Manual Therapy 16 (2011) 109e117

Contents lists available at ScienceDirect

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

Systematic review

Manual therapy for osteoarthritis of the hip or knee e A systematic review

H.P. French a, *, A. Brennan b, B. White c, T. Cusack d
a

School of Physiotherapy, Royal College of Surgeons in Ireland, Dublin 2, Ireland

Physiotherapy department, Adelaide and Meath hospital Dublin (incorporating the National Childrens hospital), Dublin 24, Ireland
c
Physiotherapy department, Mater Misericordiae University hospital, Dublin 7, Ireland
d
UCD School of Public Health, Physiotherapy and Performance Science, University College Dublin, Dublin 4, Ireland
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 1 June 2010
Received in revised form
12 October 2010
Accepted 25 October 2010

The aim of this systematic review was to determine if manual therapy improves pain and/or physical
function in people with hip or knee OA. Eight databases were searched for randomised controlled trials
(RCTs). Data were extracted and risk of bias assessed by independent reviewers. Four RCTs were eligible
for inclusion (280 subjects), three of which studied people with knee OA and one studied those with hip
OA. One study compared manual therapy to no treatment, one compared to placebo intervention, whilst
two compared to alternative interventions. Meta-analysis was not possible due to clinical heterogeneity
of the studies. One study had a low risk of bias and three had high risk of bias. All studies reported shortterm effects, and long-term effects were measured in one study. There is silver level evidence that
manual therapy is more effective than exercise for those with hip OA in the short and long-term. Due to
the small number of RCTs and patients, this evidence could be considered to be inconclusive regarding
the benet of manual therapy on pain and function for knee or hip OA.
2010 Elsevier Ltd. All rights reserved.

Keywords:
Osteoarthritis
Manual therapy
Pain
Physical function

1. Introduction
Osteoarthritis is the most common form of arthritis and one of
the leading causes of pain and disability worldwide (NICE, 2008;
Reginster, 2002). It is characterised by a progressive loss of articular cartilage, joint space narrowing, sclerosis of subchondral bone
and osteophyte formation. Soft tissues such as the capsule can
undergo soft tissue contracture and brosis (Cameron and Macnab,
1975; Jacobs, 1960). These changes can result in pain, impaired
mobility, reduced muscle strength, limitation in activities of daily
living (Steultjens et al., 2000, 2001) and reduced quality of life
(Salaf et al., 2005), with the knee and hip joints most commonly
affected. Management aims to control pain and reduce disability
(Hinton et al., 2002; Sarzi-Puttini et al., 2005). Non-pharmacologic
measures such as education, weight loss, physical therapies, and
exercise should be tried rst, with adjunctive pharmacologic
intervention (Brandt, 2000). Manual therapy is a physical treatment
used by physiotherapists, chiropractors, osteopaths and other
practitioners to treat musculoskeletal pain and disability, and
includes massage therapy, joint mobilisation and manipulation.
Recently published clinical guidelines on the management of OA
recommended manual therapy as an adjunctive therapy to exercise

* Corresponding author. Tel.: 353 1 4028591; fax: 353 1 4022471.

E-mail address: hfrench@rcsi.ie (H.P. French).
1356-689X/$ e see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2010.10.011

for OA (NICE, 2008; RACGP, 2009). Of note is that in one of the

guidelines, this recommendation was based on just one study
(Hoeksma et al., 2004) which compared manual therapy head-tohead with exercise rather than as a supplementary intervention.
Some of its proposed mechanisms include mechanical alteration of
tissues, neurophysiological effects and a psychological inuence
(Hoving et al., 2004).
No known systematic reviews have evaluated the effect of
manual therapy in the management of hip or knee OA. The aim of
this review was to determine if manual therapy is benecial for
people with hip or knee OA in reducing pain or improving physical
function.

2. Methods
An extensive computerised literature search of MEDLINE (Ebsco,
PubMed, OVID) (January 1966 e October 2009) and the following
databases from inception date to October 2009: CINAHL (Ebsco),
EMBASE, PEDro, Cochrane Library, ISI Web of Science and PsychINFO was undertaken by all four authors.
Subject headings and keywords based around population,
interventions and study design were used to identify potential
citations (Appendix A). The references of all included publications
and systematic reviews of physical therapy modalities related to
hip and/or knee OA were screened for further relevant articles.

110

H.P. French et al. / Manual Therapy 16 (2011) 109e117

Table 1
Levels of Evidence (Tugwell and OShea, 2004).
Level of evidence Description
Platinum

Gold
Silver

Bronze

At least two RCTs with sample size of at least 50 per group.

If studies did not nd a statistically signicant difference,
they must have been adequately powered for a 20% relative
difference in the relevant outcome. Both patients and
assessors must have been blinded for outcomes and
treatment allocation concealed. Withdrawals >80% should
have had appropriate follow-up including imputations
based on methods such as last observation carried
forward (LOFC).
At least one RCT fullling the same criteria as platinum.
A randomised trial that did not meet the above criteria.
Could also include evidence from at least one study of
nonrandomised cohorts who did and did not receive
the therapy. A randomised trial with a head to head
comparison of agents was considered silver level,
unless a reference was provided to one of the agents
to a placebo showing at least a 20% relative difference.
At least one high-quality case series without controls,
including before/after studies where subjects acted as
their own controls or if the conclusion was based on
expert opinion from clinical experience.

2.1. Inclusion criteria

Studies were included if they 1) were a randomised clinical trial
(RCT), 2) included subjects with clinical or radiographic diagnosis of
hip or knee OA, 3) one study group received manual therapy alone
compared to either no intervention or another intervention, and 4)
if pain and/or physical function outcomes were measured.
Manual therapy was dened for the purposes of the review as
manipulation (high velocity, low amplitude), mobilisation (low
velocity small or large amplitude), massage or other manual therapy
techniques, excluding no-touch techniques (Reid and Rivett, 2005).
2.2. Exclusion criteria
A study was excluded if it was not published as full-text, not in
English language, participants were post-surgery, it was not an RCT
or participants were not humans.
2.3. Data extraction and analysis
Data extraction was carried out independently by three
researchers (HF, AB, TC) using a standardised data extraction form,
which was modied from a Cochrane review group form (Cochrane
Cystic Fibrosis and Genetic Disorder Group, 2009). Means and standard deviations (SDs) for the relevant outcomes were extracted for
short-term (post-treatment) and long-term follow-up (at least six
months) and reported as standardised mean difference (SMD). Hedges effect size (ES) which provides an adjustment for small sample
bias was used (Deeks et al., 2007). Change scores were used where
available. All data were entered in RevMan version 5 (Copenhagen:
The Nordic Cochrane Centre, 2008). Efforts were made to contact the
original authors to obtain the data and when this was not available,
pooled SDs were estimated from the condence interval (CIs) using
methods recommended by the Cochrane Collaboration (Higgins and
Green, 2009). Statistical signicance was set at p < 0.05. Cohens
effect sizes were used to describe the size of the effect, where 0.2e0.49
represented a small effect size, 0.5e0.79 represented a moderate
effect size and >0.8 denoted a large effect size (Cohen, 1977).
2.4. Risk of bias assessment
Two authors (HF, AB or TC) independently assessed the risk of
bias for each article, using criteria recommended by the Cochrane

Back Pain Review Group (van Tulder et al., 2003). A study with a low
risk of bias was dened as a trial fullling 6 or more of the 11
criteria, whilst a study with less than 6 of the criteria was classied
as having a high risk of bias. In the case of disagreements authors
tried to reach consensus and if necessary a third author assessed
the article to resolve disagreements. Just one item in one study
required adjudication from the third author.
A qualitative analysis of the levels of evidence using the grading
system described by Tugwell and OShea (2004) and recommended
by the Cochrane Musculoskeletal Group was performed (Table 1).
3. Results
3.1. Trials
The search yielded 1209 potentially eligible studies. Following
the exclusion process (Fig. 1), eight articles which had manual
therapy as a treatment component remained. Four articles which
examined manual therapy in combination with exercise were
excluded (Bennell et al., 2005; Deyle et al., 2005, 2000; Stoneman,
2001). The four remaining studies included manual therapy as
a sole intervention and met the inclusion criteria.
3.2. Characteristics of the included studies
All four studies were single-centre trials conducted in the
Netherlands (Hoeksma et al., 2004), Australia (Pollard et al., 2008),
the USA (Perlman et al., 2006) and South Africa (Tucker et al., 2003).
Three studies were parallel design (Hoeksma et al., 2004; Pollard
et al., 2008; Tucker et al., 2003), while one was a crossover
design where a control group received the intervention after an 8week period (Perlman et al., 2006).
Only one study had a long-term follow-up period of 29 weeks
(Hoeksma et al., 2004). One had a medium-term follow-up of 16
weeks (Perlman et al., 2006), however, as this was due to the crossover of control participants to the intervention, between-group data
were available only for the short-term follow-up, whilst two had
short-term follow-ups only (Pollard et al., 2008; Tucker et al., 2003).
Three studies recruited people with knee OA (Perlman et al.,
2006; Pollard et al., 2008; Tucker et al., 2003) and one recruited
participants with hip OA (Hoeksma et al., 2004). Sample sizes ranged
from 43 to 109 individuals. Two studies undertook power calculations to estimate sample size (Hoeksma et al., 2004; Perlman et al.,
2006). Recruitment of participants varied between studies. Two
used media advertisements (Pollard et al., 2008; Tucker et al., 2003),
one recruited from orthopaedic and rheumatology clinics (Hoeksma
et al., 2004) and one recruited from primary care physicians, senior
living facilities and rheumatology centres (Perlman et al., 2006). Two
studies recruited participants with mild to moderate OA (Pollard
et al., 2008; Tucker et al., 2003) and one included subjects of all
degrees of severity (Hoeksma et al., 2004). Perlman et al. (2006)
included those with WOMAC pain scores between 40 and 90 on
the 100 mm VAS scale. One study excluded end-stage OA on the basis
that the manipulation treatment was contra-indicated (Tucker et al.,
2003). Mean baseline pain, as measured with VAS was lower than
40 mm (on a 100 mm scale) in three studies (Hoeksma et al., 2004;
Tucker et al., 2003; Pollard et al., 2008), but between 60 and 70 mm
in the study by Perlman et al. (2006). Comparison of baseline function between the four trials was not possible due to the differences in
function outcomes used. Two studies (Hoeksma et al., 2004;
Perlman et al., 2006) used the American College of Rheumatology
criteria (Altman et al., 1991, 1986) whilst the other two (Pollard et al.,
2008; Tucker et al., 2003) used other criteria (Forman et al., 1983;
Manek and Lane, 2000) to diagnose OA.

H.P. French et al. / Manual Therapy 16 (2011) 109e117

Step 1

111

Computer database search

(n=2557)
Medline (PubMed) (n=591)
Medline (EBSCO) (n=258)
Medline (OVID) (n=268)
Cinahl (EBSCO) (n=234)
ISI (Web of Knowledge) (n=378)
Cochrane Library (n=109)
PsychINFO (n=166)
PEDro (n=234)
AMED (n=53)
Embase (n=266)

Duplicate articles excluded (n=1348)

Step 2

Screening of title for inclusion

exclusion (n= 1209)
Unrelated articles excluded based
on title (n=1123)
Not manual therapy (n=634)
Not RCT (n=259)
Not OA (n=73)
Not lower limb (n=42)
Not humans (n=13)
Not written in english (n=26)
Abstract only (n=76)
Retrieved abstracts for review (n=86)
Unrelated articles excluded by
abstract (n=63)
Not manual therapy (n=4)
Not RCT (n=47)
Not OA (n=2)
Not written in english (n=8)
Abstract only (n=2)

Step 3

Retrieved full articles for full quality

review (n= 23)
Articles excluded (n=15)
Not manual therapy (n 9)
Not RCT (n =2)
Not OA (n =2)
Not written in english (n=1)

Step 4

Articles with manual therapy

component (n=8)

Articles excluded (n=4)

Multimodal treatment (n=4)
Articles included in systematic review
manual therapy alone (n=4)
Fig. 1. Flow diagram of selection process of studies.

3.3. Interventions
The types of manual therapy differed between studies. Two used
chiropractic manipulation (Pollard et al., 2008; Tucker et al., 2003).
Manipulation and muscle stretching delivered by physiotherapists
was used by Hoeksma et al. (2004), and massage therapy using
Swedish full-body therapeutic massage was provided by massage

therapists (Perlman et al., 2006). Details of the interventions are in

Table 2.
All four studies had two group allocations but comparisons
differed. One compared manual therapy to exercise therapy
(Hoeksma et al., 2004), one compared to Meloxicam (a non
steroidal anti-inammatory drug) (NSAID) (Tucker et al., 2003) and
one compared to a placebo intervention of palmar contact followed

112

H.P. French et al. / Manual Therapy 16 (2011) 109e117

Table 2
Study characteristics of Included Trials.

InterIntervention manual therapy group; Mmale; Ffemale; SDstandard deviation; ROM range of motion; yyears; VASvisual analogue scale; NS non signicant.

by interferential therapy set at zero. The fourth study compared

massage therapy to a wait-list control group where treatment was
delayed for an 8-week period (Perlman et al., 2006).
Dosage of manual therapy differed between the studies. The
number of treatment sessions varied from six to 12, and number of
weeks of treatment varied from two to eight (Table 2).
With respect to outcome measurement, pain intensity was
measured with a Visual Analogue Scale (VAS) in all studies. A range
of self-report physical function outcomes was used in all studies,
while two studies also used observed physical function measures
(Hoeksma et al., 2004; Perlman et al., 2006). No studies used the
same functional outcomes (Table 2).
Adverse events were recorded in two studies which both
reported increased symptoms (Hoeksma et al., 2004; Perlman et al.,
2006) in a small number of participants (n 6/177) (3.4%).
Only one of the four studies reported on treatment adherence
(Hoeksma et al., 2004), where 7/106 (6.6%) patients were noncompliant, with similar numbers in both interventions.
3.4. Risk of bias
The number of criteria assessed for risk of bias ranged from 4 to 9
out a maximum score of 11, where lower scores indicated higher risk
of bias (Table 3). Although all studies were randomised, allocation

concealment was unclear in three (Perlman et al., 2006; Pollard et al.,

2008; Tucker et al., 2003). No study blinded the care provider. One
study administered a placebo intervention of palmar contact to the
joint instead of manual therapy (Pollard et al., 2008), but it is not
clear if patients were blinded. In the remaining studies, blinding of
the patient was not possible due to the nature of the comparison
interventions used. Just one study blinded the outcome assessor
(Hoeksma et al., 2004). Baseline comparability could not be ascertained in one study as no description of prognostic variables was
provided (Pollard et al., 2008). Two studies carried a low risk of bias
(Hoeksma et al., 2004; Perlman et al., 2006) and two had a high risk
of bias (Pollard et al., 2008; Tucker et al., 2003).
3.5. Effects of manual therapy
3.5.1. Comparison with no treatment
One study with high risk of bias (Perlman et al., 2006) showed
that massage therapy was signicantly better than no intervention for pain and function in individuals with mild to moderate
knee OA. Pain was measured on a 100 mm VAS and the mean
difference between the groups at 8 weeks was 20.62 mm in
favour of manual therapy, resulting in a large effect size of 0.86
[95% CI 0.37,1.36]. A minimal clinically important difference
(MCID) score of 1.7 points on a 100 mm VAS has been reported in

H.P. French et al. / Manual Therapy 16 (2011) 109e117

Table 3
Risk of bias assessment.

Randomisation
Allocation concealment
Blinding of patient
Blinding of care provider
Blinding of outcome assessor
Baseline comparability
Co-interventions avoided or similar
Similar timing of outcome assessments
Acceptable drop-out rate
Intention-to-treat analysis
Compliance acceptable?

Hoeksma Perlman Pollard

Tucker

Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Yes
Unclear
No
No
No
Yes
Unclear
Yes
Yes
No
Unclear

Yes
Unclear
No
No
Unclear
Yes
Unclear
Yes
Yes
Yes
Unclear

Yes
Unclear
No
No
Unclear
Unclear
Unclear
Yes
Yes
Yes
Unclear

low back pain (Coelho et al., 2008). Pain was also measured using
the 100 mm numerical rating scale version of WOMAC, and
a comparable difference of 20.11 mm in pain severity was
reported between the two groups. This resulted in a large effect
size of 0.94 [95% CI 0.44,1.44]. An MCID of 9.7 mm has been
reported in the WOMAC pain subscale (Ehrich et al., 2000) so
these results indicate a signicant benet of manual therapy
compared with no treatment for pain relief. Effect sizes for
changes in self-report function using the WOMAC and 50 foot
walk test were smaller at 0.60 [95% CI 0.11, 1.08] and 0.62 [95% CI
0.14, 1.11] respectively (Table 4). The mean difference for WOMAC
function subscale was 15.48 mm which was greater than the
established MCID of 9.3 mm (Ehrich et al., 2000), resulting in
a signicant effect size of 0.6 [95% CI 0.1, 1.08].
3.5.2. Comparison with placebo
One study with a high risk of bias (Pollard et al., 2008) found
a mean difference of 10.2 mm on a 100 mm VAS scale in favour of
manual therapy. Although this was greater than 1.7 mm change
considered to be clinically important, it resulted in a non-signicant effect size of 0.58 [95% CI 0.04,1.20]. Function was measured
using a self-report VAS of how well activities could be performed. A
mean difference of 2.7 cm in favour of the manual therapy group
was reported, with a signicant large effect size of 0.81 [95% CI
0.17,1.43].

3.5.3. Comparison with exercise therapy

One study with a low risk of bias (Hoeksma et al., 2004) found
a mean difference in favour of manual therapy of 9.1 mm on

113

a 100 mm VAS scale of pain at rest in individuals with hip OA,

resulting in a small effect size of 0.48 [95% CI 0.08, 0.87]. A similar
result occurred for pain during walking (mean difference of 9.6 mm
and effect size of 0.48 [95% CI 0.09, 0.87]). Both of these were
greater than the previously identied MCID of 1.7 mm score
(Coelho et al., 2008). A mean change of 1.5 points was found in the
100 point Bodily Pain subscale of the SF-36 health survey. This is
lower than the MCID of 7.8 points found in patients with lower limb
OA (Angst et al., 2001) and resulted in a small and non-signicant
effect size of 0.13 [95% CI 0.26,0.52]. A mean difference of 12.1
points was found on the Harris Hip Score resulting in a large effect
size of 0.85 [95% CI 0.45,1.25], whilst walking speed over 80 m
resulted in a small effect size of 0.40 [95% CI 0.01,0.79], both in
favour of manual therapy. The Physical Function SF-36 subscale
effect size was 0.09 [95% CI 0.3,0.48] which was non-signicant,
whilst the Role Physical subscale had a non-signicant effect size of
0.39 [95% CI 0.78, 0.00] in favour of exercise therapy over
manual therapy.
This was the only study to evaluate the long-term effects of
manual therapy (Table 5). A mean difference for pain at rest was
7 mm and 12.7 mm for pain when walking resulting in effect sizes
of 0.22 [95% CI 0.2,0.64] and 0.48 [95% CI 0.06,0.90] respectively.
The effect size on pain at rest was reduced from short-term followup and was non-signicant, but increased for pain with walking
(Table 5). There was a marginal reduction in effect size for SF-36
Bodily Pain subscale (ES 0.07 [95% 0.35,0.49]) which remained
non-signicant. Effect sizes for Harris hip score and 80 m walk test
in favour of manual therapy reduced at long-term follow-up,
resulting in effect sizes of 0.49 [95% CI 0.07, 0.91] and 0.41 [95% CI
0.01, 0.83] respectively. This resulted in a non-signicant longterm effect of manual therapy on walking speed, whilst Harris hip
score remained signicant. Effect sizes for both SF-36 physical
function (ES 0.18[95% CI 0.24, 0.60]) and role physical (ES 0.05
[95% CI 0.37, 0.47]) subscales improved from short-term followup but remained non-signicant (Table 5).
3.5.4. Comparison with Meloxicam (NSAID)
One study with a high risk of bias (Tucker et al., 2003) found
a mean difference of 2.07 mm in favour of manual therapy on
a 100 mm Visual Analogue Scale (VAS) pain scale (ES 0.18 [95% CI
0.33,0.69]) and no difference between groups on a 0e10
Numerical Rating Scale (NRS) (ES 0.2 [95% CI 0.31,0.71]). Both of
these effect sizes were non-signicant. A difference of 1.5 on an 11-

Table 4
Effect Sizes at short-term follow-up (Post-intervention).
Author

Outcome

Studies with Low Risk of Bias

Hoeksma
Pain
Pain
Pain
Function
Function
Function
Function
Perlman
Pain
Pain
Function
Function
Studies with High Risk Of Bias
Pollard
Pain
Function
Tucker
Pain
Pain
Function

Outcome Measure

Manual
Therapy (n)

Control (n)

Effect Size [95%CI]

Favours Manual Therapy

VAS rest pain

VAS walking pain
SF-36 Bodily Pain subscale
SF-36 Physical Function subscale
SF-36 Role Physical Subscale
Harris Hip Score
80 metre walk test
Pain VAS
WOMAC Pain subscale
WOMAC Function subscale
50 foot walk test

53
53
53
53
53
53
53
34
34
34
34

50
50
50
50
50
50
50
34
34
34
34

0.48 [0.08, 0.87]

0.48 [0.09, 0.87]
0.13 [0.26, 0.52]
0.09 [0.3. 0.48]
0.39 [0.78, 0.00]
0.85 [0.45, 1.25]
0.40 [0.01, 0.79]
0.86 [0.37, 1.36]
0.94 [0.44,1.44]
0.60 [0.11, 1.08]
0.62 [0.14, 1.11]

Yes
Yes
No difference
No difference
No difference
Yes
Yes
Yes
Yes
Yes
Yes

VAS
VAS
NRS
VAS
Patient Specic Functional Scale

26
26
30
30
30

17
17
30
30
30

0.58
0.81
0.18
0.20
0.15

No difference
Yes
No difference
No difference
No difference

[0.04, 1.20]
[0.17, 1.43]
[0.33, 0.69]
[0.31, 0.71]
[0.36, 0.66]

VAS Visual Analogue Scale, NRS Numerical Rating Scale, SF-36 Short Form-36, WOMAC Western Ontario and McMaster Universities Index of Osteoarthritis.

114

H.P. French et al. / Manual Therapy 16 (2011) 109e117

Table 5
Effect Sizes at long-term follow-up.
Author

Outcome

Studies with Low Risk of Bias

Hoeksma
Pain
Pain
Pain
Function
Function
Function
Function

Outcome Measure Used

Manual
Therapy (n)

Control (n)

Effect Size [95% CI]

Favours Manual
Therapy

VAS Pain at rest

VAS Pain with walking
SF-36 Bodily Pain subscale
SF-36 Physical Function subscale
SF-36 Role Physical Subscale
Harris Hip Score
80 metre walk test

44
44
44
44
44
44
44

44
44
44
44
44
44
44

0.22
0.48
0.07
0.18
0.05
0.49
0.41

No difference
Yes
No difference
No difference
No difference
Yes
No difference

[0.2, 0.64]
[0.06, 0.90]
[0.35, 0.49]
[0.24, 0.60]
[0.37, 0.47]
[0.07, 0.91]
[0.01, 0.83]

VAS Visual Analogue Scale, SF-36 Short Form-36, WOMAC Western Ontario and McMaster Universities Index of Osteoarthritis.

point PSFS scale was found in favour of manual therapy, resulting in

an effect size of 0.15 [95% CI 0.33,0.66] which was non-signicant.
4. Discussion
This is the rst known systematic review to evaluate the effect of
manual therapy in hip or knee OA. The small number of included
studies indicates the limited research to date. Although patients
were randomly allocated to groups in all trials, the method of
randomisation in three of the trials which used variations of preprepared sealed envelopes can be subject to bias compared with
computer-generated numbers (Pocock, 1983; Schulz, 1995). Blinding of treatment providers and study participants was a aw in all
studies but is difcult to achieve in non-pharmacological studies
(Boutron et al., 2003, 2004). Therefore, none of the evidence could
be graded as platinum or gold grade (Table 1). It was unclear from
three studies (Perlman et al., 2006; Pollard et al., 2008; Tucker et al.,
2003) if allocation concealment was used. Effect estimates in
studies with subjective outcomes can be exaggerated with unclear
or lack of allocation concealment or blinding (Wood et al., 2008).
There were some differences in demographic characteristics of
the patients (Table 2). Mean age of patients in the trials by Hoeksma
et al. (2004) and Perlman et al. (2006) were higher than the other
two studies. This may be related to the inclusion of patients with
more severe OA in these two trials as age is the strongest risk factor
for the development of OA (Sharma et al., 2006). There was
a greater proportion of females to males in three trials (Hoeksma
et al., 2004; Perlman et al., 2006; Tucker et al., 2003) which is
unsurprising as the prevalence of OA is higher in women (Felson,
2006).
There was variation in the severity of patients included in the
four trials. It was not clear from the two studies which only
included patients with mild to moderate OA (Pollard et al., 2008;
Tucker et al., 2003) how this was ascertained. Hoeksma et al.
(2004) used a recognised radiographic severity scale (Kellgren
and Lawrence, 1957) to dene severity, whilst cut-off scores of
the self-report WOMAC outcome were used to include participants
by Perlman et al. (2006) although no explanation of how and why
these criteria were chosen was provided. Currently, there are no
established clinical denitions of what constitutes mild, moderate
or severe OA although work is ongoing within international OA
research groups to create a composite measure based on pain,
physical function and structural changes to dene states of OA
severity (Gossec et al., 2007).
All four studies were heterogeneous regarding the types of
manual therapy and comparison interventions used. Findings
suggest that manual therapy may have a benecial short-term
effect in reducing pain and improving physical function for patients
with knee OA compared with no intervention, and in hip OA
compared with exercise therapy. There is also evidence that effects
of manual therapy on pain and function can be sustained long-term

(up to 6 months) in hip OA, although the size of the effect reduced
for most outcomes. Effect size varied between studies, with larger
effect sizes in the study where the contrast was no treatment
(Perlman et al., 2006) compared with an active intervention
(Hoeksma et al., 2004; Tucker et al., 2003). When manual therapy
was compared to placebo, although the effect size was large, the
condence intervals around the effect size resulted in a nonsignicant result (Pollard et al., 2008). A type II error may explain
these results, as sample size was small and no sample size estimation was done. Effect estimates can also be exaggerated in trials
of lower methodological quality (Moher et al., 1998).
Pain and physical function are recommended as core outcome
measures in rheumatology research (Bellamy et al., 1997). All
studies measured pain severity using a VAS but there was inconsistency in physical function measurement across all studies. A
range of physical function measures was used including diseasespecic measures (WOMAC, Harris Hip Score), general health (SF36 subscales, patient specic functional scale), functional tests (50
foot walk and 80 m walk) and VAS-scaled questions. Both Perlman
et al. (2006) and Hoeksma et al. (2004) used self-report and
physical performance based measures of function which is recommended as they measure difference aspects of function in OA
(Stratford and Kennedy, 2006).
In this review a broad denition of manual therapy was used.
There is a lack of clear description of what constitutes manual
therapy in the literature. This is evidenced by different criteria for
inclusion of studies in recent systematic reviews of manual
therapy (Camarinos and Marinko, 2009; Ho et al., 2009; Miller
et al., 2010; Reid and Rivett, 2005). In this review, three studies
used manipulation (Hoeksma et al., 2004; Pollard et al., 2008;
Tucker et al., 2003) applied to the target joint, whilst one delivered whole body soft tissue massage (Perlman et al., 2006). Jointbased techniques have a role in activating pain inhibitory cortical
systems (Skyba et al., 2003). Manual therapy has previously been
shown to induce immediate hypoalgesia in individuals with knee
OA, compared with placebo and control conditions, with concurrent improvements in function (Moss et al., 2006). Cadaveric
(Arvidsson, 1990; Harding et al., 2003) and human (Byrd and Chern,
1997) studies demonstrated that some joint movement does occur
with manual therapy applied to the hip but the effect may depend
on the force applied (Byrd and Chern, 1997). Massage therapy is
purported to have an effect on pain threshold, possibly through
endorphin release. It can also increase blood ow which may
release local pain mediators (Ernst, 1999). Psychological inuences
may also play a role due to practitioner and patient interaction
(Williams et al., 2007).
Manual therapy dosage, which varied between the studies, may
have inuenced results. The two studies which showed greater
effect provided treatments over ve (Hoeksma et al., 2004) and
eight weeks (Perlman et al., 2006), whilst the other two studies
treatment duration was two (Pollard et al., 2008) and three weeks

H.P. French et al. / Manual Therapy 16 (2011) 109e117

(Tucker et al., 2003). Manual therapy techniques can be applied in

different dosages in terms of force, amplitude, rate, repetition and
duration. There is a lack of evidence on the different aspects of
treatment dosage which could inuence clinical decisions. In many
manual therapy trials of musculoskeletal disorders, the dosage and
type of technique is based on the individual clinical presentation
(Deyle et al., 2005; Juni et al., 2009; van den Dolder and Roberts,
2006) and there is minimal research into the optimal dosage
required to produce a treatment effect. It is possible that results of
manual therapy trials are awed by this issue (Jull and Moore,
2002). Preliminary work has reported differential effects of
different manual therapy techniques in other musculoskeletal
disorders (Cleland et al., 2007; Leaver et al., 2010; Vermeulen et al.,
2006), but not in OA.
Manual therapy appears to be a safe intervention as increased
symptoms was the only adverse event reported, although they
were reported by two studies.
Experience of the treatment provider is an important factor in
non-pharmacological studies (Boutron et al., 2003). No study
provided detail on the experience of the treatment providers or
specic training which is an important consideration in nonpharmacological studies (Boutron et al., 2005). Treatment was
delivered by more than one practitioner in two studies (Hoeksma
et al., 2004; Perlman et al., 2006) and in one of these, training
was provided in the treatment protocol (Hoeksma et al., 2004).
Information on the treatment provider(s) was not provided in the
other studies (Pollard et al., 2008; Tucker et al., 2003). Difference in
expertise and lack of standardised treatment approaches may have
compromised study results. The different treatment approaches
and underlying philosophy of the manual therapy approaches used
by different health professionals (van de Veen et al., 2005) should
also be acknowledged and could contribute to discordant results.
This systematic review has some limitations. As only English
language and full-text articles were included, it is possible that
relevant studies may have been excluded from the review. Studies
that combined manual therapy with other interventions were not
included although four studies which combined manual therapy
with exercise therapy were identied during the search.
4.1. Implications for practice
The implementation of manual therapy as an independent
intervention cannot be justied based on the current ndings.
Clinical guidelines have recommended the use of manual therapy
as an adjunctive therapy to exercise in the management of OA
(NICE, 2008; RACGP, 2009), these recommendations are based on
limited evidence and there is a need to investigate the specic
effects of manual therapy to understand its role in management of
musculoskeletal function.
Although results demonstrated that manual therapy improves
pain and physical function, in the short and long-term, compared
with exercise for those with hip OA, and massage therapy provides
short-term benet in pain and function over no treatment for those
with knee OA, these ndings are based on just two RCTs of varying
quality. There was no evidence that manual therapy is better than
placebo or NSAID (Meloxicam) for pain and function in knee OA.
However, these results should be interpreted with caution due to
the high risk of bias.
4.2. Implications for research
There is a need for further research to conrm these results.
Methodological quality should be enhanced with adequate randomisation, concealed allocation and blinding of outcome assessor,
even if blinding of care provider or patients is not possible. Analysis

115

should be based on the intention-to-treat principle. Future studies

should include long-term outcome assessments (>6 months). In
light of recommendations from clinical guidelines, which are
currently based on limited evidence, consideration should be given
to evaluating manual therapy, either on its own or with other
treatments such as exercise therapy using factorial designs. These
would allow for the investigation of the specic effects of manual
therapy and its interaction with other interventions. More research
is required to compare different types of manual therapy, to
establish if one form of manual therapy is more benecial over
another and what dosage is required to have a meaningful effect.
Future research should also assess the effect of the skill level and
experience of the care provider on manual therapy effectiveness.
Researchers should try to use standardised and recognised
outcome measures, particularly in the measurement of both selfreport and observed physical function measures. Groups such as
OMERACT, which is an international collaboration initiative in the
eld of rheumatology, are working towards standardising outcome
measures in clinical trials of OA (Tugwell et al., 2007) but there is
a responsibility on researchers to use outcomes with strong
psychometric properties and international recognition.
5. Conclusion
From the available data, there is silver level evidence that
manual therapy has a benecial effect compared with exercise
therapy in those with hip OA, both in the short and long-term for
pain reduction and increased physical function. These results are
based on one study with a low risk of bias and a small sample. There
is less convincing evidence from three studies with a high risk of
bias that manual therapy is no better than placebo or Meloxicam
medication, but manual therapy, in the form of massage therapy, is
effective compared to no intervention in knee OA.
Author contributions
Study conception and design: French, Brennan, White, Cusack.
Acquisition of data: French, Brennan, White, Cusack.
Analysis and interpretation of data: French, Brennan, Cusack.
Manuscript preparation: French, Brennan, Cusack.
All authors were involved in critically appraising the manuscript
for intellectual content and approved the nal version prior to
submission for publication. Ms French had full access to all of the
data in the study and takes responsibility for the integrity of the
data and accuracy of the data analysis.
Acknowledgements
We are extremely grateful to Professor Adam Perlman and Dr
James Brantingham for providing specic data from their trials. We
thank Professor Ronan Conroy, Royal College of Surgeon in Ireland
(RCSI) for his statistical advice. Sincere thanks also to Ms Grainne
McCabe, Librarian, RCSI, Anne Murphy, Librarian AMNCH, Kathryn
Smith, Deputy Head of Library Academic Services Life Sciences,
UCD and Ms Angela Rice, Librarian, MMUH for their assistance with
literature searching.
Appendix A. Sample search strategy
1.
2.
3.
4.
5.
6.

hip osteoarthritis (Subject/Mesh heading)

knee osteoarthritis (Subject/Mesh heading)
coxarthrosis (free text heading)
gonarthrosis (free text heading)
Knee
hip

116

7.
8.
9.
10.
11.
12.
13.
14.

15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.

H.P. French et al. / Manual Therapy 16 (2011) 109e117

osteoarthrosis (free text heading)

osteoarthrit* (free text heading)
Combine knee and osteoarthrit* (#5 and #7)
Combine knee and osteoarthrosis (#5 and #8)
Combine hip and osteoarthrit* (#6 and #7)
Combine hip and osteoarthrosis (#6 and #8)
Combine 1or 2 or 3 or 4 or 9 or 10 or 11 or 12.
musculoskeletal manipulations (MeSH Heading)/Manual
therapy (CINAHL subject heading)/manipulative medicine
(EMBASE heading)
manipulat*
manual ther* (free text heading)
joint mobili?ation
massage (free text heading)
muscle energy (free text heading)
manual tract* (free text heading)
neuromuscular technique (free text heading)
proprioceptive neuromuscular facilitation (free text heading)
chiropract* (free text heading)
osteopath* (free text heading)
Combine 14e24 (using OR)
Combine 13 and 25

References
Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, Brandt K, et al. The
American College of Rheumatology criteria for the classication and reporting
of osteoarthritis of the hip. Arthritis and Rheumatism 1991;34(5):505e14.
Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, et al. Development of
criteria for the classication and reporting of osteoarthritis. Classication of
osteoarthritis of the knee. Diagnostic and therapeutic criteria committee of the
american rheumatism association. Arthritis and Rheumatism 1986;29
(8):1039e49.
Angst F, Aeschlimann A, Stucki G. Smallest detectable and minimal clinically
important differences of rehabilitation intervention with their implications for
required sample sizes using WOMAC and SF-36 quality of life measurement
instruments in patients with osteoarthritis of the lower extremities. Arthritis
and Rheumatism 2001;45(4):384e91.
Arvidsson I. The hip joint: forces needed for distraction and appearance of the
vacuum phenomenon. Scandinavian Journal of Rehabilitation Medicine 1990;22
(3):157e61.
Bellamy N, Kirwan J, Boers M, Brooks P, Strand V, Tugwell P, et al. Recommendations
for a core set of outcome measures for future phase III clinical trials in knee, hip
and hand osteoarthritis. Consensus development at OMERACT III. Journal of
Rheumatology 1997;24(4):799e802.
Bennell KL, Hinman RS, Metcalf BR, Buchbinder R, McConnell J, McColl G, et al.
Efcacy of physiotherapy management of knee joint osteoarthritis: a randomised, double blind, placebo controlled trial. Annals of the Rheumatic Diseases
2005;64(6):906e12.
Boutron I, Moher D, Tugwell P, Giraudeau B, Poiraudeau S, Nizard R, et al. A checklist
to evaluate a report of a nonpharmacological trial (CLEAR NPT) was developed
using consensus. Journal of Clinical Epidemiology 2005;58(12):1233e40.
Boutron I, Tubach F, Giraudeau B, Ravaud P. Methodological differences in clinical
trials evaluating nonpharmacological and pharmacological treatments of hip
and knee osteoarthritis. Journal of the American Medical Association 2003;290
(8):1062e70.
Boutron I, Tubach F, Giraudeau B, Ravaud P. Blinding was judged more difcult to
achieve and maintain in nonpharmacologic than pharmacologic trials. Journal
of Clinical Epidemiology 2004;57(6):543e50.
Brandt KD. The role of analgesics in the management of osteoarthritis pain.
American Journal of Therapeutics 2000;7(2):75e90.
Byrd JW, Chern KY. Traction versus distension for distraction of the joint during hip
arthroscopy. Arthroscopy 1997;13(3):346e9.
Camarinos J, Marinko L. Effectiveness of manual physical therapy for painful
shoulder conditions: a systematic review. Journal of Manual and Manipulative
Therapy 2009;17(4):206e15.
Cameron HU, Macnab I. Observations on osteoarthritis of the hip joint. Clinical
Orthopaedics and Related Research 1975;108:31e40.
Cleland JA, Glynn P, Whitman JM, Eberhart SL, MacDonald C, Childs JD, et al. Shortterm effects of thrust versus nonthrust mobilization/manipulation directed at
the thoracic spine in patients with neck pain: a randomized clinical trial.
Physical Therapy 2007;87(4):431e40.
Cochrane Cystic Fibrosis and Genetic Disorder Group. Data extraction form, www.
cfgd.cochrane.org/./Study%20selection%20and%20%20extraction%20form.doc;
2009 (accessed 3/3/2010).

Coelho RA, Siqueira FB, Ferreira PH, Ferreira ML. Responsiveness of the BrazilianPortuguese version of the Oswestry Disability Index in subjects with low back
pain. European Spine Journal 2008;17(8):1101e6.
Cohen J. Statistical power: analysis for the behavioural sciences. New York:
Academic Press; 1977. pp. 1e27.
Deeks J, Altman DG, Bradburn M. Statistical methods for examining heterogeneity
and combining resuls from several studies in meta-analysis. In: Egger M,
Smith GDavey, Altman DG, editors. Systematic reviews in healthcare: metaanalysis in context. London: BMJ Books; 2007. pp. 285e312.
Deyle GD, Allison SC, Matekel RL, Ryder MG, Stang JM, Gohdes DD, et al.
Physical therapy treatment effectiveness for osteoarthritis of the knee:
a randomized comparison of supervised clinical exercise and manual
therapy procedures versus a home exercise program. Physical Therapy
2005;85(12):1301e17.
Deyle GD, Henderson NE, Mateke RL, Ryder MG, Garber MB, Allison SC, et al.
Effectiveness of manual physical therapy and exercise in osteoarthritis of the
knee. A randomized controlled trial. Annals of Internal Medicine 2000;132
(3):173e81.
Ehrich EW, Davies GM, Watson DJ, Bolognese JA, Seidenberg BC, Bellamy N, et al.
Minimal perceptible clinical improvement with the Western Ontario and
McMasterUniversities osteoarthritis index questionnaire and global assessments in patients with osteoarthritis. Journal of Rheumatology 2000;27
(11):2635e41.
Ernst E. Massage therapy for low back pain: a systematic review. Journal of Pain and
Symptom Management 1999;17(1):65e9.
Felson DT. Epidemiology of osteoarthritis. In: Brandt K, Doherty M, Lohmander LS,
editors. Osteoarthritis. Oxford University Press; 2006. pp. 13e22.
Forman MD, Malamet R, Kaplan D. A survey of osteoarthritis of the knee in the
elderly. Journal of Rheumatology 1983;10(2):282e7.
Gossec L, Hawker G, Davis AM, Maillefert JF, Lohmander LS, Altman R, et al.
OMERACT/OARSI initiative to dene states of severity and indication for joint
replacement in hip and knee osteoarthritis. Journal of Rheumatology 2007;34
(6):1432e5.
Harding L, Barbe M, Shepard K, Marks A, Ajai R, Lardiere J, et al. Posterior-anterior
glide of the femoral head in the acetabulum: a cadaver study. Journal of
Orthopaedic and Sports Physical Therapy 2003;33(3):118e25.
Higgins J, Green S. Cochrane handbook for systematic reviews of interventions,
version 5.0.0. The Cochrane Collaboration, www.cochrane-handbook.org; 2009
(accessed 10/4/2010).
Hinton R, Moody RL, Davis AW, Thomas SF. Osteoarthritis: diagnosis and therapeutic considerations. American Family Physician 2002;65(5):841e8.
Ho C-YC, Sole G, Munn J. The effectiveness of manual therapy in the management of
musculoskeletal disorders of the shoulder: a systematic review. Manual
Therapy 2009;14(5):463e74.
Hoeksma HL, Dekker J, Ronday HK, Heering A, van der Lubbe N, Vel C, et al.
Comparison of manual therapy and exercise therapy in osteoarthritis of the hip:
a randomized clinical trial. Arthritis Care and Research 2004;51(5):722e9.
Hoving J, Jull G, Koes B. Methodological and practical issues in clinical trials on
manual therapy. In: Boyling JD, Jull G, editors. Grieves modern manual
therapy: the vertebral column. Edinburgh: Elsevier Churchill Livingstone;
2004. pp. 581e90.
Jacobs JH. Osteoarthritis of the hip-joint. Annals of Physical Medicine
1960;5:318e26.
Jull G, Moore A. What is a suitable dosage of physical therapy treatment? Manual
Therapy 2002;7(4):181e2.
Juni P, Battaglia M, Nuesch E, Hammerle G, Eser P, van Beers R, et al. A randomised
controlled trial of spinal manipulative therapy in acute low back pain. Annals of
the Rheumatic Diseases 2009;68(9):1420e7.
Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Annals of the
Rheumatic Diseases 1957;16(4):494e502.
Leaver AM, Maher CG, Herbert RD, Latimer J, McAuley JH, Jull G, et al. A randomized
controlled trial comparing manipulation with mobilization for recent onset
neck pain. Archives of Physical Medicine and Rehabilitation 2010;91(9):1313e8.
Manek NJ, Lane NE. Osteoarthritis: current concepts in diagnosis and management.
American Family Physician 2000;61(6):1795e804.
Miller J, Gross A, DSylva J, Burnie SJ, Goldsmith CH, Graham N, et al. Manual therapy and
exercise for neck pain: a systematic review. Manual Therapy 2010;15(4):334e54.
Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M, et al. Does quality of reports
of randomised trials affect estimates of intervention efcacy reported in metaanalyses? Lancet 1998;352(9128):609e13.
Moss P, Sluka K, Wright A. The initial effects of knee joint mobilization on osteoarthritic hyperalgesia. Manual Therapy 2006;12(2):109e18.
NICE (National Institute of Clinical Excellence). The care and management of osteoarthritis in adults, http://www.nice.org.uk/nicemedia/pdf/CG59NICEguideline.
pdf; 2008 (accessed 20/11/2009).
Perlman AI, Sabina A, Williams AL, Njike VY, Katz DL. Massage therapy for osteoarthritis of the knee: a randomized controlled trial. Archives of Internal
Medicine 2006;166(22):2533e8.
Pocock S. Clinical trials: a practical approach. Chichester, UK: John Wiley; 1983.
Pollard H, Ward G, Hoskins W, Hardy K. The effect of a manual therapy knee
protocol on osteoarthritic knee pain: a randomised controlled trial. Journal of
the Canadian Chiropractic Association 2008;52(4):229e42.
RACGP (Royal Australian College of General Practitioners). Guideline for the nonsurgical management of hip and knee osteoarthritis, http://www.racgp.org.au/
guidelines/osteoarthritis; 2009 (accessed 10/6/2010).

H.P. French et al. / Manual Therapy 16 (2011) 109e117

Reginster JY. The prevalence and burden of arthritis. Rheumatology 2002;41
(Suppl. 1):3e6.
Reid SA, Rivett DA. Manual therapy treatment of cervicogenic dizziness: a systematic review. Manual Therapy 2005;10(1):4e13.
Salaf F, Carotti M, Grassi W. Health-related quality of life in patients with hip or
knee osteoarthritis: comparison of generic and disease-specic instruments.
Clinical Rheumatolology 2005;24(1):29e37.
Sarzi-Puttini P, Cimmino MA, Scarpa R, Caporali R, Parazzini F, Zaninelli A, et al.
Osteoarthritis: an overview of the disease and its treatment strategies. Seminars in Arthritis and Rheumatism 2005;35(Suppl. 1):1e10.
Schulz KF. Subverting randomization in controlled trials. Journal of the American
Medical Association 1995;274(18):1456e8.
Sharma L, Kapoor D, Issa S. Epidemiology of osteoarthritis: an update. Current
Opinion in Rheumatology 2006;18(2):147e56.
Skyba DA, Radhakrishnan R, Rohlwing JJ, Wright A, Sluka KA. Joint manipulation
reduces hyperalgesia by activation of monoamine receptors but not opioid or
GABA receptors in the spinal cord. Pain 2003;106(1e2):159e68.
Steultjens MP, Dekker J, van Baar ME, Oostendorp RA, Bijlsma JW. Range of joint
motion and disability in patients with osteoarthritis of the knee or hip. Rheumatology 2000;39(9):955e61.
Steultjens MP, Dekker J, van Baar ME, Oostendorp RA, Bijlsma JW. Muscle strength,
pain and disability in patients with osteoarthritis. Clinical Rehabilitation
2001;15(3):331e41.
Stoneman PD. Effect of manual therapy and exercise on pain, stiffness and function
in persons with knee osteoarthritis. Brigham Young University; 2001.
Stratford PW, Kennedy DM. Performance measures were necessary to obtain
a complete picture of osteoarthritic patients. Journal of Clinical Epidemiology
2006;59(2):160e7.

117

Tucker M, Brantingham JW, Myburg C. Relative effectiveness of a non-steroidal antiinammatory medication (Meloxicam) versus manipulation in the treatment of
osteo-arthritis of the knee. European Journal of Chiropractic 2003;50(3):163e83.
Tugwell P, Boers M, Brooks P, Simon L, Strand V, Idzerda L. OMERACT: an international
initiative to improve outcome measurement in rheumatology. Trials 2007;8:38.
Tugwell P, OShea B. Evidence-based Rheumatology. London: BMJ Books; 2004.
van de Veen EA, de Vet HC, Pool JJ, Schuller W, de Zoete A, Bouter LM. Variance in
manual treatment of nonspecic low back pain between orthomanual physicians, manual therapists, and chiropractors. Journal of Manipulative and
Physiological Therapeutics 2005;28(2):108e16.
van den Dolder PA, Roberts DL. Six sessions of manual therapy increase knee exion
and improve activity in people with anterior knee pain: a randomised
controlled trial. Australian Journal of Physiotherapy 2006;52(4):261e4.
van Tulder M, Furlan A, Bombardier C, Bouter L. Updated method guidelines for
systematic reviews in the cochrane collaboration back review group. Spine
2003;28(12):1290e9.
Vermeulen HM, Rozing PM, Obermann WR, le Cessie S, Vliet Vlieland TP.
Comparison of high-grade and low-grade mobilization techniques in the
management of adhesive capsulitis of the shoulder: randomized controlled
trial. Physical Therapy 2006;86(3):355e68.
Williams NH, Hendry M, Lewis R, Russell I, Westmoreland A, Wilkinson C.
Psychological response in spinal manipulation (PRISM): a systematic review of
psychological outcomes in randomised controlled trials. Complementary
Therapies in Medicine 2007;15(4):271e83.
Wood L, Egger M, Gluud LL, Schulz KF, Juni P, Altman DG, et al. Empirical evidence of
bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. British Medical Journal
2008;336(7644):601e5.