Manual Therapy 16 (2011) 53e65

Contents lists available at ScienceDirect

Manual Therapy
journal homepage: www.elsevier.com/math

Systematic review

A systematic literature review on the effectiveness of non-invasive therapy

for cervicobrachial painq
Emma Salt a, *, Chris Wright b, Sue Kelly b, Allison Dean a
a
Physiotherapy Department, Queen’s Hospital Foundation Trust, Burton on Trent DE13 0RB, UK
b
School of Health and Population Sciences, College of Medical and Dental Sciences, The University of Birmingham, Birmingham, UK

a r t i c l e i n f o a b s t r a c t

Article history: Cervicobrachial pain is a common cervical spine disorder. It is frequently managed through non-invasive
Received 2 September 2009 therapy. The objective of this systematic review was to assess effectiveness of non-invasive therapy for
Received in revised form the management of cervicobrachial pain, in terms of pain, function and disability.
3 September 2010
Computerised searches were performed to January 2010. Studies were selected using pre-specified
Accepted 24 September 2010
criteria. Methodological quality of included studies was assessed using PEDro and level of inter-reviewer
agreement reported using Kappa values. Meta-analyses were conducted on pain scores for similar inter-
Keywords:
ventions using DerSimonianeLaird random-effects model to allow for heterogeneity. Effect sizes and 95%
Cervicobrachial
Cervical radiculopathy
confidence intervals were reported. Qualitative analyses, based on Centre for Evidence Based Medicine
Systematic review levels of evidence, were conducted for function and disability.
Physiotherapy Eleven studies were included in the review. Interventions included general physiotherapy, cervical
traction, manual therapy, exercise therapy, and behavioural change approaches.
There was inconclusive evidence for the effectiveness of non-invasive management of cervicobrachial
pain. Potential benefits were indicated in the provision of manual therapy and exercise and behavioural
change approaches to reduce pain. General physiotherapy and traction were no more effective than
comparators in reducing pain (level A evidence). Effects of non-invasive management on function and
disability were mixed.
Future studies should identify which sub-groups of cervicobrachial pain respond to specific
interventions.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction and Fehlings, 2005). Numerous classifications have been reported

1.1. Rationale
Cervicobrachial pain is defined, for the purpose of this review, as
the presence of upper-quadrant pain associated with cervical spine
pain (Jull et al., 2008). Pain and symptoms can be referred to the
upper limb from somatic structures or radiate to the upper limb
through neuropathic mechanisms. Somatic structures include neck
muscles (Leffler et al., 2003), zygapophyseal joints (Fukui et al.,
1996) and intervertebral discs (Slipman et al., 2005). Radiating
arm pain is commonly derived from the nerve root or trunk (Carette
q Department to which work should be attributed: School of Health and Pop-
ulation Sciences, College of Medical and Dental Sciences, The University of
Birmingham.
* Corresponding author. Tel.: þ44 1283 566 333x5255.
E-mail address: EJS495@bham.ac.uk (E. Salt).
relating to cervicobrachial pain, including cervicobrachial pain
syndrome, cervical radiculopathy and neck and arm pain. Terms are
often used synonymously and interchangeably. Consequently, this
review has considered cervicobrachial pain as an overarching
classification.
Cervicobrachial pain has been estimated to be more prevalent
than neck pain in isolation (Daffner et al., 2003), suggesting that
cervicobrachial pain constitutes the majority of patients presenting
with cervical spine disorders. Radhakrishnan et al. (1994) found
recurrences in 32% of patients with cervicobrachial pain over
a period of 4.9 years. When the condition is chronic, it is likely to
become a persistent or recurring problem that impacts unfavour-
ably on an individual’s mental as well as physical health (Daffner
et al., 2003). In addition to the impact on individuals, persistent
disablement can lead to high costs for health care systems and
society (Karjalainen et al., 2006).
Management of cervicobrachial pain can be surgical or conser-
vative. Surgery does not provide long-term gains compared to

1356-689X/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2010.09.005
54 E. Salt et al. / Manual Therapy 16 (2011) 53e65
conservative management and is known to carry a 4% complication
rate (Fouyas et al., 2002; Carragee et al., 2008). Conservative
management is the initial treatment of choice for the majority of
patients with cervicobrachial pain (Fouyas et al., 2002; Daffner et al.,
2003). Exceptions to this are patients with serious local pathology
such as fractures, dislocations, myelopathy, infections or tumours
(Carette and Fehlings, 2005).
Conservative management can consist of invasive techniques
(injection therapy and acupuncture) or non-invasive therapy that
encompasses many forms of physiotherapy, osteopathy and chiro-
practics. There is some evidence to support the use of injection
therapy (Peloso et al., 2004) and acupuncture (Trinh et al., 2006) in
the management of cervicobrachial pain. The Task Force on Neck
Pain and Associated Disorders published a document in 2008
looking specifically at non-invasive interventions for neck pain, up
to 2006. This highlighted the under-representation of cervicobra-
chial pain in the non-invasive literature and recommended that
‘future efforts should focus on non-invasive interventions’ for this
patient group (Hurwitz et al., 2008, p.123). This systematic review
incorporates literature up to 2010. It was anticipated that the
evaluation of recent research would identify current evidence for
the non-invasive management of cervicobrachial pain. This was the
primary rationale for this systematic review.
1.2. Objective
Cervicobrachial pain has been found to impact substantially on
function and lead to high levels of disability (Daffner et al., 2003). This
systematic review evaluates the effectiveness of one non-invasive
approach against another or when possible, against placebo or
a control. Between-group differences in outcomes relating to pain,
function and disability establish which non-invasive approaches are
superior.
2. Methods
Methods used to conduct the systematic review were based on
guidance for undertaking reviews from the Centre for Reviews and
Dissemination (CRD, 2009). The review is reported in line with the
Preferred Reporting Items for Systematic Reviews and Meta-anal-
yses: PRISMA (Moher et al., 2009), to ensure a standardised and
comprehensive framework for reporting.
2.1. Eligibility criteria
Studies: Randomised controlled trials. No restrictions were
placed on language or publication date.
Participant inclusion criteria: Adults aged 16 years or over, of
either gender, with cervicobrachial pain. Cervicobrachial pain was
defined as the presence of upper-quadrant pain associated with
cervical spine pain (Jull et al., 2008).
Studies involving participants with cervicobrachial pain as part
of a larger generic neck pain trial were included if sub-group
analysis had been performed.
Participant exclusion criteria: Participants with cervical myelop-
athy, tumour, rheumatic disease or central neurological disorder
(e.g. multiple sclerosis) were excluded.
Interventions: Non-invasive therapies such as manual therapy,
exercise, traction, behavioural therapy and electrotherapy. Invasive
management such as surgery, spinal steroid injections and
acupuncture were excluded unless they were used as a comparator.
Other comparators included placebo treatments or no-treatment
control groups. When this was not possible, one non-invasive
approach was compared to another.
Outcome measures: Studies were included when outcome
measures were reported on at least one of the following: pain,
function or disability. Different publications from the same study
were included if different, but relevant, outcomes were reported.
Studies utilising outcomes based purely on physical measures
without also measuring pain were not included. Secondary
outcomes included reports of adverse events.
2.2. Information sources
The search strategy encompassed a wide range of information
sources (summarised below) including published and unpublished
data to reduce the potential for publication bias (CRD, 2009).
 Electronic search on computerised databases from inception to
January 2010: Cochrane central register of controlled trials
(CENTRAL), MEDLINE, EMBASE, AMED, CINAHL and the Phys-
iotherapy Evidence Database (PEDro).
 Citation search: from papers identifying cervicobrachial pain.
 Special interest groups: the Manipulative Association of Char-
tered Physiotherapists, the International Federation of Ortho-
paedic Manipulative Therapists, the British Osteopathic
Association, and the United Chiropractic Association.
2.3. Search
Keywords were specified under classification, treatment,
therapy and methodological terms. Thesaurus mapping and MESH
terms were used. All terms were exploded, where possible, in each
database. Details of the Medline search can be found in Appendix I.
2.4. Study selection and appraisal
Information sources were searched independently by two
reviewers (ES/SK) against the eligibility criteria previously outlined.
Two subject reviewers (ES/AD) independently assessed methodo-
logical quality for each included study using the PEDro scale (Maher
et al., 2003). Disagreement was resolved by a third reviewer (CW).
The PEDro scale quantifies internal validity of trials performed in
therapy. Recent evidence of the scale’s construct validity strengthens
the rationale for its use (Macedo et al., 2010). Scores lie between
0 and 10, with 10 indicating the highest methodological quality.
Attempts to contact authors were made when fulfilment of criteria
were unclear. Strength of agreement between reviewers regarding
study selection and methodological quality was evaluated using
Cohen’s Kappa (Cohen, 1960). Interpretation of Kappa values were
established using standards proposed by Landis and Koch (1977).
Data extraction tables were compiled by two independent reviewers
(ES and SK) detailing design, interventions, participant characteris-
tics and results for included studies. The Critical Appraisal Skills
Programme (CASP) tool for randomised controlled trials was used to
appraise studies (ES).
2.5. Planned method of analyses
The primary analysis evaluated the effects of non-invasive
therapy on post-intervention pain. Secondary analyses evaluated
effects on function/disability and long-term pain response.
Meta-analyses were performed on comparable outcomes across
studies that had interventions and comparators of a similar nature,
and same or similar timing of assessments. Comparable outcomes
were defined as different instruments that had been developed to
measure the same underlying construct. Two reviewers (ES/CW)
 E. Salt et al. / Manual Therapy 16 (2011) 53e65
identified combinations of studies and outcomes that were appro-
priate for quantitative synthesis.
It was recognised that direct pooling of results from different
studies would not be possible. A random-effects meta-analysis was
needed. Standard errors of study-specific estimates were adjusted to
accommodate for the amount of variation to establish intervention
effects. Analyses were conducted using standardised differences in
means to allow synthesis of data from comparable outcomes (CRD,
2009). Effect sizes and 95% confidence intervals were computed
using DerSimonianeLaird random-effects to allow for inter-study
differences (DerSimonian and Laird, 1986; CRD, 2009). Cochrane’s Q
statistic was computed to assess heterogeneity in the effect of
interventions in each analysis (CRD, 2009). Results from meta-
analysis were presented in a Forest plot. All analyses were performed
using StatsDirect software (StatsDirect, 2009).
Where meta-analyses were not possible, qualitative analysis was
conducted (Table 1) using levels of evidence as per the Centre for
Evidence Based Medicine (Howick, 2009). Grades of recommenda-
tion were used in the discussion by evaluating the levels of evidence
for each study to categorise the strength of supporting evidence.
3. Results
3.1. Study selection
Fig. 1 identifies the study selection process. Following review of
titles and abstracts, 20 articles were identified and full papers
obtained. Nine were excluded, eight had no symptoms radiating
into the arm and one did not measure pain. No relevant unpub-
lished studies were found.
Substantial to perfect inter-reviewer agreement (Landis and
Koch, 1977) was achieved for the selection of studies to be
included in the review, with Kappa values of 0.87 (SE ¼ 0.088;
p < 0.0005; n ¼ 37; SE ¼ standard error) using titles, 0.71 (SE ¼ 0.19;
p < 0.0005; n ¼ 25) using abstracts and 1.00 (SE ¼ 0.00; p < 0.0005;
n ¼ 20) using full texts.
3.2. Risk of bias within studies
Agreement in ratings across the ten PEDro items varied for
studies from moderate (Kappa ¼ 0.55, SE ¼ 0.25; Kuijper et al.,
Table 1
Levels of evidence for therapy (Centre for Evidence Based Medicine e Levels of
evidence) (Howick, 2009).
Level of evidence Description
1A: 1a SR (with homogeneity) of RCTs
1B: Individual RCT (with narrow
Confidence Interval)
2A: SR (with homogeneity) of
cohort studies
2B: Individual cohort study (including
low quality RCT; e.g.<80% follow up)
3A: SR (with homogeneity) of case-control
studies
3B: Individual case-control study
4: Case series (and poor quality cohort and
case control studies)
5 Expert opinion
Grades of recommendation Description
A Consistent level 1 studies
B Consistent level 2 or 3 studies or
extrapolations from level 1 studies
C Level 4 studies or extrapolations from
level 2 or 3 studies
D Level 5 evidence or troublingly inconsistent or
inconclusive studies of any level
Key: RCT ¼ Randomised Controlled Trial; SR ¼ Systematic Review.
55
2009) to perfect (Kappa ¼ 1.00, SE ¼ 0.00; Klaber Moffett et al.,
1990; Bernaards et al., 2007; Walker et al., 2008). This reflected
agreement on 8e10 PEDro items for each study. Agreement in
ratings across studies varied, being moderate (Kappa ¼ 0.42,
SE ¼ 0.23) for items 6 (blind assessor) and 8 (ITT analysis),
substantial (Kappa ¼ 0.76; SE ¼ 0.22) for item 3 (baseline compa-
rability) and perfect (Kappa ¼ 1.00, SE ¼ 0.00) for the remaining
items. As expected, blinding of clinician (Item 5) was not achieved
on any study.
All studies scored 6 or above on the PEDro scale (Table 2),
indicating good quality (Foley et al., 2006).
3.3. Study characteristics
Descriptive data for the ten trials (11 articles) included in the
review is summarised in Table 3.
3.3.1. Methods
All ten trials were undertaken in a clinical setting and were
randomised intervention studies. Of these, one used a ‘no-inter-
vention’ control in a cross-over trial design (Allison et al., 2002); one
used pain medication across groups to act as control for the
manipulation intervention (Howe et al., 1983) and two used placebo
intervention (Klaber Moffett et al., 1990; Young et al., 2009). The
remaining trials used comparative interventions ranging from no-
specific treatment (Kuijper et al., 2009) to a specific therapeutic
intervention (Coppieters et al., 2003). Refer to Table 3 for details.
Post-intervention assessment (defined as the first assessment
point following completion of intervention) ranged from immedi-
ately following intervention (Howe et al., 1983; Klaber Moffett et al.,
1990; Coppieters et al., 2003) to fifteen weeks post intervention
(Persson et al., 1997; Persson and Lilja, 2001). Follow-up assess-
ments ranged considerably: one study had no follow-up assess-
ment (Coppieters et al., 2003), whilst others evaluated effects up to
one year (Persson et al., 1997; Persson and Lilja, 2001; Bernaards
et al., 2007; Walker et al., 2008).
Quantity of interventions ranged from a single session of
treatment (Howe et al., 1983; Coppieters et al., 2003) to 24 sessions
(Persson et al., 1997; Persson and Lilja, 2001). Intensity and dura-
tion of intervention was also variable, from one treatment in
isolation (Coppieters et al., 2003) to 3 treatment sessions a week
(Klaber Moffett et al., 1990; Ragonese, 2009) and extending over
a period of 6 months (Bernaards et al., 2007).
3.3.2. Participants
The ten studies included in the review randomised 1036
participants. From available data, there was a broad age range from
16 to 70 years, with marginally more females than males. There was
a variable duration of symptoms, but insufficient data to detail the
range. Diagnostic methods differed across studies. The majority of
studies stated selection criteria on symptom type and pattern of
radiation.
3.3.3. Interventions
Interventions fell into five distinct groups: general physio-
therapy, traction, manual therapy, exercise therapy and behavioural
therapy.
3.3.4. Outcomes
All studies reported pain as an outcome, with the visual
analogue scale (VAS) being used in five studies, and the numeric
pain rating scale (NPRS) in four studies. Variability in time frames
existed: from evaluating ‘current pain’ (Persson et al., 1997; Persson
and Lilja, 2001) to ‘worst pain in the last four weeks’ (Bernaards
et al., 2007). Different areas for pain were requested on the scales
56 E. Salt et al. / Manual Therapy 16 (2011) 53e65

Articles identified through Articles identified

database searching through other sources

Identification
(n= 82) (n= 14)
Medline (n=20) Citation(n=9)
CINAHL (n=11) MACP (n=5)
EMBASE (n=12)
Cochrane (n=11)
AMED (n=8)
PEDro (n=20)

Articles after duplicates removed (n=37)

Screening

Records excluded
From title (n=12)
From abstract (n=5)

Articles included for full

text screening
(n=20)

Articles excluded from full

Eligibility

text (n =9)

Reasons for exclusion:

No radiating symptoms
below shoulder level (n=8)
No measure of pain (n=1)

Articles included (n= 11)

Studies included (n=10)

Included

Studies included in
quantitative synthesis
(meta analysis) (n=6)

Key: MACP= Manipulative Association of Chartered Physiotherapists

Fig. 1. Study selection flow diagram (Moher et al., 2009).

including: ‘neck, shoulder, arm pain’ (Persson et al., 1997; Persson
and Lilja, 2001), ‘pain in either neck or arm’ (Coppieters et al.,
2003), or separate outcome measures to represent pain in the
neck and arm (Kuijper et al., 2009). Allison et al. (2002) used the
Short-form McGill questionnaire in addition to a VAS. Howe et al.
(1983) used a pain scale comprising four descriptive items.
Functional and disability outcomes were used by eight of the ten
studies. The Neck Disability Index (NDI) was used by four studies.
Otherwise, there was considerable variation in measures, with a focus
on work related aspects (Bernaards et al., 2007), specific problems to
an individual (Young et al., 2009) and psychological and psychosocial
aspects of function and disability (Persson and Lilja, 2001).
3.4. Results of meta-analysis
Three meta-analyses were conducted to assess the effects of
general physiotherapy, traction and manual therapy and exercise
on post-intervention pain (Fig. 2). There were no other comparable
 E. Salt et al. / Manual Therapy 16 (2011) 53e65 57

Table 2 Walker et al. (2008) and Ragonese (2009) found reduced pain at
Quality assessment of included studies using the Physiotherapy Evidence Base three weeks following the commencement of manual therapy and
Database Scale.
exercise. Significant between-group differences were only found in
Study Criterion Total score Ragonese’s study (p < 0.01). Walker et al.’s (2008) follow-ups at 6
1 2 3 4 5 6 7 8 9 10 weeks and one year found a statistically significant reduction in
Allison et al. (2002) 1 0 0 0 0 1 1 1 1 1 6 pain from baseline for the manual therapy and exercise group
Bernaards et al. (2007) 1 0 1 0 0 1 0 1 1 1 6 (p < 0.005) (unlike the minimal intervention comparator; p ¼ 0.08),
Coppieters et al. (2003) 1 1 1 0 0 1 1 1 1 1 8 however between-group differences were not sufficient to reach
Howe et al. (1983) 1 1 1 0 0 1 1 0 1 1 7
statistical significance (p ¼ 0.21).
Klaber Moffett et al. (1990) 1 1 1 1 0 1 1 0 1 1 8
Kuijper et al. (2009) 1 1 1 0 0 1 1 1 1 1 8
Some studies focused on evaluating specific forms of manual
Persson et al. (1997) 1 1 1 0 0 0 1 1 1 1 7 therapy. Allison et al. (2002) found significant improvements in pain
Persson and Lilja (2001) 1 1 1 0 0 0 1 1 1 1 7 at 8 weeks, for participants receiving neural-biased manual therapy
Ragonese (2009) 1 1 0 0 0 1 1 0 1 1 6 (cervical lateral glide and shoulder girdle oscillation) compared to
Walker et al. (2008) 1 1 1 0 0 1 1 1 1 1 8
those receiving articular-biased treatments (glenohumeral and
Young et al. (2009) 1 1 1 1 0 1 1 1 1 1 9
thoracic mobilisation) (p ¼ 0.03). Coppieters et al. (2003) focused on
Key: Criterion: 1 ¼ random allocation; 2 ¼ concealed allocation; 3 ¼ baseline simi-
the lateral glide specifically. They found, unlike the ultrasound
larity; 4, 5 and 6 ¼ blinding of participants, therapists and assessors; 7 ¼ measures of
key outcomes from more than 85% of participants; 8 ¼ intention to treat analysis;
comparator, the lateral glide intervention resulted in a significantly
9 ¼ between-group statistical comparisons; 10 ¼ point measures and measures of improved pain response to a ‘neural tissue provocation test’ test for
variability. the upper limb immediately post intervention (p < 0.0003). Howe
Scoring: 1 ¼ yes; 0 ¼ no. Total score (0e10), with 10 representing high methodo- et al. (1983) found cervical manipulation to significantly reduce
logical quality.
neck pain immediately post-intervention compared to control
(p < 0.001), however there were variable responses to upper limb
pain (p < 0.02 in the shoulder; p ¼ 1.97 in the arm/hand). No
significant improvements were found at one week follow-up. Two
outcomes, time points or interventions across the studies included
participants in the manipulation group also had facet joint injection
in the review.
(lidocaine and hydrocortisone) to “allow the manipulation to be
Overall, there was no evidence that general physiotherapy or
carried out”. It was not reported whether participants who received
traction reduced pain post-intervention compared to comparators.
injections made any significant improvement compared to those
There was a trend towards a favourable response to manual therapy
who did not. Both injection therapy and cervical manipulation have
and exercise but this did not reach statistical significance (DerSi-
documented risk factors. Adverse events were not reported.
monianeLaird pooled effect size ¼ 0.34; 95%CI ¼ 0.79 to 0.11).
Regarding function and disability, Ragonese (2009) found
a significant improvement at three week follow-up on NDI scores
3.5. Synthesis of results for participants who had received manual therapy and exercise
compared to those who received either manual therapy or exercise.
3.5.1. General physiotherapy No other studies replicated these findings.
There was no evidence to support that general physiotherapy
was more beneficial than comparators (such as use of collar) in 3.5.4. Behavioural therapy
reducing pain post-intervention, or at longer-term follow-ups Bernaards et al. (2007) evaluated whether behavioural changes
(Persson et al., 1997; Persson and Lilja, 2001; Kuijper et al., 2009). towards working style with or without engagement in physical tasks
Regarding function, participants who received physiotherapy would be more effective that ‘usual care’. The level and type of
performed significantly better (p < 0.05) on three of the twelve intervention in ‘usual care’ was not specified. There were no signifi-
aspects of the Disability Rating Index (walking, sitting for a long time cant between-group differences on any measure post-intervention.
and heavy work) at fifteen weeks compared to those wearing a collar At one year follow-up, the group receiving the working style behav-
(Persson and Lilja, 2001). Sitting for a long time remained improved ioural change without the physical task engagement, improved
at one year (p < 0.05). Kuijper et al. (2009) found a non-significant significantly compared to the ‘usual care’ group for pain (p < 0.05).
improvement in NDI for participants receiving physiotherapy However, as acknowledged by the authors, differences were small.
compared to a ‘no-specific treatment’ comparator (p ¼ 0.090).
Regarding disability, participants who received physiotherapy 4. Discussion
had improved scores on the Sickness Impact Profile compared to
collar use at 15 week follow-up (p < 0.05), but not at one year 4.1. Summary of evidence
(p > 0.05) (Persson et al., 1997). General physiotherapy did not
improve psychological or psychosocial well being (p > 0.05) This systematic review found inconclusive evidence for the non-
(Persson and Lilja, 2001). invasive management of cervicobrachial pain. Evidence was assessed
from 10 randomised clinical trials (1036 participants), conducted
3.5.2. Traction across 5 countries, that compared outcomes on pain, disability and
There were no statistically significant differences between function for the non-invasive management of cervicobrachial pain.
groups for participants receiving traction (either intermittent or There were excellent between reviewer agreements on study selec-
sustained) compared to placebo traction in terms of pain, function or tion. Between reviewer agreements on methodological quality
disability post-intervention or at follow-up (Klaber Moffett et al., ranged from moderate to perfect. Overall methodological quality of
1990; Young et al., 2009). Potential for change using sustained involved studies was good.
traction was seen (Klaber Moffett et al., 1990).
4.1.1. General physiotherapy
3.5.3. Manual therapy and exercise Meta-analyses indicated that general physiotherapy was of no
The meta-analysis found a trend that manual therapy and additional benefit in relation to pain, post-intervention. There were
exercise was favourable in reducing pain post-intervention. Both consistent findings (level A evidence) that general physiotherapy
Table 3

58
Study characteristics.

Study Study design Participants Participant Symptom Interventions and Outcome Results
classification duration assessment tools
points
Allison et al. RCT withcrossover n ¼ 30 Cervico-brachial pain Chronic >3/12 A: cervical lateral glide, Pain: VAS pain, SF Pain: Significant
(2002) design Mean age 54 syndrome Mean 36/12 shoulder girdle oscillation, McGill. between-group
Australia 3 groups: F ¼ 20 Cervicobrachial pain with muscle re-education and Function/disability: differences in
A: neural PT with positive brachial plexus home mobilisation. B: NPQ. improvement
exercise; n ¼ 10 tension test. glenohumeral of mean
B: articular PT with and thoracic mobilisation, VAS pain for
exercise; n ¼ 10 and home A compared to
C: no treatment; n ¼ 10 exercise. B at 8/52 (p ¼
Treatment was given 0.03).
over an 8/52 No statistically
period. Quantity of significant
treatment not between-group
specified. differences on SF
Assessment at McGill (p ¼ 0.15).
baseline, 4/52 Function/disability:
and 8/52. No significant
differences between

E. Salt et al. / Manual Therapy 16 (2011) 53e65

groups at any
assessment period
(p > 0.05).
Bernaards et al. RCT n ¼ 466 Neck and upper limb symptoms Chronic IQR A: behavioural changes Pain: NPRS Pain: No significant
(2007) 3 groups: Mean age 44 Pain, tingles or stiffness in the neck, 14-16 with regard to Function/disability: between-group
Netherlands A: work related behavioural F ¼ 207 shoulders, arms, wrists and/or hands; months posture, workplace DWS differences at 6 months
change group; n ¼ 152 no local somatic disease such as tennis adjustment, breaks (numerical scale). (p > 0.05).
B: work and exercise related elbow or carpel tunnel syndrome. and coping with high Significant between-
behavioural change work demands. group differences for
group; n ¼ 156 B: as for group A. In reduced mean
C: usual care; n ¼ 158 addition, behavioural pain score
changes with regard to for A compared
engagement in with C at 12
moderate to heavy months (p < 0.05).
intensity physical Function/disability:
activities. Physical No significant
exercise was not differences in
part of the intervention. changes over time
Groups A and B received across groups or
6 meetings between-groups at
over 6 months. 6 or 12 months
Assessment at baseline, regarding disability
6/12 and one year. at work (p > 0.05).
Coppieters et al. RCT n ¼ 20 Neurogenic cervicobrachial pain An Sub-acute and A: single session of Pain: NPRS Pain: No between-
(2003) 2 groups: Age range active and passive movement chronic Range lateral glide group differences
Australia A: cervical 35e63 dysfunction related to 2/52e6/12. mobilisation were reported.
mobilisation Mean age 48. non-compliance of the to the cervical spine Significant immediate
(lateral glide); F ¼ 16 median nerve and adverse response B: single session of ther improvement in
n ¼ 10 to median nerve palpation; positive apeutic ultrasound mean pain
B: therapeutic NTPT. to most painful area. reduction pre
ultrasound; Assessment at baseline and compared to post
n ¼ 10 immediately treatment for
following treatment in A (p < 0.0003),
response to the but not for B
NTPT. (p ¼ 0.285).
Howe et al. RCT n ¼ 52 Neck and upper limb symptoms Pain Mixed range of A: quick thrust Pain: measures Pain: Significant
(1983) 2 groups: Age range in the neck, arm or hand related to a duration; details manipulation  identified between-group
UK A: cervical manipulation and 16e65. lesion in the cervical spine; no other not provided injection as ‘absent’, ‘same’, differences for
medication; n ¼ 26 F ¼ 31 causes of pain in the shoulder or arm; and treated with NSAID ( ‘better’ or immediate
B: medication; n ¼ 26 reduced cervical intervertebral joint azapropazone) ‘worse’ for improvement in
movement or palpable asymmetry. B: NSAID alone. pain in the neck, pain in neck
Quantity of treatment not shoulder and (p < 0.001)
recorded. arm/hand. and shoulder
Assessment at baseline, (p < 0.02)
immediately for A compared
following treatment, to B.
1/52 and 3/52. Differences were
not sustained
at one and
three week
follow-up
post-treatment.
No significant
differences
between-groups
for pain in the
arm/hand (p > 0.05).
n ¼ 94 Chronic A: up to 12 sessions of Pain: VAS pain Pain: No significant

E. Salt et al. / Manual Therapy 16 (2011) 53e65

Klaber Moffett RCT Neck and arm pain
et al. (1990) 2 groups: Age range Symptoms in >3/12 sustained Function/disability: between-group
UK A: sustained 39e60. the arm clinically indicative of a Mean 33/52. weighted traction VAS social differences for
traction; n ¼ 44 F ¼ 58 radiculopathy or brachialgia (between 6-18 dysfunction mean VAS
B: placebo stemming from the neck. pounds) over four weeks. scale pain at any
traction; n ¼ 50 B: up to 12 sessions of placebo and one individually assessment period
traction (2 pounds) over chosen (p > 0.01).
four weeks activity of daily Function/disability:
Neck school education living. GHQ No significant
(one hour and STAI. between-group
session) given to both groups. differences for
Assessment at baseline, mean VAS social,
immediately activity of daily
after treatment and living,
3/12 following GHQ or
treatment. STAI at any
assessment period
(p > 0.01).
(continued on next page)

59
Table 3 (continued)

60
Study Study design Participants Participant Symptom Interventions and Outcome Results
classification duration assessment tools
points
Kuijper et al. RCT n ¼ 205 Cervical radiculopathy Acute/subacute A: semi-hard collar day Pain: Two VAS pain Pain:
(2009) 3 groups: Mean age 47 Radiation of arm Mean 3/52 use for scales e neck Significant between-
Netherlands A: semi-hard F ¼ 101 pain distal to elbow. up to 6/52. pain and group differences
collar; n ¼ 69 Provocation of arm pain by at B: PT with a focus on arm pain for mean improvement
B: PT; n ¼ 70 least one mobilising Function/disability: in pain for A
C: no specific neck movement. and stabilising the NDI score compared to C for
treatment; Sensory and motor change in cervical spine was given neck pain (p ¼ 0.001)
n ¼ 66 one or more twice weekly for 6/52. and arm pain (p ¼ 0.006)
adjacent dermatome / C: advice to continue during the first 6
myotome. normal weeks post-intervention.
Diminished daily activities. Significant between-group
reflexes on Assessment at baseline, 3/52, differences for
ipsilateral side. 6/52 and 6/12. mean improvement
in B compared to
C for neck pain
(p ¼ 0.002) and arm pain
(p ¼ 0.007)
for the first 6 weeks

E. Salt et al. / Manual Therapy 16 (2011) 53e65

postintervention.
No significant
between-group
differences
for mean improvement
in pain between A and
B for the first 6 weeks
postintervention
(p > 0.05).
No significant
between-group
differences At 6
months, between
any groups
(p > 0.05).
Function/disability:
Significant between-group
differences for mean
improvement
in disability for A
compared to C
(p ¼ 0.02) for the first
6 weeks postintervention.
No significant differences
between B and C
(p ¼ 0.090)
for the first 6 weeks.
No significant between-
group differences for any
other time scale or
between any
other groups (p > 0.05).
Persson et al. RCT n ¼ 81 Cervical radiculopathy Chronic Range B: rigid collar for daytime use Pain: Two VAS pain Pain:
(1997) & 3 groups Age range Clinical and radiologic 5/12e120/12 and soft collar for night for 3 months. scales e ‘present pain’ No significant
Persson and Lilja A: surgery; n ¼ 27 28e64. findings (plane X-rays and (mean ¼ 53/12) C: general PT up to 15 and ‘worst pain between-group
(2001) B: rigid collar; F ¼ 37 magnetic resonance treatment sessions over the differences in mean
Sweden n ¼ 27 tomography) indicating over 3 months. Modalities last week’ pain scores
C: PT; n ¼ 27 nerve root compression used include TENS, Function/disability: between B and C at any
corresponding to the heat, US, cold, massage, SIP, assessment point
distribution of pain. traction, mobilisation, MACL, HADs, CS, (p > 0.05).
exercises, stretches, aerobic and DRI. Function/disability:
exercise, Significant between-group
rest, differences for reduced
relaxation and advice on disability in mean
ergonomics, SIP score for C
posture compared to B
and co-ordinated exercise. (p < 0.05) at
Most 15 weeks, but not at
commonly one year.
used treatments were Significant between-group
traction, difference in improvement
mobilisation in mean function/ reduction
and heat. in disability in some but not
Assessment at baseline, all of the DRI for C
15/52 compared to B (p < 0.05)

E. Salt et al. / Manual Therapy 16 (2011) 53e65

& one year. at 15 weeks.
No between-group
statistical differences
between B and C for
any other scale at any
of the assessment points
(p > 0.05).
Ragonese (2009) RCT n¼30 Cervical radiculopathy Not stated A: cervical lateral glide Pain: NPRS Pain: Significant between-
USA 3 groups: A: manual Age range or Definition: and thoracic Function/disability: group differences for
PT; n ¼ 10 mean not stated. neck and or mobilisations and neural NDI. reduced mean pain for
B: exercise PT; n ¼ 10 F ¼ 19 upper extremity symptoms. dynamic C compared to A and
C: combined Presence of either a positive techniques B (p < 0.01) at 3 week
manual and Spurlings test or for the median nerve. follow-up.
exercise PT; n ¼ 10 cervical distraction test or B: deep neck flexor, lower Function/disability:
ipsilateral cervical and middle Significant between-group
rotation less than 60 trapezius differences for improved
or brachial plexus and serratus anterior function for C compared
tension test. strengthening to A and B (p < 0.05) at
C: combined approaches three week follow-up.
of A and
B Treatment
given three times a week
for 3/52 for
each intervention group.
Assessment at baseline,
1/52, 2/52 and 3/52.
(continued on next page)

61
Table 3 (continued)

62
Study Study design Participants Participant Symptom Interventions and Outcome Results
classification duration assessment tools
points
Walker et al. (2008) RCT n ¼ 58 (a sub-group Neck and upper limb symptoms Mean in days: A: any manual therapy Pain: Pain:
USA 2 groups: A: manual from 98 patients Mechanical A: 1082 technique commonly used VAS pain scale No significant between-
therapy and with general neck pain with neck and B: 521 in clinical practice for upper limb pain group difference in mean
exercise; n ¼ 31 neck pain). unilateral upper (mobilisation, manipulation, Function/disability: pain at any assessment
B: minimal Mean age 48. extremity symptoms muscle energy and stretching NDI point (p ¼ 0.21).
intervention; Gender specific to techniques) and home and GROC. A had a statistically
n ¼ 27 subgroup not known. exercise programme directed significant reduction
at stabilising and mobilising in pain from baseline
the cervical spine by at each follow-up
physical therapists. (p < 0.005). B did
Most commonly used not have significant
technique(s) reduction in pain
not specified. compared to baseline
B: advice from GP on beyond 3/52 (p ¼ 0.08).
posture, continuation of Function/disability:
normal daily activities No results were reported
and neck movement, and specific to sub-group of
medication. Placebo patients with upper

E. Salt et al. / Manual Therapy 16 (2011) 53e65

therapeutic limb symptoms
ultrasound to the with regard to the
neck and range of NDI. A had statistically
motion exercises significant perceived
were provided improvement on GROC
by physical therapists. at all follow-ups
Six sessions of treatment (p < 0.01)
over three weeks
for both groups.
Assessment at baseline,
3/52,
6/52 and one year.
Young et al. (2009) RCT n ¼ 81 Cervical radiculopathy Mixed A: intermittent mechanical Pain: NPRS Pain:
USA 2 groups: Age range Unilateral <3/12 ¼ 42 traction from approx. Function/disabilit: No significant between-
A: intermittent 18e70 upper-extremity >3/12 ¼ 39 20 pounds to 35 pounds. NDI and PSFS. group difference on
traction; n ¼ 45 Mean age 47 pain, parasthesia and B: placebo mechanical mean NRPS at 2/52
B: placebo F ¼ 55 numbness. traction using 5 (p ¼ 0.24) or 4/52
traction; n ¼ 36 3 out of 4 positive tests pounds or less. (p ¼ 0.38) follow-up.
for the following: Both groups received Function/disability:
1. Spurling test postural education, No significant between-
2. distraction test manual therapy and group difference on mean
3. brachial plexus tension test exercises directed to NDI at 2/52 (p ¼ 0.31) or
4. ipsilateral neck rotation the cervical spine 4/52 (p ¼ 0.56) follow-up.
<60 and scapula. No significant between-
Treatment given group difference on mean
on average 7 PSFS at 2/52 (p ¼ 0.91) or
occasions over 4/52 (p ¼ 0.66) follow-up
an average of 4.2 weeks.
Assessment at baseline,
2/52 and 4/52

Key: CS ¼ Coping Strategies Questionnaire; DRI ¼ Disability Rating Index; DWS ¼ Disability at Work Score; F ¼ female; GHQ ¼ General Health Questionnaire; GP ¼ general practitioner; GROC ¼ Global Rating of Change scale;
HADs ¼ Hospital Anxiety and Depression scale; IQR ¼ inter-quartile range; MACL ¼ Mood Adjective Check List; NDI ¼ Neck Disability Index; NPRS ¼ Numeric Pain Rating Scale; NPQ ¼ Northwick Park Questionnaire;
NSAID ¼ non steroidal anti-inflammatory drug; NTPT ¼ neural tissue provocation test; n ¼ number; PSFS ¼ Patient Specific Functional Scale; PT ¼ physiotherapy; RCT ¼ randomised controlled trial; SF McGill ¼ Short-form McGill
questionnaire; SIP ¼ Sickness Impact Profile; STAI ¼ State Trait Anxiety Inventory; TENS ¼ transcutaneous electrical nerve stimulation; US ¼ ultrasound; VAS ¼ Visual Analogue Scale.
 E. Salt et al. / Manual Therapy 16 (2011) 53e65
Fig. 2. Forest plot of pain (intervention v. comparator).
was no more effective in reducing pain than the use of a collar or ‘no
specific treatment’ at long-term follow-up.
There was conflicting evidence that general physiotherapy
reduced disability or improved function (Level D evidence).
It is clear that the composite nature of generalised physio-
therapy (including electrotherapy, traction, manual therapy and
exercise) has not been shown to be effective. Future research
should focus on more homogeneous treatment approaches.
4.1.2. Traction
Meta-analyses indicated traction (intermittent or sustained)
was of no additional benefit in reducing pain post-intervention.
There were consistent findings (level A evidence) that cervical
traction was no more effective in reducing pain than placebo
traction at long-term follow-up.
There was level A evidence that traction did not improve func-
tion or disability either in the short or long-term.
4.1.3. Manual therapy and exercise
Meta-analyses indicated a trend for manual therapy and exer-
cise to be beneficial in reducing pain post-intervention, but this did
not reach statistical significance.
There was conflicting evidence (level D) that manual therapy
and exercise was advantageous in long-term pain reduction.
There was conflicting evidence (level D) that manual therapy and
exercise affected function and disability.
4.1.4. Behavioural therapy
Behavioural therapy improved pain in the long-term compared
to ‘usual care’. This one study shows potentially moderate benefit
(level B), but further studies are needed to confirm this. Without
having clear details on what the ‘usual care’ group received, it is
difficult to comment on how this treatment effect occurred. It is
unclear why this effect became evident at the one year follow-up
and not post-intervention.
There was level B evidence that behavioural therapy did not
improve function or disability.
63
It is difficult to compare findings of this review with others
because this is the first study to evaluate the effects of intervention
specific to cervicobrachial pain. There is agreement with a previous
systematic review (Graham et al., 2008) that there is little evidence
to support effectiveness of traction on neck pain. A previous review
evaluating manual therapy reported insufficient evidence for
manipulation and mobilisation for neck disorders with radicular
findings (Gross et al., 2004). Whilst it is acknowledged that the
evidence is not strong, there is some evidence from this review to
support a positive influence of manual therapy and exercise on
cervicobrachial pain post-intervention.
4.2. Limitations
Although this review searched literature for studies using pub-
lished and unpublished sources, a potential weakness, as for most
reviews, was the risk of incomplete retrieval of relevant literature.
Cervicobrachial pain is a heterogeneous term and, consequently,
there was variation in participants’ criteria across studies. Some
studies aimed to identify cervical radiculopathy specifically. The
usefulness of identifying patho-anatomical causes, such as cervical
radiculopathy (disease pertaining to the nerve root), has been
questioned (Max, 2000). It may be more important to identify the
mechanism of pain. For cervicobrachial pain, differentiating
neuropathic from a somatic cause can be considered. The devel-
opment of tools such as the Leeds Assessment of Neuropathic
Symptoms and Signs (LANSS), which can detect neuropathic from
somatic cause in musculoskeletal pain (Bennett, 2001; Bennett
et al., 2005), may support future studies to identify which sub-
groups of cervicobrachial pain respond to specific interventions.
There was variability in the outcome measures, with a lack of
consistency in assessment time frames and location of pain
assessed. All measures reported in this review were self-report
tools. Whilst self-report tools might be considered the most
directly reliable means of information, potential issues with cred-
ible responses should be acknowledged, including socially desir-
able, acquiescent or extreme responding (Dodorico McDonald,
64 E. Salt et al. / Manual Therapy 16 (2011) 53e65

2008). In addition, participants completing self-report tools could (continued )

be considered to be the ‘outcome assessor’. This raises issues Searches Terms Results
around ‘adequate’ blinding at assessment, especially when evalu- 18 exp musculoskeletal 9563
ating immediate post-intervention effects. manipulations/
Frequency, intensity and duration of interventions varied 19 manipulation, chiropractic/ 4595
considerably. Pooling results in meta-analyses was therefore or manipulation, orthopedic/
or manipulation, osteopathic/
limited. Future research should focus on identifying effective clin-
or manipulation, spinal/
ically appropriate interventions that are cost-effective. 20 mobili*.ti,ab 119,463
In many of the studies reviewed, there was some reduction in 21 exp massage/ 3811
pain and recovery of function irrespective of treatment received. A 22 exp traction/ 5161
clear recovery pattern was depicted by Kuijper et al. (2009) for the 23 exp electrical stimulation therapy/ 30,157
24 exp transcutaneous electrical 4280
‘no-specific treatment group’ in their study. Knowledge on the nerve stimulation/
natural history of cervicobrachial pain is limited. An epidemiological 25 interferential.ti,ab 240
study by Radhakrishnan et al. (1994) over a 13 year period, for 561 26 ultrasonic therapy/ 6752
patients with cervical radiculopathy (determined by radiographic 27 short wave therapy/ 2547
or diathermy/
evaluation), indicated that symptoms largely resolved with time.
28 (laser ADJ therap*).ti,ab 4171
However, as with Kuijper et al. (2009), all patients received some 29 exp hydrotherapy/ 1865
form of treatment; therefore it is uncertain whether the condition 30 (aquatic ADJ physiotherapy*).ti,ab 4
would have resolved spontaneously without intervention. 31 exp cryotherapy/ 17,456
32 (heat adj treatment).ti,ab 6887
33 (heat adj therap*).ti,ab 123
5. Conclusions 34 8 or 9 or 10 or 11 or 12 or 13 or 14 2,863,967
or 15 or 16 or 17 or 18 or 19 or 20 or 21
There is inconclusive evidence for the effectiveness of non- or 22 or 23 or 24 or 25 or 26 or 27 or
invasive management of cervicobrachial pain. Potential benefits 28 or 29 or 30 or 31 or 32 or 33
35 (randomi* adj clinical adj trial).ti,ab 9814
were indicated in the provision of manual therapy and exercise and
36 (randomi* adj trial).ti,ab 24,377
behavioural change approaches to reduce pain. General physio- 37 randomised controlled trials 130,584
therapy and traction were no more effective than comparators in as topic/or random allocation/
reducing pain (level A evidence). Effects of non-invasive manage- 38 35 or 36 or 37 158,262
39 7 and 34 and 38 20
ment on function and disability were mixed.
40 Duplicate filtered: [7 and 34 and 38] 20
Future studies should identify which sub-groups of cervico-
brachial pain respond to specific interventions. There is a need for
further research involving high quality trials, with appropriate
sample sizes and clinically relevant outcomes, to inform practice References
guidelines on the management of cervicobrachial syndrome.
Allison G, Nagy B, Hall T. A randomised clinical trial of manual therapy for cervico-
brachial pain syndrome. Manual Therapy 2002;7(2):95e102.
Acknowledgements Bennett M. The LANSS pain scale: the Leeds assessment of neuropathic signs and
symptoms. Pain 2001;92:147e57.
Bennett M, Smith B, Torrance N, Potter J. The S-LANNS score for identifying pain of
We acknowledge support from Mrs Karen Johnston (Librarian), predominantly neuropathic origin: validation for use in clinical and postal
for advice towards the database search and the Physiotherapy research. The Journal of Pain 2005;6(3):149e58.
Department at Queen’s Hospital, Burton. Bernaards C, Ariëns G, Knol D, Hildebrandt V. The effectiveness of a work style
intervention and a lifestyle physical activity intervention on the recovery from
neck and upper limb symptoms in computer workers. Pain 2007;1(32):142e53.
Carragee E, Hurwitz E, Cheng I, Carroll L, Nordin M, Guzman J, et al. Treatment of
neck pain e injections and surgical interventions: results of the bone and joint
decade 2000e2010. Task force on neck pain and its associated disorders. Spine
Appendix I. Medline search. 2008;33(4S):S153e69.
Carette S, Fehlings M. Clinical practice. Cervical radiculopathy. The New England
Journal of Medicine 2005;353(4):392e9.
Searches Terms Results
Centre for Reviews and Dissemination [CRD]. Systematic reviews: CRD’s guidance
1 (cervico adj brachial).ti,ab 122 for undertaking reviews in healthcare. 3rd ed. CRD University of York, York
2 cervicobrach*.ti,ab 302 Publishing Services Ltd; 2009. p. 3e90.
3 (cervical adj radicul*).ti,ab 653 Cohen J. A coefficient of agreement for nominal scales. Education and Psychological
4 (cervical adj neuralg*).ti,ab 1 Measurement 1960;20:37e46.
5 (neck adj3 arm adj pain) ti.ab 133 Coppieters M, Stappaerts K, Wouters L, Janssens K. The immediate effects of
6 (trapezius adj myalgia).ti,ab 44 a cervical lateral glide treatment technique in patients with cervicobrachial
7 1 or 2 or 3 or 4 or 5 or 6 1210 pain. Journal of Orthopaedic and Sports Physical Therapy 2003;33(7):369e78.
Daffner S, Hilibrand A, Hanscom B, Brislin B, Vaccaro A, Albert T. Impact of neck and
8 management.ti,ab 490,859
arm pain on overall health status. Spine 2003;28(17):2030e5.
9 therapeutics/ 6841
DerSimonian R, Laird N. Meta-analyses in clinical trials. Controlled Clinical Trials
10 physical therapy modalities/ 22,485 1986;7:177e88.
or“physical therapy Dodorico McDonald J. Measuring personality constructs: the advantages and
(speciality)”/ disadvantages of self-reports, informant reports and behavioral assessments.
11 osteopath*.ti.ab 3270 ENQUIRE 2008;1(1):1e19.
12 exp chiropractic/ 2282 Foley N, Sanjit K, Teasell R, Bureau Y, Speechley M. Estimates of quality and reli-
13 exp exercise/or exercise therapy/ 67,341 ability with the physiotherapy evidence-based database scale to assess the
or exercise movement methodology of randomized controlled trials of pharmacological and non-
techniques/ pharmacological interventions. Physical Therapy 2006;86(6):817e24.
14 patient education as topic/ 57,110 Fouyas I, Statham P, Sandercock P, Lynch C. Surgery for cervical radiculomyopathy
15 exp health education/ 112,942 (Cochrane review). The Cochrane Library 2002;(4). Chichester, UK: John Whiley
16 advice.ti.ab 22,904 and Sons Ltd.
Fukui S, Ohseto K, Shiotani M, Ohno K, Karasawa H, Naganuma Y, et al. Referred pain
17 cognitive therapy/or 29,575
distribution of the cervical zygapophyseal joints and cervical dorsal rami. Pain
behaviour therapy/
1996;68:79e83.
 E. Salt et al. / Manual Therapy 16 (2011) 53e65
Graham N, Gross A, Goldsmith CH, Klaber Moffett J, Haines T, Burnie S, et al.
Mechanical traction for neck pain with or without radiculopathy. Art. No.:
CD006408. Cochrane Database of Systematic Reviews 2008;(3). doi:10.1002/
14651858.CD006408.pub2 [last accessed 8th February 2010].
Gross A, Hoving J, Haines T, Goldsmith C, Kay T, Aker P, et al. Manipulation and
mobilisation for mechanical neck disorders (Cochrane review). The Cochrane
Library 2004;(1). Chilchester, UK: John Wiley and Sons, Ltd.
Howe D, Newcombe R, Wade M. Manipulation of the cervical spine e a pilot
study. Journal of the Royal College of General Practitioners 1983;33
(254):574e9.
Howick J. Levels of evidence. Centre for Evidence Based Medicine, http:/www.cebm.
net/?o¼1025; 2009 [last accessed 8th February 2010].
Hurwitz E, Carragee E, van der Velde G, Carroll L, Nordin M, Guzman J, et al.
Treatment of neck pain: noninvasive interventions. Results of the bone and
joint decade 2000e2010. Task force on neck pain and its associated disorders.
Spine 2008;33(45):S123e52.
Jull G, Sterling M, Falla D, Treleaven J, O’Leary S. Differential diagnosis of cervico-
brachial pain. In: Elvey B, editor. Whiplash, headache and neck pain. London:
Churchill Livingstone; 2008. p. 131e41.
Karjalainen K, Malmivaara A, Van Tulder M, Roine R, Jauhiainen M, Hurri H, et al.
Multidisciplinary biopsychosocial rehabilitation for neck and shoulder pain
among working age adults (Cochrane review). The Cochrane Library 2006;(2).
Chichester, UK: John Wriley and Sons, Ltd.
Klaber Moffett JA, Hughes GI, Griffiths P. An investigation of the effects of cervical
traction. Part 1: clinical effectiveness. Clinical Rehabilitation 1990;4:205e11.
Kuijper B, Tans J, Beelen Nollet F, de Visser M. Cervical collar or physiotherapy
versus wait and see policy for recent onset cervical radiculopathy: randomised
trial. British Medical Journal 2009;339:b3883.
Landis J, Koch G. The measurement of observer agreement for categorical data.
Biometrics 1977;33:159e74.
Leffler A, Hansson P, Kosek E. Somatosensory perception in patients suffering from
long-term trapezius myalgia at the site overlying the most painful part of the
muscle and in an area of pain referral. European Journal of Pain 2003;7(3):267e76.
Maher C, Sherrington C, Herbert R, Moseley A, Elkins M. Reliability of the PEDro
scale for rating quality of randomized controlled trials. Physical Therapy
2003;83:713e21.
65
Max M. Is mechanism-based pain treatment attainable? Clinical trial issues. Journal
of Pain 2000;1(3):2e9.
Moher D, Liberati A, Tetzlaff J, Altman D. Preferred reporting items for systematic reviews
and meta-analysis: the PRISMA statement. PLoS Medicine 2009;6(7):e10000.
Macedo L, Elkins M, Maher C, Moseley A, Herbert R, Sherrington C. There was
evidence of convergent and construct validity of physiotherapy evidence
database quality scale for physiotherapy trials. Journal of Clinical Epidemiology
2010;63:920e5.
Peloso P, Gross A, Haines T, Trinh K, Goldsmith G, Aker P. Medicinal and injection
therapies for mechanical neck disorders (Cochrane review). The Cochrane
Library 2004;(2). Chichester, UK: John Wriley and Sons, Ltd.
Persson L, Carlsson C, Carlsson J. Long lasting cervical radicular pain managed with
surgery, physiotherapy, or a cervical collar. A prospective randomised study.
Spine 1997;22(7):751e8.
Persson L, Lilja A. Pain, coping, emotional state and physical function in patients
with chronic radicular neck pain. A comparison between patients treated with
surgery, physiotherapy or neck collar e a blinded, prospective randomized
study. Disability and Rehabilitation 2001;23(8):325e35.
Radhakrishnan K, Litchy W, O’Fallon W, Kurland L. Epidemiology of cervicalradi-
culopathy. A population based study from Rochester, Minnesota, 1976 through
to 1990. Brain 1994;117:325e35.
Ragonese J. A randomised trial comparing manual physical therapy therapeutic
exercises, to a combination of therapies for the treatment of cervical radicul-
opathy. Orthopaedic Practice 2009;21(3):71e5.
Slipman C, Plastaras C, Patel R, Isaac Z, Chow D, Garvan C, et al. Provocative cervical
discography symptom mapping. The Spine Journal 2005;5:381e8.
StatsDirect statistical software. Version 2.7.7. StatsDirect Ltd.; 2009.
Trinh K, Graham N, Gross A, Goldsmith C, Wang E, Cameron I, et al. Acupuncture for
neck disorders (Cochrane review). The Cochrane Library 2006;(3). Chilchester,
UK: John Wiley and Sons Ltd.
Walker M, Boyles R, Young B, Strunce J, Garber M, Whiteman J, et al. The effec-
tiveness of manual therapy and exercise for mechanical neck pain. Spine
2008;33(22):2371e8.
Young I, Michener L, Cleland J, Arnold J, Aguilera A, Snyder A. Manual therapy,
exercise, and traction for patients with cervical radiculopathy: a randomised
clinical trial. Physical Therapy 2009;89(7):632e42.