Mechanical traction for neck pain with or without radiculopathy

Abstract

Background

Neck pain is a frequently reported complaint of the musculoskeletal system which can be disabling and costly to society. Mechanical traction is often used as an adjunct therapy in outpatient rehabilitation.

Objectives

To assess the effects of mechanical traction for neck disorders.

Search methods

A research librarian searched computerized bibliographic databases without language restrictions up to March 2008 for randomized controlled trials (RCTs) from the medical, chiropractic, and allied health literature.

Selection criteria

The RCTs we selected examined adults with neck disorders who received mechanical traction alone or in combination with other treatments compared to a placebo or another treatment. Our outcomes of interest were pain, function, disability, global perceived effect, patient satisfaction, and quality of life measures.

Data collection and analysis

Two review authors with different backgrounds in medicine, physiotherapy, massage therapy and chiropractics independently conducted study selection, risk of bias assessment and data abstraction using pre‐piloted forms. We resolved disagreement through consensus.

Main results

Of the seven selected RCTs (total participants = 958), only one (N = 100) had a low risk of bias. It found no statistically significant difference (SMD ‐0.16: 95%CI: ‐0.59 to 0.27) between continuous traction and placebo traction in reducing pain or improving function for chronic neck disorders with radicular symptoms. Our review found no evidence from RCTs with a low potential for bias that clearly supports or refutes the use of either continuous or intermittent traction for neck disorders.

Authors' conclusions

The current literature does not support or refute the efficacy or effectiveness of continuous or intermittent traction for pain reduction, improved function or global perceived effect when compared to placebo traction, tablet or heat or other conservative treatments in patients with chronic neck disorders. Large, well conducted RCTs are needed to first determine the efficacy of traction, then the effectiveness, for individuals with neck disorders with radicular symptoms.

Plain language summary

Mechanical traction for neck pain with or without symptoms that radiate to the neck or arm

Twenty‐six to 71% of the adult population can recall experiencing an episode of neck pain or stiffness in their lifetime. Neck pain is more common in females than in males, with rates reported as high as 77.8%. The natural history is unclear. Neck pain has a costly impact on society because of visits to healthcare providers, sick leave, disability and loss of productivity. There are a number of treatments available for neck pain, one of which is mechanical traction.

Mechanical traction 'stretches' the neck. With the patient lying on their back, a head halter is placed under the back of the head and possibly the jaw and attached to a machine. The machine is set for a certain time period and specific weight for the pulling action to occur. The traction can remain on steadily for the specified time (referred to as continuous or static) or intermittently (on/off cycle) during the treatment session. Experts think that traction expands the space between the vertebrae, increases the movement of the joints and stretches the muscles and ligaments around the vertebrae. Side effects are generally few, but can include headache, nausea, fainting and injury to tissue.

We included seven randomized controlled trials (RCT), with a total of 958 participants, that looked at the effects of continuous or intermittent mechanical traction for individuals with chronic neck pain (lasting for more than three months). Some had symptoms that radiated into the arms and head (radicular symptoms), others did not.

Only one RCT (100 participants) had a low risk of bias, which means we can have confidence in the findings. This trial found that on average, there was no statistically significant difference between continuous and placebo traction in reducing pain or improving function for individuals with chronic neck disorder with radicular symptoms.

In summary, our review found no evidence from RCTs with a low potential for bias that clearly supports or refutes the use of either continuous or intermittent traction for individuals with chronic neck disorders.

Background

Neck pain is a frequently reported complaint of the musculoskeletal system which can be disabling and costly to society. The natural history of neck pain is unclear. Twenty‐six to 71% of the adult population can recall experiencing an episode of neck pain or stiffness in their lifetime (Cote 1998; Makela 1991). Prevalence of neck pain is higher in females than in males with rates as high as 77.8% (Chiu 2006; Fejer 2006; Guez 2006). A U.S. study from the National Ambulatory Medical Care Survey reported an average of 10.2 million visits to healthcare facilities for neck pain (Riddle 2007). Neck pain has a large impact on health care expenditure, attributed to visits to health care providers, sick leave, disability and the related loss of productivity (Borghouts 1999; Hoving 2001).

Mechanical traction for the cervical spine involves a pulling force applied to the neck via a mechanical system which can be applied intermittently or continuously (Kisner 1996). It is often used as an adjunct therapy in outpatient rehabilitation (Michlovitz 2005). The physiological effects of mechanical traction for the cervical spine may include separation of vertebral bodies, movement of facet joints, expansion of intervertebral foramen and stretching of soft tissue (Kisner 1996; Michlovitz 2005). Although traction may be a common treatment approach for neck pain, its effectiveness remains limited or inconclusive (Gross 2002; Graham 2006; Kjellman 1999; Peeters 2001; van der Heijden 1995).

Objectives

This systematic review was undertaken to assess the effects of mechanical traction, either alone or in combination with other treatments in adults with neck pain with or without radicular symptoms. Where appropriate, we assessed the influence of three factors: quality of study methodology, symptom duration, and subtypes of the disorder.

Methods

Criteria for considering studies for this review

Types of studies

We included any published or unpublished (including abstracts, unpublished forms of submitted or accepted manuscripts, thesis and grey literature) randomized controlled trial (RCT).

Types of participants

The participants were adults, aged 18 years or older, with acute (less than 30 days), sub‐acute (30 to 90 days) or chronic (longer than 90 days) neck disorders categorized as:

• Neck disorders with radicular symptoms including whiplash associated disorders (WAD) Grade 3 (Spitzer 1995);

• Neck disorders with headache (Olesen 1988; Olesen 1997; Sjaastad 1990); and

• Neck disorders, including WAD Grades 1 and 2 (Spitzer 1995), myofascial neck pain, and degenerative changes (Schumacher 1993).

We excluded studies if they investigated neck disorders with:

• definite or possible long tract signs (eg. myelopathies);

• neck pain caused by other pathological entities (eg. rheumatoid arthritis, ankylosing spondylitis), according to Schumacher 1993;

• neck pain related to neurological disease (eg. spasmotic torticollis);

• neck pain related to fractures and dislocations; and

• headache not of cervical origin.

Types of interventions

RCTs using mechanical traction techniques, alone or combined with other therapies and contrasted against a placebo or another treatment were included. Mechanical traction for the neck involves a stretch being applied via a head halter which is placed under the occiput and possibly the jaw. The machine is then set for a certain time period and specific weight for the pulling action to occur. The traction can remain on steadily for the specified time (referred to as continuous or static) or intermittently (on/off cycle) during the treatment session.

We considered these comparisons:

• traction versus sham or placebo (eg. traction versus placebo; traction + sham ultrasound versus sham ultrasound);

• traction versus no treatment or wait list;

• traction plus another intervention versus that same intervention (eg. traction + exercise versus exercise);

• traction versus another intervention (eg. traction versus exercise);

• one type of traction versus another type (eg. intermittent versus continuous); and

• dose of traction treatment versus another dose of the same traction treatment (eg. nine sessions of traction over three weeks versus three sessions of same type of traction over three weeks).

Types of outcome measures

Pain relief (eg. Visual Analog Scale (VAS), Numerical Rating Scale (NRS)), functional or disability measures including work‐related disability (eg. Neck Disability Index (NDI), Activity of Daily Living (ADL), return‐to‐work, patient satisfaction, global perceived effect and quality of life measure (eg. SF‐36, Euro‐Q)), were our main outcomes of interest. Adverse effects and cost of treatments were also retrieved where available. The follow‐up periods were defined as immediate post‐treatment (up to one day after treatment), short‐term (more than one day, up to three months), intermediate‐term (more than three and less than 12 months), and long term (one year or longer).

Search methods for identification of studies

A research librarian searched computerized bibliographic databases, without language restrictions, for medical, chiropractic and allied health literature. Our search was part of a comprehensive search on patient education, manual therapies, drug therapies and physical medicine modalities in which mechanical traction is included. We searched the following databases from their beginnings to March 2008: CENTRAL (The Cochrane Library 2008, Issue 1), MEDLINE (January 1966 to March 2008), EMBASE (January 1980 to March 2008), Manual Alternative and Natural Therapy (1985 to March 2008), Cumulative Index to Nursing and Allied Health Literature (CINAHL: January 1982 to March 2008), and Index to Chiropractic Literature (1980 to March 2008). We also consulted personal files, screened references, and communicated with the Cochrane Back Review Group and content experts. Unpublished literature from specific conference proceedings was scanned by authors attending relevant society meetings. Subject headings (MeSH) and key words included anatomical terms, disorder or syndrome terms, treatment terms, and methodological terms, consistent with those recommended by the Cochrane Back Review Group (van Tulder 2003) (See Additional Table 1).

Data collection and analysis

Two review authors with different backgrounds in medicine, physiotherapy, massage therapy and chiropractics independently conducted citation identification, study selection, and data abstraction using pre‐piloted forms. We resolved disagreement through consensus. Prior to consensus, we calculated agreement between review authors for study selection using the Kappa statistic. Measures less than zero indicate poor agreement, 0 to 0.20 slight agreement, 0.21 to 0.40 fair agreement, 0.41 to 0.60 moderate agreement, 0.61 to 0.80 substantial agreement and 0.81 to 1.00 almost complete agreement (Landis 1977).

Assessment of Methodological Quality

Two review authors independently assessed each selected study for risk of bias using (i) the Back Review Group's 11 criteria (van Tulder 2003) where an RCT was rated as high quality (low risk of bias) if six or more criteria were met (see Additional Table 1): (ii) the validated 5‐point Jadad 1996 criteria where a high score is three or more out of five (see Additional Table 1): and (iii) the Cochrane grading for concealment of allocation (A to D). The latter is reported in the Characteristics of Included Studies table.

1. Methodological Quality Criteria/Risk of Bias sources.

The van Tulder et al criteria and scores are as follows: (Total Score 11; high quality =6) Was the method of randomisation adequate? A random (unpredictable) assignment sequence. Examples of adequate methods are computer‐generated random numbers table and use of sealed opaque envelopes. Methods of allocation using date of birth, date of admission, hospital numbers, or alternation should not be regarded as appropriate. Was the treatment allocation concealed? 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. Was the patient blinded to the intervention? The review author determines if enough information about the blinding is given in order to score a "yes." Was the care provider blinded to the intervention? The review author determines if enough information about the blinding is given in order to score a "yes." Was the outcome assessor blinded to the intervention? The review author determines if enough information about the blinding is given in order to score a "yes." Was the drop‐out rate described and acceptable? 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 drop‐outs does not exceed 20% for immediate and short‐term follow‐ups, 30% for intermediate and long‐term follow‐ups and does not lead to substantial bias a "yes" is scored. Did the analysis include an intention‐to‐treat analysis? All randomized patients are reported/analyzed in the group to which they were allocated by randomization for the most important moments of effect measurement (minus missing values), irrespective of noncompliance and co‐interventions. Were the groups similar at baseline regarding the most important prognostic indicators? In order to receive a "yes," 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). Were co‐interventions avoided or similar? Co‐interventions should either be avoided in the trial design or be similar between the index and control groups. Was the compliance acceptable in all groups? The review author determines if the compliance to the interventions is acceptable, based on the reported intensity, duration, number and frequency of sessions for both the index intervention and control intervention(s). Was the timing of the outcome assessment in all groups similar? Timing of outcome assessment should be identical for all intervention groups and for all important outcome assessments.

The Jadad et al. criteria and scores are as follows: 1a. Was the study described as randomised? (Score 1 if yes); 1b and 1c. Was the method of randomisation described and appropriate to conceal allocation? (Score 1 if appropriate and ‐1 if not appropriate); 2a. Was the study described as double‐blinded? (Score 1 if yes); 2b and 2c. Was the method of double blinding described and appropriate to maintain double blinding? (Score 1 if appropriate and ‐1 if not appropriate); 3. Was there a description of how withdrawals and dropouts were handled? (Score1 if yes)

Analysis of Trial Results

We calculated standardized mean differences and 95% confidence intervals (SMD: 95%CI) for continuous outcomes using a random‐effects model. SMD was used because different measures are frequently used to address the same clinical outcome. A random effects model incorporates the variability both within and between studies, and as a result has wider confidence limits, providing an appropriate level of caution for readers. Generally, an effect size can be interpreted as small (0.20), medium (0.50) or large (0.80), as defined by Cohen 1988. We selected 10‐mm change in pain on a 100‐point scale (10%) as the minimal clinically important difference (MCID) (Goldsmith 1993). A minimal clinically important difference of five neck disability index units on a 50‐point scale (10%) has been established for the neck disability index (Stratford 1999).

We calculated the relative risk (RR) for dichotomous outcomes. A RR less than one represents a beneficial treatment. When neither continuous nor dichotomous data were available, we noted the findings and the statistical significance as reported by the author(s) in the original RCT. Data from controlled trials did not permit calculation of the number‐needed‐to‐treat (NNT) and treatment advantage for primary findings (Gross 2002). However, the baseline mean, end of study mean, absolute benefit, RR, and SMD are noted in the Characteristics of Included Studies table. Power analyses was conducted for each article reporting non‐significant findings (Dupont 1990). The desired level of significance was set at alpha equal to 0.05 and adequate power was defined as at least 80%. Although power analyses is most appropriate a priori for estimating sample size, we felt that the level of insufficiency determined through a post‐hoc power analyses was useful. 
 
 Prior to calculation of a pooled effect measure, a multidisciplinary team examined the reasonableness of pooling on clinical and biological grounds. If at least two trials were clinically homogeneous, we considered pooling their results. These possible sources of heterogeneity that were considered are:

• participant differences;

• symptom duration (acute versus chronic);

• subtype of neck disorder (especially WAD), intervention type (continuous, intermittent traction);

• characteristics of treatment (eg. dosage, technique); and

• outcomes (pain reduction, measures of function and disability, employment status, patient satisfaction, quality of life).

Methodological study quality was also considered as a possible source of heterogeneity. Although sensitivity analysis and meta‐regression were planned to explore the relationship between these key characteristics, they could not be performed due to insufficient data. 
 
 Regardless of whether there were sufficient data available to use quantitative analyses to summarize the data, we assessed the overall quality of the evidence for each outcome. To accomplish this, we used the GRADE approach, as recommended by the Back Review Group (Furlan 2008). The quality of the evidence for a specific outcome was based on the study design, risk of bias, consistency of results, directness (generalizability), precision (sufficient data) and reporting of the results across all studies that measure that particular outcome. The quality was considered high when RCTs with low risk of bias provided consistent, direct and precise results for the outcome, and reduced by a level for each of the domains not met.

High quality evidence = there are consistent findings among 75% (at least two high quality) of the RCTs with low potential for bias that are generalizable to the population in question. There are sufficient data, with narrow confidence intervals. There are no known or suspected reporting biases. 
 Moderate quality evidence = one of the domains is not met 
 Low quality evidence = two of the domains are not met 
 Very low quality evidence = three of the domains are not met 
 No evidence = no evidence from RCTs

Results

Description of studies

We included seven RCTs from 564 citation postings:

• seven RCTs (958 participants) studied neck disorders with some radicular signs and symptoms of the following duration: chronic (Klaber Moffett 1990; Shakoor 2002), mixed duration (Wong 1997; Zylbergold 1985), and not defined (Brewerton 1966). Wong 1997 had a duplicate publication and Klaber Moffett 1990 had sequential publications representing a single RCT.

• two studied mixed neck disorders of the following duration: subacute and chronic (Kogstad 1978), not defined (Guanygue 2001).

There were no RCTs included that examined headaches of a cervical origin.

See Characteristics of Included Studies table for specific data outlining treatment characteristics, co‐intervention, concealment of treatment allocation, baseline values, absolute benefits, reported results, SMD, RR, side effects and costs of care.

No abstracts were found for this review. No additional RCTs were yielded through searching the grey literature or from contacting content experts and no unpublished RCTs were identified. One RCT (Kogstad 1978) required translation. Pre‐piloted selection forms classified findings as "yes", "no" and "unsure". Selection phase agreement between pairs of independent review authors from varied professional backgrounds was K_w 0.86, SD 0.10 which indicates almost complete agreement (Landis 1977).

We excluded eight RCT's for the following reasons:

• two RCTs did not include our outcomes of interest (Akinbo 2006; Joghataei 2004).

• three were quasi‐RCTs (Goldie 1970; Loy 1983; Pennie 1990).

• one RCT examined mechanical traction as part of a combined therapy and traction doses were not provided (Fialka 1989).

• one RCT had a traction group with less than 15% participants as part of a multimodal treatment (Skargren 1998).

• one RCT used traction as a standardized co‐intervention for both index and comparison groups (Jensen 1995).

We also found one published protocol (Lamb 2007). This RCT is ongoing and will be assessed for inclusion when the results have been published.

Risk of bias in included studies

Risk of Bias assessment of included RCTs

Overall, our selected RCTs were rated as having a high risk of bias using either the 11 criteria listed in van Tulder 2003 (Mean 4.1, S.D. 1.9) or the 5 items listed in Jadad 1996 (Mean 2, S.D. 1.1). The major risk of bias were: inadequate or poorly described concealment of allocation (reported in 0/7) and methods of randomization (reported in 1/7), blinding of patients, care providers and outcome assessors was poor in all but one trial (Klaber Moffett 1990) and intention‐to‐treat analyses were not included. Most RCTs (6/7) had appropriately accounted for withdrawals and dropouts appropriately (see Figure 1 and Table 2 for risk of bias assessments). The quality of the evidence for the main results did not vary by the quality assessment system used.

1.

Risk of Bias summary: review authors' judgments about each methodological quality item for each included study.

2. Jadad Scale 5‐point scale.

Author	1a‐randomized	1b‐concealment app	1c‐inappropriate	2a‐double blind	2b‐appropriate	2c‐inappropriate	3 follow‐up	Total Score
Brewerton 1966	1	0	0	0	0	0	1	2/5
Guanguye 2001	1	0	0	1	0	‐1	1	2/5
Klaber Moffett 1990	1	0	0	1	1	0	1	4/5
Kogstad 1978	1	0	‐1	0	0	0	0	0/5
Shakoor 2003	1	0	0	0	0	0	1	2/5
Wong 1997	1	0	0	0	0	0	1	2/5
Zylbergold 1985	1	0	0	0	0	0	1	2/5

Effects of interventions

It was not possible to perform a meta‐analysis for any comparisons in our review. Quantitative data can be found in the Characteristics of Included Studies table.

1.0 Traction versus placebo

1.1 Continuous Traction

We selected two RCTs comparing continuous traction to a placebo for chronic neck pain with radicular symptoms (Brewerton 1966; Klaber Moffett 1990). Klaber Moffett added an intervention of education (neck school), collar and drugs to both the continuous traction and placebo arms. Clinically, we judged the types of traction in the two RCTs to be heterogeneous.

Pain

There was low to moderate quality evidence (2 RCTs, 287 participants) that there were no statistically significant difference in pain reduction reported by those with chronic neck pain with radicular symptoms who received continuous traction (138 participants) and those who received a placebo tablet (44 participants) (RR: 0.90: 95% CI: 0.77 to 1.04), placebo heat‐untuned short‐wave diathermy (62 participants) (RR: 1.00: 95% CI: 0.85 to 1.18) (Brewerton:tracvPl(t), Brewerton:tracvPl(h)) or placebo traction (43 participants) (SMD: ‐0.16: 95% CI: ‐0.59 to 0.27) (Klaber Moffett 1990) when measured immediately post‐treatment and short‐term respectively (see Comparison 01).

Function/Disability

There was moderate quality evidence (1 RCT, ADL group = 77 participants, Social Dysfunction group = 82 participants) that there were no statistically significant differences in function or disability reported by those with chronic neck pain with radicular symptoms who received continuous traction and those who received placebo traction when measured in the short‐term (ADL: SMD ‐0.39: 95% CI: ‐0.84 to 0.06, Social Dysfunction: SMD: 0.16: 95% CI: ‐0.27 to 0.60) (Klaber Moffett 1990). There was low quality evidence (1 RCT, placebo heat group = 62 participants, placebo tablet group = 44 participants) that there were no statistically significant difference in ability to work reported by those with chronic neck pain with radicular symptoms who received continuous traction, placebo heat or placebo tablet immediately post‐treatment (RR: 0.87: 95% CI: 0.53 to 1.44), RR: 0.87: 95% CI: 0.53 to 1.44, respectively) (Brewerton:tracvPl(h); Brewerton:tracvPl(t)) (See Comparison 01).

1.2 Intermittent Traction

No studies were found.

Summary

Two RCTs, one with low potential for bias and one with high potential for bias separately showed that when different types of continuous traction were compared to different placebos, there were no statistical differences in reducing pain or improving function for patients with chronic neck disorders with radicular symptoms.

2.0 Traction plus another intervention versus that same intervention

2.1 Continuous Traction

Pain

There was low quality evidence (1 RCT, 25 participants each arm) that there were no statistically significant differences in pain reduction reported by those with neck disorders with radicular symptoms, degenerative changes or both who received continuous traction compared to exercise, heat and patient education (SMD: ‐0.22: 95% CI: ‐0.78 to 0.34) (Zylbergold: 1 v 4) when measured immediately post‐treatment (see Comparison 02).

2.2 Intermittent Traction

Pain

There was low quality evidence (1 RCT, 25 participants each arm) that those with neck disorders with radicular symptoms, degenerative changes or both who received intermittent traction reported less pain than those who received exercise, heat and patient education (SMD: ‐0.78: 95%CI: ‐1.36 to 0.21) (Zylbergold: 2 v 4) when measured immediately post‐treatment (see Comparison 04).

Summary

There was low quality evidence that intermittent traction, but not continuous traction, was more effective than exercise, heat and patient education for pain reducing pain for individuals with neck disorders with radicular findings, degenerative changes or both.

3.0 Traction versus another intervention

There was low quality evidence for the results from all of the following comparisons. Data are available in the Characteristics of Included Studies table.

3.1 One type of traction versus another type

Pain

There was no statistical difference between groups when either intermittent (SMD: ‐0.39: 95% CI: ‐0.95 to 0.17) (Zylbergold: 2 v 3) or continuous (SMD: 0.17: 95% CI: ‐0.39 to 0.72) (Zylbergold: 1 v 3) traction was compared to manual traction or to each other (SMD: ‐0.49: 95% CI: ‐1.06 to 0.07) (Zylbergold: 2 v 1) for neck disorders with radicular findings, degenerative changes or both, in multiple arms of a single RCT (see Comparison 06).

Global Perceived Effect

There was no statistical difference between groups (1 RCT, 24 participants) when intermittent traction was compared to an adaptive intermittent traction modality with EMG Biofeedback (RR: 1.12: 95% CI: 0.67 to 1.89) (Wong 1997) (see Comparison 06).

3.2 Traction versus other interventions

Pain

There was no statistical difference between groups (1 RCT, 218 participants) when continuous traction was compared to a non‐steroidal anti‐inflammatory drug (Naproxen 250 mg) for pain reduction (SMD: ‐0.26: 95% CI: ‐0.54 to 0.01) (Shakoor 2002) (see Comparison 05).

Global Perceived Effect

There was no statistical difference between groups when, in a small RCT (N = 50), intermittent traction was compared to manual therapy (RR: 0.33: 95% CI: 0.08 to 1.32) (Kogstad: CT v MT) for neck disorders with radicular symptoms, headache or both (see Comparison 05). One large RCT (N = 536) showed that continuous traction was inferior to acupuncture for those with neck disorders with or without radicular symptoms (RR: 4.31: 95% CI: 2.93 to 6.34) (Guanygue 2001) (see Comparison 03).

Summary

Low quality evidence suggests that acupuncture may be superior to continuous traction and on head‐to‐head comparisons, one type of traction did not seem superior to another.

Adverse Events

Side effects were reported in one RCT (Klaber Moffett 1990) in which two patients reported headaches following traction. Although there is a possibility of other harmful effects, such as nausea, fainting and injury to tissue when using mechanical traction (Kisner 1996; Weinberger 1976), the included RCTs did not report any of these side effects.

Cost of Care

Cost of care using traction was not assessed in these RCTs.

Discussion

Risk of Bias

Overall, high risk of bias in the primary RCTs limited this review. Only one RCT (Klaber Moffett 1990) was rated as having a low risk of bias using both the van Tulder and Jadad scales. Blinding, concealment of allocation and methods of randomization were either not adequate or poorly described and analyses did not include intention‐to‐treat. Follow‐up and drop‐out rates were described well and considered acceptable in all but one study (Kogstad 1978). A point can be made however, that five of the seven included RCTs were published from 1966 to 1990 when reporting guidelines were different than today's standards. 
 
 With many rehabilitation interventions, blinding of patients becomes an issue. It may not be feasible to blind subjects, since even with "placebo" traction, they may be able to tell if a force was being applied. If subjects sense there is no pull, they may not feel that their symptoms will improve and this may increase the potential for bias. Many of the outcomes assessed are self‐reports, therefore the patient is the outcome assessor and is unable to be blinded. 
 
 Due to heterogeneity of disorder types, interventions and outcomes, we were unable to pool the results. However, when continuous traction was compared to three different placebos for chronic neck disorders with radicular symptoms, two RCTs separately demonstrated similar findings for the effect of traction on pain, function and disability (Brewerton 1966; Klaber Moffett 1990). Since both RCTs were underpowered and Brewerton 1966 was rated as having a high risk of bias, these results would need to be confirmed through larger, better conducted RCTs.

Comparison against other review findings

Three other reviews (Kjellman 1999; van der Heijden 1995; Verhagen 2004) showed that either no clear conclusions could be drawn or there was no benefit of traction for neck disorders. This is consistent with lumbar traction reviews (Clarke 2007, Pellecchia 1994, van der Heijden 1995). Our previous review (Graham 2006) reported one positive clinical conclusion ‐ that intermittent traction may be favourable for chronic neck disorders with or without radicular symptoms for short‐term pain relief ‐ but this review included quasi‐RCTs with high risks of bias. We agree that standardization of traction dosage and clinical parameters have not been clearly reported and high risk of bias identified in this current review affects our confidence in the results.

Adverse events and cost of care

Cost of care and rates of adverse events could not be determined with these data. Clearly, there is inadequacy in reporting these data. Authors should follow CONSORT guidelines to standardize RCT reporting.

Other Methodological Issues

'Selection bias' was not likely to be present in our review. We used pairs of independent review authors from diverse professional backgrounds to select the studies. We did not search non‐English databases, so 'Language bias' may be present in this review. Although unpublished work was sought through conference proceedings, personal files and grey literature, we did not search all databases for unpublished work nor did we write authors or agencies to elicit unpublished work, therefore 'Publication bias' was not completely guarded against either. Using the GRADE approach for single studies and developing a forest plot for one study may be excessive and does not provide added value to the results.

Authors' conclusions

Implications for practice.

The current literature does not support or refute the efficacy or effectiveness of continuous or intermittent traction for pain reduction, improved function or global perceived effect over placebo traction, tablet or heat or other conservative treatments in patients with chronic neck disorders.

Implications for research.

Large, well conducted RCTs are needed that evaluate the effect of traction as an adjunct treatment in rehabilitation.

What's new

Date	Event	Description
19 January 2011	Amended	Contact details updated.

History

Protocol first published: Issue 1, 2007
 Review first published: Issue 3, 2008

Date	Event	Description
23 November 2009	Amended	Contact details updated.
13 May 2008	Amended	Converted to new review format.

Appendices

Appendix 1. Search Strategies for EMBASE and MEDLINE

EMBASE

1 exp neck/ 
 2 exp Cervical Plexus/ 
 3 exp Cervical Spine/ 
 4 Atlantoaxial Joint/ 
 5 Atlantooccipital Joint/ 
 6 AXIS/ 
 7 ATLAS/ 
 8 "Spinal Root"/ 
 9 exp Brachial Plexus/ 
 10 (odontoid or cervical or occip: or atlant:).tw. 
 11 or/1‐10 
 12 exp ARTHRITIS/ 
 13 exp myofascial pain syndromes/ 
 14 fibromyalgia/ 
 15 spondylitis/ 
 16 exp spinal osteophytosis/ 
 17 spondylolisthesis/ 
 18 exp headache/ and cervic:.tw. 
 19 whiplash injuries/ 
 20 cervical rib syndrome/ 
 21 torticollis/ 
 22 cervico‐brachial neuralgia.mp. 
 23 exp radiculitis/ 
 24 POLYRADICULITIS/ 
 25 polyradiculoneuritis/ 
 26 thoracic outlet syndrome/ 
 27 (monoradicul: or monoradicl:).tw. 
 28 or/12‐27 
 29 random:.tw,sh. 
 30 double‐blind method/ 
 31 Single Blind Procedure/ 
 32 placebos/ 
 33 exp clinical trials/ 
 34 (clin: adj25 trial:).tw. 
 35 ((singl: or doubl: or trebl: or tripl:) adj25 (blind: or mask:)).tw. 
 36 placebo:.tw. 
 37 or/29‐36 
 38 meta‐analysis.sh,pt. or meta‐analy:.tw. or metaanal:.mp. 
 39 ((systematic: or quantitativ:) adj (review: or overview:)).tw. 
 40 (cochrane or medline or cinahl or embase or scisearch or psychinfo or psycinfo or psychlit or psyclit or (national and library)).tw. 
 41 ((handsearch: or search:) and (cochrane or medline or cinahl or embase or scisearch or psychinfo or psycinfo or psychlit or psyclit or (national and library) or (hand: or manual: or electronic: or bibliograph: or database:))).tw. 
 42 ((review or guideline).pt. or consensus.ti. or guideline:.ti. or literature.ti. or overview.ti. or review.ti.) and (40 or 41) 
 43 ((synthesis or overview or review or survey) and (systematic or critical or methodologic: or quantitative or qualitative or literature or evidence or evidence‐based)).ti. 
 44 38 or 39 or 41 or 42 or 43 
 45 44 not ((case: or report:).ti. or editorial.pt. or comment.pt. or letter.pt.) 
 46 exp combined modality therapy/ 
 47 exercise/ 
 48 exertion/ 
 49 exp exercise therapy/ 
 50 exp electric stimulation therapy/ 
 51 transcutaneous electric nerve stimulation/ 
 52 exp rehabilitation/ 
 53 ultrasonic therapy/ 
 54 exp phototherapy/ 
 55 lasers/ 
 56 exp physiotherapy/ or exp physical medicine/ 
 57 or/46‐56 
 58 exp ARTHRITIS/rh 
 59 exp myofascial pain syndromes/rh 
 60 fibromyalgia/rh 
 61 spondylitis/rh 
 62 exp spinal osteophytosis/rh 
 63 spondylolisthesis/rh 
 64 exp headache/rh and cervic:.tw. 
 65 whiplash injuries/rh 
 66 cervical rib syndrome/rh 
 67 torticollis/rh 
 68 exp radiculitis/rh 
 69 POLYRADICULITIS/rh 
 70 polyradiculoneuritis/rh 
 71 thoracic outlet syndrome/rh 
 72 or/58‐71 
 73 11 and 72 
 74 11 and 28 and 57 
 75 73 or 74 
 76 75 and 37 
 77 (2004: or 2005: or 2006:).em. 
 78 76 and 77 
 79 from 78 keep 1‐47 
 80 75 and 45 
 81 77 and 80 
 82 from 81 keep 1‐10 
 83 from 82 keep 10

MEDLINE

1 neck/ 
 2 neck muscles/ 
 3 exp cervical plexus/ 
 4 exp cervical vertebrae/ 
 5 atlanto‐axial joint/ 
 6 atlanto‐occipital joint/ 
 7 axis/ 
 8 atlas/ 
 9 spinal nerve roots/ 
 10 exp brachial plexus/ 
 11 (odontoid: or cervical or occip: or atlant:).tw. 
 12 or/1‐11 
 13 exp arthritis/ 
 14 exp myofascial pain syndromes/ 
 15 fibromyalgia/ 
 16 spondylitis/ 
 17 exp spinal osteophytosis/ 
 18 spondylolisthesis/ 
 19 exp headache/ and cervic:.tw. 
 20 whiplash injuries/ 
 21 cervical rib syndrome/ 
 22 torticollis/ 
 23 cervico‐brachial neuralgia/ 
 24 cervico‐brachial neuralgia.ti,ab. 
 25 exp radiculitis/ 
 26 polyradiculitis/ 
 27 polyradiculoneuritis/ 
 28 thoracic outlet syndrome/ 
 29 (monoradicul: or monoradicl:).tw. 
 30 or/13‐29 
 31 random:.ti,ab,sh. 
 32 randomized controlled trial.pt. 
 33 double‐blind method/ 
 34 single‐blind method/ 
 35 placebos/ 
 36 clinical trial.pt. 
 37 exp clinical trials/ 
 38 controlled clinical trial.pt. 
 39 (clin: adj25 trial:).ti,ab. 
 40 ((singl: or doubl: or trebl: or tripl:) adj25 (blind: or mask:)).ti,ab. 
 41 placebos/ 
 42 or/31‐41 
 43 exp arthritis/rh 
 44 exp myofascial pain syndromes/rh 
 45 fibromyalgia/rh 
 6 spondylitis/rh 
 47 exp spinal osteophytosis/rh 
 48 spondylolisthesis/rh 
 49 exp headache/rh and cervic:.tw. 
 50 whiplash injuries/rh 
 51 cervical rib syndrome/rh 
 52 thoracic outlet syndrome/rh 
 53 torticollis/rh 
 54 cervico‐brachial neuralgia/rh 
 55 exp radiculitis/rh 
 56 polyradiculitis/rh 
 57 polyradiculneuritis/rh 
 58 or/43‐57 
 59 exp combined modality therapy/ 
 60 exercise/ 
 61 exertion/ 
 62 exp exercise therapy/ 
 63 exp electric stimulation therapy/ 
 64 Transcutaneous Electric Nerve Stimulation/ 
 65 exp rehabilitation/ 
 66 Ultrasonic Therapy/ 
 67 exp phototherapy/ 
 68 lasers/ 
 69 exp physical therapy techniques/ 
 70 or/59‐69 
 71 12 and 58 (41) 
 72 12 and 30 and 70 (94) 
 73 71 or 72 
 74 73 and 42 RCTs 
 75 meta‐analysis.sh. 
 76 meta‐analysis.pt. 
 77 (meta‐analy: or metaanaly:).tw. 
 78 ((systematic: or quantitativ:) adj5 (review: or overview:)).tw. 
 79 (cochrane or medline or cinahl or embase or scisearch or psychinfo or psycinfo or psychlit or psyclit or (national and library)).tw. 
 80 ((Handsearch: or search:) and (cochrane or medline or cinahl or embase or scisearch or psychinfo or psycinfo or psychlit or psyclit or (national and library) or (hand: or manual: or electronic: or bibliograph: or database:))).tw. 
 81 ((review or guideline).pt. or consensus.ti. or guideline.ti. or literature.ti. or overview.ti. or review.ti.) and (79 or 80) 
 82 ((synthesis or overview or review or survey) and (systematic or critical or methodologic: or quantitative or qualitative or literature or evidence or evidence‐based)).ti. 
 83 75 or 76 or 77 or 78 or 80 or 81 or 82 
 84 83 not ((case: or report:).ti. or editorial.pt. or comment.pt. or letter.pt.) 85 73 and 84 Reviews

Data and analyses

Comparison 1. CONTINUOUS TRACTION v placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 pain intensity	1		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
1.1 chronic NDR at 4 weeks treatment + 12 weeks follow‐up	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2 pain intensity	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.1 NDR at 4 weeks treatment	2		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
3 activity of daily living	1		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
3.1 chronic NDR at 4 weeks treatment + 12 weeks follow‐up	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
4 ability to work	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.1 chronic NDR at 4 weeks treatment	2		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
5 social dysfunction	1		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
5.1 chronic NDR at 4 weeks treatment + 12 weeks follow‐up	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]

1.1. Analysis.

Comparison 1 CONTINUOUS TRACTION v placebo, Outcome 1 pain intensity.

1.2. Analysis.

Comparison 1 CONTINUOUS TRACTION v placebo, Outcome 2 pain intensity.

1.3. Analysis.

Comparison 1 CONTINUOUS TRACTION v placebo, Outcome 3 activity of daily living.

1.4. Analysis.

Comparison 1 CONTINUOUS TRACTION v placebo, Outcome 4 ability to work.

1.5. Analysis.

Comparison 1 CONTINUOUS TRACTION v placebo, Outcome 5 social dysfunction.

Comparison 2. CONTINUOUS TRACTION + another intervention v that same intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 v Exs, heat, pt education: pain intensity	1		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
1.1 chronic NDR,DC at 6 weeks treatment	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]

2.1. Analysis.

Comparison 2 CONTINUOUS TRACTION + another intervention v that same intervention, Outcome 1 v Exs, heat, pt education: pain intensity.

Comparison 3. CONTINUOUS TRACTION v another intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 v Acupuncture: global perceived effect	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.1 chronic NDR/neck pain at 4 weeks treatment	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]

3.1. Analysis.

Comparison 3 CONTINUOUS TRACTION v another intervention, Outcome 1 v Acupuncture: global perceived effect.

Comparison 4. INTERMITTENT TRACTION + another intervention v that same intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 v Exs, heat, pt education: pain intensity	1		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
1.1 chronic MND,NDR,DC at 6 weeks treatment	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]

4.1. Analysis.

Comparison 4 INTERMITTENT TRACTION + another intervention v that same intervention, Outcome 1 v Exs, heat, pt education: pain intensity.

Comparison 5. INTERMITTENT TRACTION v another intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 v NSAID: summation of pain score	1		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
1.1 Neck pain/NDR spondylosis at 6 weeks treatment	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2 v Manual therapy: global perceived effect	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.1 subacute, chronic NDR at 4 weeks treatment + 18 months follow‐up	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]

5.1. Analysis.

Comparison 5 INTERMITTENT TRACTION v another intervention, Outcome 1 v NSAID: summation of pain score.

5.2. Analysis.

Comparison 5 INTERMITTENT TRACTION v another intervention, Outcome 2 v Manual therapy: global perceived effect.

Comparison 6. ONE type of TRACTION v another type.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 INT v CONT v Manual TRACTION: pain intensity	3		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
1.1 chronic neck pain, NDR, DC at 6 weeks treatment	3		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2 INT traction v INT/EMG: symptom relief	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.1 subacute and chronic NDR at 7 weeks treatment	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]

6.1. Analysis.

Comparison 6 ONE type of TRACTION v another type, Outcome 1 INT v CONT v Manual TRACTION: pain intensity.

6.2. Analysis.

Comparison 6 ONE type of TRACTION v another type, Outcome 2 INT traction v INT/EMG: symptom relief.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Brewerton 1966.

Methods	RCT Number Analyzed/Randomized: 412/466 Intention‐to‐treat Analysis: NR Power Analysis: NR
Participants	Neck disorder with radicular symptoms (duration of disorder not specified, presumed chronic)
Interventions	INDEX TREATMENT: Traction Group (trac): (97 participants) Mechanical, continuous for 20 minutes, supine position angle of pull at discretion of therapist to maximize pain relief (usually flexed position), neck on electric heat pad COMPARISON TREATMENT: Positioning Group (pos): same positioning as with traction group, with heat but no traction applied (109 participants) Collar Group (col): felt of foam plastic (99 participants) Placebo‐heat [Pl(h)]: traction and untuned short‐wave diathermy (62 participants) Placebo‐tablet [Pl(t)]: inert tablets looking like phenylbutazone (44 participants) CO‐INTERVENTION: medication‐aspirin Treatment Schedule: 4 weeks Duration of Follow‐up: 4 weeks (immediate post‐treatment assessment) and 6 months (letter)
Outcomes	Pain: (4 point scale; gone, getting better, unchanged, getting worse) Baseline: NR Reported Results: not significant for all groups RR [trac v Pl(h)]: 1.00(95%CI:0.85 to 1.18) [power 26%] RR [trac v Pl(t)]: 0.90(95%CI:0.77 to 1.04) [power 30%] RR [trac v col]: 0.99(95%CI:0.86 to 1.15) [power 26%] RR [trac v pos]: 0.97(95%CI:0.85 to 1.11) [power 29%] Ability to work: (4‐point scale; cured, mild, moderate, severe) Baseline: NR Reported Results: not significant for all groups RR [trac v Pl(h)]: 0.87(95%CI:0.53 to 1.44) RR [trac v Pl(t)]: 0.69(95%CI:0.42 to 1.14) RR [trac v col]: 0.76(95%CI:0.50 to 1.16) RR [trac v pos]: 0.86(95%CI:0.56 to 1.32) Side Effects: NR Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding? All outcomes ‐ patients?	High risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	Low risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	Low risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	High risk	not reported
Similarity of baseline characteristics?	Low risk
co‐interventions avoided or similar?	Low risk
Compliance acceptable?	High risk
Timing outcome assessments similar?	Low risk

Brewerton:trac v col.

Methods	refer to Brewerton 1966
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Brewerton:trac v pos.

Methods	refer to Brewerton 1966
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Brewerton:tracvPl(h).

Methods	refer to Brewerton 1966
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Brewerton:tracvPl(t).

Methods	refer to Brewerton 1966
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Guanygue 2001.

Methods	RCT Number Analyzed/Randomized: 536/536 Intention‐to‐treat Analysis: NA Power Analysis: NR
Participants	Neck disorder with or without radicular symptoms (duration of disorder not specified)
Interventions	INDEX TREATMENT: (179 participants) Traction: static cervical traction (in sitting, jaw support by bandage), applied force 1.5 kg to 2.5 kg, 0.5 to 1.0 hour/session, 1 to 3 times/day for 10 days (1 course) COMPARISON TREATMENT: (357 participants) Acupuncture (Acup): acupuncture, medication‐sage root, dalbergia wood and Vitamin B12 injected into each selected point, once every other day for 7 sessions (1 course) CO‐INTERVENTION: NR Treatment Schedule: each course of treatment repeated twice with 3 to 7 day interval between Duration of Follow‐up: none
Outcomes	Global perceived effect: (3‐point scale: cured, improved, ineffective) Baseline: NA Reported Results: significant favoring acupuncture RR 4.31(95% CI: 2.93 to 6.34) Side Effects: NR Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding? All outcomes ‐ patients?	High risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	High risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	Low risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	High risk	not applicable
Similarity of baseline characteristics?	Low risk
co‐interventions avoided or similar?	Low risk
Compliance acceptable?	High risk
Timing outcome assessments similar?	Low risk

Klaber Moffett 1990.

Methods	RCT Number Analyzed/ Randomized: 94/100 Intention‐to‐treat Analysis: NR Power Analysis: NR
Participants	Chronic neck disorder with radiculopathy
Interventions	INDEX TREATMENT: Group A (41 participants) static cervical traction (technique = halter that gives even pull on the chin and the occiput, patient position = neck in slight flexion on two pillows approximately 25 degree angle of pull, weight = 6 to 15 pounds based on body weight), education (group: neck school), collar, drugs COMPARISON TREATMENT: Group B (43 participants) placebo, education (group: neck school), collar, drugs COINTERVENTION: avoided in the trial Treatment Schedule: 12 session over 4 weeks Duration Follow‐up: 12 weeks
Outcomes	Pain intensity (VAS 10‐cm) Baseline: traction 5.10, placebo 4.60 End of Study Mean: traction 2.78, placebo 3.19 Absolute Benefit: traction 2.32, placebo 1.41 Reported Results: not significant SMD ‐0.16 (95%CI: ‐0.59 to 0.27) [power 9%] Activity of daily living (VAS 0 to 10) Baseline: traction 5.86, placebo 5.74 End of Study Mean: traction 2.80, placebo 3.93 Absolute Benefit: traction 3.06, placebo 1.81 Reported Results: not significant SMD ‐0.39 (95% CI: ‐0.84 to 0.06) [power 9%] SOCIAL DYSFUNCTION (VAS 0 to 10) Baseline: traction 3.65, placebo 3.77 End of Study Mean: traction 2.10, placebo 1.86 Absolute Benefit: traction 1.55, placebo 1.91 Reported Results: not significant SMD 0.16 (95%CI: ‐0.27 to 0.60) [power 6%] Side Effects: 2 patients from traction (Group A) reported headaches Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding? All outcomes ‐ patients?	Low risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	Low risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	Low risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	High risk	not reported
Similarity of baseline characteristics?	High risk
co‐interventions avoided or similar?	High risk
Compliance acceptable?	Low risk
Timing outcome assessments similar?	Low risk

Kogstad 1978.

Methods	RCT Number Analyzed/Randomized: 50/50 Intention‐to‐treat Analysis: NA Power Analysis: NR
Participants	Subacute and chronic neck disorder with or without headache or radicular symptoms
Interventions	INDEX TREATMENT: (21 participants) Conventional: 15 kg intermittent, mechanical traction for 15 minutes (8 seconds on/off), heat, massage, isometric and home exercise COMPARISON TREATMENT 1: (13 participants) Manual: manipulation, heat, massage COMPARISON TREATMENT 2: (16 participants) Placebo: placebo tablet, ergonomic advice, patient education about disorder CO‐INTERVENTION: NR Treatment Schedule: 2 times per week for 4 weeks Duration of Follow‐up: 18 months
Outcomes	Global perceived effect (2‐point scale: improved, no change) Baseline: NR Reported Results:no significant difference at 18 months, Conventional 80% improved, Placebo 53% improved, Manual 85% improved RR [CT v Pl]: 0.43(95%CI: 0.15 to 1.200) [power 8%] RR [CT v MT]: 0.33(95%CI:0.08 to 1.32) [power 6%] Side Effects: NR Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk
Allocation concealment?	High risk	C ‐ Inadequate
Blinding? All outcomes ‐ patients?	High risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	High risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	High risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	High risk	not applicable
Similarity of baseline characteristics?	High risk
co‐interventions avoided or similar?	High risk
Compliance acceptable?	High risk
Timing outcome assessments similar?	High risk

Kogstad: CT v MT.

Methods	refer to Kogstad 1978
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Kogstad: CT v Pl.

Methods	refer to Kogstad 1978
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Shakoor 2002.

Methods	RCT Number Analyzed/Randomized: 199/218 Intention‐to‐treat Analysis: NR Power Analysis: NR
Participants	Neck disorder, spondylosis
Interventions	INDEX TREATMENT: Group A (100 participants) CT (continuous traction) for 30 minutes in sitting with 15 degrees of flexion angle pull and 15% of person's body weight COMPARISON TREATMENT: Group B (99 participants) CT without weight in sitting, NSAID (Naproxen 250 mg twice a day) and Rantidine CO‐INTERVENTION: isometric neck exercises, postural advice and collar for posture used during activity, vitamin B Treatment Schedule: every other day for 6 weeks Duration of Follow‐up: weekly but outcomes only given following 6 week treatment
Outcomes	Pain: summation of pain score, tenderness index, pain frequency score and VAS (no length provided) Baseline Mean: Group A 13.73, Group B 13.32 End of Study Mean: Group A 6.60, Group B 7.52 Absolute Benefit: Group A 7.13, Group B 5.8 Reported Results: nearly significant SMD: ‐0.26(95%CI: ‐0.54 to 0.01) [power 94%] Side Effects: NR Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding? All outcomes ‐ patients?	High risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	High risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	Low risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	High risk	not reported
Similarity of baseline characteristics?	Low risk
co‐interventions avoided or similar?	Low risk
Compliance acceptable?	Low risk
Timing outcome assessments similar?	Low risk

Wong 1997.

Methods	RCT Number Analyzed/Randomized: 24/24 Intention‐to‐treat Analysis: NA Power Analysis: NR
Participants	Subacute and chronic neck disorder with radicular symptoms
Interventions	INDEX TREATMENT: Group A (12 participants) An adaptive intermittent cervical traction modality with EMG biofeedback and moist heat: Digit‐Trac E 90 KA traction unit and head halter: 8% body weight starting force and increased by 0.25 kilograms per session, 10 seconds on/5 seconds off for 20 minutes, moist heat for 20 minutes pre‐treatment, 10 Hz surface EMG (to collect signals ranging from 10 Hz to 1 KHz in frequency) applied to C5 paravertebral muscles COMPARISON TREATMENT: Group B (12 participants) Conventional intermittent traction with moist heat: same set up and parameters as Group A without the EMG biofeedback CO‐INTERVENTION: none Treatment Schedule: every other day for 7 weeks Duration of Follow‐up: none
Outcomes	Symptom relief: (4‐point scale; excellent, good, fair, poor) Baseline: NR Reported Results: no significant difference RR 1.12 (95% CI: 0.67 to 1.89) [power 6%] Side Effects: NR Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding? All outcomes ‐ patients?	High risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	High risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	Low risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	Low risk	not applicable
Similarity of baseline characteristics?	Low risk
co‐interventions avoided or similar?	High risk
Compliance acceptable?	High risk
Timing outcome assessments similar?	Low risk

Zylbergold 1985.

Methods	RCT Number Analyzed/Randomized: 100/100 Intention‐to‐treat Analysis: NA Power Analysis: NR
Participants	Neck disorder with radicular symptoms and/or degenerative changes, strains (duration of disorder not specified but presumed chronic)
Interventions	INDEX TREATMENT 1: (25 participants) Static traction (G1): 25 pounds for 15 minutes in supine with neck angles at 25 degrees flexion, instruction in neck care, moist heat for 15 minutes, range of motion and isometric exercises INDEX TREATMENT 2: (25 participants) Intermittent traction (G2): 25 pounds for 15 minutes (10 seconds on/off) in supine with neck angles at 25 degrees flexion, instruction in neck care, moist heat for 15 minutes, range of motion and isometric exercises COMPARISON TREATMENT: (25 participants) Manual traction (G3): 15 minutes, minimum 20 pulls (timing of each pull not specified), instruction in neck care, moist heat for 15 minutes, range of motion and isometric exercises COMPARISON TREATMENT G4: (25 participants) instruction in neck care, moist heat for 15 minutes, range of motion and isometric exercises CO‐INTERVENTION: medication and collar use Treatment Schedule: 2 times a week for 6 weeks Duration of Follow‐up: none
Outcomes	Pain intensity: (McGill PQ) Baseline Mean: G1 2.02, G2 1.80, G3 1.60, G4 1.86 End of Study Mean: G1 0.74, G2 0.30, G3 0.58, G4 0.98 Absolute Benefit: G1 1.28, G2 1.50, G3 1.02, G4 0.88 Reported Results: favors intermittent traction SMD [1 v 4]: ‐0.22(95%CI: ‐0.78 to 0.34) [power 63%] SMD [2 v 4]: ‐0.78(95%CI: ‐1.36 to 0.21) SMD [1 v 3]: 0.17(95%CI: ‐0.39 to 0.72) [power 58%] SMD [2 v 3]: ‐0.39(95%CI: ‐0.95 to 0.17) [power 61%] SMD [2 v 1]: ‐0.49(95%CI: ‐1.06 to 0.07) Side Effects: NR Cost of Care: NR
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding? All outcomes ‐ patients?	High risk
Blinding? All outcomes ‐ providers?	High risk
Blinding? All outcomes ‐ outcome assessors?	Low risk
Incomplete outcome data addressed? All outcomes ‐ drop‐outs?	Low risk
Incomplete outcome data addressed? All outcomes ‐ ITT analysis?	High risk	not applicable
Similarity of baseline characteristics?	Low risk
co‐interventions avoided or similar?	Low risk
Compliance acceptable?	High risk
Timing outcome assessments similar?	Low risk

Zylbergold: 1 v 3.

Methods	refer to Zylbergold 1985
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Zylbergold: 1 v 4.

Methods	refer to Zylbergold 1985
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Zylbergold: 2 v 1.

Methods	refer to Zylbergold 1985
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Zylbergold: 2 v 3.

Methods	refer to Zylbergold 1985
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Zylbergold: 2 v 4.

Methods	refer to Zylbergold 1985
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

NR ‐ not reported 
 NA ‐ not available 
 RR ‐ relative risk 
 95% CI ‐ 95% confidence interval 
 VAS ‐ visual analogue scale 
 SMD ‐ standardized mean difference 
 McGill PQ ‐ McGill Pain Questionnaire

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Akinbo 2006	only outcome is cardiovascular which is not applicable to this review
Fialka 1989	mechanical traction part of combined therapy tailored to the indvidual patient in which specific traction doses not given
Goldie 1970	quasi‐RCT, randomization by date of birth
Jensen 1995	standardized co‐intervention for both index and comparison groups
Joghataei 2004	only outcome is grip strength which is not applicable to this review
Loy 1983	quasi‐RCT, allocated alternately to treatment
Pennie 1990	quasi‐RCT, randomization by casualty number
Skargren 1998	less than 15% of the study participants received traction

Contributions of authors

This is one review of a series conducted by the Cervical Overview Group: Aker P, Bronfort G, Burnie SJ, Cameron ID, Eddy A, Goldsmith CH, Graham N, Gross AR, Haines T, Haraldsson B, Kay T, Kroeling P, Peloso PM, Radylovick Z, Santaguida P, Trinh KV, Wang E

Primary Reviewers for this review ‐ Graham N, Gross AR, Goldsmith CH, Klaber Moffett J, Haines TA, Burnie SJ, Peloso PM

Statistician ‐ Goldsmith CH 
 Methodological Quality Assessment ‐ Goldsmith CH, Aker P, Trinh KV, Haines T, Bronfort G, Peloso P 
 Study Identification and Selection ‐ Graham N, Gross A, Trinh KV, Haraldsson B, Haines T, Hoving J 
 Research Librarian ‐ Eddy A 
 Data Abstraction, Synthesis, Manuscript Preparation, Public Responsibility, Grants, Administration ‐ primary reviewers 
 Grant Writing ‐ Gross A, Kay T 
 Final Synthesis ‐ full Cervical Overview Group

Sources of support

Internal sources

• McMaster University, Hamilton, ON, Canada.

• Meadowlands Physiotherapy Ancaster, ON, Canada.

External sources

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

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

• Hamilton Hospital Association, Hamilton, ON, Canada.

• Hamilton Health Sciences Corporation, Chedoke‐McMaster Foundation, Hamillton, ON, Canada.

Declarations of interest

One of the review authors was also author of an included RCT (Klaber Moffett 1990). She was not involved in decisions regarding the inclusion of her trial, assessment of bias, or data abstraction.

Edited (no change to conclusions)