist; Mode: manual application with hands; Dose:frequency: 3 sessions per week; duration: 5 to 20 minutes per session; Monitoring: not reported; Route: lumbosacral, thoracic and cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment — exercises: Activity context: exercises were composed of postural education, cervical stabilisation and scapulothoracic exercises. Postural Education: patients in a sitting position with 2 mirrors reflecting their front and side views to find a neutral balanced position of the lumbar, thoracic and cervical spine. Cervical Stabilisation Exercises: cervical bracing technique with the activation of deep neck flexors was trained. Cervical dynamic isometric exercises were performed including functional training with elastic resistance and exercise balls. Scapulothoracic Stabilisation Exercises: The thoracic bracing technique with postural alignment and minimal multifidus muscle activation for scapulothoracic stabilisation was taught using Theraband; Setting: university physiotherapy and rehabilitation clinic; Patient‐factor: treatment expectations were not assessed; Provider: an experienced physical therapist; Mode: soft tissue massage with hands and exercise in a group; Dose:frequency: 3 sessions per week; duration: 40 to 45 minutes; Monitoring: not reported; Route: cervical spine and shoulder Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Choksi 2021	Medium‐dose; 3 sessions 3 days	Deep transverse friction massage + exercise: Activity context: frictions were applied using a therapist’s fingers slowly with a pressure at the pressure pain threshold level in the muscle relaxed position; Setting: outpatient physiotherapy department of S. S. Agrawal Institute of Physiotherapy and Medical Care Education; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: soft tissue massage with hands; Dosage: 3 minutes; Monitoring: not reported; Route: upper trapezius muscle Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment ‐ exercise: Activity context: exercises composed of scapular movements including protraction, retraction, elevation and depression, and active neck movements with flexion, extension, side flexions and both rotations; cold pack was applied in a sitting position; Setting: Outpatient physiotherapy department of S. S. Agrawal Institute of Physiotherapy and Medical Care Education; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: exercise and cold pack; Dose:frequency: 1 session per day; duration: 30 minutes; Monitoring: not reported; Route: upper back Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Gauns 2018	Medium‐dose; 6 sessions 1 week	Myofascial release + conventional care: Activity context: gross stretch in posterior cervical muscles and the upper quarter was applied. Posterior cervical muscles were stretched by a therapist’s hand placed at the occiput and the other was at the upper thoracic spine in supine. For upper quarter stretching, a patient’s arm was pulled parallel to the floor in supine; Setting: physiotherapy outpatient department of hospital; Patient‐factor: treatment expectations were not assessed; Provider: experienced physiotherapist; Mode: manual application with hands; Dose:frequency: 1 session per day; duration: 10 to 15 minutes; Monitoring: not reported; Route: arm and neck Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment ‐ conventional care: Activity context: hot moist pack and transcutaneous electrical nerve stimulation (TENS) were applied at the neck region in sitting. Home programme of stretching and strengthening exercises of cervical musculature was also given. Ergonomic advice was provided; Setting: physiotherapy outpatient department of hospital; Patient‐factor: treatment expectations were not assessed; Provider: experienced physiotherapist; Mode: heat, TENS, advice about exercise and ergonomics; Dose:frequency: 1 session per day; duration: 35 minutes; Monitoring: not reported; Route: cervical and scapular regions Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Iqbal 2016	Medium‐dose; 3 sessions 1 week	Muscle energy technique + exercise: Activity context: in supine position with cervical spine in opposite lateral flexion to the treating part so that upper trapezius muscle fibres was kept in a lengthened position. Moderate isometric contraction for 5 seconds followed by 3 seconds of relaxation while reaching to the new barrier was performed; Setting: hospital; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: manual application; Dose:frequency: 3 sessions over a week; duration: 32 seconds; Monitoring: not reported; Route: neck Duration of follow‐up: 1 day immediately post‐treatment and 1 week post‐treatment	Same co‐treatment ‐ exercise: Activity context: hot packs for 20 minutes and active stretching exercises for 100 seconds; Setting: hospital; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Dose: frequency: 3 sessions over a week; duration: 23 minutes and 10 seconds; Monitoring: not reported; Route: neck Duration of follow‐up: 1 day immediately post‐treatment and 1 week post‐treatment	Not reported
Kashyap 2018	Low‐dose; 1 session 1 day	Manual pressure release + exercise: Activity context: examiner applied gradually increasing pressure followed by sustained pressure to trigger point to patient pain tolerance; Setting: Physiotherapy Department of the Kailash Hospital; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: manual application with hands; Dose:frequency: 1 session; duration: 20 to 30 seconds; Monitoring: not reported; Route: trigger points cervical or bilateral scapular regions Duration of follow‐up: 1 day immediately post‐treatment; short‐term at 2 weeks post‐treatment	Same co‐treatment ‐ exercise: Activity context: neck and shoulder exercises and postural advice; Setting: Physiotherapy Department of the Kailash Hospital; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: exercise; Dose:frequency: 1 session; duration: not reported; Monitoring: not reported; Route: cervical and scapular regions Duration of follow‐up: 1 day immediately post‐treatment; short‐term at 2 weeks post‐treatment	Not reported
Kostopoulos 2008	Medium‐dose; 6 sessions 2 weeks	Ischaemic compression + passive stretch: Activity context: compression was applied on the trigger point using the thumb and forefinger in a pinching motion for 3 applications of 60 seconds each with a 30‐second rest period; Setting: private physiotherapy clinic; Patient‐factor: treatment expectations were not assessed; Provider: a doctoral‐level trained physical therapist, 15 years experience in myofascial pain; Mode: manual stretching and pressure; Dosage:frequency: 3 sessions per week; duration: 4 minutes; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment — passive stretch: Activity context: the targeted muscle was stretched to the end of the range of motion. A further stretch was reached with successive exhalations allowing muscles to relax; Setting: private physiotherapy clinic; Patient‐factor: treatment expectations were not assessed; Provider: a doctoral‐level trained physical therapist, 15 years experience in myofascial pain; Mode: manual stretching; Dose:frequency: 3 sessions per week; duration: 195 seconds; Monitoring: not reported; Route: cervical spine Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Nagrale 2010	Medium‐dose; 12 sessions 4 weeks	Integrated neuromuscular inhibition techniques + muscle energy: Activity context: participants were placed in supine, a practitioner moved the upper fibres of the trapezius using a pincer grasp of any trigger points. Ischaemic compression was applied: slow increasing levels of pressure were applied until the first barrier was felt. Pressure was applied again, until a new barrier was felt. Application of strain counter‐strain: modified digital pressure was applied to the identified trigger points. The pressure was maintained until there was a 70% reduction in pain; Setting: outpatient clinic; Patient‐factor: treatment expectations were not assessed; Provider: manual physical therapist; Mode: manual application with hands; Dose:frequency: 3 sessions per week; duration: 240 seconds; Monitoring: not reported; Route: cervical spine and upper fibres of trapezius Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment — muscle energy: Activity context: participant’s head and neck were placed at the upper fibres' trapezius restriction barrier. The participants then shrugged the involved/stabilised shoulder towards their ear at submaximal and held for 7 to 10 seconds. Each isometric contraction was held followed by further contralateral side‐bending, flexion and ipsilateral rotation to maintain the soft tissue stretch; Setting: outpatient clinic; Patient‐factor: treatment expectations were not assessed; Provider: manual physical therapist, Mode: soft tissue mobilisation with hand; Dose:frequency: 3 sessions per week; duration: 150 seconds; Monitoring: not reported; Route: cervical spine and upper fibres of trapezius Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Phadke 2016	Medium‐dose; 6 sessions 1 week	Muscle energy technique + passive stretch: Activity context: post isometric relaxation technique was applied to upper trapezius and levator scapulae; Setting: not reported; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: using hands; Dose:frequency: 1 session per day; duration: 100 seconds; Monitoring: not reported; Route: neck Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment ‐ passive stretch: Activity context: passive stretching was applied to upper fibres of trapezius and levator scapulae muscles; Setting: not reported; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: soft tissue stretching with hands; Dosage:frequency: 1 session per day; duration: 100 seconds; Monitoring: not reported; Route: upper fibres of trapezius and levator scapulae Duration of follow‐up: 1 day immediately post‐treatment	Conventional exercise programme, including strengthening and stretching for neck and shoulder muscles and hot packs on cervical region
Puntumetakul 2019	Medium‐dose; 6 sessions 3 weeks	Rungthip massage + thoracic manipulation: Activity context: Rungthip massage technique was performed in a side‐lying position, with 90 degrees of hip flexion and 90 degrees of knee flexion. Gently pressure using thumb was applied along the lines from the level of the inferior angle of the scapula to the lowest rib; Setting: university; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist with training and experience in spinal manipulation; Mode: manual application with hands; Dose:frequency: 2 sessions a week; duration: not reported; Monitoring: not reported; Route: level of the inferior angle of the scapula to the lowest rib Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment ‐ thoracic manipulation: Activity context: thoracic manipulation was performed at the level of T6‐T7 zygapophyseal joints in both sides; Setting: university; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist with training and experience in spinal manipulation; Mode: manual application with hands; Dose:frequency: 2 sessions a week; duration: not reported; Monitoring: not reported; Route: T6‐T7 zygapophyseal joints Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Skillgate 2020	Medium‐dose; 6 sessions 6 weeks	Massage + exercise: Activity context: effleurage, firm motion involving compression and pressure release and deep muscle/fascia massage to areas that produced concordant symptoms; Intensity ‐ delivered to tolerance without reaching pain over visual analogue scale 50/100; Setting: research clinic (Scandinavian College of Naturopathic Manual Medicine); Patient‐factor: treatment expectations were not assessed; Provider: naprapaths with experience in massage and exercise therapy; Mode: manual application; Dosage:frequency: 1 session per week; duration: 45 minutes/25 minutes + 25 minutes; Monitoring: each study participant individually monitored. The therapists underwent training sessions (3‐hour sessions twice) and regular supervision to ensure standardised provision of all the interventions in the STONE trial; Route: cervical, upper back, jaw and chest Duration of follow‐up: 12 months post‐treatment	Same co‐treatment — exercises: Activity context: Base Element ‐ strengthening, strength‐endurance, stretch‐ROM; Modular Elements: feedback‐feedforward system; Biomechanical ‐ dynamic and static stabilisation: movements were cervical flexion, extension, scapular protraction and retraction, shoulder flexion and abduction, temporomandibular opening and closing. Postural control of neck and scapulae in standing, supine and sitting; Setting: research clinic (Scandinavian College of Naturopathic Manual Medicine); Patient‐factor: treatment expectations were not assessed; Provider: naprapaths with experience in massage and exercise therapy; Mode: exercise; Dose:frequency: 1 session per week; duration: 45 minutes; Monitoring: supervised. Participants were filmed with their smartphones to ensure proper form while at home; the therapists underwent training sessions (3‐hour sessions twice) and regular supervision to ensure standardised provision of all the interventions in the STONE trial; Route: cervical, scapular, jaw Duration of follow‐up: 12 months post‐treatment	Not reported
Zhang Jishi 2013	Medium‐dose; 10 sessions 3 weeks	Ashi point massage + acupuncture: Activity context: in Ashi points in the lesion sites (neck, shoulder and back), a continuous sliding press with the thumb tip was applied. One finger Zen pushing, pressing, kneading, rolling was used to relieve the pain; Setting: outpatient clinic of the Yukuan Hospital; Patient‐factor: treatment expectations were not assessed; Provider: doctor of traditional Chinese medicine; Mode: soft tissue massage with hands; Dosage:frequency: 1 session for every other day 3 time per week for 3 weeks; duration: not reported; Monitoring: not reported; Route: neck, back, shoulder regions Duration of follow‐up: 3 months post‐treatment	Same co‐treatment ‐ acupuncture: Activity context: acupuncture was applied at points of Fengchi, Tianzhu and Fengfu in prone with moxibustion; Setting: outpatient clinic of the Yukuan Hospital; Patient‐factor: treatment expectations were not assessed; Provider: doctor of traditional Chinese medicine; Mode: acupuncture and moxibustion. Dosage:frequency: 1 session for every other day; duration: not reported; Monitoring: not reported; Route: acupuncture points, neck, back, shoulder regions Duration of follow‐up: 3 months post‐treatment	Not reported
Zhang Jun 2013	Medium‐dose; 10 sessions 3 weeks	Massage + dynamic moxibustion: Activity context: for massage, patients were seated. A bilateral neck and shoulder massage were performed using kneading and rolling methods. Pressure point areas were released. Cervical vertebrae received traction with a palm holding technique. This was followed with gentle rubbing of neck and shoulder; Setting: hospital; Patient‐factor: treatment expectations were not assessed; Provider: practitioner of the traditional Chinese medicine; Mode: soft tissue massage with hands and moxibustion with Chinese herbal medicine; Dose:frequency: 1 session for every other day, 3 sessions per week for 3 weeks; duration: 21 minutes; Monitoring: not reported; Route: neck and shoulder Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment ‐ dynamic moxibustion: Activity context: The ignited moxa was wrapped with gauze and soaked with traditional Chinese medicine and immediately pressed onto selected acupoints in the neck region. Manipulations were performed using rubbing, pressing and point‐pressing methods; Setting: Hospital; Patient‐factor: treatment expectations were not assessed; Provider: practitioner of the traditional Chinese medicine; Mode: manual application with moxibustion; Dose:frequency: 1 session for every other day; duration: 30 minutes; Monitoring: not reported; Route: neck and shoulder Duration of follow‐up: 1 day immediately post‐treatment	Not reported
Zibiri 2019	Medium‐dose; 16 sessions 8 weeks	Muscle energy technique + education/infrared radiation: Activity context: the participant’s cervical spine was positioned in a restricted range and a resistive force was applied. Isometric contract against the counter‐force by the therapist was ceased slowly, and relaxed; Setting: general outpatient department, physiotherapy department at 2 tertiary hospitals; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: manual application; Dose:frequency: 2 sessions per week; duration: 15 seconds; Monitoring: not reported; Route: neck Duration of follow‐up: 1 day immediately post‐treatment	Same co‐treatment — education/infrared radiation: Activity context: participants were advised to relax, activate and adopt stress‐coping skills, workplace ergonomics and self‐care strategies; Setting: general outpatient department, physiotherapy department at 2 tertiary hospitals; Patient‐factor: treatment expectations were not assessed; Provider: physiotherapist; Mode: education and modality; Dose:frequency: 2 sessions per week; duration: not reported; Monitoring: not reported; Route: neck Duration of follow‐up: 1 day immediately post‐treatment	Not reported

Participants were adults aged 18 to 70 (70.2% females) who suffered predominately from subacute‐chronic neck pain, with a mean pain severity of 51.8 on a visual analogue scale 0 to 100 (SD 14.1). Duration of symptoms ranged from subacute (30 days to 90 days) to chronic (longer than 90 days) as follows:

• Non‐specific neck pain:

  • Chronic (Briem 2007; Brück 2021; Cen 2003; Forman 2014; Haller 2016; Kim 2021; Pach 2018; Puntumetakul 2019; Sherman 2014; Stieven 2021; Zhang Jun 2013; Zibiri 2019; including chronic whiplash‐associated disorder ‐ Antolinos‐Campillo 2014).

  • Subacute (Gemmell 2008; Nagrale 2010; Phadke 2016).

  • Subacute‐chronic (Blikstad 2008; Celenay 2016; Choksi 2021; Irnich 2001; Kashyap 2018; Klein 2013; Kostopoulos 2008).

  • Unknown duration (Gauns 2018; Guo 2013; Iqbal 2016; Matsubara 2011; Segura‐Orti 2016).

• Neck pain with radiculopathy: unknown duration (Afzal 2019).

• Neck pain with or without radiculopathy: subacute‐chronic (Skillgate 2020).

• Neck pain with cervicogenic headache:

  • Chronic (Abaspour 2020).

  • Subacute‐chronic (Zhang Jishi 2013).

  • Unknown duration (Jafari 2017).

The trial settings were ambulatory care‐based, including university‐clinic (n = 11), hospital (n = 8), private practice (n = 11) or setting not reported (n = 3). They took place in the following regions: Europe (n = 12), Asia (n = 11), America (n = 5), the Middle East (n = 4) and Africa (n = 1).

The massage interventions varied by type and dose. Types of massage techniques (n = 15) are described in Table 7, with ischaemic compression being the most common technique.

4. Massage techniques categorised by risk of bias and dose.

Risk of bias	High‐dose (multi‐session)	Medium‐dose (multi‐session)	Low‐dose (single session)
Low	Combined techniques: General massage, myofascial release Lymphatic drainage, etc. (Sherman 2014) Craniosacral therapy (Haller 2016)	Western massage: effleurage, pétrissage, friction, tapotement (Irnich 2001) Neuromuscular inhibition, muscle energy, ischaemic compression, strain‐counterstrain (Nagrale 2010) Tunia therapy (Pach 2018)	Occipital release (Antolinos‐Campillo 2014) Myofacial band therapy (Blikstad 2008) Pressure release (Capo‐Juan 2017) Ischaemic compression (Gemmell 2008) Strain‐counterstrain (Klein 2013) Myofascial release (Stieven 2021)
High	Chinese massage: One‐finger meditation massage and rolling (Cen 2003)	Muscle energy technique (Abaspour 2020; Zibiri 2019) Manual traction (Afzal 2019) Connective tissue massage (Celenay 2016) Active muscle therapy with massage (Forman 2014) Myofascial release (Brück 2021; Gauns 2018) Acupressure/gentle pressure (Kim 2021) Ischaemic compression/stretch/dynamic soft tissue (Choksi 2021; Iqbal 2016; Jafari 2017; Kostopoulos 2008) Muscle energy, stretch (Phadke 2016) Rungthip (Puntumetakul 2019) Strain‐counterstrain (Segura‐Orti 2016) Effleurage, deep muscle fascial, compression/pressure release (Skillgate 2020) Ashi point (Zhang Jishi 2013) Kneading/rolling massage (Zhang Jun 2013)	Inhibitive distraction (Briem 2007) Manual pressure release vs muscle energy(Kashyap 2018) Local or distal acupressure (Matsubara 2011)

The length of time per session varied from one minute to 60 minutes and the total number of sessions from one to 22. The dose was categorised as follows:

• High dose: n = 3 (versus placebo ‐ Haller 2016; versus no treatment ‐ Cen 2003; Sherman 2014).

• Medium dose: n = 21 (versus placebo — Irnich 2001; Segura‐Orti 2016; versus no treatment ‐ Brück 2021; Forman 2014; Jafari 2017; Kim 2021; Pach 2018; versus adjunct to another co‐treatment ‐ Abaspour 2020; Afzal 2019; Celenay 2016; Choksi 2021; Gauns 2018; Iqbal 2016; Kostopoulos 2008; Nagrale 2010; Phadke 2016; Puntumetakul 2019; Skillgate 2020; Zhang Jishi 2013; Zhang Jun 2013; Zibiri 2019).

• Low or single dose (one session): n = 9 (versus placebo ‐ Antolinos‐Campillo 2014; Blikstad 2008; Briem 2007; Capo‐Juan 2017; Gemmell 2008; Klein 2013; Stieven 2021; versus no treatment ‐ Matsubara 2011; versus adjunct to another co‐treatment ‐ Kashyap 2018).

Our review included the following three comparators.

• Placebo or sham (n = 10, Antolinos‐Campillo 2014; Blikstad 2008; Briem 2007; Capo‐Juan 2017; Gemmell 2008; Haller 2016; Irnich 2001; Klein 2013; Segura‐Orti 2016; Stieven 2021).

• No treatment: no intervention or wait‐list control (n = 8, Brück 2021; Cen 2003; Forman 2014; Jafari 2017; Kim 2021; Matsubara 2011; Pach 2018; Sherman 2014).

• No treatment: adjunct to another co‐treatment (active co‐treatment n = 7, Celenay 2016; Choksi 2021; Kashyap 2018; Phadke 2016; Skillgate 2020; Gauns 2018; Iqbal 2016; passive co‐treatment n = 8, Abaspour 2020; Afzal 2019; Kostopoulos 2008; Nagrale 2010; Puntumetakul 2019; Zhang Jishi 2013; Zhang Jun 2013; Zibiri 2019).

Details on the control arms of the studies are available in the Characteristics of included studies table.

Adverse effects were infrequently mentioned (in eight trials (24%)) and even less frequently measured (in two trials (6%)), hampering our analysis of important harms (i.e. serious and other).

Determination of absolute risk reduction (ARR) from extractable data was rare; six trials had a minimal narrative description. The main adverse effects were transient and minor, such as immediate pain or soreness. Less common symptoms were headache, drop in blood pressure and cervical muscle fatigue. Details from each study are in the Characteristics of included studies table.

Excluded studies

In this update, we have excluded an additional 56 further studies (60 records), making a total of 189 excluded studies (207 records) from review inception. Reasons for exclusion were ineligible clinical condition (n = 39, i.e. asymptomatic latent trigger points, healthy patient, measurement used not detailed); inappropriate study design (n = 6, i.e. quasi‐RCT); irrelevant intervention (n = 107, i.e. massage was one component within a multimodal treatment; Klaber Moffett 2006); and ineligible comparison (n = 38, i.e. massage was in both the treatment and comparison group; Wang 2015). Massage was sometimes part of a control group, as in the case of Ruiz‐Molinero 2014. Studies were also excluded if an instrumental massage was applied. For example, Lauche 2012 used a massage cup to apply massage. Studies were also excluded if asymptomatic participants, including those with latent trigger points, comprised the study population, as in Oliveira‐Campelo 2013. See the Characteristics of excluded studies table for details.

Risk of bias in included studies

In Figure 2 we present each risk of bias factor across all included studies, while Figure 3 depicts the review authors' judgements about each risk of bias item for each included study. Based on selection and detection bias, we determined low risk of bias in 3% (Haller 2016), unclear risk of bias in 3% (Klein 2013) and high risk of bias in 94% (Figure 3). We used the quadratic weighted Kappa (Kw) statistic to assess agreement on a per‐question basis (Kw 0.23 to 1.00; Graham 2012). Studies identified as having high or unclear risk of bias weakened the confidence in the reported results; the overarching biases related to blinding the patient, the care provider and the outcome assessor were performance and detection bias. The 'no treatment' controlled trials were unblinded and thus at risk of performance and detection biases. The placebo‐controlled trials in this review did not assess the success of blinding in their trials.

2.

Risk of bias graph: review authors' judgements about each risk of bias domains presented as percentages across all included studies. There were no assessor‐reported outcomes identified.

3.

Risk of bias summary: review authors' judgements about each risk of bias domains for each included study. There were no assessor‐reported outcomes identified.

Allocation

Adequate randomisation was noted in 61% (n = 20) of the studies. Adequate concealment, where participants or investigators could not influence to which treatment a participant was allocated, occurred in 18% of the studies as follows:

• central allocation (n = 2) (Pach 2018; Sherman 2014); and

• sequentially numbered, opaque, sealed envelopes (n = 4) (Blikstad 2008; Gemmell 2008; Haller 2016; Klein 2013).

Unclear reporting of randomisation was a main limitation (n = 22, 67%, Abaspour 2020; Afzal 2019; Antolinos‐Campillo 2014; Brück 2021; Capo‐Juan 2017; Celenay 2016; Cen 2003; Choksi 2021; Forman 2014; Gauns 2018; Iqbal 2016; Irnich 2001; Jafari 2017; Kashyap 2018; Matsubara 2011; Nagrale 2010; Puntumetakul 2019; Segura‐Orti 2016; Stieven 2021; Zhang Jishi 2013; Zhang Jun 2013; Zibiri 2019). Another factor that influenced selection bias was concerns about baseline differences (n = 11, 33% of some concern: Abaspour 2020; Afzal 2019; Briem 2007; Choksi 2021; Gauns 2018; Iqbal 2016; Klein 2013; Kostopoulos 2008; Phadke 2016; Skillgate 2020; Zhang Jun 2013; and n = 7, 21% at high risk: Antolinos‐Campillo 2014; Celenay 2016; Forman 2014; Jafari 2017; Kashyap 2018; Segura‐Orti 2016; Zibiri 2019).

Blinding

The most apparent bias in our studies was the inability to adequately blind the patient (n = 32, 97%), therapist (n = 33, 100%) and outcome assessor (n = 32, 97%) (Figure 3). The overarching barrier is that all the studies relied on patient‐reported outcome measures such as the self‐report of pain, disability, function and satisfaction. No performance‐based outcomes were identified in the included studies to judge the primary outcomes of pain or disability‐function. Due to the nature of these studies, blinding of the therapist was almost never possible. In many cases, blinding of the patients was not possible when there were perceptible differences between treatment groups. Likewise, the blinding of the outcome assessor (who was the participant in this case) was also less possible because of the need for self‐reporting.

Incomplete outcome data

The majority of studies adequately reported the dropout rate (n = 22, 67%, Figure 3) with the use of tables or flow charts and intention‐to‐treat‐analysis (n = 15, 45%) and were judged at low risk of attrition bias.

Selective reporting

The selective reporting of outcomes (n = 25, 76%) was a very common bias with five trials demonstrating high risk of bias:

• versus placebo: Segura‐Orti 2016 did not report electromyography data; Stieven 2021 added new pain data not stated in the published protocol;

• versus no treatment: Brück 2021 retrospectively registered their trial; Forman 2014 did not report two comparison groups; and

• versus adjunct to a co‐treatment: Puntumetakul 2019 did not report neck disability index and adverse events.

To receive a low‐risk score, all the pre‐specified outcomes should be in a previously published protocol prior to the initiation of the study. Of greatest concern were outcomes of interest either added (Stieven 2021) or not reported (Abaspour 2020; Puntumetakul 2019), which may directly influence our findings.

Other potential sources of bias

Factors that influence performance bias were co‐intervention not avoided or not reported (n = 24, 73%), and compliance not assessed or not reported (n = 22, 67%). A number of the included studies possessed other sources of bias (n = 23, 70%) including reporting errors, statistical errors and confounding. Trials did not have a statistical analysis plan (i.e. Skillgate 2020), or had statistical reporting errors (Forman 2014) and significant differences at baseline (i.e. Forman 2014; Jafari 2017). Funding bias was difficult to determine due to poor reporting (n = 21, 64%). Massage trials received research funding from university or national research institutes (15%, n = 5) and no funding in 21% (n = 7) of the trials. There was no report (n = 0) indicating conflict of interest.

Effects of interventions

See: Table 1; Table 2; Table 3

Overall, 16 massage techniques of various doses were identified (Table 7).

1. Massage compared with placebo or sham

We included 10 studies assessing nine massage approaches compared with a sham procedure, including four studies on chronic neck pain (Briem 2007; Haller 2016; Irnich 2001; Stieven 2021); one study on subacute neck pain (Gemmell 2008); two studies on subacute‐chronic neck pain (Blikstad 2008; Klein 2013); one study on chronic whiplash‐associated disorders (Antolinos‐Campillo 2014) and one study on neck pain of unknown duration (Capo‐Juan 2017; Segura‐Orti 2016). See Table 1.

We identified only one trial at low risk of selection and detection bias (Haller 2016). One trial reported dichotomous data only (Blikstad 2008). We identified one study, Stieven 2021, which was an extreme outlier where the pain outcome reported was not identified in their original protocol. We therefore removed this trial from further analysis and included it under sensitivity analysis.

Pain

Massage probably results in little to no difference in pain in the short term (MD 3.43 better, 95% CI 8.16 better to 1.29 worse on 0 to 100 scale, 0 no pain; I² = 39%; 8 studies, 403 participants; low‐certainty evidence; Figure 4; Analysis 1.1, downgraded once due to indirectness and once due to risk of selection, performance and detection bias). Mean pain in the placebo group at closest to 12 weeks follow‐up was 20.55 points on a 0 to 100 scale (where a lower score indicates less pain) compared to 17.12 points in the massage group, a difference of 3.43 points (95% CI 12.6 better to 1.29 worse). This is an absolute improvement of 3% (95% CI 8% better to 1% worse) and a relative improvement of 5% (95% CI 18% better to 2% worse; using Haller 2016 baseline values). Additional dichotomous data reinforce these findings of no difference, with an absolute change of 6.9% and a relative change of 10% (RR 0.90, 95% CI 0.62 to 1.30; I² = 70%; 3 studies; Analysis 1.2).

4.

Forest plot of comparison: Massage compared with placebo for subacute‐chronic neck pain suggests little or no difference in pain reduction, outcome: 1.2 Pain (0 to 100) at short‐term follow‐up. When trials with high risk of selection and detection bias were removed, there was a small important effect favouring high‐dose massage treatment.

1.1. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 1: Pain (short‐term)

1.2. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 2: Pain (dichotomous data; short‐term)

Most trials in the placebo comparison used suboptimal massage doses (only single sessions) and may not be applicable to clinical practice. The test for subgroup differences by dose further verified this observation (Chi² = 6.82, df = 2 (P = 0.03), I² = 70.7%). Multiple trials had unclear allocation concealment, utilised a placebo that may not be credible (i.e. designed the placebo control intervention to be as similar as possible to the tested intervention) and did not test whether blinding was effective (i.e. blinding effectiveness was not routinely assessed; Giandomenico 2022; Hohenschurz‐Schmidt 2022).

Subgroup analysis by dose showed some overlap of the CIs of the summary estimate and may in part explain the noted inconsistency, while subgroup analysis by disorder classification did not influence the heterogeneity (Analysis 1.1; Analysis 1.8). We identified that Stieven 2021 seemed to be an outlier and had a protocol infraction (the pain outcome reported was not identified in their original protocol). When this trial was removed, there was a definite and clinically important difference between high‐dose versus low‐dose massage. Additional subgroup analyses based on the presence of various disorder classifications did not alter the estimates of treatment effect to a clinically important level.

1.8. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 8: Pain (short‐term): subgroup analysis by disorder classification

1. Dose (test for subgroup differences: Chi² = 6.82, df = 2 (P = 0.03), I² = 70.7%).

• High‐dose (MD 16.20 better, 95% CI 26.42 better to 5.98 better; Haller 2016) versus low‐dose – single session (MD 2.42 better, 95% CI 6.72 better to 1.88 worse; Antolinos‐Campillo 2014; Briem 2007; Capo‐Juan 2017; Gemmell 2008).

• Medium‐dose (0.52 better, 95% CI 9.14 better to 10.19 better; Irnich 2001; Segura‐Orti 2016) versus low‐dose – single session (MD 2.42 better, 95% CI 6.72 better to 1.88 worse; Antolinos‐Campillo 2014; Briem 2007; Capo‐Juan 2017; Gemmell 2008).

2. Disorder classification (test for subgroup differences: Chi² = 0.77, df = 1 (P = 0.38), I² = 0%).

• Neck pain without radiculopathy (MD 2.60 better, 95% CI 8.31 better to 3.11 worse; Briem 2007; Capo‐Juan 2017; Gemmell 2008; Haller 2016; Irnich 2001; Klein 2013; Segura‐Orti 2016) versus neck pain associated with whiplash (MD 6.50 better, 95% CI 13.10 better to 0.10 worse; Antolinos‐Campillo 2014).

• No trials were available for radiculopathy (n = 0) or cervicogenic headache (n = 0).

3. Symptom duration had insufficient data (n = 5 studies) to perform subgroup analysis.

Sensitivity analysis showed no difference in the results when we removed the identified trial that seemed to be an outlier (with Stieven 2021 included (MD 5.46, 95% CI 11.47 better to 0.56 worse; 9 studies, 432 participants; I² = 67%) versus with Stieven 2021 removed (MD 3.43 better, 95% CI 8.16 better to 1.29 worse; I² = 39%; 8 studies, 403 participants)) on a 0 to 100 scale, where a lower score indicates less pain.

Sensitivity analysis by the characteristics of the placebo or sham comparator and by risk of bias (selection and detection bias):

• Risk of selection and detection bias ‐ adequate (MD 16.20 better, 95% CI 26.42 better to 5.98 better; Haller 2016) versus all studies (MD 3.43 better, 95% CI 8.16 better to 1.29 worse).

• Poorly described or questionably adequate comparison classifications for ‘placebo or sham’ ‐ adequate (MD 16.20 better, 95% CI 26.42 better to 5.98 better; Haller 2016) versus all studies (MD 3.43 better, 95% CI 8.16 better to 1.29 worse).

Post hoc sensitivity analysis by fixed‐effect model (MD 3.75 better, 95% CI 7.25 better to 0.25 better) versus random‐effects model (MD 3.43 better, 95% CI 8.16 better to 1.29 worse) did not alter the estimates of treatment effect to a clinically important level.

A funnel plot did not detect publication bias (Figure 5).

5.

Funnel plot of comparison: 1 Massage compared with placebo for subacute‐chronic duration, outcome: 1.2 Pain (0 to 100) at short‐term follow‐up.

Function‐disability

Massage compared with sham may result in a slight increase to no difference in function‐disability in the short term (MD 9.69 better, 95% CI 17.51 better to 1.81 better on the Neck Disability Index 0 to 100, 0 no disability; I² = 0%; 2 studies, 68 participants; low‐certainty evidence, Analysis 1.3; downgraded by two levels: once due to imprecision and once due to risk of performance bias). Mean function‐disability in the placebo group at closest to 12 weeks follow‐up was 30.90 points on a 0 to 100 rating scale (where a lower score indicates better function and lower disability) compared to 21.21 points in the massage group, a difference of 9.69 points better (95% CI 17.15 better to 1.81 better). This is an absolute improvement of 10% (95% CI 18% better to 2% better) and a relative improvement of 17% (95% CI 30% better to 3% worse), using Haller 2016 baseline values. The wide confidence interval did reflect uncertainty in this estimate. One trial (contributing few patients to the meta‐analysis), but with high risk of bias across most domains, affected the interpretation of the results; when removed, the effect estimate shifted to a moderate effect (MD 10.80 better, 95% CI 19.47 better to 2.13 better). In this case, the overall limitations were not serious, and we did not downgrade the evidence for risk of selection or detection bias. However, performance bias, where the therapist is not blind to the massage intervention, was unavoidable.

1.3. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 3: Function‐disability (short‐term)

Subgroup analysis or sensitivity analysis for function‐disability were not feasible due to a paucity of studies (n = 2).

Participant‐reported treatment success

Massage compared with sham may slightly improve patient‐reported treatment success in the short term (MD 0.80 better, 95% CI 1.39 better to 0.21 better on the Patient Global Impression of Improvement 1 to 7, 1 very much improved; 1 study, 54 participants; Analysis 1.4; low‐certainty evidence, downgraded by one level due to imprecision and one level for risk of performance and detection bias). The absolute improvement was 11% better (95% CI 20% better to 3% better). This was a study that assessed high‐dose massage. Another study where only dichotomous data were available demonstrated no evidence of a difference (RR 1.00, 95% CI 0.87 to 1.14; 1 study, 61 participants; Analysis 1.5; downgraded once due to imprecision and once due to risk of performance and detection bias); the dose in this case was low‐dose or a single session.

1.4. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 4: Participant‐reported treatment success (short‐term)

1.5. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 5: Participant‐reported treatment success at short‐term (dichotomous data)

Sensitivity analysis was not feasible due to a paucity of studies (n = 1).

Health‐related quality of life

Massage compared with sham may result in little to no difference in quality of life in the short term (MD 5.30 better, 8.24 better to 2.36 better on the Short Form‐12 0 to 100, 0 worst score; 1 study, 54 participants; Analysis 1.6; low‐certainty evidence, downgraded once due to high risk of performance and detection bias and once for imprecision). This study assessed high‐dose massage.

1.6. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 6: Health‐related quality of life (short‐term)

Sensitivity analysis was not possible due to the small number of studies for this outcome (n = 1).

Patient satisfaction

No trials reported on this outcome.

Adverse events

We are uncertain whether massage results in increased adverse events. There may be no difference in minor adverse events such as slight post‐treatment soreness, sweating or low blood pressure (RR 0.99, 95% CI 0.08 to 11.55; I² = 77%; 2 studies, 175 participants; Analysis 1.7; very low‐certainty evidence, downgraded by two levels due to inconsistency and one level due to imprecision). No data were available regarding serious adverse events.

1.7. Analysis.

Comparison 1: Massage compared with placebo or sham, Outcome 7: Minor adverse events (dichotomous data; short‐term)

2. Massage compared with no treatment

We included eight studies comparing massage with no treatment (Table 2).

• Non‐specific neck pain (chronic ‐ Brück 2021; Cen 2003; Forman 2014; Kim 2021; Pach 2018; Sherman 2014; unknown duration ‐ Matsubara 2011).

• Cervicogenic headache (unknown duration ‐ Jafari 2017).

We identified no trials at low risk of selection and detection bias; two trials of low risk selection bias but high risk of detection bias (Pach 2018; Sherman 2014); and the remaining trials at unclear or high risk selection bias and high risk of detection bias (Brück 2021; Cen 2003; Forman 2014; Jafari 2017; Kim 2021; Matsubara 2011). There were no assessor‐reported outcomes used. One trial reported the significance value but no useable data (Matsubara 2011). One trial reported only dichotomous data (Sherman 2014).

Pain

Massage compared with no treatment probably results in a large reduction in pain in the short term (MD 24.24 better, 95% CI 28.48 better to 19.99 better on 0 to 100 scale, 0 no pain; I² = 0%; 6 studies, 299 participants; moderate‐certainty evidence; Analysis 2.1; downgraded due to high risk of selection, performance and detection bias). Mean pain in the no treatment group at closest to 12 weeks follow‐up was 46.83 points on a 0 to 100 scale (where a lower score indicates less pain) compared to 22.59 points in the massage group, a difference of 24.24 points (95% CI 28.48 better to 19.99 better). This is an absolute improvement of 24% (95% CI 28% better to 20% better) and a relative improvement of 41% better (95% CI 48% better to 34% better using Pach 2018 baseline values). In one additional dose response trial (Analysis 2.2), massage compared to no treatment for chronic neck pain may result in a reduction in pain at short‐term follow‐up for higher doses (3 times per week as a 60‐minute session for 4 weeks or 2 times per week at 60 minutes for 4 weeks) but not lower doses (3 times per week for 30 minutes for 4 weeks; 2 times per week for 30 minutes for 4 weeks; 1 time per week for 60 minutes for 4 weeks). This showed 51.3% fewer (61.7 fewer to 31.9 fewer) pain events, where an event included those who continue to have substantive pain. For the highest‐dose massage, pain events were reported in 26 of 35 participants (74%) in the control groups and in 9 of 39 participants (23%) in the massage groups from one trial (RR 0.31, 95% CI 0.17 to 0.57, NNTB 2; 1 study, 74 participants; Sherman 2014). The same trial also showed that weekly 60‐minute massage sessions over six weeks may result in a very small benefit or no difference at long‐term follow‐up (MD 0.61 better, 95% CI 1.17 better to 0.05 better on VAS 0 to 100, 0 no pain; 1 study with 10 arms; Sherman 2014; 173 participants with 14 to 19 per arm). This is an absolute improvement of 1% (95% CI 1% better to 0% better).

2.1. Analysis.

Comparison 2: Massage compared with no treatment, Outcome 1: Pain (short‐term)

2.2. Analysis.

Comparison 2: Massage compared with no treatment, Outcome 2: Pain (dichotomous data; short‐term): subgroup analysis by dosage

A paucity of studies (n = 6) limited our assessment and interpretation of subgroup and sensitivity analysis. Subgroup analysis by dose did not alter the estimates of treatment effect (test for subgroup differences: Chi² = 1.07, df = 1 (P = 0.30), I² = 6.3%; Analysis 2.1). Subgroup analysis by disorder characteristic and duration was not performed as all trials addressed chronic neck pain without radiculopathy.

Sensitivity analysis by selection and detection bias showed no trials with adequate or low risk of bias. Post hoc assessment showed that there were two trials with low risk of selection bias but high risk of detection bias, which did not alter the effect estimate to a clinically important level (low risk of selection bias with high risk of detection bias: MD 23.12 better, 95% CI 32.91 better to 13.32 better).

Function‐disability

Massage compared to no treatment may result in a slight improvement in function‐disability at closest to 12 weeks follow‐up (MD 9.51 better, 95% CI 13.68 better to 5.33 better on the Neck Disability Index 0 to 100, 0 no disability; I² = 56%; 6 studies, 300 participants; moderate‐certainty evidence; Analysis 2.3; downgraded for risk of selection, performance and detection bias). There were no assessor‐reported outcomes used to measure function. Mean function‐disability in the no treatment group at closest to 12 weeks follow‐up was 31.94 points on a 0 to 100 scale (where a lower score indicates better function) compared to 22.43 points in the massage group, a difference of 9.51 points (95% CI 13.68 better to 5.33 better). This is an absolute improvement of 10% (95% CI 14% better to 5% better) and a relative improvement of 20% (95% CI 29% better to 11% better, using Pach 2018 baseline values). One dose response trial, Sherman 2014, indicated that higher‐dose massage may improve function by a small amount at short‐term follow‐up (three times per week at 60 minutes for 12 sessions over four weeks; MD 9.44 better, 95% CI 14.99 better to 3.89 better on the Neck Disability Index 0 to 100, 0 no disability; 74 participants) and two times per week at 60 minutes for eight sessions over four weeks (MD 5.76 better, 95% CI 11.21 better to 0.21 better on the Neck Disability Index 0 to 100, 0 no disability; 73 participants; low‐certainty evidence, downgraded once for risk of selection, performance and detection bias and once for imprecision; this is a small pilot study that needs to be replicated). Lower‐dose massage (three times per week for 30 minutes; two times per week for 30 minutes; once a week for 60 minutes) was comparable to no treatment. At long‐term follow‐up, the reported data did not support a lasting effect of a booster‐dose session (MD 1.44 better, 95% CI 3.21 better to 0.33 worse on the Neck Disability Index 0 to 100, 0 no disability; 1 study with 10 arms, 173 participants with 14 to 19 per arm, Sherman 2014; low‐certainty evidence; downgraded once for risk of selection, performance and detection bias and once for imprecision; this is a small pilot study that needs to be replicated).

2.3. Analysis.

Comparison 2: Massage compared with no treatment, Outcome 3: Function‐disability (short‐term)

A paucity of studies inhibited a clear assessment of the noted inconsistency (I² = 56%). The inconsistency was not explained using subgroup analysis by dose, disorder classification or disorder duration. The estimates of treatment effect had broad confidence intervals.

• By dose (Chi² = 1.18, I² = 15%):

  • high (MD 15.07 better 95% CI 27.66 better to 2.48 better);

  • medium (MD 7.72 better, 95% CI 11.95 better to 3.50 better);

  • all doses (MD 9.69 better, 95% CI 17.57 better to 1.81 better);

• By disorder classification ‐ all trials were neck pain without radiculopathy.

• By disorder duration ‐ all trials were a mixed population of subacute‐chronic neck pain.

For sensitivity analysis by selection and detection bias, there were no trials with adequate or low risk of bias for both domains. Post hoc analysis showed there were two trials with low risk of selection bias but high risk of detection bias that did not alter the effect estimate to a clinically important level but did help to explain the inconsistency (low risk of selection bias with high risk detection bias: MD 9.48 better, 95% CI 12.87 better to 6.08 better; I² = 0%) versus all studies (MD 9.51 better, 95% CI 13.68 better to 5.33 better; I² = 56%).

Health‐related quality of life

Massage compared with no treatment for neck pain without radiculopathy may result in little to no difference in quality of life at closest to 12 weeks follow‐up. For chronic neck pain at short‐term follow‐up, massage may have a small beneficial effect over no treatment (MD 5.70 better, 95% CI 8.84 better to 2.56 better on Short Form‐12, 0 to 100, 0 worst quality of life; 1 study, 89 participants; Analysis 2.4; low‐certainty evidence, downgraded once for risk of detection bias and once for imprecision). This is an absolute change of 6% better (9% better to 3% worse) and a relative change of 13% better (20% better to 6% worse, using Pach 2018 baseline values).

2.4. Analysis.

Comparison 2: Massage compared with no treatment, Outcome 4: Health‐related quality of life (short‐term)

Sensitivity analysis was not possible due to limited trial numbers (n = 1).

Participant‐reported treatment success, patient satisfaction, adverse events

No data were available on participant‐reported treatment success, patient satisfaction or adverse events (minor or serious).

3. Massage as an adjuvant to the same co‐treatment

We included 15 studies examining massage added to a co‐treatment (Table 3). The disorder classification and duration included the following four categories.

• Non‐specific neck pain (subacute ‐ Nagrale 2010; Phadke 2016; Zhang Jun 2013; Zibiri 2019; subacute‐chronic ‐ Celenay 2016; Choksi 2021; Iqbal 2016; Kashyap 2018; Kostopoulos 2008; chronic ‐ Puntumetakul 2019; unknown duration ‐ Gauns 2018).

• Cervicogenic headache (chronic ‐ Abaspour 2020; Zhang Jishi 2013).

• Radiculopathy (unknown duration ‐ Afzal 2019).

• Mixed neck pain with or without radiculopathy (chronic ‐ Skillgate 2020).

The standard non‐surgical treatments used either a passive method (Abaspour 2020; Afzal 2019; Kostopoulos 2008; Nagrale 2010; Puntumetakul 2019; Zhang Jishi 2013; Zhang Jun 2013; Zibiri 2019) or an active including exercise method (Celenay 2016; Choksi 2021; Gauns 2018; Iqbal 2016; Kashyap 2018; Phadke 2016; Skillgate 2020). We identified no trials as having low risk of selection and detection bias, one trial (7%) as having low risk of selection bias but high risk of detection bias (Skillgate 2020), and 14 trials (93%) as having high risk of selection and detection bias (Abaspour 2020; Afzal 2019; Celenay 2016; Choksi 2021; Gauns 2018; Iqbal 2016; Kashyap 2018; Kostopoulos 2008; Nagrale 2010; Phadke 2016; Puntumetakul 2019; Zhang Jishi 2013; Zhang Jun 2013; Zibiri 2019). One trial reported P values (Iqbal 2016), one trial dichotomous data (Zhang Jun 2013), and one trial reported only the outcome function‐disability (Gauns 2018).

Pain

Massage as an adjuvant to a co‐treatment for neck pain may result in little to no difference in pain in the short term (MD 7.29 better, 95% CI 11.43 better to 3.16 better on a 0 to 100 scale, 0 no pain; I² = 69%; 12 studies, 795 participants; Figure 6; Analysis 3.1; low‐certainty evidence). Mean pain in the co‐treatment group at closest to 12 weeks follow‐up was 34.41 points on a 0 to 100 scale (where a lower score indicates less pain) compared to 27.12 points in the massage group, a difference of 7.29 points (95% CI 11.43 better to 3.16 better). This is an absolute improvement of 7% (95% CI 11% better to 3% better) and a relative improvement of 12% better (95% CI 19% better to 5% better, using Skillgate 2020 baseline values). This pain reduction was not maintained at long‐term follow‐up (MD 1.40 better, 95% CI 3.24 better to 6.04 worse on a VAS 0 to 100, 0 no pain; 1 study, 293 participants; low‐certainty evidence; Analysis 3.2). One additional trial using dichotomous data reinforced this finding (Analysis 3.3). We downgraded the evidence by two levels, once for inconsistency (I² = 69%) and once for high risk of selection and detection bias.

6.

Forest plot of comparison: Massage as an adjuvant to the same co‐treatment using either passive (non‐exercise) or active (exercise‐based) methods suggests evidence of little or no difference in the reduction of persisting neck pain, outcome: 2.1 Pain (0 to 100) at short‐term follow‐up.

3.1. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 1: Pain (short‐term)

3.2. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 2: Pain (long‐term)

3.3. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 3: Pain (dichotomous data; short‐term)

Subgroup analysis may in part explain the heterogeneity, however the power to detect a difference was very limited as there were fewer than 10 studies per category (Analysis 3.9; Analysis 3.10; Analysis 3.11). The effect measures of the summary estimate for two categories (disorder classification and co‐treatment type) had wide CIs, with overlap of the CIs, and thus could be considered not statistically significant. However, the formal test for subgroup difference was significant by disorder classification but not co‐treatment type.

3.9. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 9: Pain (short‐term): subgroup analysis by dose

3.10. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 10: Pain (short‐term): subgroup analysis by disorder classification

3.11. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 11: Pain (short‐term): subgroup analysis by symptom duration

1. Dose (test for subgroup differences: Chi² = 1.77, df = 1 (P = 0.18), I² = 43.4%).

• High‐dose versus low‐dose (n = 0).

• Medium‐dose (MD 7.92 better, 95% CI 12.52 better to 3.32 better) versus low‐dose (MD 2.70 better, 95% CI 8.87 better to 3.47 worse).

2. Disorder classification (test for subgroup differences: Chi² = 7.93, df = 2 (P = 0.02), I² = 74.8%).

• Neck pain with whiplash‐associated disorders versus neck pain (n = 0).

• Neck pain with radiculopathy (MD 4.00 worse, 95% CI 4.07 better to 12.07 worse) versus neck pain (MD 6.99 better, 95% CI 10.44 better to 3.53 better).

• Neck pain with cervicogenic headache (MD 19.96 better, 95% CI 40.51 better to 0.60 worse) versus neck pain (MD 6.99 better, 95% CI 10.44 better to 3.53 better).

3. Symptom duration (test for subgroup differences: Chi² = 2.29, df = 1 (P = 0.13), I² = 56.3%).

• Acute versus chronic (n = 0).

• Subacute (MD 2.47 better, 95% CI 9.30 better to 4.36 worse) versus chronic (MD 11.93 better, 95% CI 19.68 better to 4.17 better).

4. Co‐treatment type for 'massage adjunct to co‐treatment' (test for subgroup differences: Chi² = 3.54, df = 1 (P = 0.06), I² = 71.8%).

• Active (MD 3.58 better, 95% CI 6.05 better to 1.12 better) versus passive (MD 11.37 better, 95% CI 19.10 better to 3.64 better).

Sensitivity analysis by selection and detection bias showed no trials with adequate or low risk of bias for both domains. Post hoc sensitivity analysis by fixed‐effect model (MD 5.67 better, 95% CI 7.72 better to 3.63 better) versus random‐effects model (MD 7.29 better, 95% CI 11.43 better to 3.16 better) did not alter the estimates of treatment effect to a clinically important level.

Function‐disability

Massage as an adjuvant to co‐treatment may have little to no effect on function‐disability in the short term, but the evidence is very uncertain (SMD 0.46 better, 95% CI 0.92 better to 0.01 better; I² = 79%; 7 studies, 532 participants; very low‐certainty evidence; Analysis 3.4; downgraded by one level due to extreme inconsistency and one level due to high risk of selection and detection bias). This could be re‐expressed as MD 8.28 better, 95% CI 16.56 better to 0.18 better on a 0 to 100 scale, 0 no disability (Skillgate 2020 as baseline SD). The absolute change was 8% better (17% better to 0.2% worse) and the relative improvement was 20% (95% CI 39% better to 0% better). Long‐term follow‐up showed no improvement in function (MD 0.04, 95% CI ‐0.38 to 0.46; 1 study, 293 participants; moderate‐certainty evidence; Analysis 3.5; downgraded due to high risk of selection and detection bias). One additional trial reported P values (Analysis 3.6).

3.4. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 4: Function‐disability (short‐term)

3.5. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 5: Function‐disability (long‐term)

3.6. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 6: Function‐disability (other data; short‐term)

Function‐disability (other data; short‐term)
Study	Between Arm 1 vs Arm 2 post hoc Bonferroni analysis for NDI
Medium‐dose adjuvant to a passive method
Iqbal 2016	Immediate post‐treatment P = 0.005
	Short‐term P = 0.001

Subgroup analysis may in part explain the heterogeneity, however the power to detect a difference was very limited as there were fewer than 10 studies per category (Analysis 3.12; Analysis 3.13; Analysis 3.14; Analysis 3.4). Heterogeneity may in part be explained by the disorder classification as well as symptom duration; the non‐overlap of the CIs indicates statistical significance. However, the power to detect a difference was low and thus these findings may be spurious.

3.12. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 12: Function‐disability (short‐term): subgroup analysis by dose

3.13. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 13: Function‐disability (short‐term): subgroup analysis by disorder classification

3.14. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 14: Function‐disability (short‐term): subgroup analysis by disorder duration

1. Dose (test for subgroup differences: Chi² = 1.47, df = 1 (P = 0.23), I² = 31.9%).

• High‐dose versus low‐dose (n = 0).

• Medium‐dose (MD 0.53 better, 95% CI 1.05 better to 0.02 better) versus low‐dose (MD 0.01 worse, 95% CI 0.70 better to 0.73 worse).

2. Disorder classification (test for subgroup differences: Chi² = 5.00, df = 1 (P = 0.03), I² = 80.0%).

• Neck pain with whiplash‐associated disorders versus neck pain (n = 0).

• Neck pain with radiculopathy (MD 0.29 worse, 95% CI 0.48 better to 1.06 worse) versus neck pain (MD 0.73 better, 95% CI 1.18 better to 0.28 better).

• Neck pain with cervicogenic headache versus neck pain (n = 0).

3. Symptom duration (test for subgroup differences: Chi² = 12.11, df = 1 (P = 0.0005), I² = 91.7%).

• Acute versus chronic (n = 0).

• Subacute (MD 0.91 better, 95% CI 1.35 better to 0.47 better) versus chronic (MD 0.03 better, 95% CI 0.25 better to 0.19 worse).

4. Co‐treatment type for 'massage adjunct to co‐treatment' (test for subgroup differences: Chi² = 0.00, df = 1 (P = 0.98), I² = 0%).

• Active (MD 0.46 better, 95% CI 1.05 better to 0.13 worse) versus passive (MD 0.45 better, 95% CI 1.32 better to 0.43 worse).

Sensitivity analysis by selection and detection bias showed no trials with adequate or low risk of bias for both domains.

Health‐related quality of life

Massage as an adjuvant to a co‐treatment may not reduce quality of life in the short term (SMD 0.10 better, 95% CI 0.52 better to 0.31 worse; I² = 60%; 2 studies, 352 participants; Analysis 3.7; low‐certainty evidence; downgraded due to inconsistency and high risk of selection and detection bias). This could be re‐expressed as MD 0.14 better, 95% CI 0.73 better to 0.43 worse on a 0 to 100 scale where 0 is a better score (Skillgate 2020). The absolute change was 0% better (1% better to 0% worse) and the relative change was 0.16% better (0.008% better to 0.005% worse). There was also no long‐term benefit (MD 0.80 worse, 95% CI 1.12 better to 2.72 worse on EQ‐5D 0 to 1, 0 better score; 1 study, 286 participants; Analysis 3.8; low‐certainty evidence; downgraded by one level due to high risk of selection and detection bias and one level for inconsistency). Note that as the EQ‐5D was a reversed scale, the mean in each group was multiplied by ‐1.

3.7. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 7: Health‐related quality of life (short‐term)

3.8. Analysis.

Comparison 3: Massage as adjuvant to a co‐treatment, Outcome 8: Health‐related quality of life (long‐term)

Sensitivity analysis for health‐related quality of life was not feasible due to a paucity of studies (n = 2).

Participant‐reported treatment success, patient satisfaction, adverse events

No data were available on participant‐reported treatment success, patient satisfaction or adverse events (minor, serious).

Discussion

Summary of main results

Thirty‐three studies with 2053 participants assessed the influence of massage on neck pain, neck‐related function‐disability, participant‐related treatment success, health‐related quality of life, patient satisfaction and adverse events. The evidence for the contribution of massage in managing neck pain was predominately of low certainty. Adverse events were not assessed in most trials.

When massage was compared to a placebo there was low‐certainty evidence of little or no difference in pain, function‐disability, quality of life or participant‐reported treatment success for subacute and chronic neck pain in the short term. Selection and detection bias were evident as multiple trials had unclear allocation concealment, utilised a placebo that may not be credible and did not test whether blinding was effective.

Precise information about serious harms, side effects and complications was missing. Based on two trials, there was no difference in short‐term mild adverse events (i.e. treatment soreness) compared to a placebo.

Overall completeness and applicability of evidence

Important gaps exist in terms of clinical applicability.

There were 15 massage techniques described in our review (Table 7). Monitoring of the consistency of the massage intervention was not conducted in any of the 33 studies. An ideal technique, dose and provider has not been determined in the literature to date. For massage versus placebo, subgroup analysis by dose identified single‐session massage to be distinctly different from high‐dose massage. In our comparison, for the outcome pain 58% of the trials assessed a single session of massage, two trials assessed medium‐dose and one trial assessed high‐dose massage. When compared to no treatment, the trials assessed medium (one to two sessions per week for two to six weeks of 20 to 30 minutes duration) or high‐dose (one session per week for eight weeks of 45 minutes duration or three sessions per week for four weeks of 60 minutes duration) massage protocols, but the test for subgroup difference by dose was not significant. One small dose response trial suggested that a higher dose is better (Sherman 2014). Given these diverse data, there is a need for a large clinical trial to validate and establish which dose range is most effective. Few trials reported the therapist's credentials and experience level. Finally, complete reporting of the intervention characteristics was not done, making clinical replication difficult. The TIDieR (Template for Intervention Description and Replication) checklist should be used in the design phase of massage trials and include the technique, why the intervention was performed, by whom and where, the technique parameters, such as the body region or segmental level, frequency, direction, intensity and dose (Hoffman 2014).

Second, few trials reported outcomes such as function‐disability (n = 17), quality of life (n = 5), global perceived effects (n = 2) and patient satisfaction (n = 2). For back pain, standardised outcome reporting in clinical trials of patients with non‐specific low back pain has been recommended by an international multidisciplinary panel, to include physical functioning, pain intensity and health‐related quality of life as core primary outcome domains (Chiarotto 2018). Current practice guidelines for neck pain recommend validated self‐report questionnaires for pain and function‐disability as well as psychosocial functioning (Blanpied 2017). Although pain was the primary outcome reported, most trials did not report the associated minimally important difference of interest to them. This makes it impossible to determine if a meaningful benefit has been achieved, even if a statistical benefit has been noted. Furthermore, no trials reported sick leave, return to work or economic evaluations such as health care utilisation or cost of care.

Third, adverse events were adequately measured in only 6% (n = 2/33) of the trials, hampering our assessment of important harms both serious and other. Clinical application of benefits with harm trade‐offs is standard practice. Although massage is generally presumed to be safe, there was a wide spectrum of techniques seen in this review. One author found that the main adverse events and complications of massage for varied musculoskeletal disorders were disc herniation, soft tissue trauma, neurologic compromise, spinal cord injury, dissection of the vertebral arteries and others (Yin 2014). From our search on harms and massage therapy, we noted two recent case reports supporting Yin's data. Both case reports were in young adults and indicated dissection of a cervical artery following massage in the cervical region (Birkett 2020; Zuhorn 2020). Although causality may not be established, serious, rare and potentially devastating side effects (i.e. stroke) may be associated with the application of massage to the cervical region. It is therefore important that each trial assess and report the occurrence of adverse events to gain insight into the spectrum of harms by technique.

Fourth, compliance or adherence to therapy protocol reporting was almost universally lacking in the included trials, leaving readers to wonder if there was a lack of adherence and, if so, whether it was related to adverse events or other factors.

Finally, we explored symptom duration and disorder subtype using subgroup analysis. For those trials comparing massage against a sham or no treatment, most of the evidence examined chronic neck pain and subtype neck pain without radiculopathy. There were no trials evaluating acute neck pain. There were a few trials for each subtype ‐ cervicogenic headache (n = 3), whiplash‐associated disorder (n = 1) and neck pain with radiculopathy (n = 2). Our sensitivity analysis did not find that disorder characteristics influenced our results, but the low number of trials in each category limits the inferences that can be made from these data. We acknowledge the biological difference resulting from a traumatic mechanism of injury or neural impingement (neck pain with radiculopathy) albeit the root source of pain remains non‐specific. We also acknowledge a varied recovery trajectory based on prognostic indicators for whiplash‐associated disorder (Elliott 2016; Sterling 2014) and a prolonged (four to six months) trajectory for neck disorder with radiculopathy (Wong 2014). In our future updates, we will continue to consider subgroup analysis for aetiologic and prognostic classifications.

Quality of the evidence

The certainty of the evidence, based on the GRADE approach, varied from very low to moderate; the overarching concerns were:

• biases related to allocation concealment;

• biases related to blinding the patient, the care provider and outcome assessor (in other words, performance and detection bias);

• biases related to a placebo of adequate quality (in other words, control for equality assumptions ‐ contextual, patient‐related and practitioner‐related factors ‐ for hands‐on care (Giandomenico 2022)); and

• variation in the dose of trials by the characteristics of the comparator. A sham control was mostly compared to single‐dose massage (67%); no treatment or wait‐list was compared to high‐dose or medium‐dose massage (100%); and massage added to the same co‐treatment was compared against medium‐dose massage (92%).

Massage versus placebo or sham

We downgraded the evidence for pain once due to indirectness and once due to risk of selection and detection bias. When we explored subgroup analysis by dose, there was some overlap of the CIs of the summary estimates, the test for subgroup differences had an I² value of 70.7% and the result was significant when the one trial with protocol infractions was removed (Stieven 2021). Although it may not fully explain the heterogeneity, we identified six of nine studies to be single‐session placebo‐controlled trials. We believe that a single session of massage is a suboptimal dose, potentially not used in therapeutic massage practice. Equally of concern was the possible presence of selection and detection bias. Sensitivity analysis by selection and detection bias showed some, albeit small, overlap of the CI and may partially explain the heterogeneity. The use of poorly described or questionably adequate classifications for the placebo may partially explain the heterogeneity. Specific to hands‐on care, bias may be introduced by a placebo of inadequate quality. One of the elements in a placebo of adequate quality is the robustness of the 'equality assumption'. The intervention and placebo groups have to be similar in patient‐related, context‐related and practitioner‐related equality assumption items as follows: 1) patient‐related factors (the same patient‐operator relationship in the experimental and placebo groups), 2) contextual factors (dose and frequency of sessions) and 3) practitioner‐related factors (years of practitioner experience). Finally, blinding of the outcome assessor had a small overlap of CIs, which may indicate the influence of detection bias. Characteristics of the participants, such as disorder classification or symptom duration, did not help to explain the heterogeneity.

We downgraded the evidence for function‐disability, participant‐reported treatment success and health‐related quality of life for imprecision and risk of selection and detection bias, as already noted above for pain. The confidence intervals were wide, did not rule in or rule out a clinically important change and reflected uncertainly in these estimates.

We downgraded the evidence for adverse events once for inconsistency and once for imprecision. Adverse events were infrequently and inconsistently reported. Serious adverse events may not be detected in very small trials and were not even assessed.

Massage versus no treatment

The overarching findings across all participant‐reported outcomes were at risk of performance and detection bias and an overestimation of the treatment effect size.

We downgraded the evidence for pain to moderate certainty due to high risk of selection and detection bias. Allocation concealment was often unclear. The therapist performing the massage was not blind and the participant‐reported outcomes had detection bias. Subgroup analysis by dose had a paucity of studies and showed a substantive overlap of the CIs of the summary estimate. Sensitivity analysis was not feasible due to a paucity of studies. One dose‐response trial was downgraded for imprecision. This was a very small pilot study exploring dose‐response and needed to be replicated.

We downgraded the evidence for function‐disability once for risk of bias and once for inconsistency. Performance and detection bias were present because blinding of the participant and outcome assessor for participant‐reported outcomes was not possible. There were no assessor‐reported outcome measures for function identified. Subgroup analysis by dose showed the dose across most trials was consistent with either high‐ or medium‐dose and showed similar results. Sensitivity analysis was not feasible due to a paucity of studies.

We downgraded the evidence for health‐related quality of life for imprecision and risk of detection bias. The confidence intervals were wide, did not rule in or rule out a clinically important change and reflected uncertainly in this estimate.

There were no data assessing participant‐reported treatment success, patient satisfaction or adverse events.

Massage as an adjuvant to the same co‐treatment

We downgraded the evidence for pain once for high risk of selection, performance and detection bias, as well as once for inconsistency. There were greater concerns regarding selection bias (allocation concealment was unclear in 92% of the trials) and adequate randomisation (42% rated as unclear or high risk). Additionally, all trials were at risk of performance and detection bias. Subgroup analysis by dose did not help to explain the heterogeneity. Heterogeneity may be influenced by the disorder classification and the co‐treatment type. In other words, there may be lesser interaction effects between massage added to an active exercise‐based treatment and massage added to a passive treatment. Additionally, those with radiculopathy may not respond as well to the addition of massage to a standard co‐treatment. The power to detect a subgroup difference was low and was observational, thus these may be spurious findings. Sensitivity analysis by selection and detection bias indicated that there were no trials of adequate quality or low risk of bias. Simply put, there was a risk of overestimating the treatment effect.

Similarly, we downgraded the evidence for function‐disability and health‐related quality of life due to risk of bias (e.g. selection, performance and detection bias as noted above) and inconsistency. Again, there were no high‐dose massage trials, 92% of trials were of medium‐dose massage and one trial used a single low dose. Subgroup analysis by dose did not detect a difference. Subgroup analyses by disorder classification and duration were significant, but due to the small number of trials, these may be spurious findings. Sensitivity analysis was not feasible due to the paucity of trials.

Potential biases in the review process

Our approach to summarising the literature had several strengths. We conducted a comprehensive, research librarian‐directed search of multiple databases. A minimum of two people independently extracted data, which we then cross‐referenced for accuracy. We assessed the external validity of the studies using the GRADE method. Two raters collaboratively performed the GRADE assessment to ensure accuracy, followed by group consensus.

There were several limitations of the primary studies. The majority of studies were of low methodological quality or high risk of bias. One study provided insight into the dose of massage and needs replication (Sherman 2014). There was unclear information regarding which massage technique was most beneficial, due to heterogeneity in the types of techniques used. Many studies assessed only the immediate post‐treatment effect of a single application of treatment.

Agreements and disagreements with other studies or reviews

Our findings build on our previous Cochrane review update, adding 28 new studies to make a total of 33 (Patel 2012). In this update, the addition of subgroup analysis by dose advanced our understanding of treatment characteristic variables possibly impacting the outcome. Nevertheless, the contribution of massage remains uncertain given the predominance of low‐certainty evidence for most outcomes. Our findings are in agreement with one systematic review and meta‐analysis (Cheng 2014) and an update of the Bone and Joint Decade Task Force on Neck Pain and its Associated Disorders by the OPTIMa collaboration (Wong 2016). The former review reported on 5 out of 16 trials with lower risk of bias and noted that massage was an effective intervention. The conclusion of the Cheng 2014 review stated that there was moderate‐certainty evidence that massage was more effective in reducing immediate pain than inactive therapies. The conclusions of Wong 2016 were limited to low risk of bias trials and identified that massage had added benefit to self‐care advice.

Authors' conclusions

Implications for practice.

Based on low‐certainty evidence from 33 trials, we are not confident in our findings on the use of massage for neck pain. These emerging findings suggest that there are mixed results across varied doses and techniques. Many of the results may have low clinical applicability as the studies only looked at short‐term effects after a single session. There was little or no difference in pain reduction when massage was compared to a sham treatment in the short term (5% better), but massage probably results in a large reduction in pain (24% better) when contrasted with no treatment. This diversity in results may be related to the predominance of selection and detection bias in the studies.

Low‐certainty evidence shows that when compared to a placebo, massage may result in a little improvement to no difference in disability‐function, health‐related quality of life and participant‐reported treatment success. The current evidence used varying doses of massage: high‐dose (one session per week for eight weeks of 45 minutes duration or three sessions per week for four weeks of 60 minutes duration) or medium‐dose (one to two sessions per week for two to six weeks of 20 to 30 minutes duration). This indicates that optimising the dosage of massage may make a difference. Information on important harm outcomes is missing; no difference in the risk of mild adverse events (i.e. soreness) was noted. The relative value of potentially small benefits and equivocal transient harms is important for the decision‐maker and patients.

There were several challenges in this review. Most notably, study quality was low as all included studies were at risk of performance and detection bias, and the number of participants in most trials was small. Most studies lacked information on the definition and dose of the massage technique used. The heterogeneity within the current literature makes it difficult to draw firm conclusions on the effectiveness of massage for neck pain.

Implications for research.

Trials newly included in this review could move massage research forward if their results can be replicated and expanded. Focused planning for large, adequately dosed, well‐designed trials is needed. We postulate that there is a need to assess higher‐dose massage in future randomised controlled trials (RCTs). Optimal dose determination should be examined.

Attention to blinding and use of a viable placebo or sham is needed. The majority of the studies in this review were rated at high risk for performance and detection bias. This bias was inherent in the majority of trials in that blinding of patient, care provider and outcome assessor was often not achieved. Blinding of the therapist is impossible in the case of massage. Additionally, blinding of the patient is difficult due to the perceptible differences in group allocation. Patients would easily have been able to recognise if they had received massage therapy compared to none at all. Lack of blinding can lead to exaggerated treatment effect estimates; however, there could be ways to mitigate this, so that the potential bias due to lack of blinding is limited. First, trials could examine equally plausible treatments, i.e. the participants in both the index and control groups feel that they are receiving the best treatment. Second, trials may measure the success of blinding amongst the participants (did they think they received the best treatment?). Finally, an effort to blind the outcome assessor is required. When the patient themselves is the outcome assessor for self‐reported outcomes, such as pain, they cannot be blinded if they are aware of their treatment allocation. This was the case for most studies in this review, where the primary outcome was a self‐reported outcome such as pain, participant‐reported treatment success, disability‐function or health‐related quality of life. The bottom line is that blinding may be impossible for some procedures unless a viable placebo procedure exists or is developed in the future. Objective observer‐based outcome measures need to be employed in these trials. There are other design features that are available to attempt to compensate for the lack of blinding. Treatments need to be equally credible and acceptable to patients to minimise placebo effects and high dropout rates. It is also necessary to nest qualitative research into these trials and to collect information on patients' previous experience with massage, and on their expectation of massage, to assess the impact of expectations.

Several difficulties undermined our understanding of the contribution of massage to the overall effect when combined with other treatment techniques. For instance, the study designs were not such that the potential influence of interaction effects could be determined. Factorial designs would be needed to tease out the contribution of massage to the interaction effects with other therapies, and these studies were not done. For example, two commonly used neck pain modalities are deep tissue massage and cervical manipulation. A 2 x 2 factorial design study that randomised participants to the four treatment combinations would allow comparison of four experimental situations: no treatment, massage alone, manipulation alone and the two treatments combined. The 2 x 2 design would demonstrate the main effects as well as their interaction, with confidence intervals and magnitudes of importance to clinicians and patients.

Future trialists should register a protocol prospectively, include core outcome domains and consider additional outcomes such as cost of care and health care utilisation. In addition, in the planning stages, researchers should consider and adequately report on the qualifications and experience of the practitioner providing massage and use the TIDieR (Template for Intervention Description and Replication) checklist and CONSORT extension for non‐pharmacological interventions to enhance the replication of massage characteristics and dose. Studies should be multi‐centred and include monitoring for consistency and compliance with the administration of the massage intervention to increase precision and reduce inconsistency. Running trials at individual centres without guidance from a central 'expert body' will perpetuate the current state of the evidence. To assess massage combined with another treatment, factorial designs with randomisation would be more informative to delineate both the individual and additive/subtractive effect of massage. Attention to Cochrane’s risk of bias domains must underpin future trials. Single‐session pre‐post RCT designs should be discontinued as they offer little or no value. Trials should instead build on appropriate theoretical and modelling work previously completed.

What's new

Date	Event	Description
25 July 2024	Amended	Typographical errors corrected.

History

Protocol first published: Issue 3, 2004
Review first published: Issue 3, 2006

Date	Event	Description
28 February 2024	New search has been performed	This review was amended and conducted according to the guidelines recommended by the Cochrane Musculoskeletal Editorial Board (Ghogomu 2014). We updated the search to 1 October 2023 from our last search date of February 2012. We updated the methods to meet the Methodological Expectations of Cochrane Intervention Reviews (MECIR) (2021), the Cochrane Handbook for Systematic Reviews of Interventions guidelines (Handbook 2021) and the requirements of the Cochrane Style Manual (2021). We modernised our title from 'mechanical neck disorders' to 'neck pain'. We also updated our author list: we removed four authors and added one. Thirty‐three studies met our inclusion criteria. Massage may result in little or no difference in improvement of function‐disability, quality of life and global perceived effect compared to sham in the short term. Pain reduction varied by comparison. A large pain reduction was noted when massage was compared against no treatment, while little to no difference was noted when massage was compared against a placebo/sham or when massage was added to the same co‐treatment. Subgroup analysis by dose may suggest a higher dose is better. Focused planning for larger, adequately dosed, well‐designed trials is needed.
28 February 2024	New citation required and conclusions have changed	We added 28 new trials in this update. The contribution of massage to the management of neck pain remains uncertain given the predominance of low‐certainty evidence in this field.
9 April 2013	Amended	Sherman 2009 had reported adverse events in their study in the form of adverse experiences. We have now included the adverse events that were reported in the study. In the review, Sherman 2009 was reported as 'one trial, three arms, 60 participants' but this has been corrected to 'one trial, two arms, 60 participants.'
9 July 2012	New search has been performed	Fifteen studies were included in the review. We excluded studies that assessed massage as a part of multimodal treatment. In the previous review, methodological quality was assessed using the van Tulder and Jadad scale, but for this review, the risk of bias assessment tool was used. Furthermore, the level of evidence quality was assessed using the GRADE approach.
9 July 2012	New citation required but conclusions have not changed	Conclusions have not changed since the previous update. There is still a limitation in the quality of studies on massage therapy for neck pain. Furthermore, due to the variability in massage techniques, there continues to be a challenge in compiling results from different studies.
12 June 2008	Amended	Converted to new review format.
22 May 2006	New citation required and conclusions have changed	Substantive amendment

Appendices

Appendix 1. Massage terminology

Massage research: styles or techniques allowed with no restrictions (Sherman 2006)

1. Application of cold

2. Application of heat

3. Compression – pumping or static

4. Craniosacral

5. Friction or direct pressure

6. Cross‐fibre friction

7. Gliding (effluerage) ‐ Swedish

8. Gliding – deep (effluerage, stripping) ‐ clinical

9. Holding

10. Kneading (petrissage)

11. Lymphatic drainage

12. Percussion (tapotement)

13. Rocking, jostling, shaking, vibration

14. Range of motion – active or resistive (also active assisted or resisted stretching, muscle energy technique, proprioceptive neuromuscular fascilitation – consisting of three types of resistive stretching: lengthening, contracting the agonist; lengthening the agonist, contracting the antagonist; and lengthening the agonist, contracting agonist and antagonist)

15. Range of motion – passive (passive stretching, positional release)

16. Skin rolling

17. Stretching (manual)

18. Traction (manual) or pull

19. Trigger point therapy

MASSAGE TERMINOLOGY

ACUPRESSURE ‐ a system of balancing the body’s energy by applying pressure to specific acupoints to release tension and increase circulation. The many hands‐on methods of stimulating the acupressure points can strengthen weaknesses, relieve common ailments, prevent health disorders and restore the body’s vital life force (https://www.massage-exam.com/massageglossary.php#AAB).

CONNECTIVE TISSUE MASSAGE ‐ connective tissue massage. Massage therapy. A form of massage that specifically manipulates the connective tissues (e.g. fascia, ligaments and tendons of the musculoskeletal system), with the aim of enhancing circulation and, by extension, healing (https://medical-dictionary.thefreedictionary.com/connective+tissue+massage).

CRANIOSACRAL TECHNIQUE ‐ craniosacral therapy is a pain management technique focusing on the manipulation of the membranes and cerebrospinal fluid surrounding the brain. The technique combines alternative healing practices with traditional cranial osteopathy by applying massage to the head and spine with the goal of unblocking points of tension. Craniosacral therapy is based on the notion that releasing restrictions in the craniosacral system will improve the functioning of the central nervous system.

EFFLEURAGE ‐ Effleurage is the most common stroke in Swedish massage. It is a free‐flowing and gliding movement towards the heart, tracing the contours of the body using the palm of one or both hands. Oil is applied with this stroke to begin the first stage of massage. The therapist applies a light or medium constant pressure. This stroke is used to warm up the muscles, relax the body, calm the nerves, improve blood circulation and heart function, and improve lymphatic drainage (https://www.encyclopedia.com/medicine/encyclopedias-almanacs-transcripts-and-maps/swedish-massage).

FRICTION ‐ Friction strokes work on deeper muscles than effleurage and petrissage. The friction technique is a pressure stroke and is the deepest that is used in Swedish massage. The massage therapist applies pressure by placing the weight of his or her body on the flat of the hand and the pads of the thumbs, knuckles, fingers or the back of the forearms, and then releases the pressure slowly and gently. This movement should be a continuous sliding motion or a group of alternating circular motions (https://www.encyclopedia.com/medicine/encyclopedias-almanacs-transcripts-and-maps/swedish-massage).

ISCHAEMIC COMPRESSION – ischaemic compression is a therapy technique used in physical therapy, where blockage of blood in an area of the body is deliberately made, so that a resurgence of local blood flow will occur upon release.

Ischaemic compression is commonly applied to trigger points, in what is known as trigger point therapy, where enough sustained pressure is applied to a trigger point with a tolerable amount of pain, and as discomfort is reduced, additional pressure is gradually given (ischaemic compression).

MUSCLE ENERGY TECHNIQUE – muscle energy techniques describe a broad class of manual therapy techniques directed at improving musculoskeletal function or joint function, and improving pain. Historically, the concept emerged as a form of osteopathic manipulative diagnosis and treatment in which the patient's muscles are actively used on request, from a precisely controlled position, in a specific direction, and against a distinctly executed physician counterforce. It was first described in 1948 by Fred Mitchell, Sr, D.O. Muscle energy techniques are used to treat somatic dysfunction, especially decreased range of motion, muscular hypertonicity and pain.

MYOFASCIAL RELEASE ‐ myofascial release is a manipulative treatment that attempts to release tension in the fascia due to trauma, posture or inflammation. Connective tissues called fascia surround the muscles, bones, nerves and organs of the body. Points of restriction in the fascia can place a great deal of pressure on nerves and muscles, causing chronic pain.

Practitioners of myofascial release employ long stretching strokes meant to balance tissue and muscle mechanics and improve joint range of motion in order to relieve pain (https://www.spine-health.com/glossary/myofascial-release).

PETRISSAGE ‐ from the French word 'petrir' meaning to knead, this massage technique is used after the tissue is warmed, because of the focused work. It includes muscle squeezing, muscle stripping, wringing, picking up, skin rolling and a variety of kneading techniques (https://www.massage-exam.com/massage-e.php).

PERCUSSION (TAPOTEMENT) is a quick, choppy, rhythmic movement that has a stimulating or toning effect. The following are variations of percussion/tapotement:

Cupping: the therapist forms the hands into a cup shape with fingers straight but bending only at the lower knuckles; the thumbs are kept close to the palms. The therapist strikes the flesh with the flat of the hands one after another in quick succession.

Hacking: this technique is similar to cupping. The therapist uses the sides of the hands with palms facing one another to make a chopping movement.

Pummelling: for this stroke, the therapist makes loose fists in both hands and applies them rapidly in succession over the thighs and buttocks.

Tapotement techniques are invigorating for most people but may be too intense for some. When prolonged, tapotement leads to overstimulation and even exhaustion of the nerves and muscles. In addition, it should not be used over varicose veins or directly above bony structures (https://www.encyclopedia.com/medicine/encyclopedias-almanacs-transcripts-and-maps/swedish-massage).

POLARITY THERAPY ‐ in the bodywork part of this therapy, the practitioner works with the client's energy field ‐ electromagnetic patterns expressed in mental, emotional and physical experience ‐ to facilitate greater health. Developed by Dr. Randolph Stone (https://www.massage-exam.com/massage-p.php).

POSITIONAL RELEASE ‐ positional release therapy is also known by its parent term, strain counterstrain. See strain counterstrain.

RANGE OF MOTION ‐ the range, usually expressed in degrees, through which a joint can move or be moved (https://www.massage-exam.com/massage-r.php).

REIKI ‐ a form of energy healing. Reiki involves gentle touch that directs Chi (Qi or Ki) for the purpose of strengthening the client's energy system. Qi or Chi (in China) and Ki (in Japan) is the term used by Traditional Chinese Medicine or Complementary Alternative Medicine practitioners, meaning universal life force energy (https://www.massage-exam.com/massage-r.php).

SHIATSU ‐ Shiatsu massage is an alternative technique that involves manual pressure applied to specific points on the body in an attempt to relieve tension and pain. According to the Japanese healing treatment, Shiatsu (meaning finger pressure) massage uses thumbs, fingers, elbows and knees to concentrate pressure at certain energy pathways in the body called meridians. The massage, also known as acupressure, works to reduce muscle tension and fatigue and is thought to improve blood circulation in the body as well as improve function of the lymphatic system.

STRAIN COUNTERSTRAIN ‐ within manual therapy, strain‐counterstrain is a type of "passive positional release" created in 1955 by Lawrence Jones, D.O. It is a hands‐on treatment that attempts to alleviate muscle and connective tissue tightness by the use of very specific treatment positions held for 90 seconds (can be held for up to three minutes in neurological patients). During the procedure, the involved tissue is "slackened", causing a relaxation of the "spasm" which, in turn, allows local areas of inflammation, trapped within the painful tissue to dissipate. Following this "release", there is an immediate reduction of pain and tension in the involved tissue. This relaxation aims to restore normal joint mobility and release other structures in the region that may have been compressed (Wikipedia/strain counterstrain [https://Wikipedia/strain counterstrain]).

SUBOCCIPITAL INHITIBITIVE DISTRACTION – a technique in which the therapist uses the fingertips of both hands to exert a sustained ventrocranial force on the occiput just caudal to the superior nuchal line in order to inhibit the muscles inserting into the nuchal line.

SWEDISH MASSAGE – Swedish massage is a manual therapy that emphasises long strokes, kneading and friction motions to encourage proper circulation. The technique involves rubbing muscles together in the same direction as the flow of blood to the heart. Swedish massage works to improve circulation, relieve muscle tension, stretch tight ligaments and reduce emotional stress (https://www.spine-health.com/glossary/myofascial-release).

TRADITIONAL CHINESE MASSAGE (TUINA) ‐ The term tui na (pronounced "twee naw"), which literally means "pinch and pull," refers to a wide range of Traditional Chinese Medicine therapeutic massage and body work. Tui na is not generally used for pleasure and relaxation, but rather as a treatment to address specific patterns of disharmony.

TRAGER ‐ hands‐on movement education that uses gentle movements to release physical and emotional tension patterns and facilitate relaxation, increased physical mobility and mental clarity. Created and developed by Dr. Milton Trager (https://www.massage-exam.com/massage-t.php).

TRIGGER POINTS ‐ a hyper‐irritable area within a taut band of skeletal muscle, located in the muscular tissue and/or its associated fascia. The spot is painful on compression and can cause characteristic referred pain and autonomic phenomena (https://www.massage-exam.com/massage-t.php).

VIBRATION ‐ to effect vibration, the massage therapist gently shakes or trembles the flesh with the hand or fingertips, then moves on to another spot and repeats this stroke. Vibration is designed to release muscle tension in small muscle areas, such as those on the face or along the spine (https://www.encyclopedia.com/medicine/encyclopedias-almanacs-transcripts-and-maps/swedish-massage).

Appendix 2. Search strategy

MEDLINE (Ovid)

1. randomized controlled trial.pt.
2. controlled clinical trial.pt.
3. randomized.ab.
4. placebo.ab.
5. randomly.ab.
6. trial.ab.
7. groups.ab.
8. 1 or 2 or 3 or 4 or 5 or 6 or 7
9. exp animals/ not humans.sh.
10. 8 not 9
11. Neck Pain/
12. exp Brachial Plexus Neuropathies/
13. exp neck injuries/ or exp whiplash injuries/
14. cervical pain.mp.
15. neckache.mp.
16. whiplash.mp.
17. cervicodynia.mp.
18. cervicalgia.mp.
19. brachialgia.mp.
20. brachial neuritis.mp.
21. brachial neuralgia.mp.
22. neck pain.mp.
23. neck injur*.mp.
24. brachial plexus neuropath*.mp.
25. brachial plexus neuritis.mp.
26. thoracic outlet syndrome/ or cervical rib syndrome/
27. Torticollis/
28. exp brachial plexus neuropathies/ or exp brachial plexus neuritis/
29. cervico brachial neuralgia.ti,ab.
30. cervicobrachial neuralgia.ti,ab.
31. (monoradicul* or monoradicl*).tw.
32. or/11‐31
33. exp headache/ and cervic*.tw.
34. exp genital diseases, female/
35. genital disease*.mp.
36. or/34‐35
37. 33 not 36
38. 32 or 37
39. neck/
40. neck muscles/
41. exp cervical plexus/
42. exp cervical vertebrae/
43. atlanto‐axial joint/
44. atlanto‐occipital joint/
45. Cervical Atlas/
46. spinal nerve roots/
47. exp brachial plexus/
48. (odontoid* or cervical or occip* or atlant*).tw.
49. axis/ or odontoid process/
50. Thoracic Vertebrae/
51. cervical vertebrae.mp.

52. cervical plexus.mp.
53. cervical spine.mp.

54. (neck adj3 muscles).mp.
55. (brachial adj3 plexus).mp.
56. (thoracic adj3 vertebrae).mp.
57. neck.mp.

58. (thoracic adj3 spine).mp.
59. (thoracic adj3 outlet).mp.
60. trapezius.mp.
61. cervical.mp.
62. cervico*.mp.
63. 61 or 62
64. exp genital diseases, female/
65. genital disease*.mp.
66. exp *Uterus/
67. 64 or 65 or 66
68. 63 not 67
69. 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or 51 or 52 or 53 or 54 or 55 or 56 or 57 or 58 or 59 or 60 or 68
70. exp pain/
71. exp injuries/
72. pain.mp.
73. ache.mp.
74. sore.mp.
75. stiff.mp.
76. discomfort.mp.
77. injur*.mp.
78. neuropath*.mp.
79. or/70‐78
80. 69 and 79
81. Radiculopathy/
82. exp temporomandibular joint disorders/ or exp temporomandibular joint dysfunction syndrome/
83. myofascial pain syndromes/
84. exp "Sprains and Strains"/
85. exp Spinal Osteophytosis/
86. exp Neuritis/
87. Polyradiculopathy/
88. exp Arthritis/
89. Fibromyalgia/
90. spondylitis/ or discitis/
91. spondylosis/ or spondylolysis/ or spondylolisthesis/
92. radiculopathy.mp.
93. radiculitis.mp.
94. temporomandibular.mp.
95. myofascial pain syndrome*.mp.
96. thoracic outlet syndrome*.mp.
97. spinal osteophytosis.mp.
98. neuritis.mp.
99. spondylosis.mp.
100. spondylitis.mp.
101. spondylolisthesis.mp.
102. or/81‐101
103. 69 and 102
104. exp neck/
105. exp cervical vertebrae/
106. Thoracic Vertebrae/
107. neck.mp.
108. (thoracic adj3 vertebrae).mp.
109. cervical.mp.
110. cervico*.mp.
111. 109 or 110
112. exp genital diseases, female/
113. genital disease*.mp.
114. exp *Uterus/
115. or/112‐114
116. 111 not 115
117. (thoracic adj3 spine).mp.
118. cervical spine.mp.
119. 104 or 105 or 106 or 107 or 108 or 116 or 117 or 118
120. Intervertebral Disk/
121. (disc or discs).mp.
122. (disk or disks).mp.
123. 120 or 121 or 122
124. 119 and 123
125. herniat*.mp.
126. slipped.mp.
127. prolapse*.mp.
128. displace*.mp.
129. degenerat*.mp.
130. (bulge or bulged or bulging).mp.
131. 125 or 126 or 127 or 128 or 129 or 130
132. 124 and 131
133. intervertebral disk degeneration/ or intervertebral disk displacement/
134. intervertebral disk displacement.mp.
135. intervertebral disc displacement.mp.
136. intervertebral disk degeneration.mp.
37. intervertebral disc degeneration.mp.
138. 133 or 134 or 135 or 136 or 137
139. 119 and 138
140. 38 or 80 or 103 or 132 or 139
141. animals/ not (animals/ and humans/)
142. 140 not 141
143. exp *neoplasms/
144. exp *wounds, penetrating/
145. 143 or 144
146. 142 not 145
147. Neck Pain/rh, th [Rehabilitation, Therapy]
148. exp Brachial Plexus Neuropathies/rh, th
149. exp neck injuries/rh, th or exp whiplash injuries/rh, th
150. thoracic outlet syndrome/rh, th or cervical rib syndrome/rh, th
151. Torticollis/rh, th
152. exp brachial plexus neuropathies/rh, th or exp brachial plexus neuritis/rh, th
153. or/147‐152
154. Radiculopathy/rh, th
155. exp temporomandibular joint disorders/rh, th or exp temporomandibular joint dysfunction syndrome/rh, th
156. myofascial pain syndromes/rh, th
157. exp "Sprains and Strains"/rh, th
158. exp Spinal Osteophytosis/rh, th
159. exp Neuritis/rh, th
160. Polyradiculopathy/rh, th
161. exp Arthritis/rh, th
162. Fibromyalgia/rh, th
163. spondylitis/rh, th or discitis/rh, th164. spondylosis/rh, th or spondylolysis/rh, th or spondylolisthesis/rh, th
165. or/154‐164
166. 69 and 165
167. acupuncture/ or chiropractic/
168. exp Musculoskeletal Manipulations/
169. massage.tw.
170. mobili?ation.tw.
171. Acupuncture Therapy/
172. (acupuncture or acu‐puncture or needling or acupressure or mox?bustion).tw.
173. ((neck or spine or spinal or cervical or chiropractic* or musculoskeletal* or musculo‐skeletal*) adj3 (adjust* or manipulat* or mobiliz* or mobilis*)).tw.
174. (manual adj therap*).tw.
175. (manipulati* adj (therap* or medicine)).tw.
176. (massag* or reflexolog* or rolfing or zone therap*).tw.
177. Nimmo.mp.
178. exp Vibration/tu [Therapeutic Use]
179. (vibration adj5 (therap* or treatment*)).tw.
180. (Chih Ya or Shiatsu or Shiatzu or Zhi Ya).tw.
181. (flexion adj2 distraction*).tw.
182. (myofascial adj3 (release or therap*)).tw.
183. muscle energy technique*.tw.
184. trigger point.tw.
185. proprioceptive Neuromuscular Facilitation*.tw.
186. cyriax friction.tw.
187. (lomilomi or lomi‐lomi or trager).tw.
188. aston patterning.tw.
189. (strain adj counterstrain).tw.
190. (craniosacral therap* or cranio‐sacral therap*).tw.
191. (amma or ammo or effleuurage or petrissage or hacking or tapotment).tw.
192. Complementary Therapies/
193. ((complement* or alternat* or osteopthic*) adj (therap* or medicine)).tw.
194. (Tui Na or Tuina).tw.
195. or/167‐194
196. 146 and 195
197. 153 or 166 or 196
198. 10 and 197
199. limit 198 to ed=20180326‐20181002

EMBASE (Ovid)

1. neck pain/

2. brachial plexus neuropathy/

3. neck injury/ or whiplash injury/

4. cervical pain.mp.

5. neckache.mp.

6. whiplash.mp.

7. cervicodynia.mp.

8. cervicalgia.mp.

9. brachialgia/

10. brachialgia.mp.

11. brachial neuritis.mp.

12. brachial neuralgia.mp.

13. neck pain.mp.

14. neck injur*.mp.

15. brachial plexus neuropath*.mp.

16. brachial plexus neuritis.mp.

17. thorax outlet syndrome/

18. torticollis/

19. cervico brachial neuralgia.ti,ab.

20. cervicobrachial neuralgia.ti,ab.

21. (monoradicul* or monoradicl*).tw.

22. or/1‐21

23. exp headache/ and cervic*.tw.

24. exp gynecologic disease/

25. genital disease*.mp.

26. exp *uterine cervix/

27. or/24‐26

28. 23 not 27

29. 22 or 28

30. neck/ or neck muscle/

31. cervical plexus/

32. cervical spine/

33. atlantoaxial joint/

34. atlantooccipital joint/

35. atlas/

36. "spinal root"/

37. brachial plexus/

38. (odontoid* or cervical or occip* or atlant*).tw.

39. odontoid process/

40. cervical vertebra.mp.

41. cervical vertebrae.mp.

42. cervical plexus.mp.

43. cervical spine.mp.

44. (neck adj3 muscles).mp.

45. (brachial adj3 plexus).mp.

46. (thoracic adj3 vertebra?).mp.

47. neck.mp.

48. (thoracic adj3 spine).mp.

49. (thoracic adj3 outlet).mp.

50. trapezius.mp.

51. cervical.mp.

52. cervico*.mp.

53. 51 or 52

54. exp gynecologic disease/

55. genital disease*.mp.

56. exp *uterine cervix/

57. 54 or 55 or 56

58. 53 not 57

59. 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or 58

60. exp pain/

61. exp injury/

62. pain.mp.

63. ache.mp.

64. sore.mp.

65. stiff.mp.

66. discomfort.mp.

67. injur*.mp.

68. neuropath*.mp.

69. or/60‐68

70. radiculopathy/

71. temporomandibular joint disorder/

72. myofascial pain/

73. spondylosis/ or cervical spondylosis/

74. neuritis/

75. exp arthritis/

76. fibromyalgia/

77. exp spondylitis/

78. diskitis/

79. spondylolisthesis/

80. radiculopathy.mp.

81. radiculitis.mp.

82. temporomandibular.mp.

83. myofascial pain syndrome*.mp.

84. spinal osteophytosis.mp.

85. neuritis.mp.

86. spondylosis.mp.

87. spondylitis.mp.

88. spondylolisthesis.mp.

89. or/70‐88

90. 59 and 89

91. neck/

92. cervical spine/

93. neck.mp.

94. (thoracic adj3 vertebra?).mp.

95. cervical.mp.

96. cervico*.mp.

97. exp gynecologic disease/

98. genital disease*.mp.

99. exp *uterine cervix/

100. or/97‐99

101. 95 or 96

102. 101 not 100

103. (thoracic adj3 spine).mp.

104. cervical spine.mp.

105. 91 or 92 or 93 or 94 or 102 or 103 or 104

106. intervertebral disk/

107. (disc or discs).mp.

108. (disk or disks).mp.

109. 106 or 107 or 108

110. 105 and 109

111. herniat*.mp.

112. slipped.mp.

113. prolapse*.mp.

114. displace*.mp.

115. degenerat*.mp.

116. (bulge or bulged or bulging).mp.

117. 110 or 111 or 112 or 113 or 114 or 115 or 116

118. 110 and 117

119. intervertebral disk hernia/

120. intervertebral disk degeneration/

121. intervertebral disc degeneration.mp.

122. intervertebral disk degeneration.mp.

123. intervertebral disc displacement.mp.

124. intervertebral disk displacement.mp.

125. 119 or 120 or 121 or 122 or 123 or 124

126. 105 and 125

127. 59 and 69

128. 29 or 90 or 118 or 126 or 127

129. exp *neoplasm/

130. exp *penetrating trauma/

131. 129 or 130

132. 128 not 131

133. neck pain/rh, th

134. brachial plexus neuropathy/rh, th

135. neck injury/ or whiplash injury/rh, th

136. brachialgia/rh, th

137. thorax outlet syndrome/rh, th

138. Torticollis/rh, th

139. Radiculopathy/rh, th

140. temporomandibular joint disorder/rh, th

141. myofascial pain/rh, th

142. spondylosis/rh, th or cervical spondylosis/rh, th

143. neuritis/rh, th

144. exp arthritis/rh, th

145. Fibromyalgia/rh, th

146. exp spondylitis/rh, th

147. diskitis/rh, th

148. spondylolisthesis/rh, th

149. acupuncture/ or acupressure/ or acupuncture analgesia/

150. exp manipulative medicine/

151. massage.tw.

152. mobili?ation.tw.

153. (acupuncture or acu‐puncture or needling or acupressure or mox?bustion).tw.

154. ((neck or spine or spinal or cervical or chiropractic* or musculoskeletal* or musculo‐skeletal*) adj3 (adjust* or manipulat* or mobiliz* or mobilis*)).tw.

155. (manual adj therap*).tw.

156. (manipulati* adj (therap* or medicine)).tw.

157. (massag* or reflexolog* or rolfing or zone therap*).tw.

158. Nimmo.tw.

159. (vibration adj5 (therap* or treatment*)).tw.

160. (Chih Ya or Shiatsu or Shiatzu or Zhi Ya).tw.

161. (flexion adj2 distraction*).tw.

162. (myofascial adj3 (release or therap*)).tw.

163. muscle energy technique*.tw.

164. trigger point.tw.

165. proprioceptive Neuromuscular Facilitation*.tw.

166. cyriax friction.tw.

167. (lomilomi or lomi‐lomi or trager).tw.

168. aston patterning.tw.

169. (strain adj counterstrain).tw.

170. (craniosacral therap* or cranio‐sacral therap*).tw.

171. (amma or ammo or effleuurage or petrissage or hacking or tapotment).tw.

172. alternative medicine/

173. ((complement* or alternat* or osteopthic*) adj (therap* or medicine)).tw.

174. (Tui Na or Tuina).tw.

175. (swedish massage or rolfing).tw.

176. therapeutic touch.mp.

177. massotherapy.tw.

178. effleurage.mp.

179. or/149‐178

180. 132 and 179

181. 133 or 134 or 135 or 136 or 137 or 138

182. or/139‐148

183. 59 and 182

184. 180 or 181 or 183

185. randomised controlled trial/

186. controlled clinical trial/

187. (random* or sham or placebo*).tw.

188. placebo/

189. randomisation/

190. single blind procedure/

191. double blind procedure/

192. ((singl* or doubl* or trebl* or tripl*) adj5 (blind* or dumm*or mask*)).ti,ab.

193. (rct or rcts).tw.

194. (control* adj2 (study or studies or tiral*)).tw.

195. or/185‐194

196. human/

197. nonhuman/

198. animal/

199. animal experiment/

200. or/197‐199

201. 200 not (200 and 196)

202. 195 not 201

203. 184 and 202

204. limit 203 to yr="2006 ‐Current"

205. limit 203 to yr="1928 ‐ 2005"

206. guidelines as topic/

207. practice guidelines as topic/

208. (guideline? or guidance or recommendations).ti.

209. consensus.ti.

210. or/206‐209

211. 184 and 210

212. limit 211 to yr="2006 ‐Current"

213. limit 211 to yr="1928 ‐ 2005"

214. meta analysis/

215. systematic review/

216. (meta analy* or metaanaly* or met analy* or metanaly*).tw.

217. (collaborative research or collaborative review* or collaborative overview*).tw.

218. (integrative research or integrative review* or intergrative overview*).tw.

219. (quantitative adj3 (research or review* or overview*)).tw.

220. (research integration or research overview*).tw.

221. (systematic* adj3 (review* or overview*)).tw.

222. (methodologic* adj3 (review* or overview*)).tw.

223. biomedical technology assessment/

224. (hta or thas or technology assessment*).tw.

225. ((hand adj2 search*) or (manual* adj search*)).tw.

226. ((electronic adj database*) or (bibliographic* adj database*)).tw.

227. ((data adj2 abstract*) or (data adj2 extract*)).tw.

228. (data adj3 (pooled or pool or pooling)).tw.

229. (analys* adj3 (pool or pooled or pooling)).tw.

230. mantel haenszel.tw.

231. (cochrane or Pubmed or pub med or medline or embase or psycinfo or psyclit or psychinfo or psychlit or cinahl or science citation index).ab.

232. or/214‐231

233. 184 and 232

234. limit 233 to yr="2006 ‐Current"

235. limit 233 to yr="1928 ‐ 2005"

236. (ae or co or si or to).fs.

237. (safe or safety or unsafe).tw.

238. (side effect* or side event*).tw.

239. ((adverse or undesirable or harm* or injurious or serious or toxic) adj3 (effect* or event* or reaction* or incident* or outcome*)).tw.

240. (abnormalit* or toxicit* or complication* or consequence* or noxious or tolerabilit*).tw.

241. or/236‐240

242. 184 and 241

243. limit 242 to yr="2006 ‐Current"

244. limit 242 to yr="1928 ‐ 2005"

245. limit 203 to em=201027‐201216

246. limit 211 to em=201027‐201216

247. limit 233 to em=201027‐201216

248. limit 242 to em=201027‐201216

CENTRAL (Ovid)

1 Neck Pain/

2 exp Brachial Plexus Neuropathies/

3 exp neck injuries/ or exp whiplash injuries/

4 cervical pain.mp.

5 neckache.mp.

6 whiplash.mp.

7 cervicodynia.mp.

8 cervicalgia.mp.

9 brachialgia.mp.

10 brachial neuritis.mp.

11 brachial neuralgia.mp.

12 neck pain.mp.

13 neck injur*.mp.

14 brachial plexus neuropath*.mp.

15 brachial plexus neuritis.mp.

16 thoracic outlet syndrome/ or cervical rib syndrome/

17 Torticollis/

18 exp brachial plexus neuropathies/ or exp brachial plexus neuritis/

19 cervico brachial neuralgia.ti,ab.

20 cervicobrachial neuralgia.ti,ab.

21 (monoradicul* or monoradicl*).tw.

22 or/1‐21

23 exp headache/ and cervic*.tw.

24 exp genital diseases, female/

25 genital disease*.mp.

26 or/24‐25

27 23 not 26

28 22 or 27

29 neck/

30 neck muscles/

31 exp cervical plexus/

32 exp cervical vertebrae/

33 atlanto‐axial joint/

34 atlanto‐occipital joint/

35 Cervical Atlas/

36 spinal nerve roots/

37 exp brachial plexus/

38 (odontoid* or cervical or occip* or atlant*).tw.

39 axis/ or odontoid process/

40 Thoracic Vertebrae/

41 cervical vertebrae.mp.

42 cervical plexus.mp.

43 cervical spine.mp.

44 (neck adj3 muscles).mp.

45 (brachial adj3 plexus).mp.

46 (thoracic adj3 vertebrae).mp.

47 neck.mp.

48 (thoracic adj3 spine).mp.

49 (thoracic adj3 outlet).mp.

50 trapezius.mp.

51 cervical.mp.

52 cervico*.mp.

53 51 or 52

54 exp genital diseases, female/

55 genital disease*.mp.

56 exp *Uterus/

57 54 or 55 or 56

58 53 not 57

59 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or 58

60 exp pain/

61 exp injuries/

62 pain.mp.

63 ache.mp.

64 sore.mp.

65 stiff.mp.

66 discomfort.mp.

67 injur*.mp.

68 neuropath*.mp.

69 or/60‐68

70 59 and 69

71 Radiculopathy/

72 exp temporomandibular joint disorders/ or exp temporomandibular joint dysfunction syndrome/

73 myofascial pain syndromes/

74 exp "Sprains and Strains"/

75 exp Spinal Osteophytosis/

76 exp Neuritis/

77 Polyradiculopathy/

78 exp Arthritis/

79 Fibromyalgia/

80 spondylitis/ or discitis/

81 spondylosis/ or spondylolysis/ or spondylolisthesis/

82 radiculopathy.mp.

83 radiculitis.mp.

84 temporomandibular.mp.

85 myofascial pain syndrome*.mp.

86 thoracic outlet syndrome*.mp.

87 spinal osteophytosis.mp.

88 neuritis.mp.

89 spondylosis.mp.

90 spondylitis.mp.

91 spondylolisthesis.mp.

92 or/71‐91

93 59 and 92

94 exp neck/

95 exp cervical vertebrae/

96 Thoracic Vertebrae/

97 neck.mp.

98 (thoracic adj3 vertebrae).mp.

99 cervical.mp.

100 cervico*.mp.

101 99 or 100

102 exp genital diseases, female/

103 genital disease*.mp.

104 exp *Uterus/

105 or/102‐104

106 101 not 105

107 (thoracic adj3 spine).mp.

108 cervical spine.mp.

109 94 or 95 or 96 or 97 or 98 or 106 or 107 or 108

110 Intervertebral Disk/

111 (disc or discs).mp.

112 (disk or disks).mp.

113 110 or 111 or 112

114 109 and 113

115 herniat*.mp.

116 slipped.mp.

117 prolapse*.mp.

118 displace*.mp.

119 degenerat*.mp.

120 (bulge or bulged or bulging).mp.

121 115 or 116 or 117 or 118 or 119 or 120

122 114 and 121

123 intervertebral disk degeneration/ or intervertebral disk displacement/

124 intervertebral disk displacement.mp.

125 intervertebral disc displacement.mp.

126 intervertebral disk degeneration.mp.

127 intervertebral disc degeneration.mp.

128 123 or 124 or 125 or 126 or 127

129 109 and 128

130 28 or 70 or 93 or 122 or 129

131 animals/ not (animals/ and humans/)

132 130 not 131

133 exp *neoplasms/

134 exp *wounds, penetrating/

135 133 or 134

136 132 not 135

137 Neck Pain/rh, th [Rehabilitation, Therapy]

138 exp Brachial Plexus Neuropathies/rh, th

139 exp neck injuries/rh, th or exp whiplash injuries/rh, th

140 thoracic outlet syndrome/rh, th or cervical rib syndrome/rh, th

141 Torticollis/rh, th

142 exp brachial plexus neuropathies/rh, th or exp brachial plexus neuritis/rh, th

143 or/137‐142

144 Radiculopathy/rh, th

145 exp temporomandibular joint disorders/rh, th or exp temporomandibular joint dysfunction syndrome/rh, th

146 myofascial pain syndromes/rh, th

147 exp "Sprains and Strains"/rh, th

148 exp Spinal Osteophytosis/rh, th

149 exp Neuritis/rh, th

150 Polyradiculopathy/rh, th

151 exp Arthritis/rh, th

152 Fibromyalgia/rh, th

153 spondylitis/rh, th or discitis/rh, th

154 spondylosis/rh, th or spondylolysis/rh, th or spondylolisthesis/rh, th

155 or/144‐154

156 59 and 155

157 acupuncture/ or chiropractic/

158 exp Musculoskeletal Manipulations/

159 massage.tw.

160 mobili?ation.tw.

161 Acupuncture Therapy/

162 (acupuncture or acu‐puncture or needling or acupressure or mox?bustion).tw.

163 ((neck or spine or spinal or cervical or chiropractic* or musculoskeletal* or musculo‐skeletal*) adj3 (adjust* or manipulat* or mobiliz* or mobilis*)).tw.

164 (manual adj therap*).tw.

165 (manipulati* adj (therap* or medicine)).tw.

166 (massag* or reflexolog* or rolfing or zone therap*).tw.

167 Nimmo.mp.

168 exp Vibration/tu [Therapeutic Use]

169 (vibration adj5 (therap* or treatment*)).tw.

170 (Chih Ya or Shiatsu or Shiatzu or Zhi Ya).tw.

171 (flexion adj2 distraction*).tw.

172 (myofascial adj3 (release or therap*)).tw.

173 muscle energy technique*.tw.

174 trigger point.tw.

175 proprioceptive Neuromuscular Facilitation*.tw.

176 cyriax friction.tw.

177 (lomilomi or lomi‐lomi or trager).tw.

178 aston patterning.tw.

179 (strain adj counterstrain).tw.

180 (craniosacral therap* or cranio‐sacral therap*).tw.

181 (amma or ammo or effleuurage or petrissage or hacking or tapotment).tw.

182 Complementary Therapies/

183 ((complement* or alternat* or osteopthic*) adj (therap* or medicine)).tw.

184 (Tui Na or Tuina).tw.

185 or/157‐184

186 136 and 185

187 143 or 156 or 186

188 animals/ not (animals/ and humans/)

189 187 not 188

190 limit 189 to yr="2010 ‐ 2012"

CINAHL (EBSCO)

S139 S115 and S131 Limiters ‐ Published Date: 20121231‐20140330
S138 S115 and S131
S137 S109 and S131 Limiters ‐ Published Date: 20121231‐20140330
S136 S109 and S131
S135 S94 and S131 Limiters ‐ Published Date: 20121231‐20140330
S134 S94 and S131
S133 S91 and S131 Limiters ‐ Published Date: 20121231‐20140330
S132 S91 and S131
S131 S82 and S130
S130 S116 or S117 or S118 or S119 or S120 or S121 or S122 or S123 or S124 or S125 or S126 or S127 or S128 or S129
S129 TX ( ((complement* or alternat* or osteopthic*) N1 (therap* or medicine)) ) OR TX ( (Tui Na or Tuina) )
S128 TX (strain N1 counterstrain) OR TX ( (craniosacral therap* or cranio‐sacral therap*) ) OR TX ( (amma or ammo or effleuurage or petrissage or hacking or tapotment) )
S127 TX cyriax friction OR TX ( (lomilomi or lomi‐lomi or trager) ) OR TX aston patterning
S126 TX muscle energy technique* OR TX trigger point OR TX proprioceptive Neuromuscular Facilitation*
S125 TX ( (Chih Ya or Shiatsu or Shiatzu or Zhi Ya) ) OR TX (flexion N2 distraction*) OR TX ( (myofascial N3 (release or therap*)) )
S124 (MH "Vibration/TU")
S123 TX ( (massag* or reflexolog* or rolfing or zone therap*) ) OR TX Nimmo OR TX ( (vibration N5 (therap* or treatment*)) )
S122 TX (manipulati* N1 (therap* or medicine))
S121 TX ((neck or spine or spinal or cervical or chiropractic* or musculoskeletal* or musculo‐skeletal*) N3 (adjust* or manipulat* or mobiliz* or mobilis*))
S120 TX mobili?ation OR TX ( (acupuncture or acu‐puncture or needling or acupressure or mox?bustion) ) OR TX manual therapy
S119 (MH "Massage+") OR (MH "Deep Tissue Massage") OR (MH "Neuromuscular Massage") OR (MH "Sports Massage") OR (MH "Massage Therapists") OR (MH "Swedish Massage")
S118 (MH "Manipulation, Chiropractic") OR (MH "Manual Therapy+")
S117 (MH "Chiropractic+") OR (MH "Manipulation, Chiropractic") OR (MH "Chiropractic Practice") OR (MH "Chiropractors")
S116 (MH "Acupuncture+") OR (MH "Acupuncture Points") OR (MH "Acupuncturists")
S115 S110 or S111 or S112 or S113 or S114
S114 TX toxic reaction OR TX allergic reaction OR TX complications Limiters ‐ Published Date: 20100101‐20121231
S113 TX adverse outcome* OR TX adverse incident* Limiters ‐ Published Date: 20100101‐20121231
S112 TX adverse event* OR TX adverse effect* OR TX adverse reaction* Limiters ‐ Published Date: 20100101‐20121231
S111 TX ( (safe or safety or unsafe) ) OR TX ( (side effect* or side event*) ) Limiters ‐ Published Date: 20100101‐20121231
S110 (MH "Adverse Drug Event") Limiters ‐ Published Date: 20100101‐20121231
S109 S95 or S96 or S97 or S98 or S99 or S100 or S101 or S102 or S103 or S104 or S105 or S106 or S107 or S108
S108 TX PsycINFO or TX psycLIT or TX PsychINFO or TX psychLIT or TX CINAHL
S107 TX cochrane or TX pubmed or TX pub med or TX medline or TX embase
S106 TX mantel haenszel
S105 TX data N2 pool* or TX analys* N2 pool*
S104 TX data N2 abstract* or TX data N2 extract*
S103 TX electronic N2 database* or TX bibliographic database*
S102 TX hand N2 search* or TX manual N2 search
S101 TX hta or TX htas or TX technology assessment*
S100 TX methodologic* N3 review* or TX methodologic* N3 overview*
S99 TX systematic* N3 review* or TX systematic* N3 overview*
S98 TX quantitative research or TX quantitative review* or TX quantitative overview*
S97 TX meta analy* or TX metaanaly* or TX met analy* or TX metanaly*
S96 (MH "Meta Analysis")
S95 PT systematic review
S94 S92 or S93
S93 TI guideline* or TI guidance or TI recommendations or TI consensus
S92 (MH "Practice Guidelines")
S91 S83 or S84 or S85 or S86 or S87 or S88 or S89 or S90
S90 TX control* N2 study or TX control* N2 studies or TX control N2 trial*
S89 TX RCT or TX RCTs
S88 TX (singl* N1 (blind* OR dumm* OR mask*))
S87 (MH "Random Sample+")
S86 (MH "Placebos")
S85 TX random* or TX sham or TX placebo*
S84 PT clinical trial or PT randomised controlled trial
S83 (MH "Clinical Trials+")
S82 S78 NOT S81
S81 S79 or S80
S80 (MM "Pregnancy+")
S79 (MM "Abortion, Induced+")
S78 S74 NOT S77
S77 S75 or S76
S76 (MM "Wounds, Penetrating+")
S75 (MM "Neoplasms+")
S74 S16 or S41 or S56 or S69 or S73
S73 S63 and S72
S72 S70 or S71
S71 TX intervertebral disk displacement or TX intervertebral disc displacement or TX intervertebral disk degeneration or TX intervertebral disc degeneration
S70 (MH "Intervertebral Disk Displacement")
S69 S67 and S68
S68 TX herniat* or TX slipped or TX prolapse* or TX displace* or TX degenerat* or TX ( bulged OR bulge OR bulging )
S67 S63 and S66
S66 S64 or S65
S65 TX disc or TX discs or TX disk or TX disks
S64 (MH "Intervertebral Disk")
S63 S61 NOT S62
S62 (MM "Genital Diseases, Female+") or ( (MM "Cervix") or (MM "Cervix Diseases") )
S61 S57 or S58 or S59 or S60
S60 TX thoracic N3 spine or TX cervical spine or TX cervico*
S59 TX neck or TX thoracic N3 vertebr*
S58 (MH "Thoracic Vertebrae")
S57 (MH "Neck")
S56 S34 and S55
S55 S42 or S43 or S44 or S45 or S46 or S47 or S48 or S49 or S50 or S51 or S52 or S53 or S54
S54 TX neuritis or TX spondylosis or TX spondylitis or TX spondylolisthesis
S53 TX myofascial pain syndome* or TX thoracic outlet syndrome* or TX spinalosteophytosis
S52 TX radiculopathy or TX radiculitis or TX temporomandibular
S51 (MH "Spondylolysis") or (MH "Spondylolisthesis+")
S50 (MH "Fibromyalgia")
S49 (MH "Arthritis+")
S48 (MH "Polyradiculopathy")
S47 (MH "Neuritis+")
S46 (MH "Spinal Osteophytosis")
S45 (MH "Sprains and Strains+")
S44 (MH "Myofascial Pain Syndromes+")
S43 (MH "Temporomandibular Joint Diseases+") or (MH "Temporomandibular Joint Syndrome")
S42 (MH "Radiculopathy")
S41 S34 and S40
S40 S35 or S36 or S37 or S38 or S39
S39 (MH "Neuralgia")
S38 TX stiff or TX discomfort or TX injur* or TX neuropath*
S37 TX pain or TX ache* or TX sore
S36 (MH "Wounds and Injuries+")
S35 (MH "Pain+")
S34 S33 NOT S32
S33 S17 or S18 or S19 or S20 or S21 or S22 or S23 or S24 or S25 or S26 or S27 or S28 or S29 or S30 or S31
S32 (MM "Genital Diseases, Female+") or ( (MM "Cervix") or (MM "Cervix Diseases") )
S31 TX trapezius or TX cervico*
S30 TX thoracic N3 spine or TX thoracic N3 outlet
S29 TX neck
S28 TX thoracic N3 verteb*
S27 TX brachial N3 plexus
S26 TX neck n3 muscles
S25 (MH "Thoracic Vertebrae")
S24 TX ondontoid* or TX cervical or TX occip* or TX atlant*
S23 (MH "Brachial Plexus+")
S22 (MH "Spinal Nerve Roots+")
S21 (MH "Atlanto‐Axial Joint") or (MH "Atlanto‐Occipital Joint")
S20 (MH "Cervical Vertebrae+") or (MH "Cervical Atlas")
S19 (MH "Cervical Plexus+")
S18 (MH "Neck")
S17 (MH "Neck Muscles+")
S16 S10 or S15
S15 S11 NOT S14
S14 S12 or S13
S13 (MM "Cervix") or (MM "Cervix Diseases")
S12 (MM "Genital Diseases, Female+")
S11(MH "Headache+") and TX cervic*
S10 S1 or S2 or S3 or S4 or S5 or S6 or S7 or S8 or S9
S9 (MH "Brachial Plexus Neuritis")
S8 TX cervical brachial neuralgia
S7 TX cervical rib sydrome* or TX cervico brachial neuralgia or TX cervicobrachial neuralgia or TX monoradicul* or TX monoradicl*
S6 (MH "Thoracic Outlet Syndrome") or (MH "Torticollis")
S5 TX brachial neuralgia or TX neck pain or TX neck injur* or TX brachial plexus neuropath* or TX brachial plexus neuralgia or TX brachial plexus neuritis
S4 TX cervicalgia or TX brachialgia or TX brachial neuritis
S3 TX cervical pain or TX neckache or TX neck ache or TX whiplash or TX cervicodynia
S2 (MH "Neck Injuries+")
S1 (MH "Neck Pain") or (MH "Brachial Plexus Neuropathies") or (MH "Brachial Plexus Neuritis")

MANTIS (Ovid)

1 neck pain.mp. [mp=title, abstract, descriptors]

2 brachial plexus neuropathies.mp. [mp=title, abstract, descriptors]

3 neck injuries.mp. [mp=title, abstract, descriptors]

4 cervical pain.mp. [mp=title, abstract, descriptors]

5 neckache.mp. [mp=title, abstract, descriptors]

6 whiplash.mp. [mp=title, abstract, descriptors]

7 cervicodynia.mp. [mp=title, abstract, descriptors]

8 cervicalgia.mp. [mp=title, abstract, descriptors]

9 brachialgia.mp. [mp=title, abstract, descriptors]

10 brachial neuritis.mp. [mp=title, abstract, descriptors]

11 brachial neuralgia.mp. [mp=title, abstract, descriptors]

12 brachial plexus neuropath*.mp. [mp=title, abstract, descriptors]

13 brachial plexus neuritis.mp. [mp=title, abstract, descriptors]

14 (thoracic outlet syndrome or cervical rib syndrome).mp. [mp=title, abstract, descriptors]

15 torticollis.mp. [mp=title, abstract, descriptors]

16 cervico brachial neuralgia.mp. [mp=title, abstract, descriptors]

17 (monoradicul* or monoradicl*).tw.

18 or/1‐17

19 headache.mp. and cervic*.tw. [mp=title, abstract, descriptors]

20 genital diseases, female.mp. [mp=title, abstract, descriptors]

21 genital disease*.mp. [mp=title, abstract, descriptors]

22 or/20‐21

23 19 not 22

24 18 or 23

25 neck.mp. [mp=title, abstract, descriptors]

26 neck muscles.mp. [mp=title, abstract, descriptors]

27 cervical plexus.mp. [mp=title, abstract, descriptors]

28 cervical vertebrae.mp. [mp=title, abstract, descriptors]

29 atlanto‐axial joint.mp. [mp=title, abstract, descriptors]

30 atlanto‐occipital joint.mp. [mp=title, abstract, descriptors]

31 cervical atlas.mp. [mp=title, abstract, descriptors]

32 spinal nerve roots.mp. [mp=title, abstract, descriptors]

33 brachial plexus.mp. [mp=title, abstract, descriptors]

34 (odontoid* or cervical or occip* or atlant*).tw.

35 (axis or odontoid process).mp. [mp=title, abstract, descriptors]

36 thoracic vertebrae.mp. [mp=title, abstract, descriptors]

37 cervical vertebrae.mp. [mp=title, abstract, descriptors]

38 cervical plexus.mp. [mp=title, abstract, descriptors]

39 cervical spine.mp. [mp=title, abstract, descriptors]

40 (neck adj3 muscles).mp. [mp=title, abstract, descriptors]

41 (brachial adj3 plexus).mp. [mp=title, abstract, descriptors]

42 (thoracic adj3 vertebrae).mp. [mp=title, abstract, descriptors]

43 (thoracic adj3 spine).mp. [mp=title, abstract, descriptors]

44 (thoracic adj3 outlet).mp. [mp=title, abstract, descriptors]

45 trapezius.mp. [mp=title, abstract, descriptors]

46 cervical.mp. [mp=title, abstract, descriptors]

47 cervico*.mp. [mp=title, abstract, descriptors]

48 46 or 47

49 genital diseases, female.mp. [mp=title, abstract, descriptors]

50 genital disease*.mp. [mp=title, abstract, descriptors]

51 uterus.mp. [mp=title, abstract, descriptors]

52 49 or 50 or 51

53 48 not 52

54 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 53

55 pain.mp. [mp=title, abstract, descriptors]

56 injuries.mp. [mp=title, abstract, descriptors]

57 ache.mp. [mp=title, abstract, descriptors]

58 sore.mp. [mp=title, abstract, descriptors]

59 stiff.mp. [mp=title, abstract, descriptors]

60 discomfort.mp. [mp=title, abstract, descriptors]

61 injur*.mp. [mp=title, abstract, descriptors]

62 neuropath*.mp. [mp=title, abstract, descriptors]

63 or/55‐62

64 54 and 63

65 radiculopathy.mp. [mp=title, abstract, descriptors]

66 (temporomandibular joint disorders or temporomandibular joint dysfunction syndrome).mp. [mp=title, abstract, descriptors]

67 myofascial pain syndromes.mp. [mp=title, abstract, descriptors]

68 "sprains and strains".mp. [mp=title, abstract, descriptors]

69 spinal osteophytosis.mp. [mp=title, abstract, descriptors]

70 neuritis.mp. [mp=title, abstract, descriptors]

71 polyradiculopathy.mp. [mp=title, abstract, descriptors]

72 arthritis.mp. [mp=title, abstract, descriptors]

73 fibromyalgia.mp. [mp=title, abstract, descriptors]

74 (spondylitis or discitis).mp. [mp=title, abstract, descriptors]

75 (spondylosis or spondylolysis or spondylolisthesis).mp.,[mp=title, abstract, descriptors]

76 radiculitis.mp. [mp=title, abstract, descriptors]

77 tempomandibular.mp. [mp=title, abstract, descriptors]

78 myofascial pain syndrome*.mp. [mp=title, abstract, descriptors]

79 thoracic outlet syndrome*.mp. [mp=title, abstract, descriptors]

80 spinal osteophytosis.mp. [mp=title, abstract, descriptors]

81 neuritis.mp. [mp=title, abstract, descriptors]

82 spondylosis.mp. [mp=title, abstract, descriptors]

83 spondylitis.mp. [mp=title, abstract, descriptors]

84 spondylolisthesis.mp. [mp=title, abstract, descriptors]

85 or/65‐84

86 54 and 85

87 neck.mp. [mp=title, abstract, descriptors]

88 cervical vertebrae.mp. [mp=title, abstract, descriptors]

89 thoracic vertebrae.mp. [mp=title, abstract, descriptors]

90 (thoracic adj3 vertebrae).mp. [mp=title, abstract, descriptors]

91 cervical.mp. [mp=title, abstract, descriptors]

92 cervico*.mp. [mp=title, abstract, descriptors]

93 91 or 92

94 genital diseases, female.mp. [mp=title, abstract, descriptors]

95 genital disease*.mp. [mp=title, abstract, descriptors]

96 uterus.mp. [mp=title, abstract, descriptors]

97 or/94‐96

98 93 not 97

99 (thoracic adj3 spine).mp. [mp=title, abstract, descriptors]

100 cervical spine.mp. [mp=title, abstract, descriptors]

101 87 or 88 or 89 or 90 or 98 or 99 or 100

102 intervertebral disk.mp. [mp=title, abstract, descriptors]

103 (disc or discs).mp. [mp=title, abstract, descriptors]

104 (disk or disks).mp. [mp=title, abstract, descriptors]

105 102 or 103 or 104

106 101 and 105

107 herniat*.mp. [mp=title, abstract, descriptors]

108 slipped.mp. [mp=title, abstract, descriptors]

109 prolapse*.mp. [mp=title, abstract, descriptors]

110 displace*.mp. [mp=title, abstract, descriptors]

111 degenerat*.mp. [mp=title, abstract, descriptors]

112 (bulge or bulged or bulging).mp. [mp=title, abstract, descriptors]

113 107 or 108 or 109 or 110 or 111 or 112

114 106 and 113

115 intervertebral disk displacement.mp. [mp=title, abstract, descriptors]

116 intervertebral disc displacement.mp. [mp=title, abstract, descriptors]

117 intervertebral disk degeneration.mp. [mp=title, abstract, descriptors]

118 intervertebral disc degeneration.mp. [mp=title, abstract, descriptors]

119 115 or 116 or 117 or 118

120 101 and 119

121 24 or 64 or 86 or 106 or 114 or 120

122 (animals not (animals and humans)).mp. [mp=title, abstract, descriptors]

123 121 not 122

124 neoplasms.mp. [mp=title, abstract, descriptors]

125 wounds, penetrating.mp. [mp=title, abstract, descriptors]

126 124 or 125

127 123 not 126

128 rehabilitation.mp. [mp=title, abstract, descriptors]

129 therapy.mp. [mp=title, abstract, descriptors]

130 128 or 129

131 (neck pain or brachial plexus neuropathies or neck injuries or whiplash or thoracic outlet syndrome or cervical rib syndrome or torticollis

or brachial plexus neuritis).mp. [mp=title, abstract, descriptors]

132 (temporomandibular joint disorder or temporomandibular joint dysfunction syndrome).mp. [mp=title, abstract, descriptors]

133 (myofascial pain syndromes or "sprains and strains").mp. [mp=title, abstract, descriptors]

134 (radiculopathy or osteophytosis or neuritis or polyradiculopathy or arthritis or fibromyalgia or spondylitis or spondylosis or spondylolysis or spondylolisthesis).mp.

135 131 or 132 or 133 or 134

136 130 and 135

137 54 and 136

138 (acupuncture or chiropractic).mp. [mp=title, abstract, descriptors]

139 musculoskeletal manipulation*.mp. [mp=title, abstract, descriptors]

140 massage.mp. [mp=title, abstract, descriptors]

141 mobili?ation.mp. [mp=title, abstract, descriptors]

142 acupuncture therapy.mp. [mp=title, abstract, descriptors]

143 (acupuncture or acu‐puncture or needling or acupressure or mox?bustion).mp. [mp=title, abstract, descriptors]

144 ((neck or spine or spinal or cervical or chiropractic* or musculoskeletal*) adj3 (adjust* or manipulat* or mobiliz* or mobilis*)).mp. [mp=title, abstract, descriptors]

145 (manual adj therap*).mp. [mp=title, abstract, descriptors]

146 (manipulati* adj (therap* or medicine)).mp. [mp=title, abstract, descriptors]

147 (massag* or reflexolog* or rolfing or zone therap*).mp. [mp=title, abstract, descriptors]

148 Nimmo.mp. [mp=title, abstract, descriptors]

149 vibration therapy.mp. [mp=title, abstract, descriptors]

150 (vibration adj5 (therap* or treatment*)).mp. [mp=title, abstract, descriptors]

151 (ChihYa or Shiatsu or Shiatzu or ZhiYa).mp. [mp=title, abstract, descriptors]

152 (flexion adj2 distraction*).mp. [mp=title, abstract,descriptors]

153 (myofascial adj3 (release or therap*)).mp. [mp=title, abstract, descriptors]

154 muscle energy technique*.mp. [mp=title, abstract, descriptors]

155 trigger point.mp. [mp=title, abstract, descriptors]

156 proprioceptive neuromuscular facilitation*.mp. [mp=title, abstract, descriptors]

157 cyriax friction.mp. [mp=title, abstract, descriptors]

158 (lomilomi or lomi‐lomi or trager).mp. [mp=title, abstract, descriptors]

159 aston patterning.mp. [mp=title, abstract, descriptors]

160 (strain adj counterstrain).mp. [mp=title, abstract, descriptors]

161 (craniosacraltherap* or cranio‐sacral therap* or craniosacral therap*).mp. [mp=title, abstract, descriptors]

162 (amma or ammo or effleuurage or effleurage or petrissage or hacking or tapotment).mp. [mp=title, abstract, descriptors]

163 complementary therapies.mp. [mp=title, abstract, descriptors]

164 ((complement* or alternat* or osteopathic*) adj (therap* or medicine)).mp. [mp=title, abstract, descriptors]

165 (Tui Na or Tuina).mp. [mp=title, abstract, descriptors]

166 or/138‐165

167 127 and 166

168 136 or 137 or 167

169 (animals not (animals and humans)).mp. [mp=title, abstract, descriptors]

170 168 not 169

171 randomised controlled trial*.mp. [mp=title, abstract, descriptors]

172 controlled clinical trial*.mp. [mp=title, abstract, descriptors]

173 (random* or sham or placebo*).mp. [mp=title, abstract, descriptors]

174 placebos.mp. [mp=title, abstract, descriptors]

175 random allocation.mp. [mp=title, abstract, descriptors]

176 single blind method.mp. [mp=title, abstract, descriptors]

177 double blind method.mp. [mp=title, abstract, descriptors]

178 ((singl* or doubl* or trebl* or tripl*) adj25 (blind* or dumm* or mask*)).mp. [mp=title, abstract, descriptors]

179 (rct or rcts).mp. [mp=title, abstract, descriptors]

180 (control* adj2 (study or studies or trial*)).mp. [mp=title, abstract, descriptors]

181 or/171‐180

182 170 and 181

183 (guideline* or practice guideline*).mp. [mp=title, abstract, descriptors]

184 (guideline* or guidance* or recommendation*).ti.

185 consensus.ti.

186 183 or 184 or 185

187 170 and 186

188 meta‐analysis.mp. [mp=title, abstract, descriptors]

189 (metaanaly* or meta analy* or met analy* or metanaly*).mp. [mp=title, abstract, descriptors]

190 (collaborative research or collaborative review* or collaborative overview*).mp. [mp=title, abstract, descriptors]

191 (integrative research or integrative review* or integrative overview*).mp. [mp=title, abstract, descriptors]

192 (quantitative adj3 (research or review* or overview*)).mp. [mp=title, abstract, descriptors]

193 (research integration or research overview*).mp. [mp=title, abstract, descriptors]

194 (systematic* adj3 (review* or overview*)).mp. [mp=title, abstract, descriptors]

195 (methodologic* adj3 (review* or overview*)).mp. [mp=title, abstract, descriptors]

196 technology assessment biomedical.mp. [mp=title, abstract, descriptors]

197 (hta or thas or technology assessment*).mp. [mp=title, abstract, descriptors]

198 ((hand adj2 search*) or (manual* adj search*)).mp. [mp=title, abstract, descriptors]

199 ((electronic adj database*) or (bibliographic* adj database*)).mp. [mp=title, abstract, descriptors]

200 ((data adj2 abstract*) or (data adj2 extract*)).mp. [mp=title, abstract, descriptors]

201 (analys* adj3 (pool or pooled or pooling)).mp. [mp=title, abstract, descriptors]

202 mantel haenszel.mp. [mp=title, abstract, descriptors]

203 (cochrane or pubmed or pub med or medline or embase or psycinfo or psyclit or psychinfo or psychlit or cinahl or science citation index).ab.

204 or/188‐203

205 170 and 204

206 182 or 187 or 205

207 limit 206 to yr="2009 ‐Current"

Index to chiropractic literature

S1 Subject: "BACK PAIN" OR "BACK INJURIES" OR "NECK INJURIES" OR "NECK PAIN" OR "SPINAL DISEASES" OR "SPINAL INJURIES" OR "SCIATICA" OR All Fields:backpain* or backache* OR "back pain" OR "back ache" OR "back pains" OR "back aches" OR neckpain* OR neckache* OR "neck pain" OR neck ache" OR "neck pains" OR "neck aches" OR All Fields:Spondylolys* or spondylolisthes* or Spondylisthes* or Discitis or diskitis or Spondylod* OR Sciatica OR ischialgia* OR cervicalgia OR Cervicodynia

S2 All Fields:whiplash* or "whip lash" OR "whip lashes" or radiculomyelopath* or "radiculo‐myelopathy" OR "radiculo‐myelopathies" OR All Fields:"failed back" or "back surgery syndrome" or "back surgery syndromes" or FBSS OR All Fields:lumbago or dorsalgia or "myofascial pain" OR "myofascial ache"

S3 All Fields:"cervical pain" OR "cervical ache" OR "vertebrogenic pain syndrome" OR "vertebrogenic pain syndromes" OR All Fields:"degenerated disk" OR "degenerative disk" OR "degenerated disks" OR "degenerative disks" OR All Fields:"degenerated disc" OR "degenerative disc OR "degenerated discs" OR "degenerative discs"

S4 All Fields:"prolapsed disk" OR "prolapsed disks" OR "prolapsed disc" OR "prolapsed discs" OR "disk prolapse" OR "disc prolapse" "herniated disk" OR "herniated disks" OR "herniated disc" OR "herniated discs" OR All Fields:"displaced disk" OR "displaced disks" OR "displaced disc" OR "displaced discs" OR "osteoporotic compression fracture" OR "osteoporotic compression fractures" OR All Fields::"lumbar stenosis" OR "lumbar stenoses" OR "spinal stenosis" OR "spinal stenoses" OR "cervicogenic headache" OR "cervicogenic headaches" OR "cervico‐genic headache" OR "cervico‐genic headaches"

S5 All Fields:radiculomyelopathy OR radiculomyelopathies OR "radiculo‐myelopathy" OR "radiculo‐myelopathies" OR All Fields:"Zygapophyseal joint syndrome" OR "Zygapophyseal joint syndromes" OR "Z‐joint syndrome" OR "Z‐joint syndromes" OR "facet joint syndrome" OR "facet joint syndromes" OR All Fields:"thoracic pain" OR "thoracic ache" OR "spinal pain" OR "spinal ache" OR "lumbar pain" OR "lumbar ache"

S6 S1 OR S2 OR S3 OR S4 OR S5

S7 Subject:"ACUPUNCTURE" OR "ACUPRESSURE" OR "ACUPUNCTURE THERAPY" OR "ELECTROACUPUNCTURE" OR "MANIPULATION, LUMBAR" OR "MANIPULATION, CERVICAL" OR "MANIPULATION, CHIROPRACTIC" OR "MANIPULATION, SPINAL" OR "MANIPULATION, THORACIC" OR Subject:"MASSAGE" OR "CHIROPRACTIC" OR All Fields:acupuncture or "acu‐puncture" or electroacupuncture or "electro‐acupuncture" or "electric acupuncture" or "electric acu‐puncture" or needling or acupressure or "acu‐pressure" or moxibustion or moxabustion

S8 All Fields:"manual therapy" OR "manual therapies" OR massag* or reflexolog* or rolfing or "zone therapy" or "zone therapies" OR All Fields:"Chih Ya" or Shiatsu or Shiatzu or "Zhi Ya" or "Flexion distraction" OR "Trigger point" OR "Trigger points" OR "Proprioceptive Neuromuscular Facilitation" OR "Proprioceptive Neuromuscular Facilitations" OR All Fields:"myofascial release" or "myofascial therapy" OR "myofascial therapies" OR "Muscle energy technique" OR "Muscle energy techniques" OR "Cyriax Friction"

S9 All Fields:: Lomilomi or "lomi‐lomi" or trager or "Aston patterning" or "Strain counterstrain" or "Alexander technique" or "Alexander techniques" or "Tui Na" or Tuina OR All Fields:Craniosacral Therapy" or "Craniosacral Therapies" or "Cranio‐sacral Therapy"or "Cranio‐sacral Therapies" or amma or ammo or Effleurage or Petrissage or hacking or Tapotment OR All Fields:manipulat* or mobiliz* or mobilis*

S10 All Fields:"complementary therapy" OR "complementary therapies" OR "complementary medicine" OR All Fields:"alternative therapy" OR "alternative therapies" OR "alternative medicine" OR All Fields:"osteopathic therapy" OR "osteopathic therapies" OR "osteopathic medicine"

S11 S7 OR S8 OR S9 OR S10

S12 S6 AND S11

S13, Publication Type:Randomised Controlled Trial

S14 Subject:"RANDOMISED CONTROLLED TRIALS AS TOPIC" OR "CONTROLLED CLINICAL TRIALS" OR "PLACEBOS" OR All Fields:random* or sham or placebo* or RCT or RCTs or CCT or CCTs OR All Fields:"controlled clinical trial" or "controlled clinical trials" or "controlled study" or "controlled studies" or "control study" or "controlled studies"

S15 S12 AND S14

S16 S13 OR S15

Appendix 3. Imputation for missing data

1. The preliminary assumption made for imputation of missing values was that data were missing completely at random (Little 1987). In other words, it was assumed that data were not missing due to some factors confounded with the treatment effect.

Since information was solely available on change scores:

• change score treatment (T) is the difference between follow‐up treatment pain score (mFT) and baseline treatment pain score (mBT);

• change score control (C) is the difference between follow‐up control pain score (mFC) and pre‐baseline‐control pain score (mBC).

	Baseline	Follow‐up	Difference
Treatment	mBT	mFT	T = mFT ‐ mBT
Control	mBC	mFC	C = mFC ‐ mBC
	EB = mBT ‐ mBC	EF = mFT ‐ mFC	E = T ‐ C

If EB = 0 then EF is equal to E.

The mean difference was calculated with the assumption that there were no baseline differences in scores (EB = 00). For the conversion, the mean post‐score difference was assumed to be due to a difference in post‐score values (EF), which then equals the post follow‐up difference (E). The related SD was imputed from neighbouring trials by averaging their SD for the treatment and control group.

2. We denoted transformation of data from 10‐point scales to 100‐point scales with the assumption that there were no differences between type of tool used to measure pain (i.e. NRS 0 to 10 transformed to NRS 0 to 100, VAS 0 to 10, transformed to VAS 0 to 100; NDI 0 to 50 transformed to NDI 0 to 100). We indicated transformations in the footnote as * transformation. 
3. Dr. CH Goldsmith (senior statistician, Simon Fraser University) on 2017‐03‐04 for Cervical Overview Group (COG) Electrotherapy Review developed computations to determine OIS calculations using PASS15 (power and sample size) software for:
a) VAS (0 to 100) — COG1 
b) VAS (0 to 10) — COG2 
This calculation identifies that the optimal information size with a sample size closer to n = 250 to 300 is ideal, a sample size closer to n = 100 to 150 is marginal and one that is n < 100 is extremely imprecise. Please contact grossa@mcmaster.ca for a copy.

Data and analyses

Comparison 1. Massage compared with placebo or sham.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Pain (short‐term)	8	403	Mean Difference (IV, Random, 95% CI)	‐3.43 [‐8.16, 1.29]
1.1.1 High‐dose	1	54	Mean Difference (IV, Random, 95% CI)	‐16.20 [‐26.42, ‐5.98]
1.1.2 Medium‐dose	2	128	Mean Difference (IV, Random, 95% CI)	0.52 [‐9.14, 10.19]
1.1.3 Low‐dose (single session)	5	221	Mean Difference (IV, Random, 95% CI)	‐2.42 [‐6.72, 1.88]
1.2 Pain (dichotomous data; short‐term)	3	198	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.62, 1.30]
1.2.1 High‐dose	1	54	Risk Ratio (M‐H, Random, 95% CI)	0.44 [0.22, 0.89]
1.2.2 Medium‐dose	1	114	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.88, 1.39]
1.2.3 Low‐dose (single session)	1	30	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.76, 1.32]
1.3 Function‐disability (short‐term)	2	68	Mean Difference (IV, Random, 95% CI)	‐9.69 [‐17.57, ‐1.81]
1.3.1 High‐dose	1	54	Mean Difference (IV, Random, 95% CI)	‐10.80 [‐19.47, ‐2.13]
1.3.2 Medium‐dose	1	14	Mean Difference (IV, Random, 95% CI)	‐4.40 [‐23.32, 14.52]
1.4 Participant‐reported treatment success (short‐term)	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
1.4.1 High‐dose	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
1.5 Participant‐reported treatment success at short‐term (dichotomous data)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.5.1 Low‐dose (single session)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.6 Health‐related quality of life (short‐term)	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
1.6.1 High‐dose	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
1.7 Minor adverse events (dichotomous data; short‐term)	2	175	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.08, 11.55]
1.7.1 Medium‐dose	1	114	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.11, 0.97]
1.7.2 Low‐dose (single session)	1	61	Risk Ratio (M‐H, Random, 95% CI)	4.13 [0.49, 34.89]
1.8 Pain (short‐term): subgroup analysis by disorder classification	8	403	Mean Difference (IV, Random, 95% CI)	‐3.43 [‐8.16, 1.29]
1.8.1 Neck pain without radiculopathy	7	363	Mean Difference (IV, Random, 95% CI)	‐2.60 [‐8.31, 3.11]
1.8.2 Whiplash‐associated disorder	1	40	Mean Difference (IV, Random, 95% CI)	‐6.50 [‐13.10, 0.10]

Comparison 2. Massage compared with no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Pain (short‐term)	6	299	Mean Difference (IV, Random, 95% CI)	‐24.24 [‐28.48, ‐19.99]
2.1.1 High‐dose	1	74	Mean Difference (IV, Random, 95% CI)	‐28.00 [‐36.31, ‐19.69]
2.1.2 Medium‐dose	5	225	Mean Difference (IV, Random, 95% CI)	‐22.91 [‐27.85, ‐17.96]
2.2 Pain (dichotomous data; short‐term): subgroup analysis by dosage	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.2.1 3 x 60 minutes/week conventional massage versus no treatment (short‐term)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.2.2 2 x 60 minutes/week conventional massage versus no treatment (short‐term)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.2.3 3 x 30 minutes/week conventional massage versus no treatment (short‐term)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.2.4 2 x 30 minutes/week conventional massage versus no treatment (short‐term)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.2.5 1 x 60 minutes/week conventional massage versus no treatment (short‐term)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.3 Function‐disability (short‐term)	6	300	Mean Difference (IV, Random, 95% CI)	‐9.51 [‐13.68, ‐5.33]
2.3.1 High‐dose	2	94	Mean Difference (IV, Random, 95% CI)	‐15.07 [‐27.66, ‐2.48]
2.3.2 Medium‐dose	4	206	Mean Difference (IV, Random, 95% CI)	‐7.72 [‐11.95, ‐3.50]
2.4 Health‐related quality of life (short‐term)	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
2.4.1 Medium‐dose	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

Comparison 3. Massage as adjuvant to a co‐treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Pain (short‐term)	12	795	Mean Difference (IV, Random, 95% CI)	‐7.29 [‐11.43, ‐3.16]
3.1.1 Massage as an adjunct to a passive treatment method	7	305	Mean Difference (IV, Random, 95% CI)	‐11.37 [‐19.10, ‐3.64]
3.1.2 Massage as an adjunct to an active (including exercise) treatment method	5	490	Mean Difference (IV, Random, 95% CI)	‐3.58 [‐6.05, ‐1.12]
3.2 Pain (long‐term)	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.2.1 Medium‐dose adjuvant to an active (including exercise) method	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.3 Pain (dichotomous data; short‐term)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
3.3.1 Medium‐dose adjuvant to a passive method	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
3.4 Function‐disability (short‐term)	7	532	Std. Mean Difference (IV, Random, 95% CI)	‐0.46 [‐0.92, ‐0.01]
3.4.1 Massage adjuvant to a passive method	3	109	Std. Mean Difference (IV, Random, 95% CI)	‐0.45 [‐1.32, 0.43]
3.4.2 Massage adjuvant to an active (including exercise) method	4	423	Std. Mean Difference (IV, Random, 95% CI)	‐0.46 [‐1.05, 0.13]
3.5 Function‐disability (long‐term)	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.5.1 Medium‐dose adjuvant to an active (including exercise) method	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.6 Function‐disability (other data; short‐term)	1		Other data	No numeric data
3.6.1 Medium‐dose adjuvant to a passive method	1		Other data	No numeric data
3.7 Health‐related quality of life (short‐term)	2		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
3.7.1 Medium‐dose adjuvant to an active (including exercise) method	2	352	Std. Mean Difference (IV, Random, 95% CI)	0.10 [‐0.31, 0.52]
3.8 Health‐related quality of life (long‐term)	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.8.1 Medium‐dose adjuvant to an active (including exercise) method	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.9 Pain (short‐term): subgroup analysis by dose	12	795	Mean Difference (IV, Random, 95% CI)	‐7.29 [‐11.43, ‐3.16]
3.9.1 Medium‐dose	11	765	Mean Difference (IV, Random, 95% CI)	‐7.92 [‐12.52, ‐3.32]
3.9.2 Low‐dose	1	30	Mean Difference (IV, Random, 95% CI)	‐2.70 [‐8.87, 3.47]
3.10 Pain (short‐term): subgroup analysis by disorder classification	11	495	Mean Difference (IV, Random, 95% CI)	‐8.03 [‐12.79, ‐3.28]
3.10.1 Radiculopathy	1	26	Mean Difference (IV, Random, 95% CI)	4.00 [‐4.07, 12.07]
3.10.2 Cervicogenic headache	2	106	Mean Difference (IV, Random, 95% CI)	‐19.96 [‐40.51, 0.60]
3.10.3 Neck pain without radiculopathy	8	363	Mean Difference (IV, Random, 95% CI)	‐6.99 [‐10.44, ‐3.53]
3.11 Pain (short‐term): subgroup analysis by symptom duration	8	635	Mean Difference (IV, Random, 95% CI)	‐10.04 [‐15.32, ‐4.77]
3.11.1 Subacute	2	116	Mean Difference (IV, Random, 95% CI)	‐5.27 [‐9.04, ‐1.49]
3.11.2 Chronic	6	519	Mean Difference (IV, Random, 95% CI)	‐11.93 [‐19.68, ‐4.17]
3.12 Function‐disability (short‐term): subgroup analysis by dose	7	532	Std. Mean Difference (IV, Random, 95% CI)	‐0.46 [‐0.92, ‐0.01]
3.12.1 Medium‐dose	6	502	Std. Mean Difference (IV, Random, 95% CI)	‐0.53 [‐1.05, ‐0.02]
3.12.2 Low‐dose (single session)	1	30	Std. Mean Difference (IV, Random, 95% CI)	0.01 [‐0.70, 0.73]
3.13 Function‐disability (short‐term): subgroup analysis by disorder classification	6	235	Std. Mean Difference (IV, Random, 95% CI)	‐0.57 [‐1.05, ‐0.08]
3.13.1 Neck pain without radiculopathy	5	209	Std. Mean Difference (IV, Random, 95% CI)	‐0.73 [‐1.18, ‐0.28]
3.13.2 Radiculopathy	1	26	Std. Mean Difference (IV, Random, 95% CI)	0.29 [‐0.48, 1.06]
3.14 Function‐disability (short‐term): subgroup analysis by disorder duration	4	436	Std. Mean Difference (IV, Random, 95% CI)	‐0.53 [‐1.12, 0.05]
3.14.1 Subacute	2	116	Std. Mean Difference (IV, Random, 95% CI)	‐0.91 [‐1.35, ‐0.47]
3.14.2 Chronic	2	320	Std. Mean Difference (IV, Random, 95% CI)	‐0.03 [‐0.25, 0.19]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abaspour 2020.

Study characteristics
Methods	Type of trial: RCT, restricted randomisation Number analysed/randomised: 30/30 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: none Declaration of interest: no competing interests
Participants	Disorder: cervicogenic headache Radicular signs/symptoms: absent Sex: 80% female Age: mean (SD): Arm 1 29.93 (8.53), Arm 2 28.87 (11.25) Severity: (Headache Index 0 to 100, mean (SD)): Arm 1 45.77 (22.20), Arm 2 40.92 (22.67) Duration of complaints: more than 3 months Setting: neurology or physiotherapy clinics Country: Iran
Interventions	INDEX TREATMENT Arm 1: muscle energy technique in cervical muscles and infrared radiation. Activity: the line of the participant’s look was towards the contraction side in the hold time of contraction and towards the opposite side in the relaxation phase. The technique was applied with 20% to 30% of maximal voluntary contraction in a supine position. Each technique was repeated 3 times for both sides; Provider: 1 expert physiotherapist; Mode: manual application and modality; Dose: 9 minutes; Monitoring: not reported; Route: cervical muscles COMPARISON TREATMENT Arm 2: infrared radiation. Activity: in the prone position, infrared radiation from a single flame light (a power of 250 Watts and voltage 230 to 250 volts) was adjusted perpendicular to the exposed upper cervical region; Provider: an expert physiotherapist; Mode: modality; Dose: 15 minutes; Monitoring: not reported; Route: cervical spine Treatment schedule: 3/week over 2 weeks Duration of follow‐up: 1 day immediately post‐treatment, 2 weeks post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: Headache Index including pain intensity, frequency and duration; 0 to 100 Baseline mean (SD): Arm 1 45.77 (22.20), Arm 2 40.92 (22.67) Reported results: no significant difference between groups in the treatment process and the repeated measure test (P = 0.056) MD (Arm 1 vs Arm 2) at short‐term follow‐up: ‐8.78, 95% CI ‐24.27 to 6.71 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	We wrote to the author for unpublished data for the Headache Disability Index on 18 February 2023.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: Page 2, "Restricted randomization method (Random allocation rule) was used for randomization"; page 4," … but two persons withdrew, and randomization was performed in 30 subjects" Comment: there was insufficient information provided. It was not clear how randomisation was conducted.
Allocation concealment (selection bias)	Unclear risk	Quote: Page 2, “Restricted randomization method (Random allocation rule) was used for randomization in this study” Comment: there was insufficient information about the allocation process described.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of intervention and the study design.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of intervention and the study design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible due to the study design; patients were the outcome assessors. A patient‐reported outcome measure (Headache Index) was used.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 4, "... but two persons withdrew, and randomization was performed ..." Comment: there was insufficient description about procedures and reasons for dropouts.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Comment: the authors did not specify how many participants were analysed.
Selective reporting (reporting bias)	Low risk	Quote: Page 8, "IRCT2015031421459N1". Comment: consistent with registered protocol.
Baseline similarities	Unclear risk	Quote: Page 4, Table 1. Comment: the duration of pain and function‐disability was not reported.
Similar or avoided co‐interventions	Unclear risk	Comment: there was no mention of tracking other co‐interventions including medication.
Acceptable compliance	Unclear risk	Comment: compliance was not reported.
Similar outcome assessment	Low risk	Quote: Page 2, "The subjects were evaluated in three times at the beginning of the first session, after the end of the sixth session and two weeks after the end of the treatment." Comment: there was similar outcome assessment at baseline, immediate post‐treatment and 2 weeks after the intervention was completed.
Other potential sources of bias	Unclear risk	Funding source: none; conflict of Interest: none; outcome measurement: validity of Headache Index was not described including its minimally important difference. Other: the study appeared to be free of other sources of bias.

Afzal 2019.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 39/40 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: none Declaration of interest: no competing interests
Participants	Disorder: neck pain with radicular symptoms Radicular signs/symptoms: present 100% Sex: 70% female Age: mean (SD): Arm 1 40.95 (7.32), Arm 2 42.41 (6.86), Arm 3 42.50 (5.77) Severity: numeric pain rating scale 0 to 10, mean (SD): Arm 1 7.5 (0.67), Arm 2 7.1 (1.03), Arm 3 7.5 (0.89) Duration of complaints: not reported Setting: Fauji Foundation Hospital Country: Pakistan
Interventions	INDEX TREATMENT Arm 1: cervical manual traction adjunct to mobilisation. Activity: cervical manual traction and mobilisation; Provider: an experienced orthopaedic manual physical therapist; Mode: manual application; Dose: 10 minutes traction, 3 sets of 10 repetitions of mobilisation; Monitoring: not reported; Route: cervical spine COMPARISON TREATMENT Arm 2: mobilisation. Activity: the intervertebral foramen technique with movement; pulling over the restricted area of the neck with performing movement into opening; Provider: an experienced orthopaedic manual physical therapist; Mode: manual application; Dose: 3 sets of 10 repetitions; Monitoring: not reported; Route: cervical spine Arm 3: cervical manual traction. Activity: the patient's head was cradled by a physiotherapist from chin and the occiput in a supine position, and traction force applied in 25 degree neck flexion; Provider: an experienced orthopaedic manual physical therapist; Mode: manual traction; Dose: 10 minutes with 10 seconds pull and 5 seconds rest; Monitoring: not reported; Route: cervical spine Treatment schedule: 3/week for 3 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: hot pack for 15 minutes
Outcomes	PAIN: numeric pain rating scale 0 to 10 Baseline mean: Arm 1 7.5, Arm 2 7.1, Arm 3 7.5 Reported results: in pre‐ and post‐analysis of different groups, pain intensity showed significant result (P < 0.05) MD (Arm 1 vs Arm 2) at immediate post‐treatment: 5.80, 95% CI ‐0.71 to 12.31 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean: Arm 1 22.4, Arm 2 19.5, Arm 3 22.06 Reported results: not reported MD (Arm 1 vs Arm 2) at immediate post‐treatment: 1.10, 95% CI ‐1.62 to 3.82 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Arm 2 and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	High risk	Quote: Page 1238 "The subjects were recruited using nonprobability purposive sampling technique and were divided into groups I, II and III using the sealed envelope method." Comment: purposive sampling technique that uses subjective judgement.
Allocation concealment (selection bias)	Unclear risk	Comment: no information was provided about allocation concealment.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the treatment. However, we are not sure if patients were naive; the duration of treatment may alert the patient to differences.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the study design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: patients were the outcome assessors due to use of participant‐reported outcome measures (i.e. Neck Disability Index, numeric pain rating scale).
Incomplete outcome data (attrition bias)	Unclear risk	Quote: Page 1239, Figure 1. Comment: there was 1 dropout (lost to follow‐up) in Arm 3 observed, however there was no description of dropouts provided.
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Page 1239 figure 1. Comment: no intention‐to‐treat analysis was performed.
Selective reporting (reporting bias)	Unclear risk	Comment: there was no study protocol or trial registration reported.
Baseline similarities	Unclear risk	Quote: Page 1239, Table 2. Comment: there was insufficient information provided to judge comparability between groups, including duration of pain.
Similar or avoided co‐interventions	Unclear risk	Comment: there was no relevant information reported.
Acceptable compliance	Unclear risk	Comment: there was no relevant information provided.
Similar outcome assessment	Low risk	Quote: Page 1238, right "Assessment was made at baseline and after completion of 3 consecutive weeks of intervention." Comment: similar assessments were made at baseline and 3 weeks post‐treatment.
Other potential sources of bias	Low risk	Funding source: none; conflict of interest: none; outcome measurement: Neck Disability Index, numeric pain rating scale ‐ both outcomes are valid. Other: the study appeared to be free of other sources of bias.

Antolinos‐Campillo 2014.

Study characteristics
Methods	Type of trial: RCT parallel group design Number analysed/randomised: 40/40 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: chronic whiplash‐associated disorder (grade I and II) Radicular signs/symptoms: not reported Sex: 42% female Age: mean (SD): Arm 1 36 (5.2), Control 36 (5.3) Severity: visual analogue scale 0 to 100, mean (SD): Arm 1 52.0 (9.0), Control 56.2 (10.6) Duration of complaints: > 3 months Setting: private practice Country: Spain
Interventions	INDEX TREATMENT Arm 1: suboccipital muscle inhibition. Activity: suboccipital muscle inhibition was performed with the participant lying supine on a bed; the therapist placed their hands under the participant’s head and placed their fingers between the suboccipital condyles and C2; constant painless pressure was applied upwards and towards the therapist; participants were asked to keep their eyes closed and to avoid eye movements; Provider: the therapist and the evaluator were senior physical therapists with 6 years of experience in the field of manual therapy; Mode: manual application with hands; Dose: 4 minutes; Monitoring: not reported; Route: suboccipital muscles COMPARISON TREATMENT Placebo: mock treatment. Activity: active range of motion of the hip and knee; on opposite side of the body to the more affected side; Provider: the therapist and the evaluator were senior physical therapists with over 6 years of experience in the field of manual therapy; Mode: active range of motion; Dose: 4 minutes; Monitoring: not reported; Route: hip and knee Treatment schedule: 1 session Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 100 Baseline mean: Arm 1 52.0, Control 56.2 Reported results: no statistical significance between groups MD (Arm 1 vs Control) at immediate post‐treatment: ‐6.50, 95% CI ‐13.10 to 0.10 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐REALTED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: Page 250, "using a randomised number table designed by an Internet website (randomized.com)" Comment: Oour senior statistician questions the accuracy of this website (randomized.com). Inconsistency were noted with the registered protocol as follows: "were recruited and distributed into two study groups: intervention group (IG) (n = 20) and control group (CG) (n = 20)". "A randomised [using a randomised number table designed by an Internet website (randomized.com)]"... "The subjects were selected according to non‐probabilistic convenience sampling techniques."
Allocation concealment (selection bias)	Unclear risk	Quote: page 250, "An external consultant prevented access to the sequence for those participating in the study." "Subjects and evaluators, who collected or analysed data, were unaware of the treatment allocation group." Comment: insufficient information about allocation process; no data on method of concealment of allocation (including opaque, sealed envelopes, sequential numbering) described.
Blinding ‐ patient (performance bias)	High risk	Quote: page 252‐3, "The sham (placebo) intervention consisted of performing active movement of flexion/extension of the hip and knee joints on the opposite side" Comment: blinding was not possible for participants; suboccipital muscle inhibition massage and active range of motion in hip and knee are distinguishable. The success of blinding was not tested or reported.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the study design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible due to the study design and the nature of the intervention; hip and knee joint active movement in placebo group.
Incomplete outcome data (attrition bias)	Low risk	Comment: this was a pre‐post design; no missing outcome data.
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: this was a pre‐post design with no attrition possible.
Selective reporting (reporting bias)	Low risk	Comment: consistent with registered protocol, except inconsistency with randomisation process, minor deviation from study protocol and described in trial.
Baseline similarities	High risk	Quote: Table 1. Comment: differences in elbow extension were noted.
Similar or avoided co‐interventions	Low risk	Comment: this was not reported but due to the pre‐post design, it was not possible.
Acceptable compliance	Low risk	Comment: this was a pre‐post design, attending one session.
Similar outcome assessment	Low risk	Comment: this was a pre‐post design with one outcome assessment.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale, valid; other: the study appears to be free of other sources of bias. No major threat to the study’s internal validity.

Blikstad 2008.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 45/45 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific cervical disorder/mechanical neck pain Radicular signs/symptoms: not reported Sex: 55.5% female Age: mean (SD): Arm 1 22.60 (2.38), Arm 2 23.90 (3.92), Control 24.90 (5.44) Severity: numeric rating scale 0 to 10, mean (SD): Arm 1 4.6 (0.50), Arm 2 4.6 (0.63), Control 4.7 (0.96) Duration of complaints: 4 weeks but no longer than 12 weeks Setting: Anglo‐European College of Chiropractic, Bournemouth Country: UK
Interventions	INDEX TREATMENT Arm 1: myofascial band therapy. Activity: firm thumb pressure in a slow stroking motion from lateral shoulder to mastoid process along the upper trapezius muscle and through active trigger points; Provider: chiropractor; Mode: soft tissue massage with hands; Dose: 1 session for 1 minute; Monitoring: not reported; Route: upper fibres of trapezius COMPARISON TREATMENT Arm 2: activator trigger point (manipulation). Activity: hand held device was placed perpendicular over the trigger point using a force setting of 3 (170 Newton); Provider: chiropractor; Mode: using activator adjusting instrument. Dosage: 3 (170 Newton), 10 thrusts, 1 thrust per session; Monitoring: not reported; Route: trigger points on upper fibres of trapezius Placebo: sham ultrasound. Activity: Medi Link Systems ultrasound machine was used. Patients were informed that a pulsed ultrasound was going to be applied, they were notified that they should not feel any heat or pain with the ultrasound; Provider: chiropractor; Dose: 2 minutes; Monitoring: not reported; Route: upper fibres of trapezius Treatment schedule: 1 session Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: numeric rating scale 0 to 10 Baseline mean: Arm 1 4.6, Arm 2 4.6, Control 4.7 Reported results: significant results in pain reduction favouring Arm 2 RR (Arm 1 vs Control) at immediate post‐treatment: 1.00, 95% CI 0.76 to 1.32 NNTB = 3 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (sham) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 25, left paragraph 1, "The randomisation scheme was generated by using the website Randomisation.com. To ensure equal numbers in the groups, subjects were randomised in blocks of three." Comment: an acceptable randomisation process was observed.
Allocation concealment (selection bias)	Low risk	Quote: Page 25, left paragraph 1, "Sealed opaque envelopes were prepared containing the assigned treatment and numbered consecutively. Subjects were allocated to the next available envelope number." Comment: assignment generated by an independent person not responsible for determining the eligibility of the patients.
Blinding ‐ patient (performance bias)	High risk	Quote: page 25, "A detuned Medi‐Link Systems ultrasound machine from Electro‐Medical Supplies (Greenham) Ltd. was used as the sham control..." Comment: although the participants were randomised to 3 groups, they would be able to tell if they were receiving activator trigger point therapy, myofascial band therapy or sham ultrasound (control group) based on the information provided to them in order to join the study. The success of blinding was not tested amongst the patients and was not reported to be successful. a) Placebo described: the "sham ultrasound" is not similar to the massage intervention. b) Contextual factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described, but the hands‐on element of the massage group was not replicated. c) Patient‐factors: not reported. d) Practitioner factors: "the clinician"; details not reported. e) Was the success of blinding tested amongst patients and successful?: not reported
Blinding ‐ care provider (performance bias)	High risk	Comment: it was hard to blind the caregiver during a study examining manual therapy techniques.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Quote: page 25, "A detuned Medi‐Link Systems ultrasound machine from Electro‐Medical Supplies (Greenham) Ltd. was used as the sham control..." Comment: the participant was the outcome assessor due to participant‐reported outcome measures, but the participant was not credibly blinded in this study due to the type of sham.
Incomplete outcome data (attrition bias)	Low risk	Comment: all the participants completed the study.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Tables 3 to 5. Comment: clinically significant improvement was determined using an odds ratio and number needed to treat with 95% confidence intervals.
Selective reporting (reporting bias)	Unclear risk	Comment: no previously published (and referenced in current paper) protocol initiated prior to the start of the study.
Baseline similarities	Low risk	Quote: Table 1. Comment: characteristics including pain intensity and duration, and disability of participants were similar in both groups.
Similar or avoided co‐interventions	Low risk	Comment: because of the immediate pre‐post design, co‐interventions during the study were not possible.
Acceptable compliance	Low risk	Comment: participant adherence to the intervention was specified in the procedures section of the study.
Similar outcome assessment	Low risk	Quote: Tables 3 to 5. Comment: all outcomes were assessed at the same time for each group. Follow‐up was 5 minutes post‐treatment.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: numeric pain rating scale, valid; other: the study appears to be free of other sources of bias. No major threat to the study’s internal validity.

Briem 2007.

Study characteristics
Methods	Type of trial: RCT parallel group design Number analysed/randomised: 40/40 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific neck pain with/without cervicogenic headache (pain in the cervical region) Radicular signs/symptoms: not reported Sex: 77% female Age: mean: 34.7 Severity: numeric pain rating scale 0 to 10, mean: Arm 1 4.0, Control 3.7 Duration of complaints: more than 6 months Setting: private physiotherapy department Country: Iceland
Interventions	INDEX TREATMENT Arm 1: inhibitive distraction. Activity: patient in supine, therapist placed fingertips onto suboccipital musculotendinous structures just caudal to the superior nuchal line and induce a sustained force in a ventrocranial direction, thus exerting compressive forces as well as distraction to cervical and suboccipital structures. Pressure was applied slowly into perpendicular to longitudinal axis of muscles and tendons. Amount of pressure was at the patient's tolerance as reflected by muscle response; Provider: 1 physical therapist, 10 years experience, 6 years experience with technique; Mode: manual approach, ranged from light pressure and no distraction force applied with the weight of the participant's head partially supported by therapist's thenar eminence to the full weight of the participant's head resting on therapist's fingertips; Dose: 3 to 3.5 minutes one time; Monitoring: not reported; Route: cervical spine COMPARISON TREATMENT Control: placebo. Activity: patient supine and rested their head in palms of clinician for the same duration to mimic treatment position. Participants received effects of touch, warmth and rest; Provider: physical therapist; Mode: manual application; Dose: 3 to 3.5 minutes; Monitoring: not reported; Route: cervical spine Treatment schedule: 1 session, 1 day Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: no other treatment was offered to either group
Outcomes	PAIN: numeric pain rating scale, 0 to 10 Baseline mean: Arm 1 4.0, Control 3.7 Reported results: MD (Arm 1 vs Control) at immediate post‐treatment: 0.70, 95% CI ‐0.57 to 1.97 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Communication: author provided pre‐ and post‐data on pain rating for treatment and control group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Comment: no information on randomisation was provided.
Allocation concealment (selection bias)	High risk	Quote: Page 84, "receptionist supervised a list of 40 consecutive numbers, to which an intervention or placebo treatment had been randomly assigned" Comment: allocation was not concealed.
Blinding ‐ patient (performance bias)	Unclear risk	Quote: page 84, "Those in the control group rested their heads in the palms of the clinician for the same duration to mimic the treatment position as much as possible. In this way, these subjects received the effects of touch, warmth, and rest, without the actual proposed mechanical effects of the experimental ID intervention." Comment: the sham was adequate to ensure blinding. The massage and sham groups are likely indistinguishable for the patients. The success of blinding was not tested amongst the patients and success of blinding was not reported. a) Placebo described: the sham seemed similar to the intervention. b) Contextual‐factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described and the same as the intervention group. c) Patient‐related factors: experience replicated but no additional questionnaire on expectation/therapeutic alliance etc. was reported d) Practitioner‐related factors: reported e) Was the success of blinding tested amongst patients and successful?: not reported
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the study design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	Unclear risk	Comment: the patients may or may not be blinded due to the sham; if so, the patients as the outcome assessor for the self‐reported pain instrument were blinded adequately.
Incomplete outcome data (attrition bias)	Low risk	Comment: all the participants completed the study.
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: this was a pre‐post design with no follow‐up.
Selective reporting (reporting bias)	Unclear risk	Quote: page 83, "The National Bioethics Committee of Iceland and the Data Protection Authority of Iceland approved this study." Comment: the protocol was not previously published (and referenced in current paper); a protocol submission to the Bioethics Committee was initiated prior to the start of the study.
Baseline similarities	Unclear risk	Comment: this was not clearly outlined in the study regarding pain duration; the author accepted a broad range of possible neck conditions for inclusion in the study.
Similar or avoided co‐interventions	Low risk	Comment: this was a pre‐post study design. There was no opportunity for co‐intervention
Acceptable compliance	Low risk	Comment: this was a pre‐post study design.
Similar outcome assessment	Low risk	Comment: numeric pain rating scale and cervical range of motion measurements were collected directly after intervention for both groups.
Other potential sources of bias	Unclear risk	Funding source: the work was supported by a grant from the research fund of the Association of Icelandic Physiotherapists in Iceland; conflict of interest: not reported; outcome measurement: numeric pain rating scale, valid; other: no major threat to the study’s internal validity was observed.

Brück 2021.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 60/60 Intention‐to‐treat analysis: calculated Power analysis: calculated, 90% Funding source: not reported Declaration of interest: no competing interests
Participants	Disorder: mechanical neck pain Radicular signs/symptoms: not reported Sex: 78.33% female Age: mean (SD): Arm 1 48.9 (9.4), Arm 2 50.1 (11.8), Control 1 45.7 (9.4) Severity: (VAS 0 to 10): Arm 1 3.8 (2.1), Arm 2 4.0 (2.0), Control 1 3.7 (2.0) Duration of complaints: Arm 1 13.4 (7.0), Arm 2 13.5 (8.0), Control 1 10.2 (9.2) Setting: private practices for physiotherapy and osteopathy/primary Country: Germany
Interventions	INDEX TREATMENT Arm 1: Fascial treatment. Activity: 8 different fasciae of the neck and cervical region were gently stretched: (1) superficial layer of the cervical fascia, (2) middle layer of the cervical fascia, (3) deep layer of the cervical fascia, (4) nuchal fascia, (5) fascia of trapezius, (6) fascia of sternocleidomastoids, (7) fascia of scalene and (8) fascia of the pharyngeal muscles. Within each treatment, each of the fasciae were treated only once; Provider: 2 osteopathic practitioners with 10 years of experience in physiotherapy, manual therapy and osteopathy; Mode: soft tissue massage with hands; Dose: 20 minutes; Monitoring: not reported; Route: neck, cervical region COMPARISON TREATMENT Arm 2: manual therapy. Activity: a high‐velocity, low‐amplitude manipulation treatment was applied at the cervical and upper thoracic spine. The cervical spine was treated by a commonly used technique in rotation. The thoracic spine was treated by a commonly used “drop technique”. Within each treatment, each restricted segment was treated only once; Provider: 2 osteopathic practitioners with 10 years of experience in physiotherapy, manual therapy and osteopathy; Mode: manual therapy; Dose: 20 minutes; Monitoring: not reported; Route: cervical, upper thoracic spine Control: no intervention; Monitoring: not reported Treatment schedule: 2 weeks, 2 sessions Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean (SD): Arm 1 3.8 (2.1), Arm 2 4.0 (2.0), Control 3.7 (2.0) Reported results: a significant time effect (F = 41.57; P < 0.001; η 2 = 0.42) and a significant time group interaction (F = 10.12; P < 284 0.001; η 2 = 0.26) were observed MD (Arm 1 vs Control) at short‐term follow‐up: ‐22.0, 95% CI ‐34.41 to ‐9.59 FUNCTION/DISABILITY: Neck Pain and Disability Scale 0 to 100 Baseline mean (SD): Arm 1 40.7 (23.3), Arm 2 47.1 (16.3), Control 38.9 (17.4) Reported results: a significant time effect (F = 24.66; P < 0.001; η 2 = 0.30) and a significant time group interaction (F = 8.21; P = 0.001; η 2 = 0.22) were shown MD (Arm 1 vs Control) at short‐term follow‐up: 10.50, 95% CI ‐21.81 to 0.81 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (no intervention) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 999, under randomisation, "randomly allocated by drawing lots to one of the three groups" Comments: appropriate randomisation was performed.
Allocation concealment (selection bias)	Unclear risk	Quote: Page 999, under randomisation, "sealed envelopes were in one box ... envelop was drawn by a blinded assessor." Comment: not enough information was provided about allocation concealment.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the treatment.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the treatment and study design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: patients were the outcome assessors due to use of participant‐reported outcome measures (i.e. Neck Pain and Disability Scale, visual analogue scale).
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 1001, Figure 1. Comment: there was no dropout (loss to follow‐up) reported.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Page 1001, Figure 1. Comment: there was no dropout and all participants in each group were included in the analysis.
Selective reporting (reporting bias)	High risk	Quote: Page 998, under study design, "The study was retrospectively registered in the German Clinical Trials Register (DRKS00010398)." Comment: the study protocol was retrospectively registered, and the number of treatments was reported differently between the study (2 treatment sessions) and the protocol (4 treatment sessions).
Baseline similarities	Low risk	Quote: Page 1001, Table 1. Comment: baseline characteristics between groups were acceptable.
Similar or avoided co‐interventions	Unclear risk	Comment: there was no relevant information reported.
Acceptable compliance	Unclear risk	Comment: there was no relevant information reported.
Similar outcome assessment	Unclear risk	Quote: Page 999, under intervention, "two treatment groups (FT, MT) received two treatment sessions within a time period of two weeks" Comment: control group was assessed after 2 weeks of the treatment period whereas treatment groups were after the last treatment session. The time point of outcome measurement could be different between groups.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: none; outcome measurement: Neck Pain and Disability Scale, visual analogue scale, both outcomes were valid.

Capo‐Juan 2017.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 75/75 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific mechanical cervical disorder (myofacial pain syndrome) Radicular signs/symptoms: not reported Sex: 80% female Age: ranged between 23 and 54 years Severity: numeric pain rating scale 0 to 10, mean (SD): Arm 1 5.36 (0.37), Arm 2 5.32 (0.42), Control 5.04 (0.48) Duration of complaints: not reported Setting: hospital or tertiary Country: Spain
Interventions	INDEX TREATMENT Arm 1: Pressure release. Activity: pressure release on myofascial on the sternoclavicular mastoid muscle; Provider: physiotherapist; Mode: soft tissue massage with hands; Dose: number of sets ‐ not reported, duration of a session ‐ not reported, rest intervals ‐ not reported, order of care ‐ not reported; Monitoring: not reported; Route: sternoclavicular mastoid muscle COMPARISON TREATMENT Arm 2: Kinesiotaping. Activity: kinesiotaping applied to bilateral sternocleidomastoid muscle from distal to proximal at 25% strain; Provider: physiotherapist; Mode: using kinesiotape; Dose: number of sets ‐ not reported, duration of a session ‐ not reported, rest intervals ‐ not reported, order of care ‐ not reported; Monitoring: not reported; Route: neck Control: placebo pressure using algometer. Activity: application of algometric bilateral pressure on sternocleidomastoid muscle for placebo; Provider: physiotherapist; Dose: number of sets ‐ not reported; Monitoring: not reported; Route: sternocleidomastoid muscle Treatment schedule: 1 session Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: numeric pain rating scale 0 to 10 Baseline mean (SD): Arm 1 5.36 (0.37), Arm 2 5.32 (0.42), Control 5.04 (0.48) Reported results: significant difference between control and Arm 2 and between Arm 1 and Arm 2 MD (Arm 1 vs placebo) at immediate post treatment: ‐0.16, 95% CI ‐0.38 to 0.06 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: Short Form‐12, 0 to 47 Baseline mean (SD): Arm 1 24.36 (1.93), Arm 2 25 (2.18), Control 24.40 (2.00) Reported results: quality of life is only significant between control and Arm 2 MD (Arm 1 vs Placebo) at immediate post treatment: ‐2.88, 95% CI ‐3.81 to ‐1.95 PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (placebo) and the risk of bias judgements was based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 70, "patients were assigned to different groups according to an allocation number generated from a random table". Comment: an acceptable randomisation process was observed.
Allocation concealment (selection bias)	Unclear risk	Comment: no information about the allocation process was described.
Blinding ‐ patient (performance bias)	High risk	Quote: page 70, "Application of algometric bilateral pressure for the placebo group (group A)" Comment: not possible due to design for Arm 1 (pressure release) vs placebo (algometric pressure). No detailed context, practitioner, or patient description of sham control was provided. The success of blinding was not tested or reported.
Blinding ‐ care provider (performance bias)	High risk	Comment: not possible due to design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: not possible due to design; there was no description of the sham and so there was no ability to judge if the participant as the outcome assessor was blind.
Incomplete outcome data (attrition bias)	Low risk	Comment: outcome data for second and third appointment were described in tables 2 and 3. No dropouts were observed.
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: outcome data with sample size for the second and third appointment were described in tables 2 and 3. All randomised participants were analysed in each group.
Selective reporting (reporting bias)	Unclear risk	Comment: no reference to a registered trial or published protocol was made.
Baseline similarities	Low risk	Comment: baseline data were described in table 1. Baseline characteristics of participants were similar between groups.
Similar or avoided co‐interventions	Unclear risk	Comment: insufficient information, not described. For the time interval between 1 session of treatment and assessment 1 week later, there was no description of co‐intervention.
Acceptable compliance	Low risk	Comment: the sample size per group was described in tables 2 and 3. One session of pressure release and kinesiotaping was assessed. The number of participants in each treatment session and assessment session was similar between groups.
Similar outcome assessment	Low risk	Comment: outcomes were assessed at baseline and 1 week post treatment.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: numeric pain rating and Short Form‐36, valid; other: the study appears to be free of other sources of bias. No major threat to the study’s internal validity.

Celenay 2016.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 60/62 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific mechanical neck pain Radicular signs/symptoms: not reported Sex: 62% female Age mean (SD): Arm 1 50.10 (10.98), Arm 2 45.20 (10.98) Severity: VAS 0 to 10, mean (SD): Arm 1 3.65 (3.32), Arm 2 4.64 (3.19) Duration of complaints: at least 3 months Setting: university physiotherapy and rehabilitation clinic Country: Turkey
Interventions	INDEX TREATMENT Arm 1: connective tissue massage adjunct to stabilisation exercises. Activity 1: connective tissue massage started with a series of short strokes over the sacrum, lumbar spine and posterolateral pelvis, which were developed into longer paravertebral and subcostal strokes. The treatment was started with the application on the lumbosacral area and once the condition of subcutaneous tissues in the massaged region returned to normal, treatment progressed to the thoracic (scapular and inter‐scapular area) and then cervical (cervical‐occipital area). Connective tissue massage was applied according to the vascular response of the tissue. After 2 to 4 sessions the next region was treated. For each treatment the patient was sitting on a stool in the upright position, with 90 degrees flexion of the hips, knees and ankles. The thighs and feet were fully supported; Provider: an experienced physical therapist; Mode: manual application with hands; Dose: 3 times a week, 5 to 20 minutes per session; Monitoring: not reported; Route: lumbosacral, thoracic and cervical spine; Activity 2: composed of: postural education, cervical stabilisation and scapulothoracic exercises General: 10‐minute warm up, 20 to 25 minutes exercises, 5 to 10 minutes cool down and stretching; all exercises performed in a group up to 5 people. Postural education:sessions began with postural education by placing patients in a spot in sitting position with 2 mirrors reflecting their front and side views for them to find a neutral balanced position of the lumbar, thoracic and cervical spine. They were taught to perform the contraction of deep neck flexor muscle activity with the help of Stabiliser Biofeedback. Cervical stabilisation exercises: cervical bracing technique with the activation of deep neck flexors was performed. Patients were asked to maintain a neutral spine throughout the day as much as possible. The exercises to improve kinaesthetic awareness and conscious activation of the deep muscles included work‐outs of the bracing in neurodevelopment stages (supine, prone, quadrupedal and bipedal) in the cervical spine. Each contraction was kept for 10 seconds at each position for 6 to 10 repetitions. Cervical dynamic isometric exercises were performed directly forward, obliquely, towards right and left, and directly backward by maintaining stable spine with elastic resistance bands with 10 repetitions involving 6 to 10‐second hold period. Upper extremity exercises while maintaining a stable spine in each position for 8 to 12 repetitions. The exercises included functional training with elastic resistance and exercise balls on unstable surfaces to enhance unconscious activation of the muscles. Scapulothoracic stabilisation exercises:the thoracic bracing technique with postural alignment and minimal multifidus muscle activation for scapular orientation for scapulothoracic stabilisation was taught. Patients were asked to maintain the positions and contractions during the exercises including scapular retraction, eccentric scapular retraction, combined scapular retraction with shoulder lateral rotation, forward punch and dynamic hug. Patients began using yellow or red latex bands and a 200 cm long precut section of theraband with mild or medium tension. They performed 10 repetitions holding for 6 to 10 seconds each; Mode: soft tissue massage with hands and exercise in a group; Dose: 40 to 45 minutes; Monitoring: not reported; Route: cervical spine and shoulder COMPARISON TREATMENT Arm 2: stabilisation exercise. Activity 1: see above for stabilisation exercise protocol (postural education, cervical stabilisation and thoracic stabilisation exercises); Provider: an experienced physical therapist; Mode: manual application with hands; Dose: 40 to 45 minutes per session; Monitoring: not reported; Route: muscles of cervical and thoracic spine Treatment schedule: 4 weeks (12 sessions) Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean (SD): pain at rest: Arm 1 3.65, Arm 2 4.64; Pain during activity: Arm 1 5.39, Arm 2 6.72; Pain at night Arm 1: 6.29, Arm 2 3.20 Reported results: no significant difference detected between groups MD (Arm 1 vs Arm 2 pain during activity) at immediate post treatment: ‐0.76, 95% CI ‐2.13 to 0.61 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: Short Form‐36 (Physical Component Subscale), 0 to 100 Baseline mean: Arm 1 40.56, Arm 2 37.83 Reported results: no significant difference detected between groups MD at immediate post‐treatment: ‐3.63, 95% CI ‐8.31 to 1.05 PATIENT SATISFACTION: not reported ADVERSE EVENT: no adverse event observed in either group COST OF CARE: not reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 145, "Block randomisation was carried out by a computer generated random number list prepared by an investigator with no direct involvement with the trial." Comment: an acceptable randomisation process was observed.
Allocation concealment (selection bias)	Unclear risk	Comment: no information about the allocation process was described.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention i.e. massage vs exercise.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible as participants were the outcome assessors. A participant‐reported outcome measure (VAS) was used.
Incomplete outcome data (attrition bias)	Low risk	Quote: Figure 1, "analysed n = 30". Comment: all participants were analysed.
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Figure 1. Comment: dropouts were excluded from analysis.
Selective reporting (reporting bias)	Unclear risk	Comment: trial registration was not completed until approximately 1 month prior to completion of the study.
Baseline similarities	High risk	Quote: Table 1, Table 2. Comment: groups demonstrate differences in gender, age, exercise habits and smoking; large difference in night pain on visual analogue scale.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not reported.
Acceptable compliance	Unclear risk	Comment: compliance was not reported.
Similar outcome assessment	Low risk	Comment: similar outcome assessment at baseline and 4 weeks.
Other potential sources of bias	High risk	Funding bias: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale, no minimal important difference considered yet mentioned in pilot study, valid; Other: statistical error: analysis was not performed properly ‐ blocking was not consistent in analysis.

Cen 2003.

Study characteristics
Methods	Type of trial: RCT, cross‐over randomised trial (1st period data used) Number analysed/randomised: 28/31 Intention‐to‐treat analysis: not calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: chronic mechanical neck disorder Radicular signs/symptoms: not reported Sex: 74% female Age mean (SD): Arm 1 47 (11), Arm 2 48 (13), Control 51 (7) Severity: Northwick Park Pain Questionnaire 0 to 100, mean (SD): Arm 1 32.4, Arm 2 27.8, Control 31.5 Duration of complaints: 12 months or more Setting: California State University, Northridge Country: USA
Interventions	INDEX TREATMENT Arm 1: massage. Activity: traditional Chinese therapeutic massage using the following 2 techniques: one finger mediation massage that uses tip and/or whole surface of thumb, rolling massage uses the fifth metacarpophalangeal joint and hypothenar eminence, both use swinging back and forth motion 120 times per minute; Provider: 1 acupuncturist trained in TCTM for 5 years ‐ the principal investigator; Mode: soft tissue massage with hands; Dose: 3 times per week, 30 minutes; Monitoring: not reported; Route: cervical spine and upper shoulder COMPARISON TREATMENT Arm 2: therapeutic exercise programme after heat application: specific stretching (head tilt, trapezius stretch, neck flexion, shoulders and neck rolls) for 10 minutes directed by physician with weekly follow‐up for 6 weeks. Provider: 1 acupuncturist trained in TCTM for 5 years ‐ the principal investigator; Mode: home exercise programme and heat application; Dose: 10 minutes heat application followed with 10 minutes exercise; Monitoring: not reported; Route: cervical spine and shoulder Control: no treatment Treatment schedule: Arm 1 6 weeks (18 sessions), Arm 2 one initial visit and 5 telephone follow‐ups Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: not reported FUNCTION/DISABILITY: Northwick Park Neck Pain Questionnaire (NPQ) 0 to 100 Baseline mean: Arm 1 32.4, Arm 2 27.8, Control 31.5 Reported results: Arm 1 had significantly lower scoring to Arm 2 and Control groups immediately after phase I MD (Arm 1 vs Control) at immediate post treatment: ‐22.40, 95% CI ‐33.13 to ‐11.67 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (no treatment) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: Page 89 under trials design, “The participants were randomly assigned to three groups (A, B, C) by a pre‐generated assignment schema.” Comment: no detailed information on random generation was provided.
Allocation concealment (selection bias)	Unclear risk	Comment: not explicitly mentioned or described in the study.
Blinding ‐ patient (performance bias)	High risk	Comment: use of a control group to ensure blinding, however it was hard to determine whether the participants in the exercise group or the traditional Chinese therapeutic massage group were unaware of the type of treatment they were receiving.
Blinding ‐ care provider (performance bias)	High risk	Comment: it is hard to blind the care provider during a study examining manual therapy techniques or therapeutic exercise as an intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: both the participants in the traditional Chinese therapeutic massage and exercise groups would be able to detect a change in their cervical range of motion and pain post intervention. Therefore, it would have been hard to blind the participants in the traditional Chinese therapeutic massage and exercise programme from what effects they should be experiencing.
Incomplete outcome data (attrition bias)	Low risk	Quote: Table 3. Comment: 3 out of 31 participants dropped out. The researchers provided an adequate explanation as to why the participants dropped out from the study, such as orthopaedic surgeon’s recommendation and/or for personal reasons.
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Table 3. Comment: intention‐to‐treat analysis was not performed, and the study does not indicate how soon the measurement effects were obtained.
Selective reporting (reporting bias)	Unclear risk	Quote: "The study was approved by the Institutional Research Board of Calstate University Northridge" Comment: the protocol was not previously published or referenced in the current paper; the protocol was initiated prior to the start of the study.
Baseline similarities	Low risk	Quote: Table 2 and 3. Comment: the tables reported data on age, gender proportion, causes of neck pain, medication taking, pain and function‐disability level and were similar between groups.
Similar or avoided co‐interventions	Low risk	Comment: not indicated or explicitly stated within the study, no reduction of medication dosage reported in all groups.
Acceptable compliance	High risk	Comment: during phase 1 of the study, only the participants in the exercise group were checked on by the physician to ensure their compliance with the intervention. However, the study does not mention if this occurred for the remaining groups. Additionally, compliance with the interventions was not stated for Phase 2.
Similar outcome assessment	Low risk	Comment: Northwick Park Pain Questionnaire and cervical range of motion measurements were collected at the same time for all groups, e.g. baseline, post Phase 1 and 2.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: Northwick Park Pain Questionnaire, valid; other: the study appeared to be free of other sources of bias. No major threat to the study’s internal validity was observed.

Choksi 2021.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 66/66 Intention‐to‐treat analysis: calculated Power analysis: calculated, 90% Funding source: none Declaration of interest: no competing interests
Participants	Disorder: myofascial pain syndrome Radicular signs/symptoms: not reported Sex: 92.42% female Age: mean (SD): Arm 1 20.40 (2.59), Arm 2 20.59 (3.00), Arm 3 26.68 (12.60) Severity VAS 0 to 10: Arm 1 5.91 (1.69), Arm 2: 5.16 (1.49), Arm 3 5.14 (1.48) Duration of complaints: more than 15 days Setting: outpatient physiotherapy department of S. S. Agrawal Institute of Physiotherapy and Medical Care Education or primary care Country: India
Interventions	INDEX TREATMENT Arm 1: deep transverse friction massage adjunct to conventional therapy: Activity: frictions were applied slowly with a pressure slightly painful, approximately at the pressure pain threshold level of each patient; Provider: physiotherapist; Mode: soft tissue massage with hands; Dose: 3 minutes; Monitoring: not reported; Route: upper trapezius muscle Arm 2: ischaemic compression adjunct to conventional therapy. Activity: sustained deep pressure with the thumb applied to upper trapezius trigger point for 30 seconds to 1 minute; Provider: physiotherapist; Mode: soft tissue with hands; Dose: 1 minute; Monitoring: not reported; Route: upper trapezius muscle COMPARISON TREATMENT Arm 3: conventional therapy (conventional exercises and cold pack): Activity: conventional exercises composed of scapular movements including protraction, retraction, elevation and depression, and active neck movements with flexion, extension, side flexions and both rotations 10 repetitions and 5 second hold, cold pack was applied over upper trapezius region in a sitting position for 20 minutes; Mode: exercise and cold pack; Dose: 30 minutes; Monitoring: not reported; Route: upper back Treatment schedule: 3 days, 3 sessions Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean (SD): Arm 1 5.91 (1.69), Arm 2: 5.16 (1.49), Arm 3 5.14 (1.48) Reported results: there was no difference between groups; states that all the 3 arms were equally effective in reducing tenderness of trapezius MD (Arm 1 vs Arm 3) at short‐term follow‐up: 10.20, 95% CI ‐1.54 to 21.94 MD (Arm 2 vs Arm 3) at short‐term follow‐up: 1.60, 95% CI ‐8.75 to 11.95 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Arm 3 (conventional care) as well as Arm 2 vs Arm 3 (conventional care) and the risk of bias judgements were based on both of these 2 comparisons.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Comment: no information was provided.
Allocation concealment (selection bias)	Unclear risk	Quote: Page 48, "Randomly allocated in three groups in sequence" Comment: not enough information was provided.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention and study design.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention and study design.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible as the participants were the outcome assessors. A participant‐reported outcome measure (VAS) was used.
Incomplete outcome data (attrition bias)	Unclear risk	Quote: Figure 1 and table 4 Comment: it was not clearly described.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Quote: Figure 1 Comment: no specific information was provided.
Selective reporting (reporting bias)	Unclear risk	Comment: no study protocol or registration was provided.
Baseline similarities	Unclear risk	Comment: not enough information was provided, including the duration of neck pain.
Similar or avoided co‐interventions	Unclear risk	Comment: no information was provided.
Acceptable compliance	Unclear risk	Comment: no information was provided.
Similar outcome assessment	Unclear risk	Quote: Page 47, under abstraction, "Pre and post treatment visual analogue scale ... were taken." Comment: it was not clear when post‐treatment assessment was done.
Other potential sources of bias	Low risk	Funding source: none; conflict of interest: none; outcome measurement: VAS, valid.

Forman 2014.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 27/27 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific neck and shoulder pain (myofascial pain syndrome) Radicular signs/symptoms: not specified Sex: Arm 1 93.75% female, Control 81.82% female Age: mean (SD): Arm 1 49.44 (11.34), Control 48.27 (9.47) Severity: NPRS 0 to 10, mean (SD): Arm 1 4.7 (1.8), Control 3.5 (1.3) Duration of complaints: chronic Setting: university Country: USA
Interventions	INDEX TREATMENT Arm 1: soft tissue mobilisation. Activity: effleurage/gliding strokes using warming massage oil applied on cervical extensors, sternocleidomastoids, upper trapezii and levator scapulae for 2 minutes, then a head strap was attached to participant's head for upper trapezius work. Therapists used the flats of the knuckles, finger pads and/or thumb to perform the active muscle therapy massage strokes. Three 6‐second gross passes (covering the entire muscle from proximal to distal attachment) of eccentric resistance and stripping were conducted. Three additional 6‐second gross stripping passes from origin to insertion followed the focused work. The same process was repeated on the opposite side upper trapezius. For the active muscle therapy massage on the cervical extensors, participant cervical extensor muscles were passively shortened, and then they were instructed to slowly and actively contract the neck flexor muscles to flex their necks as the cervical paraspinals were stripped with the finger pads from distal to proximal attachments. Ten 6‐second passes were completed. No focused passes were performed on the cervical extensors. For the active muscle therapy massage on the levator scapulae, 3 gross passes, 10 focused passes and 3 more gross passes were administered to both levator scapulae; Provider: therapist; Mode: soft tissue massage with hands; Dose: 1 treatment for 30 minutes; Monitoring: not reported; Route: cervical spine and shoulder CONTROL TREATMENT Control: no treatment. Activity: no treatment, instruction or intervention was given; Mode: none; Dose: none; Monitoring: not reported; Route: cervical spine and shoulder Treatment schedule: 1 session per week for 6 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: numeric pain rating scale 0 to 10 Baseline mean (SD): Arm 1 4.7 (1.8), Control 3.5 (1.3) Reported results: no differences in improvement between both groups MD (Arm 1 vs Control) at immediate post treatment: ‐2.50, 95% CI ‐4.17 to ‐0.82 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean (SD): Arm 1 28.0 (9.7), Control 18.40 (9.7) Reported results: between‐group differences are not provided MD (Arm 1 vs Control) at immediate post‐treatment: ‐0.06, 95% CI ‐7.77 to 6.57 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: Page 6 under design, “a randomised study was conducted”. Comment: the method of randomisation was not described.
Allocation concealment (selection bias)	Unclear risk	Comment: the method of treatment allocation concealment was not described.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible due to the study design: use of participant‐reported outcome measures. The participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 18, paragraph 1 under results, “All sixteen participants completed the Active Muscle Therapy."
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: there were no dropouts.
Selective reporting (reporting bias)	High risk	Comment: there were inconsistencies between the protocol and study report. The protocol reports on a home‐based exercise programme as well as 4 groups with 60 participants, whereas this study reported on 2 groups with 27 participants and no home‐based exercise programme.
Baseline similarities	High risk	Quote: Page 19, Table 3. Comment: baseline Neck Disability Index scores were not similar.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐interventions were not discussed.
Acceptable compliance	Low risk	Quote: Page 18, paragraph 1 under results, “All sixteen participants completed the active muscle therapy intervention consisting of six weekly treatments over six weeks and 11 participants completed the control group participation”.
Similar outcome assessment	Low risk	Comment: there were similar outcomes at baseline and 6 weeks.
Other potential sources of bias	High risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: numeric pain rating scale and Neck Disability Index, reliable and valid; other: a repeated measures ANOVA was conducted for data analysis, yet we have no F‐values reported or any indication of main effect differences or interactions. They only reported % differences.

Gauns 2018.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 40/41 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: mechanical neck pain with radiation into the upper extremity (myofascial pain syndrome) Radicular signs/symptoms: absent Sex: 65% female Age: mean (SD): Arm 1 36.27 (11.43), Control: 33.1 (12.96) Severity: Northwick Park Neck Pain Questionnaire, mean (SD): Arm 1 28.83 (6.61), Control 32.58 (8.47) Duration of complaints: not specified Setting: hospital physiotherapy outpatient department Country: India
Interventions	INDEX TREATMENT Arm 1: myofascial release adjunct to conventional care. Activity: gross stretch of the upper quarter and gross stretch of posterior cervical spine; Provider: physiotherapist; Mode: manual application with hands; Dose: hot moist pack 15 minutes, myofascial release 10 to 15 minutes/session, transcutaneous electrical nerve stimulation (TENS) 20 minutes; Monitoring: not reported; Route: arm and neck COMPARISON TREATMENT Control: conventional care. Activity: hot moist pack, TENS, and stretching and strengthening exercise; Provider: physiotherapist; Mode: heat, TENS, stretch, exercise; Dose: hot moist pack 20 minutes, TENS 15 minutes, home exercise programme; Monitoring: not reported; Route: cervical and scapular regions Treatment schedule: 6 sessions, 6 days Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: not reported FUNCTION/DISABILITY: Northwick Park Neck Pain Questionnaire (NPQ) 0 to 100 Baseline mean (SD): Arm 1 28.83 (6.61), Control 32.58 (8.47) Reported results: statistical significance favouring Arm 1 MD (Arm 1 vs Control) at immediate post treatment: ‐9.10, 95% CI ‐13.37 to ‐4.83 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: Page 53 under methods, "Using envelope method, the subjects were segregated into two groups" Comment: there was not sufficient detail on randomisation.
Allocation concealment (selection bias)	Unclear risk	Comment: there was no mention of how treatment allocation was concealed.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding is not possible due to the nature of the intervention, massage vs conventional treatment including hot pack and exercises.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding is not possible due to the nature of the intervention, massage vs conventional treatment including hot pack and exercises.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible due to the use of participant‐reported outcome measures (self‐report questionnaires): the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 55, CONSORT chart. Comment: 1 dropout in the control group was reported.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Quote: Page 55 consort chart. Comment: there was no intention‐to‐treat analysis performed; one dropout in control group was not included.
Selective reporting (reporting bias)	Unclear risk	Comment: there was no clinical trial number reported.
Baseline similarities	Unclear risk	Quote: Page 56, Table 4. Comment: there was no report of the duration of pain.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not reported.
Acceptable compliance	Unclear risk	Comment: there was no report on home exercise programme compliance.
Similar outcome assessment	Low risk	Quote: Page 53, paragraph 2 under methods. Comment: the assessment was done at the beginning (1st day) and at the end (6th day).
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: Northwick Park Neck Pain Questionnaire. No details of outcome measures including psychometric properties or relevant references were provided.

Gemmell 2008.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 45/45 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific cervical disorders (myofascial pain syndrome) Radicular signs/symptoms: not reported Sex: not reported Age: mean (SD): Arm 1 24 (3.3), Arm 2 24 (4.6), Control 23 (1.5) Severity: VAS 0 to 100, mean (SD): Arm 1 41.3 (7.8), Arm 2 43.6 (8.8), Control 38.1 (8.8) Duration of complaints (mean (SD)): less than 3 months Setting: Anglo European College of Chiropractic Country: UK
Interventions	INDEX TREATMENT Arm 1: soft tissue mobilisation. Activity: ischaemic compression: sustained deep pressure with thumb to trigger point. Pressure was released when there was a decrease in tension in the trigger point or when the trigger point was no longer tender or when 1 minute had passed. Done to upper fibres of trapezius; Provider: chiropractor; Mode: soft tissue massage with hands; Dose: 1 treatment 30 to 60 seconds; Monitoring: not reported; Route: upper fibres of trapezius COMPARISON TREATMENT Arm 2: trigger point pressure release. Activity: a non‐painful pressure with the thumb was applied slowly until a tissue resistance barrier was felt. The pressure was increased until a new barrier was felt. This process was repeated until there was no trigger point tension or tenderness or until 90 seconds had passed. Applied to trigger points in upper fibres of trapezius; Provider: chiropractor; Mode: manual application with thumb; Dose: 90 seconds maximum; Monitoring: not reported; Route: upper trapezius Control: sham ultrasound: a detuned Medi‐Link Systems ultrasounds machine was used. Patients were told pulsed ultrasound was going to be used. Ultrasound was applied over upper fibres of trapezius. Provider: chiropractor; Mode: manual application of ultrasound; Dose: 2 minutes; Monitoring: not reported; Route: upper fibres of trapezius Treatment schedule: 1 session Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 100 Baseline mean: Arm 1 41.3, Arm 2 43.8, Control 38.1 Reported results: no statistically significant results between groups. There was a clinically significant difference between Arm 1 and Control in the number needed to treat (NNT). Arm 1 NNT = 3; Control NNT = 5 MD (Arm 1 vs Control) at immediate post treatment: 0.26, 95% CI ‐7.42 to 7.94 MD (Arm 2 vs Control) at immediate post treatment: 4.46, 95% CI ‐4.82 to 13.74 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (sham) as well as Arm 2 vs Control (sham) and the risk of bias judgements were based on these 2 comparisons.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 31, “The randomisation scheme was generated by using the website Randomisation.com.” Comment: this was acceptable.
Allocation concealment (selection bias)	Low risk	Quote: Page 31, “To ensure equal numbers in the groups, participants were randomised in blocks of three. Sealed opaque envelopes were prepared containing the assigned treatment and numbered consecutively. Participants were allocated to the next available envelope number.” Comment: this was appropriately performed: sealed, opaque envelopes, consecutively numbered.
Blinding ‐ patient (performance bias)	High risk	Quote: page 32, "A detuned Medi‐Link Systems ultrasound machine from Electro‐Medical Supplies (Greenham) Ltd. was used..." Comment: this was not possible because the interventions (hands‐on manual therapies vs sham ultrasound) are perceptibly different. The success of blinding was not tested among the patients, and its success was not reported. a) Placebo described: the "sham ultrasound" is not similar to the hands‐on massage intervention. b) Contextual factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described but the hands‐on element of the massage group was not stated. c) Patient factors: not reported. d) Practitioner factors: described as "the clinician"; poor reporting d) Study assessed the fidelity of randomisation: not reported.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention. It is hard to blind the caregiver during a study examining manual therapy techniques. The treatment provider cannot be blinded to the manual therapies for trigger points and use of sham ultrasound.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: the participant was the outcome assessor due to participant‐reported outcome measures, but the participant was not credibly blinded in this study due to the type of sham.
Incomplete outcome data (attrition bias)	Low risk	Comment: all participants completed the study; no dropouts were reported.
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: this is a pre‐post design of 1 session.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol was available or referenced.
Baseline similarities	Low risk	Quote: Table 1. Comment: characteristics of the participants, including age and pain intensity, were observed to be similar between groups.
Similar or avoided co‐interventions	Low risk	Comment: this was a pre‐post study design with no opportunity for co‐intervention.
Acceptable compliance	Low risk	Comment: this was a pre‐post study design.
Similar outcome assessment	Low risk	Comment: the measure was based on immediate effect; baseline, post‐treatment.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale, valid; other: the study appears to be free of other sources of bias.

Haller 2016.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 54/54 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific cervical neck disorders/mechanical neck pain Radicular signs/symptoms: not reported Sex: 81.5% females Age: mean (SD): total 44.6 (10.0), Arm 1 44.2 (9.7), Arm 2 45.0 (10.5) Severity: VAS 0 to 100, mean (SD): Arm 1 64.1(12.8), Arm 2 64.4 (13.3) Duration of complaints: mean (SD): Arm 1 9.9 (9.2), Arm 2 9.3 (8.8) years Setting: Department of Internal and Integrative Medicine, University of Duisburg‐Essen Country: Germany
Interventions	INDEX TREATMENT Arm 1: craniosacral therapy. Activity: purpose was to release restrictions of the cranium of the spine up to the pelvis and the sacrum using standardised application of gentle fascial traction, release, unwinding techniques; Provider: not reported; Mode: soft tissue massage with hands; Dose: 1 session per week, 1 treatment for 45 minutes; Monitoring: treatment steps were recorded by therapists using a structured log; Route: cranium of the spine up to the pelvis and sacrum COMPARISON TREATMENT Control: light touch sham treatment. Activity: sham was applied on standardised anatomic areas, equal to those treated with craniosacral therapy, for 2 minutes each time. Body awareness instructions were given to stimulate craniosacral therapy dialogue techniques; Provider: not reported; Mode: manual application with hands. Dose: 45 minutes; Monitoring: treatment steps were recorded by therapists using a structured log; Route: cranium of the spine up to the pelvis and sacrum Treatment schedule: 8 sessions, 8 weeks Duration of follow‐up: immediately post‐treatment (2 days post‐treatment), short‐term (12 weeks) CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale (pain during last 7 days) 0 to 100 Baseline mean: Arm 1 64.1, Control 64.4 Reported results: significant difference favouring Arm 1 at immediate post‐treatment and 12 weeks post‐treatment MD (Arm 1 vs Control) at short‐term follow‐up: ‐16.20, 95% CI ‐26.42 to ‐5.98 Pain on Movement Questionnaire, 0 to 100 MD (Arm 1 vs Control) at short‐term follow‐up: ‐13.50, 95% CI ‐22.77 to ‐4.23 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean: Arm 1 32.4, Control 29.3 Reported results: significant difference favouring Arm 1 at immediate post‐treatment and short‐term follow‐up MD (Arm 1 vs Control) at short‐term follow‐up: ‐5.40, 95% CI ‐9.73 to ‐1.07 PARTICIPANT‐REPORTED TREATMENT SUCCESS: Patient's Impression of Global Improvement 1 to 7 MD (Arm 1 vs Control) at short‐term follow‐up: ‐0.80, 95% CI ‐1.39 to ‐0.21 HEALTH‐RELATED QUALITY OF LIFE: Short Form‐12 Physical Component, 0 to 100 Baseline mean: Arm 1 38.0, Control 41.2 Reported results: significant difference favouring Arm 1 immediate post‐treatment and short‐term follow‐up MD (Arm 1 vs Control) at short‐term follow‐up: ‐5.30, 95% CI ‐8.24 to ‐2.36 Physical Well Being, 0 to 5 MD (Arm 1 vs Control) at short‐term follow‐up: ‐0.30, 95% CI ‐0.54 to ‐0.06 PATIENT SATISFACTION: not reported ADVERSE EVENT: no serious adverse events were reported. Minor adverse events during or after the treatment were reported by 6 patients in the craniosacral therapy group and included increased neck pain in 2, and pain in the jaw area, shivering, tiredness, strong emotional reactions and weeping in 1 patient, respectively. Within the sham group, 8 patients reported minor side effects, which included transient headache or migraine in 7 patients, worsened neck pain in 3, tingling sensations in 2 and dizziness in 1. COST OF CARE: not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: page 442 under randomisation, “statistician who was not involved in conducting the study generated a non stratified allocation sequence with randomly varying block lengths using the random number generator.” Comment: this was acceptable.
Allocation concealment (selection bias)	Low risk	Quote: page 442 under randomisation, “...prepared sealed and opaque envelopes sorted in the ascending order of group assignments, the envelope with the lowest number was opened directly after baseline assessment”. Comment: this was acceptable.
Blinding ‐ patient (performance bias)	Low risk	Quote: page 442 under interventions, "light touch was applied on standardised anatomic areas, equal to those treated with craniosacral therapy, for 2 minutes each time. In addition, body awareness instructions were given to simulate craniosacral therapy dialog techniques." "Treatment comprised 8 units of craniosacral therapy techniques or sham treatment once a week lasting 45 minutes. Participants were told they would receive one of 2 different Craniosacral therapy techniques." Comment: the sham was credible. The massage and sham groups are likely indistinguishable for the patients. The success of blinding was tested among the patients and was reported to be successful. a) Placebo described: the placebo (light contact) seemed similar to the intervention. b) Contextual factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described and the same as the intervention group. Department of Internal and Integrative Medicine, Kliniken Essen‐Mitte, University of Duisburg‐Essen, Essen, Germany. c) Patient‐related factors: reported ‐ Credibility/Expectancy Questionnaire and quality of the therapeutic alliance, measured by the Helping Alliance Questionnaire; patients in both groups received initial structural craniosacral therapy examination. d)Practitioner‐related factors: reported 4 licensed physiotherapists with advanced CST qualification, and on average 6 years of clinical practice. d) Was the success of blinding tested among patients and was it successful?: Yes. "treatment expectancy was assessed as part of the Credibility/Expectancy Questionnaire on a 9‐point rating scale from 1 (not at all) to 9 (very much). Treatment credibility and quality of the therapeutic alliance, measured by the Helping Alliance Questionnaire, were analysed and reported separately." Patients’ expectancy, credibility and therapeutic alliance did not appear to affect study outcomes, blinding patients to group allocation was possible, and sham intervention was tolerable and safe.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention: the same physical therapist treated both groups.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	Low risk	Comment: the participant was the outcome assessor due to participant‐reported outcome measures (e.g. pain, disability) and the participant was likely blinded to the sham treatment.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 444, Figure 1. Comment: the report provided an excellent dropout analysis.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Page 443 under statistical analysis, “All analyses were based on the intention‐to‐treat population including all patients who were initially randomised”. Comment: this was acceptable.
Selective reporting (reporting bias)	Low risk	Comment: the protocol number was NCT01526447. The report did not report on week 20.
Baseline similarities	Low risk	Quote: Page 445, Table 2. Comment: the reported data on age, gender proportion, pain intensity and functioning level were similar between groups.
Similar or avoided co‐interventions	High risk	Quote: Page 443 and 445, Figure 2. Comment: the sham group used acupuncture and massage more than the craniosacral therapy group. Use of medication and other co‐interventions appeared higher in the sham control group. The authors did not control for patient use of co‐interventions.
Acceptable compliance	Unclear risk	Comment: this was not adequately reported.
Similar outcome assessment	Low risk	Quote: Page 442, under trial design and registration. Comment: baseline, 8 weeks and 20 weeks were reported for all groups.
Other potential sources of bias	Low risk	Funding source: not reported; conflict of interest: none; outcome measurement: reliability and validity of outcomes were not discussed; other: the study appears to be free of other sources of bias. No major threat to the study’s internal validity was observed.

Iqbal 2016.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 90/90 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific cervical neck disorder (myofacial pain syndrome) Radicular signs/symptoms: not reported Sex: 0% females Age: ranged between 19 and 38 years old Severity: not reported Duration of complaints: not reported Setting: hospital Country: India
Interventions	INDEX TREATMENT Arm 1: ischaemic compression and muscle energy technique adjunct to conventional treatment. Activity: in supine position with cervical spine in opposite lateral flexion to the treating part so that upper trapezius muscle fibres were kept in a lengthened position. Gradual increasing pressure to the myofacial trigger point pain until participant perceived first noticeable pain. Pressure maintained until discomfort/pain eased by 50%, then pressure increased until discomfort appeared again; Provider: physiotherapist; Mode: soft tissue massage with hands; Dose: 90 seconds/session; Monitoring: not reported; Route: neck COMPARISON TREATMENT Arm 2: muscle energy technique adjunct to conventional treatment. Activity: in supine position with cervical spine in opposite lateral flexion to the treating part so that upper trapezius muscle fibres were kept in a lengthened position. Moderate isometric contraction for 5 seconds followed by 3 seconds of relaxation while reaching to the new barrier was performed; Provider: physiotherapist; Mode: using hands; Dose: 32 seconds/session; Monitoring: not reported; Route: neck Arm 3: conventional treatment. Activity: hot packs for 20 minutes and active stretching exercises; Dose: 5 repetitions per session, 100 seconds/session; Monitoring: not reported; Route: neck Treatment schedule: 3 sessions, 1 week Duration of follow‐up: 1 day immediately post‐treatment and short‐term (1 week post‐intervention) CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 The study has data representing “Between group post hoc Bonferroni analysis for VAS and NDI and not mean and SD which is required by our analysis.” FUNCTION/DISABILITY: Neck Disability Index scale not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Could not contact the author due to lack of email identification; no response when the editor of the journal was contacted. Data from this trial were not included in our analysis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 173, “assigned randomly via chit method to any of the three groups” Comment: this was an acceptable form of randomisation.
Allocation concealment (selection bias)	Unclear risk	Comment: this was not explicitly mentioned or described in the study.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding of participants was not possible as ischaemic compression, acupuncture and sham laser were distinguishable.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: the trial used participant‐reported outcome measures. Blinding of the outcome assessor was not possible as the participants were the outcome assessors.
Incomplete outcome data (attrition bias)	Unclear risk	Comment: this was not described.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Comment: this was not described.
Selective reporting (reporting bias)	Unclear risk	Comment: no trial registration or protocol was reported.
Baseline similarities	Unclear risk	Comment: the baseline data were not described.
Similar or avoided co‐interventions	Unclear risk	Comment: this was not described.
Acceptable compliance	Unclear risk	Quote: Page 173 under interventions in methods, "all the three groups received the treatment on three alternate days for one week" Comment: in the results, no information about how many participants actually had the allocated intervention was provided.
Similar outcome assessment	Low risk	Quote: Page 173 under procedure, "Visual Analogue Scale were taken pre‐intervention and after two minutes of post ‐intervention".
Other potential sources of bias	High risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale, valid; other: statistical reporting: No analysed data including mean, SD were provided but P values from post hoc Bonferroni analysis were.

Irnich 2001.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 165/177 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: subacute‐chronic mechanical neck disorder Radicular signs/symptoms: not reported Sex: 66% females Age: mean (SD): Arm 1 52.3 (13.3), Arm 2 52.7 (11.5), Control 52.2 (13.2) Severity: visual analogue scale 0 to 100, mean: Arm 1 54.71, Arm 2 54.15, Control: 57.15 Duration of complaints: more than 1 month Setting: University in Munich and Würzburg Country: Germany
Interventions	INDEX TREATMENT Arm 1: soft tissue mobilisation. Activity: massage: conventional western massage(effleurage, pétrissage, friction, tapotement); Provider: massages were performed by 5 experienced physiotherapists; Mode: soft tissue massage with hands; Dose: 30 minutes; Monitoring: not reported; Route: cervical spine and shoulder COMPARISON TREATMENT Arm 2: acupuncture. Activity: traditional Chinese approach (ear acupuncture and dry needling); Provider: acupuncture and sham laser acupuncture were performed by 4 experi­enced, licensed medical acupuncturists. Mode: manual application with acupuncture needles; Dose: 30 minutes sessions, 3 times per week for total of 5 sessions; Monitoring: not reported; Route: acupuncture points Control: sham laser. Activity: laser pen that was inactivated by manufacturer Seirin International; Provider: acupuncture and sham laser acupuncture were performed by 4 experi­enced, licensed medical acupuncturists; Mode: manual application with laser pen; Dose: 30 minutes; Monitoring: not reported; Route: acupuncture points Treatment schedule: 5 sessions, 3 weeks Duration of follow‐up: 1 week immediately post‐treatment, short‐term (3 months) CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale (pain related to motion) 0 to 100 Baseline mean: Arm 1 54.71, Arm 2 54.15, Control 57.15 Reported results: significant favouring Arm 2, acupuncture, compared to Arm 1 (P = 0.0052) at 1 week post intervention; not significant at 3 months MD (Arm 1 vs Control) at short‐term follow‐up: ‐0.31, 95% CI ‐10.65 to 10.03 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: slight pain or lowered blood pressure reported by 4 patients in Arm 1, 17 patients in Arm 2, 12 patients in Control RR (Arm 1 vs Control) at short‐term follow‐up: 0.33, 95% CI 0.11 to 0.97 COST OF CARE: not reported
Notes	Communication: author provided additional details on treatment technique and data on immediate post and short‐term (3 months) follow‐up. Analysis included Arm 1 vs Control (placebo) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 2 under randomisation, "A block randomisation stratified for two centres was performed by using a validated software program (PC Random, Biometric Center for Therapeutic Studies, Munich)." Comment: randomisation was adequate.
Allocation concealment (selection bias)	Unclear risk	Comment: allocation was not described.
Blinding ‐ patient (performance bias)	High risk	Comment: participants reported to be blinded, but blinding of participants was not possible as acupuncture, massage and sham laser acupuncture were distinguishable.
Blinding ‐ care provider (performance bias)	High risk	Comment: care providers were reported to be blinded but blinding of treatment providers was not possible as acupuncture, massage and sham laser acupuncture are distinguishable.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Quote: "Placebo—Sham laser acupuncture was performed with a laser pen, which was inactivated by the manufac­turer (Laser Pen, Seirin International, Fort Lauder­dale). Only red light was emitted. Patients were not informed about the inactivation of the laser pen. To strengthen the power of this sham procedure, visual and acoustic signals common for this type of laser pen accompanied the red light emission. Criteria for selec­tion of points were identical with those used in the acu­puncture group, including palpation of acupuncture points for diagnostic reasons. Every point was treated for 2 minutes, with the pen at a distance of 0.5­1 cm from the skin." Comment: the participant was the outcome assessor due to participant‐report outcome measures, but the participant was not credibly blinded in this study due to the type of sham. a) Placebo described: the "sham laser" is not similar to the massage intervention. b) Contextual factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described, but the hands‐on element of the massage group was not replicated. c) Patient naive to procedure: not reported. d) Study assessed the fidelity of randomisation: yes. "To evaluate the adequacy of control treatments we assessed patients' beliefs about the treatment. After randomisation and before the first treatment they had to answer four questions on a 100 point visual analogue scale..." "We were surprised by the results of sham laser acupuncture compared with massage. They could be explained by an enhanced placebo effect, but the assessment of credibility showed no differences between therapies before treatment. Sham."
Incomplete outcome data (attrition bias)	Low risk	Quote: Flow chart. Comment: an explanation was provided for dropouts.
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: intention‐to‐treat analysis was performed and described within the study.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol was reported.
Baseline similarities	Low risk	Quote: Table 1. Comment: baseline similarities were present.
Similar or avoided co‐interventions	Low risk	Comment: although pain medication was avoided in the protocol, they did not monitor co‐interventions.
Acceptable compliance	High risk	Comment: compliance was not monitored.
Similar outcome assessment	Low risk	Comment: primary outcome measure, improvement of pain related to motion, was assessed 1 week after treatment and compared with baseline measurements.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale, valid; other: the study appeared to be free of other sources of bias. No major threat to the study’s internal validity was observed.

Jafari 2017.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 19/19 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: Tarbiat Modares University Declaration of interest: not reported
Participants	Disorder: cervicogenic headache Radicular signs/symptoms: not reported Sex: 100% females Age: mean (SD): total 37.11(12.25), Arm 1 38.66 (13.20), Control 35.7 (1.86) Severity: visual analogue scale 0 to 10, mean (SD): Arm 1 6.23 (1.30), Control 5.81 (1.67) Duration of complaints: not reported Setting: outpatient headache clinic Country: Iran
Interventions	INDEX TREATMENT Arm 1: ischaemic compression. Activity: myofacial trigger point was grasped between the physiotherapist's thumb and index fingers. The sustained and constant pressure for ischaemic compression was applied to maximum tolerable level before reproducing of headache pattern. When the patient reported a 50% reduction of pain, the pressure was increased to maximum tolerable level again; Provider: physiotherapist; Mode: manual approach; Dose: 3 sets of 30 to 60 seconds compression, with 30 seconds rest intervals between compressions; Monitoring: not reported; Route: cervical spine COMPARISON TREATMENT Control: no treatment applied Treatment schedule: 4 sessions, 1 week (8 days) Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean (SD): Arm 1 6.23 (1.30), Control 5.81 (1.67) Reported results: Arm 1 had lesser intensity, frequency and shorter duration in their headaches than those of control MD (Arm 1 vs Control) at immediate post‐treatment: ‐2.10, 95% CI ‐3.77 to ‐0.43 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Comment: randomisation was not described.
Allocation concealment (selection bias)	Unclear risk	Comment: allocation was not explicitly mentioned or described in the study.
Blinding ‐ patient (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: this was not possible. Participant‐reported outcome measures were used; the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Unclear risk	Comment: attrition was not described.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Comment: this was not described.
Selective reporting (reporting bias)	Unclear risk	Comment: no trial registration or protocol was found.
Baseline similarities	High risk	Quote: Table 1. Comment: characteristics of participants, including age and pain intensity, were similar between groups.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not described.
Acceptable compliance	Unclear risk	Comment: compliance was not described.
Similar outcome assessment	Unclear risk	Quote: Page 4 under study protocol. Comment: there were insufficient data; similarities between groups were not well described.
Other potential sources of bias	Low risk	Funding source: Tarbiat Modares University; conflict of interest: not reported; outcome measurement: visual analogue scale, valid; other: the study appeared to be free of other sources of bias. No major threat to the study’s internal validity was observed.

Kashyap 2018.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 45/51 Intention‐to‐treat analysis: not calculated Power analysis: not reported Funding source: Deanship of Scientific Research, King Saudi University Declaration of interest: no competing interests
Participants	Disorder: subacute‐chronic non‐specific neck disorder (myofascial pain syndrome) Radicular signs/symptoms: absent Sex: 100% females Age: mean (SD): total 21.49 (3.66) Severity: visual analogue scale 0 to 10, mean (SD): Arm 1 4.73 (1.79), Arm 2 4.93 (1.39), Arm 3 4.80 (1.47) Duration of complaints: not specified Setting: physiotherapy department of the Kailash Hospital, Noida Country: Saudi Arabia
Interventions	INDEX TREATMENT Arm 1: manual pressure release adjunct to conventional care. Activity: examiner applied gradually increasing pressure followed by sustained pressure to trigger point to patient pain tolerance; Provider: physiotherapist; Mode: manual application with hands; Dose: 20 to 30 seconds, 1 session; Monitoring: not reported; Route: trigger points cervical or bilateral scapular regions COMPARISON TREATMENT Arm 2: muscle energy technique adjunct to conventional care. Activity: participants in a supine position with the cervical spine in opposite lateral flexion to the area of the body being treated. Moderate isometric contraction (~75% of maximal) of the upper trapezius muscles was elicited; Provider: physiotherapist; Mode: manual application with hands; Dose: 5‐second contraction, 3‐second relaxation, 4 repetitions, 1 session; Monitoring: not reported; Route: trigger points cervical and scapular regions Arm 3: conventional care. Activity: neck exercises and postural advice; Provider: physiotherapist; Mode: exercise; Dosage 3 times daily for 10 repetitions; Monitoring: not reported; Route: cervical and scapular regions Treatment schedule: 1 session, 1 day Duration of follow‐up: 1 day immediately post‐treatment; short‐term at 2 weeks post‐treatment CO‐INTERVENTION: avoided in trial design
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean: Arm 1 4.73, Arm 2 4.93, Arm 3 4.80 Reported results: no statistical significance for all groups MD (Arm 1 vs Arm 3) at immediate post‐treatment: ‐0.27, 95% ‐0.89 to 0.35 MD (Arm 1 vs Arm 2) at immediate post‐treatment: 4.00, 95% CI ‐6.59 to 14.59 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean: Arm 1 21.56, Arm 2 22.19, Arm 3 17.51 Reported results: no statistical significance for all groups MD (Arm 1 vs Arm 3) at immediate post treatment: 0.10, 95% CI ‐5.83 to 6.03 MD (Arm 1 vs Arm 2) at immediate post treatment: ‐1.70, 95% CI ‐6.38 to 2.98 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Arm 3 (conventional care) as well as Arm 2 vs Arm 3 (conventional care) and the risk of bias judgements were based on these 2 comparisons.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 315, "using an online randomisation tool, Randomzation.com (http://www.randomization.com)." Comment: a credible online randomisation tool was used.
Allocation concealment (selection bias)	Unclear risk	Comment: allocation was not explicitly mentioned or described in the study.
Blinding ‐ patient (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention, i.e. manual pressure release vs muscle energy technique vs control.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: this was not possible as the participant is the outcome assessor due to use of patient‐reported outcome measures (visual analogue scale, Neck Disability Index).
Incomplete outcome data (attrition bias)	Unclear risk	Quote: Page 5155 under results, "Of the 51 participants, three dropped out with valid reasons, two failed to complete the intervention, and one complained of aggravated pain after the day 1 intervention and dropped out of the study." Comment: 3 dropouts of 51 participants was acceptable. One dropout was due to an adverse event, however it did not indicate from which group.
Randomised participants analysed in their groups (reporting bias)	High risk	Comment: no intention‐to‐treat analysis was conducted.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol or trial registration was reported.
Baseline similarities	High risk	Comment: pressure pain threshold was not similar at baseline. Large standard deviations for groups B and C were observed. Information on gender or duration of pain was not given.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not reported.
Acceptable compliance	Unclear risk	Comment: compliance was not reported.
Similar outcome assessment	Low risk	Quote: Page 3153 and Figure 1 on Page 3154, "Postintervention and follow‐up data were collected on study days 1, 5, 10, and 15 as described in Figure 1." Comment: similar timing of outcome assessment was reported for days 1, 5, 10 and 15.
Other potential sources of bias	Low risk	Funding source: the Deanship of Scientific Research, King Saud University; conflict of interest: not reported; outcome measurement: visual analogue scale, Neck Disability Index, valid; other: statistical reporting errors: it was unclear if a sample size/power calculation was completed. We suspect they were under sampled and may have committed a type II error for the between‐group comparisons. Also, there were inappropriate conclusions based on significant differences between groups indicating benefits of all interventions. Without between‐group differences they cannot say that the effect was equal or better.

Kim 2021.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 50/54 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: none Declaration of interest: no competing interests
Participants	Disorder: non‐specific neck disorder/mechanical neck pain Radicular signs/symptoms: absent Sex: 70% females Age: mean (SD): Arm 1 50.1 (5.5), Control 53.8 (5.4) Severity: visual analogue scale 0 to 100, mean (SD): Arm 1 57.0 (10.8), Control 58.3 (13.1) Duration of complaints: more than 3 months Setting: not reported Country: Korea
Interventions	INDEX TREATMENT Arm 1: Acupressure. Activity: in the supine position, the therapist placed the middle fingers on 3 unilateral acupoints located in the cervical spine. Each acupoint gentle pressure maintained about 10 seconds each time and repeated 5 times then gentle pressure was applied to 9 bilaterally symmetrical acupoints with the thumb; Provider: 1 therapist, a licensed aesthetician certified by Korean Ministry of Health and Welfare with a master’s degree majoring in manipulative therapy. In addition, the therapist completed the 2‐year acupressure apprenticeship programme with the Korea Association of Meridian Acupressure; Mode: manual application; Dose: 30 minutes; Monitoring: all selected acupoints were evaluated and validated by a Traditional Korean Medical doctor; Route: neck and shoulder COMPARISON TREATMENT Control: no treatment. Participants were asked not to change their lifestyle over 4 weeks and were given KRW 20,000 (about USD 18). Activity: none; Provider: 1 therapist as above; Mode: none; Dose: not applicable; Monitoring: not reported; Route: not applicable Treatment schedule: 8 session, 4 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 100 Baseline mean: Arm 1 57.0 (10.8), Control 58.3 (13.1) Reported results: significantly decreased in the acupressure group compared with changes in the control group MD (Arm 1 vs Control) at immediate post‐treatment: ‐27.60, 95% CI ‐35.90 to ‐19.30 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean (SD): Arm 1 32.7 (11.4), Control 28.6 (12.2) Reported results: significantly decreased in the acupressure group compared with changes in the control group MD (Arm 1 vs Control) at immediate post‐treatment: ‐9.80, 95% CI ‐15.55 to ‐4.05 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	High risk	Quote: Page 2, "randomly assigned 1:1.25 to either the experimental or control groups. ....Each subject was assigned to the experimental group or control group depending on the labelled paper in the envelope." Comment: method was not clearly described.
Allocation concealment (selection bias)	High risk	Quote: Page 2, "Each subject pulled out a sealed opaque envelope from the box." Comment: no serial numbering mentioned and procedure not clearly described. This method could be tampered with.
Blinding ‐ patient (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention, acupressure vs no treatment.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: this was not possible as the participants were the outcome assessors; participant‐reported outcome measures (Neck Disability Index, Visual Analogue Scale) were used.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 2, "One subject in the experimental group dropped out due to failure to adhere to the treatment schedule. .... four subjects in the control group withdrew from the study because they did not want to be in the control group" Comment: the dropout rates were acceptable.
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Page 2, "Fifty‐four women enrolled in the study……data from 24 women in the acupressure group and 26 women in the control group were analyzed." Comment: no intention‐to‐treat analysis was performed.
Selective reporting (reporting bias)	Unclear risk	Quote: Page 2, "KCT0005363". Comment: information was not available. We searched for the study protocol but it could not be found.
Baseline similarities	Low risk	Quote: Page 5, Table 1, Table 2. Comment: baseline characteristics between groups were observed to be comparable.
Similar or avoided co‐interventions	Low risk	Quote: Page 5, Table 1. Comment: 37 participants (74%) were not on any medication. Table 1 depicts those on extra medication. It seems balanced. There was no significant difference between groups for this factor at the start of care. The authors did not measure or screen for other health conditions. There was no monitoring for co‐intervention during care.
Acceptable compliance	Unclear risk	Quote: Page 2, "Twenty‐three subjects in the acupressure group underwent all eight intervention sessions." "Subjects in the control group were untreated and asked not to change their usual lifestyle over 4 weeks." Comment: there was no description of compliance in the control group participants, who were asked not to change their usual lifestyle over 4 weeks. We believe the control group should be monitored for adherence to "usual lifestyle".
Similar outcome assessment	Low risk	Quote: Page 1, "Before and after intervention, pain intensity and physical disability were measured by visual analogue scale and neck disability index" Comment: data were reported at baseline and immediate post‐treatment for both groups.
Other potential sources of bias	Low risk	Funding source: none; conflict of interest: none; outcome measurement: Neck Disability Index, visual analogue scale, valid. other: the study appeared to be free of other sources of bias. No major threat to the study’s internal validity was observed.

Klein 2013.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 61/61 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific neck disorder/mechanical neck pain Radicular signs/symptoms: not reported Sex: 73.8% females Age: mean (SD): Arm 1 49.9 (10.1), Control 41.9 (10.4) Severity: Pain Intensity Scale 0 to 5, mean (SD): Arm 1 2.2 (1.2), Control 2.2 (1.1) Duration of complaints: not specified Setting: private general practice Country: Germany
Interventions	INDEX TREATMENT Arm 1: strain counter‐strain. Activity: the affected body parts were positioned to the free direction contrary to the restriction. To activate neurophysiologic reflex mechanisms, this position was held for 90 seconds while the tender point was monitored by using the finger of the therapist in the position with the minimal tension of the tender point. Afterwards, a slow reposition to basic position was carried out; Provider: first author, general physician and Osteopathic Medicine practitioner, 8 years of experience in using osteopathic treatments; Mode: soft tissue stretch with hands; Dose: number of sets ‐ 1, duration of session ‐ 90 seconds but no information on total treatment time reported; Monitoring: not reported; Route: cervical spine and shoulder COMPARISON TREATMENT Control: sham treatment (hand on neck movements). Activity: the finger of the therapist was placed at the height of C4 para‐vertebrally on the right‐hand side of the dorsal part and the head was rotated by 30 degrees to the left to basic position without any flexion, extension or lateral flexion. This position was also held for 90 seconds. Afterwards, a slow reposition to basic position was carried out; Provider: as noted above; Dose: number of sets ‐ 1, duration of session ‐ 90 seconds but no information on total treatment time reported; Monitoring: not reported; Route: neck Treatment schedule: 1 session, 1 day Duration of follow‐up: immediately post‐treatment (2 to 3 days post‐treatment) CO‐INTERVENTION: a full individualised osteopathic treatment
Outcomes	PAIN: Pain Intensity Scale 0 to 5 Baseline mean (SD): Arm 1 2.2 (1.2), Control 2.2 (1.1) Reported results: pain was reduced significantly after the strain‐counterstrain but not after sham treatment MD (Arm 1 vs Control) at immediate post treatment: ‐0.30, 95% CI ‐0.90 to 0.30 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean (SD): Arm 1 40.4 (17.9), Control 39.4 (18.1) Reported results: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: RR (Arm 1 vs Control) at immediate post‐treatment: 1.00, 95% CI 0.87 to 1.14 HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: Arm 1 mild transient adverse effect (pain, dizziness) in 4/30, Arm 2 mild transient adverse effect (pain) in 1/31 COST OF CARE: not reported
Notes	Only baseline values were available for function.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 2, "using Research Randomizer (www.randomizer.org) with variable block sizes of 8, 10 and 12". Comment: this was acceptable.
Allocation concealment (selection bias)	Low risk	Quote: Page 2, "sequentially numbered, opaque, sealed envelopes prepared". Comment: this was acceptable.
Blinding ‐ patient (performance bias)	Unclear risk	Quote: Page 2, "To carry out the sham treatment, the finger of the therapist was placed at the height of C4 prevertebrally on the right‐hand side of the dorsal part and the head was rotated by 30 degrees to the left to basic position without any flexion, extension or lateral flexion. This position was also held for 90 seconds. Afterwards, a slow reposition to basic position was carried out." Comment: this was a credible sham technique. The massage and sham groups are likely indistinguishable for the patients. The success of blinding was not tested among the patients and is not reported. a) Placebo described: the placebo (light contact and positioning) is likely similar to the intervention. b) Contextual factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described and the same as the intervention group. c) Patient factors: not reported; no additional questionnaire on expectation, therapeutic alliance etc. was reported. d)Practitioner factors: "All treatments were performed by the first author, a general practitioner with additional qualifications in sports medicine and manual therapies... full osteopathic curriculum (postgraduate) of the Deutsch‐Amerikanische Akademie für Osteopathie (German‐American Academy of Osteopathic Medicine) in cooperation with the Philadelphia College of Osteopathic Medicine (Certificate and Diploma Osteopathic Medicine, EROP Diploma Osteopathic Medicine TM) and has 8 years of experience in using osteopathic treatments" d) Was the success of blinding tested among patients and was it successful?: not reported.
Blinding ‐ care provider (performance bias)	High risk	Comment: this was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	Unclear risk	Comment: the participant was the outcome assessor due to participant‐report outcome measures (e.g. pain, disability) and the participant may or may not be blinded to the sham treatment.
Incomplete outcome data (attrition bias)	Low risk	Quote: Figure 1. Comment: this was a pre‐post design with no follow‐up.
Randomised participants analysed in their groups (reporting bias)	Low risk	Comment: this was a pre‐post design with no follow‐up.
Selective reporting (reporting bias)	Low risk	Comment: trial registry checked and primary outcomes were reported.
Baseline similarities	Unclear risk	Quote: Page 4, "patients in the sham group were considerably more often female and younger. Patients in the control group tended to have a slightly longer history of neck complaints." Comment: this was adequately reported.
Similar or avoided co‐interventions	Unclear risk	Quote: Page 2, "individualized osteopathic intervention they would have received in routine practice outside the study." Comment: this was a pre‐post design but description of individualised treatment raised questions.
Acceptable compliance	Low risk	Comment: this was a pre‐post design with no follow‐up.
Similar outcome assessment	Low risk	Comment: this was a pre‐post design with a 1‐day follow‐up.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: Pain Intensity Scale, Neck Disability Scale, valid; other: the study appeared to be free of other sources of bias.

Kostopoulos 2008.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 90/90 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not clearly reported. This study was sponsored by St. Matthew’s University School of Medicine Declaration of interest: not reported
Participants	Disorder: subacute‐chronic non‐specific neck pain or headaches Radicular signs/symptoms: not reported Sex: 60% females Age: not specified Severity: VAS 0 to 10, mean (SD): Arm1 6.80 (1.77), Arm 2 7.08 (1.54), Arm 3 7.33 (1.48) Duration of complaints: greater than 1 month to 2 years Setting: private physiotherapy clinic Country: USA
Interventions	INDEX TREATMENT Arm 1: ischaemic compression adjunct to passive stretch. Activity: referred to Arm 2 and Arm 3 activity description; frequency alternate between ischaemic compression and passive stretch with 30 seconds rest intervals; Provider: a doctoral‐level trained physical therapist, 15 years experience in myofascial pain; Mode: manual stretching and pressure; Dose: alternate between ischaemic compression and passive stretch with 30 seconds rest intervals; Monitoring: not reported; Route: cervical spine COMPARISON TREATMEN Arm 2: passive stretch. Activity: the targeted muscle is stretched until tension is sensed at the end of range of motion. The patient exhales, allowing the muscle to relax, increasing the stretch. The newly gained range is held while the patient inhales. A further stretch is reached with successive exhalations allowing muscles to relax; Provider: as per Arm 1; Mode: manual stretching; Dose: 45 seconds at the rate of 3 to 4 mm/seconds; Monitoring: not reported; Route: cervical spine Arm 3: ischaemic compression. Activity: application of slowly increasing non‐painful pressure over a trigger point until a barrier of tissue resistance is encountered. Contact is maintained until the tissue barrier releases and pressure is increased to reach a new barrier to eliminate trigger point tension and tenderness; Provider: as per Arm 1; Mode: soft tissue pressure with hands; Dose: 15 minutes per session; Monitoring: not reported; Route: upper fibres of trapezius Treatment schedule: 6 sessions, 2 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean: Arm 1 6.8, Arm 2 7.08, Arm 3 7.33 Reported results: at 2 weeks from randomisation found a significant improvement within groups (P < 0.05 in all groups); Arm 3 was significantly better than Arm 1 (P < 0.05) and Arm 2 (P < 0.01) for pain MD (Arm 1 vs Arm 2) at immediate post treatment: ‐1.27, 95% CI ‐2.15 to ‐0.39 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Arm 2 and the risk of bias judgements were based on only these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	High risk	Quote: Page 268 under the sample, "Participants were randomly assigned to one of three treatment groups". Comment: insufficient information was provided; it appears to be an alternating sequence.
Allocation concealment (selection bias)	High risk	Quote: Page 268 under the sample, "Group assignment was conducted serially with the first patient assigned to .... until all participants had been assigned a group" Comment: insufficient information was provided. There appeared to be an alternating sequence.
Blinding ‐ patient (performance bias)	High risk	Comment: although participants were told the type of treatment they were to receive, they did not know that other participants were assigned to alternative groups receiving other forms of intervention. Blinding of participants was not possible as ischaemic compression and passive stretching were distinguishable.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding of treatment providers was not possible due to the nature of the study design; ischaemic compression and passive stretching were distinguishable.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding of outcome assessor was not possible as participant‐reported outcomes were applied. The paritipant was the outcome assessor.
Incomplete outcome data (attrition bias)	Low risk	Quote: Table 4. Comment: all participants completed the study; no dropouts were reported.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Table 4. Comment: there were no dropouts.
Selective reporting (reporting bias)	Unclear risk	Comment: there was no reference to a previously published protocol; a protocol appeared to be initiated prior to the start of the study.
Baseline similarities	Unclear risk	Quote: Table 1, 2, 4. Comment: pain intensity was similar between groups as reported in Table 4. There was insufficient information on other characteristics including age, gender, pain duration, functioning level, race and educational level noted in Table 1 and 2; summary data across both groups together were reported; data for the treatment or control group were not reported separately or compared between groups.
Similar or avoided co‐interventions	Unclear risk	Comment: there was insufficient information; this was not reported in the study.
Acceptable compliance	Unclear risk	Comment: compliance was not reported in the study.
Similar outcome assessment	Low risk	Comment: the measure was based on immediate effect, baseline and after last intervention.
Other potential sources of bias	Unclear risk	Funding source: this study was sponsored by St. Matthew’s University School of Medicine (http:// www.stmatthews.edu/); conflict of Interest: not reported; outcome measurement: visual analogue scale, valid; other: statistical reporting and errors could not be judged by our senior statistician due to the inadequate/poor reporting of data in this trial.

Matsubara 2011.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 33/33 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific neck disorder/mechanical neck pain Radicular signs/symptoms: not reported Sex: not reported Age: mean (SD): Arm 1 34.8 (4.0), Arm 2 35.5 (6.4), Control 37.2 (7.0) Severity: verbal rating scale 0 to 3, mean (SD): Arm 1 2.1 (0.5), Arm 2 1.7 (0.8), Control 1.8 (0.6) Duration of complaints: not reported Setting: Nihon Fukushi University Country: Japan
Interventions	INDEX TREATMENT Arm 1: acupressure to local points. Activity: local acupressure was performed. Acupressure was performed to 3 local points on both sides of the individual's neck and shoulders. The therapist used the pulp of the right thumb and treated with 20 to 25 cycles per minute in each point; Provider: 1 investigator, department of rehabilitation; Mode: acupuncture; Dose: 3 sets for 30 seconds (1.5 minutes) and bilaterally applied, 3 total minutes; Monitoring: not reported; Route: cervical spine and shoulder COMPARISON TREATMENT Arm 2: acupressure to distal points. Activity: acupressure was performed to 3 distal points on both of the individual's forearm. The therapist used the pulp of the right thumb and treated with 20 to 25 cycles per minute in each point; Provider: 1 investigator; Mode: manual application with pulp of right thumb; Dose: 3 sets of 30 seconds; Monitoring: not reported; Route: forearm Control: no treatment performed Treatment schedule: 1 session, 1 day Duration of follow‐up: immediately post‐treatment (1 day post‐treatment) CO‐INTERVENTION: not reported
Outcomes	PAIN: verbal rating scale 0 to 3 Baseline mean: Arm 1 2.1, Arm 2 1.7, Control 1.8 Reported results: authors reported significant results favouring Arm 2 vs Control immediately post‐treatment. Between‐group differences were not significant for Arm 2 vs Control at 1 day post‐treatment. Between‐group differences for Arm 1 vs Arm 2 and Arm 1 vs Control were not found to be significant at any time points. FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean: Arm 1 9.4, Arm 2 7.6, Control 7.9 Reported results: authors reported significant results favouring Arm 2 vs Control at 1 day post‐treatment. Between‐group difference were not significant for Arm 1 vs Arm 2 or Arm 1 vs Control at 1 day post‐treatment PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: verbal rating scale 0 to 3 Reported Results: no significant differences between Arm 1 vs Arm 2 at both time points ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Authors were contacted on 5 November 2016. Unable to obtain post‐treatment data for the mean and SDs of verbal rating scale and Neck Disability Index. Data for these analyses were not available or used.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: page 2, "The subjects were randomly allocated to three groups." Comment: it was not adequately described.
Allocation concealment (selection bias)	Unclear risk	Comment: there was insufficient information. Allocation was not adequately described.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding participants was not possible due to the nature of the intervention (massage vs no treatment) and location of applications (local vs distal).
Blinding ‐ care provider (performance bias)	High risk	Comment: it was not possible due to the nature of the intervention (massage vs no treatment) and location of applications (local vs distal).
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: due to the use of participant‐reported outcome measures, the participant was the outcome assessor and could not be blinded.
Incomplete outcome data (attrition bias)	Low risk	Comment: this was a pre‐post design.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Comment: it was unclear if all data were collected 1 day after intervention.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol or registration was reported.
Baseline similarities	Low risk	Quote: Table 1. Comment: characteristics, including age, pain and functional level, at baseline between groups were comparable.
Similar or avoided co‐interventions	Unclear risk	Comment: there was no description of co‐intervention between study date and day 1 follow‐up.
Acceptable compliance	Low risk	Comment: compliance was acceptable as this was a pre‐post design.
Similar outcome assessment	Low risk	Comment: the timing of the outcomes was similar at baseline, post‐treatment and 1 day post‐treatment.
Other potential sources of bias	High risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: verbal rating scale, no evidence of validated outcome was noted; the intensity of neck pain or stiffness was evaluated on a numerical scale from 0 to 3 (0: no pain, 1: mild pain, 2: moderate pain and 3: severe pain); Neck Disability Index, valid; other: statistical reporting and error is questioned due to inadequate and very poor reporting at the 1 day follow‐up endpoint.

Nagrale 2010.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 60/60 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific neck disorder (myofascial pain syndrome) Radicular signs/symptoms: not reported Sex: 58% female Age: mean (SD): Arm 1 28.2 (4.78), Arm 2 27.6 (4.28) Severity: VAS 0 to 10, mean (SD): Arm 1 8.2 (0.66), Arm 2 8.36 (0.55) Duration of complaints: less than 3 months Setting: outpatient clinic Country: India
Interventions	INDEX TREATMENT Arm 1: integrated neuromuscular inhibition techniques adjunct to muscle energy technique. Activity: participants placed in supine and their arm was placed in slight shoulder abduction with the elbow bent and hand resting on their stomach. Practitioner moved throughout the upper fibres of trapezius using a pincer grasp and made note of any trigger points. Once the trigger points were identified, ischaemic compression was applied by the therapist utilising pincer grip placing the thumb and index finger over the active trigger point. Slow increasing levels of pressure were applied until the first barrier was felt. Pressure was applied again, until a new barrier was felt. This was repeated until tension/tenderness could not be identified, or 90 seconds had elapsed. All active trigger points were treated. Application of strain counter‐strain: modified digital pressure was applied to the identified trigger points as participants rated their pain on a scale of 1 to 10 pressure was increased if pain could not be identified. If pain was reproduced, the pressure was maintained over the trigger points until there was a 70% reduction in pain. Once this was identified, it was held for 20 to 30 seconds and repeated 3 to 5 times. Muscle energy technique towards the upper fibres of trapezius: each isometric contraction was held for 7 to 10 seconds and followed by further contralateral side‐bending, flexion and ipsilateral rotation to maintain the soft tissue stretch. Each stretch was held for 30 seconds and was repeated 3 to 5 times; Provider: manual physical therapist; Mode: manual application with hands; Dose: 3 times per week; Monitoring: not reported; Route: cervical spine and upper fibres of trapezius COMPARISON TREATMENT Arm 2: muscle energy techniques (the same treatment as above). Activity: participants were placed supine and practitioner stabilised the shoulder on the affected side with one hand, while the ear/mastoid area of the affected side was helped by opposite hand. The head and neck were then side bent towards the contralateral side, flexed and rotated ipsilaterally, placing the participant just short of the upper fibres' trapezius restriction barrier. Muscle energy technique:the participants then shrugged the involved/stabilised shoulder towards their ear at submaximal, pain‐free effort (20% of their strength), which was held for 7 to 10 seconds while a normal breathing rhythm was maintained. During relaxation phase, the head and neck were eased into increasing degrees of side bending, flexion and rotation; resistance phase = 7 to 10 seconds isometric phase, relaxation phase = 3 to 5 repetitions of each stretch, stretch held for 30 seconds each; Provider: manual physical therapist, Mode: soft tissue mobilisation with hand; Dose: 3 times per week; Monitoring: not reported; Route: cervical spine and upper fibres of trapezius Treatment schedule: 4 weeks, 12 sessions Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean: Arm 1 8.2, Arm 2 8.36 Reported results: significantly greater improvement in pain favouring Arm 1 (P < 0.05) MD at immediate post‐treatment: ‐0.82, 95% CI ‐1.12 to ‐0.52 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean: Arm 1 42.96, Arm 2 42.9 Reported results: significantly greater improvement in disability favouring Arm 1 (P < 0.05) MD (Arm 1 vs Arm 2) at immediate post‐treatment: ‐4.69, 95% CI ‐6.75 to ‐2.63 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	We wrote to the author for data clarification on 23 April 2022. The SD seemed very small and we suspected a reporting error where the SD can be confused with the SE.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 39, "a computer generated randomised block of numbers was performed before the study." Comment: adequate reporting noted.
Allocation concealment (selection bias)	Unclear risk	Quote: Page 39. "A sealed, opaque envelope was then opened indicating the treatment group to which the subjects were randomized." Comment: sequential numbering was not mentioned.
Blinding ‐ patient (performance bias)	High risk	Comment: it was not possible to blind participants due to the nature of the intervention.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the treatments. The treatments were distinctly different between groups.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: due to the use of participant‐reported outcome measures, the participant was the outcome assessor and could not be blinded.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 40, Figure 5. Comment: no dropouts were reported.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Page 40, Figure 5. Comment: participants were analysed in their groups.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol or registration was available.
Baseline similarities	Low risk	Quote: Page 40, table 1. Comment: groups were equivalent at baseline.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not reported.
Acceptable compliance	Unclear risk	Comment: compliance was not reported.
Similar outcome assessment	Low risk	Comment: both groups were assessed at baseline, 2 weeks and 4 weeks from initiation of treatment.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale, Neck DisabiIity Index, valid; other: there may be a reporting error. The standard deviation seemed very small and we suspected this might be a standard error. The author was written to for data clarification.

Pach 2018.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 89/92 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: none Declaration of interest: no competing interests
Participants	Disorder: chronic non‐specific neck pain Radicular signs/symptoms: absent Sex: 85.3% female Age: mean (SD): Arm 1 46.2 (9.6), Control 44.7 (9.9) Severity: VAS 0 to 100, mean (SD): Arm 1 55.8 (10.2), Control 59.5 (12.6) Duration of complaints: mean (SD): Arm 1 11.5 (9.1) years, Control 11.2 (11.4) years Setting: university outpatient clinic specialised in Integrative Medicine Country: Germany
Interventions	INDEX TREATMENT Arm 1: Tuina therapy. Activity: meridian point manipulation, soft tissue massage at neck region, local muscle stretching, mobilisation and traction of cervical spine, manipulation of local pain (trigger points); Provider: 1 German physiotherapist with 2 years of part‐time Tuina training by Prof. Dr. A. Meng, head of the Austrian Tuina working group and 8 years of expertise in Tuina therapy; Mode: manual application with hands; Dose: 30 minutes; 2 sessions per week; Monitoring: not reported; Route: meridian points, wrist, elbow, shoulder, neck (cervical spine), trigger points COMPARISON TREATMENT Control: no treatment Treatment schedule: 6 sessions, 3 weeks Duration of follow‐up: immediately post‐treatment at 1 week; short‐term at 9 weeks post‐treatment CO‐INTERVENTION: avoided in trial design ‐ both groups were instructed not to receive other treatments
Outcomes	PAIN: visual analogue scale 0 to 100 Baseline mean: Arm 1 55.8, Control 59.5 Reported results: pain intensity after 12 weeks differed significantly between groups MD (Arm 1 vs Control) at short‐term follow‐up: ‐18.00, 95% CI ‐26.84 to ‐9.16 FUNCTION/DISABILITY: Neck Disability Index 0 to 100 Baseline mean: Arm 1 45.5, Control 46.5 Reported results: significant differences in favour of Arm 1 for function MD (Arm 1 vs Control) at short‐term follow‐up: ‐0.51, 95% CI ‐0.87 to ‐0.14 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: Short Form‐12 (physical), 0 to 100 Baseline mean: Arm 1 43.8, Control 43.6 Reported results: significant differences in favour of Arm 1 for the physical aspect of quality of life MD (Arm 1 vs Control) at short‐term follow‐up: ‐0.75, 95% CI ‐1.18 to ‐0.32 PATIENT SATISFACTION: not reported ADVERSE EVENT: aching muscles, tensions, dizziness, headache, sleepiness, mood swings, painful point, difficulties staying asleep, slight nausea and head not movable (mild transient) COST OF CARE: quality‐adjusted life years over 12 weeks 0.006, 95% CI ‐0.005 to 0.017; no significant difference
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 232, "Randomisation performed with SAS/BASE software". Comment: it was adequately reported.
Allocation concealment (selection bias)	Low risk	Quote: Page 232, "database (Microsoft Corporation, Redmond, WA), where it was implemented in the background, which was not accessible to anyone in study involved with randomisation or treatment. Randomisation performed within database to ensure concealment." Comment: It was considered as an adequate allocation process.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention (Tunia vs no intervention). These treatments are distinguishable.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention (Tunia vs no intervention).
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: participant‐reported outcome measures (Neck Disability Index, visual analogue scale) were used. The participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 234 Figure 1, and page 235. Comment: there were 3 from the wait‐lists that dropped out.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Page 233, Page 234 Figure 1. Comment: intention‐to‐treat analysis was conducted.
Selective reporting (reporting bias)	Low risk	Quote: Page 232. Comment: the clinical trial number was reported and the outcome measures listed match those reported.
Baseline similarities	Low risk	Quote: Page 235 table 1. Comment: baseline characteristics between groups observed were comparable.
Similar or avoided co‐interventions	Low risk	Quote: Page 233 paragraph 3, page 235 paragraph 1 and page 235 paragraph 2. "Accompanying therapies, including concurrent therapies, were not significantly different between both groups." Comment: concurrent therapies were similar and those patients that had an adverse event only used self‐care. Self‐care was vague and did not instil confidence regarding the reporting of co‐interventions.
Acceptable compliance	Low risk	Quote: Page 233 paragraph 3, Page 234 paragraph 1. Comment: patients received 5.3 ± 1.2 treatments out of a possible 6.
Similar outcome assessment	Low risk	Quote: Page 232 paragraph 5. Comment: outcomes reported at baseline, 4 weeks and 12 weeks were comparable.
Other potential sources of bias	Low risk	Funding source: none; conflict of interest: none; outcome measurement: visual analogue scale, Neck Pain and Disability Scale and Short Form‐12 (Physical) measures, valid; other: the study appeared to be free of other sources of bias. No major threat to the study's internal validity was observed.

Phadke 2016.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 60/60 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: not reported Declaration of interest: no competing interests
Participants	Disorder: non‐specific mechanical cervical disorder Radicular signs/symptoms: not reported Sex: 59% female Age: ranged between 18 and 50 years Severity: visual analogue scale 0 to 10, mean (SD): Arm 1 5.75 (1.17), Arm 2 5.5 (1.20) Duration of complaints: 4 to 12 weeks Setting: not reported Country: India
Interventions	INDEX TREATMENT Arm 1: muscle energy technique adjunct to passive stretch. Activity: post‐isometric relaxation technique was applied to upper trapezius and levator scapulae; Provider: physiotherapist; Mode: using hands; Dose: timing: not reported, frequency: once a day, dose: number of sets ‐ 5, duration of a session‐ 100 seconds, rest intervals ‐ not reported, order of care ‐ not reported, nuration: 6 days; Monitoring: not reported; Route: neck COMPARISON TREATMENT Arm 2: passive stretch. Activity: passive stretching was applied to upper fibres of trapezius and levator scapulae muscles; Provider: physiotherapist; Mode: soft tissue stretching with hands; Dose: timing: not reported, frequency: once a day, dose: number of sets ‐ 5, duration of a session‐ 100 seconds, rest intervals ‐ not reported, order of care ‐ not reported, duration: 6 days; Monitoring: not reported; Route: upper fibres of trapezius and levator scapulae Treatment schedule: 6 sessions, 1 week Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: strengthening exercise for deep neck flexors, rhomboids, lower trapezius and serratus anterior, and stretching exercise for pectoralis muscle and Hydrocollator hot packs for both arms
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean (SD): Arm 1 5.75 (1.17), Arm 2 5.5 (1.20) Reported results: significant difference was seen in both groups in terms of change in pain intensity on VAS MD (Arm 1 vs Arm 2) at immediate post‐treatment: ‐0.68, 95% CI ‐1.22 to ‐0.14 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean (SD): Arm 1 17.21 (2.75), Arm 2 17.28 (2.86) Reported results: significant difference was seen in both groups in terms of change in functional disability on NDI MD (Arm 1 vs Arm 2) at immediate post‐treatment: ‐0.69, 95% CI ‐1.23 to ‐0.15 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	The author provided post‐intervention data for pain and function.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 6 under randomisation, "randomly allocated to Group A or Group B, using chit method without replacement." Comment: this was an adequate randomisation method.
Allocation concealment (selection bias)	High risk	Quote: Page 6 under randomisation, "The allocation was conducted by the primary investigator prior to the baseline assessment." Comment: allocation was conducted by the investigator, raising suspicions of bias.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding of the care provider was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: this was not possible due to the use of participant‐reported outcome measures; the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Unclear risk	Quote: Page 7 under results, "Two participants from Group A dropped out due to personal reasons and two participants from Group B withdrew because they could not comply with the treatment and assessment schedule." Comment: one reason for dropout was identified as "personal reasons". This was insufficient information and did not provide enough detail.
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Page 7 under results, "Intention‐to‐treat analysis was not used."
Selective reporting (reporting bias)	Low risk	Comment: the protocol was consistent with the research paper.
Baseline similarities	Unclear risk	Comment: the duration of pain was not described.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not described.
Acceptable compliance	Unclear risk	Comment: compliance was not described.
Similar outcome assessment	Low risk	Comment: outcome assessments were similar, performed at baseline and 6 days.
Other potential sources of bias	Low risk	Funding source: not reported; conflict of interest: none; outcome measurement: visual analogue scale and Neck Disability Index, valid.

Puntumetakul 2019.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 30/30 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: Research Centre of Back, Neck, Other Joint Pain, and Human Performance Declaration of interest: no competing interests
Participants	Disorder: non‐specific mechanical neck pain Radicular signs/symptoms: absent Sex: 73.3% female Age: mean (SD): 23.17 (3.65) Severity: visual analogue scale 0 to 100, mean (SD): Arm 1 42.33 (7.72), Arm 2 45.29 (11.53) Duration of complaints: 4 to 12 weeks Setting: Department of Physical Therapy, Faculty of Associated Medical Sciences, Khon Kaen University Country: Thailand
Interventions	INDEX TREATMENT Arm 1: Rungthip massage technique adjunct to thoracic manipulation. Activity: Rungthip massage technique was performed with the participants in the side‐lying position, with 90 degrees of hip flexion and 90 degrees of knee flexion. The therapist gently pressed her thumb along the treatment lines from the level of the inferior angle of the scapula to the lowest rib; Provider: physiotherapist with training and experience in spinal manipulation; Mode: manual application with hands; Dose: 3 repetitions; Monitoring: not reported; Route: level of the inferior angle of the scapula to the lowest rib COMPARISON TREATMENT Arm 2: thoracic manipulation. Activity: thoracic manipulation was performed directly on both sides of the T6–T7 zygapophyseal joints. The participants were asked to lie in the prone position on the examination table and instructed to inhale and exhale deeply. During exhalation, the therapist performed thoracic manipulation (screw thrust technique) at the T6–T7 zygapophyseal joints. If a popping sound was not heard on the first attempt, the therapist repositioned the participant and performed a second manipulation; Provider: physiotherapist with training and experience in spinal manipulation; Mode: manual application with hands; Dose: 1 repetition per session, 2 sessions a week; Monitoring: not reported; Route: T6–T7 zygapophyseal joints; Mode: manual application with hands; Dosage: 1 to 2x/session, 2 sessions/week; Monitoring: not reported; Route: T6 to T7 zygapophyseal joints Treatment schedule: 6 sessions, 3 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 100 Baseline mean: Arm 1 42.33 (7.72), Arm 2 45.29 (11.53) Reported results: significant, favouring Arm 1 MD (Arm 1 vs Arm 2) at immediate post‐treatment: ‐11.04, 95% CI ‐18.31 to ‐3.77 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	We wrote to the author for unpublished data on the Neck Disability Index and adverse events on 18 February 2023.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 197 paragraph 3, "Computer generated randomised table of numbers". Comment: adequate randomisation was identified.
Allocation concealment (selection bias)	Unclear risk	Quote: Page 197 paragraph 3, "Sequentially numbered cards placed in opaque sealed envelopes" Comment: sequentially numbered cards were inside the opaque, sealed envelopes. The allocation remained unclear.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention, massage vs manipulation.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the intervention, massage vs manipulation.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: blinding was not possible due to participant‐reported outcome measures (visual analogue scale); the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Unclear risk	Quote: Page 198 Figure 1, Tables 1 and 2. Comment: no dropouts were reported in the figure (n = 15), but tables reported a different number of participants (n = 14).
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Quote: Page 198, Figure 1, Tables 1 and 2. Comment: we were not sure if there were dropouts and there was no indication that an intention‐to‐treat analysis was conducted.
Selective reporting (reporting bias)	High risk	Quote: Page 196 paragraph 1, "NCT03187808". Comment: the clinical trial number was reported. Secondary outcomes of cervical range of motion, Neck Disability Index and adverse events were not reported.
Baseline similarities	Low risk	Quote: Page 199 Table 1. Comments: baseline characteristics were similar.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not reported.
Acceptable compliance	Unclear risk	Quote: Page 198, Figure 1. Comment: there was no indication that all participants completed the treatment beyond Figure 1; it indicated that all participants completed the study. But the discrepancy in Tables 1 and 2 indicated a dropout or missing data that were not reported.
Similar outcome assessment	Low risk	Quote: Page 197 "... at baseline before commencing treatment and on completion of the intervention" Comment: the timing of outcome assessment was similar between groups.
Other potential sources of bias	Low risk	Funding source: the Research Center of Back, Neck, Other Joint Pain, and Human Performance; conflict of interest: none; outcome measurement: visual analogue scale, valid; other: the study appeared to be free of other sources of bias.

Segura‐Orti 2016.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 24/34 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: not reported Declaration of interest: no competing interests
Participants	Disorder: non‐specific mechanical cervical disorder (myofascial pain syndrome) Radicular signs/symptoms: absent Sex: 73% female Age: mean (SD): 32.7 (10.3) Severity: visual analogue scale 0 to 100, mean (SD): Arm 1 46.9 (20.9), Arm 2 32.2 (22.5), Control 34.2 (17.5) Duration of complaints: more than 4 months Setting: not reported Country: Spain
Interventions	INDEX TREATMENT Arm 1: strain counter‐strain. Activity: investigators placed the selected upper trapezius muscle in a shortened position for 90 seconds and "the subject's cervical spine in lateralisation ipsilateral to and rotation contralateral to the symptomatic myofascial trigger point". Once palpation of the selected myofascial trigger point no longer elicited a pain response, investigators passively maintained that position for 90 seconds; Provider: physiotherapist; Mode: manual application; Dose: timing: not reported, frequency: 6 sessions, number of sets ‐ not reported, duration of sessions ‐ unclear, rest intervals ‐ not reported, order of care ‐ not reported, duration: 3 weeks; Monitoring: not reported; Route: cervical spine COMPARISON TREATMENT Arm 2: dry needling plus exercise. Activity: a quick in and out technique was preferred to ensure an upper trapezius local twitch response and to promote effectiveness. Needling at the specific myofascial trigger point was continued until the local twitch response was exhausted; after the dry needling was completed, participants performed 8 repetitions of active shoulder abduction and shrugging followed by a passive upper trapezius stretch; Provider: physiotherapist; Mode: acupuncture needle; Dose: timing: not reported, frequency: 3 sessions, dose: not reported, duration: 3 weeks; Monitoring: not reported; Route: upper fibres of trapezius Control: sham, strain counter‐strain. Activity: no therapeutic intervention was intended for this group. Study participants were placed in a supine position while a myofascial trigger point was located on the subject’s upper trapezius muscle. Investigators were careful to only apply slight digital pressure below the pain threshold over the selected myofascial trigger point. Gentle, slow lateral and circular digital motions were performed over the myofascial trigger point over 90 seconds. Pain response was not elicited; Provider: physiotherapist; Mode: soft tissue stretching with hands; Dose: timing: not reported, frequency: not reported, dose: number of sets ‐ not reported, duration of sessions: unclear, rest intervals: not reported, order of care: not reported, duration: 3 weeks; Monitoring: not reported; Route: cervical spine Treatment schedule: 3 to 6 sessions, 3 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: no reported
Outcomes	PAIN: visual analogue scale 0 to 100 Baseline mean (SD): Arm 1: 46.9 (20.9), Arm 2 32.2 (22.5), Control 34.2 (17.5) Reported results: post‐intervention, the results of the mixed ANOVA revealed no differences between the 3 groups for any of the variables studied (P = 0.412) MD (Arm 1 vs Control) at immediate post treatment: 6.30, 95% CI ‐20.90 to 33.54 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean (SD): Arm 1 10.2 (7.7), Arm 2 7.2 (3.4), Control 8.8 (4.0) Reported results: post‐intervention, the results of the mixed ANOVA revealed no differences between the 3 groups for any of the variables studied (P = 0.739) MD (Arm 1 vs Control) at immediate post‐treatment: ‐4.40, 95% CI ‐23.32 to 14.52 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (sham) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 172 under study design, "using blocked randomisation by gender and age (http://www.randomization.com)". Comment: acceptable detail was provided.
Allocation concealment (selection bias)	Unclear risk	Comment: this was not described.
Blinding ‐ patient (performance bias)	High risk	Quote: "Study subjects were placed in a supine position while a myofascial trigger point was located on the subject’s upper trapezius muscle. Investigators were careful to only apply slight digital pressure below the pain threshold over the selected myofacial trigger point. Gentle, slow lateral and circular digital motions were performed over the myofascial trigger point over 90 seconds. Pain response was not elicited." Comment: blinding of participants was not possible as massage and acupuncture were distinguishable, however massage and sham‐massage could be blinded, albeit 50% of the sham group dropped out. The success of blinding among patients was not tested and its success was not reported. a) Placebo described: the massage placebo was light contact and light motion; it seemed similar to the intervention. b) Contextual‐factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was not well described. c) Patient‐related factors: not well reported; no additional questionnaire on expectation, therapeutic alliance etc. was reported. d) Practitioner‐related factors: "Six postgraduate programme physical therapists (PTs) were equipped to perform specific techniques and measures." e) Was the success of blinding tested among patients and successful?: not reported
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding was not possible due to the nature of the interventions.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: the participant was the outcome assessor due to participant‐reported outcome measures (e.g. pain, disability) and the participants were likely not blinded adequately.
Incomplete outcome data (attrition bias)	High risk	Quote: Figure 1. Comment: the dropout rate in the sham group was 50%.
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Figure 1. Comment: no intention‐to‐treat analysis was applied.
Selective reporting (reporting bias)	High risk	Quote: Table 3. Comment: no mention of the outcome ‐ electrical activity (EMG) ‐ and the number of candidates differed in the protocol.
Baseline similarities	High risk	Quote: Tables 2 and 3. Comment: baseline characteristics for gender, medication, work related to upper extremities, work compensation claim and visual analogue scale were similar between groups; there was no EMG or pain duration noted at baseline.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not described.
Acceptable compliance	Unclear risk	Comment: compliance was not described.
Similar outcome assessment	Unclear risk	Comment: the timing of the outcomes was similar between groups at baseline and 3 weeks post intervention. Although the description on page 173 under intervention group might suggest that the last day of treatment may differ between groups (2x/week vs 1x/week).
Other potential sources of bias	High risk	Funding source: not reported; conflict of interest: none; outcome measurement: visual analogue scale and Neck Disability Index, valid; other: statistical errors were suspected. Standard deviation (SD) and standard error (SE) can sometimes be confused in the reports of studies, and the terminology may be used inconsistently. We transformed SE into a SD using the RevMan calculator for both the visual analogue scale and Neck Disability Index. Additionally, with a loss of 50% of sham participants and the use of the full sample (n = 12) as noted in Table 3 to calculate P values, we were unclear if a statistical analysis error resulted.

Sherman 2014.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 221/228; Arm 1 (1 x 60 minutes per week) 38/38; Arm 2 (2 x 30 minutes per week) 38/38; Arm 3 (2 x 60 minutes per week) 38/39; Arm 4 (3 x 30 minutes per week) 34/37; Arm 5 (3 x 60 minutes per week) 38/39; Control (on waiting list) 35/37 Intention‐to‐treat analysis: not calculated Power analysis: calculated Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific mechanical neck pain Radicular signs/symptoms: absent Sex: 71% female Age: mean (SD): Arm 1 50.2 (10.90), Arm 2 42.3 (11.3), Arm 3 48.7 (11.5), Arm 4 45.7 (11.5), Arm 5 49.0 (9.9), Control 44.4 (12.2) Severity: NPRS 0 to 10, mean (SD): Arm 1 5.9 (1.5), Arm 2 5.8 (1.4), Arm 3 5.6 (1.1), Arm 4 6.1 (1.5), Arm 5 5.7 (1.2), Control 5.6 (1.3) Duration of complaints: more than 3 months Setting: group health integrated health care system, Seattle Country: USA
Interventions	INDEX TREATMENT forSherman 2014 Arm 1,Arm 3 and Arm 5:60‐minute massage protocol. Activity: consisting of 3 minutes of cervical range of motion assessment, 1 minute of hands‐on check‐in/tissue warming, 2 to 4 minutes of lymph drainage, 12 to 20 minutes of neck work consisting of: friction, myofacial techniques, effleurage, proprioceptive neuromuscular facilitation, muscle energy technique, active assisted stretching, 15 to 20 minutes for compensatory patterns, 6 to 10 minutes for neck work part two, 8 to 18 minutes for full body integration and 1 minute for completion; Provider: massage therapists, 5 years experience, completed special hands‐on training on the protocol and on the performance of massage in a research setting, including applicable regulations; Mode: soft tissue massage with hands; Dose:Arm 1 one per week, Arm 3 two per week, Arm 5 three per week; Monitoring: not reported; massage therapists practised the protocol prior to the study to ensure that they are able to deliver each protocol in a comfortable and accurate manner; Route: cervical spine and upper back and chest Arm 2 and Arm 4: 30‐minute massage protocol. Activity: 3 minutes of cervical range of motion assessment, 1 minute of hands‐on check‐in/tissue warming, 2 minutes of lymph drainage, 12 to 15 minutes of neck work consisting of friction, myofacial techniques, effleurage, proprioceptive neuromuscular facilitation, muscle energy technique, active assisted stretching, 4 to 8 minutes for compensatory patterns, and 1 minute for completion; Provider: massage therapists, 5 years experience; Mode: soft tissue massage with hands; Dose:Arm 2 two per week, Arm 4 three per week; Monitoring: not reported; Route: cervical spine and upper back and chest Control: 4‐week period on waiting list BOOSTER TREATMENT for Cook 2015 (companion follow‐up report) Booster: randomisation of participants to receive an additional 6 massages, 60 minutes 1 x/week. Primary randomisation (as noted in Sherman 2014) to 1 of 5 groups receiving 4 weeks of massage (30 minutes 2 x per week or 3 x per week; 60 minutes 1 x, 2 x or 3 x per week). Control: no additional booster massage Treatment schedule: varied session, 4 weeks with 2 to 3 massages per week for 30‐minute and 60‐minute treatments depending on the Arm Duration of follow‐up: 1 week after end of treatment; 12 and 26 weeks from the first randomisation; 12 weeks means 1 week post‐treatment for booster arm and 8 weeks post‐treatment for no further treatment arms; 26 weeks means 15 weeks post‐treatment for booster arms and 22 weeks post‐treatment for no further treatment arms CO‐INTERVENTION: not reported
Outcomes	PAIN: numeric pain rating scale (NPRS) 0 to 10 Baseline mean: Arm 1 (1 x 60 minutes per week): 5.9, Arm 2 (2 x 30 minutes/week): 5.8, Arm 3 (2 x 60 minutes per week): 5.6, Arm 4 (3 x 30 minutes per week): 6.1, Arm 5 (3 x 60 minutes per week): 5.7, Control (on waiting list): 5.6 Reported results: the adjusted likelihood of improvement in neck pain intensity was 2.30 (95% CI 1.26 to 4.18; P = 0.007) and 2.73 (95% CI 1.52 to 4.91; P = 0.001), respectively. For each additional weekly massage, there was an estimated ‐0.75‐point improvement in neck pain intensity (95% CI –1.01 to –0.47; P < 0.001). Cook: reported results: there was a main effect of the booster treatment at 1 week post‐treatment but not at 15 weeks. No statistically significant effect of the primary treatment group (group of first randomisation) was found at 1 or 15 weeks. Calculated results: adjusted mean change difference (booster treatment across all boosters vs no further treatment across all no treatment group): At 12 weeks post‐booster treatment: ‐0.61, 95% CI ‐1.17 to ‐0.05 At 26 weeks post‐booster treatment: ‐0.01, 95% CI ‐0.66 to 0.64 FUNCTION/DISABILITY: Neck Disability Index (NDI) 0 to 50 Baseline mean: Arm 1 (1 x 60 minutes per week): NDI 14; Arm 2 (2 x 30 minutes per week): NDI 13.4; Arm 3 (2 x 60 minutes per week): NDI 13.7; Arm 4 (3 x 30 minutes per week): NDI 13.1; Arm 5 (3 x 60 minutes per week): NDI 14.3; Control (on waiting list): NDI: 13.4 Reported results: the adjusted likelihood of improvement in NDI score for the 2 times weekly and 3 times weekly groups relative to the control group was RR 3.41 (95% CI 1.05 to 11.08; P = 0.04) and RR 4.98 (95% CI 1.64 to 15.17; P = 0.005), respectively. For each additional weekly massage, there was an estimated ‐1.81‐point improvement in NDI (95% CI –2.52 to –1.10; P < 0.001). Calculated results: change scores were used. SDs were imputed using the 5 trial SDs from this comparison. MD (Arm 5 vs Control): ‐9.44 NDI 0 to 100, 95% CI ‐14.99 to ‐3.89 Cook: reported results: there was a main effect of the booster treatment at 1 week but not at 15 weeks post‐treatment. No statistically significant effect of the primary treatment group (group of first randomisation) was found at 1 or 15 weeks. Calculated results: adjusted mean change difference (booster treatment across all boosters vs no further treatment across all no treatment group): At 1 week post‐booster treatment: ‐2.87, 95% CI ‐4.37 to ‐1.36 At 15 weeks post‐booster treatment: ‐1.44, 95% CI ‐3.21 to 0.33 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: very satisfied with care: yes/no Reported results: there were no differences observed by primary treatment group (group of first randomisation) at 1 or 15 weeks post‐booster treatment. Those receiving booster sessions had improvements in both dysfunction and pain at 1 week post‐booster treatment, but those were not significant at 15 weeks post‐booster treatment. ADVERSE EVENT: post‐treatment pain primarily spine pain 10 participants (5.2% of the total) reported 14 adverse events (11 mild and 3 moderately severe). Adverse event frequencies were similar in participants attending 30‐minute and 60‐minute treatments (4% vs 6%, respectively) and in those attending 1, 2 or 3 times per week (7.9% vs 2.6% vs 6.7%, respectively). COST OF CARE: not reported
Notes	We wrote to the author but she had passed away and data could not be accessed. Our analysis included primarily the comparison between Arm 5 vs Control (wait‐list) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Sherman 2014: Quote: page 113, "randomisation done using the R software (version 2.11.0)" Cook 2015: Quote: page 2208, "Randomisation using a computer‐assisted telephone interviewing program. Treatment assignments were generated using R (version 2.11.0;R core team (2013), Vienna, Austria." Comment: an acceptable method of randomisation was reported.
Allocation concealment (selection bias)	Low risk	Cook 2015: Quote: page 2208, "Randomisation using a computer‐assisted telephone interviewing program. Treatment assignments were generated using R (version 2.11.0; R core team (2013), Vienna, Austria." Comment: concealment of allocation using a computer‐assisted telephone interviewing program.
Blinding ‐ patient (performance bias)	High risk	Comment: blinding was not possible due to the nature of the interventions.
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding of the care provider was not possible due to the nature of the interventions.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: due to the use of a participant‐reported outcome measure, the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Unclear risk	Sherman and Cook: Figure 1. Comment: the dropout rate was reported and acceptable but the reasons for dropouts were not reported.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Sherman: Comment: dropouts were excluded from analysis; intention‐to‐treat analysis was not performed. Cook: Comment: unclear reporting, as Figure 1 shows that patients were excluded from analysis, but the authors state they performed an intention‐to‐treat analysis.
Selective reporting (reporting bias)	Low risk	Comment: the protocol was reported.
Baseline similarities	Low risk	Quote: Table 1. Comment: Table 1 showed similarities in the study population between groups.
Similar or avoided co‐interventions	Unclear risk	Quote: Cook 2015: page 2210 "The use of medication as a non‐study treatment varied across groups...." Comment: the medication use was different between groups, other types of intervention were not reported.
Acceptable compliance	Low risk	Sherman: Quote: page 115, "under non‐study treatment". Cook: Quote: Figure 1 and page 2210. Comment: there was 100% compliance in the primary treatment period, 90.9% in the 30‐minute 3 times a week group and 87% in the booster dose group.
Similar outcome assessment	Low risk	Sherman: Comment: the timing of outcome assessment was similar at baseline and 5 weeks. Cook: Comment: the timing of outcome assessment was similar at baseline, 12 weeks and 26 weeks from primary group randomisation. This meant follow‐up times differed between the 2 groups: 1 week post‐treatment for booster dose, but 8 weeks post‐treatment for the no treatment group. Note: as this was the second phase of a randomised controlled trial, the authors did not state whether baseline data were re‐measured, or if they used the first phase's post‐treatment data as the new baseline.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; Sherman: outcome measurement: numeric pain rating scale valid Cook: Neck Disability Index, valid; other: no major threat to the study’s internal validity was observed.

Skillgate 2020.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 545/621 Intention‐to‐treat analysis: not calculated Power analysis: not reported Funding source: Swedish Research Council, Working Life and Welfare, Swedish Naprapathic Association Declaration of interest: no competing interests
Participants	Disorder: non‐specific mechanical neck pain Radicular signs/symptoms: not specified Sex: 69% female Age: mean (range): 46 (25 to 64) Severity: numerical rating scale 0 to 100: all groups greater than 10/100 Duration of complaints: subacute‐chronic Setting: research clinic (Scandinavian College of Naturopathic Manual Medicine, Stockholm) Country: Sweden
Interventions	INDEX TREATMENT Arm 3: Massage adjunct to exercise. Activity: base element: strengthening, strength‐endurance, stretch‐range of motion; modular elements: feedback‐feedforward system; biomechanical ‐ dynamic and static stabilisation; Provider: naprapaths with experience in massage and exercise therapy; Mode: manual resistance; Movement: cervical flexion, extension, scapular protraction and retraction, shoulder flexion and abduction, temporomandibular opening and closing. Postural control of neck and scapulae in standing, supine and sitting; Dose: intensity was similar exercise and massage intensities as Arm 1 and Arm 2; 60 minutes (included 25 minutes exercise, 25 minutes and 10 minutes of anamnesis), 6 sessions over 6 weeks; Monitoring: supervised exercise. Participants were filmed with their smartphones; massages individually monitored. The therapists underwent training sessions (3‐hour sessions twice) and regular supervision to ensure a standardised provision of all the interventions in the STONE trial; Route: same as Arm 1 and 2 COMPARISON TREATMENT Arm 1: Massage. Activity: effleurage, firm motion involving compression and pressure release and deep muscle/fascia massage to areas that produced concordant symptoms; intensity – delivered to tolerance without reaching pain over visual analogue scale 50/100; Provider: naprapaths (medical professionals who focus on the management of musculoskeletal pain with a combination of manual therapy, exercise and/or advice) either under training (but with experience in the provision of massage and physical training) or licensed; Mode: manual application; Dose: 45 minutes (included 10 minutes of anamnesis), up to 6 sessions over 6 weeks; Monitoring: each study participant individually monitored. The therapists underwent training sessions (3‐hour sessions twice) and regular supervision to ensure a standardised provision of all the interventions in the STONE trial; Route: cervical, upper back, jaw and chest Arm 2: Exercises. Activity: (base element‐strengthening, strength‐endurance, stretch‐ROM; modular elements: feedback‐feedforward system; biomechanical ‐ dynamic and static stabilisation); Provider: naprapaths with experience in massage and exercise therapy; Mode: manual resistance; Movement: cervical flexion, extension, scapular protraction and retraction, shoulder flexion and abduction, temporomandibular opening and closing. Postural control of neck and scapulae in standing, supine and sitting; Dose: intensity depended on participants’ ability, tolerance and stamina; 45 minutes (included 10 minutes of anamnesis), 6 sessions over 6 weeks; Monitoring: supervised. Participants were filmed with their smartphones to ensure proper form while at home. The therapists underwent training sessions (3‐hour sessions twice) and regular supervision to ensure a standardised provision of all the interventions in the STONE trial; Route: cervical, scapular, jaw. Arm 4: Advice. Activity: motivating discussion regarding spinal pain with booklet; Provider: naprapaths with experience in massage and exercise therapy; Mode: booklet; Dose: 30 minutes (included 10 minutes of anamnesis), up to 3 sessions; Monitoring: each participant individually monitored. The therapists underwent training sessions (3‐hour sessions twice) and regular supervision to ensure a standardised provision of all the interventions in the STONE trial; Route: not specified Treatment schedule: 3 to 6 sessions, 6 weeks Duration of follow‐up: 1 day immediately post‐treatment, 12 months CO‐INTERVENTION: not avoided
Outcomes	PAIN: numerical rating scale 0 to 10 Baseline mean: Arm 1 5.9 (1.3), Arm 2 6.1 (1.3), Arm 3 6.1 (1.5), Arm 4 5.8 (1.5) Reported results: improvement in pain intensity favouring massage and combined therapy compared to advice MD (Arm 3 vs Arm 2) at immediate post‐treatment: ‐2.70, 95% CI ‐6.88 to 1.48 MD (Arm 3 vs Arm 2) at long‐term follow‐up: 1.40, 95% CI ‐3.24 to 6.04 DISABILITY/FUNCTION: Chronic Pain Grade Questionnaire (CPGQ) 0 to 10 Baseline mean: Arm 1 4.4 (2.0), Arm 2 4.2 (1.8), Arm 3 4.3 (1.7), Arm 4 4.3 (1.7) Reported results: we found no consistent differences in pain‐related disability or sickness absence MD (Arm 3 vs Arm 2) at immediate post‐treatment: ‐0.01, 95% CI ‐0.45 to 0.43 MD (Arm 3 vs Arm 2) at long‐term follow‐up: 0.04, 95% CI ‐0.38 to 0.46 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: EuroQ‐5D‐3L 0 to 1 transformed to 0 to 100 scale Baseline mean: Arm 1 85.1, Arm 2 85.9, Arm 3 84.6, Arm 4 86.2 Reported results: no differences between groups were observed (all P values > 0.05) MD (Arm 3 vs Arm 2) at immediate post‐treatment: 0.50 (95% CI ‐1.58 to 2.58) MD (Arm 3 vs Arm 2) at long‐term follow‐up: 0.80, 95% CI ‐1.12 to 2.72 PATIENT SATISFACTION: not reported ADVERSE EVENT: 1 reported in massage group – highly bothersome dizziness COST OF CARE: not reported
Notes	Email ‐ response (31 March 2020) ‐ additional unpublished data provided on end of study mean (SD) at all time points. Analysis included Arm 3 vs Arm 2 (exercise) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 2 under 2.2 randomisation, ".... in blocks of 160 containing one of the three index groups......Cards indicating the treatment arm were placed in a sealed urn and randomly selected one by one to be placed in the envelopes." Comment: this was adequately described.
Allocation concealment (selection bias)	Low risk	Quote: Page 2 under 2.2 randomisation, "...sequentially numbered sealed envelopes in blocks of 160 containing one of the three index groups (deep tissue massage, strengthening and stretching exercises or, combined therapy) or the reference group (40 each). Cards indicating the treatment arm were placed in a sealed urn and randomly selected one by one to be placed in the envelopes." Comment: this was adequately described.
Blinding ‐ patient (performance bias)	High risk	Comment: it was not possible due to the nature of the interventions, massage vs exercises vs advice.
Blinding ‐ care provider (performance bias)	High risk	Comment: it was not possible due to the nature of the interventions, massage vs exercises vs advice.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: it was not possible due to the use of participant‐reported outcome measures. The participant was the outcome assessor.
Incomplete outcome data (attrition bias)	High risk	Comment: dropout ratio was acceptable, but no descriptions of dropouts were reported. Two self‐excluded patients were also noted.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Comment: insufficient reporting was noted.
Selective reporting (reporting bias)	Low risk	Quote: page 2 under 2.1 study design, setting and participants, "ISRCTN01453590". Comment: clinical trial number was reported and outcome measures listed match those reported.
Baseline similarities	Unclear risk	Comment: it was generally acceptable but also questionable due to the difference in the duration of pain and medication.
Similar or avoided co‐interventions	High risk	Quote: Page 8, "participants in the advice group used additional health services to a larger extent during the first three months (they spent on average 290 Euros, compared to 207–220 Euros in the other groups, p‐value < 0.001) which we believe may have underestimated the difference in early effect between advice and the other interventions." Comment: co‐intervention was identified as a limitation.
Acceptable compliance	High risk	Quote: Page 3 under results, "The average number of visits was 2.4 for advice to stay active, 5.8 for massage, 5.0 for exercise and 5.5 for massage and exercise." Comment: compliance was clearly reported but notably different between groups.
Similar outcome assessment	Low risk	Comment: the timing of the outcome assessment was similar at baseline, 7 weeks, 12 weeks, 26 weeks and 52 weeks in all groups.
Other potential sources of bias	High risk	Funding source: national institutes; conflict of interest: none. outcome measure: numeric pain rating scale, Chronic Pain Questionnaire, Euro Qual‐5D3L, valid; other: statistical error was suspected. Our senior statistician believed the analysis of this study was incorrect. It used 3 correlated contrasts as massage vs advice, exercise vs advice and massage + exercise vs advice; however, it did not have an inflated sample size for the advice group that should be root(3) = 1.73 times what the other 3 groups had and used a Dunnett test for testing and confidence intervals. Since they made the sample sizes the same, they should have used a 2 x 2 factorial design, and it would have had more precision for the main effects for exercise and massage, as well as being able to estimate the interaction between them. See the paper listed below showing what they did is NOT correct: Montgomery AA, Peters TJ, Little P: Design, analysis and presentation of factorial randomised controlled trials. BMC Medical Research Methodology 2003;3:26. HTTP://www.biomedcentral.com/1471‐2288/3/26. The Statistical Analysis Plan was not written as part of the trial registration even though the trial was properly registered. He thinks they may not have had adequate statistical help in the analysis and writing.

Stieven 2021.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 44/44 Intention‐to‐treat analysis: calculated Power analysis: calculated Funding source: none Declaration of interest: no competing interests
Participants	Disorder: non‐specific mechanical neck pain (myofascial pain syndrome) Radicular signs/symptoms: not reported Sex: 66% female Age: mean (SD): Arm 1 26.6 (3.9), Arm 2 27.1 (4.7), Control 28.2 (9.4) Severity: numeric pain rating scale 0 to 10, mean (SD): Arm 1 4.79 (1.19), Arm 2 4.50 (1.18), Control 4.71 (1.33) Duration of complaints: more than 3 months Setting: university’s laboratory in Porto Alegre Country: Brazil
Interventions	INDEX TREATMENT Arm 1: myofascial release. Activity: the therapist palpated the upper trapezius muscle using thumb and index fingers and patient was in a sitting position. Myofascial release was performed until a gentle release of the taut band was perceived by the therapist. The upper trapezius muscle was passively stretched for 30 seconds. The combination of myofascial release plus stretching was performed 3 times; Provider: physiotherapist; Mode: manual application; Dose: 270 seconds; Monitoring: not reported; Route: upper trapezius muscle COMPARISON TREATMENT Arm 2: dry needling. Activity: vertical postioning of the needle on the skin was performed with more than 3 times a local twitch response. The needle was inserted on the myofascial trigger point for approximately 10 seconds with no rotational movement applied; Provider: physiotherapist; Mode: dry needling; Dose: 30 seconds; Monitoring: not reported; Route: upper trapezius muscle Control: sham dry needling. Provider: physiotherapist; Activity: the therapist manipulated the tissues adjacent to the myofascial trigger point on the symptomatic side with a light touch using the thumb and index finger. No needle puncture applied; Mode: dry needling; Dose: 10 seconds; Monitoring: not reported; Route: neck muscles Treatment schedule: 1 session, 1 day Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: numeric pain rating scale 0 to 10 Baseline mean (SD): Arm 1 4.79 (1.19), Arm 2 4.50 (1.18), Control 4.71 (1.33) Reported results: 2 x 3 mixed‐model ANOVA showed significant reduction in pain scores immediately after the intervention for all 3 groups MD (Arm 1 vs Control) at immediate post‐treatment: ‐2.15, 95% CI ‐3.06 to ‐1.24 MD (Arm 1 vs Arm 2) at immediate post‐treatment: 0.14, 95% CI ‐0.62 to 0.9.0 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: no adverse effects were reported in all groups COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Control (sham) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 96, "the simple randomization procedure via Internet‐based randomization software (www.random.org)" Comment: randomisation software was used.
Allocation concealment (selection bias)	Unclear risk	Quote: Page 96, "using sealed and opaque envelopes" Comment: sequential numbering on the envelope was not fully described.
Blinding ‐ patient (performance bias)	High risk	Quote: Page 97, "Then the needle guide tube was positioned on the MTrP, where a gentle tap was performed. No needle puncture occurred.” Comment: there was no sham massage; sham acupuncture may not be believable as no needle puncture occurred and patient might be able to distinguish it from true acupuncture. a) Placebo description: the placebo was a needle‐guided tube with no needle puncture. This did not simulate the hands‐on element of the massage intervention. b) Contextual factors (process of care, positive framing may include reputation, confidence and therapeutic alliance): the context was described and similar. c) Patient factors: not reported. d) Practitioner factors: physiotherapist with no further detailing; not reported. e) Was the success of blinding tested among patients and successful?: not reported. They noted in the discussion section that they "did not varify the participants awarness regarding the treatment received if it was not a validated sham procedure".
Blinding ‐ care provider (performance bias)	High risk	Comment: blinding of the care provider was not possible due to the nature of the intervention.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: the participant was the outcome assessor due to self‐reported outcome measures but the participant was not credibly blinded in this study due to the type of sham.
Incomplete outcome data (attrition bias)	Low risk	Quote: Page 97, Figure 1. Comment: one‐time treatment, pre‐post design with no dropouts was performed.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Page 97, Figure 1. Comment: one‐time treatment, pre‐post design was performed.
Selective reporting (reporting bias)	High risk	Comment: the study protocol was available at https://ensaiosclinicos.gov.br/rg/RBR-9p4f7d. Secondary outcome, pain intensity, was never mentioned in the protocol. Only pressure pain threshold of upper fibre trapezius and elbow radial head were included.
Baseline similarities	Low risk	Quote: Page 98, Table 1. Comment: baseline characteristics were similar between groups.
Similar or avoided co‐interventions	Low risk	Quote: Page 97, Figure 1. Comment: one‐time treatment, pre‐post design implied no co‐intervention.
Acceptable compliance	Low risk	Quote: Page 97, Figure 1. Comment: one‐time treatment, pre‐post design was performed.
Similar outcome assessment	Low risk	Quote: Page 97, Figure 1. Comment: the timing of outcome assessment was at baseline and immediately post‐treatment.
Other potential sources of bias	Low risk	Funding source: none; conflict of interest: none; outcome measurement: numeric pain rating scale, valid.

Zhang Jishi 2013.

Study characteristics
Methods	Type of trial: RCT, restricted randomisation Number analysed/randomised: 76/76 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: chronic cervicogenic headache Radicular signs/symptoms: present 100% Sex: 52% female Age: mean (range): Arm 1 40.41 (22 to 66) years, Arm 2 41.11 (22 to 66) years Severity: visual analogue scale 0 to 10, mean (SD): Arm 1 7.35 (1.69), Arm 2 7.45 (1.17) Duration of complaints: mean (range): Arm 1 4.22 months (5 days to 14 months), Arm 2 4.15 months (10 days to 13 months) Setting: outpatient clinic of the Yukuan Hospital of Traditional Chinese Medicine Country: China
Interventions	INDEX TREATMENT Arm 1: Ashi point massage adjunct to warm needle acupuncture. Activity: after selection of the correct Ashi points in the lesion sites (neck, shoulder and back), a continuous sliding press with the thumb top was applied while gradually increasing the intensity of pressure; Provider: doctor of traditional Chinese medicine; Mode: soft tissue massage with hands; Dose: every point was manipulated for 30 seconds to 1 minute, with relaxation increments 15 seconds between; Monitoring: not reported; Route: Ashi points selected in the neck, back, shoulder regions COMPARISON TREATMENT Arm 2: warm needle acupuncture. Activity: acupuncture needle (diameter of.40 mm and length of 75 mm) were used. Points selected were Fengchi, Tianzhu and Fengfu. Patients were in prone and needles were inserted with the tip towards the interior and upper direction, straight into the bone; then drawn in small increments into the sub‐periosteum, with a total depth of no more than 2 cm to 2.5 cm. A protective towel was placed around the area, as well as 2 cm long moxa cones. 95% ethanol was applied to the moxa and then set on the needle handles for moxibustion. After the moxa burnt out and the needle bodies were cool, the needles could be pulled out; Provider: doctor of traditional Chinese medicine; Mode: using acupuncture needles and moxibustion. Dose: 10 sessions over every other day; Monitoring: not reported; Route: acupuncture points neck, back, shoulder regions Treatment schedule: 10 sessions, 3 weeks Duration of follow‐up: 3 months post treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale 0 to 10 Baseline mean: Arm 1 7.35, Arm 2 7.45 Reported results: significant difference favouring Arm 1 (P < 0.05) MD (Arm 1 vs Arm 2) at short‐term follow‐up: ‐2.98, 95% CI ‐3.17 to ‐2.79 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: no adverse effects were reported in all groups COST OF CARE: not reported
Notes	An email was written to request clarification of reported small SD (23 April 2022). We suspect an error in reporting where the SE can be confused with the SD.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Unclear risk	Quote: page 117, "They were randomised into the treatment group and the control group". Comment: randomisation process was not adequately reported.
Allocation concealment (selection bias)	Unclear risk	Comment: no information on allocation was reported.
Blinding ‐ patient (performance bias)	High risk	Comment: it was not possible due to the nature of the intervention; the interventions were distinguishably different, massage vs acupuncture.
Blinding ‐ care provider (performance bias)	High risk	Comment: it was not possible due to the nature of the intervention, massage vs acupuncture.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: due to the use of participant‐reported outcome measures (visual analogue scale), the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Low risk	Quote: Table 1. Comment: no dropouts were observed.
Randomised participants analysed in their groups (reporting bias)	Low risk	Quote: Table 1. Comment: no dropouts were observed, and all participants were analysed in the groups they were randomised to.
Selective reporting (reporting bias)	Unclear risk	Comment: no registration or protocol was reported.
Baseline similarities	Low risk	Quote: Page 117 under clinical data and table 2, "ln the treatment group (11 = 40), patients (19 males and 21 females) ranged ín age from 25 to 66....In the control group (17 = 36) patients (17 males and 19 females) ranged in age from 20 to 66..." Comment: baseline characteristics between groups were observed to be comparable.
Similar or avoided co‐interventions	Unclear risk	Comment: no data on co‐intervention were reported.
Acceptable compliance	Unclear risk	Comment: no data on treatment compliance were reported.
Similar outcome assessment	Low risk	Comment: the timing of outcome assessment was similar between groups at baseline and 3 months for all participants.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: visual analogue scale valid; other: we suspected a reporting error. The standard deviations were unusually small and we suspected these might be standard errors. We have written to the author for data clarification.

Zhang Jun 2013.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 60/60 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: chronic non‐specific neck pain (cervical spondylosis) Radicular signs/symptoms: not reported Sex: not reported Age: ranged between 18 and 60 years Severity: VAS 0 to 10: more than 3 points Duration of complaints: more than 3 months Setting: hospital Country: China
Interventions	INDEX TREATMENT Arm 1: Massage adjunct to dynamic moxibustion. Activity: for massage, patients were seated. A bilateral neck and shoulder massage was performed using kneading and rolling methods. Pressure point areas were released. Cervical vertebrae received traction with a palm holding technique. This was followed with gentle rubbing of neck and shoulder; Provider: practitioner of the traditional Chinese medicine; Mode: soft tissue massage with hands and moxibustion with Chinese herbal medicine; Dose: 5 treatments every other day; 3 to 4 sessions per week; 2 courses of treatment; Monitoring: not reported; Route: neck and shoulder COMPARISON TREATMENT Arm 2: dynamic moxibustion. Activity: patients were seated with their neck exposed. The ignited moxa was wrapped with gauze and soaked with traditional Chinese medicine and immediately pressed onto selected acupoints. Manipulations were performed, mainly the rubbing method, pressing method and point‐pressing methods; Provider: practitioner of traditional Chinese medicine; Mode: manual application with moxibustion; Dose: 5 treatments every other day; 3 to 4 sessions per week; 2 courses of treatment; Monitoring: not reported; Route: neck and shoulder Treatment schedule: 10 sessions, 3 weeks Duration of follow‐up: 1 day immediate post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: curative rate ‐ Ninmodipine Score, 4‐point scale; we dichotomised into cured vs marked improvement/improved/not improved McGill Scale: not reported Baseline mean: not reported Reported results: risk ratio indicates significant protective effect favouring Arm 1 compared to Arm 2 for number of individuals labelled as cured RR (curative rate ‐ Ninmodipine Score) at immediate post‐treatment: 0.46, 95% CI 0.29 to 0.73 FUNCTION/DISABILITY: not reported PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: Neck Pain Questionnaire, 0 to 36: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Attempted to contact authors on 3 January 2016: invalid email address. Unable to obtain data for McGill Scale and Neck Pain Questionnaire.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 123 under clinical data, "According to the random number table method, they were divided into the treatment and control group."; in the abstract "randomly divided". Comment: the randomisation process was acceptable.
Allocation concealment (selection bias)	Unclear risk	Comment: it was not adequately described.
Blinding ‐ patient (performance bias)	High risk	Comment: it was not possible due to the nature of the intervention and perceivable difference between interventions, massage vs moxibustion.
Blinding ‐ care provider (performance bias)	High risk	Comment: it was not possible due to the nature of the intervention, massage vs moxibustion.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: due to the use of patient‐reported outcome measures, the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Low risk	Comment: there were no dropouts.
Randomised participants analysed in their groups (reporting bias)	Unclear risk	Comment: insufficient information was provided.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol or registration was reported.
Baseline similarities	Unclear risk	Comment: no baseline data were reported.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not reported.
Acceptable compliance	High risk	Comment: compliance was not reported.
Similar outcome assessment	Unclear risk	Comment: it was unclear when outcome measures were recorded.
Other potential sources of bias	High risk	Funding source: not reported; conflict of interest: not reported; outcome measurement: no data for outcomes of Neck Pain Questionnaire and McGill Scale were not reported, validity of curative rate was unclear; other: there was a lack of statistical reporting and end of study data.

Zibiri 2019.

Study characteristics
Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 35/41 Intention‐to‐treat analysis: not calculated Power analysis: not reported Funding source: none Declaration of interest: no competing interests
Participants	Disorder: chronic non‐specific neck pain Radicular signs/symptoms: not reported Sex: 65.7% female Age: mean (SD): Arm 1 49.50 (17.50), Arm 2 42.00 (14.58), Arm 3 49.27 (11.32) Severity: numeric pain rating scale 0 to 10, mean (SD): Arm 1 5.92 (2.11), Arm 2 7.67 (1.83), Arm 3 5.36 (1.91) Duration of complaints: more than 3 months Setting: general outpatient department, physiotherapy department at 2 tertiary hospitals Country: Nigeria
Interventions	INDEX TREATMENT Arm 1: muscle energy technique adjunct to conventional treatment (neck care education and infrared radiation). Activity: a therapist localised the joint or the body tissue into the position of the initial range of motion resistance to a specific movement. A participant’s cervical spine was positioned in a restricted range and a resistive force was applied. Then the patient was asked to isometrically contract for 5 seconds, and then the counter‐force by the therapist ceased slowly, and the participant was asked to relax. The therapist then took the joint to a new barrier and the same procedure was repeated 3 times; Provider: physiotherapist; Mode: manual application; Dose: 15 seconds with 3 repetitions; Monitoring: not reported; Route: neck COMPARISON TREATMENT Arm 2: neck stabilisation exercise adjunct to conventional treatment (neck care education and infrared radiation). Activity: chin tuck, cervical extension, shoulder shrugs, shoulder rolls, scapular retraction 15 repetitions each for 30 minutes; Provider: physiotherapist; Mode: neck stabilisation exercise; Dose: 30 minutes; Monitoring: not reported; Route: chin, cervical, shoulder, scapular area Arm 3: conventional treatment (neck care education and infrared radiation). Activity: participants were advised to relax, activate and adopt stress‐coping skills, workplace ergonomics and self‐care strategies; Provider: physiotherapist; Mode: education and modality; Dose: not reported; Monitoring: not reported; Route: neck Treatment schedule: 16 sessions, 8 weeks Duration of follow‐up: 1 day immediately post‐treatment CO‐INTERVENTION: conventional treatment including neck care education, infrared radiation
Outcomes	PAIN: numeric pain rating scale 0 to 10 Baseline mean (SD): Arm 1 5.92 (2.11), Arm 2 7.67 (1.83), Arm 3 5.36 (1.91) Reported results: significant difference between the pre‐ and post‐treatment interventions for the outcome measures across the groups MD (Arm 1 vs Arm 2) at immediate post‐treatment: ‐0.08 95% CI ‐0.84 to 0.68 MD (Arm 1 vs Arm 3) at immediate post‐treatment: ‐1.43 95% CI ‐2.41 to ‐0.45 FUNCTION/DISABILITY: Neck Disability Index 0 to 50 Baseline mean: Arm 1 28.67 (14.53), Arm 2 34.50 (13.81), Arm 3 27.27 (11.32) Reported results: a significant difference between the pre‐ and post‐treatment interventions for the outcome measures across the groups MD (Arm 1 vs Arm 2) at immediate post‐treatment: ‐4.83 95% CI ‐8.88 to ‐0.78 MD (Arm 1 vs Arm 3) at immediate post‐treatment: ‐4.20 95% CI ‐13.02 to 4.62 PARTICIPANT‐REPORTED TREATMENT SUCCESS: not reported HEALTH‐RELATED QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Analysis included Arm 1 vs Arm 3 (conventional care) and the risk of bias judgements were based only on these 2 groups.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate randomisation	Low risk	Quote: Page 2, "using the computer‐generated random number sequence" Comment: computer‐generated random numbers was used.
Allocation concealment (selection bias)	Unclear risk	Comment: this was not described.
Blinding ‐ patient (performance bias)	High risk	Comment: it was not possible due to the nature of the interventions; the interventions are distinguishable (massage vs exercise, or neck care education and infrared radiation).
Blinding ‐ care provider (performance bias)	High risk	Comment: it was not possible due to the nature of the interventions, massage vs exercise, or neck care education and infrared radiation.
Blinding outcome assessor ‐ self‐reported outcomes (detection bias)	High risk	Comment: due to the use of patient‐reported outcome measures (numeric pain rating scale, Neck Disability Index), the participant was the outcome assessor.
Incomplete outcome data (attrition bias)	Unclear risk	Quote: Page 3, Figure 1. Reason for dropout "indisposed, transferred" Comment: the dropout rate was acceptable at 15% (n = 6 of 41) but the reason for the dropouts is unclear; more detail was needed for the terms "indisposed, transferred".
Randomised participants analysed in their groups (reporting bias)	High risk	Quote: Page 3, Figure 1. Comment: no intention‐to‐treat analysis was applied.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol or trial registration was reported.
Baseline similarities	High risk	Quote: Page 5, Tables 1 and 2. Comment: differences in group 3 for pain intensity and Neck Disability Index were observed; duration of pain was not reported.
Similar or avoided co‐interventions	Unclear risk	Comment: co‐intervention was not described.
Acceptable compliance	Unclear risk	Comment: compliance was not described.
Similar outcome assessment	Low risk	Quote: Page 2, "Assessment of pain, neck disability, depression, anxiety, and sleep disturbance was done at baseline, end of 4th and 8th weeks" Comment: the timing of outcome assessment was similar between groups at baseline, 4 weeks and 8 weeks.
Other potential sources of bias	Unclear risk	Funding source: not reported; conflict of interest: none; outcome measurement: Neck Disability Index, numeric pain rating scale, Insomnia Severity Index, Hospital Anxiety Depression Scale, psychometric information or relevant references not provided.

MD: mean difference
NDI: Neck Disability Index
NPRS numeric pain rating scale
NNTB: number needed to treat for an additional beneficial outcome
RCT: randomised controlled trial
RR: risk ratio
SD: standard deviation
TCTM: traditional Chinese therapeutic massage
VAS: visual analogue scale

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abdel‐Aal 2021	Intervention: massage applied with an instrument
Abu Taleb 2016	Intervention: use of massage instrument
Ahmad 2015	Participant: chronic brachial neuralgia; there was no indication that this brachial pain originated from the neck
Alghadir 2020	Ineligible clinical condition: inappropriate measurement using pain pressure threshold
Amini 2017	Intervention: surrogate measurement variable
Ammer 1990	Intervention: multimodal, which includes massage
Antunez Sanchez 2016	Intervention: multimodal approach includes massage; we are unable to determine the unique contribution of the massage. The treatment consisted of individual treatment: infrared heat therapy, massage and passive stretching. The treatment group consisted of a programme of active mobilisation, isometric contractions, self‐stretching and postural recommendations.
Arab 2018	Ineligible comparison: exercise
Arjona 2021	Intervention: massage applied with an instrument
Arsh 2020	Intervention: multimodal approach includes massage
Bakar 2014	Ineligible clinical condition: inappropriate measurement using only pain pressure threshold
Battecha 2021	Intervention: massage was applied for all intervention groups
Bernal‐Utrera 2020	Intervention: multimodal approach includes massage
Boonruab 2021	Ineligible comparison: exercise
Braun 2011	Intervention: use of a massage instrument, not manual
Brodin 1985	Intervention: multimodal, which includes massage
Bukhari 2016	Intervention: use of cervical traction
Buyukturan 2018	Intervention: massage in both arms
Büyükturan 2021	Intervention: multimodal approach includes massage
Cabak 2016	Ineligible clinical condition: measurement variable does not apply to all the arms in the study
Cabak 2017	Ineligible clinical condition: surrogate measurement variable
Cabrera‐Martos 2020	Intervention: multimodal approach that includes self‐myofascial release
Campa‐Moran 2015	Ineligible comparison: exercise and manual therapy; dry needling
Celenay 2016b	Intervention: mechanical traction
Celenay 2021	Ineligible comparison: kinesiotape
Cheng 2021	Ineligible comparison: massage was in both treatment arms
Cohen 2014	Intervention: multimodal approach includes massage
Coppieters 2000	Intervention: the main mode of therapy was cervical joint mobilisation using neural tension positioning as a secondary part of the treatment. No distinction was made between the two modalities in the data.
Costello 2016	Ineligible comparison: ultrasound
Cui 2017	Intervention: manipulation of cervical spine
Daher 2021	Intervention: massage was applied for all intervention groups
De Meulemeester 2017	Intervention: use of a massage instrument
Desmoulin 2012	Intervention: massage instrument (Khan Kinetic Treatment Device) used
Ding 2012	Intervention: joint manipulations; comparison: inappropriate
Dissanayaka 2016	Intervention: effects of myofacial release treatment cannot be extracted.
Domingues 2019	Intervention: intervention is multimodal manual therapies
Donoyama 2010	Ineligible clinical condition: healthy participants without pain
Durianova 1977	Ineligible clinical condition: the measure used was not clearly stated
El‐Gendy 2019	Intervention: intervention is multimodal manual therapies
Elserty 2020	Ineligible clinical condition: inappropriate measurement using pain pressure threshold
Espi‐Lopez 2016	Ineligible clinical condition: both arms include massage, so it is difficult to get insight on the massage intervention
Fernandez 2006	Ineligible comparison: transverse friction massage
Fernandez‐delasPenas 2004a	Intervention: multimodal, which includes massage
Fernandez‐delasPenas 2004b	Intervention: multimodal, which includes massage; craniosacral and myofascial trigger point manual therapy
Fernandez‐Perez 2012	Intervention: trigger point identification manual pressure assessment
Ferrante 2005	Intervention: both the active and control group received myofascial release technique
Fialka 1989	Intervention: multimodal, which includes massage
Field 2014	Ineligible design: quasi‐RCT
Fitz‐Ritson 1994	Ineligible clinical condition: unsure, sample not adequately described (query whiplash‐associated neck disorder) Irrelevant intervention: no soft tissue therapy was used
Fryer 2005	Ineligible clinical condition: neck tenderness not pain
Gam 1998	Intervention: multimodal, which includes massage
Garcia‐Perez‐Juana 2018	Intervention: massage was not the active treatment
Gemmell 2008a	Ineligible comparison: activator trigger point manipulation
Ghodrati 2020	Intervention: intervention is multimodal manual therapies
Gillani 2020	Ineligible comparison: exercise
Guo 2013	Ineligible comparisons: acupuncture; traditional massage
Gurumoorthy 2000	Intervention: no soft tissue therapy used
Haas 2010	Intervention: both arms received massage
Haas 2010b	Intervention: massage was used in both arms as an attention‐control. Quote: "To relax the neck and upper back in preparation for spinal manipulation, the chiropractor administered a moist heat pack for 5 minutes and conducted a light massage for 2 minutes."
Hadamus 2021	Ineligible clinical condition: inappropriate measurements used, i.e. sleep quality
Hakkinen 2007	Intervention: multimodal including massage
Haleema 2021	Intervention: massage was applied for all intervention groups
Haller 2014	Ineligible clinical condition: no pain or function data reported
Hanten 1997	Ineligible clinical condition: neck tenderness was examined instead of pain intensity
Hanten 2000	Intervention: mechanical massage device ‐ J Knob instrument used as massage technique and education versus education
Hemmila 2005	Intervention: multimodal including massage; massage and bone setting
Hou 2002	Intervention: For Stage 1: 8 women were randomised to each group. 14 trigger points were assessed per group. Although the unit of randomisation was 8 women per group, the analysis was done on 14 trigger points per group ‐ a different unit of analysis. For Stage 2: all treatments were multimodal with one treatment item being massage.
Hoving 2002	Intervention: multimodal, which includes massage
Iaroshevskyi 2019	Intervention: multimodal approach includes massage
Ibrahim 2021	Intervention: massage was applied for all intervention groups
Iqbal 2020	Intervention: massage was applied in both intervention arms
Jahanshahi 1991	Ineligible clinical condition: no sample with neck disorder meeting inclusion criteria (torticollis)
Jamal 2016	Intervention: multiple arms in this trial; massage is in the control treatment of all arms
Jellad 2009	Intervention: multimodal, which includes massage; manual traction, which includes standard care
Jiang 2012	Intervention: cervical fixed point traction and manipulation
Jordan 1998	Intervention: multimodal which includes massage
Junaid 2020	Intervention: multimodal approach applied
Karlberg 1996	Intervention: multimodal, which includes massage
Kaur 2017	Design: quasi‐RCT
Khan 2022	Ineligible comparison: one massage technique versus another massage technique
Kim 2018	Ineligible clinical condition: inappropriate measurement using pain pressure threshold
Kim SA 2013	Ineligible clinical condition: pressure algometry
Kim TH 2012	Intervention: instrument massage (cupping)
Kirthika 2021	Ineligible comparison: kinesiotape; manual therapy
Klaber Moffett 2006	Intervention: multimodal, which includes massage; this is the control arm of this trial
Koes 1991	Intervention: multimodal, which includes massage
Kogstad 1978	Intervention: multimodal, which includes massage
Kojidi 2016	Ineligible clinical condition: participants had latent symptoms
Kojidi 2016a	Ineligible clinical condition: participants had latent symptoms
Konig 2003	Ineligible clinical condition: no pain, range of motion
Lalruatliana 2021	Intervention: multimodal approach includes massage
Lauche 2011	Intervention: instrumental massage (cupping)
Lauche 2012	Intervention: instrument massage (cupping)
Lauche 2013	Intervention: instrument massage (cupping)
Leboeuf 1987	Ineligible clinical condition: no sample with neck disorder meeting inclusion criteria (repetitive strain injury of upper limb)
Lee 2014	Ineligible clinical condition: no pain, tenderness pressure algometry
Lee 2019	Intervention: used an instrument ‐ S. alba seeds
Levoska 1993	Intervention: multimodal which includes massage
Lin 2004	Intervention: both groups received massage
Lin 2012	Intervention: massage manipulation vs control (oral ginseng treatment) for 4 weeks; the intervention appeared to be two massage techniques plus exercise (stretch) of the cervical spine
Llamas‐Ramos 2014	Ineligible comparison: acupuncture
Madson 2010	Intervention: no comparison to control group
Manzoor 2017	Ineligible clinical condition: healthy participants with no current neck pain
Manzoor 2021	Ineligible comparison: manual therapy
McReynolds 2005	Intervention: multimodal, which includes massage
Meseguer 2006	Ineligible clinical condition: tenderness
Mezaki 1995	Inappropriate study design: unsure RCT Ineligible clinical condition: no participants with neck disorder meeting inclusion criteria (spasmodic torticollis)
Mishra 2018	Ineligible comparison: one massage versus another massage
Mohamadi 2017	Ineligible clinical condition: surrogate measurement variable
Moraska 2017	Ineligible clinical condition: surrogate measurement variable
Moretti 2004	Intervention: multimodal approach includes massage
Morikawa 2017	Intervention: excluded because non‐interventional trial exploring biological mechanisms related to myofacial trigger point release
Moustafa 2016	Intervention: multimodal approach with no arm having massage only as an intervention
Moustafa 2018	Intervention: used a massage instrument ‐ denneroll
Murphy 2010	Intervention: multimodal approach includes massage
Mylonas 2021	Intervention: massage was applied with an instrument
Nalawade 2020	Intervention: multimodal approach includes massage
Nasb 2020	Intervention: use of instrument (cupping massage)
Nilsson 1997	Intervention: multimodal, which includes massage
Oliveira‐Campelo 2013	Ineligible clinical condition: latent trigger points, no neck pain
Osama 2020	Ineligible comparison: muscle energy technique; exercise
Otadi 2020	Ineligible comparison: dry needling and patient education
Paanalahti 2016	Intervention: population also had back pain and no arm had only massage therapy as intervention
Padula 2018	Ineligible clinical condition: participants had pain at arm, neck, shoulder and lower back
Palmgren 2006	Intervention: multimodal, which used massage (myofascial technique)
Paoloni 2013	Intervention: multimodal approach includes massage
Parkin‐Smith 1997	Intervention: unclear how many participants received "non therapeutic pre manipulative soft tissue massage" for muscle spasm
Patil 2020	Ineligible comparison: one massage versus another massage technique
Persson 2001	Intervention: multimodal, which used massage; massage use varied between patients in physiotherapy group
Picelli 2011	Ineligible comparison: mobilisation and exercise
Pillastrini 2018	Intervention: intervention is multimodal manual therapies
Provinciali 1996	Intervention: multimodal, which used massage
Pérez‐Cabezas 2020	Intervention: multimodal approach includes massage
Pérez‐Martínez 2020	Intervention: massage instrument (INYBI) used
Razzaq 2020	Intervention: multimodal approach includes massage
Reginiussen 2000	Intervention: multimodal, which used massage
Rodriguez‐Fuentes 2016	Ineligible comparison: manual therapy
Rodriguez‐Huguet 2018	Intervention: multimodal approach, where massage cannot be differentiated
Rodríguez‐Huguet 2020	Intervention: massage instrument (vacuum myofascial therapy) used
Rodríguez‐Huguet 2020a	Intervention: multimodal approach includes massage
Rohe 2015	Ineligible comparison: one massage versus another massage technique
Rotter 2020	Intervention: multimodal approach includes massage
Rudolfsson 2014	Ineligible comparison: exercise coordination; exercise strengthening
Ruiz‐Molinero 2014	Intervention: massage was part of control
Saadat 2018	Intervention: no massage in the traditional physiotherapy
Sachdeva 2019	Ineligible comparison: mobilisation
Sadria 2017	Ineligible comparisons: one massage technique versus another massage technique
Saha 2017	Intervention: instrumental massage
Sarrafzadeh 2012	Inappropriate study design: not an RCT
Savva 2016	Intervention: use of cervical traction
Schenk 1994	Ineligible clinical condition: no sample with neck disorder meeting inclusion criteria (normal cervical spine)
Schnabel 2002	Intervention: multimodal, which used massage
Schumann 2012	Intervention: instrumental massage (cupping)
Schwerla 2008	Intervention: multimodal approach includes massage
Shady 2021	Ineligible comparison: exercise
Shafique 2019	Intervention: intervention is multimodal manual therapies
Shah 2013	Intervention: traction
Sherman 2009	Ineligible comparison: education
Simoni 2021	Intervention: multimodal approach includes massage
Skargren 1997	Intervention: multimodal, which used massage; only 36% of physiotherapy group received massage
Stephens 2020	Intervention: instrument used for massage
Strunk 2009	Inappropriate study design: not an RCT
Sun 2014	Ineligible comparison: acupuncture
Susana 2020	Intervention: self‐application with auto‐acupressure
Topolska 2012	Inappropriate study design: not an RCT
Tozzi 2011	Ineligible clinical condition: trial participants included those with neck and lumbar pain. We wrote to the author (contacted date 4 December 2016); there was no response received when we rechecked 20 May 2020. We therefore excluded the trial.
Turel 2015	Intervention: multimodal approach includes massage
van Dongen 2016	Intervention: multimodal approach includes massage
Vasseljen 1995	Intervention: no soft tissue therapy used
Vassiliou 2006	Intervention: multimodal, including massage
Vavrek 2010	Ineligible clinical condition: mixed conditions including tension‐type headache, migraine, cervicogenic headache
Ventegodt 2004	Intervention: multimodal including massage; combination of alternative therapies ‐ gestalt therapy, Rosen Body Work, craniosacral therapy
Walker 2013	Ineligible clinical condition: patient had spinal pain of unspecified origin
Wang 2015	Comparison: massage is in both arms (massage + magnetic sticking at acupuncture points vs massage)
Wu 2019	Ineligible comparison: one type of massage versus another type
Yagci 2004	Ineligible clinical condition: tenderness
Yip 2006	Intervention: multimodal, including massage
Ylinen 2007	Intervention: multimodal approach includes massage
Youssef 2013	Ineligible comparisons: mobilisation
Zaproudina 2007	Ineligible comparison: conventional physical therapy; mobilisation
Zhang 2005a	Intervention: both groups received massage
Zhang 2005b	Ineligible clinical condition: it is unclear which construct the outcome Cervical Spondylopathy Treatment Effect Rating Scale is measuring; it is a composite of clinical symptoms, physical examination and activities of daily life. Clarification on the outcome was sought but not resolved.
Zhang 2021	Intervention: multimodal approach includes massage
Zhang Jiafu 2011	Ineligible comparison: mechanical traction
Zhou 2005	Intervention: multimodal massage with acupuncture. Unable to extract massage data only.
Ziaeifar 2019	Ineligible comparison: acupuncture ‐ dry needling
Zuo 2008	Ineligible comparison: mechanical traction
Zutshi 2021	Intervention: massage was compared to a multimodal approach
Zylbergold 1985	Intervention: not clear if the manual traction used a halter or was performed manually

Characteristics of studies awaiting classification [ordered by study ID]

Buttagat 2021.

Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 45/45 Intention‐to‐treat analysis: calculated Power analysis: calculated, 90% Funding source: none Declaration of interest: no competing interests
Participants	Disorder: myofascial pain syndrome Radicular signs/symptoms: absent Sex: 66.67% female Age: mean (SD): Arm 1 22 (1.60), Arm 2 21.3 (1.29), Control 20.3 (1.29) Severity: visual analogue scale 0 to 10: Arm 1 4.93 (0.96), Arm 2 5.07 (1.03), Control 4.20 (0.77) Duration of complaints: more than 3 months Setting: Department of Physical Therapy, School of Integrative Medicine, Mae Fah Luang University; primary Country: Thailand
Interventions	INDEX TREATMENT Arm 1: traditional Thai massage: Activity: 8 x 15‐minute sessions of traditional Thai massage over 2 weeks applied to the posterior neck area. Gentle thumb pressing was gradually applied to the traditional Thai massage points for 5 seconds each until the patient started to feel slight discomfort. The therapist repeated this sequence several times for each massage point. Moreover, 3 x 30‐second sessions of neck stretching for the left and right sides (1.5 minutes for the left side; 1.5 minutes for the right side) were passively applied at the end of the intervention; Mode: soft tissue massage with hands; Dosage: 15 minutes; Monitoring: not reported; Route: neck COMPARISON TREATMENT Arm 2: muscle energy technique: Activity: 8 sessions of muscle energy technique treatment over a 2‐week period. The post‐isometric relaxation technique was applied to the neck extensor muscles for 3 repetitions using 20% of maximal isometric contraction. In a supine position, the participants' necks were passively flexed to the point where resistance to the movement was first noted by the therapist. The participants were asked to contract the neck extensor muscles with a mild resisted effort (approximately 20% of maximal isometric contraction) for 7 seconds against an opposing force, then a 30‐second passive stretch to the palpated barrier. It was repeated 3 times with rest periods of approximately 4 minutes after a set of muscle energy technique treatment; Mode: manual application with hands; Dosage: 15 minutes; Monitoring: not reported; Route: neck Control: no intervention; Monitoring: not reported Treatment schedule: 2 weeks, 8 sessions Duration of follow‐up: immediate post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: visual analogue scale, 0 to 10 Baseline mean (SD): Arm 1 4.93 (0.96), Arm 2 5.07 (1.03), Control 4.20 (0.77) Reported results: a significantly greater improvement in these parameters for the Arm 1 and Arm 2 technique group compared with the control group; no significant difference was noted between the Arm 1 and Arm 2 technique groups MD (Arm 1 vs Control) at short‐term follow‐up: ‐37.40, 95% CI ‐43.49 to ‐31.31 MD (Arm 2 vs Control) at short‐term follow‐up: ‐39.40, 95% CI ‐45.80 to ‐33.00 FUNCTION/DISABILITY: Neck Disability Index, 0 to 50 Baseline mean (SD): Arm 1 27.13 (10.72), Arm 2 26.00 (8.13), Control 21.20 (7.75) Reported results: a significantly greater improvement in these parameters for the Arm 1 and Arm 2 technique group compared with the control group; no significant difference was noted between the Arm 1 and Arm 2 technique groups. MD (Arm 1 vs Control) at short‐term follow‐up: ‐33.88, 95% CI ‐43.55 to ‐24.21 MD (Arm 2 vs Control) at short‐term follow‐up: ‐38.80, 95% CI ‐49.48 to ‐28.12 GLOBAL PERCEIVED EFFECT: not reported QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Extremely small SDs were identified. We believe there are errors in the reporting of the SD. We have written to the author (April 2022).

Matthieu 2013.

Methods	RCT with 3 arms
Participants	Patients with mechanical neck pain
Interventions	Manipulation group vs ischaemic compression group vs control
Outcomes	Pressure pain threshold
Notes	Only abstract available ‐ search completed 20 May 2020; wrote to the author for full‐text manuscript

Monteiro de Sobral 2010.

Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 20/20 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: not reported
Participants	Disorder: non‐specific neck disorder/mechanical neck pain Radicular signs/symptoms: not reported Sex: 100% female Age: mean (SD): Arm 1 33.5 (5.0), Control 31.50 (5.4) Severity: numeric pain rating scale 0 to 10, mean (SD): Arm 1 9.30 (0.82), Control 9.20 (0.79) Duration of complaints: not reported Setting: not reported Country: Brazil
Interventions	INDEX TREATMENT Arm 1: positional release therapy: Activity: first identification of sensitive points in the cervical muscles and classification of these points in decreasing order of intensity of pain. Technique is first performed at the point of the greatest sensitivity. Pressure is maintained on the point and then the neck of the patient is passively and slowly placed in a comfortable position. The positioning and digital pressure is maintained for an interval of 90 seconds and after this time, the digital pressure is released and the participant's head is passively and slowly repositioned to the neutral position. Each point has the technique repeated to obtain a minimal response of relief of symptoms of 70% at the end of the session; Mode: soft tissue massage with hands; Dosage: 10 sessions 3 times per week; Monitoring: not reported; Route: cervical spine and shoulder COMPARISON TREATMENT Control: no treatment applied Treatment schedule: 3 weeks (10 sessions) Duration of follow‐up: immediate post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN numeric pain rating scale 0 to 10 Baseline mean: Arm 1 9.30, Control 9.20 Reported results: differences were found to be significant at study end in the two analysis groups for pain, range of motion and muscle strength. In the analysis of pain by NPRS, 100% of patients in the intervention group reported improvement at the end of treatment. In the control group, 20% of patients reported decreased pain, 40% remained the same and 40% worsened. MD (Arm 1 vs Control) at immediate post‐treatment: ‐8.20, 95% CI ‐8.85 to ‐7.55 FUNCTION/DISABILITY: not reported GLOBAL PERCEIVED EFFECT: not reported QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	Portuguese ‐ translation completed by a Portuguese health science student and Google translator. On 16 April 2022, we wrote the author to request clarification of the SD, which seems extremely small. We suspect a reporting error.

NCT03585283.

Methods	RCT with 2 arms
Participants	Participants with mechanical neck pain
Interventions	Myofascial group vs sham group
Outcomes	Pain pressure threshold on trigger points with algometer, disability, perceived pain in the cervical region, global rating of change, cervical ROM
Notes	Study completed: 20 March 2019; checked 28 April 2020 and 1 January 2023: no study report available

Stuner 2016.

Methods	Type of trial: RCT, parallel‐group design Number analysed/randomised: 10/10 Intention‐to‐treat analysis: calculated Power analysis: not reported Funding source: not reported Declaration of interest: no competing interests
Participants	Disorder: non‐specific mechanical neck pain (myofascial pain syndrome) Radicular signs/symptoms: absent Sex: 60% female Age: mean (SD): Arm 1 70.4 (10.3), Control 63 (6.2) Severity: not reported Duration of complaints: not reported Setting: not reported Country: France
Interventions	INDEX TREATMENT Arm 1: ischaemic compression: Activity: patient was supine. Therapist applied a trigger point technique on the upper trapezius within the comfortable pain limit for 90 seconds twice; Mode: manual application; Dosage: 180 seconds; Monitoring: not reported; Route: upper trapezius muscle COMPARISON TREATMENT Control: sham (compression on deltoid muscle): Activity: compression was applied on the deltoid as a control treatment for 90 seconds twice; Mode: manual application; Dosage: 180 seconds; Monitoring: not reported; Route: deltoid Treatment schedule: 1 session Duration of follow‐up: immediately post‐treatment CO‐INTERVENTION: not reported
Outcomes	PAIN: VAS 0 to 10 Baseline mean: not reported Reported results: significant change in pain of 2.14 (SD 0.71) in Arm 1 and 0.06 (SD 0.09) in the sham group FUNCTION/DISABILITY: not reported GLOBAL PERCEIVED EFFECT: not reported QUALITY OF LIFE: not reported PATIENT SATISFACTION: not reported ADVERSE EVENT: not reported COST OF CARE: not reported
Notes	We wrote to the authors for data clarification on 7 June 2021. The end of study SDs are extremely small: Arm 1 7.32 (SD 0.32) and Sham 0.30 (SD 0.17). The mean is also high for the treatment group and the opposite direction of effect, contrasting with what is reported in the manuscript. The data suggest increased pain in the treatment group. Additionally, even when using the RevMan calculator to convert the data from SE to SD the SDs are extremely small ‐ far below all other trials. We believe there is a data reporting error.

NDI: Neck Disability Index; RCT: randomised controlled trial; ROM: range of motion; SD: standard deviation; SF36: short form 36; VAS: visual analogue scale

Characteristics of ongoing studies [ordered by study ID]

González Rueda 2017.

Study name	Effectiveness of a specific manual approach to the suboccipital region in patients with chronic mechanical neck pain and rotation deficit in the upper cervical spine: study protocol for a randomized controlled trial
Methods	RCT with 3 arms
Participants	78 participants with chronic mechanical neck pain and rotation deficit in the upper cervical spine
Interventions	Pressure group vs mobilisation group vs control group
Outcomes	Visual analogue scale, Neck Disability Index, cervical range of motion, headache intensity, global rating of change
Starting date	Not reported
Contact information	Email: martinbarra@gmail.com Jordi Gol Institute of Research on Primary Health Care, Barcelona, Spain
Notes	—

Raja 2020.

Study name	Effectiveness of deep cervical fascial manipulation and yoga postures on pain, function and oculomotor control in mechanical neck pain: A pragmatic, parallel group, randomized, controlled trial
Methods	RCT
Participants	Mechanical neck pain
Interventions	Deep cervical fascial manipulation
Outcomes	Pain
Starting date	Registered on 24 January 2020; clinical trial registry India ‐ CTRI/2020/01/022934
Contact information	Raja PG, Bhat S, Fernández De Las Peñas C, Gangavelli R, Davis F, Shankar R
Notes	Search for update 29 May 2020; trial registered, no additional information

Silva 2020.

Study name	The effect of tongue and suprahyoid muscles release in the treatment of chronic non‐specific neck pain: study protocol for a randomized controlled trial
Methods	RCT
Participants	Mechanical chronic neck pain
Interventions	Manual therapy of neck and thorax, neck stretching, scapular girdle muscle strengthening, scapulothoracic coordination, proprioceptive and postural training + tongue and suprahyoid muscle release vs manual therapy neck and thorax, neck stretching, scapular girdle muscle strengthening, scapulothoracic coordination, proprioceptive and postural training
Outcomes	Pain (numerical rating scale), function‐disability (NDI), cervical range of motion
Starting date	Not reported
Contact information	Email: politti@uni9.pro.br
Notes	NDI: Neck Disability Index; RCT: randomised controlled trial

NDI: Neck Disability Index; RCT: randomised controlled trial

Differences between protocol and review

Our protocol was published in 2004 (Haraldsson 2004), following our first publications that included massage (Gross 1996).

For our update in 2012 (Patel 2012), multimodal care (massage combined with other treatment) and quasi‐RCTs were removed from this review.

We explicitly used the Cochrane Musculoskeletal protocol template method (12 December 2022) to write this review update. The MECIR Standards and the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2021) guided the methodology of this update. We excluded trials where massage was compared to other treatments. We formally incorporated disorder classifications and duration into the heading "disorder characteristics" and added formal subgroup analysis (post hoc) by dose response; note that in our original protocol we considered factors like plausibility and dose response relationship to make final decisions. We believe that adding one 'formal' subgroup analysis clarifies our findings and makes these clinically more meaningful. We conducted sensitivity analysis by random‐effects versus fixed‐effect model post hoc to be consistent with the Cochrane Musculoskeletal protocol template.

We updated the terminology used for types of participants to "neck pain" (this was formerly mechanical neck disorder, simple or non‐specific neck pain; trauma related such as whiplash‐associated disorders I/II or work‐related injury; we clarified that radiculopathy included whiplash‐associated disorder III; neck pain with cervicogenic headaches). The diagnostic criteria for participants did not change. We clarified the exclusion criterion "fractures" to also mean grade 4 neck pain.

We clarified the type of massage interventions to be manual touch and excluded the use of instrumentation for massage application.

We classified measures of treatment effect as small, medium and large (Chou 2017).

Contributions of authors

Our primary author contributions were as follows:

• Co‐ordinators: Anita R Gross, Nejin Chacko, Geoff Gelley.

• Methodological quality assessment: Geoff Gelley, Mario Forget, P Lina Santaguida, other Cervical Overview Group Validity Team members.

• Study identification and selection: Nejin Chacko, Nadine Graham, Anita Gross, Craig Dixon, student groups from Silla University and McMaster University.

• Research librarian: Maureen Rice.

• Grants: P Lina Santaguida, Joy MacDermid, Anita R Gross, other Cervical Overview Group members.

• Data abstraction and synthesis, manuscript preparation: Haejung Lee, Jeanette Ezzo, Craig Dixon, Nejin Chacko, Annie Morien, student groups.

• Final synthesis, public responsibility, administration: Anita R Gross, Haejung Lee, Jeanette Ezzo, Nejin Chacko, Annie Morien, Geoff Gelley, Mario Forget, Lina Santiguida, Maureen Rice, Craig Dixon, Nadine Graham.

• Conferences and publication: Haejung Lee, Anita Gross, student members.

Sources of support

Internal sources

• McMaster University, Department of Clinical Epidemiology and Biostatistics, School of Rehabilitation Science, Occupational Health Program, Canada

  Research time, library, methodological experts, technology access
Conflict of interest: none

• Silla University, Department of Physical Therapy, Korea, South

  Research time, library, methodological experts, technology access
Conflict of interest: none

• Royal Canadian Chiropractic College, Canada

  Research time, library, technology access
Conflict of interest: none

• University of Manitoba, Faculty of Health Sciences, College of Rehabilitation Sciences, Canada

  Research time, technology support
Conflict of interest: none

External sources

• Canadian Institute of Health Research ‐ Meetings, Planning & Dissemination Grant: Knowledge Translation Supplement, Canada

  Funding period: 2011‐10/2012‐10. Partial financial support was provided for the Cervical Overview Group foundational to the linking of knowledge translation efforts.
Conflict of interest: none

• Canadian Physiotherapy Association, Physiotherapy Foundation, Canadian Academy Musculoskeletal Physical Therapy Student Research Fund, Canada

  Funding period: 2018, 2020, 2021
Conflict of interest: none

• Canadian Institute of Health Research ‐ Meetings, Planning & Dissemination Grant: Planning Grant, Canada

  Funding period: 2011‐06/2012‐03. Partial financial support was provided for the Cervical Overview Group foundational to the linking of the knowledge translation efforts underpinning the dissemination of information through International Collaboration on Neck.
Conflict of interest: none

• Canadian Institute of Health Research – Knowledge Synthesis Grant, Canada

  Funding period: 2010‐01/2010‐12. Partial funding was attained through linking knowledge translation efforts. International Collaboration on Neck ‐ Synthesizing research evidence into clinical recommendations for managing neck pain.
Conflict of interest: none

• Industry Partner Grant ‐ Lifemark Health, Hamilton, Canada

  Funding period: 2010‐6/2011‐06. Partial funding was received by the Cervical Overview Group from an industry partner. Lifemark Health is an interdisciplinary health care provider employing interdisciplinary healthcare team/professionals: physicians, physiotherapists, chiropractors, occupational therapists, massage therapists, speech language pathologists.
Conflict of interest: the researchers had full possession of the systematic review process from planning to reporting without this funder, with potential conflict of interest, having any possibility to interfere in the process. Anita Gross is a physiotherapist employed (part‐time as a clinician) in one of the industry partner's community services.

• Canadian Institute of Health Research ‐ Meetings, Planning & Dissemination Grant: Knowledge Translation, Canada

  Funding period: 2009‐05/2010‐04. Partial funding was attained through linking knowledge translation efforts.
Conflict of interest: none

• Sunnybrook and Women's College Health Sciences Foundation ‐ Problem‐based Research Award, Toronto, Canada

  Funding period: 2004‐03/2006‐06. Conservative management of mechanical neck disorders: a series of systematic reviews.
Conflict of interest: none

• Hamilton Hospital Association, Hamilton, Canada

  Funding period: 2002‐1/2003‐1. Conservative management for neck disorders: a series of systematic reviews.
Conflict of interest: none

• National Institute of Health ‐ Consortial Center for Chiropractic Research, Bethesda, MD, USA

  Funding period: 2002‐2/2003‐2. Conservative management for neck disorders: a series of systematic reviews.
Conflict of interest: none

• Chedoke‐McMaster Foundation ‐ Hamilton Health Sciences Corporation, Hamilton, Canada

  Funding period: 1997‐10/1999‐3. Conservative management of neck disorders: a meta‐analysis.
Conflict of interest: none

• University of Saskatchewan ‐ Clinical Teaching and Research Award, Canada

  Funding period: 1997‐9/1998‐10. Conservative management of neck disorders: a meta‐analysis.
Conflict of interest: none

• McGregor Fund, Hamilton Foundation, Hamilton, Canada

  Funding period: 1993‐3/1994‐2. Conservative management of mechanical neck disorder: a systematic overview and meta‐analysis.
Conflict of interest: none

Declarations of interest

The review authors have had full possession of the review process from planning to reporting without funders with potential conflict of interest having any possibility to interfere in the process. No conflicts of interest are declared by each of the authors: Anita Gross, Haejung Lee, Jeanette Ezzo, Nejin Chacko, Geoff Gelley, Mario Forget, Annie Morien, Nadine Graham, Pasqualina L Santaguida, Maureen Rice, Craig Dixon

Anita Gross declares that she was a part‐time employee of Lifemark Health up to 2018. Lifemark is an interdisciplinary healthcare service and was the Cervical Overview Group’s industry partner for one year (2010 to 2011). We acknowledge support of office supplies and telephone costs made in kind for that year.

Craig Dixon has the following acknowledgement on his form: Health and Performance Centre ‐ University of Guelph ‐ Eramosa Physiotherapy Associates.

Edited (no change to conclusions)