VOLUME 12 NUMBER 3 PAGES 199– 296 August 2007

Editors International Advisory Board

Ann Moore PhD, GradDipPhys, FCSP, CertEd, FMACP K. Bennell (Victoria, Australia)
Clinical Research Centre for Health Professions K. Burton (Hudders¢eld, UK)
University of Brighton B. Carstensen (Frederiksberg, Denmark)
Aldro Building, 49 Darley Road E. Cruz (Setubal Portugal)
Eastbourne BN20 7UR, UK L. Danneels (Mar|¤ akerke, Belgium)
Gwendolen Jull PhD, MPhty, Grad Dip ManTher, FACP S. Durrell (London, UK)
Department of Physiotherapy S. Edmondston (Perth, Australia)
University of Queensland J. Endresen (Flaktvei, Norway)
Brisbane QLD 4072, Australia L. Exelby (Biggleswade, UK)
J. Greening (London, UK)
Associate Editor’s C. J. Groen (Utrecht,The Netherlands)
Darren A. Rivett PhD, MAppSc, (ManipPhty) A. Gross (Hamilton, Canada)
T. Hall (West Leederville, Australia)
GradDipManTher, BAppSc (Phty)
W. Hing (Auckland, New Zealand)
Discipline of Physiotherapy
M. Jones (Adelaide, Australia)
Faculty of Health S. King (Glamorgan, UK)
The University of Newcastle B.W. Koes (Amsterdam,The Netherlands)
Callaghan, NSW 2308, Australia J. Langendoen (Kempten, Germany)
Tim McClune D.O. D. Lawrence (Davenport, IA, USA)
Spinal Research Unit. D. Lee (Delta, Canada)
University of Hudders¢eld R. Lee (Brighton, UK)
30 Queen Street C. Liebenson (Los Angeles, CA, USA)
Hudders¢eld L. Ma¡ey-Ward (Calgary, Canada)
HD12SP, UK E. Maheu (Quebec, Canada)
C. McCarthy (Coventry, UK)
Editorial Committee J. McConnell (Northbridge, Australia)
Masterclass Editor S. Mercer (Queensland, Australia)
Karen Beeton PhD, MPhty, BSc(Hons), MCSP D. Newham (London, UK)
MACP ex o⁄cio member J. Ng (Hung Hom, Hong Kong)
Associate Head of School (Professional Development) L. Ombregt (Kanegem-Tielt, Belgium)
School of Health and Emergency Professions N. Osbourne (Bournemouth, UK)
University of Hertfordshire M. Paatelma (Jyvaskyla, Finland)
College Lane N. Petty (Eastbourne, UK)
Hat¢eld AL10 9AB, UK A. Pool-Goudzwaard (The Netherlands)
M. Pope (Aberdeen, UK)
Case reports & Professional Issues Editor G. Rankin (London, UK)
Je¡rey D. Boyling MSc, BPhty, D. Reid (Auckland, New Zealand)
GradDipAdvManTher, MCSP, MErgS A. Rushton (Birmingham, UK)
Je¡rey Boyling Associates C. Shacklady (Manchester, UK)
Broadway Chambers M. Shacklock (Adelaide, Australia)
Hammersmith Broadway D. Shirley (Lidcombe, Australia)
LondonW6 7AF, UK V. Smedmark (Stenhamra, Sweden)
Book Review Editor W. Smeets (Tongeren, Belgium)
Raymond Swinkels MSc, PT, MT C. Snijders (Rotterdam,The Netherlands)
Ulenpas 80 R. Soames (Dundee, UK)
P. Spencer (Barnstaple, UK)
5655 JD Eindoven
M. Sterling (St Lucia, Australia)
The Netherlands
P. Tehan (Victoria, Australia)
M. Testa (Alassio, Italy)
M. Uys (Tygerberg, South Africa)
P. van der Wu¡ (Doorn,The Netherlands)
P. van Roy (Brussels, Belgium)
B.Vicenzino (St Lucia, Australia)
H.J.M.Von Piekartz (Wierden,The Netherlands)
M.Wallin (Spanga, Sweden)
M.Wessely(Paris, France)
A.Wright (Perth, Australia)
M. Zusman (Mount Lawley, Australia)

Visit the journal website at http://www.intl.elsevierhealth.com/journals/math

doi:10.1016/S1356-689X(07)00094-X
 ARTICLE IN PRESS
Manual Therapy 12 (2007) 199
Editorial
Physiotherapists: Closer together, not worlds apart
Since the last issue of Manual Therapy Journal, the
Fifteenth World Congress of Physical Therapy (WCPT)
has taken place in Vancouver, British Columbia,
Canada. The conference saw the meeting of up to
4000 physical therapists’ minds! One thousand six
hundred and forty-six presentations were made at
Congress by representatives from the five regions of
WCPT. The largest proportion of presentations came
from Europe with the Americas and the Antipodes and
South East Asia also forming a large proportion of the
physiotherapy contingent. Musculoskeletal physiother-
apy research and development presentations formed
the largest proportion of the specialities represented.
The 4-yearly WCPT events provide a rich and fertile
environment for the exchange of views, ideas, knowl-
edge and research findings between physiotherapists
from across the world.
During Congress, an important event took place—the
inaugural meeting of the International Society of
Physiotherapy Journal Editors (ISPJE). Forty phy-
siotherapy journals were represented at this event and
we are delighted to say that the editors of Manual
Therapy Journal were part of the development of this
international society and were represented at the event.
The ISPJE aims to promote the sharing of knowledge,
expertise and good practice among international phy-
siotherapy editors and raise the standards within and the
academic profile of physiotherapy publication across the
world. It is also hoped that by working together, we can
facilitate more access to publications for those people
living in developing countries. It was a privilege to be
part of this innovation and we look forward to the
opportunities that such collaborations can provide for
Manual Therapy Journal, but also for the physiotherapy
profession as a whole.
WCPT once again demonstrated the goodwill and
bonhomie that occurs when physiotherapists from
around the world get together. Research collaborations
flourished, new developments in technology and practice
were shared, new research findings were taken to all
corners of the earth for further dissemination by the
message bearers. The conference makeup this time
1356-689X/$ - see front matter r 2007 Published by Elsevier Ltd.
doi:10.1016/j.math.2007.06.001
www.elsevier.com/locate/math
provided a number of learning opportunities, not just
from the multi-speciality content perspective, but the
method of delivery, e.g. platform presentations, viper
discussions, poster presentations, workshops, symposia,
discussion panels, debates, pre- and post-congress
courses, clinical visits and last-but-not-least, and prob-
ably most importantly, the social events and informal
networking opportunities. It was indeed a magical
learning environment for everyone.
Importantly, WCPT highlights the positive benefits of
physiotherapists from all specialities learning together
and the fact that as a profession we possess many skills
which are transferable across speciality boundaries and
many issues and developments occurring in some areas
of practice have strong applications and relevance in
other areas. We can all learn from experiences that have
occurred in other specialities and we should not forget
to do so.
Manual Therapy Journal will be publishing a Special
Issue, including selected key musculoskeletal papers
from Congress before the end of 2007. We are pleased to
announce that Associate Professor Darren Rivett of the
Editorial Board for Manual Therapy will be guest editor
for this Special Issue.
While we celebrate the benefits and successes of
WCPT, we must now prepare for the next international
event and the most important event in the musculoske-
letal calendar—IFOMT 2008 which will be held in June
2008 in Rotterdam.
This conference presents us with a superb opportunity
to share, learn, collaborate, debate and discuss issues of
importance to musculoskeletal physiotherapists world-
wide. We should all look forward to hearing about the
latest cutting-edge research and taking part in the
philosophical debate that will inevitably occur at this
most exciting event. We look forward to seeing you all
in Rotterdam.
(Co-Editors)
Ann Moore, Gwen Jull
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 200–208

www.elsevier.com/locate/math

Review

Iliotibial band friction syndrome—A systematic review

Richard Ellis, Wayne Hing, Duncan Reid
Health Rehabilitation Research Centre, Division of Rehabilitation and Occupation Studies, Faculty of Health & Environmental Sciences, AUT
University, Private Bag 92006, Auckland, New Zealand
Received 19 October 2004; received in revised form 19 July 2006; accepted 30 August 2006

Abstract

Iliotibial band friction syndrome (ITBFS) is a common injury of the lateral aspect of the knee particularly in runners, cyclists and
endurance sports. A number of authors suggest that ITBFS responds well to conservative treatment, however, much of this opinion
appears anecdotal and not supported by evidence within the literature. The purpose of this paper is to provide a systematic review of
the literature pertaining to the conservative treatment of ITBFS.
A search to identify clinical papers referring to the iliotibial band (ITB) and ITBFS was conducted in a number of electronic
databases using the keyword: iliotibial. The titles and abstracts of these papers were reviewed to identify papers specifically detailing
conservative treatments of ITBFS. The PEDro Scale, a systematic tool used to critique randomized controlled trials (RCTs), was
employed to investigate both the therapeutic effect of conservative treatment of ITBFS and also to critique the methodological
quality of available RCTs examining the conservative treatment of ITBFS.
With respect to the management of ITBFS, four RCTs were identified. The interventions examined included the use of non-
steroidal anti-inflammatory drugs (NSAIDs), deep friction massage, phonophoresis versus immobilization and corticosteroid
injection.
This review highlights both the paucity in quantity and quality of research regarding the conservative treatment of ITBFS. There
seems limited evidence to suggest that the conservative treatments that have been studied offer any significant benefit in the
management of ITBFS. Future research will need to re-examine those conservative therapies, which have already been examined,
along with others, and will need to be of sufficient quality to enable accurate clinical judgements to be made regarding their use.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Iliotibial band; Iliotibial band friction syndrome; Systematic review; Conservative treatment

1. Introduction (Kirk et al., 2000; Brosseau et al., 2004). However, it is

Iliotibial band friction syndrome (ITBFS) was first
specifically described by Renne (1975) as a pain felt on
the lateral aspect of the knee with lower limb activities
such as running and cycling. Following an increase in the
popularity of running and other endurance multi-
disciplinary sports, since the 1980s, ITBFS has become
more common (Anderson, 1991; Kirk et al., 2000). The
overall incidence of ITBFS can range from between 1.6%
and 52% depending on which population you examine
Corresponding author. Tel.: +64 9 921 9999x7800;
fax: +64 9 921 9620.
E-mail address: wayne.hing@aut.ac.nz (W. Hing).
generally accepted that ITBFS is the most common
running injury of the lateral knee, and has an incidence of
between 1.6% and 12% (Orava, 1978; McNicol et al.,
1981; Messier et al., 1995; Fredericson et al., 2000;
Taunton et al., 2002). Within cycling, ITBFS is believed
to account for 15–24% of overuse injuries (Farrell et al.,
2003; Holmes and Pruitt, 1993). Its incidence in military
recruits may range from 1% to 5.3% (Jordaan and
Schwellnus, 1994; Almeida et al., 1999). ITBFS is
uncommon in the inactive population (Orava, 1978).
The aetiology of ITBFS is multi-factorial with
representation of both intrinsic and extrinsic factor
(McNicol et al., 1981; Kirk et al., 2000; Taunton et al.,
2002). ITBFS in a non-traumatic overuse injury caused

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.08.004
 ARTICLE IN PRESS
R. Ellis et al. / Manual Therapy 12 (2007) 200–208
by friction/rubbing of the distal portion of the iliotibial
band (ITB) over the lateral femoral epicondyle (LFE)
with repeated flexion and extension of the knee. Orchard
et al. (1996) describe an ‘impingement zone’ which occurs
at approximately 301 of knee flexion during foot-strike
and early stance phase. At approximately 301 and greater,
of knee flexion, the ITB passes over and posterior to the
LFE (Renne, 1975; McNicol et al., 1981; Anderson, 1991;
Barber and Sutker, 1992; Aronen et al., 1993; Puniello,
1993; Messier et al., 1995; Novacheck, 1998; Fredericson
et al., 2002; Farrell et al., 2003). During the impingement
period, eccentric contraction of the tensor fascia lata
(TFL) and gluteus maximus, to decelerate the leg whilst
running, exert great tension through the ITB (Orchard
et al., 1996; Kirk et al., 2000; Farrell et al., 2003). Farrell
et al. (2003) described a similar impingement zone for
cycling.
The pathogenesis of ITBFS involves inflammation and
irritation of the lateral synovial recess (Renne, 1975; Orava,
1978; McNicol et al., 1981; Ekman et al., 1994; Nemeth
and Sanders, 1996; Nishimura et al., 1997; Kirk et al., 2000;
Levin, 2003), as well as continued irritation of the posterior
fibres of the ITB (Ekman et al., 1994; Fredericson et al.,
2000; Kirk et al., 2000; Austermuehle, 2001; Fredericson
et al., 2002) and inflammation of the periosteum of the
LFE (McNicol et al., 1981; Noble et al., 1982; Nishimura et
al., 1997; Fredericson et al., 2002), all of which describes
the pathogenesis of ITBFS. Kirk et al. (2000) suggest that
with repetitive soft tissue irritation there is simply not
enough time for the body to repair these damaged tissues.
This may lead to further irritation and injury which, in
theory, would extend the area of the impingement zone and
increase the risk of irritation (Levin, 2003).
A number of authors have commented that ITBFS
responds well to conservative treatment (Anderson, 1991;
Kirk et al., 2000; Levin, 2003) with success rates reported
as high as 94% (McNicol et al., 1981). A number of
different treatment options are reported in the literature,
however, it should be questioned whether these treat-
ments are delivered based on sound evidence.
The purpose of this paper is to perform a systematic
review, evaluating the efficacy of conservative treatment
of ITBFS, in order to highlight key concepts to guide
evidence-based practice in the management of ITBFS.
Relevant functional anatomical and biomechanical
contributions to the aetiology and pathomechanics of
ITBFS will also be discussed and related back to the
findings of the RCTs available.
2. Methodology
2.1. Literature search strategy
A search to identify clinical papers, clinical reviews
and clinical trials pertaining to the ITB and ITBFS, was
201
conducted in electronic databases, subscribed to by the
Auckland University of Technology (AUT) library,
which included MEDLINE via PubMed (from 1966
onwards), Cumulative Index to Nursing and Allied
Health Literature (CINAHL) (from 1983 onwards),
The Cochrane Controlled Trials Register in the
Cochrane Library (latest edition), SPORT-Discus
(from 1830 onwards), Allied and Complementary
Medicine Database (AMED) (from 1985 onwards),
Blackwell-Synergy, Master FILE (from 1975 onwards),
Expanded Academic ASAP (from 1980 onwards),
Index New Zealand (INNZ) (from 1987 onwards),
Lippincott 100 Nursing and Health Science Collection,
Physiotherapy Evidence Database (PEDro) (from 1953
onwards), ProQuest 5000 International, ProQuest
Health and Medical Complete, Web of Science (from
1945 onwards), Wiley Interscience–Life and Medical
Sciences Titles. This search was conducted in August–
September 2004.
The ITB and ITBFS were deemed to be relatively
narrow fields to search, therefore only one Medical
Subject Heading (MESH) was used as a keyword:
iliotibial. There was no limitation regarding date or
language leading to 1260 citations being identified of
which many were repeated across databases.
The titles and/or abstracts of these citations were
reviewed to identify papers specifically detailing the
aetiology and conservative treatment of ITBFS and the
anatomy and biomechanics of the ITB. The bibliogra-
phies of each paper were also used for cross-referencing
to identify other relevant papers.
2.2. Study selection
Inclusion criteria: The following criteria were used in
order to select relevant papers to be included within this
review:
Type of participant: Participants to be 18 years of age
and older, of either gender and have a clinical
diagnosis of ITBFS for greater than 14 days duration.
Type of study design: Randomized controlled trials.
Type of intervention: Conservative treatment of
ITBFS, i.e. non-surgical.
Outcome measurements: To include at least one of the
following outcome measurements: pain rating (e.g.
Visual Analogue Scale (VAS)), function-specific VAS
(i.e. work or sport related pain), time from diagnosis
until symptom free, return to work and/or sport
status.
Exclusion criteria: The following criteria were used to
eliminate papers from this review: papers written in non-
English languages, non-RCTs, RCTs which utilized
non-conservative treatment, i.e. surgical interventions.
 ARTICLE IN PRESS
202 R. Ellis et al. / Manual Therapy 12 (2007) 200–208

2.3. Review of methodological quality paper are included in Table 2.

Three reviewers independently assessed each of the
RCTs identified for their respective methodological
quality. The PEDro Scale (see Table 1), developed by
The Centre of Evidence-Based Physiotherapy (CEBP)
was utilized to assess each paper. The PEDro Scale is an
11-item scale, which is a validated and versatile tool
used to rate RCTs for the PEDro Database (Clark et al.,
1999; Maher et al., 2003).
An overall score of methodological quality, or quality
score (QS), was determined for each paper by each of
the three reviewers as a total of positive scores for each
of the 11 items. A consensus method was used to discuss
and resolve discrepancies between the markings of each
paper between the reviewers. The agreed QS for each
Table 1
PEDro scale (modified from Maher et al., 2003)
Criteria Score
1. Eligibility criteria were specified No (0)
2. Subjects randomly allocated to groups No (0)
3. Allocation was concealed No (0)
4. Groups similar at baseline regarding No (0)
the most important prognostic factors
5. Blinding of all subjects No (0)
6. Blinding of all therapists who No (0)
administered therapy
7. Blinding of all assessors who No (0)
measured at least one outcome
8. Measures of at least one key outcome No (0)
were obtained from more than 85% of
initially allocated subjects
9. All subjects for whom outcome No (0)
measures were available received
treatment or control as allocated, or if
this was not the case, at least one
outcome measure analysed using
‘intention to treat’ analysis
10. The results of between-group No (0)
statistical comparisons are reported for
at least one key outcome
11. The study provides both point No (0)
measures and measures or variability for
at least one key outcome
Total
The PEDro Scale is an 11-item scale. The various
items deal with differing aspects of RCT analysis
including internal validity, external validity and
statistics. In order to allow quantitative analysis of
the overall methodological quality of each study,
seven items which relate to internal validity were
identified. These seven items include the following items
numbers 2, 3, 5, 6, 7, 8, 9 (refer to Table 1). The positive
scores of each of these seven items is added together to
calculate an Internal Validity Score (IVS) (Reid and
Rivett, 2005).
2.4. Analysis of methodological quality
Based on the IVS of each paper, it is possible to make
a qualitative assessment about the methodological
quality. In the instance whereby the RCTs reviewed
are not clinically heterogeneous, it is appropriate to use
a qualitative method of analysis as quantitative analysis
Yes (1) is made difficult in that the RCTs may not be directly
Yes (1) comparing like interventions (Reid and Rivett, 2005;
Yes (1) van Tulder et al., 1997).
Yes (1)
The qualitative assessment used within this review is
Yes (1) an adaptation of those used by several authors
Yes (1) (Karjalainen et al., 2001; Reid and Rivett, 2005)
modified specifically for IVS obtained in this review
Yes (1) using the PEDro Scale:
Level 1: Strong evidence—when provided by generally
Yes (1)
consistent findings in multiple RCTs of high quality
(IVS ¼ 6–7).
Yes (1) Level 2: Moderate evidence—when provided by
generally consistent findings in one RCT of high quality
(i.e. IVS ¼ 6–7) and one or more lower-quality RCTs
(i.e. IVSp5);
Level 3: Limited evidence—when provided by gen-
Yes (1) erally consistent findings in one RCT of moderate
quality (i.e. IVS ¼ 4–5) and one or more low-quality
RCTs (i.e. IVSp3).
Yes (1)
Level 4: Insufficient evidence—when provided by
generally consistent findings of one or more RCTs of
x/11
limited quality (i.e. IVSp3), no RCTs available or
conflicting results.

Table 2
Randomized controlled trials of the conservative treatment of ITBFS in order of PEDro score

Scores for PEDro criteria QS Methodological quality IVS

1 2 3 4 5 6 7 8 9 10 11

Gunter and Schwellnus (2004) Corticosteroid injection 1 1 1 1 1 0 0 1 0 1 1 9 Moderate 4

Schwellnus et al. (1991) NSAID’s 1 1 0 1 1 1 0 1 0 1 1 8 Moderate 4
Schwellnus et al. (1992) Deep transverse friction massage 1 1 1 0 0 0 1 0 1 1 1 7 Moderate 4
Bischoff et al. (1995) Phonophoresis versus immobilization 1 1 1 1 0 0 0 0 1 1 1 7 Limited 3

Note: QS ¼ overall quality score; IVS ¼ internal validity score.

Table 3
Randomized controlled trials of the conservative treatment of ITBFS

Author Patient demographics Intervention Control Outcome Results IVS QS

R. Ellis et al. / Manual Therapy 12 (2007) 200–208

Gunter and  n ¼ 18 9 subjects with ITBFS 9 subjects with ITBFS Baseline test: a previous Decrease (P ¼ 0.01) in running 4 9
Schwellnus (2004)  Aged between 20 and validated treadmill- pain from day 7 to 14 in a
Corticosteroid 50 running test—pain whilst validated treadmill running test

ARTICLE IN PRESS
injection  Control group mean running (VAS)—every and VAS, intervention group
age 28.975, minute compared to controls
experimental group 40 mg methylprednisilone, 20 mg 1% lignocaine Running speed at subjects
mean age 29.076.5 10 ml 1% lignocaine hydrochloride injection previous best speed,
 Duration of hydrochloride, injection deep to the ITB at knee maintained for 30 min
symptoms—within 14 deep to the ITB at knee lateral joint line (after 5 min warm-up) or
days lateral joint line until the pain on VAS
reached 8
‘‘Total Daily Pain’’:
evening pain (VAS) over
14 day period
‘‘Total pain during
running’’ (VAS): repeated
on days 7 and 14 after
injection (Mean taken)
Schwellnus et al.  Over 18 yr incl (a) 14 subjects with 13 subjects with ITBFS As for Gunter and Decrease (Po0.05) in overall pain 4 8
(1991) NSAID’s  Final incl 43 pts ITBFS 50 mg diclofenac, Schwellnus (2004) experienced over trial period.
 Age range not given— physiotherapy Decrease (Po0.05) in running
group 1 mean 2275, (b) 16 subjects with Physiotherapy (ITB For total daily pain pain from day 0 to 3 in a validated
group 2 mean age ITBFS 400 mg ibuprofen, stretch, ultrasound, deep (VAS)—mean pain treadmill running test for (b)
2476, group 3 mean 500 mg paracetamol, friction massage) only measured over 14 days intervention group. Increase
age 2272 20 mg codeine phosphate, (Po0.05) in running distance in all
 Duration of physiotherapy (ITB groups from days 3 to 7, and in (b)
symptoms—group 1 stretch, ultrasound, deep intervention group from days
mean 6.877.1 weeks, friction massage) 0 to 7
group 2 mean 6.178.1
weeks, group 3 mean
7.4713.1 weeks

203
Table 3 (continued )

204
Author Patient demographics Intervention Control Outcome Results IVS QS

Schwellnus et al.  n ¼ 20 over 18 yr 9 subjects with ITBFS 8 subjects with ITBFS Total running pain 4 7
(1992) Deep  Group 1 mean 2576, (VAS)—treadmill test
transverse friction group 2 mean 2975, performed on days 0, 3, 7,
massage no range for ages 14 (Mean taken)
given Deep transverse friction Daily stretching and As for Gunter and Decreases in daily pain (VAS) and
 Duration of massage to the distal ITB  2—daily ice therapy Schwellnus (2004) running pain during a validated
symptoms group 1 from days 3 to 14 daily from days 0 to 14. treadmill running test in both
23717 weeks, group 2 stretching and twice-daily Ultrasound to the distal groups. However, no significant
74795 weeks ice therapy from days 0 to ITB days 3–14 difference between groups
14. Ultrasound to the
distal ITB from days 3 to
14
For total daily pain
(VAS)—mean pain

R. Ellis et al. / Manual Therapy 12 (2007) 200–208

measured over 14 days.
Total running pain
(VAS)—treadmill test

ARTICLE IN PRESS
performed on days 0, 3, 7,
14 (Mean taken)
Bischoff et al.  n ¼ 25, Navy students 13 subjects with ITBFS 13 subjects with ITBFS ‘‘Number of days from Less days to become pain free 3 7
(1995)  Group 1 mean 22 yr, Phonophoresis diagnosis to pain-free whilst treadmill running for
Phonophoresis group 2 mean 23 yr (ultrasound through 10% examination’’ intervention group compared to
versus  Duration of hydrocortisone cream) control group. More subjects from
immobilization symptoms—group 1 over 2 weeks intervention group recovered from
mean 17.5 days, group Rest, ice, stretching and Knee immobilization over ‘‘Number of days from ITBFS in 10 days compared to
2 mean 15.0 days ibuprofen 2 weeks diagnosis to symptom- control group
free 1 mile run’’—
symptom-free ¼ ‘‘without
pain or stiffness’’
Rest, ice, stretching and Subjects were examined
ibuprofen daily through the study
 ARTICLE IN PRESS
R. Ellis et al. / Manual Therapy 12 (2007) 200–208
For this review it was decided amongst the reviewers
that in using a seven-item IVS, taken from the initial
PEDro score (QS), a study of high methodological
quality was one with an IVS of between 6 and 7,
moderate quality between 4 and 5, limited quality
between 0 and 3.
2.5. Analysis
When RCTs were clinically and therapeutically
heterogeneous, there is no available method to
directly assess the relative benefit (or lack thereof) of
one intervention versus another. In this instance,
previous systematic reviews have decided to not
include quantitative analysis for this very reason
(Reid and Rivett, 2005). Therefore, it was decided not
to perform any quantitative analysis, as no direct
comparison could be made to determine clinical or
therapeutic benefit between the RCTs and interventions
examined.
3. Results
3.1. Selection of studies
Four RCT’s regarding conservative management of
ITBFS meeting the inclusion criteria were identified
following the electronic and cross-referencing searches.
These studies are summarized in Table 3.
3.2. Methodological quality
The methodological quality, statistically represented
by the IVS, for each paper is detailed within Table 2.
Three of the four RCTs reviewed were given an IVS of
four. This suggests that the authors felt that these
studies were of moderate methodological quality. One of
the RCTs was given an IVS of three, suggesting the
authors felt this study was of limited methodological
quality.
All of the four RCTs satisfied the item relating to
random allocation of subjects (Item 2). Otherwise, there
were no clear trend towards any of the other internal
validity rated items (3, 5, 6, 7, 8, 9) either being
universally satisfied or not.
3.3. Study characteristics
The first important point to note is that all of the four
RCTs assess different therapeutic interventions. There-
fore, they were clinically and therapeutically hetero-
geneous. See Table 3 for detail of the each study’s
characteristics.
205
3.4. Efficacy
Because four different therapeutic interventions were
used, it is difficult to make direct comparison of
therapeutic benefit using quantitative analysis. How-
ever, qualitative analysis is possible when assessing the
methodological quality of the RCTs examining con-
servative treatment of ITBFS. Of the four RCTs
identified, three had IVS’s of 4 (refer to Table 2). Using
the qualitative rating system, as mentioned earlier, it
appears there is limited evidence (Level 3) to support the
use of conservative interventions in the treatment of
ITBFS. Some discussion of the key features of these
studies is reported by intervention as follows.
3.4.1. Non-steroidal anti-inflammatory drugs (NSAIDs)
Schwellnus et al. (1991) conducted a RCT on 43
patients with unilateral ITBFS with pain that was severe
enough to limit running or who had had to stop running
as a consequence of the pain. Subjects were randomly
allocated to three groups. Initial treatment to all subjects
consisted of rest, ice application and medication from
day 0 to 7. From day 3 to 7 all subjects received
standard physiotherapy treatment consisting of ultra-
sound, transverse friction massage (on days 3, 5 and 7)
and daily ITB stretching. The medication was delivered
over the 7 days in a double blind, placebo-controlled
fashion with Group 1 given a placebo anti-inflammatory
medication, Group 2 an anti-inflammatory only (50 mg
diclofenac) and Group 3 a combined anti-inflammatory/
analgesic (400 mg ibuprofen, 500 mg paracetamol, 20 mg
codeine phosphate) medications. Outcome measures
included both daily pain and running pain, each
measured via the visual analogue scale (VAS). Running
pain was measured by a validated treadmill test at 3 and
7 days after treatments commenced.
Results of this study demonstrated that during the
first week of treatment, physiotherapy in conjunction
with combined anti-inflammatory/analgesic medication
was the most effective management. Significant differ-
ences were seen in the combined group with decreased
running pain and increased running time/distance from
0 to 7 days, compared to the other experimental groups.
The combined group was also the only group to show a
significant decrease in running pain at the 3-day test. It
was of interest to note that there was a significant
reduction in daily pain seen across all groups.
3.4.2. Deep transverse friction massage (DTFM)
Schwellnus et al. (1992) commented that the use of
DTFM, in the treatment of ITFBS, is often reported on
the basis of anecdotal evidence that it might be effective.
Schwellnus et al. (1992) also commented that it seems
contradictory that friction techniques may be beneficial
in an injury where the mechanism of the injury is
friction. In order to test these two statements Schwellnus
 ARTICLE IN PRESS
206 R. Ellis et al. / Manual Therapy 12 (2007) 200–208
et al. (1992) conducted an RCT to establish the
therapeutic benefit of DTFM. Twenty subjects with
chronic ITBFS (414 days duration) were randomly
divided into two groups. Both groups received treatment
consisting of rest (apart from treadmill running exercise
tests), ice twice a day and baseline physiotherapy
treatment of daily stretching exercises to the ITB and
5 min of therapeutic low dose ultrasound on days 3, 5, 7
and 10. The intervention group were also given DTFM
for 10 min on the treatment days whereas the control
group received only the general physiotherapy treatment
on the same days. Results of this study found that daily
pain and treadmill running pain levels were both
significantly reduced (P ¼ 0.0005) in both groups with
the authors concluding that the addition of deep friction
massage did not alter the therapeutic outcome of the
condition.
3.4.3. Phonophoresis versus immobilization
Bischoff et al. (1995) conducted an RCT comparing
phonophoresis (using 10% hydrocortisone cream as the
active drug) and knee immobilization, over a 2-week
period in a group of navy diving students who had
developed ITBFS as a result of rigorous physical
training involving a significant amount of running. All
subjects were of similar age (22–23 years) and had
symptoms for 15–17 days prior to entering the trial. The
subjects were randomly assigned to either the knee
immobilization group (three panel knee immobilizer) or
the phonophoresis group. All subjects received ice
massage and non-steroidal anti-inflammatory medica-
tion. Outcome measures in this study were the number
of days required until pain free on examination and the
ability to run on a treadmill at 6.5 miles per hour.
Results of this study concluded subjects in the phono-
phoresis group recovered from the injury in fewer than
10 days and had significantly less pain during the
treadmill running test than the immobilization group.
3.4.4. Corticosteroid injection
Gunter and Schwellnus (2004) conducted an RCT
looking at 18 runners with an acute onset of ITBFS
(o14 days duration). Subjects were randomly allocated
into two groups: Group A receiving an injection of
corticosteroid (40 mg methylprednisilone and 10 mg 1%
lignocaine hydrochloride) deep to the distal ITB, and
Group B receiving a placebo injection (20 mg 1%
lignocaine hydrochloride). Subjects were instructed not
to run for 14 days following the injection and to apply
ice to the area twice daily at 12 h intervals for 30 min. No
physiotherapy treatment was provided to subjects in this
study. Outcome measures were pain measured with a
VAS and an ability to perform a treadmill running test
for 30 min at the subjects best recent 10 km running
speed on days 7 and 14 following the injection.
Although there was a clinical improvement in both
groups, a significant (P ¼ 0.01) decrease (30%) in
running pain (measured with a VAS following a
treadmill test) was observed in the cortisone injection
group compared to control group.
4. Discussion
4.1. The conservative management of ITBFS
The results of this review identified only four RCTs
regarding the conservative management of ITBFS.
These RCTs investigated four different types of treat-
ments including NSAIDs deep friction massage, pho-
nophoresis versus immobilization, and corticosteroid
injection. Some discussion of the key features of these
studies is pertinent. Following the qualitative statistical
analysis, the authors of this review concluded that there
is limited evidence to suggest that the conservative
treatments analysed here are beneficial in the treatment
of ITBFS.
From this review, it is evident that in the majority of
studies a course of physiotherapy treatment was used as
baseline, which involved a combination of ice, ultra-
sound, deep friction massage and stretching. Indeed, it is
not uncommon to find reference to the conservative
treatments, within the literature pertaining to treatment
of ITBFS. In light of the analysis contained within this
systematic review, it seems ironic that many of these
interventions are commonly used within clinical practice
and their use appears to be based on no firm evidence-
and research-based rationale.
4.2. Methodological quality
Three of the four RCTs reviewed (Schwellnus et al.,
1991, 1992; Gunter and Schwellnus, 2004) were given an
IVS of four suggesting the authors felt that these studies
were of moderate methodological quality. Analysis of
these studies, indicate there appears to be some benefit
from using NSAIDs/analgesics and corticosteroid injec-
tions and no benefit from using DTFM.
The fourth RCT (Bischoff et al., 1995) examined
phonophoresis versus immobilization. This study con-
cluded that phonophoresis was more beneficial com-
pared to immobilization. However, there was no
blinding evident throughout this RCT and the present
authors deemed that this study was of limited metho-
dological quality. It is very difficult to therefore deem
this study worthy of consideration when making
educated judgement as to the true effectiveness of these
interventions in the management of ITBFS.
Of most interest was the lack of attention of all the
studies to the various aspects of blinding. For example,
only two studies (Schwellnus et al., 1991; Gunter and
Schwellnus, 2004) satisfied the items relating to subject
 ARTICLE IN PRESS
R. Ellis et al. / Manual Therapy 12 (2007) 200–208
blinding (Item 5). The other two RCTs either did not
adequately blind the subjects or did not mention this.
Only one of the four RCTs satisfied the respective items
relating to therapist blinding/Item 6 (Schwellnus et al.,
1991) and assessor blinding/Item 7 (Schwellnus et al.,
1992). The study by Bischoff et al. (1995) did not satisfy
any of the items relating to blinding, either because there
was no blinding or that blinding was not mentioned.
4.3. Future research
Following the extensive literature search, carried out
for this review, there is an obvious paucity of research
concerning the conservative management of ITBFS. Not
only is there a lack in quantity of such research, upon
dissection of the scarce research that is available, there
seems to also be a paucity of quality.
It now seems apparent that for any of the many
varieties of conservative therapies, for treatment of
ITBFS, that there is no research base available to
conclude any clear benefit from the clinical use of any of
the conservative therapies mentioned. If this is indeed the
case, then future research must attempt to fill this void.
From the RCTs that were available, it seems that the
methodological quality of all these studies was well
below a level that allowed any credible conclusions or
answers to be sought. Additionally, common to all these
studies was a lack of systematic blinding. It would be
advisable for future research to acknowledge this
problem and attempt to organize more robust metho-
dology in order to answer the important research
questions asked.
Not only were the interventions heterogeneous
through the four RCTs reviewed, so to were a number
of other key features including outcome measures and
duration of subjects symptoms. With regard to duration
of symptoms, some papers looked at the more acute
stages of ITBFS (i.e. Bischoff et al. (1995) and Gunter
and Schwellnus (2004) within 2 weeks) compared to
more chronic duration (i.e. Schwellnus et al. (1992) at
approximately 2 months or greater). It would be
pertinent for future research to acknowledge clearly
the duration of symptoms (i.e. acute versus chronic) as it
is likely that some conservative treatments may have
relatively greater or lesser impact at different patholo-
gical stages throughout the course of ITBFS presenta-
tion. For example, the studies looking at corticosteroid
and NSAID use (Schwellnus et al., 1991; Gunter and
Schwellnus, 2004) may have more application in an
early phase of ITBFS where acute inflammation may be
more of a clinical problem and needing to be addressed.
Further to this point, for more chronic presentations of
ITBFS, it may be more appropriate to guide research to
look at more rehabilitation management, such as ITB
stretching, pelvic and knee muscle stabilization, DTFM,
orthotics prescription, etc.
207
With respect to outcome measures, it is not only
important to gain some homogeneity in intervention
selection but also consistency in outcome measures
selected if there is going to be quantitative analysis of
therapeutic benefit of conservative treatments for
ITBFS. As Reid and Rivett (2005) have stated, direct
quantitative comparison, within the realms of systematic
review, is very difficult when interventions, and also
outcome measurements for that matter, are heteroge-
neous. Throughout three of the four RCTs reviewed
(Schwellnus et al., 1991, 1992; Gunter and Schwellnus,
2004) the same previously validated treadmill running
test was used to score running pain. This outcome
measure seems to be appropriate for ITBFS and is also
becoming more widely used. Perhaps a validated test
like this could become a standard test in ITBFS
research.
From a biomechanical and pathological perspective,
the knowledge base regarding ITBFS seems to be
healthy. The clinical application of such theories is both
possible and plausible. There now needs to be research
of sufficient quality and quantity to enable these theories
to be challenged and either accepted or discarded.
5. Conclusion
ITBFS is a common repetitive strain injury of the
lateral aspect of the knee. The pathomechanics and
clinical presentation are well understood. However,
trying to determine the most appropriate choice of
conservative therapy has been made difficult by paucity
in quality and quantity of RCTs to examine therapeutic
benefit.
The aetiology of ITBFS is multifactorial, with a
combination of intrinsic and extrinsic factors. The
causes of ITBFS are in response to the complex
functional anatomy of the ITB and its action as an
independent structure and indirectly through the mus-
cles that it provides attachment to.
Reviewing the efficacy of the conservative manage-
ment of ITBFS has highlighted that there are a small
number of RCTs investigating the effects of therapeutic
interventions on ITBFS. Within the acute stage of the
presenting symptoms (less than 14 days duration)
corticosteroid injection alone appears to be beneficial
with subjects able to return to running pain-free with 14
days of the intervention. In the more chronic presenta-
tions (greater than 14 days duration), there appears to
be benefit gained from using both combined anti-
inflammatory/analgesic medication over anti-inflamma-
tories alone. The inclusion of DTFM to a standard
physiotherapy programme of ultrasound and stretching
exercises, does not appear to produce any additional
benefit. In all of the reviewed trials this generalized
physiotherapy programme proved to be beneficial in
 ARTICLE IN PRESS
208 R. Ellis et al. / Manual Therapy 12 (2007) 200–208
reducing both daily pain and pain experienced on
treadmill running. Although this provides positive
confirmation regarding the benefits of conservative
treatment for ITBFS, it is unfortunate that there are
no RCTs examining the benefit of these different
modalities specifically or in isolation. When investigat-
ing the novel delivery of anti-inflammatory medication
via phonophoresis, a significant reduction in pain was
accomplished when compared to immobilization.
The evidence for the use of conservative treatment in
the management of ITBFS appears to be limited and of
insufficient quality. The research that is available is
heterogeneous and inconsistent. Further examination of
the clinical effect of conservative therapies, in an ITBFS
population, will be of great importance to evidence-
based management of this condition and must direct
future research.
References
Almeida S, Williams K, Shaffer R, Brodine S. Epidemiological
patterns of musculoskeletal injuries and physical training. Medicine
and Science in Sports and Exercise 1999;31(8):1176–82.
Anderson G. Iliotibial band friction syndrome. The Australian Journal
of Science and Medicine in Sport 1991;23(3):81–3.
Aronen J, Cronister R, Regan K, Hensien M. Practical, conservative
management of iliotibial band syndrome. The Physician and Sports
Medicine 1993;21(6):59–69.
Austermuehle P. Common knee injuries in primary care. Nurse
Practitioner 2001;26(10):26–40.
Barber F, Sutker A. Iliotibial band syndrome. Sports Medicine
1992;14(2):144–8.
Bischoff C, Prusaczyk W, Sopchick T, Pratt N, Goforth H.
Comparison of phonophoresis and knee immobilization in treating
iliotibial band syndrome. Sports Medicine, Training and Rehabi-
litation 1995;6:1–6.
Brosseau L, Casimiro L, Milne S, Robinson V, Shea B, Tugwell P, et
al. Deep transverse friction massage for treating tendinitis
(Cochrane Review). The Cochrane Library 2004(Issue 2).
Clark HD, Wells GA, Huet C, McAlister FA, Salmi LR, Fergusson D,
et al. Assessing the quality of randomized trials: reliability of the
Jadad scale. Controlled Clinical Trials 1999;20:448–52.
Ekman E, Pope T, Martin D, Curl W. Magnetic resonance imaging of
iliotibial band syndrome. The American Journal of Sports
Medicine 1994;22(6):851–4.
Farrell K, Reisinger K, Tillman M. Force and repetition in cycling:
possible implications for iliotibial band friction syndrome. The
Knee 2003;10:103–9.
Fredericson M, Guillet M, DeBenedictus L. Quick solutions for
iliotibial band syndrome. The Physician and Sports Medicine
2000;28(2):52.
Fredericson M, White JJ, MacMahon JM, Andriacchi TP. Quantita-
tive analysis of the relative effectiveness of 3 iliotibial band
stretches. Archives of Physical Medicine and Rehabilitation
2002;83:589–92.
Gunter P, Schwellnus M. Local corticosteroid injection in iliotibial
band friction syndrome in runners: a randomised controlled trial.
British Journal of Sports Medicine 2004;38:269–72.
Holmes J, Pruitt A. Iliotibial band syndrome in cyclists. The American
Journal of Sports Medicine 1993;21(3):419–24.
Jordaan G, Schwellnus M. The incidence of overuse injuries in military
recruits during basic military training. Military Medicine 1994;
159(6):421–6.
Karjalainen K, Malmivaara A, van Tulder M, Roine R, Jauhiainen M,
Hurri H. Multidisciplinary biopsychosocial rehabilitation for
subacute low back pain in working-age adults: a systematic review
within the framework of the Cochrane Collaboration Back Review
Group. Spine 2001;26(3):262–9.
Kirk K, Kuklo T, Klemme W. Iliotibial band friction syndrome.
Orthopedics 2000;23(11):1209–15.
Levin J. Run down: battling IT band syndrome in long distance
runners. Biomechanics 2003;1:22–5.
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(8):713–21.
McNicol K, Taunton J, Clement D. Iliotibial tract friction syndrome
in athletes. Canadian Journal of Applied Sport Science 1981;
6(2):76–80.
Messier S, Edwards D, Martin D, Lowery R, Cannon D, James M,
et al. Etiology of iliotibial band friction syndrome in distance
runners. Medicine and Science in Sports and Exercise 1995;27(7):
951–60.
Nemeth W, Sanders B. The lateral synovial recess of the knee:
anatomy and role in chronic iliotibial band friction syndrome.
Arthroscopy 1996;12(5):574–80.
Nishimura G, Yamato M, Tamai K, Takahashi J, Uetani M. MR
findings in iliotibial band syndrome. Skeletal Radiology 1997;
26(9):533–7.
Noble H, Hajek M, Porter M. Diagnosis and treatment of iliotibial
band tightness in runners. The Physician and Sports Medicine
1982;10(4):67–74.
Novacheck T. Running injuries: a biomechanical approach. Journal of
Bone and Joint Surgery 1998;80(8):1220–32.
Orava S. Iliotibial tract friction syndrome in athletes-an uncommon
exertion syndrome on the lateral side of the knee. British Journal of
Sports Medicine 1978;12(2):69–73.
Orchard J, Fricker P, Abud A, Mason B. Biomechanics of iliotibial
band friction syndrome in runners. American Journal of Sports
Medicine 1996;24(3):375–9.
Puniello M. Iliotibial band tightness and medial patellar glide
in patients with patellofemoral dysfunction. The Journal
of Orthopaedic and Sports Physical Therapy 1993;17(3):
144–8.
Reid S, Rivett D. Manual therapy treatment of cervicogenic dizziness:
a systematic review. Manual Therapy 2005;10:4–13.
Renne J. The iliotibial band friction syndrome. The Journal of Bone
and Joint Surgery 1975;57–A(8):1110–1.
Schwellnus M, Theunissen L, Noakes T, Reinach S. Anti-inflamma-
tory and combined anti-inflammatory/analgesic medication in the
early management of iliotibial band friction syndrome. South
African Medical Journal 1991;79:602–6.
Schwellnus M, Mackintosh L, Mee J. Deep transverse frictions in the
treatment of iliotibial band friction syndrome in athletes: a clinical
trial. Physiotherapy 1992;78(8):564–8.
Taunton J, Ryan M, Clement D, McKenzie D, Lloyd-Smith D,
Zumbo B. A retrospective case–control analysis of 2002
running injuries. British Journal of Sports Medicine 2002;36(2):
95–101.
van Tulder M, Assendelft W, Koes B, Bouter L. Method guidelines for
systematic reviews in the Cochrane Collaboration Back Review
Group for spinal disorders. Spine 1997;22(20):2323–30.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 209–218

www.elsevier.com/locate/math

Original article

Changes in pelvic floor and diaphragm kinematics and respiratory

patterns in subjects with sacroiliac joint pain following a motor
learning intervention: A case series
Peter B. O’Sullivan, Darren J. Beales
Curtin University of Technology, School of Physiotherapy, GPO Box U1987, Perth, WA 6845, Australia
Received 24 May 2005; received in revised form 16 February 2006; accepted 2 June 2006

Abstract

This study was a case series design. The objectives of the study were to investigate the ability of a motor learning intervention to
change aberrant pelvic floor and diaphragm kinematics and respiratory patterns observed in subjects with sacroiliac joint pain
(SIJP) during the active straight leg raise (ASLR) test.
The ASLR test is a valid and reliable tool to assist in the assessment of load transference through the pelvis. Irregular respiratory
patterns, decreased diaphragmatic excursion and descent of the pelvic floor have been reported in subjects with SIJP during this test.
To date the ability to alter these patterns has not been determined.
Respiratory patterns, kinematics of the diaphragm and pelvic floor during the ASLR test and the ability to consciously elevate the
pelvic floor in conjunction with changes in pain and disability levels were assessed in nine subjects with a clinical diagnosis of SIJP.
Each subject then undertook an individualized motor learning intervention. The initial variables were then reassessed.
Results showed that abnormal kinematics of the diaphragm and pelvic floor during the ASLR improved following intervention.
Respiratory patterns were also influenced in a positive manner. An inability to consciously elevate the pelvic floor pre-treatment was
reversed. These changes were associated with improvement in pain and disability scores.
This study provides preliminary evidence that aberrant motor control strategies in subjects with SIJP during the ASLR can be
enhanced with a motor learning intervention. Positive changes in motor control were associated with improvements in pain and
disability. Randomized controlled research is required to validate these results.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Diaphragm; Low back pain; Motor control; Pelvic floor; Respiration; Sacroiliac joint

1. Introduction established in subjects with clinically diagnosed SIJ pain

The sacroiliac joint (SIJ) and surrounding ligamen-
tous structures are reported to be a source of symptoms
in subjects with a diagnosis of non-specific chronic low
back pain (Young et al., 2003). Recent research has
focused on a test that investigates the ability of a subject
to transfer load between the lower limb and the trunk,
called the active straight leg raise (ASLR) test. The
validity and reliability of this test procedure has been
Corresponding author. Tel.: +61 8 9266 3629;
fax: +61 8 9266 3699.
E-mail address: P.Osullivan@curtin.edu.au (P.B. O’Sullivan).
(SIJP) (Mens et al., 2001, 1999; O’Sullivan et al., 2002a).
This test involves lying supine and raising the leg 5 cm
off the supporting surface. The test is positive when
accompanied by a primary sensation of profound
heaviness of the leg (7pain), which is relieved with
the application of compression across the ilium. This
test is reported to be positive in a sub-group of subjects
with SIJP (Mens et al., 1999; Pool-Goudzwaard et al.,
2005). It has been proposed that the reduction in
the sensation of heaviness with the application of
compression across the ilia reflects enhanced force
closure through the SIJ (Pool-Goudzwaard et al.,
1998; O’Sullivan et al., 2002a).

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.006
 ARTICLE IN PRESS
210 P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218
Recent research has documented motor control
deficits in the presence of SIJP (O’Sullivan et al.,
2002a). O’Sullivan et al. (2002a) reported in a group
of SIJP subjects with a positive ASLR the presence of
aberrant motor control strategies observed during the
ASLR test when compared to pain-free controls. Using
real time ultrasound and spirometry, the authors’
demonstrated decreased diaphragmatic motion, in-
creased descent of the pelvic floor, increased minute
ventilation and respiratory rate, and altered breathing
patterns in the pain subjects during the ASLR. These
aberrant motor control strategies were eliminated with
the addition of manual compression through the ilia
applied during the ASLR.
It was hypothesized that these disruptions might
represent a deficit in local motor control (pelvic floor,
transverse abdominal wall) within the lumbopelvic
region in these subjects. This manifested as the adoption
of splinting or bracing strategies of the abdominal wall
with associated disrupted patterns of respiration during
the ASLR, not observed in the normal subjects
(O’Sullivan et al., 2002a). Furthermore the normal-
ization of these patterns with the application of
compression supported this notion. The adoption of
these splinting strategies appears to represent an under-
lying deficit in the motor control systems ability to
provide adequate local compression, or force closure, to
the SIJs during the ASLR (O’Sullivan et al., 2002a).
This concept is also supported by the report that
abdominal bracing is less effective than preferential
activation of the transverse abdominal wall muscles for
increasing the compression across the SIJs (Richardson
et al., 2002).
To test the validity of this hypothesis we proposed
that the application of a motor learning intervention
directed to the local stabilizing muscles of the pelvis
would result in the normalization of the aberrant motor
control strategies displayed by these subjects, with
associated reductions in pain and disability.
Previous studies have reported motor learning inter-
ventions to be effective in altering specific motor control
deficits in the presence of chronic low back (O’Sullivan
et al., 1997, 1998) and knee pain (Cowan et al., 2002),
but to date no study has investigated these specific
changes with SIJP during the ASLR test.
2. Methods
Nine subjects (8 female and 1 male) with a clinical
diagnosis of SIJP and a positive ASLR test were
recruited for this study. These subjects were recruited
directly from a previous study by O’Sullivan et al.
(2002a) providing a series of clinical case studies. Four
of the 13 subjects from the original study declined to be
involved in the intervention aspect of the study as they
were already under different forms of management. The
inclusion criteria included pain over the SIJ without
proximal referral (Maigne et al., 1996; Young et al.,
2003) present for a duration of at least 3 months and
showing no signs of abating, no impairment of spinal
range of motion, a positive ASLR test (Mens et al.,
1999, 2001) and at least three out of five positive pain
provocation tests which include: (1) posterior shear test
(Ostgaard et al., 1994; Laslett et al., 2005); (2) sacral
torsion test (Laslett et al., 2005); (3) sacral thrust test
(Laslett et al., 2005); (4) distraction test (Laslett et al.,
2005); and (5) tenderness on palpation of the long dorsal
SIJ ligament (Vleeming et al., 2002). Potential subjects
were excluded if they had a specific radiological
diagnosis for their pain disorder, the presence of
radicular pain, neurological deficits or disorders, hip
joint pathology, an inflammatory disorder, a history of a
significant respiratory disorder, were pregnant or less
than 6 months post partum and/or had a body mass
index of greater than 31 kg/m as previously described by
O’Sullivan et al. (2002a, b). The demographic data for
this group is displayed in Table 1.
The methodology used in this paper has been
previously utilized and described in detail (O’Sullivan
et al., 2002a). Respiratory rate, tidal volume, diaphrag-
matic motion and pelvic floor kinematics were measured
in resting supine, during the ASLR and during the
ASLR with the application of manual compression
through the ilia. In addition pelvic floor kinematics were
measured while subjects were instructed to consciously
elevate their pelvic floor muscles as described in detail
elsewhere (Thompson and O’Sullivan, 2003). A Stead–
Wells water-sealed spirometer (60 Hz, serial number:
3657, Collins, USA) was used to record respiratory rate
and tidal volume from which minute ventilation was
calculated. Movement of the diaphragm was recorded
utilizing real-time ultrasound to visualize the leading
edge of the diaphragm (Cohen et al., 1994). The probe
was positioned transversely and angled superiorly below
the right costal margin in the midclavicular line. Pelvic
floor motion was also recorded using real-time ultra-
sound with the probe positioned trans-abdominally and
Table 1
Demographic data of subjects
Age in years 34.9 (11.2)
Gender 8 female/1 male
Duration of symptoms in months 44.1 (41.2)
Weight (kg) 64.8 (6.5)
Height (cm) 164.3 (6.5)
BMI (kg/m2) 24.2 (3.4)
Number of subjects post-partum 4
Number of subjects post-trauma 9
Number of subjects with a subjective 9
complaint of bladder dysfunction
Data expressed as mean (standard deviation).
 ARTICLE IN PRESS
P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218
angled inferiorly and posteriorly (Walz and Bertermann,
1990). This allowed visualization of the bladder. Due to
the anatomical relationship of the bladder to the pelvic
floor (DeLancey, 1994) motion of the inferior bladder
can be interpreted as motion of the pelvic floor. For
both variables inbuilt electronic calipers were used to
record motion. A Toshiba Sonolayer SSA-250A (serial
number: 32926, Shimoishigami, Japan) real-time ultra-
sound unit (3.75 MHz probe) in movement mode was
used for this purpose. The reliability of these methods
have been previously reported (O’Sullivan et al., 2002a;
Thompson and O’Sullivan, 2003; Sherburn et al., 2005;
Thompson et al., 2005).
Subjects then underwent a motor learning interven-
tion tailored to their individual clinical presentation
over a 12-week period. Three treating physiotherapists
were involved in the study. The specifics of this
intervention are discussed below. Following the inter-
vention the dependant variables were reassessed.
Measurements were carried out in a clinical practice
by one of two physiotherapists, one of whom was also
involved in the intervention process. In conjunction
with these physiological measures the Short Form
McGill Pain Questionnaire (Melzack, 1987) including
a visual analogue scale for pain severity and the
Oswestry Disability Index (Fairbank et al., 1980) were
employed to document changes in pain and functional
status following the treatment period. The study was
approved by the Human Research Ethics Committee of
Curtin University of Technology and written informed
consent was obtained from all participants prior to
testing.
Initial data analysis involved visual inspection of the
spirometry traces. Statistical analysis was then per-
formed with SPSS Version 10.0 for Windows. Sono-
graphy and spirometry data were analysed using a 2
group (pre/post-treatment) by 3 condition (resting
supine, ASLR, ASLR with compression) analysis of
variance. Simple contrasts were performed between all
possible pairs of the three conditions. Paired t-tests were
performed on the pain and disability measures as well as
the pelvic floor kinematics during conscious pelvic floor
contraction. A critical alpha value of 0.05 was used to
determine statistical significance.
3. Intervention model
The motor learning intervention model utilized in this
study was adapted from work described elsewhere
(O’Sullivan et al., 1997; Richardson et al., 1999;
O’Sullivan, 2005b). This model is directed by the specific
classification of a group of disorders where deficits in
motor control appear to be a mechanism for increased
strain and resultant ongoing pain (Elvey and O’Sullivan,
2005; O’Sullivan, 2005a). Within this management
211
model the impairments of motor control that are
considered to be linked to the pain disorder are
identified, and correct patterns are trained within a
cognitive as well as a physical framework. The aim of
the intervention was to retrain the local stabilizing
muscles of the pelvis in a functional and automatic
manner while gaining pain control and enhancing
functional capacity. The reported functional impair-
ments for each individual specifically directed their
intervention although all patients reported pain aggra-
vation during sustained sitting and standing.
Components of the intervention paradigm are pre-
sented in Fig. 1. Each subject was initially educated that
they had specific deficits in their local stabilizing muscles
of the pelvis (pelvic floor, transverse fibres of the
abdominal wall) that had resulted in increased strain
across the pain sensitive structures of the pelvis. These
deficits were identified as a potential mechanism for
ongoing pain and disability. Training of each subject
began in supine crook lying with a semi-full bladder
utilizing transverse abdominal real-time ultrasound
imaging of the pelvic floor. This was conducted as a
means of providing visual feedback in order to teach
each subject to achieve an elevating contraction of
the pelvic floor with simultaneous co-contraction
of the lower transverse abdominal wall (transverse
abdominis and the transverse fibres of internal
oblique) without associated breath holding and/or
global bracing of the abdominal wall (Thompson et
al., 2006a). Once the correct pattern of contraction had
been achieved, the holding capacity of the muscles was
trained for up to 30 s at a time. This stage took up to 4
weeks of training.
This new motor pattern was then progressed to
upright sitting, with the pelvis in slight anterior tilt, a
neutral lumbar lordosis and the thorax in a relaxed
neutral position. The exact sitting position was con-
sidered critical to enable pain control and facilitate
automatic activation of the local stabilizing muscles
(O’Sullivan et al., 2002b; O’Sullivan et al., 2006). The
holding capacity in this posture was then trained so that
the posture could be maintained for long periods of
time, such as sitting in a non-supported chair for up to
30 min while watching TV or reading, in order to
improve the endurance of the trunk postural muscles.
Concurrently subjects were instructed in moving from
sitting to standing while maintaining appropriate
lumbopelvic alignment, to enable pain-free transfer of
load during functional movement tasks. Subjects were
then taught to alter their standing posture to align the
thorax over the pelvis with a neutral lumbar lordosis,
and avoid ‘sway’ postures known to inhibit the local
stabilizing muscles of the lumbopelvic region (O’Sulli-
van et al., 2002b). This new posture was then trained in
its holding capacity during single leg standing followed
by walking. Initially subjects were instructed to walk
 ARTICLE IN PRESS
212 P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218

PHYSICAL COMPONENTS OF MOTOR LEARNING

INTERVENTION

Elevating contraction of the pelvic floor with simultaneous co-contraction of the lower
transverse abdominal wall (transverse abdominis and the transverse fibres of internal
oblique) without associated breath holding and/or global bracing of the abdominal wall

Train neutral lordosis in sitting with relaxed thoracic postures

Train load transfer such as sit to stand

Train aligned standing posture (avoiding sway standing) with neutral lumbar lordosis
and relaxed thorax

Integrate postural alignment into single leg stance and walking

Train trunk loading such as bending and lifting as directed by aggravating factors
reported by patient

COGNITIVE COMPONENTS OF INTERVENTION

Enhanced understanding of the pain mechanism

Enhanced body awareness and control

Learning strategies to develop pain control

Enhancement of positive coping strategies and beliefs regarding disorder

Self management of disorder

Enhancement of functional capacity

Independence from passive treatment

Fig. 1. The first part of this figure presents a paradigm for clinical utility of the physical dimension of the intervention model. This is complimented
by list of the cognitive components of the intervention model.

with control until they could hold the motor pattern for
up to 30 min at a time.
Other functional movement tasks were then identified
and retrained based on pain provoking activities
reported by each of the subjects. It is important to note
that each subject reported pain control when they were
able to adopt the new postures while maintaining their
motor pattern.
Each subject was seen weekly over a period of 12
weeks and instructed to carry out a home exercise
programme on a daily basis. Three subjects wore SIJ
belts during the first 3 weeks of the training period until
they had achieved functional activation of their local
stabilizing muscles at which time they reported that they
no longer required the belt. No other co-interventions
were carried out during the study period.

P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218
4. Results
The individual pre-treatment data for respiratory
rate, tidal volume, diaphragmatic motion and pelvic
floor kinematics for these subjects was extracted from
our previous study (O’Sullivan et al., 2002a) and
reprocessed as a new group to provide the pre-
intervention baseline for this case intervention series.
4.1. ASLR tasks
4.1.1. Pelvic floor kinematics
As there was no pelvic floor motion in resting
supine this condition was not included for analyses.
A significant difference was found between pre- and
post-treatment (F ¼ 12.142, P ¼ 0.008) and between the
remaining two conditions (F ¼ 48.700, Po0.001). There
was an interaction between ASLR and ASLR with
manual pelvic compression (F ¼ 12.374, P ¼ 0.008).
The distinguishing feature of this interaction was the
decrease in pelvic floor descent during the ASLR
following the intervention (Fig. 2). No subject had any
descent of the pelvic floor during the ASLR with
compression post-treatment.
4.1.2. Diaphragmatic excursion
A significant difference between pre- and post-
treatment (F ¼ 6.105, P ¼ 0.039) was found for dia-
phragm excursion and between the three conditions
(F ¼ 11.915, P ¼ 0.006). An interaction was distin-
guished between resting supine and the ASLR
(F ¼ 25.928, P ¼ 0.001), and between ASLR and ASLR
with manual pelvic compression (F ¼ 19.837,
Resting Supine
ASLR ASLR With Compression
14
Mean Pelvic Floor Descent (mm)
12
10
8 12
6 10
4
2
0
Pre-Treatment
Fig. 2. Mean (standard error of the mean) measurements for pelvic
floor descent pre- and post-treatment. Note there is no bar for resting
supine as there was no pelvic floor movement during this test
condition. The graph depicts decreased descent of the pelvic floor
during the ASLR post-treatment. Note there is no error bar for pelvic
floor descent post-treatment during the ASLR with compression as all
subjects had no descent during this task.
ARTICLE IN PRESS
213
P ¼ 0.002). The main feature of this interaction was
increased diaphragmatic excursion during the ASLR
post-intervention (Fig. 3).
4.1.3. Respiratory function
Changes in minute ventilation did not reach a
statistically significant difference pre- and post-treat-
ment (F ¼ 4.966, P ¼ 0.056) nor between the conditions
(F ¼ 4.008, P ¼ 0.069). However, a trend towards
reduced minute ventilation during the ASLR post-
intervention was observed (Fig. 4A).
Subanalysis of the components of minute ventilation
was also undertaken. Respiratory rate was reduced
in post-intervention compared to pre-intervention
(F ¼ 8.563, P ¼ 0.019), however a significant difference
was not identified between the three test conditions
(F ¼ 1.267, P ¼ 0.339). No significant interaction was
found between resting supine and the ASLR (F ¼ 2.465,
P ¼ 0.155) or between the ASLR and ASLR with
compression (F ¼ 2.861, P ¼ 0.129). Fig. 4b therefore
denotes a trend towards decrease respiratory rate
during the ASLR after the intervention. No difference
in tidal volume was observed pre- and post-treatment
(F ¼ 1.900, P ¼ 0.205) or between conditions
(F ¼ 0.286, P ¼ 0.760).
The respiratory traces across all subjects were
variable. In spite of a lack of statistically significant
difference, visual inspection of the respiratory traces
highlighted interesting changes between the pre- to post-
intervention period. Three cases are depicted in Fig. 5 as
examples. (Note: pre-treatment traces were previously
reported in O’Sullivan et al. (2002a).) The notable
feature of these traces is the marked improvement of the
respiratory traces during the ASLR following interven-
tion. The increase in respiratory rate and decrease in
Resting Supine ASLR ASLR With Compression
16
Mean Diaphragmatic Excursion (mm)
14
8
6
4
Post -Treatment 2
0
Pre-Treatment Post -Treatment
Fig. 3. Mean (standard error of the mean) measurements for
diaphragmatic excursion pre- and post-treatment, denoting increased
diaphragmatic motion during the ASLR following the intervention
period.
 ARTICLE IN PRESS
214 P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218

Resting Supine ASLR ASLR With Compression conscious pelvic floor contraction with an average
16 magnitude of 6.12 mm (SE ¼ 0.97). This change was
14
significantly different (Po0.001). It was recorded in the
Mean Minute Ventilation (L/min)

treatment notes that all subjects reported improved

12 bladder function following the intervention although
this was not formally measured.
10

8 4.3. Pain and disability scores

6 Significant differences were found between pre- to

4 post-treatment for the Short Form McGill Pain Ques-
tionnaire (Po0.001), the VAS for usual pain
2 (P ¼ 0.001) and the Oswestry Low Back Pain Ques-
tionnaire (P ¼ 0.003), denoting reductions in pain and
0
(a) Pre-Treatment Post -Treatment disability associated with the intervention (Fig. 6). In
addition it was recorded in the treatment notes that all
Resting Supine ASLR ASLR With Compression subjects reported reduced heaviness during the ASLR
20 test following the intervention although this was not
formally measured.
Mean Respiratory Rate (breaths/min)

18
16
14
12
10
8 SIJP can positively change pelvic floor and diaphragm
6 kinematics and patterns of respiration observed during
4 the ASLR. These changes were associated with con-
2
0 this study is a case series and did not have a control
(b) Pre-Treatment
Fig. 4. Mean (standard error of the mean) values for: (a) minute
ventilation, and (b) respiratory rate, during the three test conditions
before and after treatment. Both denote trends for improvement
during the ASLR post-treatment.
tidal volume during the ASLR observed in Fig. 5a
before the intervention match that of the resting supine
condition post-intervention. Similarly the multiple
breath holds displayed during the ASLR pre-interven-
tion in Fig. 5b, denoted by the flat line in the respiratory
trace, are not observed after the intervention. The
erratic pattern seen in Fig. 5c during the ASLR, while
not equivalent to resting supine post-intervention, has
improved.
4.2. Conscious pelvic floor elevation task
Pelvic floor kinematics during conscious contraction
of the pelvic floor before and after intervention was only
available for eight of the nine subjects due to lost data.
Prior to the intervention all subjects exhibited descent of
the bladder with this task. The average magnitude of
this descent was 11.5 mm (SE ¼ 2.09). After the inter-
vention all subjects demonstrated elevation during
5. Discussion
This study provides preliminary evidence that a
specific motor learning intervention for subjects with
current reductions in pain and disability in a group of
chronically disabled pelvic pain subjects. However as
Post -Treatment
group or blinded independent investigators, the findings
should be viewed with caution. Randomized controlled
research is required to validate these results.
While it is well recognized that movement and motor
control impairments commonly co-exist with lumbopel-
vic pain disorders (O’Sullivan et al., 2002a; Hungerford
et al., 2003; Pool-Goudzwaard et al., 2005), the mere
presence of these impairments does not establish cause
and effect. Movement and motor control impairments
are known to occur secondary to the presence of pain
(Hodges and Moseley, 2003; van Dieen et al., 2003) as
well as pathological and psychological processes (Fry-
moyer et al., 1985; Hall and Elvey, 1999; Hodges and
Moseley, 2003; Marras, 2004; O’Sullivan, 2005a).
Attempts to simply ‘normalize’ movement or motor
control impairments in many of these disorders without
consideration for their underlying mechanism may be
inappropriate and ineffective.
There is however growing evidence that some
disorders do exist where movement and motor control
impairments appear to result in abnormal tissue loading
and pain, leaving them amenable to specific physical
therapy intervention (O’Sullivan et al., 1997; Hides et
al., 2001; Cowan et al., 2002; Stuge et al., 2004b).
Furthermore there is evidence that patterns of abnormal
 ARTICLE IN PRESS
P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218 215

Fig. 5. Respiratory patterns of three subjects before and after treatment. Traces for Subjects A and B during the ASLR post-treatment match that of
the resting supine condition. Subject C demonstrates an improved, though not fully resolved, respiratory pattern during the ASLR post-treatment.
Pre-treatment traces previously published in O’Sullivan et al (2002a) (Sup ¼ resting supine, ALSR ¼ active straight leg raise, Comp ¼ ASLR with
manual pelvic compression).

motor behaviour can be altered with specific exercise or
motor learning interventions, leading to improvements
in pain and disability in specific pain populations
(O’Sullivan et al., 1998; Cowan et al., 2002). Clearly a
priority for clinicians is the ability to identify specific
patient groups for whom motor learning interventions
are appropriate and effective.
The subjects in this current study represent a sub-
group with non-specific chronic pelvic pain as they had
no radiological diagnosis specific for their disorder.
Selection was based on specific clinical inclusion criteria.
There is growing evidence to support that this cluster of
signs and symptoms are associated with pain disorders
of the SIJ and its supporting structures (Mens et al.,
2001; Young et al., 2003; Stuge et al., 2004a; Laslett et
al., 2005; Pool-Goudzwaard et al., 2005).
In our previous paper we proposed that the altered
pelvic floor and diaphragm kinematics and patterns of
respiration in subjects with SIJP during an ASLR that
were normalized with manual pelvic compression
(O’Sullivan et al., 2002a) may reflect loss of force
closure within the pelvis, secondary to a deficit in the

216 P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218
Pre-Treatment Post-Treatment
100
90
Pain and Disability Scales
80
70
60
50
40
30
20
10
0 stabilizing strategy was adopted in the subjects following
McGill (x/45) VAS (mm)
Fig. 6. Mean (standard error of the mean) scores pre- and post-
treatment for the Short Form McGill Pain Questionnaire, the Visual
Analogue Scale for usual pain and the Oswestry Low Back Pain
Questionnaire. Significant improvements were found for all three of
these variables.
local muscles such as the pelvic floor and transverse
abdominal wall. Biomechanical studies show that the
pelvic floor and transverse abdominal wall have the
capacity to locally compress or stabilize the SIJs
(Snijders et al., 1993a, b; Richardson et al., 2002; Pool-
Goudzwaard et al., 2004; van Wingerden et al., 2004).
Growing evidence suggests that dysfunction of these
muscles is present in subjects with SIJP (Avery et al.,
2000; O’Sullivan et al., 2002a; Hungerford et al., 2003;
Pool-Goudzwaard et al., 2005). The improvement of the
altered motor control patterns in the current study,
following a motor learning intervention targeting these
local force closure muscles, lends support for this
hypothesis. Further to this the clinical reports of the
reduction of the ‘heaviness’ associated with the ASLR
may be suggestive of an enhanced motor control
strategy for load transfer across the pelvis during the
ASLR.
Recent research has documented that depression of
the pelvic floor is associated with generation of high
levels of intra-abdominal pressure and global activation
of the pelvic floor, abdominal wall and chest wall
muscles (Thompson et al., 2006a). These bracing
strategies have been shown less able to locally stabilize
the SIJs (Richardson et al., 2002) and have been
reported to be associated with reduced muscle activity
of the pelvic floor during pelvic floor muscle contraction
in women with bladder control disorders (Thompson
et al., 2006b). Recent research has also documented a
relationship between pelvic pain and bladder control
disorders with increased pelvic floor muscle activation
(Pool-Goudzwaard et al., 2005). Further research into
the functioning of the pelvic floor muscles in conjunc-
tion with the other muscles of the abdomino-pelvic
cavity is required to further identify patterns of altered
motor control in subgroups with SIJP. In contrast a
lifting contraction of the pelvic floor (as was trained in
ARTICLE IN PRESS
this study), is associated with high levels of activation of
the pelvic floor and transverse abdominal with minimal
activation of the external oblique, rectus abdominis and
chest wall muscles, minimal increase in intra-abdominal
pressure and allows relaxed respiration (Sapsford et al.,
2001; Thompson et al., 2006a). This local stabilizing
strategy has been shown to enhance the stability of the
SIJs (Richardson et al., 2002) and is also considered
important for the control of continence (Bo et al., 1988;
Thompson et al., 2006a). In light of this research, the
findings of the current study support that a more local
Oswestry (%) the training period, compared to a straining pattern
prior to the intervention. This trained strategy closely
reflects a normal motor control pattern associated with
the ASLR under the pelvic compression condition, and
that previously documented in a pain-free population
during ASLR (O’Sullivan et al., 2002a). It was also
interesting to note the subjective reports that bladder
control symptoms reduced in subjects following the
intervention period. This may be suggestive of a positive
change in the motor control strategies associated with
the control of intra-abdominal pressure and activation
of the pelvic floor muscles associated with the control of
continence. Further research is warranted to further
investigate these issues.
Stuge and co-workers have recently reported long-
term benefits from a specific stabilizing exercise pro-
gramme directed to the lumbopelvic region in subjects
with post partum pelvic pain (Stuge et al., 2004a, b).
Interestingly these authors reported that the normal-
ization of the ASLR test was associated with increased
functional mobility and reductions in pain in this group
of subjects. These findings suggest that the change in the
ASLR was predictive of outcome in these subjects. In
contrast Mens et al. (2000) reported that global training
of the trunk muscles did not result in reduction of pain
and disability in subjects with pelvic pain.
It should be noted that improvement of the motor
patterns associated with the ASLR in this study did
not fully resolve the pain disorder, but rather was
associated with reductions in pain and disability.
Furthermore the intervention had both a functional
and cognitive component to it, with subjects being
taught to utilize their local stabilizing muscles so as to
enhance their functional capacity with pain control.
These findings may support that other physical, neuro-
physiological and cognitive factors may also be asso-
ciated with these pain disorders. Such factors may
include underlying disruption to the pelvic ligaments
resulting in ongoing compromise to the form closure
mechanism of the pelvis, central nervous system
adaptation resulting in ongoing tissue sensitization due
to a chronic pain state and cognitive factors such as
anxiety, fear avoidance behaviour and poor coping
strategies.
 ARTICLE IN PRESS
P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218
Although it cannot be ruled out that the improve-
ments observed were as a result of a change in the
subjects’ natural history, this is unlikely given that the
subjects had reported chronic disabling pelvic pain for
an average of 44 months with no sign of abating prior to
the intervention. Furthermore there is a possibility that
the changes observed were due to a placebo effect
secondary to the expectation of improvement with
treatment. Certainly further research is indicated to
repeat this study using a randomized-controlled clinical
trial design with long-term follow-up and greater subject
numbers. Furthermore, repeating this study with con-
current measures of IAP and muscle activity would also
provide further insight into the exact motor control
strategies that were utilized in these subjects before and
after the intervention period.
In conclusion this study provides preliminary case-
based evidence that altered kinematics of the pelvic floor
and diaphragm, as well as disrupted respiratory
patterns, observed in subjects with SIJP can be shifted
towards those patterns observed in pain-free individuals
with a motor learning intervention. Furthermore the
improvement of these motor patterns was associated
with functional improvements and decreased symptoms.
These findings may provide some insight into the
relationship between motor control strategies and the
ASLR test in subjects with SIJP, however further
research is required to validate this.
Acknowledgments
We would like to acknowledge the statistical support
of Marie Blackmore and the assistance of Julie Beetham,
Anita Avery, Ivan Lin, Beatrice Tucker, Jillian Crisp,
Felicitas Graf and Chris Perkin.
References
Avery AF, O’Sullivan PB, McCallum M. Evidence of pelvic floor
muscle dysfunction in subjects with chronic sacro-iliac joint pain
syndrome. In: Proceedings of the seventh scientific conference of
the international federation of orthopaedic manipulative therapists,
Perth, WA, Australia, 2000. p. 35–8.
Bo K, Larsen S, Oseid S, Kvarstein B, Hagen RR, Jorgensen
J. Knowledge about and ability to correct pelvic floor muscle
exercises in women with urinary stress incontinence. Neurourology
and Urodynamics 1988;7:261–2.
Cohen E, Mier A, Heywood P, Murphy K, Boultbee J, Guz
A. Excursion–volume relation of the right hemidiaphragm
measured by ultrasonography and respiratory airflow measure-
ments. Thorax 1994;49(9):885–9.
Cowan SM, Bennell KL, Crossley KM, Hodges PW, McConnell
J. Physical therapy alters recruitment of the vasti in patellofemoral
pain syndrome. Medicine and Science in Sports and Exercise
2002;34(12):1879–85.
DeLancey J. Functional anatomy of the pelvic floor and urinary
continence mechanism. In: Schussler B, Laycock J, Norton P,
217
Stanton S, editors. Pelvic floor re-education. Principles and
practice. London: Springer; 1994. p. 9–21 [chapter 1.1].
Elvey R, O’Sullivan PB. A contemporary approach to manual therapy.
In: Jull GA, Boyling JD, editors. Grieve’s modern manual therapy.
3rd ed. Edinburgh: Churchill Livingstone; 2005. p. 471–93 [chapter
33].
Fairbank JC, Couper J, Davies JB, O’Brien JP. The Oswestry low back
pain disability questionnaire. Physiotherapy 1980;66(8):271–3.
Frymoyer JW, Rosen JC, Clements J, Pope MH. Psychologic factors
in low-back-pain disability. Clinical Orthopaedics and Related
Research 1985(195):178–84.
Hall TM, Elvey RL. Nerve trunk pain: physical diagnosis and
treatment. Manual Therapy 1999;4(2):63–73.
Hides JA, Jull GA, Richardson CA. Long-term effects of specific
stabilizing exercises for first-episode low back pain. Spine
2001;26(11):E243–8.
Hodges PW, Moseley GL. Pain and motor control of the lumbopelvic
region: effect and possible mechanisms. Journal of Electromyo-
graphy and Kinesiology 2003;13(4):361–70.
Hungerford B, Gilleard W, Hodges P. Evidence of altered lumbopelvic
muscle recruitment in the presence of sacroiliac joint pain. Spine
2003;28(14):1593–600.
Laslett M, Aprill CN, McDonald B, Young SB. Diagnosis of sacroiliac
joint pain: validity of individual provocation tests and composites
of tests. Manual Therapy 2005;10(3):207–18.
Maigne JY, Aivaliklis A, Pfefer F. Results of sacroiliac joint double
block and value of sacroiliac pain provocation tests in 54 patients
with low back pain. Spine 1996;21(16):1889–92.
Marras WS. State-of-the-art research perspectives on musculoskeletal
disorder causation and control: the need for an intergraded
understanding of risk. Journal of Electromyography and Kinesiol-
ogy 2004;14(1):1–5.
Melzack R. The short-form McGill Pain Questionnaire. Pain
1987;30(2):191–7.
Mens JM, Vleeming A, Snijders CJ, Stam HJ, Ginai AZ. The active
straight leg raising test and mobility of the pelvic joints. European
Spine Journal 1999;8(6):468–74.
Mens JM, Snijders CJ, Stam HJ. Diagonal trunk muscle exercises in
peripartum pelvic pain: a randomized clinical trial. Physical
Therapy 2000;80(12):1164–73.
Mens JM, Vleeming A, Snijders CJ, Koes BW, Stam HJ. Reliability
and validity of the active straight leg raise test in posterior pelvic
pain since pregnancy. Spine 2001;26(10):1167–71.
Ostgaard HC, Zetherstrom G, Roos-Hansson E. The posterior pelvic
pain provocation test in pregnant women. European Spine Journal
1994;3(5):258–60.
O’Sullivan P. Diagnosis and classification of chronic low back pain
disorders: maladaptive movement and motor control impairments
as underlying mechanism. Manual Therapy 2005a;10(4):242–55.
O’Sullivan PB. ‘Clinical Instability’ of the lumbar spine: its patholo-
gical basis, diagnosis and conservative management. In: Jull GA,
Boyling JD, editors. Grieve’s modern manual therapy. 3rd ed.
Edinburgh: Churchill Livingstone; 2005b. p. 311–31 [chapter 22].
O’Sullivan PB, Twomey LT, Allison GT. Evaluation of specific
stabilizing exercise in the treatment of chronic low back pain with
radiologic diagnosis of spondylolysis or spondylolisthesis. Spine
1997;22(24):2959–67.
O’Sullivan PB, Twomey L, Allison GT. Altered abdominal muscle
recruitment in patients with chronic back pain following a specific
exercise intervention. The Journal of Orthopaedic and Sports
Physical Therapy 1998;27(2):114–24.
O’Sullivan PB, Beales DJ, Beetham JA, Cripps J, Graf F, Lin IB, et al.
Altered motor control strategies in subjects with sacroiliac
joint pain during the active straight-leg-raise test. Spine 2002a;
27(1):E1–8.
 ARTICLE IN PRESS
218 P.B. O’Sullivan, D.J. Beales / Manual Therapy 12 (2007) 209–218
O’Sullivan PB, Grahamslaw KM, Kendell M, Lapenskie SC, Moller
NE, Richards KV. The effect of different standing and sitting
postures on trunk muscle activity in a pain-free population. Spine
2002b;27(11):1238–44.
O’Sullivan PB, Dankaerts W, Burnett A, Booth R, Carlsen C, Chen D,
Schultz A. Evaluation of the flexion relaxation phenomenon in the
trunk muscles in sitting. Spine 2006;31(17):2009–16.
Pool-Goudzwaard AL, Vleeming A, Stoeckart R, Snijders CJ, Mens
JMA. Insufficient lumbopelvic stability: a clinical, anatomical and
biomechanical approach to ‘‘a-specific’’ low back pain. Manual
Therapy 1998;3(1):12–20.
Pool-Goudzwaard A, Hoek Van Dijke G, Van Gurp M, Mulder P,
Snijders C, Stoeckart R. Contribution of pelvic floor muscles
to stiffness of the pelvic ring. Clinical Biomechanics 2004;
19(6):564–71.
Pool-Goudzwaard AL, Slieker Ten Hove MC, Vierhout ME, Mulder
PH, Pool JJ, Snijders CJ, et al. Relations between pregnancy-
related low back pain, pelvic floor activity and pelvic floor
dysfunction. International Urogynecology Journal and Pelvic
Floor Dysfunction 2005;16(6):468–74.
Richardson CA, Jull GA, Hodges PW, Hides JA. Therapeutic exercise
for spinal segmental stabilization in low back pain. Scientific basis
and clinical approach. Sydney: Churchill Livingstone; 1999.
Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J.
The relation between the transversus abdominis muscles, sacroiliac
joint mechanics, and low back pain. Spine 2002;27(4):399–405.
Sapsford RR, Hodges PW, Richardson CA, Cooper DH, Markwell SJ,
Jull GA. Co-activation of the abdominal and pelvic floor muscles
during voluntary exercises. Neurourology and Urodynamics
2001;20(1):31–42.
Sherburn M, Murphy CA, Carroll S, Allen TJ, Galea MP. Investiga-
tion of transabdominal real-time ultrasound to visualise the
muscles of the pelvic floor. Australian Journal of Physiotherapy
2005;51(3):167–70.
Snijders C, Vleeming A, Stoeckart R. Transfer of lumbosacral load to
iliac bones and legs. Part 1: biomechanics of self-bracing of the
sacroiliac joints and its significance for treatment and exercise.
Clinical Biomechanics 1993a;8(6):285–94.
Snijders C, Vleeming A, Stoeckart R. Transfer of lumbosacral load to
iliac bones and legs. Part 2: loading of the sacroiliac joints when
lifting in a stooped posture. Clinical Biomechanics 1993b;8(6):295–
301.
Stuge B, Laerum E, Kirkesola G, Vollestad N. The efficacy of a
treatment program focusing on specific stabilizing exercises for
pelvic girdle pain after pregnancy: a randomized controlled trial.
Spine 2004a;29(4):351–9.
Stuge B, Veierod MB, Laerum E, Vollestad N. The efficacy of a
treatment program focusing on specific stabilizing exercises for
pelvic girdle pain after pregnancy: a two-year follow-up of a
randomized clinical trial. Spine 2004b;29(10):E197–203.
Thompson JA, O’Sullivan PB. Levator plate movement during
voluntary pelvic floor muscle contraction in subjects with incon-
tinence and prolapse: a cross-sectional study and review. Interna-
tional Urogynecology Journal and Pelvic Floor Dysfunction
2003;14(2):84–8.
Thompson JA, O’Sullivan PB, Briffa NK, Neumann P. Differences in
muscle activation patterns during pelvic floor muscle contraction
and Valsalva manouevre. Neurourology and Urodynamics
2006a;25(2):148–55.
Thompson JA, O’Sullivan PB, Briffa K, Neumann P. Altered muscle
activation patterns in symptomatic women during pelvic floor
muscle contraction and Valsalva manouevre. Neurourology and
Urodynamics 2006b;25(3):268–76.
Thompson JA, O’Sullivan PB, Briffa K, Neumann P, Court
S. Assessment of pelvic floor movement using transabdominal
and transperineal ultrasound. International Urogynecology Jour-
nal and Pelvic Floor Dysfunction 2005;16(4):285–92.
van Dieen JH, Cholewicki J, Radebold A. Trunk muscle recruitment
patterns in patients with low back pain enhance the stability of the
lumbar spine. Spine 2003;28(8):834–41.
van Wingerden JP, Vleeming A, Buyruk HM, Raissadat K. Stabiliza-
tion of the sacroiliac joint in vivo: verification of muscular
contribution to force closure of the pelvis. European Spine Journal
2004;13(3):199–205.
Vleeming A, de Vries HJ, Mens JM, van Wingerden JP. Possible role
of the long dorsal sacroiliac ligament in women with peripartum
pelvic pain. Acta Obstetricia et Gynecologica Scandinavica
2002;81(5):430–6.
Walz P, Bertermann H. Ultrasound examination of bladder and
prostate. Urologia Internationalis 1990;45(4):217–30.
Young S, Aprill C, Laslett M. Correlation of clinical examination
characteristics with three sources of chronic low back pain. The
Spine Journal 2003;3(6):460–5.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 219–225

www.elsevier.com/locate/math

Original article

Activation of vastus medialis oblique is not delayed in patients with

osteoarthritis of the knee compared to asymptomatic participants
during open kinetic chain activities
John Dixona,, Tracey E. Howeb
a
Teesside Centre for Rehabilitation Sciences, University of Teesside, The James Cook University Hospital, Marton Road, Middlesbrough, UK
b
HealthQWest, Glasgow Caledonian University, Cowcaddens Road, Glasgow, UK

Received 7 July 2005; received in revised form 23 February 2006; accepted 1 June 2006

Abstract

This study investigated whether the onset of electromyographic (EMG) activity of vastus medialis oblique (VMO) was delayed
relative to that of vastus lateralis (VL) in patients with osteoarthritis (OA) of the knee compared to asymptomatic participants
during open kinetic chain activities. An exploratory observational cross sectional study was carried out. Two groups were tested,
symptomatic OA knee patients, diagnosed by an orthopaedic surgeon, ðn ¼ 17Þ, mean (SD) age 66.0 (7.6) years, and asymptomatic
participants ðn ¼ 17Þ, 56.7 (8.6) years. Surface EMG activity of VMO and VL was measured, during concentric contractions
extending the knee from 901 flexion, and during maximal voluntary isometric contractions at 601 knee flexion. The EMG onset times
of VMO and VL were determined visually and by algorithm. The onset timing difference (OTD) between the two muscles was
calculated for each subject, by subtracting the onset time of VL from VMO. Mann–Whitney U-tests revealed that the OTD between
VMO and VL was not significantly different between the groups during either contraction type (both p40.05). The results of this
exploratory study may have implications for rehabilitation programmes aimed at developing preferential activation of VMO
compared to VL in OA knee patients.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Electromyography; Knee; Quadriceps; Osteoarthritis

1. Introduction is evidence to suggest that VMO may be preferentially

Osteoarthritis (OA) is a common cause of disability in
adults (Badley and Tennant, 1993). Although previously
thought of being solely due to joint wear and tear, it is now
believed that muscle dysfunction is an important factor in
OA knee (Shrier, 2004). Impairments that may increase
joint damage over time, such as arthrogenous muscle
inhibition, quadriceps weakness, and slowed protective
reflexes have been implicated (Hurley, 1999). Clinicians
observe what appears to be selective atrophy of muscle
vastus medialis oblique (VMO) in OA knee patients. There
Corresponding author. Tel.: +44 1642 384125.
E-mail address: John.dixon@tees.ac.uk (J. Dixon).
affected in patellofemoral pain syndrome patients, and its
activation delayed relative to that of vastus lateralis (VL)
when compared to healthy subjects (Witvrouw et al., 1996;
Cowan et al., 2001, 2002). Theoretically this could lead to
a biomechanical imbalance at the knee joint. Despite the
prevalence of OA knee, there is a scarcity of research
about VMO activation in OA knee patients. This has been
studied during stair climbing (Hinman et al., 2002), and
during the patellar tendon reflex response (Dixon et al.,
2004). In both studies the activation timing of VMO
relative to VL did not differ in OA knee patients compared
to asymptomatic participants.
However, it has been proposed that muscle activation
order may be affected by the type of muscle contraction

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.012
 ARTICLE IN PRESS
220 J. Dixon, T.E. Howe / Manual Therapy 12 (2007) 219–225
(Grabiner et al., 1994; Stensdotter et al., 2003), and
voluntary muscle activation during open kinetic chain
(OKC) activities remain to be investigated in this knee
pathology. This type of activity is often used to evaluate
quadriceps function in rehabilitation. Hence the aim of
this study was to investigate onset of VMO and VL
EMG activity in OA knee patients and asymptomatic
participants during two OKC activities, seated non-
weightbearing knee extensions (concentric contractions)
extending the knee from 901 flexion, and during
maximal voluntary isometric contractions (MVIC) at
601 knee flexion. We tested the hypothesis that onset of
VMO EMG activity, relative to that of VL, would be
delayed in OA knee patients when compared to similarly
aged asymptomatic participants. We also measured
torque produced by the quadriceps during MVICs using
a load cell to allow a comparison with previously
published studies (Tan et al., 1995; Cheing and Hui-
Chan, 2001) reporting deficits in maximal quadriceps
strength in patients with OA knee.
2. Methods
An exploratory observational cross sectional study
was carried out, in conjunction with a study into reflex
EMG activity (Dixon et al., 2004).
2.1. Participants
Ethical approval was granted by local research ethics
committees. All subjects gave written and verbal in-
formed consent before taking part in the study. Two
subject groups were tested, a group of symptomatic OA
knee patients, and a group of similarly aged asympto-
matic participants. Asymptomatic participants comprised
a convenience sample from the local area and OA
patients were recruited from South Tees Hospitals NHS
Trust, UK outpatients orthopaedic clinics. Diagnosis of
OA knee was made by an orthopaedic surgeon, according
to the American College of Rheumatology criteria
(Altman et al., 1986), using clinical signs and symptoms
and the presence of osteophytes determined by weight-
bearing radiographs. Asymptomatic participants were
individuals who reported having no history of knee pain.
Participants were excluded if they presented with
significant cognitive, cardiorespiratory, neurological, or
musculoskeletal impairments (excepting OA knee in the
patient group) that limited functional ability, or reported
use of medication affecting neuromuscular function, or if
they could not carry out the activities.
2.2. Procedure
For testing, each participant was seated on an
adjustable padded ‘quadriceps chair’ (Tornvall, 1963;
Edwards et al., 1977; Dixon et al., 2004) with the
posterior aspect of the knee positioned at the front edge
of the seat and arms folded across the chest. The chair
back-rest was set at 701 of hip flexion (Doxey and
Eisenman, 1987). In bilateral OA knee patients and
asymptomatic participants, data were recorded from the
dominant limb, which was defined as the limb with
which they would kick a ball (Cheing and Hui-Chan,
2001). The symptomatic limb of patients with unilateral
OA was tested.
After cleaning the skin with isopropyl alcohol, and
shaving the area if necessary, active surface EMG
recording electrodes (BIOPAC Inc., USA, TSD150B,
Ag/Ag Cl, diameter 11.4 mm, electrode spacing 20 mm
centre to centre, with a built in 350  amplification and
a 3 dB bandpass of 12–500 Hz) were placed on VMO
and VL at standardized sites. The electrodes were
oriented in the estimated direction of the muscle fibres
(Lieb and Perry, 1968). The VL electrode was sited on
the muscle belly, at one-third the distance from the
superior border of the patella to the greater trochanter
(Mannion and Dolan, 1996), oriented 12–151 laterally
from the long axis of the femur. For VMO placement
was on the muscle belly 5 cm from the superior medial
border patella border (Callaghan et al., 2001), oriented
50–551 medially from the long axis of the femur.
Hypoallergenic conductive gel (Lectron II, Pharmaceu-
tical Innovations Inc., USA) was applied to the
electrodes to facilitate electrical contact with the skin
surface. All electrodes were taped to the skin to prevent
movement artifacts. A ground electrode (Red dot TM,
3M Healthcare, USA) was attached to the patella of the
untested leg.
2.2.1. Concentric contraction
An electrogoniometer (Type SG150, Biometrics Ltd,
UK) was used to measure knee angle during the
concentric contraction. This was calibrated prior to
use, and attached to the knee joint of the participant
with doubled sided tape. Each participant was asked to
perform extension of the knee joint, from 901 knee
flexion to full extension (or as far as possible) and back.
The instruction ‘‘straighten your leg’’ was given. This
was followed by ‘‘and relax’’ when the participant
reached full extension. This was carried out ten times,
each contraction separated by 20 s to minimize fatigue
effects.
2.2.2. Maximal voluntary isometric contraction (MVIC)
Force was measured using a load cell (TC601, Amber
Instruments Ltd, UK). The load cell output was fed to a
channel of the BIOPAC system and the voltage output
was pre-calibrated by hanging known weights vertically
from the load cell. The MVIC activity was carried out at
least 3 min after the concentric contraction. The knee
was positioned at 601 flexion (Kannus et al., 1987;
 ARTICLE IN PRESS
J. Dixon, T.E. Howe / Manual Therapy 12 (2007) 219–225
Powers et al., 1996). This position has been shown to
produce a higher torque than other knee angles (Lieb
and Perry, 1971). This leg position was standardized by
the attachment to the chair of an adjustable arm that
supported the leg of the subject at mid-calf level. The
angle of flexion was verified with a goniometer and the
leg support adjusted accordingly. A comfortable non-
extensible ankle strap was attached to the leg of the
subject at the level of the lateral malleolus. The load cell
under the chair was adjusted so that the chain passed
from the leg at an angle of 901. The length of the lever-
arm, being the distance from the centre of the knee joint
to the load cell strap at the lateral malleolus, was
measured to allow torque calculations.
Participants carried out five maximal isometric con-
tractions of the quadriceps with 30 s rest between each
contraction to minimize fatigue. Each participant was
asked to straighten the leg as forcefully as possible
against the resistance of the chain for 3 s (Soderberg and
Knutson, 2000) and then told to relax. The investigator
gave standardized instructions of ‘ready, steady, push,
push, push, relax’ to elicit a contraction of 3 s. It was
confirmed that participants understood the instructions
and they were given practice attempts to ensure
familarization with the procedure. Verbal encourage-
ment was given, as encouragement has been shown to
increase maximal voluntary contraction values (McNair
et al., 1996). The peak force was utilized as the MVIC
for the purposes of the study.
All data were sampled at 2048 Hz using a physiolo-
gical data acquisition system (BIOPAC Inc., USA)
comprising an MP100 workstation with a high-level
transducer HLT100 and dedicated analysis software
(AcqKnowledge 3.5.3).
2.3. Onset timing determination
Onset of EMG activity, the ‘earliest rise beyond the
steady state’ of the raw EMG signal (Hodges and Bui,
1996), was evaluated both visually and by computer
algorithm. Both methods were used because of age-
related changes that affect the quality of the EMG signal
in older people and which could affect onset times
determined by algorithm, e.g. lower amplitude and
poorer signal to noise ratio than younger people. For
visual evaluation, the peak or trough of the first spike of
the raw EMG was identified as the point of onset, by
viewing the digitally stored data in the AcqKnowledge
software.
For computerized evaluation, an algorithm based on
the work of Hodges and Bui (1996) was used to identify
the EMG onset time for each electrode site (Dixon,
2004). Here, EMG data was saved as an ASCII text file
with a resting baseline prior to contraction of 500 ms
duration, and imported into Microsoft Excel. In Excel,
the raw EMG data were full wave rectified and 50 Hz
221
low-pass filtered using a single pole first order digital
filter. The algorithm calculated the mean and standard
deviation (SD) of the first 100 data points (the baseline
window, 48.83 ms). The SD value was then used as the
threshold to determine EMG onset time. The algorithm
identified onset as the time point at which the mean
value of a moving window of 50 data points (24.41 ms)
exceeded the mean of the baseline by 1SD of the baseline
window. The mid-point of the moving window was the
time determined as onset. The algorithm was also tested
with 2 and 3 SDs as the threshold value, but it was
found that the 1 SD threshold had best agreement with
visually determined onset times using Bland and Altman
limits of agreement (Dixon, 2004). The 1SD algorithm
onset times were therefore used, as in previous literature
(Karst and Willett, 1995). On the few occasions when
the baseline window contained a marked artifact, its
position was moved to ensure that artifacts did not
artificially increase the SD and hence the threshold.
The EMG onset time difference (OTD) between the
VMO and VL was then calculated for both visually and
computer evaluated results, by subtraction of the time
for VL from that of VMO. A negative value therefore
indicates that VMO onset is before that of VL. The
median OTD (of 10 concentric and 5 MVICs) was then
calculated for each subject for both activities. This was
used in preference to the mean, as data for some
participants were skewed.
For the concentric contraction, data were excluded
for one asymptomatic participant due to an electrode
problem. Concentric and MVIC EMG data for one OA
knee patient were excluded due to the poor signal
quality possibly arising from obesity (Marks et al.,
1994), which prevented successful onset determination.
However this patient could perform the MVIC and their
torque data were not excluded from the torque results.
2.4. Statistical analysis
Data were analysed using SPSS V 10. The Mann–
Whitney U-test was used to test for between group
differences, as the data sets displayed either differences
in variance or non-normal distributions, as has been
described in other work (Karst and Willett, 1995).
Values of po0.05 were regarded as statistically sig-
nificant.
3. Results
Descriptive characteristics of the groups are shown in
Table 1. The groups were significantly different in age,
mass and body mass index (Mann–Whitney U-test, all
po0.05), but not in height ðp ¼ 1:00Þ, with the OA knee
group being on average 9.3 years older, 9.4 kg heavier
and having greater body mass index by 3.4 kg/m2.
 ARTICLE IN PRESS
222 J. Dixon, T.E. Howe / Manual Therapy 12 (2007) 219–225

Table 1
Description of participants

Group Age (years) Height (m) Mass (kg) BMI (kg/m2) Sex
Mean (SD) Mean (SD) Mean (SD) Mean (SD) (M:F)

Asymptomatic 56.7 (8.6) 1.68 (0.09) 74.4 (10.9) 26.3 (2.7) 8:9
OA knee 66.0 (7.6) 1.67 (0.09) 83.8 (12.9) 29.7 (3.8) 11:6

Group
asymptomatic
OA knee
150 150
Onset timing difference (ms)

100 100

50 50

0 0

-50 -50

-100 -100
N- 16 16 16 16 N- 17 16 17 16
visual algorithm visual algorithm
(a) Onset determination method (b) Onset determination method

Fig. 1. Onset timing difference during initiation of concentric contraction (1a) and MVIC (1b). Negative value indicates VMO activated before VL.

Fig. 1 shows the results for OTD, onset of VMO Table 2

EMG activity relative to that of VL, for the concentric Median (interquartile range) onset timing differences (ms) by contrac-
contraction and for the MVIC, in graphic format. The tion type and evaluation method
median and interquartile range values are also presented Group Concentric MVIC
in Table 2. For both contraction types, the differences
between the groups in the OTD were not statistically Visual Algorithm Visual Algorithm
significant for either visual or computerized evaluation Asymptomatic 2.7 (23.9) 2.5 (38.7) 4.6 (21.2) 13.6 (44.3)
methods (Mann–Whitney U-test, all p40.05). We noted OA knee 7.9 (52.4) 3.6 (87.9) 1.5 (27.5) 5.3 (45.1)
greater variability in the OA knee patients than in the p valuea 0.41 0.50 0.06 0.07
asymptomatic participants in both contraction types, as a
Mann–Whitney U-test.
can be seen clearly in Fig. 1.
The MVIC torque produced by the OA knee patients
was on average 30% less than the asymptomatic
participants (120.6 Nm compared to 84.6 Nm). This OA knee patients compared to similarly aged asympto-
difference was statistically significant (Mann–Whitney matic participants, during maximal voluntary isometric
U-test, po0.05). and concentric muscle contractions of the quadriceps.
We were not able to report confidence intervals for We did this by calculating the OTD by subtracting the
the differences between the groups because the data EMG onset time of VL from VMO. We found that
breached the assumptions of normality and non-para- OTD was not significantly different between the OA
metric tests were used (Bland, 2000). knee patients and the asymptomatic participants in
either activity. This agrees with the small amount of
literature in this area on OA knee patients during other
4. Discussion tasks (Hinman et al., 2002; Dixon et al., 2004). We also
found that MVIC torque of OA knee patients was
Our aim was to investigate whether the onset of EMG significantly less than the similarly aged asymptomatic
activity of VMO was delayed relative to that of VL in group, as has been shown in previous studies (Tan et al.,
 ARTICLE IN PRESS
J. Dixon, T.E. Howe / Manual Therapy 12 (2007) 219–225
1995; Fisher and Pendergast, 1997; Cheing and Hui-
Chan, 2001), and our torque data are comparable with
these reports.
We observed marked variability both within and
between subjects (Fig. 1) in the VMO-VL OTD, which
has been mentioned in some patellofemoral pain
syndrome literature (Karst and Willett, 1995; Cowan
et al., 2001) but not in great detail. We believe that this
may be firstly related to the OKC activities studied here,
which have been shown to elicit greater variability in
onset times than closed chain type activities (Stensdotter
et al., 2003). Secondly, this may also be because of age-
related changes that affect the quality of the EMG signal
in older people. As the EMG signals of older people
often have a lower amplitude and poorer signal to noise
ratio than that of younger people, due to muscle atrophy
and increased subcutaneous fat, onset determination
may be more difficult. Possibly we could have used
greater smoothing than a 50 Hz low pass filter, but
excess smoothing can cause loss of information (Karst,
1998). Very few studies have evaluated EMG in older
people, and it would be preferable if more work were
carried out here. There is no universally accepted best
method for computerized onset determination (Hodges
and Bui, 1996; Staude and Wolf, 1999), but because
these computerized methods have set standards or
parameters, subjectivity is reduced. The algorithm used
here is based on a well-used technique (Hodges and Bui,
1996). Some studies have used variations, such as having
the threshold exceeded for a minimum of 25 ms (Hin-
man et al., 2002) rather than the point where the mean
of the window exceeds threshold, as used here. Also, we
used a first order low pass filter compared to some
studies using higher order filters. However, we believe
that this should produce very little difference in results,
as at 50 Hz filtering the phase lag is approximately 3 ms.
In addition the standard moving window technique that
follows the low pass filtering averaged the data over
24 ms periods, and the VL onset times were subtracted
from those of VMO, to produce a relative OTD.
The lack of a significant difference between the groups
could be attributable to several explanations. Firstly, the
sample size ðn ¼ 17Þ could possibly result in an under-
powered study. We did not carry out post hoc power
calculations as it is known that any analysis finding no
significant difference will be shown to have been
underpowered when post hoc power calculations are
carried out (Goodman and Berlin, 1994), and research
must start somewhere (Bacchetti, 2002). However, we
did carry out an estimation of how many participants
would be required to adequately power a future study to
detect differences between the groups in OTD. These
post hoc calculations were carried out with a significance
level of 0.05 using the larger of the group SDs from the
visually evaluated results, but it should be noted that
because the data breached the assumptions of normality,
223
these calculations were approximations only, and may
be overestimates. Due to the variability observed, it was
decided to use a value of 20 ms as the difference to be
detected between the groups. The calculations showed
that a study with 95 participants in each group would
have 80% power to detect a difference of 20 ms in OTD
for the concentric contraction. However, 142 partici-
pants in each group would be required to provide 80%
power for detection of a 20 ms difference for MVIC.
Further research with a larger sample size is therefore
recommended, and this study provides information that
will assist future studies. To avoid making a type II error
from the results observed, future studies should use a
much larger total sample size, and our estimations show
that approximately 90–140 participants may be required
per group for this type of OKC activity. The small
sample size in this study is a limitation and therefore the
results must be interpreted with caution. However, this
was an exploratory study from which the results have
enabled sample size calculations to be undertaken for
subsequent studies.
It is possible that the OA knee patients in this study
could be a heterogeneous group, as the general OA knee
population is understood to display considerable hetero-
geneity (Hurley, 1999). No attempt was made to
categorize patients in terms of aspects of OA knee
classification, or duration or level of symptoms. If this
was the case in future studies, sub-group analysis could
reveal whether specific sub-groups of the OA knee
population do exhibit an impairment in OTD that could
be clinically important. In addition, an assumption was
made that the asymptomatic participants did not have
radiographic OA. Although radiographs were not taken
from the asymptomatic participants for ethical reasons,
it is possible that they could have exhibited asympto-
matic radiographic changes at the knee joint (Felson et
al., 1987) which may have affected muscle activation.
The OKC type muscle contraction tested here may
not be representative of normal functional movement
that tends to occur in the weight bearing closed kinetic
chain manner (Callaghan and Oldham, 1996; Stensdot-
ter et al., 2003). However, this is a standard method of
evaluating quadriceps function. Future studies should
attempt to use both open and closed kinetic chain
testing. In addition, it was noticed that the MVIC
activity often generated anticipatory muscle contraction
in some subjects. As this anticipatory muscle activity
generated force (measured by the load cell), these EMG
onsets could not be validly excluded. Interestingly, using
the computerized method, onset was often detected for
VMO and not VL, or vice-versa, which could have
added to the variability discussed above.
Nevertheless, we believe that this exploratory study
adds to the evidence base about neuromuscular impair-
ments in OA knee, and will inform future research
and practice. Much of our understanding of muscle
 ARTICLE IN PRESS
224 J. Dixon, T.E. Howe / Manual Therapy 12 (2007) 219–225
dysfunction in OA knee remains speculative (Hurley,
1999). This study and those already published indicate
that the timing of VMO muscle activation relative to
that of VL in OA knee patients does not differ from that
of asymptomatic participants, in contrast to evidence in
PFPS. However there are some published studies
reporting that experimental knee effusion can preferen-
tially affect VMO (Kennedy et al., 1982; Spencer et al.,
1984; Torry et al., 2000). Therefore there is a now a need
for the intensity of VMO and VL EMG activity to be
investigated in OA knee patients, as it is possible that
this is where any preferential inhibitory effect on VMO
may be exhibited. The need for more research into
muscle function in OA knee is clear. Subsequent to
studies reporting muscle inhibition in OA knee patients
(Hurley and Newham, 1993; Hurley et al., 1997), a study
has found that not all OA knee patients exhibit muscle
inhibition but some asymptomatic participants do
(Lewek et al., 2004). Quadriceps strength and muscle
inhibition do not show the expected correlation with
severity of cartilage damage (Pap et al., 2004). Similarly,
the findings that quadriceps strengthening may actually
not benefit patients with malaligned knees (Sharma et
al., 2003) shows that further research is urgently
required. This seems particularly important with regard
to the heterogeneity and variability within the OA knee
population, if treatments are to be efficacious.
5. Conclusion
Our acceptance of the null hypothesis agrees with the
only other studies into the activation timing of VMO
and VL in OA knee (Hinman et al., 2002; Dixon et al.,
2004). This evidence of this exploratory study for the
lack of a preferential delay in VMO activation in OA
knee patients contrasts with findings of a preferential
effect in PFPS. Rehabilitation programmes for OA knee
patients should not therefore be aimed at altering the
timing of VMO activation relative to VL, as a growing
body of evidence seems to indicate that no temporal
delay exists.
Acknowledgements
The authors wish to thank Vicki Whittaker for statistical
advice and assistance, and all the participants who took
part. This work was supported by a University of Teesside
School of Health and Social Care Ph.D. Studentship.
References
Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, Christy
W, Cooke TD, Greenwald R, Hochberg M. Development of
criteria for the classification and reporting of osteoarthritis.
Classification of osteoarthritis of the knee. Diagnostic and
Therapeutic Criteria Committee of the American Rheumatism
Association. Arthritis & Rheumatism 1986;29:1039–49.
Bacchetti P. Peer review of statistics in medical research: the other
problem. British Medical Journal 2002;324:1271–3.
Badley EM, Tennant A. Impact of disablement due to rheumatic
disorders in a British population: estimates of severity and
prevalence from the Calderdale Rheumatic Disablement Survey.
Annals of the Rheumatic Diseases 1993;52:6–13.
Bland M. An introduction to medical statistics. Oxford: Oxford
University Press; 2000.
Callaghan MJ, McCarthy CJ, Oldham JA. Electromyographic fatigue
characteristics of the quadriceps in patellofemoral pain syndrome.
Manual Therapy 2001;6:27–33.
Callaghan MJ, Oldham JA. The role of quadriceps exercise in the
treatment of patellofemoral pain syndrome. Sports Medicine
1996;21:384–91.
Cheing GL, Hui-Chan CW. The motor dysfunction of patients with
knee osteoarthritis in a Chinese population. Arthritis and
Rheumatism 2001;45:62–8.
Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J.
Delayed onset of electromyographic activity of vastus medialis
obliquus relative to vastus lateralis in subjects with patellofemoral
pain syndrome. Archives of Physical Medicine and Rehabilitation
2001;82:183–9.
Cowan SM, Hodges PW, Bennell KL, Crossley KM. Altered vastii
recruitment when people with patellofemoral pain syndrome
complete a postural task. Archives of Physical Medicine and
Rehabilitation 2002;83:989–95.
Dixon J. Quadriceps activity in osteoarthritis of the knee: An
electromyographic analysis. Ph.D. Thesis, University of Teesside,
2004.
Dixon J, Howe TE, Kent JR, Whittaker VJ. VMO-VL reflex latency
difference between quadriceps components in osteoarthritic and
asymptomatic knees. Advances in Physiotherapy 2004;6:
166–72.
Doxey GE, Eisenman P. The influence of patellofemoral pain on
electromyographic activity during submaximal isometric contrac-
tions. Journal of Orthopaedic & Sports Physical Therapy
1987;9:211–6.
Edwards RH, Young A, Hosking GP, Jones DA. Human skeletal
muscle function: description of tests and normal values. Clinical
Science and Molecular Medicine 1977;52:283–90.
Felson DT, Naimark A, Anderson J, Kazis L, Castelli W, Meenan RF.
The prevalence of knee osteoarthritis in the elderly. The Framing-
ham Osteoarthritis Study. Arthritis & Rheumatism 1987;30:914–8.
Fisher NM, Pendergast DR. Reduced muscle function in patients with
osteoarthritis. Scandinavian Journal of Rehabilitation Medicine
1997;29:213–21.
Goodman SN, Berlin JA. The use of predicted confidence intervals
when planning experiments and the misuse of power when
interpreting results. Annals of Internal Medicine 1994;121:200–6.
Grabiner MD, Koh TJ, Draganich LF. Neuromechanics of the
patellofemoral joint. Medicine and Science in Sports and Exercise
1994;26:10–21.
Hinman RS, Bennell KL, Metcalf BR, Crossley KM. Temporal
activity of vastus medialis obliquus and vastus lateralis in
symptomatic knee osteoarthritis. American Journal of Physical
Medicine and Rehabilitation 2002;81:684–90.
Hodges PW, Bui BH. A comparison of computer-based methods for
the determination of onset of muscle contraction using electro-
myography. Electroencephalography & Clinical Neurophysiology
1996;101:511–9.
Hurley MV. The role of muscle weakness in the pathogenesis of
osteoarthritis. Rheumatic Diseases Clinics of North America
1999;25:283–98.
 ARTICLE IN PRESS
J. Dixon, T.E. Howe / Manual Therapy 12 (2007) 219–225
Hurley MV, Newham DJ. The influence of arthrogenous muscle
inhibition on quadriceps rehabilitation of patients with early,
unilateral osteoarthritic knees. British Journal of Rheumatology
1993;32:127–31.
Hurley MV, Scott DL, Rees J, Newham DJ. Sensorimotor changes
and functional performance in patients with knee osteoarthritis.
Annals of the Rheumatic Diseases 1997;56:641–8.
Kannus P, Jarvinen M, Latvala K. Knee strength evaluation: a
standardized test protocol and scoring scale for isokinetic and
isometric strength measurement of the knee joint for evaluation of
long-term healing results of various knee problems and their
treatment procedures. Scandinavian Journal of Sports Science
1987;9:9–13.
Karst GM. EMG onset timing. Physical Therapy 1998;78:543–6.
Karst GM, Willett GM. Onset timing of electromyographic activity in
the vastus medialis oblique and vastus lateralis muscles in subjects
with and without patellofemoral pain syndrome. Physical Therapy
1995;75:813–23.
Kennedy JC, Alexander IJ, Hayes KC. Nerve supply of the human
knee and its functional importance. American Journal of Sports
Medicine 1982;10:329–35.
Lewek MD, Rudolph KS, Snyder-Mackler L. Quadriceps femoris
muscle weakness and activation failure in patients with sympto-
matic knee osteoarthritis. Journal of Orthopaedic Research
2004;22:110–5.
Lieb FJ, Perry J. Quadriceps function. An anatomical and mechanical
study using amputated limbs. Journal of Bone and Joint Surgery
1968;50-A:1535–48.
Lieb FJ, Perry J. Quadriceps function: an electromyographic study
under isometric conditions. Journal of Bone & Joint Surgery
1971;53-A:749–58.
Mannion AF, Dolan P. Relationship between myoelectric and
mechanical manifestations of fatigue in the quadriceps femoris
muscle group. European Journal of Applied Physiology and
Occupational Physiology 1996;74:411–9.
Marks R, Percy JS, Semple J, Kumar S. Comparison between the
surface electromyogram of the quadriceps surrounding the knees of
healthy women and the knees of women with osteoarthrosis.
Clinical & Experimental Rheumatology 1994;12:11–5.
McNair PJ, Depledge J, Brettkelly M, Stanley SN. Verbal encourage-
ment: effects of maximum effort voluntary muscle activation.
British Journal of Sports Medicine 1996;30:243–5.
225
Pap G, Machner A, Awiszus F. Strength and voluntary activation of
the quadriceps femoris muscle at different severities of osteoar-
thritic knee damage. Journal of Orthopaedic Research 2004;22:
96–103.
Powers CM, Landel R, Perry J. Timing and intensity of vastus
muscle activity during functional activities in subjects with and
without patellofemoral pain. Physical Therapy 1996;76:
946–55.
Sharma L, Dunlop D, Cahue S, Song J, Hayes KW. Quadriceps
strength and osteoarthritis progression in malaligned and lax
knees. Annals of Internal Medicine 2003;138:613–9.
Shrier I. Muscle dysfunction versus wear and tear as a cause of exercise
related osteoarthritis: an epidemiological update. British Journal of
Sports Medicine 2004;38:526–35.
Soderberg GL, Knutson LM. A guide for use and interpretation of
kinesiologic electromyographic data. Physical Therapy
2000;80:485–98.
Spencer JD, Hayes KC, Alexander IJ. Knee joint effusion and
quadriceps reflex inhibition in man. Archives of Physical Medicine
and Rehabilitation 1984;65:171–7.
Staude G, Wolf W. Objective motor response onset detection in
surface myoelectric signals. Medical Engineering & Physics
1999;21:449–67.
Stensdotter A-K, Hodges PW, Mellor R, Sundelin G, Hager-Ross C.
Quadriceps activation in closed and in open kinetic chain exercise.
Medicine & Science in Sports & Exercise 2003;35:2043–7.
Tan J, Balci N, Sepici V, Gener FA. Isokinetic and isometric strength
in osteoarthrosis of the knee: a comparative study with healthy
women. American Journal of Physical Medicine and Rehabilitation
1995;74:364–9.
Tornvall G. Assessment of physical capabilities with special reference
to the evaluation of maximal voluntary isometric muscle strength
and maximal working capacity. Acta Physiologica Scandinavica
1963;58(Suppl 201):1–102.
Torry MR, Decker MJ, Viola RW, O’Connor DD, Steadman JR.
Intra-articular knee joint effusion induces quadriceps avoidance
gait patterns. Clinical Biomechanics 2000;15:147–59.
Witvrouw E, Sneyers C, Lysens R, Victor J, Bellemans J. Reflex
response times of vastus medialis oblique and vastus lateralis in
normal subjects and in subjects with patellofemoral pain syndrome.
Journal of Orthopaedic and Sports Physical Therapy 1996;24:
160–5.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 226–230

www.elsevier.com/locate/math

Original article

The validity of clinical measures of patella position

Islay McEwana, Lee Herringtonb,c,, Jeanette Thomd
a
Department of Exercise and Sports Science, Manchester Metropolitan University, UK
b
Directorate of Sport, Allerton Annexe, University of Salford, Manchester M6 6PU, UK
c
Centre for Rehabilitation and Human Performance Research, University of Salford, Manchester, UK
d
Department of Exercise and Sports Science, University of Bangor, UK
Received 4 May 2005; received in revised form 20 March 2006; accepted 27 June 2006

Abstract

Patellar taping is regarded as an important element of the treatment of patellofemoral joint pain. Key to the successful use of
patellar taping is the assessment of patella position. The reliability and validity of the techniques used to assess patella position has
been questioned. The aim of the study was to assess the validity of the clinical assessment technique of patella medio-lateral position
and patella lateral tilt against the criterion measure of MRI. Twenty-four subjects eight females and 16 males had their patella
position examined in the study (mean age 24.577.9 years, range 18–42 years). The study also assessed intra-tester reliability of the
technique. A good correlation was found between the findings of the clinical test for medio-lateral position and the MRI measure
(r ¼ 0:611, p ¼ 0:002). All of the subjects found to have a laterally tilted patella on clinical examination had a lateral patella tilt
defined by PTA of greater than 51. Those subjects with a PTA of less than 51 on clinical examination were assessed as having no
degree of patella tilt. The study undertaken shows that when undertaken by an experienced manual therapist positional assessment
of the patella can have strong criterion validity and intra-tester reliability.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Assessment; Patella; Validity; Reliability

1. Introduction
Within clinical practice, it has always been considered
essential to accurately diagnose the presenting dysfunc-
tion and then treat accordingly. This is be particularly
true when assessing patella position prior to the
application of patellar taping, a common treatment for
patellofemoral pain, as the anatomical findings will to a
large extent dictate the directional control the tape will
provide (McConnell, 1996).
Corrective taping for patella maltracking was intro-
duced nearly two decades ago by McConnell (1986), this
was one of the first treatments to have claimed to
Corresponding author. Directorate of Sport, Allerton Annexe,
University of Salford, Manchester M6 6PU, UK.
Tel.: +00 441612952326.
E-mail address: L.C.Herrington@Salford.ac.uk (L. Herrington).
achieve long term success for this often difficult to treat
condition (Dye, 2005). The primary theory underpin-
ning treatment is that corrective taping alters the
abnormal patella orientation, therefore normalizing
patella position and in so doing relieves the stress on
the patellofemoral joint (PFJ) and surrounding tissues
and thus reduces pain (Dye, 2005). Although many
authors have demonstrated that corrective taping
relieves pain, the mechanism by which this is brought
about still remains unclear (Crossley et al., 2000).
Key to the success of patella taping is the accurate
assessment of patella position in order that tape may be
applied to counter the abnormalities of position
presented (McConnell, 1996). The method described
originally by McConnell (1986) to assess patella position
has been investigated by a number of authors. The inter-
tester reliability of the assessment method has been
regarded as poor (Fitzgerald and McClure, 1995;

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.013
 ARTICLE IN PRESS
I. McEwan et al. / Manual Therapy 12 (2007) 226–230
Watson et al., 1999) along with the intra-tester reliability
(Tomisch et al., 1996). Powers et al. (1999) also
questioned the validity of the method. In contrast,
Herrington (2002) and Herrington and Nester (2004)
found the method to assess medio-lateral position of the
patella described by McConnell (1986) to be reliable
when used by experienced clinicians. Herrington (2002)
criticized the previous studies for using relatively novice
clinicians with their minimal training being the sig-
nificant factor in the test’s reliability not the test itself.
The study of Herrington (2002) was in contrast to that
of Powers et al. (1999), with Powers et al. (1999)
reporting poor validity of the assessment methods
against the criterion values produced from assessment
of patella position by MRI, whilst Herrington (2002)
showing a strong relationship between the clinical
measure and MRI findings.
The contradictory findings in the literature would appear
to indicate that there is a need for further study in this area.
The aim of this study is firstly: to assess the validity of the
clinical measures of patella position against the criterion
measure of MRI, here unlike previous studies both patella
medio-lateral position and lateral tilt will be assessed.
Secondly the intra-tester reliability of the clinical measure-
ments will be examined. If these assessment techniques are
found to be valid and reliable then decision-making
regarding the choice of taping technique will have improved
efficacy. If the techniques prove to be unreliable and not
valid then this will question the significance of any direction
specific effects for patella taping as the direction of the
effects cannot be defined clinically.
2. Method
2.1. Subjects
Twenty-four subjects, eight females and 16 males were
examined in the study, (mean age 24.577.9 years, range
Fig. 1. Clinical technique for assessing patella position. The femoral condyles are palpated and marked on the tape along with the mid point of the
patella. The distances between the condyles and the middle of the patella can be subtracted from each other to give relative medio-lateral position.
227
18–42 years). Subjects were excluded if they had any
current or previous history of knee or lower limb injury.
The subjects were recruited as a sample of convenience
from a University Exercise and Sports Science Depart-
ment and were all physically active and regular sports
participants, typical of a population from which
patellofemoral pain patients would come (Taunton et
al., 2002). The tests were performed in agreement with
the declaration of Helsinki and all subjects gave
informed written consent to participate prior to
commencing study. The study was approved by the
institutional research ethics committee.
2.2. Clinical measurement of medio-lateral patella
position
The method used was that described by McConnell
(1986) and Herrington and Nester (2004). The following
landmarks were located: medial and lateral epicondyles
of the femur and mid point of the patella on the subject’s
right knee, which was positioned and supported in 201
of knee flexion (in order to place the patella within the
trochlea groove). The examiner then marked the
distance from the ascribed position of the medial and
then the lateral epicondyle of the femur to the mid-point
of the patella on a piece of folded zinc oxide tape. The
method is shown graphically in Fig. 1. The medial and
lateral measurements were each repeated three times on
separate pieces of tape, with re-palpation of the land-
marks on each occasion. An independent assessor then
measured the distance of medial and lateral points to
mid point with the average of the three measurements
being recorded. These measurements of the tape were
then repeated one day later. The level of agreement
between days was found by ICC to be r ¼ 0:92
(po0:05).
 ARTICLE IN PRESS
228 I. McEwan et al. / Manual Therapy 12 (2007) 226–230

2.3. Clinical measurement of lateral patellar tilt

The method used was that described by McConnell

(1986) and Fitzgerald and McClure (1995). The degree
of tilt was determined by comparing the height of the
medial and lateral borders of the patella. The examiner
places their thumb and index finger on the medial and
lateral borders of the patella. The patella would be
laterally tilted if the medial border is more anterior than
the lateral and vice versa for medial tilt.

2.4. Intra-tester reliability of clinical measurements

At random intervals during the testing period (in-

between the other subjects) the examiner repeatedly
examined the patella position of a single subject (10
times in total). These measurements were then examined
to ascribe intra-tester reliability of the examiner. The
physiotherapist carrying out the testing was blind to the
results of each test. The physiotherapist carrying out all
testing was a clinical specialist with specific manual
therapy qualifications and 15 years experience of
musculoskeletal medicine.

2.5. Measurement of patella position following MRI

Transverse plane MRI images were performed on the
participant’s right knee using an extremity coil fixed 0.2-
T MRI scanner (E-Scan; ESAOTE Biomedica, Genova,
Italy). The knee was fixed at 20 degrees of knee flexion
using a wedge whilst the participant was supine. Scans
were performed using a T1 weighted 3D isotropic profile
with the following scanning parameters: time to echo:
16 ms; repetition time: 38 ms; field of view:
180 mm  180 mm; matrix: 256  192. Images were
sliced at 5 mm intervals (with a 10% gap) and the image
which included the marker was used for analysis. The
scans were digitized on a Macintosh G4 computer and
analysed off-line with ImageJ image analysis software
(National Institutes of Health, USA).
Medio-lateral component of patella orientation was
measured by finding the lateral patellar displacement
(LPD) on the MRI film, using the method described by
Larsen et al. (1995). The LPD quantifies the position of
the patella in the frontal plane relative to the medial
femoral condyle. Fig. 2 demonstrates the measurement
of LPD. A single examiner was used to measure the
position, who was blind to the findings of the clinical
examination. The LPD was measured three times with
the average being taken and compared to a repeated
measurement one day later, with the level of agreement
between days found by ICC to be r ¼ 0:99 (po0:05).
Lateral tilt component of patella orientation was
measured by finding the patella tilt angle (PTA) on the
MRI film using the method described by Guzzanti et al.
(1994). Fig. 3 demonstrates the measurement of PTA. A
Fig. 2. Schematic of the method used to assess medial/lateral position
of the patella from MRI. Position is determined by drawing lines
connecting the anterior portions of the medial and lateral femoral
condyles (AB), and the maximum width of the patella (CD). A line was
then drawn from the anterior portion of the lateral condyle so that it
was perpendicular to line AB and bisected line CD. The distance from
point E to point C represented the amount of medial or lateral
orientation.
single examiner was used to measure the position, who
was blind to the findings of the clinical examination.
This was measured three times with the average being
taken and compared to a repeated measurement one day
later. The level of agreement was found by ICC to be
r ¼ 0:94 (po0:05) between days.
2.6. Statistical analysis
The degree of agreement between the two techniques
(MRI and clinical) was quantified by calculation of
Pearsons product moment. The intra-tester reliability of
the clinical measurement was quantified by means of the
intra-class correlation coefficient. For purposes of this
study, correlation coefficients were interpreted as
follows: below 0.50 was poor, 0.50–0.75 was good, and
above 0.75 was excellent (Portney and Watkins, 1993).
Statistical analysis was undertaken on the statistical
analysis package SPSS (version 12).
 ARTICLE IN PRESS
I. McEwan et al. / Manual Therapy 12 (2007) 226–230 229

Table 1
Intratester reliability of the clinical measure of mediolateral patella
position

Lateral Medial distance Difference

Distance (mm) (mm) (mm)

Mean 8.3 8.9 0.6

Standard 0.1 0.1 0.1
Deviation
Standard error of 0.1 0.1 0.1
measurement
Confidence 8.1–8.5 8.7–9.1 0.4–0.8
interval (95%)
ICC3, k 0.86 0.82 0.9
 Statistical Signifcant (po0:01).

Table 2
Results from the examination of patella position by MRI and clinical
method

LPD (mm) PTA (1) Clinical medio-

lateral test
(difference, mm)

Mean 8.1 6.8 5.0

Fig. 3. Calculation of the patella tilt angle on MRI scan (from Standard 2.8 3.9 3.3
Guzzanti et al., 1994). The PTA is identified by the intersection of a Deviation
line parallel to the lateral facet of the patella (y) and the base line (x) Range 4.6–13.8 4–12 0–13
which runs from the deepest point of the patella to the edge of the
LPD Lateral patella displacement, PTA Patella tilt angle.
lateral condyle.  Minus value equals medial patella tilt.

3. Results 3.4. Between method agreement of patella position

3.1. Intra-tester reliability
Table 1 shows the results of the intra-tester reliability
study for the clinical measure of medio-lateral patella
position.
3.2. Patella position from MRI film
Table 2 shows the results from the MRI examination
of patella position. All 24 subjects demonstrated some
degree of lateral patella displacement as assessed by
LPD. 22 of the 24 subjects had laterally tilted patellae
assessed by PTA, with 2 subjects having a medially tilted
patella.
3.3. Clinical assessment of patella position
All subjects except one were found to have laterally
displaced patellae, the one exception was found to have
a centralized patella. Mean lateral patella displacement
was 573 mm (range 0–13 mm). A total of 17 subjects
were found to have laterally tilted patella.
The Pearson’s product moment revealed a good
correlation between the findings of the clinical test for
medio-lateral position and the MRI measure for LPD
(r ¼ 0:611, p ¼ 0:002). All of the subjects found to have
a laterally tilted patella on clinical examination had a
lateral patella tilt defined by PTA of greater than 51.
Those subjects with a PTA of less than 51 on clinical
examination were assessed as having no degree of
patella tilt.
4. Discussion
The results of the study would appear to indicate that
the clinical method used to assess patella position
described shows a good and significant relationship to
patella positions measured from MRI. Powers et al.
(1999) found the assessment technique to have poor
agreement with measures from MRI and to over-
estimate the degree of lateral displacement, these
findings were in contrast to this study and those of
Herrington (2002) with strong agreement being shown
and no over-estimation of displacement. In the Powers
et al. (1999) study the examiner had less than one years
experience of using this technique, the length of time
 ARTICLE IN PRESS
230 I. McEwan et al. / Manual Therapy 12 (2007) 226–230
since qualification and the nature of any post-graduate
training was not reported, this was in contrast
to the examiner in this study, who had over 15 years
clinical experience and specialist manual therapy
training.
The findings of the study support those of Herrington
(2002) which contended that the clinical measurement
of medio-lateral patella position used showed strong
criterion validity. Furthermore, this study also demon-
strated that the clinical measure of patella tilt out-
lined also showed strong criterion validity. Criterion
validity establishes the ability of one test to predict the
results obtained by another (George et al., 2000). This
analysis is particularly useful in this case were a
practical, quick and cheap test is proposed as an
alternative to a time consuming and cost prohibitive
procedure (MRI).
Previous studies have shown that reliability when
using these techniques varies considerably with training,
experience and skill (Herrington, 2002; Watson
et al., 1999). The clinician undertaking the measure-
ments was an experienced manual therapist, pre-
viously reported research (Herrington, 2003) has shown
manual therapy training to have a significant effect on
reliability when undertaking these tests and the results
of this study and its external validity must be viewed
in light of such information. Alongside the external
validity issue of using an experienced manual therapist
to carry out the tests, the measurements could have
been open to recall bias effecting the reliability
findings, the study aimed to minimize this with blinding
of the assessor, both to the clinical findings and those of
the MRI.
The study undertaken demonstrated that the majority
of asymptomatic physically active individuals examined
had laterally displaced and tilted patellae, which have
previously been regarded as pathological (McConnell,
1986). Further study would be required in a pathological
population to assess the extent of these findings within
this patient group. It may prove that these tests are both
reliable and valid, but lack sensitivity in detecting those
individuals with PFJ pain and so have little ecological
validity (George et al., 2000), this requires further
investigation. Alternatively, it may be that the lateral
deviation of the patella found in these normal indivi-
duals relates to a sub-optimal performance of the PFJ
and would lead to potential pathology with the
appropriate alteration of environmental circumstances
(Dye, 2005), if this is the case then the tests may prove
an appropriate screening test for PFJ pain, this again
requires further investigation.
5. Conclusion
Previous research has questioned the validity of the
methods of assessing patella position described by
McConnell (1986), this study would appear to refute
those findings. The study undertaken shows that when
undertaken by an experienced manual therapist posi-
tional assessment of the patella can have strong criterion
validity and intra-tester reliability.
References
Crossley K, Cowan S, Bennell J, McConnell J. Patellar taping: is
clinical success supported by scientific evidence. Manual Therapy
2000;5(3):142–50.
Dye S. The pathophysiology of patellofemoral pain. Clinical
Orthopaedic and Related Research 2005;436:100–10.
Fitzgerald K, McClure P. Reliability of measurements obtained with
four tests for patellofemoral alignment. Physical Therapy
1995;75:84–92.
George K, Batterham A, Sullivan I. Validity in clinical research:
a review of basic concepts and definitions. Physical Therapy in
Sport 2000;1:19–27.
Guzzanti V, Gigante A, Di Laazzaro A, Fabbriciani A. Patellofemoral
malalignment in adolescents. American Journal of Sports Medicine
1994;22:55–60.
Herrington L. The inter-tester reliability of a clinical measurement
used to determine the medial lateral orientation of the patella.
Manual Therapy 2002;7:163–7.
Herrington L. The reliability of a clinical measurement used to
determine the medial/lateral orientation of the patella World
Confederation of Physical Therapy 14th International Conference,
Barcelona, June 2003.
Herrington L, Nester C. Q-angle undervalued? The relationship
between Q angle and medio-lateral position of the patella. Clinical
Biomechanics 2004;19:1070–3.
Larsen B, Andersen E, Urter A, Mickelson M, Newhouse K. Patellar
taping: a radiological examination of medial glide taping. Amer-
ican Journal of Sports Medicine 1995;23:465–71.
McConnell J. The management of Chondromalacia Patellae: a long term
solution. Australian Journal of Physiotherapy 1986;32:215–22.
McConnell J. Management of patellofemoral problems. Manual
Therapy 1996;1:60–6.
Portney L, Watkins M. Foundations of clinical research: applications
to practice. Norwalk, Connecticut: Appleton and Lange; 1993.
Powers C, Mortenson S, Nishimoto D, Simon D. Criterion-related
validity of a clinical measurement to determine the medial/lateral
component of patellar orientation. Journal of Orthopaedic and
Sports Physical Therapy 1999;29:372–7.
Taunton J, Ryan M, Clement D, McKenzie D, Lloyd-Smith D,
Zumbo B. A retrospective case-control analysis of 2002 running
injuries. British Journal of Sports Medicine 2002;36:95–101.
Tomisch D, Nitz A, Threlkeld J, Shapiro R. Patellofemoral alignment:
reliability. Journal of Orthopaedic and Sports Physical Therapy
1996;23:200–8.
Watson C, Propps M, Galt W, Redding A, Dobbs D. Reliability of
McConnell’s classification of patellar orientation in symptomatic
and asymptomatic subjects. Journal of Orthopaedic and Sports
Physical Therapy 1999;29:378–85.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 231–239

www.elsevier.com/locate/math

Original article

Strain on the repaired supraspinatus tendon during manual traction

and translational glide mobilization on the glenohumeral joint:
A cadaveric biomechanics study
Takayuki Murakia,, Mitsuhiro Aokib, Eiichi Uchiyamac, Tomoya Miyasakaa,
Gen Murakamic, Shigenori Miyamotob
a
Graduate School of Health Sciences, Sapporo Medical University, Sapporo, Japan
b
Department of Physical Therapy, Sapporo Medical University School of Health Sciences, Sapporo, Japan
c
Department of Anatomy, Sapporo Medical University School of Medicine, Sapporo, Japan
Received 21 July 2005; received in revised form 4 May 2006; accepted 27 June 2006

Abstract

There has been no report on the mechanical effects of joint mobilization on rotator cuffs. The purpose of this study was to
determine whether it is safe to use grade 3 joint mobilization techniques after rotator cuff repair. Nine fresh frozen cadaveric
shoulders were used in this study. The strains on the artificially repaired supraspinatus tendon during joint mobilization were
measured at 01 and 301 of shoulder abduction and were compared with those at the maximal stretching position and relaxing
position. Additionally, gap distances were measured during this experiment. The strain at 301 of abduction of the repaired tendon
during each joint mobilization was significantly smaller than that at 01 abduction (Po0:05). At 301 of abduction, the strain during
joint mobilization was not statistically different from that of the shoulder in the relaxing position, except during the inferior glide
technique. Gap distances were 0 mm at 301, while the distances were 1.06–1.46 mm at 01. Our findings suggest that joint mobilization
techniques, except inferior glide, can be performed safely without significantly straining the repaired tendon at 301 of abduction, if
rotator cuff repair is performed at 01 of abduction.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Strain; Supraspinatus tendon; Rotator cuff tear; Fresh cadaver

1. Introduction tion (Bruzga and Speer, 1999; Mangus et al., 2002)

Shoulder joint contracture can occur after repair of
rotator cuff tears. This can be caused by capsular
contracture, tendon shortening, scars, and adhesions of
the subacromial and the subscapularis-conjoined tendon
regions (Warner and Greis, 1998; Hatakeyama et al.,
2001a; Matsen et al., 2004). Therefore, intensive physical
therapy using range of motion exercise (ROM exercise)
(Cofield, 1985; Matsen et al., 2004) and joint mobiliza-
Corresponding author. Tel.: +81 11 611 2111;
fax: +81 11 611 2150.
E-mail address: tmuraki@sapmed.ac.jp (T. Muraki).
immediately after surgery is important.
Generally, in order to prevent and treat joint
contracture, limited movement of the joint itself is used
in ROM exercise. However, when this movement causes
pain, effective stretching of the connective tissue that
limits motion becomes difficult (Quillen et al., 1992).
For example, when contracture of the posterior capsule
occurs in the glenohumeral joint, antero-superior
translation of the humeral head occurs during arm
abduction and, this translation consequently leads to
subacromial impingement (Harryman et al., 1990;
Warner et al., 1990). In addition, contracture of the
anterior capsule (Flatow et al., 1994) and the inferior

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.017
 ARTICLE IN PRESS
232 T. Muraki et al. / Manual Therapy 12 (2007) 231–239
capsule (Cofield, 1985; Hjelm et al., 1996) often lead
to impingement. Therefore, in a shoulder joint with
a capsular contracture, this exercise can actually
worsen any injuries of the joint capsule and rotator
cuff.
On the other hand, joint mobilization, such as
traction and glide, is used to stretch the tendon,
ligament, and capsule and to improve the physiological
accessory movement. Traction is the technique that
distracts one articular surface perpendicular to the
other, and glide techniques translationally glide one
articular surface parallel to the other (Kaltenborn,
1999). These techniques are considered capable of
stretching the particular connective tissues that limit
joint motion without impingement, resulting in an
improvement of the limited ROM and reduction in pain
(Johns and Wright, 1962; Quillen et al., 1992).
The effects of joint mobilization have been previously
reported. Hsu et al. (2000a, b, 2002a) demonstrated
that anterior–posterior and inferior translational
gliding improved the range of abduction and external
rotation in cadaveric glenohumeral joints. Conroy
and Hayes (1998) investigated the effects of joint
mobilization and compared them to general physical
therapy on a patient with subacromial impinge-
ment syndrome. They then reported that joint mobiliza-
tion was not effective on ROM and any functional
outcome of the shoulder joint, but it provided effective
pain relief.
However, care should be taken when joint mobiliza-
tion is used on the shoulder joint after surgery. In order
to prevent and treat joint contracture after rotator cuff
repair, knowledge of the mechanical stresses of joint
mobilization on the repaired rotator cuff is required.
Hatakeyama et al. (2001a) and Zuckerman et al. (1991)
observed the effects of shoulder positions on the rotator
cuff under several different conditions. Although the
effects and safety of joint mobilization were studied
from the physiological aspect of the shoulder (Hsu et al.,
2002b; Gokeler et al., 2003), the mechanical effect of
joint mobilization on the rotator cuff was unclear
because connective tissues of the shoulder that had
different mechanical properties were not observed
individually in these studies. Therefore, the mechanical
effects of joint mobilization on intact and repaired
rotator cuffs should be clarified first in order to perform
joint mobilization without excessive stretching stress on
the rotator cuff.
The purpose of this study was to measure the strain
on both intact and repaired supraspinatus tendons,
which is the primary site of rotator cuff tears, and
the gap distance of the repair site during joint
mobilization by using fresh frozen cadaveric shoulders.
Furthermore, based on these results, we discuss about
safely applying joint mobilization after rotator cuff
repair.
2. Methods
2.1. Preparation of specimens
Nine frozen shoulders (four left shoulders and five
right shoulders) harvested from nine fresh cadavers were
used in the experiment. The mean age at death was 80
years (71–91 years of age). Any shoulder with macro-
scopic evidence of rotator cuff tears or osteoarthritis
was excluded. However, none of these shoulders
conformed to these criteria.
The shoulders, disarticulated from their thoraxes,
were stored in a freezer at 20 1C. The thawing of the
specimens at room temperature started 12 h prior to
experimentation. Then, soft tissues—except the rotator
cuff muscles, biceps brachii muscles, coracoacromial
ligaments, and capsules—were carefully removed to
avoid the loss of intra-articular negative pressure in the
glenohumeral joint. The distal third of the humerus was
exposed, and an acrylic stick was inserted perpendicular
to the shaft indicating the direction of the forearm.
Next, the humerus was amputated above the elbow.
During the experiment, the specimens were kept moist
by spraying saline solution on them every 5–10 min. The
room temperature was maintained at 22 1C.
2.2. Testing apparatus
A wooden jig, consisting of a wooden board and a
square timber, was used for this experiment. The scapula
of the specimen was fixed on the wooden jig so that the
medial border of the scapula was perpendicular to the
ground (Culham and Peat, 1993) (Fig. 1). Two anchors
(Fastin RC threaded suture anchor, Mitek, Tokyo,
Japan) were inserted into the bony insertion of the
infraspinatus and subscapularis tendons to apply a
compressive force of 11 N to each thread (total 22 N)
against the glenoid fossa. In previous cadaveric studies,
this compressive force was used as the minimum force
required, preventing subluxation of the humeral head on
application of translational loads (Warner et al., 1992;
Tibone et al., 1998). Using this system, the humeral head
was maintained in concentric position onto the glenoid
fossa after joint mobilization was performed.
2.3. Measurement device
The strain data on the supraspinatus tendon was
obtained from a precise displacement sensor (Pulse
Coder, LEVEX, Kyoto, Japan) (Fig. 2a). This Pulse
Coder consisted of a coil sensor and a brass pipe.
The displacement was measured by detecting the
position of the brass pipe relative to the coil sensor that
generated the magnetic field. Analogue data of the
displacement was represented and recorded on a digital
scaling meter (HV35, Allied Control, Tokyo, Japan)
 ARTICLE IN PRESS
T. Muraki et al. / Manual Therapy 12 (2007) 231–239 233

3SPACE sensors

Pulse corder

Fig. 1. Schematic illustration of the experimental setup. A scapula of a specimen was mounted vertically on the wooden jig. A Pulse Coder was
attached to the supraspinatus tendon. Sensors of the 3SPACE device were attached to the acromion and the humerus.

Fig. 2. Photographs of the Pulse Coder. (a) The precise displacement sensor ‘‘Pulse Coder’’ used to measure the strain on tendons by changes in the
length between points; (b) Pulse Coder attached to the supraspinatus tendon with sutures.

that converted the obtained data into a digital form.
The non-linearity of this sensor was 0.25%/full scale,
and the range of measurement was 14 mm according
to the manufacturer’s instructions. The sensors with
fishhook-like barbed points were attached to the greater
tuberosity and the proximal part of the supraspinatus
tendon. The sensor was placed parallel to the
tendon fibre (Fig. 2b). Changes in the length between
the points of the coil sensor and the brass pipe allowed
the sensor to measure the strain on the supraspinatus
 ARTICLE IN PRESS
234 T. Muraki et al. / Manual Therapy 12 (2007) 231–239
tendon. By using a digital caliper, the accuracy of
preliminary calibration of the Pulse Coder, which was
attached in the supraspinatus tendon, was 0.1 mm root
mean square (RMS). A similar type device, which uses
the same principle, was previously used to measure
strain on the supraspinatus tendon (Hatakeyama et al.,
2001a, b).
A six-degree-of-freedom electromagnetic tracking
device (3SPACE FASTRACK, Polhemus, Colchester,
Vermont) was used to monitor the precise glenohumeral
angles during this measurement. This device enabled
measurement of the three-dimensional position and
orientation of the sensors relative to the absolute
coordination generated by the source (An et al., 1988).
One sensor was placed on the acromion and the other
was placed on the middle portion of the humerus
(Fig. 1). In this system, the angle of arm abduction and
extension was defined as the angle between the plane of
the glenoid fossa and the longitudinal axis of the
humerus. The rotation angle was defined as the rotation
of the humerus along the longitudinal axis. With a
750-mm range of measurement from the source, the
positional accuracy was 0.8 mm RMS, and the angular
accuracy was 0.51 RMS. The 3SPACE device kinema-
tically monitored the glenohumeral angles during the
strain measurement.
Fig. 3. Photographs of the supraspinatus tendon repair. (a) Artificial supraspinatus tear (width 2.0 cm, length 1.5 cm). Small arrows indicate the site
of supraspinatus tendon tear; (b) Repair of torn supraspinatus tendon. The threads were clamped with a 3-kg force at the outlet from the bone (a
large arrow). Small arrows indicate the repair site.
2.4. Experimental procedure
The measurements were first performed on intact
supraspinatus tendons to assess the difference of tendon
conditions and then on the repaired tendon model. In
the case of the model, the supraspinatus tendon was
excised (width 2.0 cm, length 1.5 cm) (Hatakeyama et al.,
2001a, b)(Fig. 3a) from the greater tuberosity, which
simulated the retracted supraspinatus tendon, and was
repaired with #2 polyester threads (Ethibond, Ethicon
Inc. Somerville, NJ) passed through drill holes in the
greater tuberosity. The threads were pulled out with a 3-
kg force and clamped at the outlet from the bone with
the arm in 301 of external rotation at 01 of arm
abduction (Hatakeyama et al., 2001a, b) (Fig. 3b).
Acromioplasty and removal of the coracoacromial
ligament was also performed because this technique is
often used in rotator cuff repair and enabled the Pulse
Coder to move in the subacromial space. All these
techniques were performed by an orthopaedic surgeon
who was familiar with shoulder surgery.
The neutral position was defined as 301 of external
rotation at 01 of abduction in the glenohumeral joint,
because the scapula internally tilts 301 relative to the
coronal plane in vivo (Culham and Peat, 1993; Itoi et al.,
2004). For reliable comparison, the maximal stretching
 ARTICLE IN PRESS
T. Muraki et al. / Manual Therapy 12 (2007) 231–239
position (adduction at extension) (Evjenth and Ham-
berg, 1984) and relaxing position of the supraspinatus
tendon (neutral position at 301 of abduction) (Zucker-
man et al., 1991; Hatakeyama et al., 2001a) were used as
reference positions producing the highest and lowest
tendon strain. The shoulder positions examined in this
mobilization experiment were at 01 and 301 of arm
abduction in the scapular plane. In order to reproduce
shoulder positions during the measurement, the posi-
tions were monitored by the 3SPACE FASTRACK
device.
Four joint mobilization techniques, i.e. traction,
inferior, anterior and posterior glide, were performed
at each position by a physical therapist who had more
than 5 years of experience in correcting shoulder
disorders by using joint mobilization. The therapist
had no information regarding the hypotheses of this
study, and the strain data represented on the digital
scaling meter was masked. The grade of joint mobiliza-
tion was set at 3 as defined by Kaltenborn (1999). Grade
3 is used to stretch connective tissues by applying force
to the final stop during joint mobilization. At these
positions, the strains and gap distances on the supras-
pinatus tendon were measured and compared with those
during joint mobilization techniques. The measurements
were performed 3 times during each joint mobilization
technique in each of the two positions. The holding time
during each measurement was set at 20 s. In a clinical
setting, joint mobilization for 20–60 s was used to stretch
the connective tissue (Quillen et al., 1992; Conroy and
Hayes, 1998; Mangus et al., 2002). Hsu et al. (2000a, b,
2002a, b) reported that the ROM improved after joint
mobilization was done for 10–30 s. Therefore, we
decided that the holding time should be set at 20 s as
the minimal time that is effective in biomechanical study
and clinical practice.
2.5. Data analysis
First, the length between the barbed points of the coil
sensor and the brass pipe on the sensors at the neutral
position was recorded. Next, the longitudinal displace-
ment of the sensors at the measurement area of the
tendon was recorded when the measurement positions
were held for 20 s. The displacement was defined as the
length change from the neutral position. The strain on
the tendon was calculated by the following formula:
Strainð%Þ ¼ DLðmmÞ=LðmmÞ  100.
L indicates the length between the points at the
neutral position, and DL indicates the displacement
from L. A positive strain value indicates that the
supraspinatus tendon was stretched from the neutral
position, and a negative strain value indicates a
slackening of the tendon.
235
Gap distance was regarded as the distance between
the proximal edge of the torn supraspinatus tendon and
its distal edge. Therefore, we defined the detected
displacement by the sensor, which was placed across
the repair site as gap distance (DL), as Pruitt et al. (1991)
defined it in a previous study regarding flexor tendon
repair. In the neutral position, the gap distance was
0 mm with a 3-kg tensile force on the repair site.
2.6. Statistical analysis
Intra-class correlation coefficients were calculated to
determine test–retest reliability in each condition. A
two-way repeated measures analysis of variance was
used to determine the effects of supraspinatus tendon
conditions (intact tendon and repaired tendon),
shoulder positions (01 and 301 of shoulder abduction),
and joint mobilization techniques (traction and inferior/
anterior/posterior glide). To detect the differences
among the joint mobilization techniques, Bonferroni’s
multiple comparison procedure was used. Moreover,
Dunnett’s multiple comparison test was performed to
compare the strain during joint mobilization techniques
occurring in both the relaxing and stretching positions.
The a level was set at 0.05. All statistical analyses were
performed on SPSS for Windows ver.11.5J. (SPSS Japan
Inc., Tokyo, Japan).
3. Results
3.1. Reliability of these measurements
Intra-class correlation coefficients of these measure-
ments in each condition ranged from 0.809 to 0.984.
These values corresponded to almost perfect (Landis
and Koch, 1977).
3.2. Strain on the intact supraspinatus tendon
The strains on the intact supraspinatus tendon are
shown in Fig. 4. The strains during each joint
mobilization technique at 301 of arm abduction were
significantly smaller than those occurring at 01
(Po0:005). There were no significant differences among
joint mobilization techniques both at 01 and 301 of arm
abduction. At 01 of arm abduction, the strains during
each joint mobilization technique did not show sig-
nificant differences from those occurring in the stretch-
ing position, and were significantly larger than those
occurring in the relaxing position (Po0:001). At 301 of
arm abduction, the strains during each joint mobiliza-
tion technique were significantly smaller than those
occurring in the stretching position (Po0:05), while
these strains did not show significant differences from
those in the relaxing position.
 ARTICLE IN PRESS
236 T. Muraki et al. / Manual Therapy 12 (2007) 231–239

10

5
Trac

0 Inf
Strain (%)

Ant
-5
Post

-10 Add.Ext

30 Abd
-15

-20
Abduction 0° Abduction 30° Reference positions

Fig. 4. Strain on intact supraspinatus tendon. The values and bars represent mean strain and standard deviation, respectively. Direction of the
glenohumeral mobilization and reference positions: Trac, traction; Inf, inferior glide; Ant, anterior glide; Post, posterior glide; Add. Ext, adduction at
extension; 30 Abd, 301 of arm abduction with neutral rotation.

80

60
Trac

40 Inf
Strain (%)

Ant
20
Post

0 Add.Ext

30 Abd
-20

-40
Abduction 0° Abduction 30° Reference positions

Fig. 5. Strain on repaired supraspinatus tendon. The values and bars represent mean strain and standard deviation, respectively. Direction of the
glenohumeral mobilization and reference positions: Trac, traction; Inf, inferior glide; Ant, anterior glide; Post, posterior glide; Add. Ext, adduction at
extension; 30 Abd, 301 of arm abduction with neutral rotation.

3.3. Strain on the repaired supraspinatus tendon
The strains on the repaired supraspinatus tendon are
shown in Fig. 5. While the strain on the repaired
supraspinatus tendon during joint mobilizations was
significantly larger than that on the intact tendon at 01
of abduction (Po0:05), the strain on the repaired
tendon during joint mobilizations was not different
from that occurring on the intact tendon at 301 of
abduction (Figs. 4 and 5). The strains during each joint
mobilization technique at 301 of arm abduction were
significantly smaller than those occurring at 01
(Po0:005). Among joint mobilization techniques, there
were no significant differences in the strains at both 01
and 301 of arm abduction.
At 01 of arm abduction, the strains during each joint
mobilization technique were significantly smaller than
those occurring in the stretching position (Po0:001),
 ARTICLE IN PRESS
T. Muraki et al. / Manual Therapy 12 (2007) 231–239
Table 1
Gap distance on the repaired supraspinatus tendon during joint
mobilization, stretching position, and relaxing positiona
Gap distance (mm)
01 of abduction 301 of abduction
Traction 1.371.2 0 issue, the stress on particular tissues such as the
Inferior glide 1.571.3 0 supraspinatus tendon by joint mobilization should be
Anterior glide 1.371.3 0 quantitatively determined.
Posterior glide 1.171.0 0 In this study, the strain and gap distance on the
Stretching position 4.574.0
Relaxing position 0
a
The values represent mean gap distance and standard deviation.
while these strains were significantly larger than those
occurring in the relaxing position (Po0:001). At 301 of
arm abduction, although the strains during each joint
mobilization technique were significantly smaller than
those occurring in the stretching position, only inferior
glide showed a significantly larger strain than that
occurring in the relaxing position (Po0:01). There were
no significant differences in the strains between other
techniques and the relaxing position.
3.4. Gap distances
The gap distances of the repaired site during joint
mobilization techniques at each abduction angle are
listed in Table 1. Gaps were formed in the stretching
position (4.47 mm) and during all joint mobiliza-
tion techniques at 01 of abduction (1.06–1.46 mm).
Conversely, no gap was observed in the relaxing
position and all joint mobilization techniques at 301 of
abduction.
4. Discussion
Application of joint mobilization techniques immedi-
ately after rotator cuff repair has not been clarified.
Quillen et al. (1992) considered the stress on immature
repaired tissues as a contraindication of joint mobiliza-
tion. This indicated that joint mobilization should be
prohibited immediately after surgery if it stresses such
tissues. Bruzga and Speer (1999) recommended joint
mobilization techniques, which are performed at grade 1
or 2, after surgery if these would be useful in reducing
pain and promoting joint nutrition. However, they also
stated that joint mobilization at grade 3 or 4 should be
used only after the healing of repaired tissue. Based on
these opinions, joint mobilization techniques, which
exert stress on the repaired tissue, should be avoided
immediately after rotator cuff repair. On the other hand,
based on the concept of joint mobilization, it is
hypothesized that this technique can stretch specific
237
tissues and can reduce the stress on tissues that should
not be stretched (Conroy and Hayes, 1998). Accord-
ingly, while stress on the repaired supraspinatus tendon
can be avoided, a particular part of the capsule, which is
responsible for joint contracture, may be stretched by
joint mobilization techniques. In order to resolve this
repaired supraspinatus tendons were measured to
estimate the stress undergone by these during joint
mobilization. For ideal repair, minimization of gap
formation at the repair site and maintenance of
mechanical stability of the repaired tendon until solid
healing occurs are important (Gerber et al., 1994).
Therefore, the criteria for safely performing joint
mobilization are that (1) no gap is formed, and (2) the
repaired tendon relaxes more than in the neutral
position in which the repair was performed.
A gap was formed in the stretching position (4.5 mm)
and in all joint mobilization techniques at 01 abduction
(1.1–1.5 mm) in this study. In comparison with a 10-mm
gap distance, which is regarded as complete gap
formation (Burkhart et al., 1997a, b), joint mobilization
techniques at 01 abduction corresponded to 11–15%,
while the stretching position was 45%. These techniques
have the risk of increasing gap formation and can lead
to the failure of healing (Burkhart et al., 1998; Gerber
et al., 1994, 1999). Therefore, joint mobilization at 01
abduction should be avoided immediately after tendon
repair.
On the other hand, during joint mobilization at 301
abduction, strains were negative compared to neutral
position and no gaps were observed. The supraspinatus
tendon relaxes after abducting the arm beyond 301. By
using three-dimensional analysis using magnetic reso-
nance imaging, Nakajima et al. (2004) observed that
intact supraspinatus tendons were relaxed beyond 301
arm abduction. Zuckerman et al. (1991) demonstrated
that the strain on repaired tendons, with both small and
large tears, remained small above 301 abduction
irrespective of the position of flexion/extension or
rotation. Hatakeyama et al. (2001a) determined strain
on repaired supraspinatus tendons under the same
repair condition as ours. They concluded that arm
abductions above 301 in the scapular plane seemed to be
safe even immediately after the repair, because the strain
decreased in these positions and the estimated tensile
forces were less 0.5 kg. According to the estimation from
their data, tensile forces caused by joint mobilization at
301 were 0–0.5 kg, compared to 3 kg in the neutral
position. Therefore, joint mobilization techniques can be
safely performed keeping the tendon relaxed and not
forming a gap at the repair site if the torn tendon is
repaired at 01.
 ARTICLE IN PRESS
238 T. Muraki et al. / Manual Therapy 12 (2007) 231–239
However, repetitive joint mobilization at 301 abduc-
tion might lead to gap formation on the repair site.
The repaired tendon has less endurance to mechanical
stress than the intact tendon because these two have
different mechanical properties. The failure load of
the repaired tendon was reported to be 72–605 N
(Rossouw et al., 1997; Hatakeyama et al., 2001a; Ma
et al., 2004), while that of the intact tendon was 600–
800 N (Itoi et al., 1995). In addition, cyclic loading
might lead to gap formation even if the mechanical
properties of the repair site are strong, or the cyclic
stress is small (Rossouw et al., 1997; Petit et al., 2003).
Hatakeyama et al. (2001a) concluded that internal
rotation at 301 abduction should be postponed until
the rotator cuff healing progresses, because the strain
and tensile force on the repair site during internal
rotation significantly increased from neutral rotation.
In this study, the strain during inferior glide was
significantly larger than that occurring in the relaxing
position. Therefore, of these four joint mobiliza-
tion techniques, repeated use of the inferior glide
technique should be avoided immediately after rotator
cuff repair.
Strain on the supraspinatus tendon is commonly
measured by a small linear displacement sensor. The
sensor, ‘‘Pulse corder’’ in this study, is capable of
directly measuring tendon behaviour along with its fibre
orientation during joint mobilization, while measure-
ments by imaging techniques are limited. Because of
distorted three-dimensional movement of imaging mar-
kers buried in the cuff tendons during shoulder motion,
accurate detection of strain between each marker
becomes extremely difficult. In addition, the sensor
could accurately measure the strain with little resistance
because preliminary calibration of the Pulse Coder,
which was attached in the supraspinatus tendon,
demonstrated the high accuracy of 0.1 mm RMS.
There is concern that our experiment was performed
on a cadaveric model. Since muscle tension in fresh
cadavers is different from that of in vivo and some soft
tissues were removed, the stress on the tendon due to
muscle tension might also differ between them. How-
ever, the strain on the tendon in fresh cadavers might be
the same or smaller than that of in vivo because
physiologic muscle tone and muscle contraction stabilize
the humeral head and decrease its movement (Warner
et al., 1999).
This study had a few limitations. First, the strain
obtained from this study was not always consistent
because the specimens were obtained from aged
cadavers. The supraspinatus tendon in younger adults
has great endurance to stress because the connective
tissues in adults are more flexible and have greater
failure load than that in the aged cadavers (Reeves,
1968). Therefore, we believe that our findings are
applicable to younger to middle-aged adults.
Second, the safety of joint mobilization could not be
determined with regard to the strain–stress curve
and endurance to repetitive loading because tensile
force to the tendon was not measured. In this
study, instead, the strains and gap distances in the
stretching and relaxing positions were observed and
tensile forces on the repaired tendon were estimated
from the data in a previous study (Hatakeyama et al.,
2001a).
Finally, our finding can be applied only to rotator cuff
tears that are of the same size as or smaller than those
studied here. Strain on a massive tear including the
infraspinatus and/or subscapularis tendons during joint
mobilization may be different from the results of this
study. Therefore, further studies are required to clarify
the strain on a massive tear.
Further research following our study should deter-
mine whether the joint mobilization techniques could
prevent joint contracture after rotator cuff repair.
Investigation of strain and stress on the contracted joint
capsule during mobilization techniques are necessary. In
addition, the effect of repetitive joint mobilization on
the repaired tendon and capsule should be also
determined. These studies will contribute to the making
of good decisions concerning the application of joint
mobilization after rotator cuff repair.
5. Conclusion
Our findings suggested that if rotator cuff repair is
performed at 01 of abduction, joint mobilization
techniques at 01 of arm abduction should be avoided
immediately after supraspinatus tendon repair because
these techniques produce large strain of the tendon
and form a gap at the repair site. In the same fashion,
our findings also suggest that joint mobilization
techniques at 301 of arm abduction can be used without
large strain of the tendon and forming a gap at the
repair site. However, inferior glide, even at 301 abduc-
tion, should be postponed because relatively larger
strain than that of the relaxing position may lead to
failure of tendon repair. Although our findings would
be useful to decide the application of joint mobiliza-
tion after rotator cuff repair, further study regarding
strain on the contracted joint capsule or the effect of
repetitive joint mobilization would provide even more
information.
Acknowledgments
The authors would like to thank Daisuke Suzuki
Ph.D. for his technical assistance. In addition, we would
like to thank Shuhei Takauji and Mitsuo Nakamura for
their assistance.
 ARTICLE IN PRESS
T. Muraki et al. / Manual Therapy 12 (2007) 231–239
References
An KN, Jacobsen MC, Berglund LJ, Chao EY. Application of a
magnetic tracking device to kinesiologic studies. Journal of
Biomechanics 1988;21(7):613–20.
Bruzga B, Speer K. Challenges of rehabilitation after shoulder surgery.
Clinics in Sports Medicine 1999;18(4):769–93.
Burkhart SS, Johnson TC, Wirth MA, Athanasiou KA. Cyclic loading
of transosseous rotator cuff repairs: tension overload as a possible
cause of failure. Arthroscopy 1997a;13(2):172–6.
Burkhart SS, Diaz Pagan JL, Wirth MA, Athanasiou KA. Cyclic
loading of anchor-based rotator cuff repairs: confirmation of the
tension overload phenomenon and comparison of suture anchor
fixation with transosseous fixation. Arthroscopy 1997b;13(6):720–4.
Burkhart SS, Wirth MA, Simonick M, Salem D, Lanctot D,
Athanasiou K. Loop security as a determinant of tissue fixation
security. Arthroscopy 1998;14(7):773–6.
Cofield RH. Rotator cuff disease of the shoulder. Journal of Bone and
Joint Surgery 1985;67A(6):974–9.
Conroy DE, Hayes KW. The effect of joint mobilization as a
component of comprehensive treatment for primary shoulder
impingement syndrome. Journal of Orthopaedic Sports and
Physical Therapy 1998;28(1):3–14.
Culham E, Peat M. Functional anatomy of shoulder complex. Journal
of Orthopaedic Sports and Physical Therapy 1993;18(1):342–50.
Evjenth O, Hamberg J. Muscle stretching in manual therapy: a clinical
manual. The extremities, vol. 1. Alfta, Sweden: Alfta Rehab
Forlag; 1984 p. 40 [Chapter 2].
Flatow EL, Soslowsky LJ, Ticker JB, Pawluk RJ, Hepler M, Ark J,
et al. Excursion of the rotator cuff under the acromion: patterns of
subacromial contact. The American Journal of Sports Medicine
1994;22(6):779–88.
Gerber C, Schneeberger AG, Beck M, Schlegel U. Mechanical strength
of repairs of the rotator cuff. Journal of Bone and Joint Surgery
1994;76B(3):371–80.
Gerber C, Schneeberger AG, Perren SM, Nyffeler RW. Experimental
rotator cuff repair: a preliminary study. Journal of Bone and Joint
Surgery 1999;81A(9):1281–90.
Gokeler A, Paridon-Edauw GH, DeClercq S, Matthijs O, Dijkstra PU.
Quantitative analysis of traction in the glenohumeral joint: in vivo
radiographic measurement. Manual Therapy 2003;8(2):97–102.
Harryman 2nd DT, Sidles JA, Clark JM, McQuade KJ, Gibb TD,
Matsen 3rd FA. Translation of the humeral head on the glenoid
with passive glenohumeral motion. Journal of Bone and Joint
Surgery 1990;72-A(9):1334–43.
Hatakeyama Y, Itoi E, Pradhan RL, Urayama M, Sato K. Effect of
arm elevation and rotation on the strain in the repaired rotator cuff
tendon: a cadaveric study. The American journal of Sports
Medicine 2001a;29(6):788–94.
Hatakeyama Y, Itoi E, Urayama M, Pradhan RL, Sato K. Effect of
superior capsule and coracohumeral ligament release on strain in
the repaired rotator cuff tendon: a cadaveric study. The American
Journal of Sports Medicine 2001b;29(5):633–40.
Hjelm R, Draper C, Spencer S. Anterior-inferior capsular length
insufficiency in the painful shoulder. Journal of Orthopaedic Sports
and Physical Therapy 1996;23(3):216–22.
Hsu AT, Ho L, Ho S, Hedman T. Immediate response of
glenohumeral abduction range of motion to a caudally directed
translational mobilization: a fresh cadaver simulation. Archives of
Physical Medicine and Rehabilitation 2000a;81(11):1511–6.
Hsu AT, Ho L, Ho S, Hedman T. Joint position during anterior–
posterior glide mobilization: its effect on glenohumeral abduction
range of motion. Archives of Physical Medicine and Rehabilitation
2000b;81(2):210–4.
Hsu AT, Hedman T, Chang JH, Vo C, Ho L, Ho S, et al. Changes in
abduction and rotation range of motion in response to simulated
239
dorsal and ventral translational mobilization of the glenohumeral
joint. Physical Therapy 2002a;82(6):544–56.
Hsu AT, Ho L, Chang JH, Chang GL, Hedman T. Characterization of
tissue resistance during a dorsally directed translational mobiliza-
tion of the glenohumeral joint. Archives of Physical Medicine and
Rehabilitation 2002b;83(3):360–6.
Itoi E, Berglund LJ, Grabowski JJ, Schultz FM, Growney ES, Morrey
BF, et al. Tensile properties of the supraspinatus tendon. Journal of
Orthopaedic Research 1995;13(4):578–84.
Itoi E, Morrey BF, An KN. Biomechanics of the shoulder. In:
Rockwood Jr CA, Matsen 3rd FA, Wirth MA, Lippitt SB, editors.
The Shoulder. 3rd ed. Philadelphia, PA: WB Saunders; 2004.
p. 223–67 Chapter 6.
Johns RJ, Wright V. Relative importance of various tissues in joint
stiffness. Journal of Applied Physiology 1962;17(5):824–30.
Kaltenborn FM. Manual mobilization of the joints: the Kaltenborn
method of joint examination and treatment. The extremities, 5th
ed. Oslo, Norway: Olaf Norlis bokhandel; 1999 p. 21–8 [Chapter 2].
Landis JR, Koch GG. The measurement of observer agreement for
categorical data. Biometrics 1977;33(2):159–74.
Ma CB, MacGillivray JD, Clabeaux J, Lee S, Otis JC. Biomechanical
evaluation of arthroscopic rotator cuff stitches. Journal of Bone
and Joint Surgery 2004;86A(6):1211–6.
Mangus BC, Hoffman LA, Hoffman MA, Altenburger P. Basic
principles of extremity joint mobilization using a Kaltenborn
approach. Journal of Sport Rehabilitation 2002;11(4):235–50.
Matsen 3rd FA, Titelman RM, Lippitt SB, Wirth MA, Rockwood Jr.
CA. Rotator cuff. In: Rockwood Jr. CA, Matsen 3rd FA, Wirth
MA, Lippitt SB, editors. The shoulder. 3rd ed. Philadelphia, PA:
WB Saunders; 2004. p. 795–878 Chapter 15.
Nakajima T, Hughes RE, An KN. Effects of glenohumeral rotations
and translations on supraspinatus tendon morphology. Clinical
Biomechanics 2004;19(6):579–85.
Petit CJ, Boswell R, Mahar A, Tasto J, Pedowitz RA. Biomechanical
evaluation of a new technique for rotator cuff repair. The
American Journal of Sports Medicine 2003;31(6):849–53.
Pruitt DL, Manske PR, Fink B. Cyclic stress analysis of flexor tendon
repair. Journal of Hand Surgery 1991;16A(4):701–7.
Quillen WS, Halle JS, Rouillier LH. Manual therapy: mobilization of
the motion-restricted shoulder. Journal of Sport Rehabilitation
1992;1(3):237–48.
Reeves B. Experiments on the tensile strength of the anterior capsular
structures of the shoulder in man. Journal of Bone and Joint
Surgery 1968;50B(4):858–65.
Rossouw DJ, McElroy BJ, Amis AA, Emery RJ. A biomechanical
evaluation of suture anchors in repair of the rotator cuff. Journal
of Bone and Joint Surgery 1997;79B(3):458–61.
Tibone JE, McMahon PJ, Shrader TA, Sandusky MD, Lee TQ.
Glenohumeral joint translation after arthroscopic, nonablative,
thermal capsuloplasty with a laser. The American Journal of Sports
Medicine 1998;26(4):495–8.
Warner JJ, Greis PE. The treatment of stiffness of the shoulder after repair
of the rotator cuff. Instruction Course Lecture 1998;47(8):67–75.
Warner JJ, Micheli LJ, Arslanian LE, Kennedy J, Kennedy R.
Patterns of flexibility, laxity, and strength in normal shoulders and
shoulders with instability and impingement. The American Journal
of Sports Medicine 1990;18(4):366–75.
Warner JJ, Deng XH, Warren RF, Torzilli PA. Static capsuloligamentous
restraints to superior–inferior translation of the glenohumeral joint.
The American Journal of Sports Medicine 1992;20(6):675–85.
Warner JJ, Bowen MK, Deng X, Torzilli PA, Warren RF. Effect of
joint compression on inferior stability of the glenohumeral joint.
Journal of Shoulder and Elbow Surgery 1999;8(1):31–6.
Zuckerman JD, Leblanc JM, Choueka J, Kummer F. The effect of arm
position and capsular release on rotator cuff repair: a biomecha-
nical study. Journal of Bone and Joint Surgery 1991;73B(3):402–5.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 240–248

www.elsevier.com/locate/math

Original article

Changes in postural activity of the trunk muscles following spinal

manipulative therapy
Manuela L. Ferreiraa,b,c, Paulo H. Ferreiraa,b,d, Paul W. Hodgesa,
a
Division of Physiotherapy, The University of Queensland, Brisbane QLD 4072, Australia
b
School of Physiotherapy, The University of Sydney, Sydney, Australia
c
Departamento de Fisioterapia, Pontifı´cie Universidade Católica de Minas Gerais, Belo Horizonte, Brazil
d
Departamento de Fisioterapia, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil

Received 6 May 2005; received in revised form 10 May 2006; accepted 27 June 2006

Abstract

Spinal manipulative therapy (SMT) is common in the management of low back pain (LBP) and has been associated with changes
in muscle activity, but evidence is conflicting. This study investigated the effect of SMT on trunk muscle activity in postural tasks in
people with and without LBP. In 20 subjects (10 with LBP and 10 controls), EMG recordings were made with fine-wire electrodes
inserted into transversus (TrA), obliquus internus (OI), and externus (OE) abdominis. Rectus abdominis (RA) and anterior deltoid
EMG was recorded with surface electrodes. Standing subjects rapidly flexed an arm in response to a light, before and after a small
amplitude end range rotational lumbar mobilization at L4-5. In controls, there was no change in trunk muscle EMG during the
postural perturbation after SMT. In LBP subjects there was an increase in the postural response of OI and an overall increase in OE
EMG. There was no change in TrA or RA EMG. These results indicate that SMT changes the functional activity of trunk muscles in
people with LBP, but has no effect on control subjects. Importantly, SMT increased the activity of the oblique abdominal muscles
with no change in the deep trunk muscle TrA, which is often the target of exercise interventions.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Spinal manipulative therapy; Motor control; Transversus abdominis; Fine-wire EMG

1. Introduction Several possible mechanisms have been discussed in

Spinal manipulative therapy (SMT), defined as
manual loading of the spine using short or long leverage
methods, is one of the most common approaches in the
management of low back pain (LBP). The efficacy of
SMT on clinical outcome measures for people with LBP
has been investigated in several systematic reviews which
conclude that it leads to clinically significant improve-
ments in pain and function (van Tulder et al., 1997;
Ferreira et al., 2002; Assendelft et al., 2003). Despite the
positive clinical benefit, the physiological mechanisms
responsible for these effects are still unclear.
Corresponding author. Tel.: +61 7 3365 2008;
fax: +61 7 3365 2775.
E-mail address: p.hodges@shrs.uq.edu.au (P.W. Hodges).
the literature. For instance, SMT has been associated
with sympathoexcitatory effects (Vicenzino et al., 1998),
changes in passive and active spinal range of motion
(Nilsson et al., 1996; Lehman and McGill, 2001), and
effects such as enhanced production of tumour necrosis
factor and substance P (Brennan et al., 1992). Alter-
natively, it has been argued that SMT changes muscle
activity. However, few studies have investigated the
effects on parameters of muscle activation and results
are contradictory. For instance increased (Herzog et al.,
1999) and decreased (Dishman and Bulbulian, 2000;
Dishman and Bulbulian, 2001) activity of the paraspinal
muscles have been reported in response to SMT.
Several factors may explain the inconsistency of the
results. First, studies have used a variety of manipulative
techniques. The term SMT is used to describe a broad

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.015
 ARTICLE IN PRESS
M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248
spectrum of techniques that may or may not include
audible cavitation. Second, a range of muscle functions,
from resting activity to reflex amplitude have been
tested. Finally, most studies have involved people
without LBP. Thus, from the present data it is not
possible to determine whether manipulation modifies
the activity of the trunk muscles in people with LBP
during functional activities.
One functional task that has been evaluated exten-
sively is the postural response of the trunk muscles to a
rapid limb movement. In this task trunk muscle activity
is initiated prior to movement to prepare the spine for
the perturbation from limb movement (Belenkii et al.,
1967; Aruin and Latash, 1995; Hodges and Richardson,
1997). This task provides an ideal model to evaluate the
effect of SMT as it provides a measure of the pre-
planned strategy used by the central nervous system to
control the trunk muscles. Furthermore, people with
recurrent LBP have been found to have changes in this
pre-planned postural adjustment. Notably, activity of
transversus abdominis (TrA), the deepest abdominal
muscles, is delayed (Hodges and Richardson, 1996;
Hodges and Richardson, 1998). Although impaired
activity of this muscle is a relatively consistent finding
in LBP, activity of superficial trunk muscles is often
increased (Arendt-Nielsen et al., 1996; Radebold et al.,
2000; Hodges et al., 2003a, b).
We hypothesized that manipulative therapy would
change the response of the trunk muscles, but on the
basis of previous data it was not possible to predict
whether activity would be increased or decreased. Thus,
the aims of the present study were, first, to determine
whether the pre-planned postural activity of the trunk
muscles could be modified by SMT and second, to
investigate whether the effect differed between people
with and without LBP.
2. Materials and methods
2.1. Subjects
Twenty subjects (10 with a history of LBP and 10
controls) volunteered for this experiment. The subject
demographics are presented in Table 1 and were not
different between the two groups. Subjects were
excluded if they had any respiratory or neurological
disorder, musculoskeletal pain elsewhere in the spine or
lower limbs or if they had been pregnant in the previous
2 years. Subjects in the control group were also excluded
if they had a history of LBP that had restricted function
or for which they had sought medical or allied health
intervention.
To be included in the LBP group, volunteers needed
to have a history of at least one episode of LBP that had
limited function or work in the past 18 months and had
241
Table 1
Demographic details
Control group LBP group P-valuea
mean (SD) mean (SD)
Age (years) 33 (11) 28 (5) P ¼ 0.21
Height (cm) 159 (38) 171 (10) P ¼ 0.35
Weight (kg) 68 (13) 69 (13) P ¼ 0.95
LBP, low back pain.
a
P-values refers to t-test for independent samples.
an episode of LBP within the past 6 months. Patients
were excluded if they had neurological signs, specific
spinal pathology (e.g. malignancy, inflammatory joint or
bone disease), if they had undergone back surgery in the
past 12 months, or if they could not tolerate a grade IV
lumbar rotary mobilization technique (no subject was
excluded on the basis of this criteria). The mean (SD)
pain intensity (past 7 days) was 3.2 (2.3) on a 10 cm
visual analogue scale (VAS) and the disability score
(Roland Morris) was 3.2 (1.8). All included patients
presented with LBP of at least 3 months duration.
The study was approved by the Institutional Medical
Research Ethics Committee and all procedures were
conducted in accordance with the declaration of
Helsinki.
2.2. Electromyography
EMG recordings were made with surface and
intramuscular fine-wire electrodes. Fine-wire electrodes
were fabricated from two strands of Teflon-coated
stainless-steel wire (75 mm diameter, A-M systems,
USA) threaded into a hypodermic needle
(0.6  32 mm) and inserted with guidance from ultra-
sound imaging into the right ventro-lateral abdominal
wall muscles: TrA, obliquus internus abdominis (OI),
and obliquus externus abdominis (OE) half-way be-
tween the iliac crest and distal border of the rib cage in
the anterior axillary line (Hodges et al., 1999). Surface
electrodes were placed on left anterior deltoid muscle
approximately in parallel with the muscle fibres, and
over the muscle belly of the right rectus abdominis (RA).
EMG data were amplified 2000 times, band-pass filtered
between 20 and 1 kHz (Neurolong, Digitimer, UK) and
sampled at 2 kHz using a Power-1401 and Spike-2
software (Cambridge Electronic Design, UK).
2.3. Procedure
Subjects rapidly flexed or extended the left upper limb
in response to visual command in a choice reaction time
task. Two directions of movement were performed to
limit the predictability of the response, but only the
flexion data was analysed. Ten repetitions in each
direction were performed in random order before and
 ARTICLE IN PRESS
242 M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248
after the spinal manipulative technique (see below). The
second set of ten arm movements commenced immedi-
ately after the cessation of the manipulative technique.
2.4. Spinal manipulative technique
A specific small amplitude rotational end-range
(grade IV) oscillatory mobilization technique was
directed to at L4–L5 with subjects in left side-lying
(Maitland et al., 2001). This procedure was standardized
for all patients because of the placement of electrodes on
the right side of the trunk. Although it is unlikely that
L4–L5 was the symptomatic level in all patients we
considered this technique to be appropriate to test the
aims of the experiment for a number of reasons. First,
although attempts are made to focus the technique to a
target segment, movement will occur over a number of
levels and second, recent data suggest that that patients
report significant decreases in pain even when mobiliz-
ing non-painful lumbar segments (Chiradejnant et al.,
2003). The technique was repeated three times for 30 s,
with oscillations at 1 Hz, with feedback from a
metronome. The mobilization technique was performed
at the same spinal level (L4–L5) for all subjects, even
though this might not have been the level of pain in
some subjects. Pain intensity was measured on a VAS
before and after the application of the technique.
2.5. Data analysis
In order to investigate both temporal and spatial
aspects of EMG, root mean square (RMS) EMG
amplitude was calculated during four 25 ms epochs
before the onset of deltoid EMG and four 25 ms epochs
after the onset of deltoid EMG. This analysis technique
measures the postural response without problems
associated with detection of EMG onset of the trunk
muscles. EMG onset detection was difficult due to the
high degree of baseline activity, particularly in OE after
the performance of the manipulative technique. Data
were normalized to the epoch with the greatest
amplitude in the pre-manipulation condition. This
normalization method provides a high sensitivity to
compare the pre- and post-manipulation conditions for
each group, but does not permit comparison of EMG
amplitude between muscles or between groups, although
the pattern of changes in pattern of activity could be
compared between groups. Data were not normalized to
EMG amplitude in a maximum voluntary contraction as
these values are not possible to obtain reliably in people
with LBP (Allison et al., 1998) and data were not
normalized to a submaximal task as people with LBP
are likely to use an abnormal strategy during submax-
imal tasks, making it an invalid reference. Furthermore,
recent data suggests that normalization to a submaximal
task may increase variability (Urquhart et al., 2005).
However, as this study was a repeated measures design
with each subject acting as their own controls normal-
ization would not affect the results, but normalization to
pre-manipulation peak amplitude normalizes all data to
a similar scale for graphical representation of the data.
Statistical analysis involved a three-way repeated
measures analysis of variance (ANOVA) with two
repeated measures factors (pre- and post-mobilization
and epoch) and one independent factor (group). Post-
hoc Duncan tests were used when appropriate. Due to
the inability to compare between muscles as a result of
the normalization procedure analyses were run sepa-
rately for each muscle (RA, OE, OI and TrA). The alpha
level was set at 0.05.
3. Results
3.1. TrA
When subjects rapidly moved the left arm, a burst of
EMG activity of TrA occurred in association with
deltoid EMG. There was a trend for the activity to occur
in an earlier epoch for TrA in the control group
compared to the LBP group. Fig. 1 shows an increase
in EMG activity greater than 10% above baseline EMG
in epochs 3 and 4 in the control group. This increase is
not as evident in the LBP group until epoch 5. When
arm movement was performed after the application of
spinal mobilization there was no change in the activity
of TrA in the control or LBP subjects compared to the
pre-mobilization condition (Fig. 1). Table 2 presents the
main effects and interactions.
3.2. OI
Similar to TrA, a burst of OI EMG activity occurred
in conjunction with arm movement. In both groups the
increase in EMG activity tended to start in the epoch
prior to the onset of deltoid EMG. Following the
mobilization technique there was no change in the
postural response for the control group (P ¼ 0:20).
Main effects and interactions are shown in Table 2.
Although there was no consistent effect for the control
group, OI EMG was increased after mobilization in 2
subjects. This increased the variability of the response
(Fig. 2). In contrast, OI EMG activity was increased
following the onset of deltoid EMG (Epochs E5: 0–25,
E6: 25–50, E7: 50–75, E8: 75–100 ms) for subjects in the
LBP group (all: Po0:001). Notably, mobilization did
not increase the background activity prior to arm
movement, only the postural adjustment associated with
the limb movement. Variability was increased in this
group. Three subjects had an accentuated increase in
muscle activity, and two subjects had a decrease in
EMG activity in epochs 1–3.
 ARTICLE IN PRESS
M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248 243

LBP Control subjects

Pre-mobilisation
1.4
Post-mobilisation

Prop. pre-mobilisation EMG

1.2

1.0

0.8

0.6

0.4

0.2

0
E1 E2 E3 E4 E5 E6 E7 E8 E1 E2 E3 E4 E5 E6 E7 E8

Fig. 1. Change in TrA EMG with SMT. EMG amplitude is shown for four 25-ms epochs prior to onset of deltoid EMG (E1–E4) and four 25-ms
epochs after the onset of deltoid (E5–E8). The dashed line divides the epochs before and after the onset of deltoid EMG. EMG is presented as a
proportion of peak activity recorded across the eight epochs during the pre-mobilization trial. Confidence intervals (95%) are shown.

Table 2
Main effects and interactions for comparison between groups, between epochs and pre- and post-manipulative technique (Manip)

Main effect/interaction TrA OI OE RA

F P F P F P F P

Epoch 5.27 o0.01* 2.44 0.02* 0.5 0.83 1.03 0.42

Group 0.32 0.57 5.07 0.03* 29.9 o0.01* 10.88 o0.01*
Epochgoup 0.28 0.96 0.51 0.82 0.07 0.99 0.42 0.89
Manip 0.40 0.53 16.81 o0.01* 11.99 o0.01* 2.61 0.11
Manipepoch 0.17 0.99 0.45 0.87 0.17 0.99 0.02 1.00
Manipgroup 3.33 0.07 4.73 0.03* 16.25 o0.01* 1.17 0.28
Manipepochgroup 0.23 0.98 0.59 0.76 0.25 0.97 0.11 1.00

*po0.05.

LBP Control subjects

3.0 Pre-mobilisation
Post-mobilisation
2.5
Prop. pre-mobilisation EMG

2.0

1.5
*
1.0 * * *
0.5

0.0

-0.5
E1 E2 E3 E4 E5 E6 E7 E8 E1 E2 E3 E4 E5 E6 E7 E8

Fig. 2. Change in OI EMG with SMT. EMG amplitude is shown for four 25-ms epochs prior to onset of deltoid (E1–E4) and four after the onset of
deltoid. The dashed line divides the epochs before and after the onset of deltoid EMG. EMG is presented as a proportion of peak activity recorded
across the eight epochs during the pre-mobilization trial. Confidence intervals (95%) are shown. An increase in EMG activity was observed post-
manipulation, after the onset of deltoid in the LBP group. *Po0.001.
 ARTICLE IN PRESS
244 M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248

3.3. OE 3.4. RA

When subjects with and without LBP flexed the left EMG activity for RA was increased for both groups
upper limb in response to a visual stimulus, there was an prior to the onset of deltoid activity (E4: -25 to 0). No
increase OE EMG. There was a tendency for this burst change EMG was observed after the mobilization in the
to initiate prior to the onset of deltoid EMG in both control subjects. Although not significant, there was a
groups. After the application of SMT, no change in OE trend for RA activity to increase in people with LBP
activity was observed as a consequence of arm move- after manipulation (Fig. 4). Main effects and interac-
ment for the control subjects (P ¼ 0:68). Main effects tions are shown in Table 2.
and interactions are shown in Table 2. In two subjects
there was a decrease in OE EMG in epochs 2–5 after the
mobilization, which explains the increase in variability 3.5. Pain
in this group. In contrast, when subjects with LBP flexed
their left arm after the mobilization, an overall increase Average pain intensity immediately before the
in muscle activity was observed for all epochs (Po0:01). manipulation was 1.1 (1.7) and immediately
That is, OE EMG was increased even prior to arm after the procedure was 0.7 (2.2). There was no
movement. Likewise, the group with LBP presented significant change in pain intensity following SMT
greater variability after the mobilization (Fig. 3). (P ¼ 0:34).

LBP Control subjects

3.0 Pre-mobilisation
Post-mobilisation
2.5
Prop. pre-mobilisation EMG

2.0
1.5
1.0
* * * *
* * * *
0.5
0.0
-0.5
-1.0
E1 E2 E3 E4 E5 E6 E7 E8 E1 E2 E3 E4 E5 E6 E7 E8

Fig. 3. Change in OE EMG with SMT. EMG amplitude is shown for four 25-ms epochs prior to onset of deltoid (E1–E4) and four after the onset of
deltoid. The dashed line divides the epochs before and after the onset of deltoid EMG. EMG is presented as a proportion of peak activity recorded
across the eight epochs during the pre-mobilization trial. Confidence intervals (95%) are shown. An overall increase in EMG activity was observed
post-manipulation only in the LBP group. *Po0.001.

LBP Control subjects

Pre-mobilisation
1.75
Post-mobilisation
Prop. pre-mobilisation EMG

1.5

1.25

1.0

0.75

0.5

0.25

0
E1 E2 E3 E4 E5 E6 E7 E8 E1 E2 E3 E4 E5 E6 E7 E8

Fig. 4. Change in RA EMG with SMT. EMG amplitude is shown for four 25-ms epochs prior to onset of deltoid (E1–E4) and four after the onset of
deltoid. The dashed line divides the epochs before and after the onset of deltoid EMG. EMG is presented as a proportion of peak activity recorded
across the eight epochs during the pre-mobilization trial. Confidence intervals (95%) are shown.
 ARTICLE IN PRESS
M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248
4. Discussion
The results of this study demonstrate that pre-planned
postural activity of the trunk muscles can be modified by
SMT performed as a small amplitude oscillation without
cavitation. This was evidenced by the increased ampli-
tude of OE and OI EMG as a component of postural
adjustment associated with arm movement after SMT.
However, this change was only observed in people with
LBP. In contrast to the superficial oblique abdominal
muscles, SMT did not affect the postural response of
TrA in either group.
There are several possible mechanisms for the increase
in EMG activity following SMT that are related to the
effect of manual techniques on motoneuron excitability
via direct effects of stimulation of joint and muscle
afferents, the effect of manual techniques on pain, and
other neurophysiological mechanisms. However, there
has been considerable debate and findings are often
inconsistent.
One extensively argued possibility is that excitability
of spinal muscle motoneurons is changed by afferent
input from stimulation of mechanoreceptors in the joint
capsules (Indahl et al., 1997), ligaments (Indahl et al.,
1997), muscles (Herzog et al., 1999) and cutaneous
receptors (Herzog et al., 1999). However, some studies
report increased excitability (Herzog et al., 1999; Keller
and Colloca, 2000), while others report decreased
activity (Dishman and Bulbulian, 2000; Lehman and
McGill, 2001; Lehman et al., 2001). These discrepancies
may be explained to some extent by the methods that
have been used to evaluate muscle activity. For instance,
Keller and Colloca (2000) reported increased erector
spinae activity during trunk extension in prone, Herzog
et al. (1999) reported short latency excitatory muscle
activity in response to manipulative thrusts, whereas
Dishman and Bulbulian (2000) reported reduced Hoff-
man-reflex (H-reflex) amplitude in the gastrocnemius
muscle in the leg, and Lehman and McGill (2001)
reported decreased muscle response to a painful
mechanical stimulus applied over the spinous process.
Thus, few studies have measured the same parameter
and it is difficult to predict from existing data how trunk
muscle activity would change during a postural adjust-
ment in association with arm movement.
Short latency responses in paraspinal muscles to
manipulative and mobilization techniques have been
extensively reported (Herzog et al., 1999; Colloca and
Keller, 2001; Shirley et al., 2002). These responses have
been argued to be mediated by stimulation of muscle
spindles and have short duration (100–400 ms) (Herzog
et al., 1999). However, it is unclear whether these
responses are associated with changes in ongoing EMG
activity, although ongoing reduction in abdominal
muscle activity has been identified in a single case report
by Herzog et al. (1999) following initial excitation.
245
Similarly it is difficult to interpret data regarding
amplitude of H-reflexes in lower leg muscles. Although
these responses have been shown to be reduced (Dish-
man and Bulbulian, 2000), and it is argued that the
motoneurons that innervate these muscles lie in the
lumbosacral spinal cord, it is unlikely that these
responses would reflect the properties of the distinct
motoneurons to the spinal muscles. Furthermore,
H-reflex amplitude is influenced by pre-synaptic effects
and may not reflect changes in motoneuron excitability
(Rudomin, 2002). Finally, the response of the paraspinal
muscles to a mechanical stimulus to the segment may
reflect changes in local reflex mechanisms and not
functional activity of the trunk muscles.
Studies that have investigated trunk muscle activity
during trunk movements also have divergent results.
Keller and Colloca (2000) reported increased activity
during back extension. In contrast, Lehman and McGill
(2001) found no consistent change in muscle activity
during trunk movements in standing, but reported a
tendency for decreased activity in some muscles in some
subjects. Consistent with the data of Keller and Colloca
(2000), we identified a significant increase in activity of
OE and OI, and a trend for increased RA EMG. The
task used in the present study involved evaluation of the
activity of the trunk muscles in association with arm
movement. These responses are considered to be pre-
programmed by the nervous system as they are initiated
in advance of limb movement (Belenkii et al., 1967;
Bouisset and Zattara, 1981). As such, they reflect the
strategy used by the nervous system to prepare the body
for the perturbation resulting from the movement
(Belenkii et al., 1967; Hodges et al., 1999). Notably,
temporal and spatial parameters of the anticipatory
postural adjustments are matched to the timing,
direction and amplitude of the perturbation from limb
movement (Bouisset and Zattara, 1981; Aruin and
Latash, 1995; Hodges and Richardson, 1997; Hodges
et al., 1999). Thus, changes in EMG amplitude could
reflect either changes in motoneuron excitability (i.e.
larger response initiated for the same descending drive),
or changes in descending drive (i.e. increased output
from higher centres). In terms of modification of
descending drive to the motoneuron pool, if SMT
modifies the afferent discharge from the peripheral
mechanoreceptors this may modify the descending
postural response if input regarding the status of the
spine was changed. It has been shown that excitation of
1a afferents by mechanical stimulation such as vibration
modifies the perception of the position of the spine
(Brumagne et al., 1999) and leads to modified postural
responses (Kasai et al., 2002). Although the present data
do not allow differentiation between spinal or suprasp-
inal mechanisms, the data do indicate that the postural
response of the trunk muscles is modified by SMT.
Importantly; the data indicate that this is not a general
 ARTICLE IN PRESS
246 M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248
response, as not all muscles responded in a similar
manner. Why no change was identified in the response
of TrA following SMT is unclear. However, the function
of this muscle has been shown to be impaired in people
with LBP (Hodges and Richardson, 1996; Hodges and
Richardson, 1998), and this may not be resolved by the
mechanical stimulus.
This study evaluated a rotational spinal mobilization
without thrust. This technique is likely to stretch the
abdominal muscles, and thus stimulate muscle spindle
afferents in the oblique muscles. It is possible that this
stimulus may have modified excitability of abdominal
motoneurons. However, the rotational mobilization
technique applied in this study rotated the pelvis and
spine to the right which would shorten the right OI and
stretch the OE, from which EMG was recorded. As
activity of both OI and OE was increased the change in
length of he muscles with the procedure (opposite for
these muscles) is unlikely to explain the increased in
activity observed for both muscles.
The present data indicate that changes following SMT
were only present in people with LBP, and no changes
were identified in healthy control subjects. This is
consistent with previous studies. For instance, decreased
paraspinal muscle activity in response to painful
stimulation over the spinous process following SMT
has been shown to occur only at painful segments
(Lehman and McGill, 2001). Although the reason that
responses were only changed in people LBP is unclear, it
may relate to the function of the mechanoreceptors or
due to the already abnormal control of the trunk
muscles in people with pain. Several studies have
reported that proprioception in the spine is reduced in
people with LBP (Gill and Callaghan, 1998; Brumagne
et al., 2000), thus afferent stimulation from SMT may
lead to different responses. In terms of control, it is
increasingly accepted that motor control of the trunk
muscles is modified in people with LBP. For instance
activity of the deep abdominal muscle, TrA is delayed
(Hodges and Richardson, 1996) or reduced (Hodges et al.,
2004), reflex responses of the paraspinal muscles are
modified (Leinonen et al., 2001), co-contraction of the
oblique abdominal and paraspinal muscles is increased
(Radebold et al., 2000), and paraspinal muscle activity is
increased during gait (Arendt-Nielsen et al., 1996) and
trunk flexion (Zedka et al., 1999). Thus, changes may be
apparent in this population, but not pain-free controls,
due to pre-existing changes in muscle coordination. For
example, activity of the quadriceps has been shown to
increase as a result of reduced inhibition following SMT
to the sacroiliac joint (Suter et al., 1999). Thus, the
change that was apparent due to pre-existing deficit in
the ability to drive the muscle, and no change would be
expected if the motoneurons were not inhibited initially.
Thus, the pre-existing status of the trunk muscles may
determine whether activity is modified by SMT.
Previous authors have argued that changes in muscle
activity may be explained by reduction in pain. For
instance cervical mobilization produces a hypoalgesic
effect (i.e. increased pressure pain thresholds and
decreased visual analogue scores), which has been
associated with a sympatho-excitatory effect and de-
creased superficial neck flexor muscle activity (Sterling
et al., 2001). Those data are consistent with the
hypothesis that SMT activates descending inhibitory
pathways mediated through the midbrain periaqueduc-
tal grey region (Vicenzino et al., 1998), which could also
be responsible for the motor response associated with
SMT. Consistent with the present data, animal studies
indicate that activation of the dorsal periaqueductal
grey region induces motor facilitation (Lovick, 1992).
In the lumbar spine, previous studies have identified
that muscle responses to mechanical pain stimuli are
changed by SMT only when it is applied to a painful
spinal segment (Lehman and McGill, 2001). However,
this mechanism is unlikely to explain the result of our
study as pain levels did not change significantly after
the SMT.
A notable finding of the present study was the
variability between individuals. This is consistent with
several previous studies (Herzog et al., 1999; Lehman
et al., 2001). Without complete understanding of the
mechanism for SMT to change postural responses it is
difficult to speculate on the factors contributing to the
variability. However, the LBP population was hetero-
geneous and differences in pathology, pain intensity and
functional presentation may be responsible for the
differences. Furthermore, the SMT technique applied
in this study was aimed at a consistent segment, L4–L5,
and this is unlikely to be symptomatic in all individuals.
As previous studies have found that responses are only
modified by SMT at the painful segment (Lehman and
McGill, 2001), this may account for some variability. In
addition, although pain relief from the mobilization was
not significant across the group, some individuals did
report a reduction in symptoms and this may have
contributed to individual variation.
Was the change in muscle activity observed after SMT
positive for clinical improvement? Although SMT has
been shown to reduce pain in people with acute LBP
(Ferreira et al., 2002), whether the change in muscle
response induced by SMT is beneficial for people with
chronic LBP is unclear. Increased activity of the oblique
abdominals has been identified as a common strategy
used by people with LBP to increase the stability of the
spine (Radebold et al., 2000; van Dieen et al., 2003).
Whether a further increase in activity of these muscles is
beneficial is debatable. For instance, augmented activity
of these muscles increases spinal loading (McGill et al.,
2003), which may have negative consequences in the
long term. In fact, some contemporary exercise inter-
ventions, which have been shown to be effective in
 ARTICLE IN PRESS
M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248
management of people with acute and chronic LBP
(O’Sullivan et al., 1997; Hides et al., 2001), aim to
normalize (i.e. reduce) activity of the superficial muscles
(Richardson et al., 1999). Furthermore, these interven-
tions argue that activity of deeper trunk muscles, such as
TrA, which have been shown to be impaired in people
with LBP (Hodges and Richardson, 1996; Hodges and
Richardson, 1998), and have been shown to contribute
to inter-vertebral control (Hodges et al., 2003a, b),
should be augmented. However, SMT did not change
the recruitment of this muscle. These results suggest
manipulative therapy is unlikely to be an alternative to
specific exercise interventions to improve the recruit-
ment of deep abdominal muscles, although it has been
shown that SMT can change superficial abdominal
muscle recruitment. Further, biomechanical and clinical
studies are required to determine whether the change in
muscle activity is beneficial to the health of the spine, or
leads to further effects.
Three methodological issues require consideration.
First, EMG was normalized to the peak activity
recorded for each muscle before SMT. As a result it is
not possible to compare between muscles or between
subject groups. Moreover, it does not allow comparison
of ongoing background activity between groups. How-
ever, it is not practical to normalize to maximum
contractions in people with LBP as they do not perform
true maximal efforts (Allison et al., 1998). The objective
of the study was to evaluate the effect of SMT on the
activity of the trunk muscles and normalization to peak
activity increases the sensitivity to change in a repeated
measures analysis. Second, treatment technique was not
pragmatically performed, since the targeted level was
not always the most symptomatic and patients were not
always positioned on the non-painful side. This may
have influenced the results. However, it is unlikely that
this would explain the variability in response as
mobilization to L4–L5 will affect adjacent segments. A
follow-up study would clarify this issue. Moreover,
thrust techniques may induce different changes to the
mobilization technique used here, even though, it has
been reported that neurophysiologic effects, including
motor response, are not dependent on audible cavitation
and it has been reported that similar changes occur with
mobilization and manipulative thrust techniques (Her-
zog et al., 1999; Dishman and Bulbulian, 2000). Third,
the population included in the study presented with low
levels of pain and disability. These findings need to be
replicated in a more disabled population.
Acknowledgements
We thank Rebecca Mellor for assistance with plan-
ning of the study and data collection. PH was supported
by the NHMRC of Australia.
247
References
Allison GT, Godfrey P, Robinson G. EMG signal amplitude
assessment during abdominal bracing and hollowing. Journal of
Electromyography and Kinesiology 1998;8(1):51–7.
Arendt-Nielsen L, Graven-Nielsen T, Svarrer H, Svensson P. The
influence of low back pain on muscle activity and coordination
during gait: a clinical and experimental study. Pain 1996;64(2):
231–40.
Aruin AS, Latash ML. Directional specificity of postural muscles in
feed-forward postural reactions during fast voluntary arm move-
ments. Experimental Brain Research 1995;103:323–32.
Assendelft WJ, Morton PC, Yu EI, et al. Spinal manipulative therapy
for low back pain: a meta-analysis of effectiveness relative to other
therapies. Annals of Internal Medicine 2003;138:871–81.
Belenkii V, Gurfinkel VS, Paltsev Y. Elements of control of voluntary
movements. Biofizika 1967;12(1):135–41.
Bouisset S, Zattara M. A sequence of postural adjustments precedes
voluntary movement. Neuroscience Letters 1981;22:263–70.
Brennan P, Triano J, McGregor M. Enhanced neutrophil respiratory
burst as a biological marker for manipulation forces: duration of
the effect and association with substance P and tumor necrosis
factor. Journal of Manipulative and Physiological Therapeutics
1992;15(2):83–9.
Brumagne S, Lysens R, Swinnen S, Verschueren S. Effect of paraspinal
muscle vibration on position sense of the lumbosacral spine. Spine
1999;24(13):1328–31.
Brumagne S, Cordo P, Lysens R, Verschueren S, Swinnen S. The role
of paraspinal muscle spindles in lumbosacral position sense in
individuals with and without low back pain. Spine 2000;25(8):
989–94.
Chiradejnant A, Maher C, Latimer J. Efficacy of "therapist-selected"
versus "randomly selected" mobilisation techniques for the treat-
ment of low back pain: a randomised controlled trial. Australian
Journal of Physiotherapy 2003;49:233–41.
Colloca CJ, Keller TS. Stiffness and neuromuscular reflex response of
the human spine to posteroanterior manipulative thrusts in patients
with low back pain. Journal of Manipulative and Physiological
Therapeutics 2001;24(8):489–500.
Dishman JD, Bulbulian R. Spinal reflex attenuation associated with
spinal manipulation. Spine 2000;25(19):2519–24 discussion 2525.
Dishman JD, Bulbulian R. Comparison of effects of spinal manipula-
tion and massage on motoneuron excitability. Electromyography
and Clinical Neurophysiology 2001;41(2):97–106.
Ferreira M, Ferreira P, Latimer J. Does spinal manipulative therapy
help people with chronic low back pain? Australian Journal of
Physiotherapy 2002;48:277–84.
Gill KP, Callaghan MJ. The measurement of lumbar proprioception
in individuals with and without low back pain. Spine 1998;23(3):
371–7.
Herzog W, Scheele D, Conway PJ. Electromyographic responses of
back and limb muscles associated with spinal manipulative therapy.
Spine 1999;24(2):146–52 discussion 153.
Hides JA, Jull GA, Richardson CA. Long term effects of specific
stabilizing exercises for first episode low back pain. Spine
2001;26:243–8.
Hodges PW, Richardson CA. Inefficient muscular stabilisation of the
lumbar spine associated with low back pain: a motor control
evaluation of transversus abdominis. Spine 1996;21:2640–50.
Hodges PW, Richardson CA. Feedforward contraction of transversus
abdominis in not influenced by the direction of arm movement.
Experimental Brain Research 1997;114:362–70.
Hodges PW, Richardson CA. Delayed postural contraction of of
transversus abdominis associated with movement of the lower limb
in people with low back pain. Journal of Spinal Disorders 1998;
11(1):46–56.
 ARTICLE IN PRESS
248 M.L. Ferreira et al. / Manual Therapy 12 (2007) 240–248
Hodges PW, Cresswell AG, Thorstensson A. Preparatory trunk
motion accompanies rapid upper limb movement. Experimental
Brain Research 1999;124:69–79.
Hodges P, Kaigle Holm A, Holm S, Ekstrom L, Cresswell A, Hansson
T, et al. Intervertebral stiffness of the spine is increased by evoked
contraction of transversus abdominis and the diaphragm: in vivo
porcine studies. Spine 2003a;28(23):2594–601.
Hodges PW, Moseley GL, Gabrielsson AH, Gandevia SC. Acute
experimental pain changes postural recruitment of the trunk
muscles in pain-free humans. Experimental Brain Research 2003b;
151:262–71.
Hodges PW, Saunders S, et al. Reduced tonic activity of the deep trunk
muscle during locomotion in people with low back pain. Porto,
Portugal: Spineweek; 2004.
Indahl A, Kaigle AM, Reikeras O, Holm SH. Interaction between the
porcine lumbar intervertebral disc, zygapophysial joints, and
paraspinal muscles. Spine 1997;22(24):2834–40.
Kasai T, Yahagi S, Shimura K. Effect of vibration-induced postural
illusion on anticipatory postural adjustment of voluntary arm
movement in standing humans. Gait Posture 2002;15(1):94–100.
Keller TS, Colloca CJ. Mechanical force spinal manipulation increases
trunk muscle strength assessed by electromyography: a compara-
tive clinical trial. Journal of Manipulative and Physiological
Therapeutics 2000;23(9):585–95.
Lehman GJ, McGill SM. Spinal manipulation causes variable spine
kinematic and trunk muscle electromyographic responses. Clinical
Biomechanics (Bristol, Avon) 2001;16(4):293–9.
Lehman GJ, Vernon H, McGill SM. Effects of a mechanical pain
stimulus on erector spinae activity before and after a spinal
manipulation in patients with back pain: a preliminary investiga-
tion. Journal of Manipulative and Physiological Therapeutics
2001;24(6):402–6.
Leinonen V, Kankaanpaa M, Luukkonen M, Hanninen O, Airaksinen
O, Taimela S. Disc herniation-related back pain impairs feed-
forward control of paraspinal muscles. Spine 2001;26(16):E367–72.
Lovick TA. Midbrain influences on ventrolateral medullo-spinal
neurons in the rat. Experimental Brain Research 1992;90:147–52.
Maitland G, Banks K, English K, Hengeveld E. Lumbar spine.
Maitland’s Vertebral Manipulation. Oxford: Butterworth-Heine-
mann; 2001.
McGill SM, Grenier S, Kavcic N, Cholewicki J. Coordination of
muscle activity to assure stability of the lumbar spine. Journal of
Electromyography and Kinesiology 2003;13(4):353–9.
Nilsson N, Christensen HW, Hartvigsen J. Lasting changes in passive
range of motion after spinal manipulation: a randomized, blind,
controlled trial. Journal of Manipulative and Physiological
Therapeutics 1996;19:165–8.
O’Sullivan PB, Twomey LT, Allison GT. Evaluation of specific
stabilizing exercise in the treatment of chronic low back pain with
radiologic diagnosis of spondylolysis or spondylolisthesis. Spine
1997;22(24):2959–67.
Radebold A, Cholewicki J, Panjabi MM, Patel TC. Muscle response
pattern to sudden trunk loading in healthy individuals and in
patients with chronic low back pain. Spine 2000;25(8):947–54.
Richardson CA, Jull GA, Hodges PW, Hides JA. Therapeutic exercise
for spinal segmental stabilisation in low back pain: Scientific basis
and clinical approach. Edinburgh: Churchill Livingstone; 1999.
Rudomin P. Selectivity of the central control of sensory information in
the mammalian spinal cord. Advances in Experimental Medicine
and Biology 2002;508:157–70.
Shirley D, Ellis E, Lee M. The response of posteroanterior lumbar
stiffness to repeated loading. Manual Therapy 2002;7(1):19–25.
Sterling M, Jull G, Wright A. Cervical mobilisation: concurrent effects
on pain, sympathetic nervous system activity and motor activity.
Manual Therapy 2001;6(2):72–81.
Suter E, McMorland G, Herzog W, Bray R. Decrease in quadriceps
inhibition after sacroiliac joint manipulation in patients with
anterior knee pain. Journal of Manipulative and Physiological
Therapeutics 1999;22(3):149–53.
Urquhart DM, Hodges PW, Allen TJ, Story IH. Abdominal muscle
recruitment during a range of voluntary exercises. Manual Therapy
2005;10(2):144–53.
van Dieen JH, Selen LP, Cholewicki J. Trunk muscle activation in low-
back pain patients, an analysis of the literature. Journal of
Electromyography and Kinesiology 2003;13(4):333–51.
van Tulder M, Koes BW, Bouter LM. Conservative treatment of acute
and chronic nonspecific low back pain - a systematic review of
randomized controlled trials of the most common interventions.
Spine 1997;22(18):2128–56.
Vicenzino B, Collins D, Benson H, Wright A. An investigation of the
interrelationship between manipulative therapy-induced hypoalge-
sia and sympathoexcitation. Journal of Manipulative and Physio-
logical Therapeutics 1998;21(7):448–53.
Zedka M, Prochazka A, Knight B, Gillard D, Gauthier M. Voluntary
and reflex control of human back muscles during induced pain.
Journal of Physiology (London) 1999;520(Pt 2):591–604.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 249–255

www.elsevier.com/locate/math

Original article

Assessment of fine motor control in patients with occupation-related

lateral epicondylitis
Darrell K. Skinnera,, Sandra L. Curwinb
a
Okanagan College, 1000 KLO Road, Kelowna, BC, Canada V1Y 4X8
b
100-1314 Tower Road, Halifax, NS, Canada B3H 4S7
Received 30 October 2005; received in revised form 31 May 2006; accepted 27 June 2006

Abstract

Lateral epicondylitis (LE) is a common overuse injury related to a mechanical overload of the wrist extensors’ origin; however,
some patients also complain of clumsiness suggesting a possible motor control problem. The purpose of this study was to examine
for differences in fine motor control ability between subjects with LE and matched control subjects. Subtests of the Purdue Pegboard
Test (PPT) and the Complete Manual Dexterity Test (CMDT) were administered to 28 subjects with LE, and 28 age, gender, and
hand dominance-matched control subjects. The LE group demonstrated a significant decrease in fine motor control ability on both
measures, compared with the control group on both the PPT, F(1,52) ¼ 9.98, P ¼ 0.003, and the CMDT, F(1,52) ¼ 18.11,
P ¼ 0.001. There appeared to be no effect for the length of time since injury. There were significant differences in fine motor control
ability between individuals with LE and a matched control group for both measures used. These results suggest that tests of fine
motor control should be considered in the assessment of clients with LE. The mechanism related to the deficit is unknown and
warrants further research.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Tendonitis; Lateral epicondylitis; Motor control; Neuroplasticity; Repetitive strain

1. Introduction 2001; Smidt et al., 2003). LE is considered a self-limiting

Lateral epicondylitis of the elbow (LE) is a common
soft tissue injury which involves the tendinous origin of
the wrist and finger extensors (Nirschl and Ashman,
2003). Epidemiological studies have suggested an
incidence of approximately 4–7% in the general
population (Assendelft et al., 1996). Workers in
occupations involving repetitive hand-intensive work
appear particularly at risk for developing this condition
(Kivi, 1984; Dimberg et al., 1989; Kurppa et al., 1991;
Chiang et al., 1993). Despite the high incidence of LE,
there is little consensus on the best treatment approach
to guide the clinician (Labelle et al., 1992; Bowen et al.,
Corresponding author. Tel.: +1 250 762 5445x4464;
fax: +1 250 862 5461.
E-mail address: dskinner@okanagan.bc.ca (D.K. Skinner).
condition that will resolve gradually with adequate rest
and time, however, there is a smaller subpopulation of
patients with chronic or recurrent problems and
increased disability who present a significant treatment
challenge (Burgess, 1990; Boyer and Hastings, 1999;
Haahr and Andersen, 2003).
The ‘‘overuse’’ or ‘‘biomechanical’’ model of LE
focuses on the repetitive mechanical overloading of the
tendon beyond its adaptive and reparative capacity as
the primary cause of signs and symptoms (Jarvinen
et al., 1997; Melborn, 1998; Moore, 2002). Few signs of
inflammation are typically present in chronic cases of
tendonitis, and it is now appreciated that the traditional
soft tissue healing model with a prolonged inflammatory
response does not fully explain the pathology involved
(Assendelft et al., 1996; Green et al., 2002; Smidt et al.,
2002). The effect of repetitive movement on the

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.018
 ARTICLE IN PRESS
250 D.K. Skinner, S.L. Curwin / Manual Therapy 12 (2007) 249–255
somatosensory cortex has interested researchers. Highly
stereotypical repetitive movement patterns have been
shown in animal models to produce cortical changes
which may lead to impaired motor performance (Byl
et al., 1996a, b, 1997; Byl and Melnick, 1997). It is
possible that these changes in movement strategies and
altered muscle recruitment patterns, whether due to pain
or central alterations in motor control, may result in
increased loading of the already compromised structures
within the muscles and tendon attachments (Byl et al.,
2000a, b; Barr et al., 2004; Byl, 2004; Ervilha et al.,
2004). Patients in the clinical setting often describe a
feeling of ‘‘clumsiness’’ associated with LE, suggesting
the possible existence of a fine motor control problem.
Few researchers have measured motor control in
patients with tendonitis. Viikari-Juntura et al. (1994)
compared reaction time, movement time, visual atten-
tion, and visuospatial ability in 26 meat cutters and
packers with a history of two or more episodes of wrist
tenosynovitis to a control group, and found no
significant differences. Subjects in the experimental
group had a history of tenosynovitis, however, were
asymptomatic at the time of testing (Viikari-Juntura
et al., 1994). These results suggest that poor manual
dexterity may not be a predictor of wrist tenosynovitis.
Pienimaki et al. (1997) measured the gross motor
control of patients with chronic unilateral LE, and
concluded that reaction speed and speed of movement
were decreased bilaterally in patients when compared to
age and gender-matched control subjects; however, were
unable to explain the results (Pienimaki et al., 1997).
The purpose of the present study was to systematically
examine whether there were differences in fine motor
control ability between individuals with LE and control
subjects matched on age, gender, and hand dominance,
using valid and reliable measures. This study was
reviewed and approved by the Ethics Boards of
the University of Alberta, and Okanagan University
College.
2. Subjects
Subjects in the LE group (n ¼ 28) were injured
workers between the ages of 30 and 53 years with LE,
attending a Cumulative Activity-Related Disorder
(CARD) programme at Millard Health in Edmonton
(n ¼ 15) or Orion Health in Calgary (n ¼ 13). All the
subjects in the LE group were on a Worker’s Compen-
sation claim due to a work-related repetitive strain
injury of the upper extremity. A work site visit was
completed by a staff of the CARD programme to
confirm the presence of identifiable risk factors to
substantiate the claim. The length of time since injury
was self-reported by the subject as the time in weeks
since they first experienced symptoms related to their
claim. The mean time since injury for subjects in the LE
group was 30.54 weeks (SD 36.69 weeks). The 28 control
subjects were individuals also between 30 and 53 years,
attending physiotherapy treatment at Sun City Phy-
siotherapy (n ¼ 21), or Columbia Health (n ¼ 7) in
Kelowna, for a non-upper extremity-related condition.
The procedure for subject recruitment and reasons for
attrition are illustrated in a flow chart (Fig. 1).
3. Methods
The diagnosis of LE was confirmed clinically by: (1)
positive Cowen’s test (pain reproduced with resisted
wrist extension with the forearm in pronation), (2)
lateral elbow pain with an extended grasp, (3) accom-
panying localized tenderness to palpate near the origin
of the extensor carpi radialis brevis and extensor carpi
ulnaris tendons just distal to the lateral epicondyle
(Burgess, 1990; Boyer and Hastings, 1999). All three of
these tests were required to be positive for inclusion in
this study as this gave the strongest reassurance of the
presence of LE. Subjects with cervical radiculopathy
were excluded by clinical examination by a physiothera-
pist including a clinical neurological examination of the
cervical spine. Subjects with referred pain or elbow
pathology, other than LE, were excluded by clinical
examination of the shoulder and elbow joints. Subjects
with medical factors affecting motor control including:
carpal tunnel syndrome and peripheral nerve conditions,
multiple sclerosis, Parkinson’s disease, brain tumours,
cerebral vascular accidents, peripheral neuropathies and
any other known neurological condition were excluded
from the LE or control groups. Individuals on medica-
tions such as tricyclic antidepressants, neuroleptic and
antipsychotic drugs, or those in withdrawal from
alcohol or street drugs were also excluded. The presence
of medical conditions and medications was determined
by chart review and subject interview. Any subject who
complained of pain or discomfort during testing was
excluded from this study.
The outcome measures used in this study were the
subtest of the Purdue Pegboard Test (PPT) Model
32020, and the One-handed Turning and Placing test of
the Complete Manual Dexterity Test (CMDT) Model
32023A. Both tests are manufactured by Lafayette
Instrument (3700 Sagmore Parkway N, PO Box 5729,
Lafayette, IN 47903). The PPT consists of a rectangular
board with two vertical lines of pin holes, and four
shallow wells containing pins at the top of the board.
The One-handed Pin Placement assesses a person’s
ability to place as many pins as possible over a 30 s
period using predominantly a finger tip to tip grip. The
total number of pins placed by subjects over three trials
of the One-handed Pin Placement test was recorded. The
CMDT consists of a folding board with 60 wells, into
 ARTICLE IN PRESS
D.K. Skinner, S.L. Curwin / Manual Therapy 12 (2007) 249–255
Informed Consent
Lateral Epicondylitis (n = 38)
Assessed for Eligibility
1 subject excluded due to possible
radial tunnel syndrome
1 subject excluded due to
language difficulties
Testing of Motor Control (n = 36)
1 subject excluded due to
elbow discomfort
1 subject excluded due to
shoulder discomfort
Matching (n = 34)
6 subjects excluded as no age or
gender match in the control group
Analysed (n = 28)
12 males, 16 females
Mean age 41.93 (SD 6.44)
Fig. 1. Subject enrolment.
which medium sized cylindrical blocks are placed. The
One-handed Turning and Placing test involves using the
fingers to turn over the block, and then reaching
forward to place each block consecutively into a
well. The total time in seconds to complete four trials
of the One-handed Turning and Placing test was
recorded.
The experience of many therapists who regularly use
the PPT is that the original norms supplied with the test
are often higher than those test scores observed in
clinical practice. Hamm and Curtis (1980) have also
recommended that a clinical sample be compared to its
own control group (Hamm and Curtis, 1980). The
literature suggested that age, gender, and hand dom-
inance were probable confounding variables which
could influence the test results and therefore were
controlled for in this study.
251
Control Group (n = 31)
Assessed for Eligibility
Testing of Motor Control (n = 31)
1 subject excluded due to symptoms
of possible ulnar neuropathy
Matching (n = 30)
2 subjects excluded as no age or
gender match in the LE group
Analysed (n = 28)
12 males, 16 females
Mean age 42.36 (SD 6.44)
4. Matching
The LE and control groups were matched on the basis
of age, gender, and hand dominance. Both genders were
included in this study and there were 12 males and 16
females in both the LE and control groups. Matching
for age was achieved by individually selecting the
control subject with the closest available match for age
to the individual with LE. Twenty out of 28 of the
subjects (71%) in the LE group had LE in their
dominant limb. The effect of hand dominance was
controlled by matching the affected limb (whether
dominant or non-dominant) of the subject with LE to
the corresponding dominant or non-dominant limb of
the control subject. The dominant limb was defined as
the hand that the individual chooses to write with. The
affected limb was the limb that the subject was attending
 ARTICLE IN PRESS
252 D.K. Skinner, S.L. Curwin / Manual Therapy 12 (2007) 249–255
treatment for. Matching of the LE and control groups
by right handedness or left handedness was not
performed, as the intent was to control for the possible
effect of hand dominance, not the effect of cerebral
lateralization. In reality, only one subject in the LE
group was left-hand dominant and the inclusion or
exclusion of this subject would not have substantially
altered the results of this study.
5. Data analyses
For reasons described in Fig. 1, 10 subjects in the LE
group and three subjects in the control group were
excluded from the data. In particular, two subjects in the
LE group were excluded due to discomfort during
testing. Statistical analyses of the data were performed
using the SPSS Ver 12.0 software (SPSS for Windows,
SPSS, Chicago, IL). The level for statistical significance
was set at Po0.05.
Standard descriptive statistical analyses were used to
describe the characteristics of both groups. An indepen-
dent t-test was used to compare the mean ages of the LE
and control groups to illustrate that there was no
significant difference between the mean ages of the two
groups. Two-way ANOVAs were used for each measure
to test for the main effect for the factors of hand
dominance and group, and the interaction effect
between group and affected limb. The LE group was
then divided into two subgroups: those who had LE for
12 weeks or less (acute group, n ¼ 13), and those
individuals who had LE for greater than 12 weeks
(chronic group, n ¼ 15). Chronicity is often defined as
pain lasting more than 3 or 6 months, or simply longer
than the expected course of healing for an acute disease
process (Russo and Brose, 1998; Starcke, 2005).
The cut-off point of 12 weeks was used in this study
for the purpose of examining if there was a greater
deficit of fine motor control in the more chronic cases of
LE. A one-way ANOVA was performed with these two
subgroups.
Table 1
Mean test scores and standard deviations for lateral epicondylitis and control groups
N Mean
Purdue* lateral epicondylitis control 28 41.43a
28 46.43a
CMDT** lateral epicondylitis control 28 354.82b
28 303.79b
*Po.005; **Po.001.
a
Cumulative score of the number of pins placed in three 30 s trials.
b
Cumulative score in seconds to complete four trials of the Complete Manual Dexterity Test (CMDT).
6. Results
There were no significant differences between the
mean ages of the LE group (41.9876.44 years) and the
control group (42.3676.44 years) (t2 ¼ 0.25,
P ¼ 0.80). The gender distribution was identical be-
tween the groups by design.
A two-way ANOVA indicated no main effect
for hand dominance for either the Purdue scores
F(1,52) ¼ 1.66, P ¼ 0.20, or the CMDT scores
F(1,52) ¼ 1.52, P ¼ 0.22, and no interaction effect
between the group and the affected limb for the Purdue
F(1,52) ¼ 0.003, P ¼ 0.96, or the CMDT F(1,52) ¼
0.10, P ¼ 0.75. Therefore, the scores on the PPT and
CMDT for the dominant and non-dominant limbs
were collapsed to represent only two groups, n ¼ 28
(Table 1). Subjects in the LE group placed on average
5.00 fewer pins over three trials of the PPT, and were on
average 51.03 s slower to complete four trials of the
CMDT, than individuals in the control group. Statistical
comparison indicated a significant difference between
the mean scores of the LE group and the control group,
for both the PPT F(1,52) ¼ 9.98, P ¼ 0.003, and the
CMDT F(1,52) ¼ 18.11, P ¼ 0.001.
In summary, the LE group demonstrated a significant
decrease in fine motor control ability on both the PPT
and CMDT compared with the age- and gender-
matched control groups. There was no significant
difference between the acute LE group and the chronic
LE group for both the PPT, F(1,26) ¼ 0.087, P ¼ 0.77;
and the CMDT, F(1,26) ¼ 0.094, P ¼ 0.76.
7. Discussion
Repetitive strain injuries of the wrist and forearm are
common work-related conditions that can be disabling
and costly to treat (Barr and Barbe, 2002). The
commonly assumed model of injury for work-related
LE is a biomechanical overloading of the tendons of the
common extensor origin; however, resistance of some
cases of LE to various forms of conservative treatment
SD 95% confidence interval for the mean
Lower Upper
6.84 38.78 44.08
3.65 45.01 47.84
46.91 336.63 373.01
38.49 288.86 318.71
 ARTICLE IN PRESS
D.K. Skinner, S.L. Curwin / Manual Therapy 12 (2007) 249–255
may imply that this explanation is not true for all cases.
Complaints of clumsiness from some patients with LE
suggests that there may be alterations in motor control
ability in these individuals (Byl et al., 1996a, b, 1997; Byl
and Melnick, 1997).
The mean age of the LE subjects (41.9876.44 years)
and the propensity for the dominant limb to be most
often affected (71%) was consistent with the findings of
previous epidemiological studies, suggesting that the
sample was likely representative of the larger population
of individuals with LE (Coonrad and Hooper, 1973;
Assendelft et al., 1996). The results of this study
demonstrated a significant decrease in fine motor
control in subjects with LE compared to the control
group. Results from both the PPT and the CMDT
showed similar decreases in fine motor control increas-
ing the concurrent validity of this study. The design of
this present study did not allow determination of
whether LE preceded the observed deficits, or whether
the deficits in fine motor control preceded the develop-
ment of the LE. Although the sample size was small in
both the acute and chronic subgroups, comparison of
the two groups suggested that the time since injury did
not affect fine motor control performance, and raises
some questions regarding the possible time course of the
difficulty in motor control. Further research with an
estimated sample size of approximately 25 subjects in
each group would be needed to confirm the current
findings. The results from animal research by Barbe and
Barr suggest that multiple pathomechanisms may exist
during the development of tendonitis, and that simulta-
neous pathophysiological changes may occur in the
musculoskeletal system along with neurological system
adaptation (Barr and Barbe, 2002; Barbe et al., 2003).
The mechanisms involved in the decrease in fine
motor control in the subjects with LE in this study are
not known. The literature related to this area does,
however, offer some possible theories. Experimental
primate studies (Jenkins et al., 1990; Byl et al., 1996a, b),
and human studies with musicians (Elbert et al., 1998;
Byl et al., 2000a b; Pantev et al., 2001), have demon-
strated that reorganization of the cortical representation
of the somatosensory cortex (Brodman’s area 3b) can
occur with repetitive use. These neuroplastic changes
potentially may result in decreased fine motor control
and increased load on the tendon structures. The
difficulties in fine motor control found in subjects with
LE in this study, however, does not explicitly mean that
this is due to cortical remodelling. Other possible
explanations for the results of this study include altered
motor performance due to the effect of a chronically
painful condition, or individual behavioural responses
(Lund et al., 1991; Smeulders et al., 2001; Flor, 2003;
Giamberardino, 2003). In this present study, pain did
not appear to be a significant factor during testing as
only two subjects in the LE group complained of
253
discomfort during testing and were excluded. The verbal
instructions given to each subject are standardized to
help minimize differences in individual performance due
to communication and interpretation.
8. Limitations
The selection of the LE group was a non-randomized
convenience sample, and the results of this study can
only be generalized to a population of similar char-
acteristics. Sampling bias was minimized through the
use of age, gender, and dominance-matched control
subjects. The examiner was not blinded to the subject
group assignment, and this may possibly have influenced
the subjects’ performance. Strength or endurance
limitations may affect the results of this study; however,
both tests are short and non-strenuous tests of finger
dexterity. Motivational factors are a potential factor in
any test of human performance. The Purdue Pegboard
and the CMDT tests are well standardized and
structured as to the verbal instructions given to each
subject; however, the response of subjects to the
standardized instructions could vary.
9. Clinical implications
Failure to recognize or address problems with fine
motor control may be one explanation for the persis-
tence of difficult cases of LE that are resistant to existing
treatment approaches usually aimed at the peripheral
musculotendinous structures. The PPT and the CMDT
are commonly used clinical assessment tools, and a
practical method for clinicians working with patients
with LE to similarly identify if there are problems with
motor control. This factor could then be addressed in
the patient’s rehabilitation programme. The PPT and
the CMDT appear equally sensitive in detecting
differences in fine motor control ability in these subjects
with LE; however, the PPT is less expensive and time
consuming for the practicing clinician.
The PPT and CMDT are measures of impairment,
and it is not known how the observed deficits would
affect the individual functionally. A study examining the
relationship between and the degree of impairment and
the level of functional disability would be helpful. The
results of this study do not explain or imply causation
for the decrease in fine motor control; however, this
study may offer some explanation of why some subjects
complain of clumsiness in their affected limb. The
possibility of dysfunction in neuromuscular control
should be considered in assessment and treatment of
repetitive strain disorders (Byl et al., 2000a, b; McKenzie
et al., 2003; Barr et al., 2004). Measurement of fine
motor control may help select subjects with LE who
 ARTICLE IN PRESS
254 D.K. Skinner, S.L. Curwin / Manual Therapy 12 (2007) 249–255
could benefit from specific motor control training. It is
not known as yet, however, if improvements in fine
motor control will have a direct impact on the resolution
of LE, and this warrants further research.
Acknowledgements
I am very thankful to my advisor, Sandra Curwin
Ph.D., for her encouragement and advice throughout
this research project. I thank my committee members
and Bonnie Dobbs, Ph.D. for the important contribu-
tion of their time and experience. The support of
management and clinicians from the Workers Compen-
sation Board of Alberta, Orion Health, Sun City
Physiotherapy, and Columbia Health was invaluable
in making this research project possible.
References
Assendelft WJ, Hay EM, et al. Corticosteroid injections for lateral
epicondylitis: a systematic overview. British Journal of General
Practice 1996;46(405):209–16.
Barbe MF, Barr AE, et al. Chronic repetitive reaching and grasping
results in decreased motor performance and widespread tissue
responses in a rat model of MSD. Journal of Orthopaedic Research
2003;21(1):167–76.
Barr AE, Barbe MF. Pathophysiological tissue changes associated
with repetitive movement: a review of the evidence. Physical
Therapy 2002;82(2):173–87.
Barr AE, Barbe MF, et al. Work-related musculoskeletal disorders of
the hand and wrist: epidemiology, pathophysiology, and sensor-
imotor changes. Journal of Orthopaedic Sports and Physical
Therapy 2004;34(10):610–27.
Bowen RE, Dorey FJ, et al. Efficacy of nonoperative treatment for
lateral epicondylitis. American Journal of Orthopaedics 2001;30(8):
642–6.
Boyer MI, Hastings II H. Lateral tennis elbow: is there any science out
there? Journal of Shoulder Elbow Surgery 1999;8(5):481–91.
Burgess RC. Tennis elbow. Journal of Ky Medical Association
1990;88(7):349–54.
Byl N. Focal hand dystonia may result from aberrant neuroplasticity.
Advances in Neurology 2004;94:19–28.
Byl NN, Melnick M. The neural consequences of repetition: clinical
implications of a learning hypothesis. Journal of Hand Therapy
1997;10(2):160–74.
Byl N, Wilson F, et al. Sensory dysfunction associated with repetitive
strain injuries of tendinitis and focal hand dystonia: a comparative
study. Journal of Orthopaedic Sports and Physical Therapy
1996a;23(4):234–44.
Byl NN, Merzenich MM, et al. A primate genesis model of
focal dystonia and repetitive strain injury: I. Learning-induced
dedifferentiation of the representation of the hand in the primary
somatosensory cortex in adult monkeys. Neurology 1996b;47(2):
508–20.
Byl NN, Merzenich MM, et al. A primate model for studying focal
dystonia and repetitive strain injury: effects on the primary
somatosensory cortex. Physical Therapy 1997;77(3):269–84.
Byl N, Nagarajan SS, et al. Effect of sensory discrimination training on
structure and function in a musician with focal hand dystonia.
Physical Therapy Case Report 2000a;3:94–113.
Byl NN, McKenzie A, et al. Differences in somatosensory
hand organization in a healthy flutist and a flutist with focal hand
dystonia: a case report. Journal of Hand Therapy 2000b;13(4):
302–9.
Chiang HC, Ko YC, et al. Prevalence of shoulder and upper-
limb disorders among workers in the fish-processing industry.
Scandinavian Journal of Work Environment and Health 1993;
19(2):126–31.
Coonrad RW, Hooper WR. Tennis elbow: its course, natural history,
conservative and surgical management. Journal of Bone Joint
Surgery America 1973;55(6):1177–82.
Dimberg L, Olafsson A, et al. The correlation between work
environment and the occurrence of cervicobrachial symptoms.
Journal of Occupational Medicine 1989;31(5):447–53.
Elbert T, Candia V, et al. Alteration of digital representations in
somatosensory cortex in focal hand dystonia. Neuroreport
1998;9(16):3571–5.
Ervilha UF, Arendt-Nielsen L, et al. The effect of muscle pain on
elbow flexion and coativation tasks. Experiments in Brain Research
2004;156(2):174–82.
Flor H. Cortical reorganisation and chronic pain: implications
for rehabilitation. Journal of Rehabilitation Medicine 2003;
35(Suppl 41):66–72.
Giamberardino MA. Referred muscle pain/hyperalgesia and central
sensitisation. Journal of Rehabilitation Medicine 2003;35
(Suppl 41):85–8.
Green S, Buchbinder R, et al. Non-steroidal anti-inflammatory drugs
(NSAIDs) for treating lateral elbow pain in adults. Cochrane
Database System Review 2002(2):CD003686.
Haahr JP, Andersen JH. Prognostic factors in lateral epicondylitis:
a randomized trial with one-year follow-up in 266 new cases
treated with minimal occupational intervention or the usual
approach in general practice. Rheumatology (Oxford) 2003;42(10):
1216–25.
Hamm N, Curtis D. Normative data for the purdue pegboard on a
sample of adult candidates for vocational rehabilitation. Perceptual
and Motor Skills 1980;50:309–10.
Jarvinen M, Jozsa L, et al. Histopathological findings in chronic
tendon disorders. Scandinavian Journal of Medicine and Science in
Sports 1997;7(2):86–95.
Jenkins WM, Merzenich MM, et al. Functional reorganization
of primary somatosensory cortex in adult owl monkeys after
behaviorally controlled tactile stimulation. Journal of Neurophy-
siology 1990;63(1):82–104.
Kivi P. Rheumatic disorders of the upper limbs associated with
repetitive occupational tasks in Finland in 1975–1979. Scandina-
vian Journal of Rheumatology 1984;13(2):101–7.
Kurppa K, Viikari-Juntura E, et al. Incidence of tenosynovitis or
peritendinitis and epicondylitis in a meat-processing factory.
Scandinavian Journal of Work Environment and Health 1991;
17(1):32–7.
Labelle H, Guibert R, et al. Lack of scientific evidence for
the treatment of lateral epicondylitis of the elbow. An attempted
meta-analysis. Journal of Bone and Joint Surgery—British Volume
1992;74(5):646–51.
Lund JP, Donga R, et al. The pain-adaptation model: a discussion
of the relationship between chronic musculoskeletal pain and
motor activity. Canadian Journal of Physiology and Pharmacology
1991;69(5):683–94.
McKenzie AL, Nagarajan SS, et al. Somatosensory representation of
the digits and clinical performance in patients with focal hand
dystonia. American Journal of Physical Medicine and Rehabilita-
tion 2003;82(10):737–49.
Melborn JM. Cumulative trauma disorders and repetitive strain
injuries. Clinical Orthopaedics and Related Research 1998;
351:107–28.
 ARTICLE IN PRESS
D.K. Skinner, S.L. Curwin / Manual Therapy 12 (2007) 249–255
Moore JS. Biomechanical models for the pathogenesis of specific distal
upper extremity disorders. American Journal of Industrial Medi-
cine 2002;41(5):353–69.
Nirschl RP, Ashman ES. Elbow tendinopathy: tennis elbow. Clinical
Sports in Medicine 2003;22(4):813–36.
Pantev C, Engelien A, et al. Representational cortex in musicians.
Plastic alterations in response to musical practice. Annals of the
New York Academy of Sciences 2001;930:300–14.
Pienimaki TT, Kauranen K, et al. Bilaterally decreased motor
performance of arms in patients with chronic tennis elbow.
Archives of Physical Medicine and Rehabilitation 1997;
78(10):1092–5.
Russo CM, Brose WG. Chronic pain. Annual Review of Medicine
1998;49:123–33.
255
Smeulders MJ, Kreulen M, et al. Fine motor assessment in chronic
wrist pain: the role of adapted motor control. Clinical Rehabilita-
tion 2001;15(2):133–41.
Smidt N, van der Windt DA, et al. Corticosteroid injections,
physiotherapy, or a wait-and-see policy for lateral epicondylitis:
a randomised controlled trial. Lancet 2002;359(9307):657–62.
Smidt N, Assendelft WJ, et al. Effectiveness of physiotherapy
for lateral epicondylitis: a systematic review. Annali Medici 2003;
35(1):51–62.
Starcke AM. Managing the elusive beast of chronic pain: a challenge
for PTs. Magazine of Physical Therapy 2005;13(11):48–52.
Viikari-Juntura E, Hietanen M, et al. Psychomotor capacity and
occurrence of wrist tenosynovitis. Journal of Occupational
Medicine 1994;36(1):57–60.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 256–262

www.elsevier.com/locate/math

Original article

The diagnostic validity of the cervical flexion–rotation

test in C1/2-related cervicogenic headache
Mark Ogince, Toby Hall, Kim Robinson, A.M. Blackmore
School of Physiotherapy, Curtin University of Technology, C/o 54 Bunya Street, Noranda, Perth, WA 6062, Australia
Received 8 March 2005; received in revised form 21 March 2006; accepted 27 June 2006

Abstract

This single-blind comparative group design aimed to investigate the sensitivity and specificity of the cervical flexion–rotation test
in the diagnosis of C1/2-related cervicogenic headache.
This study tested 23 cervicogenic headache, 23 asymptomatic controls and 12 migraine with aura subjects, all aged 18–66 years. In
stage 1, an experienced manipulative physiotherapist who did not partake in the flexion–rotation test procedure identified C1/2
dysfunction using passive segmental mobility tests in the cervicogenic headache group. Those with C1/2 dysfunction participated in
stage 2. In stage 2, using the flexion–rotation test, subjects were tested by two experienced manipulative physiotherapists blinded to
the subjects’ group allocation. Each therapist stated whether the test was positive or not based on the therapist’s interpretation of
range of motion.
The sensitivity and specificity of the flexion–rotation test was 91% and 90%, respectively (Po.001), with an overall diagnostic
accuracy of 91% (Po.001). The cervical flexion–rotation test significantly assists in the differential diagnosis of cervicogenic
headache and in the identification of movement impairment at the C1/2 segment in patients with cervicogenic headache.
r 2006 Published by Elsevier Ltd.

Keywords: Sensitivity; Specificity; Manual examination; C1/2 segment

1. Introduction Headache presents a diagnostic challenge due to

Cervicogenic headache has been identified as a
distinct subgroup by the International Headache Society
(IHS) (Headache Classification Committee of the
International Headache Society, 2004) and is caused
by disease or dysfunction of structures in the neck
(Edmeads, 2001).
The diagnostic criteria for cervicogenic headache
outlined by the IHS include subjective features together
with evidence of impairment of cervical function on
physical examination. Such impairment includes atlan-
to-axial motion segment (C1/2 level) dysfunction (Hall
and Robinson, 2004) identified by the flexion–rotation
test.
Corresponding author. Tel./fax: +61 8 9375 3224.
E-mail address: ogincem@gmail.com (M. Ogince).
similarities of signs and symptoms among the many
types of headache (Nicholson and Gaston, 2001). In
particular, distinguishing between cervicogenic head-
ache and migraine is problematic (Lewit, 1977, Sjaastad
and Bovim, 1991, Vernon et al., 1992; Blau and
MacGregor, 1994).
The IHS subjective classification criteria are com-
monly used to classify cervicogenic headache however
this process fails to identify the segmental source of
pain, which is important for the application of manual
therapy treatment. Radiological examinations are not
effective (Jensen et al., 1990; Edmeads, 2001), and nerve
block procedures are often impractical, particularly in
the upper cervical region (Bogduk et al., 1985).
Impaired neck mobility is a diagnostic criterion in
cervicogenic headache (Mersky and Bogduk, 1994;
Sjaastad et al., 1998), however several studies have

1356-689X/$ - see front matter r 2006 Published by Elsevier Ltd.

doi:10.1016/j.math.2006.06.016
 ARTICLE IN PRESS
M. Ogince et al. / Manual Therapy 12 (2007) 256–262
found that active cervical mobility is unreliable in
differential diagnosis (Jull et al., 1988; Jensen et al.,
1990; Treleaven et al., 1994; Sandmark and Nisell, 1995;
Placzek et al., 1999; Hall and Robinson, 2004).
Conversely, manual examination has been shown to
detect symptomatic cervical joint dysfunction in a
number of studies of cervical headache (Jull et al.,
1988; Jaeger, 1989; Jensen et al., 1990; Watson and
Trott, 1993; Treleaven et al., 1994; Whittingham et al.,
1994; Schoensee et al., 1995). Clinically, these tests,
described by Maitland et al. (2001), are used as the
current reference standard; however, they involve a high
degree of skill on the part of the therapist.
Additionally, most external measures of cervical
motion incorporate movements of both the upper and
lower cervical regions simultaneously (Amiri et al.,
2003). However, as cervicogenic headache has a primary
involvement in the upper cervical segments, measure-
ment of rotation purportedly biased to the upper
cervical region could be a relevant clinical measure in
the differential diagnosis of cervicogenic headache
(Amiri et al., 2003). Furthermore, determining the
dysfunctional cervical segment facilitates a more accu-
rate treatment approach.
The cervical flexion–rotation test is an easily applied
method of manual examination that is said to provide a
means of determining the presence of joint dysfunction at
the C1/2 level (Stratton and Bryan, 1994). The flexion–
rotation test is conducted with the cervical spine fully
flexed in an attempt to block as much rotational move-
ment as possible above and below C1/2. The head is then
rotated to the left and the right. If firm resistance is
encountered and range is limited before the expected end
range, then this is said to be significant, with a presumptive
diagnosis of limited rotation of the atlas on the axis
(Stratton and Bryan, 1994). Anecdotally, pain provocation
during the flexion–rotation test is also a feature of a
positive test result, however, pain is not a feature in
asymptomatic subjects (Hall and Robinson, 2004).
It is deducible from the frequently reported over-
lapping characteristics seen with cervicogenic headache
and migraine that in the clinical realm many cervico-
genic patients are currently misdiagnosed as having
migraine headache and migraine patients misdiagnosed
as having cervicogenic headache (Sjaastad and Bovim,
1991). Consequently, it is likely that treatment is
unsubstantiated and a poor prognosis will follow. Thus,
there appears to be a need to identify physical tests that
are valid, reliable and sensitive in assisting the diagnosis
of cervicogenic headache.
Accordingly, the aims of this study were to determine
the sensitivity and specificity of the cervical flexion–
rotation test. An additional aim was to determine if a
relationship exists between cervicogenic headache sever-
ity and the extent of restriction demonstrated by the
flexion–rotation test.
257
2. Methods
2.1. Subjects
A single-blind comparative group design was used to
determine differences between asymptomatic subjects,
migraine with aura subjects and those with C1/2-related
cervicogenic headache. Based on the previous findings of
Hall and Robinson (2004), in order to detect a 101
difference of rotation with the cervical spine in flexion,
with alpha at .05, power of 80% and a standard
deviation of 81, this study required at least 10 subjects
per group. In total, 23 cervicogenic headache subjects (3
males, 20 females, mean age ¼ 46 years), 23 asympto-
matic controls (8 males, 15 females, mean age ¼ 40
years) and 12 migraines with aura subjects (9 males,
3 females, mean age ¼ 37 years) participated in the
study. Subjects ranged in age from 18 to 66 years. This
study had approval by the appropriate Human Research
Ethics Committee.
3. Materials
The cervical range of motion device (CROM)
(Performance Attainment Associates. 958 Lydia Drive,
Roseville, Minnesota, USA. 55113) was modified to
measure cervical rotation in a fully flexed cervical spine
position. Two Velcro straps were fixed to the subject’s
head, traversing the transverse and coronal planes,
respectively (Fig. 1). The CROM goniometer was
attached to the centre of the coronal Velcro strap to
measure cervical rotation in maximal flexion (Fig. 1).
The CROM and modified CROM have been shown to
have good intratester and intertester reliability (Capua-
no-Pucci and Rheault, 1991; Hall and Robinson, 2004;
Rheault and Albright, 1992). A headache questionnaire
Fig. 1. The cervical range of motion device (modified to measure
cervical rotation in flexion) and the flexion–rotation test.
 ARTICLE IN PRESS
258 M. Ogince et al. / Manual Therapy 12 (2007) 256–262
(Niere and Robinson, 1997) was used to assess the
severity of headache.
3.1. Procedures
Subjects were recruited from medical specialists and
by advertisements. In response, 325 headache and 23
asymptomatic subjects expressed interest in participa-
tion. Using the IHS (Headache Classification Commit-
tee of the International Headache Society, 1988)
subjective diagnostic criteria as well as criteria outlined
by Sjaastad et al. (1998), Bogduk (1994) and Lord et al.
(1994), telephone interviews were conducted to identify
subjects with cervicogenic headache, migraine with aura
headache and asymptomatic controls.
The inclusion criteria for the cervicogenic subjects
were unilateral or side dominant headache without side-
shift, (Sjaastad et al., 1998) headache associated with
neck pain or reported stiffness (Bogduk, 1994; Sjaastad
et al., 1998) neck symptoms preceding or co-existent
with the onset of headache (Lord et al., 1994; Sjaastad
et al., 1998) and pain precipitated or aggravated by
specific neck movements or sustained posture (Head-
ache Classification Committee of the International
Headache Society, 1988). Additionally, passive segmen-
tal mobility tests reveal symptomatic C1/2 dysfunction
(Maitland et al., 2001) as well as headache frequency at
least an average of one per week and history of episodic
semicontinuous or continuous headache for at least the
previous 3 months. The exclusion criteria for the
cervicogenic headache group were diagnostic criteria
of headache that are not of cervical origin (Headache
Classification Committee of the International Headache
Society, 1988).
Asymptomatic subjects had no history of subjective
features of cervicogenic headache, migraine, migraine
with aura headache, episodic headache, and neck pain
or stiffness. The following exclusion criteria related to
all 3 groups; dizziness on cervical spine movement,
inability to tolerate the flexion–rotation test position,
failure to provide informed consent, insufficient fluency
in English, known congenital fusion, history of cervical
surgery, cervical or cranial trauma, rheumatoid arthritis
and Downs Syndrome.
The inclusion criteria for the migraine with aura
group were based on the IHS classification (Headache
Classification Committee of the International Headache
Society, 1988). The exclusion criteria for this group were
neck pain, discomfort or stiffness.
Of the 325 symptomatic subjects interviewed, 82%
reported overlapping cervicogenic and migraine symp-
toms and were rejected from the study. The remaining
symptomatic subjects consisted of 12 migraine with aura
and 46 cervicogenic subjects. As a sample of conve-
nience 23 cervicogenic, 23 asymptomatic and 12
migraine with aura subjects who matched the criteria
were recruited into the study. Telephone interviews and
the allocation of subjects to their respective groups were
conducted by the principal investigator who did not
partake in the flexion–rotation test procedures.
Cervicogenic headache subjects were required to
attend a preliminary session to determine the presence
of C1/2 dysfunction. This was defined as stage 1.
Thereafter, all subjects attended a single session for
data collection, defined as stage 2. Prior to the start of
the study all subjects were required to provide informed
consent.
In stage 1, an experienced teacher of manipulative
physiotherapy, with 12 years experience who did not
take part in the flexion rotation test procedure,
identified 23 out of 34 cervicogenic headache subjects
with C1/2 as the dominant level of dysfunction using
passive segmental mobility tests. This method of
assessment has been shown to be a valid means of
identifying the symptomatic cervical level in a previous
study (Jull et al., 1988). These 23 subjects were then
allocated to the cervicogenic headache group. The
flexion–rotation test was not used in this part of the
assessment. Segmental mobility was assessed using
passive accessory and physiological intervertebral move-
ments (PAIVMs and PPIVMs) as described by Maitland
et al. (2001). In congruence with the clinical setting,
assessing segmental dysfunction using PAIVMs and
PPIVMs served as the reference standard in this study.
In stage 2, using the flexion–rotation test subjects
were tested by two experienced manipulative phy-
siotherapists who had 16 and 13 years postgraduate
manipulative therapy experience and were blinded to the
subjects’ group allocation. The flexion–rotation test was
conducted with the subject relaxed and recumbent. The
cervical spine was fully flexed with the occiput resting
against the examiner’s abdomen (Fig. 1). The head was
then rotated to the left and right. Each therapist was
required to interpret the flexion–rotation test results and
state whether the test was positive or not, based on the
therapist’s interpretation of range of motion. The range
of motion was considered limited when the therapist
determined a firm end feel and based on the therapists
interpretation there was a minimum of a 101 reduction
from the expected range. Pain provocation during the
flexion–rotation test was not tested in this study.
Thereafter, the goniometer was fixed and the flexion–
rotation test was repeated twice in each direction, with
one therapist ascertaining range in reverse order and
measurements recorded on each occasion. An interval of
30 s elapsed between each trial. Additionally, each
therapist was blinded to each other’s interpretation
and recordings.
At the time of assessment, cervicogenic and migraine
subjects completed a headache questionnaire (Niere and
Robinson, 1997) detailing their history of headache
frequency, intensity and duration. This questionnaire
 ARTICLE IN PRESS
M. Ogince et al. / Manual Therapy 12 (2007) 256–262 259

has been shown to be reliable and enables an index of Table 1

headache severity to be calculated (Niere and Robinson, The frequencies used to calculate sensitivity and specificity
1997). Additionally, the intensity of headache at the Cervicogenic headache Total
time of the assessment was determined on a 10 cm visual
analogue scale (VAS). The results of the questionnaire Positive Negative
and VAS were used to determine if the degree of a
Cervical flexion Positive 21/21 3/4a 24/25a
restriction of rotation in flexion was related to the Rotation test Negative 2/2a 32/31a 34/33a
severity of the headache symptoms in the cervicogenic Total 23/23a 35/35a 58
headache group. a
Therapist 1/Therapist 2.

3.2. Statistical analysis

50
All data were analysed using SPSS Version 11.0
statistical software (SPSS, Inc., Chicago, IL). In all cases
40
alpha was set at the .05 level. An analysis of variance

Range of Motion (degrees)

(ANOVA) and planned orthogonal comparisons were
used to analyse range of rotation with the cervical spine
in flexion between the 3 subject groups. An ANCOVA
was used to determine whether age and gender
accounted for the difference in range of rotation with
the cervical spine in flexion between the 3 groups. The
sensitivity and specificity of the flexion–rotation test
were analysed using cross tabulation and were deter-
mined with a receiver operating characteristic (ROC)
curve. To calculate the sensitivity and specificity the
migraine with aura and asymptomatic groups were
combined and then compared to the cervicogenic
headache group. The frequencies used to calculate
sensitivity and specificity are given in Table 1. The
dichotomous variables used to determine the sensitivity
and specificity were the therapists’ identification of the
presence or absence of C1/2 dysfunction and thereby
cervicogenic headache. The ROC curve was created with
the flexion–rotation range of motion values. The inter-
tester reliability of the flexion–rotation test was calcu-
lated from a cross-tabulation using kappa.
A variable headache severity index was calculated
using the method described by Niere and Robinson
(1997). The greater the score on the headache severity
index the greater the severity of headache. To determine
the relationship of the headache severity index score and
the VAS score to the range of rotation in flexion,
Pearson’s correlation analyses were used.
4. Results
The average range of unilateral rotation for both sides
was 391 (SD ¼ 6.9), 391 (SD ¼ 6.5) and 201 (SD ¼ 11)
for the migraine with aura, asymptomatic and cervico-
genic headache groups, respectively. For the cervico-
genic headache group the average range of unilateral
rotation refers to the most restricted side. The difference
between groups was significant (Po.001) (Fig. 2). The
results indicate that the range of rotation was signifi-
cantly reduced in the cervicogenic headache group when
30
20
10
0
Cervicogenic Migraine Asymptomatic
with aura
Fig. 2. The mean range of motion and 95% confidence interval of
unilateral cervical rotation in maximal flexion to each side for each
subject group.
compared to the migraine with aura and asymptomatic
subjects (Po.001). There was no significant difference in
range between the migraine with aura subjects and
asymptomatic subjects (P ¼ :971) (Fig. 2). There was a
significant negative correlation between age and ROM
(r ¼ :404, P ¼ :002), and the difference in ROM
between males (M ¼ 35:1, SD ¼ 9.2) and females
(M ¼ 29:7, SD ¼ 13.7) was close to significance,
tð52Þ ¼ 1:76, P ¼ :084. Therefore, age and gender were
both included as covariates in an ANCOVA, and a
significant result on ANCOVA (Po.001) revealed that
neither age nor gender accounted for the differences
between groups.
Sensitivity is the test’s ability to obtain a positive test
when the target condition is really present (Portney and
Watkins, 1993). Specificity is the test’s ability to obtain a
negative test when the condition is really absent
(Portney and Watkins, 1993). Positive predictive value
is the likelihood that person who tests positive actually
has the disease and negative predictive value is the
probability that a person who tests negative is actually
disease-free (Portney and Watkins, 1993). A positive
likelihood ratio indicates the increase in odds favouring
the condition given a positive test result and a negative
 ARTICLE IN PRESS
260 M. Ogince et al. / Manual Therapy 12 (2007) 256–262

Table 2 the ROC curve indicated that the test value, which
The sensitivity, specificity, positive and negative predictive values as provides the highest sensitivity and the lowest
well as likelihood ratios of the cervical flexion–rotation test
1-specificy, is 321 (cut-off score). That is, if the flexion
Therapist 1 Therapist 2 rotation-test value is less than or equal to 321 the test is
positive.
Sensitivity (%) 91.3 91.3 In examining the relationship between the headache
Specificity (%) 91.4 88.6
severity index score and the range of rotation in flexion,
Positive predictive value (%) 87.5 84.0
Negative predictive value (%) 94.1 93.9 there was no significant correlation rð23Þ ¼ :24, P ¼ :27.
Positive likelihood ratio 10.65 7.99 Additionally, when examining the relationship between
Negative likelihood ratio 0.095 0.098 severity of headache at the time of assessment on a VAS
to range of rotation in flexion, there was no significant
correlation rð23Þ ¼ :09, P ¼ :960.

1.00
5. Discussion

This study found the range of cervical rotation in

0.75 flexion was significantly reduced in the presence of C1/2
dominant cervicogenic headache when compared to a
control group of either asymptomatic or migraine with
Sensitivity

aura. These results concur with Hall and Robinson

0.50
0.25
0.00
0.00 0.2 0.50 0.75 1.00
1 – Specificity
Fig. 3. The diagnostic accuracy of the cervical flexion–rotation test.
The area under the curve is .91 (Po.001, CI ¼ 82%, CI ¼ 100%).
likelihood ratio indicates the change in odds favouring
the condition given a negative test result (Irwig et al.,
1994). Table 2 outlines the sensitivity, specificity,
positive predictive value, and negative predictive value
as well as likelihood ratios for each therapist. The mean
sensitivity and specificity for the two therapists were
91% and 90%, respectively (Po.001). Cross tabulation
revealed that 98.3% of the time both testers agreed on
their identification of C1/2 movement restriction and the
absence or presence of cervicogenic headache (Po.001).
The kappa value for the cervical flexion–rotation test
was .81, indicating excellent agreement (Landis and
Koch, 1977).
The ROC curve (Fig. 3) shows the relationship
between sensitivity and specificity. The area under the
curve represents the ability of the test to discriminate
between the diseased and nondiseased state. An ROC
curve revealed that presented with a randomly chosen
pair of patients, using the flexion–rotation test, the
clinician is able to make the correct diagnosis 91% of
the time (Po.001) (Fig. 3). Additionally, coordinates on
(2004) who evaluated the flexion–rotation test compar-
ing only cervicogenic headache subjects and asympto-
matic controls. Hall and Robinson (2004) demonstrated
that the average range of unilateral rotation to each side
was 281 and 451 for the cervicogenic and asymptomatic
groups, respectively. This study showed that average
range of unilateral rotation was 201 and 391 for the
cervicogenic and asymptomatic groups, respectively.
Additionally, the range of rotation in flexion in
asymptomatic controls is comparable with Amiri et al.
(2003) who found the average range of unilateral
rotation was 421.
This study also found there was no difference in range
of flexion between the migraine with aura and the
asymptomatic subjects. Migraine with aura has been
described as a disturbance of brain function or a
neurovascular event (Sanchez-del-Rio and Reuter,
2004) and consequently does not involve impairment
of the cervical spine. Hence, this result was not
unexpected and helps to confirm the lack of cervical
involvement in headache with aura patients.
The results of this study suggest that the clinical
method purported to bias rotation to the upper cervical
region is an accurate and reliable clinical measure in the
identification of dysfunction at the C1/2 level and in the
differential diagnosis of cervicogenic headache. Thus, in
accordance with Amiri et al. (2003) it would be
reasonable to recommend that the flexion–rotation test
be used in the assessment of patients with C1/2
dysfunction, for purposes of differential diagnosis and
assessment of treatment outcomes.
This is the first study to establish the sensitivity and
specificity of the cervical flexion–rotation test in
cervicogenic headache diagnosis. The average sensitivity
and specificity was 91% and 90%, respectively, with an
 ARTICLE IN PRESS
M. Ogince et al. / Manual Therapy 12 (2007) 256–262
overall diagnostic accuracy of 91%. This indicates that
the cervical flexion–rotation test has very high accuracy
in determining the presence of C1/2 involvement in
cervicogenic headache and for headache differential
diagnosis. Furthermore, the average kappa statistic of
.81 suggests excellent therapist agreement (Landis and
Koch, 1977). Bogduk (1997) concludes that for most
good clinical tests in physical examination, the kappa
value should range between .4 and .6. This study has also
determined a range of 321 as the cut-off value at which
the flexion rotation-test is deemed positive. This enables
the clinician to confidently interpret the results of the test
in clinical practice. Consequently, these results help to
establish the flexion–rotation test as a reliable measure
that assists in differential headache diagnosis and
determining the presence of C1/2 dysfunction.
It is important to note that at the time of testing
each therapist interpreted the flexion–rotation test
results by stating whether the test was positive or not,
based on the therapist’s interpretation of range of
motion. That is, the range of motion was considered
limited when the therapist determined a firm end feel
and based on the therapists interpretation there was a
minimum of a 101 reduction from the expected range.
Since it is impractical to use a CROM device in the
clinical environment this method more accurately
reflects those used by physiotherapists in the clinical
environment.
In contrast to Hall and Robinson (2004), this study
found that the severity of cervicogenic headache is not
related to the degree of restriction of rotation with the
cervical spine in maximal flexion. Given that the sample
of cervicogenic headache subjects was similar in both
studies it appears that this difference may be due to
alternative factors. Silberstein et al. (2001) contend that
many patients are not good observers of their own
complaints, even when those complaints are chronic.
Certainly, the authors observed that whilst subjects
completed the questionnaire they expressed difficulty in
recalling the intensity, frequency and duration of
symptoms. This may have skewed the results of the
questionnaire and thus account for the different results
in this study.
Using the VAS to assess the influence of intensity of
headache on range of rotation at the time of testing, this
study indicated that the severity of headache at the time
of assessment is not related to the degree of restriction
with the cervical spine in flexion. Clearly, this appears to
be advantageous as the diagnostic accuracy of the
flexion–rotation test is not influenced by the patient’s
head pain at the time of assessment.
When comparing the cervicogenic (n ¼ 23) and
asymptomatic groups (n ¼ 23), we were only able to
recruit 12 subjects in the migraine with aura group
(n ¼ 12). This sample corresponds to Ziegler and
Hassanein (1990) who showed that at most, only 30%
261
of the migraine population have migraine with aura. Of
the symptomatic subjects interviewed, 82% reported
overlapping cervicogenic and migraine symptoms and
thus were rejected from the study. This data corresponds
with Sjaastad and Bovim (1991), Fishbain et al. (2001)
and Nicholson and Gaston (2001) who report frequent
overlapping characteristics seen with cervicogenic and
migraine patients. This study’s population was limited
to cervicogenic and migraine with aura patients who
have symptoms characteristic of no overlap. Based on
the study sample, it appears that those who participated
represent a small percentage of the symptomatic
population who present in the clinical setting.
The study’s case control design and spectrum bias is
by no means a limitation, as in order to determine the
diagnostic accuracy of physical test the subjects must
represent either a purely positive disease state, that is,
cervicogenic headache or symptoms manifest of a
disease free state. Thus, this study has demonstrated
the cervical flexion–rotation test significantly assists in
C1/2 cervicogenic headache diagnosis when the sub-
jective criteria follow those of the IHS. Consequently,
this study sets the groundwork for future studies to
evaluate the cervical flexion–rotation test on a subject
population whose symptoms overlap cervicogenic and
migraine headache.
The authors acknowledge a number of limitations of
the study. In particular, the use of a single assessor to
identify C1/2 dysfunction using manual diagnosis. At
the time of this study no credible alternative reference
standard was available for determining the segmental
level of involvement for the high cervical spine. The only
other potential reference standard, double blind anaes-
thetic blocks, has not been attempted in the high cervical
spine. In addition, selection bias was unavoidable in
obtaining a large sample size. Not all subjects presenting
with the relevant condition were included in order of
entry neither was the selection random.
6. Conclusion
This study has demonstrated that the cervical flexion–
rotation test is extremely reliable and has high sensitivity
and specificity in detecting the presence of C1/2 rotation
restriction in patients with cervicogenic headache. The
cut of value for a positive test is range of rotation less
than 321. The flexion–rotation test is a simple, non-
invasive test that can be easily applied in the clinical
setting.
Acknowledgements
The authors thank Mr. Wim Dankaerts for assisting
with the testing procedures. The authors also acknowl-
 ARTICLE IN PRESS
262 M. Ogince et al. / Manual Therapy 12 (2007) 256–262
edge the School of Physiotherapy at Curtin University
of Technology for providing assistance to present the
research findings at the WCPT Conference in Barcelona
2003 as well as the Australian Physiotherapy Associa-
tion WA Branch for awarding the MJ Rosen Scholar-
ship to present this research at the Annual Conference
of the APTA in Chicago 2004.
References
Amiri M, Jull G, Bullock-Saxton J. Measuring range of active cervical
rotation in a position of full head flexion using the 3D Fastrak
measurement system: an intra-tester reliability study. Manual
Therapy 2003;8:176–9.
Blau J, MacGregor E. Migraine and the neck. Headache 1994;34:88–90.
Bogduk N. Cervical causes of headache and dizziness. In: Boyling J,
Palastanga N, editors. Grieves modern manual therapy: the
vertebral column. Edinburgh: Churchill Livingstone; 1994.
p. 317–32.
Bogduk N. Musculoskeletal pain: toward precision diagnosis. In:
Proceedings of the eighth world congress on pain. Seattle: IASP
Press; 1997. p. 507–26.
Bogduk N, Corrigan B, Kelly P. Cervical headache. Medical Journal
of Australia 1985;143:202–7.
Capuano-Pucci D, Rheault W. Intratester reliability of the cervical
range of motion device. Archives of Physical Medicine and
Rehabilitation 1991;72:338–40.
Edmeads J. Disorders of the neck: cervicogenic headache. In:
Silberstein SD, Lipton RB, Dalessio DJ, editors. Wolff’s headache
and other head pain. Oxford: Oxford University Press; 2001.
p. 447–58.
Fishbain D, Cutler R, Cole B, Rosomoff H, Rosomoff R. Interna-
tional Headache Society headache diagnostic patterns in pain
facility patients. Clinical Journal of Pain 2001;17:78–93.
Hall T, Robinson K. The flexion–rotation test and active cervical
mobility—a comparative measurement study in cervicogenic head-
ache. Manual Therapy 2004;9:197–202.
Headache Classification Committee of the International Headache
Society. Classification and diagnostic criteria for headache
disorders, cranial neuralgias and fascial pain. Cephalalgia 1988;8:
1–96.
International Headache Society. The International classification of
headache disorders 2nd edition. Cephalalgia 2004;24(Suppl 1):9–160.
Irwig L, Tosteson A, Gatsonis C. Guidelines for metanalyses
evaluating diagnostic tests. Annals of Internal Medicine 1994;120:
667–76.
Jaeger B. Are cervicogenic headaches due to myofascial pain and
cervical spine dysfunction? Cephalalgia 1989;9:157–64.
Jensen O, Nielsen F, Vosmar L. An open study comparing manual
therapy with the use of cold packs in the treatment of post
concussional headache. Cephalalgia 1990;10:241–9.
Jull G, Bogduk N, Marsland A. The accuracy of manual diagnosis for
cervical zygapophyseal joint pain syndromes. Medical Journal of
Australia 1988;148:233–6.
Landis J, Koch G. The measurement of observer agreement for
categorical data. Biometrics 1977;33:159–74.
Lewit K. Pain arising in the posterior arch of the Atlas. European
Neurology 1977;16:263–9.
Lord S, Barnsely L, Wallis B, Bogduk N. Third occipital headache: a
prevalence study. Journal of Neurology, Neurosurgery and
Psychiatry 1994;57:1187–90.
Maitland G, Hengeveld E, Banks K, English K. Vertebral manipula-
tion, 6th ed. London: Butterworth Heinemann; 2001.
Mersky H, Bogduk N. Classification of chronic pain. 2nd ed. Seattle:
IASP Press; 1994.
Nicholson G, Gaston J. Cervical headache. Journal of Orthopedic and
Sports Physical Therapy 2001;31:184–93.
Niere K, Robinson P. Determination of manipulative physiotherapy
treatment outcome in headache patients. Manual Therapy 1997;2:
199–205.
Placzek J, Pagett B, Roubal P, Jones B, McMichael H, et al. The
influence of cervical spine on chronic headache in women: a pilot
study. The Journal of Manual and Manipulative Therapy 1999;7:
33–9.
Portney G, Watkins M. Foundation of clinical research: applications
to practice. London: Appleton and Lange; 1993.
Rheault W, Albright B. Intertester reliability of the cervical range of
motion device. Journal of Orthopaedic and Sports Physical
Therapy 1992;15:147–50.
Sanchez-del-Rio M, Reuter U. Migraine aura: new information on
underlying mechanisms. Current Opinion in Neurology 2004;17(3):
289–93.
Sandmark H, Nisell R. Validity of five common manual neck pain
provocation tests. Scandinavian Journal of Rehabilitation Medi-
cine 1995;27:131–6.
Schoensee H, Jensen G, Nicholson G, Gossman M, Katholi C. The
effect of mobilisation on cervical headaches. Journal of Orthopae-
dic and Sports Physical Therapy 1995;21:181–96.
Silberstein S, Lipton R, Dalessio D. Overview, diagnosis and
classification of headache. In: Silberstein S, Lipton R, Dalessio
D, editors. Wolff’s headache and other head pain. Oxford: Oxford
University Press; 2001. p. 6–26.
Sjaastad O, Bovim G. Cervicogenic headache: the differentiation from
common migraine, an overview. Functional Neurology 1991;6:
93–100.
Sjaastad O, Fredriksen T, Pfaffenrath V. Cervicogenic headache:
diagnostic criteria. Headache 1998;38:442–5.
Stratton S, Bryan J. Dysfunction, evaluation, and treatment of the
cervical spine and thoracic inlet. In: Donatelli R, Wooden M,
editors. Orthopaedic physical therapy. New York: Churchill
Livingstone; 1994. p. 77–122.
Treleaven J, Jull G, Atkinson L. Cervical musculoskeletal dysfunction
in post-concussional headache. Cephalalgia 1994;14:273–9.
Vernon H, Steiman I, Hagino C. Cervicogenic dysfunction in muscle
contraction headache and migraine: a descriptive study. Journal of
Manipulative and Physiological Therapeutics 1992;15:418–29.
Watson D, Trott P. Cervical headache. An investigation of natural
head posture and upper cervical flexor muscle performance.
Cephalalgia 1993;13:272–84.
Whittingham W, Ellis W, Molyneux T. The effect of manipulation
(toggle recoil techniques) for headaches with upper cervical joint
dysfunction: a pilot study. Journal of Manipulative and Physiolo-
gical Therapeutics 1994;17:369–75.
Ziegler D, Hassanein R. Specific headache phenomena: their frequency
and coincidence. Headache 1990;30:152–6.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 263–270

www.elsevier.com/locate/math

Original article

Accuracy and reliability of observational motion analysis in

identifying shoulder symptoms
Brendan W. Hickeya, Stephan Milosavljevica,, Melanie L. Bellb, Peter D. Milburna
a
Centre for Physiotherapy Research, School of Physiotherapy, University of Otago, P.O. Box 56, Dunedin, New Zealand
b
Department of Preventive and Social Medicine, University of Otago, P.O. Box 56, Dunedin, New Zealand
Received 22 December 2004; received in revised form 20 February 2006; accepted 8 May 2006

Abstract

Introduction: Aberrations in shoulder movement patterns are believed to be associated with the presence of shoulder symptoms.
However, the detection of movement aberrations has not been rigorously investigated. It is possible that manipulative
physiotherapists use the clinical history to prejudge the existence of aberrations, rather than the actual observation of the movement
pattern itself. There is a need to determine whether physiotherapists, in the absence of a clinical history, can relate observed
anomalies of shoulder movement to the presence of symptoms and to determine the reliability for observation of such anomalous
shoulder movement.
Methods: The sample comprised of 9 symptomatic subjects recruited from four physiotherapy clinics in Christchurch, New Zealand
and a further 11 asymptomatic subjects recruited from Christchurch’s general population. They were videotaped performing
shoulder flexion, abduction, and scapular plane abduction. The video-recordings were evaluated by 11 manipulative
physiotherapists who did not know which subjects were symptomatic and who were thus required to judge the symptomatic
status as: asymptomatic, symptomatic left, symptomatic right or symptomatic both. Additionally, each physiotherapist completed a
survey on each of the 20 subjects regarding the type of movement anomaly that was perceived (e.g. too much scapular elevation, too
little glenohumeral movement, etc). Classification accuracy (percentage of correct responses) and agreement (k) among
physiotherapists were computed.
Results: Out of the 220 responses by the physiotherapists regarding symptomatic status, 58% were correct, with 68%
asymptomatic, 71% symptomatic left and 30% symptomatic right subjects correctly classified. Reliability analysis showed k
statistics for all subjects was 0.23, for asymptomatic subjects 0.22, symptomatic left 0.34, and symptomatic right 0.17. Only five
subjects had two or more evaluators agree on the type of anomalous movement.
Conclusions: Although movement analysis is considered an integral part of a physiotherapist’s skill this research has shown that a
sample of experienced manipulative physiotherapists had difficulty in determining the symptomatic status of patients with clinically
diagnosed shoulder complaints by movement analysis alone.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Shoulder; Movement; Accuracy; Reliability

1. Introduction aberrations in shoulder movement patterns are thought

Shoulder pain is a common musculoskeletal com-
plaint with an estimated lifetime prevalence of 10% in
the adult population (van der Heijden, 1999). Although
Corresponding author. Tel.: +64 3 479 7460; fax: +64 3 479 8414.
E-mail address: stephan.milosavljevic@otago.ac.nz
(S. Milosavljevic).
to be associated with shoulder symptoms (Sahrmann,
2002) it is not clear how this interaction occurs. It is
common clinical practice for physiotherapists to use
bilateral assessment and comparison of scapulo-humeral
movement to predict the dysfunctional status of the
shoulder yet there is little evidence to indicate that such
observation is either associated with shoulder dysfunc-
tion or can be agreed upon between physiotherapists.

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.05.005
 ARTICLE IN PRESS
264 B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270
There is a need to determine whether manipulative
physiotherapists, in the absence of a clinical history, can
accurately observe whether a given subject has shoulder
symptoms, can decide which shoulder is symptomatic
and, can reliably agree on the observed anomaly of
shoulder movement.
2. Background
Shoulder complex movement has been described as a
synchronous contribution of the glenohumeral, acro-
mioclavicular, sternoclavicular, and scapulothoracic
joints (Inman et al., 1944; Culham and Peat, 1993)
and the kinematic relationship between the scapula and
humerus has been researched extensively. Although
earlier literature considered that the humerus and
scapula moved in a 2:1 ratio, respectively (Codman,
1934) recent research has demonstrated variable hu-
merus-to-scapula movement ratios, ranging from 0.91:1
to 7.51:1 depending on the phase of movement (Bagg
and Forrest, 1988; Talkhani and Kelly, 1997; Manda-
lidis et al., 1999).
Three-dimensional (3D) changes in scapular orienta-
tion have also been observed during shoulder elevation
including changes in lateral and external rotation and
posterior tilting (Ludewig et al., 1996). Concentric and
eccentric activity, external loads, changes in velocity,
different spinal postures, and age are also thought to
influence the scapular and humeral relationship (Lude-
wig and Cook, 1996; Kebaetse et al., 1999; Pascoal et al.,
2000; Talkhani and Kelly, 2001; Borstad and Ludewig,
2002; Sugamoto et al., 2002; Finley and Lee, 2003).
In symptomatic individuals, changes in scapular
orientation including decreased posterior tilt and
decreased external rotation (‘scapular winging’) have
been noted during elevation in the scapular plane
(Warner et al., 1992; Lukasiewicz et al., 1999; Ludewig
and Cook, 2000; Borstad and Ludewig, 2002; Hébert
et al., 2002). Such reductions are thought to reduce the
available sub-acromial space during shoulder elevation,
and increase the risk of sub-acromial impingement
(Kamkar et al., 1993).
As movement timing changes have been associated
with disorders within the stabilizing and activating
muscular complex of the shoulder (Kibler, 1998;
Sahrmann, 2002) the evaluation of movement timing
and the active correction of asymmetry are considered
important for successful rehabilitation (Hayes et al.,
2001; Sahrmann, 2002; Magarey and Jones, 2003).
Although detailed descriptions of clinical assessment
of shoulder movement exist, the ability of physiothera-
pists to diagnose movement disorders within the
shoulder complex has yet to be rigorously examined
(Kapandji, 1982; Magarey and Jones, 2003). The issues
of reliability and accuracy of diagnosis of movement
disorder in particular need to be ascertained.
Poor reliability, high costs and limited clinical
applicability of scapular orientation measurement with
3D digital tracking devices have led to the use of video-
tape recording of movement as a tool for shoulder
complex movement analysis (Barnett et al., 1999; Yanai
and Hay, 2000; Johnson et al., 2001; Karduna et al.,
2001; Ackermann et al., 2002; Magarey and Jones, 2003;
Lowe, 2004). However, only two studies have assessed
how well physiotherapists can visually evaluate a
shoulder complex for movement anomalies (Babyar,
1996; Kibler et al., 2002). A four category classification
system for qualitatively describing video-recorded scap-
ular movement anomalies was used by Kibler et al.
(2002) on 20 symptomatic and six asymptomatic
subjects during scapular and frontal plane abduction.
Fair-to-moderate agreement for intra- and inter-rater
reliability (Landis and Koch, 1977) was noted for two
orthopaedic surgeons (k ¼ 0:59 and k ¼ 0:31), and two
physical therapists (k ¼ 0:49 and k ¼ 0:42). However,
despite the use of both symptomatic and asymptomatic
subjects, no attempt was made to relate observed
movement patterns to the presence of symptoms.
In order to consider how well physiotherapists can
visually identify the presence of symptomatic shoulder
movement disorders the aims of this study are:
(1) to determine whether the observation skills of
manipulative physiotherapists alone allow them to
correctly decide;
(a) if a given subject has shoulder symptoms,
(b) which shoulder is symptomatic, and
(2) to describe the nature of the observed aberration and
whether physiotherapists can agree on this descrip-
tion.
3. Methods
Following approval from the Canterbury Regional
Ethics Committee 9 subjects with symptomatic
shoulders were recruited from the clinical caseloads of
the manipulative physiotherapists working in four
private physiotherapy clinics in Christchurch, New
Zealand. The recruiting physiotherapists were asked to
evaluate patients with shoulder symptoms using their
standard clinical examination and to provisionally
recruit them if they met the inclusion criteria described
below. They were also screened by the principal
investigator (BWH) to determine eligibility.
Inclusion criteria
(1) shoulder symptoms below the acromion for greater
than 1 month,
 ARTICLE IN PRESS
B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270
(2) currently on active physiotherapy treatment for their
shoulder,
(3) aged between 18 and 55 and,
(4) capable of at least 1501 of shoulder flexion and 1401
of shoulder abduction in both shoulders, measured
against the vertical plane.
(5) For asymptomatic subjects—no shoulder symptoms
requiring treatment in the previous year.
Exclusion criteria
To exclude shoulder dysfunction due to spinal or non-
mechanical causes, and to minimize for asymmetric
postural anomalies and surface marks.
(1) clinical reproduction or evidence of shoulder symp-
toms with any cervical spine movement,
(2) a history of systemic inflammatory disease,
(3) a history of neurological symptoms including
descriptions of numbness, tingling and/or other
sensory disturbance in the shoulder and upper limb
in the presence of upper limb weakness,
(4) a scoliosis producing a visible rib hump in neutral
standing posture,
(5) any prior history of shoulder surgery,
(6) any hearing loss limiting their ability to understand
verbal instructions, and
(7) any tattoos on the posterior aspect of the trunk
between the cervico-thoracic and thoraco-lumbar
junctions.
A further 11 asymptomatic subjects with no history of
previous shoulder complaint in the previous 12 months
were recruited from the general population of the
Christchurch metropolitan area. These subjects were
chosen so that their ages and anthropometric character-
istics were similar to the symptomatic subjects.
Following pilot testing all subject recording took
place in the same room at one of the participating
clinics, in order to maintain consistency of visual
background and to control for environmental factors.
Sex, age, hand dominance, height, weight and body
mass index (BMI) were recorded for each subject.
Subjects were then instructed on the following three
shoulder movements to be completed for evaluation:
shoulder flexion (about a coronal axis), scapular plane
abduction (approximated as between 301 and 401
anterior to the frontal plane), and shoulder abduction
(about a sagittal axis). Following instruction and
movement familiarization each subject was positioned
for video recording. The subject stood barefoot, with
feet placed on a predetermined mark and arms by their
side. Each subject stood equidistant from the overhead
lighting and closest corners of the room. A video camera
(Sony Digital Handicam 8 mm, PCR-TRV320E PAL)
was placed on a tripod 1.2 m high and 2 m behind the
subjects. The field of view was adjusted to include from
265
the occiput to the thoraco-lumbar junction, and the
olecranon processes on either side.
Once recording had begun, each subject used the
investigator’s verbal cues to initiate and then repeat the
shoulder flexion, abduction and scapular plane abduction
movements. The between-subject order in which these
movements were video-recorded was randomized prior to
recording to counter any possible within group summa-
tive effects of movement on symptom reproduction. So
that the physiotherapists could not determine sympto-
matic status by order of presentation, the master
videotape was constructed with the order of presentation
randomized for the presence or absence of symptoms.
The master tape was copied onto 11 standard VHS
tapes and given to the 11 participating manipulative
physiotherapists. These physiotherapists were selected
from the staff and graduates of the University of Otago
and were based in Dunedin, New Zealand. The
advantage of a considerable geographical distance
between the two cities thus minimized the likelihood of
evaluator recognition of subjects. The physiotherapists
had all completed graduate-level qualifications in
manipulative physiotherapy between 1985 and 2003
and had been practising as a physiotherapist for a
minimum of 5 years. They were instructed not to confer
with each other prior to or after their videotape
evaluation session, and to perform their evaluation
session in isolation.
Each therapist was provided with subject evaluation
forms and an information sheet which detailed the
purpose of the study, the terminology to be used for
classifying any identified anomalous movements, and
instructions on completing the evaluation forms. The
nomenclature and description of the movement anoma-
lies (see Table 1) were adapted from a reference text on
movement impairment syndromes (Sahrmann, 2002).
Upon viewing the video, each evaluator was required
to nominate: (1) whether they thought the subject was
symptomatic or not; (2) which shoulder(s) they con-
sidered were symptomatic, if any, and; (3) to state what
movement anomalies they used to judge the shoulder(s)
as having symptoms. If an evaluator judged a subject to
be symptomatic, they were instructed to choose the one
anomalous movement that most strongly caused them
to believe the subject had a symptomatic shoulder. If the
subject was thought to be asymptomatic, they were
instructed not to complete the remainder of the form.
The evaluators were allowed to observe each subject’s
recording as often as required in order to make their
best possible decision.
4. Statistical analyses
Our quantitative data consisted of the rating (asymp-
tomatic, symptomatic right, symptomatic left, sympto-
 ARTICLE IN PRESS
266 B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270

Table 1
Definitions of scapular movement for evaluators

Movement Definition

Scapular lateral rotation Combined lateral and superior movement of the inferior angle of the scapula, through an axis located
perpendicular to the frontal plane
Scapular elevation Superior translation of the whole scapula, with minimal to no medial or lateral movement
Scapular abduction Lateral translation of the vertebral border of the scapula, with minimal to no superior or rotation
movement
Scapular anterior tilt Tilting of the scapula on an oblique angle, such that the acromion moves anterior and inferior, and the
inferior angle of the scapula moves posteriorly away from the rib cage
Scapular winging Internal rotation of the scapula, such that the whole vertebral border of the scapula moves posteriorly
away from the ribcage
Hypermobile glenohumeral range Excessive movement at the glenohumeral joint, in the absence of scapular movement anomalies
of motion
Hypomobile glenohumeral range Diminished movement at the glenohumeral joint, in the absence of scapular movement anomalies
of motion
Other Another movement anomaly not covered by the criteria already listed

(Adapted from Sahrmann, 2002).

matic both) that each of the 11 physiotherapists gave to
each of the 20 subjects, as well as their true status, as
determined by the recruiting physiotherapists and the
primary investigator (BWH). All statistical analyses
were performed using SAS statistical software (SAS
Institute, 1999).To compute chance adjusted agreement
(k) between multiple evaluators the SAS MAGREE
macro (SAS Technical Support, 2004) was used.
Generalized linear mixed (logistic) models, as imple-
mented in the SAS macro GLIMMIX were used to
model accuracy, i.e. the proportion of correct responses.
These models allow for the lack of independence in the
data, i.e. that there are multiple observations for each
evaluator and for each subject, by including evaluator
and subject as random effects. For more on these
models see Brown and Prescott (1999). Backward
selection for the fixed effect variables was used to
determine the best model, starting with the symptomatic
status (left, right, neither), age, BMI and sex. Statistical
significance of Po0:05 was used.
The qualitative data on the nature of the anomalous
movement, as judged by each of the physiotherapists,
was summarized by frequencies of the survey responses.
5. Results
Table 2 shows subject characteristics. Although the
asymptomatic subjects were on average 4.7 years
younger and had a wider weight distribution, the
comparison of height, weight and BMI revealed similar
mean scores. Four of the 220 evaluation sheets
submitted by the 11 physiotherapists were incompletely
recorded for choosing a subject’s symptomatic status,
and these responses were not used in the analyses. Fig. 1
shows the symptomatic status chosen by each of the 11
physiotherapists for each subject, as well as the true
status. Individual physiotherapist accuracy ranged from
45% to 75%.
Table 3 shows the proportions and confidence
intervals of evaluator accuracy by subject status, as
estimated by the logistic mixed model. Evaluators were
most accurate classifying symptomatic left subjects. The
proportion of these responses that were correct was
estimated to be 71% (i.e. 71% of evaluators classified a
subject as being symptomatic on their left side after
viewing video footage of a symptomatic left subject).
This was comparable to the asymptomatic group, with a
proportion of 68% correctly classifying the subject. The
right-sided symptomatic group was accurately judged by
only 30% of responses. The evaluator accuracy was not
statistically significantly different between asympto-
matic and left-sided symptomatic groups (P ¼ 0:8302).
However, evaluators were statistically significantly more
accurate for these groups compared to the right-sided
symptomatic group (P ¼ 0:0281 for left vs right and P ¼
0:0070 for asymptomatic vs right). Subject age, BMI and
sex were not significant predictors of accuracy in the
model.
Table 4 shows the k values for inter-rater agreement
by response for the evaluators. The overall level of
agreement was k ¼ 0:23. When examined relative to
subject status, the asymptomatic responses demon-
strated a k value of 0.22, the symptomatic left responses
had a k value of 0.34 and the symptomatic right
responses had a k value of 0.17. It should be noted that
these values only represent the agreement of responses
by evaluators, and do not take into account whether
these responses were correct.
There were only a small number of instances where
two or more evaluators agreed within a subject on the
nature of anomaly. Two subjects had three evaluators
who agreed, two subjects had two evaluators agree, and
one subject had two separate sets of agreement within
 ARTICLE IN PRESS
B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270 267

Table 2
Subject demographic means and standard deviations, mean (SD)

Total (n ¼ 20) Symptomatic (n ¼ 9) Asymptomatic (n ¼ 11)

Age (years) 32.7 (12.4) 35.1 (14.4) 30.7 (10.8)

Height (metres) 1.72 (0.08) 1.73 (0.10) 1.72 (0.08)
Weight (kg) 77.3 (13.6) 76.3 (9.0) 78.1 (16.9)
Body mass index (kg/m2) 26.1 (3.8) 25.9 (3.5) 26.2 (4.1)
Symptomatic side Right ¼ 6 Not applicable
Left ¼ 3

Fig. 1. Distribution of evaluator ratings for individual subjects. (+)
Marks the subject’s actual symptomatic status. (o) Represents one
evaluator’s choice for classifying the subject’s status.
Table 3
Accuracy of evaluators as estimated by the logistic mixed model
Status Proportion (95% CI)
Fig. 2. Frequency of anomalous movements selected by evaluators.
Key: MSE, Too much scapular elevation; LSE, Too little scapular
elevation; MSA, Too much scapular abduction; LSA, Too little
scapular abduction; MSLR, Too much scapular lateral rotation;
LSLR, Too little scapular lateral rotation; MSW, Too much scapular
winging; MSAT, Too much scapular anterior tilt; LGHM, Too little
glenohumeral movement; MGHM, Too much glenohumeral move-
ment; Other.

Asymptomatic 68.0% (53.2–79.9)

Symptomatic right 30.1% (15.7–49.9) to describe the anomalous movement (up to five criteria
Symptomatic left 71.0% (43.3–88.7) for some subjects), despite instructions to choose only
one criterion. These responses were included in the
Total 57.5% (43.6–70.3)
frequencies of anomalies. These results are displayed in
Fig. 2, as the number of instances each scapular
Table 4 anomaly was chosen by the evaluators.
k statistics for agreement amongst evaluators by response

Status Kappa (95% CI)

6. Discussion
Asymptomatic 0.22 (0.16–0.28)
Symptomatic right 0.17 (0.11–0.23) We have modelled the accuracy, as defined by the
Symptomatic left 0.34 (0.29–0.40)
proportion of four possible responses that were correct
Symptomatic both 0.01 (0.05 to 0.07)
for 11 physiotherapists rating 20 subjects, by video
Overall 0.23 (0.19–0.27) recording. Additionally we modelled the inter-rater
reliability (chance corrected agreement) as defined by
the k statistic. The type of movement anomaly was also
their dataset (which did not agree with each other). recorded.
Thirty-six responses regarding the nature of a subject’s Examination of the overall accuracy of the evaluators
anomalous movement checked more than one criterion demonstrated correct classification in 57.5% of the
 ARTICLE IN PRESS
268 B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270
subjects. Unlike the k statistic, there are no guidelines
for defining nominal categories of high, good, moderate,
fair or poor for accuracy. Although this is better than
chance (25% for four possible responses), this is not a
clinically acceptable level of accuracy.
There were some trends in evaluator accuracy
dependent on the subject’s symptom status. Asympto-
matic subjects were correctly identified in over two-
thirds (71%) of instances, which indicates this group of
manipulative physiotherapists is fairly capable of
accurately identifying asymptomatic shoulder move-
ment in an adult population.
Within the symptomatic group there was a significant
divergence of accuracy dependent on which shoulder
was symptomatic. Although it is possible that the
symptomatic left group had shoulder pathologies that
more readily demonstrated anomalous movement pat-
terns, it is more likely that it was due to chance as there
were only three left-sided symptomatic subjects. As all
subjects were right-handed it is also possible that hand
dominance plays a role in the perception of anomalous
movement. It has been suggested that anomalies will
occur in the static orientation of the scapula predomi-
nantly on the dominant shoulder (Sobush et al., 1996)
and also that anomalies in static scapular orientation
relate to dysfunctional movement patterns (Sahrmann,
2002). If movement pattern differences exist at the
dominant shoulder, symptomatic movement anomalies
on the dominant side may not be easily visualized by
physiotherapists due to the confluence of relevant and
irrelevant scapular positional anomalies. These hypoth-
eses require a larger study to clarify the reasons for this
divergence in accuracy.
The overall k value for agreement regarding symptom
status was 0.23, which can be considered only fair
agreement (Landis and Koch, 1977). A comparison to
the four pre-determined movement pattern categories by
Kibler et al. (2002) shows their values of agreement
between two physical therapists achieved moderate
agreement (k ¼ 0:42). While a direct comparison cannot
be made as Kibler et al. (2002) did not ask the
evaluators to gauge symptom status, it would be
reasonable to expect that with a much larger group of
evaluators as in the present study, the levels of
agreement would diminish somewhat. Overall, this
group of manipulative physiotherapists was capable of
agreeing on subject status only mildly better than
chance.
Only five subjects were agreed on by two or more
evaluators for accurate symptom status and the nature
of anomalous movement. It is clear from these results
that evaluators, when given the freedom to classify
anomalous shoulder movements via multiple choices of
theorized movement dysfunction, find it difficult to
come to any agreement. This was further demonstrated
by the large number of responses which chose multiple
anomalies, despite instructions to check only one
anomaly from the list. Out of the 220 responses, there
were 109 symptomatic responses, and of these there
were 35 (32%) with multiple responses. Five of the
evaluators did not submit any multiple responses, three
submitted one or two multiple responses, and three
submitted the remainder of the multiple responses. We
chose to include all responses in our summary because
we felt there was valuable information within them,
although it does mean that care should be taken in
interpreting the results. It is unknown why these
evaluators recorded their evaluations this way, whether
they were poor at following instructions, the instructions
were not clear enough, or they felt that the symptoms
could not be narrowed to one movement anomaly. A
possible limitation of this study is the use of a
classification system of movement anomalies adapted
from the work of Sahrmann (2002). It is possible that
some of the participating manipulative physiotherapists
were not familiar with this classification system, used
other observational criteria for analysis, and/or did not
believe that only one movement anomaly would occur in
a dysfunctional shoulder. Although in common clinical
use, with some evidence for the presence of such
movement disorders, the reliability and validity of this
classification system for clinically identifying true
disorders of shoulder movement has yet to be deter-
mined.
We compared our qualitative movement results with
other published research. Although ‘too much scapular
elevation’ was the most frequent response it has only
been identified as an anomalous movement in one study
(Babyar, 1996) yet included as a clinical syndrome
category by Kibler et al. (2002). ‘Too much scapular
winging’ was also a common choice and its relationship
with shoulder pathology has been well established
(Ludewig and Cook, 2000; Hébert et al., 2002). ‘Too
little glenohumeral movement’ and ‘too much scapular
lateral rotation’ were also reported by our evaluators yet
decreased glenohumeral movement is only alluded to by
Babyar (1996), and excessive scapular lateral rotation
has only been reported by Graichen et al. (2001).
Surprisingly, ‘too much scapular anterior tilt’ was not a
common choice, despite the evidence for such a
relationship (Mottram, 1997; Kibler, 1998; Lukasiewicz
et al., 1999; Ludewig and Cook, 2000). The overall
frequency and variation of responses suggests manip-
ulative physiotherapists may not be looking for the few
scapular movement anomalies which have more reliably
been shown to be relevant in symptomatic shoulders,
such as scapular winging and scapular anterior tilt. Thus
it may be timely for further review of the literature by
those who practise clinical evaluation and management
of musculoskeletal dysfunction in the shoulder. It may
be that increased attention to the disorders of scapula
position known to occur during dysfunctional shoulder
 ARTICLE IN PRESS
B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270
complex movement will be a future issue in successful
management of the problematic shoulder.
A high number of responses were marked ‘Other’ as a
response for the nature of anomalous movement. The
descriptions tended to report anomalies in terms of the
poor control of the scapulothoracic muscles (e.g. ‘poor
motor control serratus anterior/lower traps’) or as some
variation in speed or eccentric control (e.g. ‘faster return
left scapula during abduction’). It is apparent that some
physiotherapists preferred to define movement anoma-
lies in terms of possible neuromuscular causes, instead
of their biomechanical consequences.
We did not to arrive at an inclusion diagnosis for the
symptomatic subjects by radiological investigation (such
as X-ray, MRI and/or ultrasound), and further inves-
tigation by orthopaedic referral was also not done. The
diagnosis and suitability of subjects with primary
shoulder dysfunction was made by the recruiting
physiotherapist in conjunction with the screening
process of the first author. Thus it is possible that some
of these subjects had conditions that did not lend
themselves to the presence of anomalous shoulder
movement, given that anomalous movement in sympto-
matic shoulders has been most consistently demon-
strated in subjects with clinically diagnosed shoulder
impingement (Lukasiewicz et al., 1999; Ludewig and
Cook, 2000; Hébert et al., 2002). Some evaluators may
also have felt limited in their ability to accurately judge
the presence of relevant movement anomalies given no
further clinical history on the subjects was provided,
however a goal of this study was to ascertain what
accuracy could be achieved in the absence of other
information.
Care should be taken in generalizing these results.
Each of the evaluating physiotherapists had graduate-
level manipulative physiotherapy training, and each was
affiliated with the University of Otago. Newer graduates
or musculoskeletal physiotherapists without manipula-
tive physiotherapy training may demonstrate different
accuracy from the results presented here, although we
expect them to be fairly similar or less accurate as a
whole.
7. Conclusions
Movement analysis is considered an integral part of a
physiotherapist’s skill. Yet this research has shown that
a sample of experienced manipulative physiotherapists
had difficulty determining the symptomatic status of a
sample of patients with clinically diagnosed shoulder
complaints by movement analysis alone. Paradoxically,
based on this movement analysis, they were relatively
successful in determining whether a person was asymp-
tomatic. The implications therefore are that they can
detect asymmetric movement but are poor at determin-
269
ing a relationship between such asymmetry and which
shoulder is symptomatic. Given that they were not
aware of the subjects’ symptomatic status and clinical
history this may not be surprising. However the low
level of accuracy allows us to determine that movement
analysis per se cannot be disassociated from the clinical
history. Manipulative physiotherapists who place a high
level of importance on movement analysis to determine
the dysfunctional status of a given shoulder complaint
need to consider whether the anomalous movement
observations they make are relevant to the presenting
shoulder complex disorder. It is easy to associate or ‘‘fit’’
a given movement pattern(s) once the clinical history is
known. However this research indicates that it is
difficult to reliably ascertain shoulder movement
anomalies in the absence of a clinical history.
Acknowledgements
The authors would like to thank the Barrington
Physiotherapy Clinic, Riccarton Physiotherapy, and
Physiosouth for acting as sources for study subjects
and the subjects who participated in the research.
References
Ackermann B, Adams R, Marshall E. The effect of scapula taping on
electromyographic activity and musical performance in profes-
sional violinists. Australian Journal of Physiotherapy
2002;48(3):197–203.
Babyar SR. Excessive scapular motion in individuals recovering from
painful and stiff shoulders: causes and treatment strategies.
Physical Therapy 1996;76(3):226–38.
Bagg SD, Forrest WJ. A biomechanical analysis of scapular rotation
during arm abduction in the scapular plane. American Journal of
Physical Medicine and Rehabilitation 1988;67(6):238–45.
Barnett ND, Duncan RDD, Johnson GR. The measurement of three
dimensional scapulohumeral kinematics–—a study of reliability.
Clinical Biomechanics 1999;14(4):287–90.
Borstad JD, Ludewig PM. Comparison of scapular kinematics
between elevation and lowering of the arm in the scapular plane.
Clinical Biomechanics 2002;17(9–10):650–9.
Brown H, Prescott R. Applied mixed models in medicine. Chinchester:
Wiley; 1999 Chapter 3.
Codman EA. The shoulder: rupture of the supraspinatus tendon and
other lesions in or about the subacromial bursa. Privately printed
1934, Boston.
Culham E, Peat M. Functional anatomy of the shoulder complex.
Journal of Orthopaedic and Sports Physical Therapy
1993;18(1):342–50.
Finley MA, Lee RY. Effect of sitting posture on 3-dimensional
scapular kinematics measured by skin-mounted electromagnetic
tracking sensors. Archives of Physical Medicine and Rehabilitation
2003;84(4):563–8.
Graichen H, Stammberger T, Bonel H, Weidemann E, Englmeier K-H,
Reiser M, et al. Three-dimensional analysis of shoulder girdle and
supraspinatus motion patterns in patients with impingement
syndrome. Journal of Orthopaedic Research 2001;19(6):1192–8.
 ARTICLE IN PRESS
270 B.W. Hickey et al. / Manual Therapy 12 (2007) 263–270
Hayes K, Walton JR, Szomor ZL, Murrell GAC. Reliability of five
methods for assessing shoulder range of motion. Australian
Journal of Physiotherapy 2001;47(4):289–94.
Hébert LJ, Moffet H, McFadyen BJ, Dionne CE. Scapular behavior in
shoulder impingement syndrome. Archives of Physical Medicine
and Rehabilitation 2002;83(1):60–9.
Inman VT, Saunders JBDC, Abbott LC. Observations on the function
of the shoulder joint. Journal of Bone and Joint Surgery—
American Volume 1944;26(1):1–30.
Johnson MP, McClure PW, Karduna AR. New method to assess
scapular upward rotation in subjects with shoulder pathology.
Journal of Orthopaedic and Sports Physical Therapy
2001;31(2):81–9.
Kamkar A, Irrgang JI, Whitney SL. Nonoperative management of
secondary shoulder impingement syndrome. Journal of Orthopae-
dic and Sports Physical Therapy 1993;17(5):212–24.
Kapandji IA. The physiology of joints. Annotated diagrams of the
mechanics of the human joints, vol 1. Upper Limb, 5th ed.
Churchill Livingstone, Edinburgh, 1982.
Karduna AR, McClure PW, Michener LA, Sennett B. Dynamic
measurements of three-dimensional scapular kinematics: a valida-
tion study. Journal of Biomechanical Engineering 2001;123(4):
184–90.
Kebaetse M, McClure P, Pratt NA. Thoracic position effect on
shoulder range of motion, strength, and three-dimensional scapular
kinematics. Archives of Physical Medicine and Rehabilitation
1999;80(8):945–50.
Kibler WB. The role of the scapula in athletic shoulder function.
American Journal of Sports Medicine 1998;26(2):325–37.
Kibler WB, Uhl TL, Maddux JWQ, Brooks PV, Zeller B, McMullen J.
Qualitative clinical evaluation of scapular dysfunction: a reliability
study. Journal of Shoulder and Elbow Surgery 2002;11(6):
550–6.
Landis JR, Koch GG. The measurement of observer agreement for
categorical data. Biometrics 1977;33(1):159–74.
Lowe BD. Accuracy and validity of observational estimates of
shoulder and elbow posture. Applied Ergonomics 2004;35(2):
159–71.
Ludewig PM, Cook TM. The effect of head position on scapular
orientation and muscle activity during shoulder elevation. Journal
of Occupational Rehabilitation 1996;6(3):147–58.
Ludewig PM, Cook TM. Alterations in shoulder kinematics and
associated muscle activity in people with symptoms of shoulder
impingement. Physical Therapy 2000;80(3):276–91.
Ludewig PM, Cook TM, Nawoczenski DA. Three-dimensional
scapular orientation and muscle activity at selected positions of
humeral elevation. Journal of Orthopaedic and Sports Physical
Therapy 1996;24(2):57–65.
Lukasiewicz AC, McClure P, Michener L, Pratt N, Sennett B.
Comparison of 3-dimensional scapular position and orientation
between subjects with and without shoulder impingement. Journal
of Orthopaedic and Sports Physical Therapy 1999;29(10):574–86.
Magarey ME, Jones MA. Dynamic evaluation and early management
of altered motor control around the shoulder complex. Manual
Therapy 2003;8(4):195–206.
Mandalidis DG, McGlone BS, Quigley RF, McInerney D, O’Brien M.
Digital fluoroscopic assessment of the scapulohumeral rhythm.
Surgical and Radiologic Anatomy 1999;21(4):241–6.
Mottram SL. Dynamic stability of the scapula. Manual Therapy
1997;2(3):123–31.
Pascoal AG, van der Helm FCT, Correia PP, Carita I. Effects of
different arm external loads on the scapulo-humeral rhythm.
Clinical Biomechanics 2000;15(Supplement 1):S21–4.
Sahrmann SA. Diagnosis and treatment of movement impairment
syndromes. St. Louis, Missouri: Mosby; 2002. p. 193–9, Chapter 5.
SAS Institute. SAS/STAT V8 software user’s guide, Cary. North
Carolina: SAS Institute; 1999.
SAS Technical Support, Data Analysis Sample Programs. /http://
ftp.sas.com/techsup/download/stat/S Retrieved 10 June, 2004.
Sobush DC, Simoneau GG, Dietz KE, Levene J, Grossman RE, Smith
WB. The Lennie test for measuring scapular position in healthy
young adult females: a reliability and validity study. Journal of
Orthopaedic and Sports Physical Therapy 1996;23(1):39–50.
Sugamoto K, Harada T, Machida A, Inui H, Miyamoto T, Takeuchi
E, et al. Scapulohumeral rhythm: relationship between motion
velocity and rhythm. Clinical Orthopaedics and Related Research
2002;401:119–24.
Talkhani IS, Kelly CP. Scapulo thoracic rhythm in normal male
volunteers. Biomedical Sciences Instrumentation 1997;34:327–31.
Talkhani IS, Kelly CP. Movement analysis of asymptomatic normal
shoulders: a preliminary study. Journal of Shoulder and Elbow
Surgery 2001;10(6):580–4.
van der Heijden GJ. Shoulder disorders: state of the art review.
Baillieres Clinical Rheumatology 1999;13(2):287–309.
Warner JJP, Micheli LJ, Arslanian LE, Kennedy J, Kennedy R.
Scapulothoracic motion in normal shoulders and shoulders with
glenohumeral instability and impingement syndrome: a study using
Moire topography. Clinical Orthopaedics and Related Research
1992;285:191–9.
Yanai T, Hay JG. Shoulder impingement in front-crawl swimming: II.
Analysis of stroking technique. Medicine and Science in Sports and
Exercise 2000;32(1):30–40.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 271–279

www.elsevier.com/locate/math

Original article

The influence of specific training on trunk muscle recruitment

patterns in healthy subjects during stabilization exercises
Veerle K. Stevens, Pascal L. Coorevits, Katie G. Bouche, Nele N. Mahieu,
Guy G. Vanderstraeten, Lieven A. Danneels
Department of Rehabilitation Sciences and Physiotherapy, Faculty of Medicine and Health Sciences, Ghent University, De Pintelaan 185, 6K3,
B9000 Ghent, Belgium
Received 14 February 2005; received in revised form 12 May 2006; accepted 10 July 2006

Abstract

Low back pain is a major problem involving high medical costs, therefore effective prevention strategies are essential.
Stabilization exercises seem to facilitate the neuromuscular control of the lumbar spine and may be useful in prevention programs.
To investigate whether specific lumbar stabilization training has an effect on muscle recruitment patterns in a healthy population, in
the present study 30 subjects were recruited to perform two types of testing exercises, i.e. bridging exercises and exercises in four-
point kneeling, both before and after training. Surface electromyographic data of different abdominal and back muscles were
obtained. After training, analysis of the relative muscle activity levels (percentage of maximal voluntary isometric contraction)
showed a higher activity of the local (segmental-stabilizing) abdominal muscles, but not of the local back muscles; minimal changes
in global (torque-producing) muscle activity also occurred. Analysis of the local/global relative muscle activity ratios revealed higher
ratios during all exercises after training, although not all differences were significant. These results indicate that muscle recruitment
patterns can be changed in healthy subjects by means of a training program that focuses on neuromuscular control. Additional
studies are needed to evaluate this type of training as a prevention strategy.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Stabilization exercise; Prevention; Lumbar stabilization training; Surface electromyography

1. Introduction treatment periods (O’Sullivan et al., 1997;Rasmussen-Barr

In 1994 52% of Belgian hospital nurses reported
musculoskeletal job-related problems that lasted longer
than 1 day; low back pain (LBP) was the major cause
(53.3%) (Clarijs et al., 1998). Effective primary and
secondary prevention strategies are needed to address
this problem because the costs for, e.g. health insurance,
employers and society, as well as the reduced quality-of-
life of the patients, are substantial.
Short- and long-term results indicate that specific
lumbar stabilizing therapy can decrease the number of
recurrent pain episodes (Hides et al., 2001) and recurrent
Corresponding author. Tel.: 32 9 2402996; fax: 32 9 2403811.
E-mail address: Veerle.Stevens@UGent.be (V.K. Stevens).
et al., 2003). Specific lumbar-stabilizing therapy involves
changing muscle recruitment patterns. Symptomatic
chronic LBP patients with clinical evidence of spondylo-
lysis or spondylolisthesis were shown to be able to activate
the deep abdominal muscles without significant co-
activation of the rectus abdominis muscle (RA) when
performing an abdominal drawing-in manoeuvre after a
10-week intervention (O’Sullivan et al., 1998). Lumbar
stabilization training that paid no specific attention to the
local muscles showed no changes in relative electromyo-
graphic (EMG) amplitudes during more complex stabili-
zation exercises after 12 weeks of training (Arokoski et al.,
2004). Apart from these studies on patients, the effect of
stabilization training has not yet been investigated in a
healthy population in relation to primary prevention.

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.07.009
 ARTICLE IN PRESS
272 V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279
Local muscles of the trunk, such as the lumbar
multifidus (MF), with their vertebrae to vertebrae
attachments (as described by Macintosh and Bogduk,
1987), are supposed to control the fine-tuning of the
positions of adjacent vertebrae (segmental stabilization)
(Bergmark, 1989; Richardson et al., 1999; Hodges and
Moseley, 2003). Because of their connection through the
thoracolumbar fascia, the transversus abdominis (TA)
and the inferior fibres of the internal oblique (IO) also
have direct attachment to the lumbar vertebrae and can
therefore also be considered as local muscles (Hodges,
1999).
Unlike the local muscles, the global muscles are
supposed to be important for torque production and
general trunk stability, because they are not directly
attached to the spine (Bergmark, 1989). The global
muscle system includes the RA, the external oblique
(EO), the gluteus maximus, the latissimus dorsi and the
thoracic part of the iliocostalis lumborum muscles
(ICLT) (Richardson et al., 1999).
To train functional stability, once an optimal local
activation has been achieved, the interplay between local
and global muscles is thought to be necessary (Hodges
and Moseley, 2003; Richardson et al., 2004). Because
biomechanical and muscle research has shown no clear
distinction between the contribution of the local and
global muscles to spine stability, this functional classi-
fication based on anatomic findings needs to be
considered with some caution (Arokoski et al., 2001;
Cholewicki and Van Vliet, 2002; Kavcic et al., 2004a;
Stevens et al., 2006). Not only is the local muscle system
important, but also the controlled co-operation between
the two muscle systems can provide a stable structure.
Consequently, the ratio of the local muscle activity to
the global muscle activity needs to be further elucidated.
According to Edgerton et al. (1996) EMG ratios can be
a sensitive discriminator of altered recruitment patterns
and muscle dysfunction. In order to highlight differences
in the synergistic activity of local versus global muscles
(e.g. IO versus RA; lumbar versus thoracic erector
spinae muscles), ratios of muscle activity levels during
various stabilization exercises have been investigated in
healthy subjects (Marshall and Murphy, 2005) and in
LBP patients (O’Sullivan et al., 1998; van Dieën et al.,
2003).
The purpose of the current study was to evaluate the
benefit derived from specific stabilization training for a
prevention program by investigating whether this
training had an effect on muscle recruitment patterns
in healthy subjects. The training was an isolated local
muscle contraction (first phase) followed by integrating
local co-contraction in different movements starting
from various positions. The exercises used for the
evaluation were bridging exercises and exercises in
four-point kneeling, both of which are often used in
clinical practice to train lumbar stability. The specific
attention paid to the local muscles during the interven-
tion aimed to increase local muscle activity and
consequently change the local/global ratio.
2. Methods
2.1. Study design
This was a cross-sectional study. The baseline EMG
test session was followed by a 3-month intervention
period and then a second EMG test session.
2.2. Subjects
Thirty healthy subjects (15 men and women) volunta-
rily participated in this study. Their mean age was 19.6
(range 19–23) years, mean height was 176.5 (range 157–
194) cm and mean weight was 66.9 (range 42–84) kg. All
subjects gave written informed consent. The study was
approved by the Ghent University Ethics Committee.
2.3. Procedures
2.3.1. Intervention period
The two EMG test sessions were separated by an
intervention period of 3 months. The subjects were
instructed in accordance with the principles often used
in training lumbar stability (Richardson and Jull, 1995;
Richardson et al., 1999; O’Sullivan, 2000). Instruction in
the basic anatomy of the TA, MF and other abdominal
and back muscles was aimed at emphasizing the
difference between the local and global trunk muscles
and to help avoid ‘substitution’ strategies of the global
muscles. In the first phase of the training, local muscle
activity was facilitated without substitution strategies of
the global muscles and with focus on the continuation of
normal breathing during the exercises. Subsequently, the
time for holding the position and the number of
repetitions were increased, and different postures
(supine, four-point kneeling, prone, sitting and stand-
ing) were added (Richardson et al., 1999). Once an
accurate and sustained contraction of the local muscles
was achieved in different postures (10 contractions with
10-s holds), the exercises progressed to the second phase
which involved applying low load to the muscles
through controlled movements of the upper and lower
extremities (Richardson and Jull, 1995). The aim during
the third phase was to integrate the motor skill into
normal static tasks and dynamic functional tasks
(Richardson et al., 1999). During the 3-month interven-
tion period, eight guided training sessions took place,
each lasting 30 min; the subjects were also asked to
perform the exercises for about 15 min each day at home
as part of the intervention. During the intervention
 ARTICLE IN PRESS
V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279 273

period, no specific attention was paid to the exercises
performed in the test sessions.
2.3.2. Test sessions
Before and after the specific stabilization training, the
subjects performed two types of testing exercises:
bridging exercises and exercises in four-point kneeling.
For both, the only instruction given during the testing
was to maintain the lumbar neutral spine position. At
the moment of the first test session, the subjects had no
knowledge or experience of stabilization principles.
2.4. Equipment and measurements
2.4.1. Electromyography
The skin was prepared by shaving excess hair and
rubbing the skin with alcohol to reduce impedance
(typically p10 kO). Disposable Ag/AgCl surface elec-
trodes (Blue Sensor, Medicotest A/S, Ølstykke, Den-
mark) were attached parallel to the muscle-fibre
orientation, as described previously (Danneels et al.,
2001b, 2002), bilaterally over the following so-called
local trunk muscles: the inferior fibres of the IO (midway
between the anterior iliac spine and symphysis pubis,
above the inguinal ligament) and the lumbar MF (lateral
to the midline of the body, above and below a line
connecting both posterior superior iliac spines).
Although the focus of stabilization training was on the
TA, it was expected that this would be reflected in the
surface EMG of the inferior fibres of the IO. Marshall
and Murphy (2003) showed that medially and inferiorly
to the anterior superior iliac spine, the fibres of the TA
and IO are blended, so that a distinction between the
muscle signals cannot be made at this location; also, at
this site the direction of the fascicles of both muscles is
similar (inferomedial) (Urquhart et al., 2005). More-
over, both the TA and the inferior fibres of the IO play a
similar role in compressing the sacroiliac joint and
consequently increasing the control of that region
(Richardson et al., 2002).
The selected so-called global trunk muscles were the
EO (15 cm lateral to the umbilicus), the RA (3 cm lateral
to the umbilicus), and the ICLT (above and below the
L1 level, midway between the midline and the lateral
aspect of the body). The maximal inter-electrode spacing
between the recording electrodes was 2.5 cm as recom-
mended by Ng et al. (1998), and each electrode had a
pick-up area of approximately 1.0 cm2. To reduce the
variability due to the electrode position, a personalized
template ensured the exact reapplication of the electro-
des (Danneels et al., 2001a).
2.4.2. Exercises during the test procedure
Maximal voluntary isometric contractions (MVIC) of
the muscles were measured before the experimental
tasks. These exercises were performed to provide a basis
for EMG signal amplitude normalization. Three differ-
ent isometric exercises against manual resistance were
performed according to the description of Danneels
et al. (2001b), and each exercise was registered three
times during 3 s.
After the MVICs, the subjects were asked to start the
experimental exercises. Six exercises (often used in
clinical practice to train lumbar stability) were per-
formed. The first group of exercises was executed in
supine position, knees bent (601 flexion) and feet on the
floor. A bridging exercise, either simple or accompanied
by leg movements, was performed (exercises 1 to 3 in
Table 1). The second group of exercises was performed
in four-point kneeling (exercises 4 to 6 in Table 1). At
the start of each exercise, the examiner determined the
subject’s lumbar neutral spine position and the subjects
were asked to hold this position throughout the exercise.
In four-point kneeling, the neutral spine position was set
about halfway between full extension and a flat spine
(Danneels et al., 2002); in supine position the anterior
and posterior superior iliac spines were in line
(Richardson et al., 2004). The exercises were performed
in a random sequence. In order to standardize the
position of the subject and the equipment, markers were
placed on the floor. The dynamic phases (i.e. lifting and
lowering of the pelvis and the extremities) lasted 2 s. The
mid-phase (i.e. extended leg/arm and lifted pelvis) was
held for 5 s. The rhythm of 60 beats/min was set by a
metronome. For each exercise three trials were per-
formed. To prevent muscular fatigue, an interval of at

Table 1
Exercises

Group 1 Bridging in supine position

Exercise 1 Bridging in supine position
Exercise 2 Ball bridge
Exercise 3 Unilateral bridging: bridging with extension of the left/right leg
Group 2 Four-point kneeling
Exercise 4 Single-leg lift, performed by extending the left/right leg out to the horizontal and returning it to the starting position
Exercise 5 The leg extension of exercise 4 coupled with the simultaneous raising of the contralateral arm to the horizontal before returning
the extended leg and arm to the original position
Exercise 6 This exercise is basically the same as exercise 5, but with the addition of moving the trunk/pelvis in a backward direction
(i.e. away from the hands), which increases the angle of hip flexion of the loaded leg by 30 1
 ARTICLE IN PRESS
274 V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279
least 15 s was allowed between the exercises; during these
periods the exercises were explained.
Before the second test session, an ultrasound (US)
evaluation was carried out to assess whether the subjects
could produce an isolated contraction of the TA. In a
clinical setting the tonic contraction of the MF is easy to
palpate, but the difference between contraction of the
IO and the TA is not always easy to detect. Therefore,
the subjects were placed in a supine position lying
crooked, and were then asked to draw in their lower
abdomen slowly and gently, without moving the spine.
The transducer was placed on the anterolateral aspect of
the abdominal wall, at the level of the umbilicus. A
Siemens Sonoline SL-1 ultrasound imaging device was
used with a linear array probe with a wave frequency of
7.5 MHz. The criterion was a slow and controlled
tensioning of the anterior fascial attachment of the TA
in a lateral direction. The TA was slightly thickened and
the IO and EO remained relatively inactive (Richardson
et al., 2004). Because the aim was to evaluate the effect
of training local muscle co-contraction on the perfor-
mance of stabilization exercises, this additional US
evaluation was useful to understand the reasons for
changes in EMG activity. This US study revealed that 5
subjects were not able to contract the TA in isolation
from the global muscle system and were thus excluded.
2.5. Data analysis
The raw surface EMG signals were measured at a
bandwidth of 10–500 Hz, using a differential amplifier
(MyoSystem 1400, Noraxon Inc, Scottsdale, USA). The
overall gain was 1000 and the common mode rate
rejection ratio was 115 dB. The signals were analogue/
digitally (A/D) (12-bit resolution) converted at 1000 Hz
and stored in a personal computer. The stored data were
full-wave rectified and smoothed with a root mean
square (RMS) with a window of 150 millis. For each of
the muscles and for each testing session, the RMS was
calculated for the 3 repetitions of the different exercises.
The mean RMS of the three MVIC trials for every
muscle was used to provide a basis for EMG signal
amplitude normalization of the data of the exercises.
The static phases of the exercises were analysed using an
interval of 4700 ms after the defined starting point of the
holding position. Noraxon MyoResearch software 2.10
was used for these analyses.
The effect on muscle recruitment patterns was
investigated in two ways. First, the changes in muscle
activity of each muscle as a result of the training were
investigated. Second, the difference in the ratio of local
muscle activity to global muscle activity (separately for
abdominal and back muscles) before and after training
was evaluated. This assessment was based on two ratios
of the abdominal muscles (the IO/RA and the IO/EO)
and one of the back muscles (the MF/ICLT).
2.6. Statistical analysis
Statistical analysis was performed using the SPSS 11.0
software package (SPSS Inc., Chicago, IL) for Win-
dows. Given the symmetry of the task during the single-
bridging and the ball-bridging exercise (PX0.1), the
EMG values of the muscles of the left and the right side
were averaged. Because there were no significant
differences (P40.05) between the ipsilateral and con-
tralateral muscle activity values during the asymmetric
bridging exercises, they were also averaged. Ipsilateral
referred to the side of the extended leg and contralateral
to the other side.
A multivariate analysis of variance (MANOVA) was
used to evaluate the changes in muscle activity as a
result of specific stabilization training (factors muscle
[5 muscles], time [before and after training], and side
[only in the asymmetric exercises]). In the event of
several significant interactions, least significance differ-
ence tests, adjusted by a (although conservative)
Bonferroni test to protect against type I errors, were
used to analyse the significant differences between the
individual muscles in each exercise. Consequently, the
level for statistical significance was set at a ¼ 0:002 for
exercises 1 to 4 (two-factor interaction) and at a ¼
0:0008 for exercises 5 and 6 (three-factor interaction).
To analyse the difference in the ratio of local muscle
activity to global muscle activity before and after
training, paired sample t-tests were used and a was set
at 0.05.
3. Results
3.1. Changes in relative muscle activity
Figs. 1–3 present the relative pre- and post-training
EMG levels (% MVIC) of the different muscles and
their respective P-values.
3.1.1. Local muscle activity
After training, the local IO showed a significantly
higher relative muscle activity (Pp0.001) during all the
bridging exercises. In contrast to the bilateral changes of
the local IO during the bridging exercises, after training
only the relative muscle activity of the ipsilateral IO
increased significantly during the asymmetric four-
point kneeling exercise 5 (Pp0.001) and exercise 6
(P ¼ 0:001).
3.1.2. Global muscle activity
After training, during the symmetric exercises 1 and 2
the relative muscle activity of the global RA was also
significantly higher (Pp0.001).
For the global EO and the local and global back
muscles, no significant differences between the EMG
 ARTICLE IN PRESS
V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279 275

70

Relative muscle activity (% MVIC)

60

50
< 0.001

40 < 0.001 0.19

0.81
0.63
Ex1 - Pre
30 0.01
Ex1 - Post
Ex2 - Pre 0.93
20 Ex2 - Post
0.26
10 0.001
0.001

0
IO MF RA EO ICLT
Muscle

Fig. 1. Mean values, SD and P-values (between pre- and post-training for each exercise) of the relative EMG activity during the symmetric exercises
(a ¼ 0:002). Ex, exercise; Pre, before training; Post, after training; MVIC, maximal voluntary isometric contraction; IO, internal oblique; MF,
lumbar multifidus; RA, rectus abdominis; EO, external oblique; ICLT, iliocostalis lumborum pars thoracis.

70
< 0.001
60
Relative muscle activity (%MVIC)

0.39

50 Ex3 - Pre
Ex3 - Post
Ex4 - Pre 0.29
40 0.11 0.38 Ex4 - Post
0.91 0.02
0.38 Ex5 - Pre 0.007
0.11
0.007 < 0.001 Ex5 - Post 0.003
30 Ex6 - Pre 0.93
0.001 Ex6 - Post
0.06
20
0.005
10 0.020.06 0.03

0
IO MF RA EO ICLT
Muscle

Fig. 2. Mean values, SD and P-values (between pre- and post-training for each exercise) of the ipsilateral relative EMG activity during the
asymmetric exercises (a ¼ 0:002 in exercises 3 and 4; a ¼ 0:0008 in exercises 5 and 6). Ex, exercise; Pre, before training; Post, after training; MVIC,
maximal voluntary isometric contraction; IO, internal oblique; MF, lumbar multifidus; RA, rectus abdominis; EO, external oblique; ICLT,
iliocostalis lumborum pars thoracis.

levels before and after training were found for any of the not in all exercises the increase in the IO/RA and the
exercises. MF/ICLT ratios was significant. All ratios were signi-
ficantly higher (Pp0.01) in the single bridging exercise
3.2. Changes in ratios (exercise 1). In the ball bridge exercise (exercise 2) a
significant difference (Pp0.001) was found for the
Because co-operation between the local and global abdominal muscles (IO/RA ratio and IO/EO ratio),
muscle systems is particularly important in creating a but not for the back muscles (MF/ICLT ratio). In the
stable structure, changes in the local/global muscle unilateral bridging exercise (exercise 3), the difference
activity ratio after stabilization training were also between the ratios before and after training was
evaluated. To detect changes in this ratio, paired sample significant (Pp0.02), except for the MF/ICLT ratio
t-tests were used; the level for statistical significance was and the ipsilateral IO/RA ratio. In the exercises in four-
set at a ¼ 0:05. point kneeling (exercises 4 – 6), the difference in the
After training, the ratio of the local to global muscle ratio local to global muscles before and after training
activity was higher in all exercises (Table 2). However, was significant (Pp0.05) only for the ipsilateral muscles.
 ARTICLE IN PRESS
276 V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279

70

0.05
60 Ex3 - Pre

Relative muscle activity (% MVIC)

Ex3 - Post
0.007 Ex4 - Pre
50 Ex4 - Post 0.43
Ex5 - Pre
Ex5 - Post
Ex6 - Pre 0.29 0.44
40 0.45 Ex6 - Post 0.09
0.52
0.28 0.05
30 0.001 0.07
0.01 0.06
0.38
0.41
20

0.02
10 0.01 0.10
0.03

0
IO MF RA EO ICLT
Muscle

Fig. 3. Mean values, SD and P-values (between pre- and post-training for each exercise) of the contralateral relative EMG activity during the
asymmetric exercises (a ¼ 0:002 in exercises 3 and 4; a ¼ 0:0008 in exercises 5 and 6). Ex, exercise; Pre, before training; Post, after training; MVIC,
maximal voluntary isometric contraction; IO, internal oblique; MF, lumbar multifidus; RA, rectus abdominis; EO, external oblique; ICLT,
iliocostalis lumborum pars thoracis.

However, the ratio of the contralateral oblique abdom- the superficial fibres of the MF during the intervention
inal muscles also increased significantly (P ¼ 0:04) in period. Moseley et al. (2002, 2003, 2004) demonstrated a
exercise 5. different activation of the deep and superficial fibres of
the MF anticipating different loading conditions in
standing. Since it has been shown that recording the
4. Discussion muscle activity of the deep fibres of the MF by surface
electrodes may be difficult (Stokes et al., 2003), this
4.1. Changes in relative muscle activity technique may not have been sufficiently accurate to
detect any changes in activity of the deep MF fibres.
Stabilization training involves isolated local muscle Whatever the reason, the training strategy used in the
contraction and an integration of the local and global present study was unable to influence the muscle
muscle systems during particular movement patterns activation patterns of the local back muscle.
(O’Sullivan, 2000). It was thought that such training
with specific attention paid to the TA and MF 4.1.2. Global muscle activity
(Richardson et al., 1999; O’Sullivan, 2000) would Not only did the local (so-called segmental-stabilizing
significantly increase the relative muscle activity of the abdominal) muscle activity levels change, but also the
local muscles in healthy subjects. relative activity of some global (so-called torque-
producing and general-stabilizing) muscles changed.
4.1.1. Local muscle activity After training, the activity of the RA was significantly
The results of the present study indicate that, since the higher during the symmetric bridging exercises.
relative local IO abdominal muscle activity was in- Co-contraction with other abdominal muscles is often
creased on both sides during bridging exercises and reported when subjects are trying to contract the TA
ipsilaterally during four-point kneeling exercises, ab- (Richardson et al., 1999). Beith et al. (2001) concluded
dominal muscle activity can be changed after a lumbar that while performing an abdominal hollowing man-
stabilization training program in healthy subjects. oeuvre, because elimination of activity in the EO
In contrast, no significant change in relative muscle muscles may be too difficult or even impossible for
activity of the local back muscle MF was found after some to achieve, it may not always be a feasible goal.
training. One reason for this finding is that it may be Studies on the effect of an isometric contraction of all
more difficult to produce an isolated contraction of the the abdominal wall muscles (known as an abdominal
MF during training. This idea is supported by the brace manoeuvre) showed a considerably higher relative
clinical experience that, in general, subjects find it easier muscle activity of the RA in exercises 1, 3, 4 and 5
to concentrate on drawing in the lower abdomen than (Kavcic et al., 2004b). However, the results of the local
on focusing on the lower back muscle contraction. Also, and global relative muscle activity changes in the present
perhaps it was not possible to train both the deep and study are limited to only those subjects who have been
 ARTICLE IN PRESS
V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279 277

Table 2
Mean values, standard deviations (SD) and P-values of the ratio local muscle activity/global muscle activity (mean) during the different exercises

Exercise 1 2

Pre Post P Pre Post P

Mean SD Mean SD Mean SD Mean SD

IO/RA 3.31 1.76 7.62 6.23 0.001* 3.19 2.19 7.38 5.93 0.001*
IO/EO 1.19 0.91 2.93 2.18 o0.001* 0.80 0.54 2.34 1.85 o0.001*
MF/ICLT 1.21 0.50 1.81 1.12 0.01* 1.27 0.56 1.59 0.97 0.09

Exercise 3 4

Pre Post P Pre Post P

Mean SD Mean SD Mean SD Mean SD

Ipsilateral
IO/RA 8.26 3.40 9.90 6.28 0.20 4.04 3.03 5.88 4.65 0.04*
IO/EO 2.70 1.96 3.35 1.95 0.02* 0.43 0.32 0.95 0.61 o0.001*
MF/ICLT 1.00 0.36 1.48 0.78 0.15 2.09 1.36 3.21 2.25 0.02*
Contralateral
IO/RA 3.34 1.86 4.37 2.91 0.01* 10.02 5.07 10.77 9.66 0.72
IO/EO 0.77 0.38 1.53 1.22 0.01* 1.83 1.50 2.43 1.72 0.11
MF/ICLT 1.48 0.79 2.48 1.92 0.31 0.86 0.39 1.25 1.99 0.32

Exercise 5 6

Pre Post P Pre Post P

Mean SD Mean SD Mean SD Mean SD

Ipsilateral
IO/RA 3.42 3.35 5.27 3.49 0.02* 3.36 2.84 4.99 3.83 0.05*
IO/EO 0.30 0.22 0.67 0.50 o0.001* 0.40 0.25 0.88 0.68 0.001*
MF/ICLT 1.63 1.05 2.58 1.50 0.004* 1.11 0.37 1.64 1.05 0.01*
Contralateral
IO/RA 10.13 6.47 11.89 9.77 0.44 8.08 5.65 8.23 6.36 0.92
IO/EO 1.65 1.33 2.43 1.75 0.04* 1.57 1.09 2.17 1.70 0.06
MF/ICLT 0.89 0.65 0.92 1.38 0.91 1.01 0.28 1.27 1.21 0.31

IO, internal oblique; MF, lumbar multifidus; RA, rectus abdominis; EO, external oblique; ICLT, iliocostalis lumborum pars thoracis.
*P-value significant at a ¼ 0.05 level.

shown able to isolate TA contraction (investigated using 4.2. Changes in ratios

real-time ultrasound).
The absence of co-contraction of the more global
abdominal muscles during the exercises in four-point
kneeling compared with the bridging exercises, might be
explained by the difference in posture and level of
difficulty between the two groups of exercises. The four-
point kneeling position provides increased awareness of
the abdominal wall due to the gravitational stretch on
the muscles, and allows complete relaxation of the
abdominal wall. This position may increase the sensi-
tivity of the stretch receptors and might enhance the
stimulus to contract abdominal muscles separately
(Richardson and Jull, 1995; Richardson et al.,
1999;Beith et al., 2001). This stretch of the abdominal
wall does not exist in the supine position, which might
make it harder to recruit the deep abdominals separately
in this position.
In general, analysis of the relative muscle activity
levels showed significant changes only in the local IO
after stabilization training. However, analysis of the
local to global muscle activity ratios revealed that all
ratios increased after training. This shows that muscle
activity patterns can be changed in a healthy population
if stabilization training is performed with specific
attention paid to the so-called local muscles (Richardson
et al., 1999;O’Sullivan, 2000).
The ratio between the local and global muscle activity
increased due to a greater increase in local muscle
activity compared with global muscle activity. This
increase in the local/global ratio was significant in most
of the bridging exercises. In the four-point kneeling
exercises, the local muscles at the side of the extended leg
seemed to be activated to higher intensities than the
 ARTICLE IN PRESS
278 V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279
global muscles. The increase in activity was also
apparent at the contralateral side, but the difference
between the ratios was not significant.
Similar stabilization training in symptomatic chronic
LBP patients with clinical evidence of spondylolysis or
spondylolisthesis also resulted in a significant increase in
the IO/RA ratio during an abdominal drawing in
manoeuvre (O’Sullivan et al., 1998). Simulations based
on MVIC contractions in several directions predicted
that an increase of the IO/RA ratio would be effective in
increasing spinal stability (Van Dieën et al., 2003).
In the present study, the relative muscle activity ratios
increased during all exercises independent of the type of
exercise or the surface on which the exercise was
performed. Marshall and Murphy (2005) also demon-
strated no change in the IO/RA ratio between stabiliza-
tion exercises performed on and off a swiss ball.
A limitation of the present study is that only static
phases of stabilization exercises were evaluated. How-
ever, during the intervention period, the exercises were
progressed to non-neutral positions (Akuthota and
Nadler, 2004) and to dynamic functional movements
with upper and lower extremity movements (Richardson
et al., 2004). In addition to the evaluation of the static
phases, further studies could also investigate the more
advanced dynamic movements. The results of our
relatively young group of participants may not be
representative for the whole population. However, the
present study was primarily designed to evaluate the
effects of a basic stabilization package that could be
used in a prevention programme.
5. Conclusion
In the present study, healthy subjects learned isolated
local muscle contractions (controlled by ultrasound) and
their integration into basic static stabilization exercises.
After training, analysis of the relative muscle activity
levels showed a higher activity of the local abdominal
muscles, but not of the local back muscles; minimal
changes in global muscle activity also occurred. Analysis
of the local/global relative muscle activity ratios
revealed higher ratios during all exercises after training,
although not all differences were significant. This
indicates that muscle recruitment patterns can be
changed in healthy subjects after a training programme
that focuses on neuromuscular control, which could be
useful in prevention programmes. More studies are
needed to substantiate these results.
Acknowledgements
The authors thank Ms. Evelien De Burck and Ms.
Wendy Van Loo for their assistance with collection of
the data, Prof. Georges van Maele for statistical advice,
and Ms. Iris Wojtowicz for linguistic corrections.
References
Akuthota V, Nadler SF. Core strengthening. Archives of Physical
Medicine and Rehabilitation 2004;85:S86–92.
Arokoski JP, Valta T, Airaksinen O, Kankaanpää M. Back and
abdominal muscle function during stabilization exercises. Archives
of Physical Medicine and Rehabilitation 2001;82:1089–98.
Arokoski JP, Valta T, Kankaanpää M, Airaksinen O. Activation of
lumbar paraspinal and abdominal muscles during therapeutic
exercises in chronic low back pain patients. Archives of Physical
Medicine and Rehabilitation 2004;85:823–32.
Beith ID, Synnott E, Newman A. Abdominal muscle activity during
the abdominal hollowing manoeuvre in the four-point kneeling and
prone positions. Manual Therapy 2001;6(2):82–7.
Bergmark A. Stability of the lumbar spine. A study in mechanical
engineering. Acta Orthopaedica Scandinavica 1989;230(Suppl.):20–4.
Cholewicki J, Van Vliet JJ. Relative contribution of trunk muscles to
the stability of the lumbar spine during isometric exertions. Clinical
Biomechanics 2002;17:99–105.
Clarijs JP, Zinzen E, Van Roy P, Duquet W, Caboor D, Verlinden M,
et al. Multi- en interdisciplinaire evaluatie van cervicale en lumbale
wervelkolomproblematiek bij ziekenhuis verpleegkundigen, met
ontwikkeling en toepassing van een primair preventieprogramma.
Eindverslag ST/03/029 Wetenschappelijk ondersteuningsprogram-
ma voor de gezondheidsbescherming van de werknemer (1994–
1998); 1998.
Danneels LA, Cagnie BJ, Vanderstraeten GG, Cambier DC,
Witvrouw EE, De Cuyper HJ. Intra-operator and inter-operator
reliability of surface electromyography in the clinical evaluation of
back muscles. Manual Therapy 2001a;6(3):145–53.
Danneels LA, Vanderstraeten GG, Cambier DC, Witvrouw EE,
Stevens VK, De Cuyper HJ. A functional subdivision of hip,
abdominal, and back muscles during asymmetric lifting. Spine
2001b;26(6):E114–21.
Danneels LA, Coorevits PL, Cools AM, Vanderstraeten GG, Cambier
DC, Witvrouw EE, et al. Differences in electromyographic activity
in multifidus muscle and the iliocostalis lumborum between healthy
subjects and patients with subacute and chronic low back pain.
European Spine Journal 2002;11:13–9.
Edgerton V, Wolf S, Levendowski D, Roy R. Theoretical basis for
patterning EMG amplitudes to assess muscle dysfunction. Medical
Science Sports Exercise 1996;28:744–51.
Hides JA, Jull GA, Richardson CA. Long-term effects of specific
stabilizing exercises for first-episode low back pain. Spine
2001;26(11):E242–8.
Hodges PW. Is there a role for transversus abdominis in lumbo-pelvic
stability? Manual Therapy 1999;4(2):74–86.
Hodges PW, Moseley GL. Pain and motor control of the lumbopelvic
region: effect and possible mechanisms. Journal of Electromyo-
graphy and Kinesiology 2003;13:361–70.
Kavcic N, Grenier S, McGill SM. Determining the stabilizing role of
individual torso muscles during rehabilitation exercises. Spine
2004a;29(11):1254–65.
Kavcic N, Grenier S, McGill SM. Quantifying tissue loads and spine
stability while performing commonly prescribed low back stabiliza-
tion exercises. Spine 2004b;29(20):2319–29.
Macintosh JE, Bogduk N. Volvo award in basic science. The
morphology of the lumbar erector spinae. Spine 1987;12(7):658–68.
Marshall PW, Murphy BA. The validity and reliability of surface
EMG to assess the neuromuscular response of the abdominal
muscles to rapid limb movement. Journal of Electromyography
and Kinesiology 2003;13:477–89.
 ARTICLE IN PRESS
V.K. Stevens et al. / Manual Therapy 12 (2007) 271–279
Marshall PW, Murphy BA. Core stability exercises on and off a Swiss
ball. Archives of Physical Medicine and Rehabilitation
2005;86:242–9.
Moseley GL, Hodges PW, Gandevia SC. Deep and superficial fibers of
the lumbar multifidus muscle are differentially active during
voluntary arm movements. Spine 2002;27:E29–36.
Moseley GL, Hodges PW, Gandevia SC. External perturbation of the
trunk in standing humans differentially activates components of
the medial back muscles. Journal of Physiology 2003;547:581–7.
Moseley GL, Nicholas MK, Hodges PW. Does anticipation of back
pain predispose to back trouble? Brain 2004;127:2339–47.
Ng JK, Kippers V, Richardson CA. Muscle fibre orientation of
abdominal muscles and suggested surface EMG electrode posi-
tions. Electromyography and Clinical Neurophysiology
1998;38(1):51–8.
O’Sullivan PB. Lumbar segmental ‘instability’: clinical presentation
and specific stabilizing exercise management. Manual Therapy
2000;5(1):2–12.
O’Sullivan PB, Twomey L, Allison GT. Evaluation of specific
stabilizing exercise in the treatment of chronic low back pain with
radiologic diagnosis of spondylolysis or spondylolisthesis. Spine
1997;22(24):2959–67.
O’Sullivan PB, Twomey L, Allison GT. Altered abdominal muscle
recruitment in patients with chronic back pain following a specific
exercise intervention. JOSPT 1998;27(2):114–24.
Rasmussen-Barr E, Nilsson-Wikmar L, Arvidsson I. Stabilizing
training compared with manual treatment in subacute and chronic
low-back pain. Manual Therapy 2003;8(4):233–41.
279
Richardson CA, Jull GA. Muscle control—pain control. What
exercises would you prescribe? Manual Therapy 1995;1:2–10.
Richardson CA, Jull GA, Hides JA, Hodges PW. Therapeutic exercise
for spinal stabilisation. Scientific basis and practical techniques.
London: Churchill Livingstone, Harcourt Brace and Company
Limited; 1999.
Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J.
The relation between the transversus abdominis muscles, sacro-iliac
joint mechanisms, and low back pain. Spine 2002;27(4):399–405.
Richardson CA, Hodges PW, Hides JA. Therapeutic exercise for
lumbopelvic stabilization. A motor control approach for the
treatment and prevention of low back pain. 2nd ed. London:
Churchill Livingstone, Harcourt Brace and Company Limited;
2004.
Stevens VK, Vleeming A, Bouche KG, Mahieu NN, Vanderstraeten
MD, Danneels LA. Electromyographic activity of trunk and hip
muscles during stabilization exercises in four-point kneeling in
healthy volunteers. European Spine Journal 2006, in press.
Stokes IAF, Henry SM, Single RM. Surface EMG electrodes do not
accurately record from lumbar multifidus muscles. Clinical
Biomechanics 2003;18:9–13.
Urquhart DM, Barker PJ, Hodges PW, Story IH, Briggs CA. Regional
morphology of the transversus abdominis and obliquus internus
and externus abdominis muscles. Clinical Biomechanics 2005;20:
233–41.
Van Dieën JH, Cholewicki J, Radebold A. Trunk muscle recruitment
patterns in patients with low back pain enhance the stability of the
lumbar spine. Spine 2003;28(8):834–41.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 280–285

www.elsevier.com/locate/math

Case report

Improved contraction of the transversus abdominis immediately

following spinal manipulation: A case study using
real-time ultrasound imaging
Norman W. Gilla,, Deydre S. Teyhenb, Ian E. Leea
a
US Army-Baylor University Post Professional Doctoral Program in Orthopaedic and Manual Physical Therapy,
Brooke Army Medical Center, San Antonio, TX 78233, USA
b
US Army-Baylor University Doctoral Program in Physical Therapy, Army Medical Department Center and School, San Antonio, TX, USA
Received 28 November 2005; received in revised form 24 February 2006; accepted 27 June 2006

1. Introduction Support exists for an association between spinal

The authors encountered a clinical dilemma when
attempting to apply a clinical prediction rule for
manipulation (Flynn et al., 2002; Childs et al., 2004)
to a patient with a history and physical examination
consistent with clinical lumbar instability (Hicks et al.,
2005). Although the patient met four of five criteria
predicting short-term success with manipulation, the
presence of symptoms suggestive of underlying clinical
instability remains a relative contraindication to thrust
manipulation (Greenman, 1996; Maitland, 2001). Could
the application of both manipulation and stabilization
be logically justified in this patient?
Despite the widespread use of spinal manipulation, the
biological mechanisms by which it produces a beneficial
effect in certain patients are not fully understood (Pickar,
2002). There is evidence which supports a reflexogenic
effect from manipulation in the paraspinal muscles as one
possible mechanism (Herzog et al., 1995, 1999; Lehman
et al., 2001; Pickar, 2002). Specifically, several researchers
have identified altered motor neuron pool excitability
following spinal manipulation (Murphy et al., 1995;
Floman et al., 1997; Herzog et al., 1999; Dishman and
Bulbulian, 2000). The effect on neural pathways asso-
ciated with manipulation has been suggested as one
possible mechanism that may improve muscle perfor-
mance (Pickar, 2002) and patient symptoms.
Corresponding author.
E-mail address: skip.gill@us.army.mil (N.W. Gill).
manipulation and improved muscle function in the
quadriceps (Suter et al., 1999), the erector spinae (Keller
and Colloca, 2000), and the deep neck flexors (Sterling
et al., 2001). Therefore, it is reasonable to hypothesize
that spinal manipulation, by a reflexogenic mechanism,
may improve the performance of the deep trunk
stabilizers. In turn, improved relaxation and contrac-
tility of the lumbar multifidus and the transversus
abdominis (TrA) theoretically could lead to improved
functional stability of the spine through enhancement of
the neurological and active subsystems as defined by
Panjabi (1992a, b).
This single case study describes changes observed in
the TrA musculature pre- to post-manipulation in a
patient that presented with a clinical paradox (symp-
toms suggestive of clinical lumbar instability but also
meeting the clinical prediction rule to succeed with
lumbar manipulation therapy). Real-time ultrasound
imaging (USI) was used to describe the changes in the
TrA musculature.
2. Methods
2.1. Patient
The patient was a 43-year-old male with a 30-day
history of right low back pain and diffuse, posterior
right thigh pain to the knee (Fig. 1). His symptoms had

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.06.014
 ARTICLE IN PRESS
N.W. Gill et al. / Manual Therapy 12 (2007) 280–285
developed insidiously while running. He reported a
history of mild recurrent low back pain for 10 years. The
patient’s most notable examination findings were con-
stant pain in the thigh and a painful catch in flexion and
in extension. Upon screening for a concurrent research
study, he was found to have signs suggestive of clinical
instability (Table 1). However, on examination he was
also noted to have four of five criteria that predict short-
term success with spinal manipulation (Table 2). The
patient consented to receive USI both before and after
spinal manipulation.
2.2. Pre-manipulation training
A previously developed technique (Teyhen et al.,
2005) was used to train an isometric contraction of the
TrA muscle through the abdominal drawing-in man-
oeuvre (ADIM). The patient was first instructed in the
ADIM via traditional training techniques (Richardson
et al., 1999). The patient was instructed to gently pull his
abdominals in toward his spine as he exhaled and
then to maintain this contracted state for 10 s. The
initial training also included tactile cueing with the
patient’s fingers palpating 1 inch medial and inferior to
the anterior superior iliac spine to help confirm
contraction of the TrA. After the patient understood
the intent of the ADIM, he underwent training using
real-time USI as biofeedback. In order to obtain a
Fig. 1. Pain body diagram as annotated on initial evaluation. P1, P2, P3 ¼ pain #1, pain #2, pain #3; C ¼ constant; I ¼ intermittent; V ¼ variable;
Sup. ¼ superficial; check mark ¼ area cleared.
281
ceiling effect, training continued until the patient
was able to perform three isolated TrA contractions
with no further increase in the thickness of the TrA
during the ADIM. The training session lasted for
approximately 15 min.
2.3. Ultrasound instrumentation and measurement
technique
Ultrasound images of the lateral abdominal muscles
(Fig. 2) were obtained both pre- and post-manipulation
using the Sonosite 180 Plus (Sonosite Inc. Bothell, WA)
with a 2–5 MHz curvilinear array in B mode. The
patient was positioned in a supine crook-lying position.
The transducer was placed approximately 1 inch super-
ior to the iliac crest along the mid-axillary line in the
transverse plane. Once an adequate image of the TrA
was obtained, a skin marker (a single line bisecting the
length of the transducer head) was used to standardize
location of the transducer between pre- and post-
manipulation images.
Measurements of muscle thickness of the TrA were
obtained pre- and post-manipulation with the patient at
rest and during the ADIM. To standardize the influence
of respiration (De Troyer et al., 1990; Hodges et al.,
1997; Hodges and Gandevia, 2000a, b) on the thickness
of the TrA, the images were collected upon completion
of the exhalation as determined by visual inspection of
 ARTICLE IN PRESS
282 N.W. Gill et al. / Manual Therapy 12 (2007) 280–285

Table 1 2.4. Manipulation

Criteria used during concurrent study for patients with symptoms
suggestive of clinical instability
The manipulation performed was a regional lumbo-
Criteria Present on examination pelvic manipulation detailed elsewhere (Flynn et al.,
2002; Childs et al., 2004). This regional technique was
1 of 5 aberrant movement signs Yes selected since it was the specific technique used to
Painful arc on flexion Yes
develop (Flynn et al., 2002) and validate (Childs et al.,
Painful arc on return from flexion Yes
Reversal of lumbo-pelvic rhythm Yes 2004) the clinical predictive rule criteria that matched
Gower’s sign No our patient. The manipulation was attempted on the
Deviation in motion No right side first but yielded no cavitation. The patient was
Prone instability test Yes repositioned and the left side was then manipulated.
Hypermobility on lumbar spring test Yes (L3) Again, no cavitation was heard or felt. Finally, a repeat
Recurrent low back pain Yes manipulation on the right yielded a cavitation and the
Straight leg raise 4901 No intervention was considered complete. The patient was
Age o40 years of age No
immediately repositioned on the table and the post-
Based, in part, on criteria predictive of success with lumbar manipulation measurements were captured using the
stabilization exercises; from Hicks et al (2005). previously describe protocol. Care was taken to place
the transducer head in the exact location that was used
to capture the pre-manipulation images.
Table 2
Criteria for clinical predictive rule for regional lumbopelvic
manipulation
2.5. Data analysis

Criteria Present on The statistical significance of the changes in muscle

Symptoms o16 days
FABQ (w) o19
No symptoms distal to knee
At least one level hypomobility in lumbar
spine
At least one hip internal rotation 4351
From Flynn et al. (2002). FABQ (w) ¼ Fear Avoidance Behaviour
Questionnaire Work Subscale.
examination thickness was analysed by considering a previously
No reported standard error of the measure (SEM). The
Yes SEM for the measurement technique described was
Yes established to be 0.031 cm (Teyhen et al., 2005). There-
Yes (L5) fore, a significant change in muscular thickness was
defined as a change greater than two SEMs (0.062 cm);
Yes
thus the authors could be 95% confident that any
change measured was due to an actual change in the
phenomenon being measured and not measurement
error.

3. Results

Changes in the TrA function were noted immediately

Fig. 2. Ultrasound image of the lateral abdominal wall (resting image
taken post-manipulation). White solid arrows indicate thickness.
SST ¼ superficial soft tissue; EO ¼ external oblique; IO ¼ internal
oblique; TrA ¼ transversus abdominis; AC ¼ abdominal contents.
the ultrasound image. The procedures used to measure
muscle thickness of the TrA are described elsewhere
(Teyhen et al., 2005).
post-manipulation. The thickness of the TrA muscle
during the resting and contracted states both pre- and
post-manipulation is provided in Table 3. The total rest-
to-contract difference post-manipulation was 0.38 cm,
nearly doubling the total rest-to-contract thickness pre-
manipulation (0.21 cm). Our patient demonstrated a
statistically significant change in all three measurements
related to the TrA thickness: a decrease in resting
thickness of the TrA, an increase in maximal thickness
during the ADIM, and an increase in total rest-to-
contract thickness of the TrA post-manipulation.
The changes in muscular function noted after the
manipulation were also accompanied by modest clinical
improvements. The patient’s painful catch moving into
extension was abolished and his constant thigh pain
became intermittent. However, his painful catch with
forward bending was only transiently improved.
 ARTICLE IN PRESS
N.W. Gill et al. / Manual Therapy 12 (2007) 280–285
Table 3
Muscular thickness of the transversus abdominis (TrA); measured in centimeters
Pre-manipulation
Resting TrA 0.51
Max contract during ADIM 0.72
Rest-to-contract difference 0.21
Significance (*)42SEMs (0.062 cm); SEM ¼ standard error of the measurement; ADIM ¼ abdominal drawing-in manoeuvre.
4. Discussion
Our results document a dramatic change in one
patient’s ability to perform a preferential TrA contraction
during an ADIM immediately following spinal manipula-
tion. It is notable that the observed improvement occurred
after an established ceiling effect in the performance of the
ADIM was reached. This improvement in the ability to
contract the TrA was also associated with improvements
in some clinical exam findings.
Manipulating a patient with a presentation suggestive
of clinical instability appears counterintuitive.
However, there exists the logical possibility that in some
patients with clinical instability, a high-velocity, low-
amplitude thrust procedure may be a reasonable
treatment option. There are at least three theoretical
mechanisms of action for spinal manipulation including
a mechanical effect on joint arthrokinematics, a neuro-
endocrine effect (such as endorphin release), and a
neurophysiological or reflexogenic effect (Herzog et al.,
1999).
The reflexogenic effect describes the process of a
stretch reflex elicited by manipulation in either the joint
mechanoreceptors or muscle spindles that eventually
leads to an inhibition, depression, or attenuation
of the alpha motor neuron pool thus causing muscle
relaxation that breaks the ‘‘spasm–ischemia–pain–
spasm’’ cycle (Murphy et al., 1995). This has been
demonstrated in the porcine model (Indahl et al., 1997).
In further support of this theory, Lehman et al. (2001)
found an exaggerated surface electromyographic (EMG)
response to painful stimuli in patients with chronic low
back pain, which was attenuated following spinal
manipulation. In another study by Lehman and McGill
(2001) with low back pain patients, they found a
decrease in muscle amplitude on surface EMG during
quiet stance post-manipulation in their most acute
subjects. More specific to our study was a case they
highlighted of a subject who showed a dramatic decrease
in surface EMG activity (41%) during quiet stance post-
manipulation in the area of the right internal oblique
which was qualitatively ‘‘in spasm’’ pre-manipulation
(Lehman et al., 2001). This may be similar to the
decrease in the resting tone of the TrA post-manipula-
tion in our patient.
283
Post-manipulation Difference
0.42 0.09*
0.80 +0.08*
0.38 +0.17*
Two groups of researchers noted a decrease in the S1
alpha motor neuron pool excitability as measured by the
H-reflex in asymptomatic subjects receiving spinal
manipulation (Murphy et al., 1995; Dishman and
Bulbulian, 2000). However, Floman et al. (1997) found
a facilitation instead of an inhibition of the H-reflex
after manipulation in patients with a unilateral her-
niated nucleus pulposus and an initially diminished
H-reflex. Therefore, activated inhibitory and facilitory
pathways may work synergistically to break the
‘‘spasm–ischemia–pain–spasm’’ cycle. In our patient,
improved relaxation of the muscle in the resting state
coupled with increased excitability during contraction
(during the ADIM) may reflect changes in motor neuron
pool excitability following manipulation.
An unexplored, yet potentially critical link exists
between the reflexogenic effect of manipulation and
spinal stabilization theory. The three components of
spinal stabilization include the passive (non-contractile)
subsystem, active (musculotendinous) subsystem, and
neural control subsystems (Panjabi, 1992a, b, 2003).
Clinicians often view instability as a dysfunction of only
the passive subsystem. However, a dysfunction arising in
the neural control or active subsystems could alone, or
together, alter the stabilization system (Panjabi,
1992a, b, 2003; van Dieen et al., 2003). It is possible,
then, that neural inhibition of key spinal stabilization
muscles, such as the TrA and lumbar multifidus,
could cause signs and symptoms consistent with
instability despite the passive constraints being normal.
Manipulation in this case, via a possible reflexogenic
mechanism, may restore or improve the neural control
and active subsystems and thereby minimize the clinical
complaints associated with instability. The authors
propose this is analogous to a ‘‘control+alt+delete’’
procedure performed on a malfunctioning software
programme.
The relationship between inhibition of primary
muscular stabilizers and pain, joint dysfunction, and
instability has already been demonstrated in the lumbar
spine (Hodges and Richardson, 1996; Richardson et al.,
2002; Hungerford et al., 2003). This concept has been
extended to other regions including the cervical spine
(Sterling et al., 2001), the shoulder (Magarey and Jones,
2003), and the knee (Suter et al., 1999; Cowan et al.,
 ARTICLE IN PRESS
284 N.W. Gill et al. / Manual Therapy 12 (2007) 280–285
2001). Demonstrating a relationship between muscle
inhibition and spinal manipulation, Suter et al. (1999)
measured a decrease in quadriceps inhibition in patients
with anterior knee pain after manipulation of the
sacroiliac joint. Furthermore, Keller and Colloca
(2000) used surface EMG activity to assess the isometric
strength of the erector spinae post-manipulation and
noted increased gains compared to a sham manipula-
tion. Therefore, it is likely that a regional lumbopelvic
manipulation could reduce muscle inhibition and/or
facilitate optimal function of the intrinsic lumbar
stabilizers via a reflexogenic mechanism; improved
performance of the TrA and multifidus muscles would,
in turn, result in improved stability to the spine.
There is growing evidence that a combination of
manipulation and lumbar stabilization is effective (Aure
et al., 2003; Niemisto et al., 2003; Childs et al., 2004).
This could be due, in part, to the potential for improved
functioning of the TrA and multifidus muscles after
manipulation. The immediate effect of manipulation
may decrease pain, restore motion, and reset the neural
control and active (muscle) subsystems. This would
facilitate the training of isolated TrA and multifidus
contractions. In essence, the manipulation would
provide a neurophysiological ‘‘window of opportunity’’
to maximize the effectiveness of the first phase of lumbar
stabilization. This is then followed by advanced and
functional lumbar stabilization training to facilitate
motor re-patterning for the long-term rehabilitative
and preventive effect (O’Sullivan et al., 1997, 1998;
Hides et al., 2001). A similar suggestion (application of
mobilization prior to muscle retraining) has been
suggested by Sterling et al. (2001) in the retraining of
the deep neck flexor muscles. Direct support for this
theory was measured in a recent trial by Childs et al.
(2004) in which those patients receiving two manipula-
tions and simple motion exercises during the first week
followed by 3 weeks of lumbar stabilization exercises
experienced greater improvement in short-term and
long-term function than those receiving lumbar stabili-
zation exercises alone.
Threats to internal validity abound in a case study
and therefore it is not possible to ascribe any changes
documented here to a cause and effect relationship.
Training effect is one significant threat in this case study
but the training effect was mitigated by imaging the
lateral abdominal muscles pre-manipulation after a
ceiling effect was noted. Furthermore, the proximity of
the pre- and post-measurements helps to minimize the
influence of a training effect. The authors are currently
investigating the effects of manipulation on motor
control of the TrA and the multifidus using a
randomized control trial. It is important to note that
the patients of interest are those with signs of
clinical instability and not those with frank or gross
instability.
5. Conclusion
In this single case study of a patient with symptoms
suggestive of clinical instability, improvement in the
contraction of the TrA was found immediately following
a regional lumbo-pelvic manipulation. It is theoretically
feasible that the manipulation provided a neurophysio-
logical ‘‘window of opportunity’’ in which to maximize
the preferential activation of the key spinal stabilization
muscles. Future researchers should investigate the
effects of manipulation on neuromuscular control of
the deep trunk muscles and on symptoms suggestive of
clinical instability.
Disclaimer
The opinions or assertions contained here in are the
private views of the Authors and are not to be construed
as official or as reflecting the views of the Departments
of the Army or Defense.
References
Aure OF, Nilsen JH, Vasseljen O. Manual therapy and exercise
therapy in patients with chronic low back pain: a randomized,
controlled trial with 1-year follow-up. Spine 2003;28(6):525–32.
Childs JD, Fritz JM, Flynn TW, Irrgang JJ, Johnson KK, Majkowski
GR, et al. A clinical prediction rule to identify patients with low
back pain most likely to benefit from spinal manipulation: a
validation study. Annals of Internal Medicine 2004;141(12):920–8.
Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J.
Delayed onset of electromyographic activity of vastus medialis
obliquus relative to vastus lateralis in subjects with patellofemoral
pain syndrome. Archives of Physical Medicine and Rehabilitation
2001;82:183–9.
De Troyer A, Estenne M, Ninane V, Van Gansbeke D, Gorini M.
Transversus abdominis muscle function in humans. Journal of
Applied Physiology 1990;68(3):1010–6.
Dishman JD, Bulbulian R. Spinal reflex attenuation associated with
spinal manipulation. Spine 2000;25(19):2519–25.
Floman Y, Liram N, Gilai AN. Spinal manipulation results in
immediate H-reflex changes in patients with unilateral disc
herniation. European Spine Journal 1997;6(6):398–401.
Flynn T, Fritz J, Whitman J, Wainner R, Magel J, Rendeiro D, et al. A
clinical prediction rule for classifying patients with low back pain
who demonstrate short-term improvement with spinal manipula-
tion. Spine 2002;27(24):2835–43.
Greenman PE. Principles of manual medicine, second ed. Philadelphia:
Lippincott Williams & Wilkins; 1996.
Herzog W, Conway PJ, Zhang YT, Gal J, Guimaraes AC. Reflex
responses associated with manipulative treatments on the thoracic
spine: a pilot study. Journal of Manipulative and Physiological
Therapeutics 1995;18(4):233–6.
Herzog W, Scheele D, Conway PJ. Electromyographic responses of
back and limb muscles associated with spinal manipulative therapy.
Spine 1999;24(2):146–53.
Hicks GE, Fritz JM, Delitto A, McGill SM. Preliminary development
of a clinical prediction rule for determining which patients with low
back pain will respond to a stabilization exercise program.
 ARTICLE IN PRESS
N.W. Gill et al. / Manual Therapy 12 (2007) 280–285
Archives of Physical Medicine and Rehabilitation 2005;86(9):
1753–62.
Hides JA, Jull GA, Richardson CA. Long-term effects of specific
stabilizing exercises for first-episode low back pain. Spine
2001;26(11):E243–8.
Hodges PW, Gandevia SC. Activation of the human diaphragm during
a repetitive postural task. Journal of Physiology 2000a;522:165–75.
Hodges PW, Gandevia SC. Changes in intra-abdominal pressure
during postural and respiratory activation of the human dia-
phragm. Journal of Applied Physiology 2000b;89(3):967–76.
Hodges PW, Richardson CA. Inefficient muscular stabilization of the
lumbar spine associated with low back pain. A motor control
evaluation of transversus abdominis. Spine 1996;21(22):2640–50.
Hodges PW, Gandevia SC, Richardson CA. Contractions of specific
abdominal muscles in postural tasks are affected by respiratory
maneuvers. Journal of Applied Physiology 1997;83(3):753–60.
Hungerford B, Gilleard W, Hodges P. Evidence of altered lumbopelvic
muscle recruitment in the presence of sacroiliac joint pain. Spine
2003;28(14):1593–600.
Indahl A, Kaigle AM, Reikeras O, Holm SH. Interaction between the
porcine lumbar intervertebral disc, zygapophysial joints, and
paraspinal muscles. Spine 1997;22(24):2834–40.
Keller TS, Colloca CJ. Mechanical force spinal manipulation increases
trunk muscle strength assessed by electromyography: a compara-
tive clinical trial. Journal of Manipulative and Physiological
Therapeutics 2000;23(9):585–95.
Lehman GJ, McGill SM. Spinal manipulation causes variable spine
kinematic and trunk muscle electromyographic responses. Clinical
Biomechanics 2001;16(4):293–9.
Lehman GJ, Vernon H, McGill SM. Effects of a mechanical pain
stimulus on erector spinae activity before and after a spinal
manipulation in patients with back pain: a preliminary investiga-
tion. Journal of Manipulative and Physiological Therapeutics
2001;24(6):402–6.
Magarey ME, Jones MA. Dynamic evaluation and early management
of altered motor control around the shoulder complex. Manual
Therapy 2003;8(4):195–206.
Maitland GD. Vertebral manipulation, sixth ed. Oxford: Butterworth-
Heinemann; 2001.
Murphy BA, Dawson NJ, Slack JR. Sacroiliac joint manipulation
decreases the H-reflex. Electromyography and Clinical Neurophy-
siology 1995;35(2):87–94.
285
Niemisto L, Lahtinen-Suopanki T, Rissanen P, Lindgren KA, Sarna S,
Hurri H. A randomized trial of combined manipulation, stabilizing
exercises, and physician consultation compared to physician
consultation alone for chronic low back pain. Spine 2003;28(19):
2185–91.
O’Sullivan PB, Twomey LT, Allison GT. Evaluation of specific
stabilizing exercise in the treatment of chronic low back pain with
radiologic diagnosis of spondylolysis or spondylolisthesis. Spine
1997;22(24):2959–67.
O’Sullivan PB, Twomey LT, Allison GT. Altered abdominal muscle
recruitment in patients with chronic back pain following a specific
exercise intervention. Journal of Orthopaedic & Sports Physical
Therapy 1998;27(2):114–24.
Panjabi M. The stabilizing system of the spine. Part I. Function,
dysfunction, adaptation, and enhancement. Journal of Spinal
Disorders 1992a;5(4):383–9.
Panjabi M. The stabilizing system of the spine. Part II. Neutral zone and
instability hypothesis. Journal of Spinal Disorders 1992b;5(4):390–7.
Panjabi M. Clinical spinal instability and low back pain. Journal of
Electromyography and Kinesiology 2003;13(4):371–9.
Pickar JG. Neurophysiological effects of spinal manipulation. Spine
Journal 2002;2(5):357–71.
Richardson CA, Jull G, Hodges P, Hides JA. Therapeutic exercise for
spinal segmental stabilization in low back pain. London: Churchill
Livingstone; 1999.
Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J.
The relation between the transversus abdominis muscles, sacroiliac
joint mechanics, and low back pain. Spine 2002;27(4):399–405.
Sterling M, Jull G, Wright A. Cervical mobilisation: concurrent effects
on pain, sympathetic nervous system activity and motor activity.
Manual Therapy 2001;6(2):72–81.
Suter E, McMorland G, Herzog W, Bray R. Decrease in quadriceps
inhibition after sacroiliac joint manipulation in patients with
anterior knee pain. Journal of Manipulative and Physiological
Therapeutics 1999;22(3):149–53.
Teyhen DS, Miltenberger CE, Deiters HM, Del Toro YM, Pulliam JN,
Childs JD, et al. The use of ultrasound imaging of the abdominal
drawing-in maneuver in subjects with low back pain. Journal of
Orthopaedic & Sports Physical Therapy 2005;35(6):346–55.
van Dieen JH, Selen LP, Cholewicki J. Trunk muscle activation in low-
back pain patients: an analysis of the literature. Journal of
Electromyography and Kinesiology 2003;13(4):333–51.
 ARTICLE IN PRESS

Manual Therapy 12 (2007) 286–294

www.elsevier.com/locate/math

Technical and measurement report

Measurement of segmental cervical multifidus contraction by

ultrasonography in asymptomatic adults
Jo-Ping Leea, Wen-Yih I. Tsengb, Yio-Wha Shauc, Chung-Li Wangd,
Hsing-Kuo Wanga, Shwu-Fen Wanga,
a
School and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, Floor 3, No.17, Xuzhou Road,
Zhongzheng District, Taipei City 100, Taipei, Taiwan, ROC
b
Center for Optoelectronic Biomedicine, College of Medicine, National Taiwan University, Taipei, Taiwan, ROC
c
Institute of Applied Mechanics, College of Engineering, National Taiwan University, Taipei, Taiwan, ROC
d
Department of Orthopedic Surgery, National Taiwan University Hospital, Taipei, Taiwan, ROC
Received 15 June 2005; received in revised form 12 May 2006; accepted 10 July 2006

Abstract
The deep muscles that play significant roles in maintaining segmental stability have been measured using ultrasonography (US).
However, few studies have been carried out to determine the reliability and validity of US for measuring the cervical multifidus
during contraction. The aims of this investigation were to evaluate the reliability of the dimensions of the cervical multifidus as
measured using US and compare the US measurements with those determined with magnetic resonance imaging (MRI), the gold
standard. Ten asymptomatic adult subjects (age, 28.573.5 years) participated in testing–retesting of muscle dimensions at rest
and during isometric head extension with the cranio-cervical spine maintained in a flexed position against individual maximum
resistance. Ten asymptomatic adult subjects (age, 28.174.1 years) participated in testing to compare US and MRI measurements of
the thickness, width, area, and shape ratio of the cervical multifidus at the C4, C5, and C6 levels. US measurements of
muscle thickness at the C4, C5, and C6 levels at rest were 0.6770.14, 0.7070.20 and 0.7370.09 cm, respectively; the corresponding
measurements as determined by MRI were 0.7070.12, 0.6770.15 and 0.7070.06 cm. The within-subject coefficient of variation
(CVw) for thickness at rest and during contraction was less than 10%, indicating acceptable reliability. US measurement of thickness
had better reliability and validity. The range of limit of agreement for muscle thickness at the C4, C5, and C6 levels was 0.20 to
0.20 cm, and the range of R2 was 0.42–0.64. The thickness of the cervical multifidus muscle increased significantly during contraction
(1.1370.20, 1.1970.20 and 1.1770.12 cm for the C4, C5, and C6 levels, Po0.05). However, no significant differences were noted
among the three levels. The results indicate that US can detect changes in segmental cervical multifidus during contraction.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Ultrasonography; Cervical multifidus; Validity; Reliability

1. Introduction neck pain had occurred during the previous 6 months

Neck pain, which has been thought to be caused by
neck trauma or occupational repetitive injuries, is
extremely prevalent in modern society (Cote et al.,
1998; Cote et al., 2000). In a Canadian study of adults,
64% of respondents reported that they had experienced
neck pain during their life time and 54% reported that
Corresponding author. Tel.: +2 33228139; fax: +2 332218161.
E-mail address: sfwang@ntu.edu.tw (S.-F. Wang).
(Cote et al., 1998). Moreover, nearly 5% of the
respondents reported that they had been considerably
disabled because of neck pain during the previous 6
months (Cote et al., 1998). These findings highlight the
need to better understand the pathogenesis of the
cervical spine. Novel measurement techniques can help
to advance this understanding.
The deep muscles play major roles in maintaining
segmental stability, (Panjabi et al., 1989; Kristjansson
and Jonsson, 2002; Kristjansson et al., 2003) and

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.07.008
 ARTICLE IN PRESS
J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294
evidence suggests an association between deep muscle
dysfunction and spinal pain (Hides et al., 1994, 1996;
Hayashi et al., 2002; Kristjansson, 2004). The impor-
tance of deep lumbar muscles (Panjabi et al., 1989;
Hides et al., 1994; Hides et al., 1996) and abdominal
muscles (Hodges and Richardson, 1996; Hodges and
Richardson, 1997) for controlling back stability has
been demonstrated in biomechanical models (Panjabi
et al., 1989) and clinical experiments (Hides et al., 1994,
1996; Hodges and Richardson, 1996, 1997). Further-
more, in patients with low back pain, intervertebral
segmental muscle atrophy has been found (Hides et al.,
1994, 1996; Campbell et al., 1998) with use of magnetic
resonance imaging (MRI) (Campbell et al., 1998) or
ultrasonography (US) (Hides et al., 1994, 1996).
Similarly, muscle atrophy has been noted after neck
injury (Hayashi et al., 2002; Kristjansson, 2004). In
addition, altered configuration of the cervical lordosis
has been found in patients with neck disorders (Ueki
et al., 1995; Kristjansson and Jonsson, 2002; Kristjans-
son et al., 2003). One possible explanation for this
altered configuration is the inability of the deep cervical
segmental muscles to maintain cervical alignment
(Kristjansson, 2004).
MRI is considered to be the gold standard (West-
brook and Kaut, 1993; Hides et al., 1995; Esformes and
Narici, 2002) for measuring the lumbar multifidus at rest
(Hides et al., 1994, 1996; Stokes et al., 2005), and US has
also been demonstrated to be a valid and reliable
method. US has the advantages of a lower cost
(Esformes and Narici, 2002; Kristjansson, 2004) and
the ability to provide non-invasive visualization of the
change in architecture of the deep cervical (Esformes
and Narici, 2002; Hodges et al., 2003; Kristjansson,
2004) and lumbar muscles (Ito et al., 1998). Further-
more, good reliability was obtained in measuring the
thickness of the transversus abdominis muscle during
the hollowing maneuver (McMeeken et al., 2004).
However, to our knowledge, there have been no reports
on the validity and reliability of US for measuring the
cervical multifidus muscle during contraction. Changes
in the dimensions of the cervical multifidus during
contraction with maximum resistance are also unknown.
US was suggested to be the reliable method for detecting
isometric contractions of low maximal voluntary con-
traction (Hodges et al., 2003), but the changes in muscle
Table 1
Demographic details of participants
Total
Men (n ¼ 11) Women (n ¼ 6)
Age (years) 26.0974.0 27.1774.0
Height (cm) 179.977.6 157.172.6
Body mass (kg) 65.675.8 56.376.4
287
dimensions and force differed among the muscles tested
(Hodges et al., 2003; McMeeken et al., 2004).
The purposes of this study were to determine the
reliability of US for measuring changes in the dimen-
sions of the cervical multifidus muscle during rest and
contraction and to evaluate the validity of US as
compared with MRI. The specific aims were to (1)
investigate the intra-rater, intersession reliability of US
for measuring the thickness, width, area, and shape ratio
of the cervical multifidus at the C4, C5, and C6 levels at
rest and during isometric head extension with cranio-
cervical spine maintained in a flexed position against
individual maximum resistance; (2) compare US and
MRI measurements of the thickness, width, area, and
shape ratio of the cervical multifidus at the C4, C5, and
C6 levels under static condition; and (3) determine the
change in the dimensions of the cervical multifidus at the
C4, C5, and C6 levels during contraction.
2. Subjects and methods
2.1. Participants
After Institutional Review Board (IRB) approval, 17
asymptomatic volunteers (age, 26.7573.8 years) were
recruited (Table 1). Most of the participants exercised
routinely (mean of 1.470.7 times per week, for a mean
of 1.571.0 h each time). Criteria for exclusion were a
history of trauma to the cervical, thoracic, or lumbar
spine, previous surgery of the spine, and neck pain
within the previous three months. All participants gave
their informed consent after the nature of the proce-
dures had been fully explained. Reliability and validity
testings were carried out with 10 participants each; thus
four subjects participated in both groups.
2.2. US
US was performed using a real-time ultrasound
apparatus with a 10 MHz linear array transducer (HDI
5000; ATL Ultrasound, Bothell, WA, USA). With this
system, the axial resolution of the image was 0.7 mm at a
depth of 3 cm and the horizontal resolution was 1.0 mm
at a depth of 4–6 cm. The measurement protocol using
US was designed through several pilot trials and was
Reliability Validity
Men (n ¼ 7) Women (n ¼ 3) Men (n ¼ 6) Women (n ¼ 4)
27.574.7 29.073.5 28.1773.8 27.0574.7
168.576.3 156.672.8 170.876.4 158.772.3
67.075.6 54.877.6 70.073.6 58.372.0
 ARTICLE IN PRESS
288 J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294
based on a fundamental knowledge of cervical anatomy
and US.
The bifurcation of the spinal process at the C2–C6
cervical levels was identified manually and on the US
images in the following order. The C2 level was
identified as the first bifurcation of the spinal process
palpated from the occipital bone downward. The C7
level was palpated as the most prominent process among
the cervical levels where no bifurcation was found on the
US image. The C4 level was identified as the second
bifurcation of the spinal process palpated downward
from the C2 level, and the C5 and C6 levels were
identified as the first and second bifurcation of the spinal
process caudal to the C4 level. The cervical spinal
processes were palpated and marked by the rater, a
physical therapist, with US guidance to identify the
bifurcation of the spinal process. This method of
identifying the segmental spinal processes has been
shown to be acceptably reliable (intra-rater within-
subject coefficient of variation [CVw] of 6.4–10.7%) (Wu
et al., 2005).
2.3. MRI
MRI was performed with a 1.5 Tesla scanner
(Magnetom Sonata, Siemens, Erlangen, Germany). A
flexible surface coil, 20  50 cm, was used as a receiver
coil and was fixed over the posterior aspect of the
participant’s neck. The participant was placed in
the same prone position as for US. The locations of
the vertebral levels were determined from parallel
images in the mid-sagittal plane. To position the slices
of MRI at the same location as those obtained by US,
multiple transverse slices were scanned from the super-
ior border of the C3–C4 intervertebral disk to the
Fig. 1. Images obtained from magnetic resonance imaging (MRI). (A) Sagittal view of the cervical spine. Multiple transverse slices were scanned
from the superior border of the C3–C4 intervertebral disk to the inferior border of the C6–C7 intervertebral disk at 5-mm intervals. (B) The square
indicates the image view relevant to the ultrasonography (US) view. The region within white line is the area of the right side of the cervical multifidus.
inferior border of the C6–C7 intervertebral disk at 5-mm
intervals (Fig. 1A). From these images, the rater, an
experienced radiologist, determined the C4, C5, and C6
levels by identifying the bifurcation of the spinal process
and the articular process (Fig. 1A). T1-weighted images
were obtained with use of half-Fourier single-shot turbo
spin-echo acquisition (HASTE); the parameters were
TR/TE/flip angle of 1500/114 ms/901; the matrix size
was 256  256, the field of view was 21 cm, the slice
thickness was 5 mm, and the number of average was 20.
2.4. Reliability of US
Reliability testing was carried out in 10 asymptomatic
participants (four women and six men; age, 28.1074.1
years) (Table 1). To standardize the position of the
participants during measurement, each participant was
asked to sit upright in a customized chair and rest their
arms on their legs. A head belt was fixed over the
forehead of the participant to maintain the head and
neck in a neutral position (Fig. 2). The participant was
asked to extend both knees, and the heels of their feet
were placed on a 15-cm height board to eliminate the
overflowing force of the lower extremities under the
conditions of rest and the isometric head extension with
the cranio-cervical spine maintained in a flexed position
against maximum resistance.
After each cervical level was identified, participants
were asked to perform isometric extension of the head
using maximum force, against a board placed behind the
head at occipital level, with the cranio-cervical spine
maintained in a flexed position. Participants practiced
this isometric head extension maneuver twice under the
supervision of the rater, to ensure that the head and
neck underwent a pure horizontal movement without
 ARTICLE IN PRESS
J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294 289

cranio-cervical spine moving toward flexion or exten-
sion. Because the participants were asymptomatic, each
individual maximum force was assumed to be the same
for each participant. The maximum contraction could
activate the neural pathway and neuromuscular junction
as well as the contractile and non-contractile tissue of
the muscles. The two practice isometric head extension
helped to maintain the condition of the cervical muscle
before measurement. US of the multifidus muscle on the
right at the C4, C5, and C6 levels, in sequence, were
recorded for the conditions of rest and isometric head
extension contraction. Proper positioning of the parti-
cipant was ensured after each test, and the same
procedure was repeated twice in a 20-min interval by
the same rater.
The multifidus muscle is located lateral to the junction
of the spinal process and lamina; dorsolateral to the
lamina of vertebrae, medial to the articular process, and
ventral to the interfaces of the fasciae of the semispinalis
cervicis (SSC) muscle (Fig. 3A) (Gerhardt and Fromm-
hold, 1991). The orientation of the transducer was
adjusted to bring it horizontal to the transverse plane.
To obtain the clearest image, the US transducer was
held by a custom-designed device and was tilted slightly
(up or down) until the clearest interface lines could be
observed. An US image was obtained for each cervical
level, and the boundary of the muscle was identified and
measured manually using a custom-written C++
computer graphic program.
The muscle thickness, width, and area of the cervical
multifidus muscle were measured at the C4–C6 levels.
The muscle thickness was measured as the largest
distance from the dorsal to the ventral boundary of
the multifidus, and the width was measured as the
largest distance between the medial and lateral bound-
aries of the muscle (Fig. 3). The area was measured as
the region along the boundary lateral to the junction of
the spinal process and lamina; dorsolateral to the lamina
of the vertebrae, medial to the articular process, and
ventral to the interfaces of the fasciae of the SSC
muscle. The shape ratio was calculated as the lateral
diameter (width) divided by the ventrodorsal diameter
(thickness).

2.5. Validity of US

Validity testing was carried out in 10 asymptomatic

Fig. 2. Experimental setup for the isometric head extension with the
cranio-cervical spine maintained in a flexed position.
participants (three women and seven men; age,
26.273.5 years). Each participant was asked to lie in a
comfortable and relaxed prone position on an examina-
tion table, with both arms resting symmetrically beside
the body. A head support was used to maintain the head
in the neutral position. Flexion and extension of the
neck were controlled by maintaining the posterior part

Fig. 3. Ultrasonography (US) images of the transverse plane of the right side of the cervical multifidus at the C5 level at rest (A) and during
contraction (the isometric head extension with the cranio-cervical spine maintained in a flexed position) (B). The borders of the multifidus muscle are:
medially, the spinous process (SP); dorsolaterally, the semispinalis cervicis (SSC); laterally, the articular process (AP); and inferiorly, the lamina. The
region within the dotted line indicates the area of the multifidus muscle. T ¼ thickness, and W ¼ width.

290
of the head and the thoracic region on a horizontal level
with use of an oil-filled inclinometer.
The spinal process level was identified, and MRI of
the cervical multifidus at each level was carried out,
followed by US. The multifidus was located lateral to
the junction of the spinal process and lamina; dorso-
lateral to the lamina of vertebrae, medial to the articular
process, and ventral to the interfaces of the fasciae of the
SSC muscle (Fig. 3A) (Gerhardt and Frommhold,
1991). As with testing for reliability, the transducer
was held by a custom-designed device and was tilted
slightly (up or down) until the clearest interface lines
could be observed. The orientation of the transducer
was adjusted to bring it perpendicular to the horizontal
plane, calibrated by a vertical line using a plume. US
and MRI images obtained for each cervical level and the
boundary of the muscle were identified and measured
manually using a custom-written C++ computer
graphic program.
2.6. Statistical analysis
The reliability of intra-rater, intersession was de-
scribed using CVw and between-subject coefficient of
variation (CVb). The variance within each subject was
presented as the percentage of the average coefficient of
variation (CV) for each subject. The variance between
subjects was presented as the percentage of the average
CV from each measurement (Rainoldi et al., 1999).
Thus,
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u ,
Pn uP m  2
t xij  xi ðm  1Þ xi
i¼1 j¼1
CVw ¼
n There were no significant interactions between the
s
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
n 
 
P
m P 2
xij  xj ðn  1Þ
j¼1 i¼1
CVb ¼
m and B) and the main effect of cervical level did not differ
In these equations, n denotes the number of subjects
(n ¼ 10), m denotes the number of repeated measure-
ments (m ¼ 2), and X ij denotes the US measurement for
the ith subject in the jth repeated measurement.
A limit of agreement was used to evaluate the
differences between the geometric parameters of the
cervical multifidus as measured on US and MRI.
Additionally, the relationships of the values obtained
with US and MRI were determined by regression
analysis.
The changes in the thickness of the cervical multifidus
muscle at the C4–6 levels and at two grades of resistance
(rest and contraction) were evaluated by two-way
analysis of variance (ANOVA) with repeated measure-
ment (three cervical levels [C4, C5, and C6]  two grades
of resistance [rest and contraction]) with use of
ARTICLE IN PRESS
J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294
SPSS software (version 10.0; SPSS Inc., Chicago,
Illinois, USA).
3. Results
3.1. Reliability and validity of US
The CVw for muscle thickness, width, area, and shape
ratio of the cervical multifidus muscle at rest were within
10%, with the exception of the shape ratio at the C4
level (CVw ¼ 11.39%). During contraction, the CVw
was within 10% for thickness at the C4–C6 levels, width
at the C4 and C5 levels, and area at the C4 level. Hence,
the CVw was less than 10% for the measurement of
muscle thickness at the C4–C6 levels, both at rest and
during contraction (Table 2).
There were significant relationships between the
muscle thickness at the C4–C6 levels as measured by
US and MRI (Table 3). Data from the three cervical
levels were pooled for regression analysis and determi-
nation of R2 values. The R2 values for thickness at the
C4, C5, and C6 levels and pooled value were 0.42, 0.64,
0.43 (Table 3, Po0:05), and 0.60 (y ¼ 1:0239x20:0005,
Po0.05). The R2 values for width were 0.02, 0.11, 0.22
(Table 3, P40:05), and 0.02 (y ¼ 0:1275x þ 1:5184,
P40.05); the R2 values for area 0.39 (Table 3, Po0:05),
0.11, 0.29, (P40.05), and 0.25 (y ¼ 0:6351x þ 0:4244,
P40.05); and the R2 values for shape ratio were 0.13,
0.28, 0.02 (Table 3, P40:05) and 0.05 (y ¼ 0:2247xþ
2:0092, P40.05).
, ffi
3.2. Change in muscle dimensions
,
measurement of muscle thickness at the three cervical
levels and the two grades of resistance (F(2, 18) ¼ 2.620,
xj P40.05). The main effect of contraction did differ
. significantly (F(1, 9) ¼ 22.875, Po0.05, Table 2; Fig. 3A
significantly (F(2, 18) ¼ 1.353, P40.05). The muscle
thickness increased significantly during isometric head
extension with the cranio-cervical spine maintained in a
flexed position against maximum resistance, indicated
that the muscle thickness changes during contraction
(Table 2, Fig. 4).
4. Discussion
In the present study, we attempted to determine
whether US was a reliable and valid method for
measuring the dimensions of the cervical multifidus at
rest and during contraction. Intra-rater, intersession
reliability for measurement of muscle thickness was
acceptable during both rest and contraction. Compared
 ARTICLE IN PRESS
J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294 291

Table 2
Coefficients of variation (within subject [CVw] and between subjects [CVb] for measurements of the dimensions of the cervical multifidus at rest and
during contractiona

Measurement CVw (%)b CVb (%)b

Session 1 Session 2

Rest
Thickness (cm)
C4 0.7270.10 0.7070.07 7.89 (4.27–11.49) 12.78 (12.21–13.35)
C5 0.6870.08 0.7170.12 7.22 (4.55–9.88) 14.57 (13.83–15.30)
C6 0.7770.09 0.8270.12 6.49 (3.81–9.16) 14.21 (13.39–15.03)
Width (cm)
C4 1.7070.20 1.7670.38 8.22 (1.45–14.99) 16.99 (14.15–19.81)
C5 1.8070.25 1.8370.34 3.88 (1.82–5.93) 16.42 (12.95–19.88)
C6 1.9070.38 2.0770.40 4.57 (0.77–8.37) 19.85 (10.16–29.54)
Area (cm2)
C4 0.9270.16 0.9170.05 7.89 (4.27–11.49) 19.11 (17.11–21.10)
C5 0.9670.16 1.0470.27 3.92 (0.02–7.88) 21.87 (19.87–23.85)
C6 1.2070.29 1.3970.33 6.49 (3.81–9.16) 24.35 (18.32–29.54)
Shape ratio (lateral/ventrodorsal dimension)
C4 2.4070.47 2.5370.60 11.39 (1.82–20.96) 21.71 (5.08–38.32)
C5 2.6770.52 2.6470.66 7.47 (1.25–13.69) 22.24 (4.38–40.08)
C6 2.5470.63 2.5870.56 5.64 (1.50–9.78) 12.47 (15.4840.77)
Contraction
Thickness (cm)
C4 1.1370.20c 1.1870.16 6.86 (3.63–10.11) 15.87 (13.84–17.88)
C5 1.197020c 1.1670.15 8.73 (1.64–15.83) 15.26 (11.86–18.64)
C6 1.1770.12c 1.1370.13 6.72 (3.58–9.85) 11.83 (10.81–12.85)
Width (cm)
C4 1.4370.34 1.4270.53 7.65 (3.71–19.81) 19.73 (14.74–24.72)
C5 1.3570.33 1.3970.55 8.67 (3.77–13.57) 18.64 (13.82–23.46)
C6 1.3470.42 1.4670.48 10.43 (0.32–20.54) 20.80 (5.77–35.81)
Area (cm2)
C4 1.3570.30 1.3770.41 8.92 (2.33–15.51) 23.89 (15.47–32.29)
C5 1.3170.31 1.2670.29 10.72 (4.64–16.79) 16.46 (7.8–25.11)
C6 1.3370.31 1.2770.26 10.64 (6.46–14.81) 13.87 (9.65–18.08)

Shape ratio (lateral/ventrodorsal dimension)

C4 1.2870.36 1.2270.51 12.89 (3.91–21.86) 16.79 (4.46–29.10)
C5 1.1870.44 1.2470.58 11.56 (4.56–18.56) 21.92 (2.42–46.27)
C6 1.1770.39 1.3270.49 11.16 (0.61–22.94) 21.63 (9.47–33.78)
a
Measurements given as the mean and standard deviation (SD).
b
Value presented as mean ((mean1 SD)(mean+1 SD)).
c
Significantly different from the value measured at rest, Po0.05.

with MRI, US had an acceptable validity for measuring
the thickness of the cervical multifidus muscle but not
for measuring its width and area. Additionally, US
detected significant increases in muscle thickness during
contraction.
A significant linear relationship of measurements was
found between the two modalities for muscle thickness
but not for muscle width, area, or shape ratio. In a
previous study, no difference was found between US
and MRI measurements of the cross-sectional area of
the lumbar multifidus (Hides et al., 1995). The cervical
multifidus muscle is much smaller than the lumbar
multifidus, and the variation in measurement may have
had a significant influence on the R2 value. In addition,
the lateral boundary of the cervical multifidus is not
distinct on either US (Kristjansson, 2004) or MRI
because its attachment, the articular processes of the
cervical vertebrae (Anderson et al., 2005), are the
common insertion areas of the semispinalis capitis
muscle (Platzer, 2005) or another passing muscle fiber
(Stokes et al., 2005). Thus, the lack of a significant
relationship between the US and MRI measurements of
width, area, and shape ratio may have been due to the
size and anatomic structure of the cervical multifidus.
Another possible reason for the lack of a significant
relationship is the anatomical alignment of the cervical
 ARTICLE IN PRESS
292 J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294

Table 3
Limit of agreement and R2 values for the dimensions of the cervical multifidus muscle at the C4, C5, and C6 levels, as measured by ultrasonography
(US) and magnetic resonance imaging (MRI)a

US MRI Mean differenceb R2 p

Thickness (cm)
C4 0.6770.14 0.7070.12 0.0170.16 0.42 0.04
C5 0.7070.20 0.6770.15 0.00370.20 0.64 0.005
C6 0.7370.09 0.7070.06 0.0270.14 0.43 0.03
Width (cm)
C4 1.7270.28 1.4370.16 0.2970.62 0.02 0.69
C5 1.6570.20 1.4170.23 0.2470.70 0.11 0.33
C6 1.7470.19 1.6470.40 0.0970.70 0.22 0.16
Area (cm2)
C4 0.9970.35 0.8870.20 0.1070.54 0.39 0.05
C5 0.9570.30 0.8270.22 0.1270.60 0.11 0.34
C6 1.0470.28 0.9970.28 0.0470.54 0.29 0.10

Shape ratio (lateral/ventrodorsal dimension)

C4 2.5070.48 2.0870.46 0.4271.06 0.13 0.31
C5 2.5170.63 2.1870.57 0.3371.17 0.28 0.11
C6 2.5070.57 2.4070.61 0.1171.80 0.02 0.66
a
Values given as the mean and standard deviation (SD).
b
Value given as the mean and two SDs.

rest identified by a hyperecho (Hides et al., 1992; Stokes

1.5
∗ contraction ∗ ∗ et al., 2005). In contrast, because the relaxation time for
fat tissue is shorter than that for water or protein, fat
thickness (cm)

1 tissue could be distinguished more clearly from muscle

0.5
0
1 2
level
Fig. 4. Comparison of thickness measured at rest and during
contraction (the isometric head extension with the cranio-cervical
spine maintained in a flexed position) at the three cervical levels.
*Significant difference between the values measured at rest and during
contraction, Po0.05.
multifidus in relation to the resolution of the US image.
The muscle tissue of the cervical multifidus was aligned
from the spinal process to the articular process, which
was perpendicular to the traveling direction of the
soundwave (Kremkau, 1989). This alignment made the
thickness at the muscle interface easily identifiable
because the axial resolution of US is better than the
horizontal resolution (Kremkau, 1989). Also, the pre-
sence of fat tissue around the muscle fascia results in
slightly different tissue boundaries seen on images from
the two modalities. Fat tissue between the layers of
muscle fascia may influence the accuracy of the US
measurement. Fat tissue could not be easily distin-
guished from muscles on US because both were
on MRI (Westbrook and Kaut, 1993). Thus, the use of
US to measure the cervical multifidus in sedentary adults
with possible fat infiltration requires further verification.
The different scanning planes of US and MRI may
also cause differences in the measurements for the two
3 modalities. All participants were placed in the same
comfortable and relaxed prone position on an examina-
tion table, with both arms resting symmetrically beside
the body and the head maintained in the neutral
position by a head support. However, due to lordosis,
as Hides et al. (1995) discussed, if MRI slices are not
manually positioned perpendicular to the muscle at each
vertebral segment, the images will pass obliquely
through the muscle and show different cross-sectional
areas. Indeed, in the cervical region, lordosis also plays a
role in the measurement of muscle thickness. Even if
osseous landmarks, such as lamina and articular
processes, are used to locate the plane of scanning by
US and MRI, the scanning plane by US would not fine
adjusted to be perpendicular to the muscle fiber to
obtain a clearest US image, which may explain the
differences between the US and MRI measurements.
The measurement for the centre segment of the cervical
lordosis (C4) was estimated with less error; this was
supported by our data (Table 3).
The present study was the first, to our knowledge, to
document contraction of the cervical multifidus with use
of US. Several cervical muscles, including the splenius
 ARTICLE IN PRESS
J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294
capitis (Soltani et al., 1996), semispinalis capitis
(Rezasoltan et al., 1998; Rezasoltan et al., 2002), and
multifidus (Kristjansson, 2004) have been measured with
US during rest, with acceptable reliability. Previous
studies have also demonstrated acceptable reliability of
US measurements of the lumbar multifidus during rest
(Hides et al., 1992; Hides et al., 1995). A significant
change in the area of the semispinalis capitis muscle has
been observed during contraction (Rezasoltan et al.,
2002). Good reliability (ICC ¼ 0.98) of the US measure-
ment of thickness of the transversus abdominis during
contraction has also been reported (McMeeken et al.,
2004), and there was a significant change in thickness
during abdominal holding (Hodges et al., 2003) and the
abdominal hollowing maneuver (McMeeken et al.,
2004). Through a strict protocol, we were able to obtain
acceptable reliability of measurements during contrac-
tion of the cervical multifidus, which is smaller than its
lumbar counterpart. Contraction was achieved with the
isometric head extension, which has been shown in this
study to activate the cervical multifidus in adults.
A limitation to using US to measure muscle thickness
is the assumption of constant sound velocity in different
tissues, including skin, fat tissue, fascia, and muscles
(Kremkau, 1989). Therefore, the muscle thickness
measured by US is an estimated value. To reduce these
errors, either normalization (dividing the value during
contraction by the initial value of the muscle geometry)
or continuous US measurements should be considered
in a future study to investigate the change pattern of the
muscle under contraction.
5. Conclusion
Using noninvasive, real-time US with a standard
protocol, the thickness of the cervical multifidus muscle
at the C4, C5, and C6 levels in asymptomatic young
adults could be quantified during contraction. The
application of this methodology for individuals with
chronic neck pain should be further explored.
References
Anderson JS, Hsu AW, Vasavada AN. Morphology, architecture, and
biomechanics of human cervical multifidus. Spine 2005;30:E86–91.
Campbell WW, Vasconcelos O, Laine FJ. Focal atrophy of the
multifidus muscle in lumbaosacral radiculopathy. Muscle & Nerve
1998;21:1350–3.
Cote P, Cassidy JD, Carroll L. The Saskatchewan Health and Back
Pain Survey. The prevalence of neck pain and related disability in
Saskatchewan adults. Spine 1998;23:1689–98.
Cote P, Cassidy JD, Carroll L. The factors associated with neck pain
and its related disability in the Saskatchewan population. Spine
2000;25:1109–17.
Esformes R, Narici MV. Measurement of human muscle volume using
ultrasonography. European Journal of Applied Physiology 2002;
87:90–2.
293
Gerhardt P, Frommhold W. Altas of anatomic correlations in CT and
MRI. New York: Thieme Medical Publishers, Inc.; 1991.
Hayashi N, Masumoto T, Abe O, Aoki S, Ohtomo K, Tajiri Y.
Accuracy of abnormal paraspinal muscle findings on contrast-
enhanced MR images as indirect signs of unilateral cervical root-
avulsion injury. Radiology 2002;223:397–402.
Hides JA, Cooper DH, Stokes MJ. Diagnostic ultrasound imaging for
measurement of the lumbar multifidus muscle in normal young
adults. Physiotherapy Theory and Practice 1992;8:19–26.
Hides JA, Stokes IA, Saide M, Jull GA, Cooper DH. Evidence of
lumbar multifidus muscle wasting ipsilateral to symptoms in
patients with acute/subacute low back pain. Spine 1994;19:165–72.
Hides JA, Richardson CA, Jull GA. Magnetic resonance imaging and
ultrasonogrpahy of the lumbar multifidus. Spine 1995;20:54–8.
Hides JA, Richardson CA, Jull GA. Multifidus muscle recovery is not
automatic after resolution of acute, first-episode low back pain.
Spine 1996;21:2763–9.
Hodges PW, Richardson C. Feedforward contraction of transversus
abdominis is not influenced by the direction of arm movement.
Experimental Brain Research 1997;114:362–70.
Hodges PW, Richardson CA. Inefficient muscular stabilization of the
lumbar spine associated with low back pain. A motor control
evaluation of transversus abdominis. Spine 1996;21:2640–50.
Hodges PW, Pengel LHM, Herbert RD, Gandevia SC. Measurement
of muscle contraction with ultrasound imaging. Muscle & Nerve
2003;27:682–92.
Ito M, Kawakami Y, Ichinose Y, Fukashiro S, Fukunaga T.
Nonisometric behavior of fascicles during isometric contractions
of human muscle. Journal of Applied Physiology 1998;85:1230–5.
Kremkau FW. Diagnostic ultrasound: principle, instruments, and
exercise. Philadelphia: W.B. Saunders Company; 1989.
Kristjansson E. Reliability of ultrasonography for the cervical multi-
fidus muscle in asymptomatic and symptomatic subjects. Manual
Therapy 2004;9:83–8.
Kristjansson E, Jonsson Jr. H. Is the sagittal configuration of the
cervical spine changed in women with chronic whiplash syndrome?
A comparative computer-assisted radiographic assessment. Journal
of Manipulative Physiological and Therapeutics 2002;25:550–5.
Kristjansson E, Leivseth G, Brinckmann P, Frobin W. Increased
sagittal plane segmental motion in the lower cervical spine in
women with chronic whiplash-associated disorders, grades I–II: a
case–control study using a new measurement protocol. Spine 2003;
28:2215–21.
McMeeken JM, Beith ID, Newham D J, Milligan P, Critchley DJ. The
relationship between EMG and change in thickness of transversus
abdominis. Clinical Biomechanics (Bristol., Avon.) 2004;19:
337–42.
Panjabi M, Abumi K, Duranceau J, Oxland T. Spinal stability and
intersegmental muscle forces. A biomechanical model. Spine
1989;14:194–9.
Platzer W. Locomotor system, color atlas of human anatomy, vol. 1.
New York: Thime Stuttgart; 2005.
Rainoldi A, Galardi G, Maderna L, Comi G, Lo Conte L, Merletti R.
Repeatability of surface EMG variables during voluntary isometric
contractions of the biceps brachii muscle. Journal of Electromyo-
graphy Kinesiology 1999;9:105–9.
Rezasoltan A, Kallinen M, Malkia E, Vihko V. Neck semispinalis
capitis muscle size in sitting and prone positions measured by real-
time ultrasonography. Clinical Rehabilitation 1998;12:36–44.
Rezasoltan A, Ylinen JJ, Vihko V. Isometric cervical extension force
and dimensions of semispinalis capitis muscle. Journal of
Rehabilitation Research and Development 2002;39:423–8.
Soltani AR, Kallinen M, Malkia E, Vihko V. Ultrasonogrpahy of the
neck splenius capitis muscle. Acta Radiologica 1996;37:647–50.
Stokes M, Rankin G, Newham DJ. Ultrasound imaging of lumbar
multifidus muscle: normal reference ranges for measurements and
 ARTICLE IN PRESS
294 J.-P. Lee et al. / Manual Therapy 12 (2007) 286–294

practical guidance on the technique. Manual Therapy 2005;10:
116–26.
Ueki J, DeBruin PF, Pride NB. In vivo assessment of diaphragm
contraction by ultrasound in normal subjects. Thorax 1995;50:
1157–61.
Westbrook C, Kaut C. MRI in practice. London: Oxford Blackwell
Scientific Publications; 1993.
Wu JP, Lee JP, Mao HY, Yu MH, Wang SF. Reliability of locating
cervical spinal levels by palpation and ultrasonography. Formosan
Journal of Physical Therapy 2005;30:80–7.

Diary of events
10th International Conference in Mechanical Diagnosis and
Therapy — The Evidence Mounts
23–25 March 2007
Queenstown, New Zealand
Honorary Chairman: Robin McKenzie
Presented by: The McKenzie Institute International
For more information visit: www.mckenziemdt.org
4th Low Back Pain Symposium
April 30–May 3, 2007
Target audience: colleagues of the following disciplines;
physical medicine and rehabilitation, orthopaedic surgery,
neurosurgery, physical-, occupational-, manual- and
Mensendieck therapy. Moreover, company doctors,
medical advisors of insurance companies and other
health care professionals interested in the topic of low
back pain.
Chairmen: Prof. Dr. Henk J. Stam, Prof. David Niv M.D.
FIPP, Prof. Dr. Mehmet Zileli
Accreditation applied for at the EACCME (Accreditation
Council) of the European Union of Medical Specialists
(UEMS)
More information and registration:
website: www.vitalmedbodrum.com
E-mail: vitalmed@vitalmedbodrum.com
6–10 July 2007, Singapore
Changing Pain and Movement Behaviour – A Classification
Based Approach to the Management of Chronic
Low Back Pain Disorder
by A/P Peter O’Sullivan for information on the workshop,
please e-mail marc@emmanuelphysio.com
1st World Congress on Manual Therapy
July 27th–August 1st, 2007
Mangalore, India
Organising Chair: Prof. Umasankar Mohanty,
President, Manual Therapy Foundation of Indias
For Registration and more information
Website: www.wcomt.com
E-mail: wcomt07@yahoo.co.in
295
Manual Therapy (2007) 12(3), 295
First international Fascia Research Congress
Basic Science and Implication for Conventional and
Complementary Health Care
4–5 October 2007
The Conference Center, Harvard Medical School Boston MA
http://www.fascia2007.com
5th International Course on the Hand
October 21–25, 2007
Target audience: colleagues of the following disciplines;
physical medicine and rehabilitation, plastic- and hand
surgery, physical- and occupational therapy and other
health care professionals, interested in the topic of the hand
Lectures include: Prof. Dr. S.E.R Hovius, Ton A.R.
Schreuders PT, PhD and G. Van Strein MSc
Accreditation applied for at the EACCME (Accreditation
Council) of the European Union of Medical Specialists
(UEMS)
More information and registration:
website: www.vitalmedbodrum.com
E-mail: vitalmed@vitalmedbodrum.com
2nd World Congress on Manual Therapy and Sport
Rehabilitation, The Spine II, in Roma Italy
6th–8th of March 2009
www.newmaster.it
Janet G. Travell, MD Seminar Series, Bethesda, USA
For information, contact: Myopain Seminars, 7830 Old
Georgetown Road, Suite C-15, Bethesda, MD 20814-2432,
USA.
Tel.: +1 301 656 0220; Fax: +1 301 654 0333;
website: www.painpoints.com/seminars.htm
E-mail: Calhoun@painpoints.com
If you wish to advertise a course/conference, please contact:
Karen Beeton, Associate Head of School (Professional
Development), School of Health and Emergency Professions,
University of Hertfordshire, College Lane, Hatfield, Herts
AL10 9AB, UK. There is no charge for this service.
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e1
Letter to the Editor
Letter to the Editor regarding a study titled ‘‘Diagnosis of
sacroiliac joint pain: Validity of individual provocation
tests and composite of tests’’
[Manual Therapy 10 (2005) 207–218]
To the Editors:
I always appreciate it when new evidence comes out
that may help clinicians detect when patients have
sacroiliac joint dysfunction. I am unsure however
whether this paper adds anything new to the existing
literature. The difficulty lies in the patient’s that were
selected for this study and their credibility as ‘‘trust-
worthy’’ respondents. The inherent problem of this
study is in choosing a patient population where
susceptibility bias is intrinsically possible. According to
Table 1 the mean duration of symptoms for this cohort
of patients was 42.1 months or over 3 years, the time off
work was nearly 18 months or one and a half years, and
1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.07.018
www.elsevier.com/locate/math
finally the Roland-Morris pain questionnaire’s mean
score was about 76%! This data suggests a distorted
assembly and one that I could hardly have faith in when
trying to determine if an 80% subjective improvement in
pain reduction occurred after sacroiliac joint injection.
The target or intended patient population of this study
should resemble characteristics of low back pain
patients frequently seen in the clinic and not one that
is reminiscent of a cohort of workman’s compensation
or litigating patients.
I realize that there is no ‘‘perfect’’ study and I applaud
the authors for trying to help improve the diagnostic
accuracy of detecting sacroiliac joint dysfunction.
Best Regards,
Michael T. Cibulka
Maryville University, Program in Physical Therapy,
St. Louis, MO 63105, USA
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e2
Book review
Neuromusculoskeletal Examination and Assessment.
Nicola J. Petty. Elsevier, Churchill Livingstone, Kidlington,
UK (2006) (413pp., price £27.99, ISBN: 044310204X)
The primary aim of this book is to provide a
systematic approach to the assessment of the neuro-
musculoskeletal system. As the third edition in the series
it covers the same scope as previous editions however
the layout is clearer with improved tables and figures
directing the reader to pertinent points. The book
devotes a chapter to each region of the neuromuscu-
loskeletal system with separate chapters on subjective
examination and physical examination. There is some
repetition of the subjective and physical assessments
section in each chapter however this only helps to
reiterate salient points and limits page turning.
This edition explores clinical reasoning in more depth
than its predecessor, both towards planning the physical
examination and for follow on treatment. The author
has also made subtle changes in this edition which helps
the book move away from it’s ‘‘joint’’ bias and embraces
the neuromuscular components of assessment and
reasoning. This is particularly evident in the clinical
doi:10.1016/j.math.2006.09.005
www.elsevier.com/locate/math
reasoning exercises. Throughout, references have been
updated however there is limited critique of the quality
of the evidence presented.
The physical examination section is an excellent
toolbox for undergraduates and practitioners with
concise instructions and updated photographs albeit
with recognition that the tests need to be flexibly
implemented to suit both patient and therapist. Com-
ment on the validity and reliability of testing procedures
would have been useful although it is helpfully
explained, particularly for undergraduates, that no one
test can be truly diagnostic for any single pathologic
entity.
Overall this is a very practically based book which
guides the reader through the assessment process with
useful prompts to assist clinical reasoning. It is therefore
a valuable resource for undergraduates, new graduates
and also as a reference text for clinicians working in any
neuromusculoskeletal setting.
Caroline Miller
University Hospital Birmingham Foundation Trust, UK
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e3
Book Review
Maintaining Body Balance Flexibility and Stability.
Chaitow L. Elsevier-Churchill Livingstone, Kidlington,
UK (2004) (194pp., price £21.99, ISBN 0443073511).
Due to daily load our body physically adapts to
certain situations. Sometimes the body adapts well,
sometimes poorly. Leon Chaitow shows us that through
an application of simple techniques we can learn a better
way of physical adaptation. The book is written for
patients as well as practitioner or therapist. It clearly
says in the preface that it is not a substitute for
professional attention and treatment, but can be used as
support and guidance material. It contains more than
180 pages of well-illustrated techniques and exercises
with separate exercise sheets inside the back cover. The
author describes a series of exercises to influence the soft
tissues. The basic idea is that, in the case of articular
problems, attention should first be given to the soft
tissues, in order to normalize joint function. It is the soft
tissues that support and move the joint. The author,
being a registered osteopathic practitioner himself, uses
osteopathy as the primary source of many of the
doi:10.1016/j.math.2006.09.004
www.elsevier.com/locate/math
methods described. According to Chaitow the book
also relies on work of medical physicians, physiothera-
pists, exercise physiologists, chiropractors, massage
therapists and others, however no references are given
throughout the text. The first chapter is the most
theoretical of all. Different forms of muscle energy
technique (MET), like reciprocal inhibition and post-
isometric relaxation, are explained. The other chapters
describe methods on testing muscle length, treatment
and self-treatment MET, trigger points, flexibility,
stability, balance, positional release technique and strain
and counter strain. The book is written in a way that
every patient can understand with clear step-by-step
instructions on each exercise method. For the therapist
it is a work which gives a lot of practical information
and inspiration for treatment.
Ulrike Van Daele
Higher Institute of Physiotherapy,
Hogeschool Antwerpen, Belgium
Faculty of Physical Education and Physiotherapy,
Vrije Universiteit Brussel, Brussel, Belgium
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e4
Book Review
Meaningful Motion: Biomechanics for Occupational
Therapists. 1st ed., S.J. Spaulding, Elsevier, Amsterdam
(2005). 187pp., ISBN:0443074399.
This book has been specifically written for members
of the occupational therapy community, as mentioned in
the preface, however, because understanding motion is
one of the cornerstones of our profession too, this book
could be useful for students, teachers and clinicians in
the field of physical therapy/manual therapy.
Spaulding summarizes in the first section of her book
the recent and seminal research findings in the area of
biomechanics, motor control and learning. The second
section of the book gives the integration of the knowl-
edge from the first section into the specific areas in
which occupational therapists work: balance training,
the environment, ergonomics and leisure. Perhaps the
most interesting parts of the book are the comprehen-
sive case studies, one in each chapter. The end of each
chapter includes questions about the case and a list of
questions in the form of laboratory exercises. This gives
doi:10.1016/j.math.2007.01.002
www.elsevier.com/locate/math
the reader the opportunity to think further about the
material in the chapter and will help to apply the
principles of biomechanics and motor control to
practice. Despite finding the case studies invaluable, as
a musculoskeletal physical therapist/manual therapist,
this reviewer would love to see more patient cases with
musculoskeletal disorders. Most cases describe patients
with neurological disorders, which is perfectly normal
for occupational health practitioners.
In summary, this is a well-written book that deserves
to be read, not only by occupational therapy students
and practicing occupational therapists. This useful
workbook could also broaden the understanding and
advancing the practice of physical therapy students and
practicing physical therapists/manual therapists.
Simon Brumagne
Faculty of Kinesiology and Rehabilitation Sciences,
University of Leuven, Leuven, Belgium
E-mail address: Simon.Brumagne@faber.kuleuven.be
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e5
Book Review
Clinical Application of Neuromuscular Techniques. Prac-
tical Case Study Exercises, L. Chaitow, J. DeLany.
Elsevier/Churchill Livingstone, New York, NY (2005).
(289pp., £21,99), ISBN: 0443100004.
Clinical Application of Neuromuscular Techniques,
Practical Case Study Exercises by Leon Chaitow and
Judith DeLany provides a collection of 34 patient cases
authored by 19 different ‘‘Neuromuscular Therapists’’
presenting a wide range of clinical presentations. Cases
are systematically presented under the headings of
Profile, Health or Family History, Presenting Com-
plaint(s), Significant Contributing Factors, Clinical
Evidence and Previous Treatment, Questions for Read-
er, Examination, Clinical Impression and Treatment or
Action Suggested, and Key Points with some cases also
including sections on Further Reading, References and
Websites Worth a Visit. Interspersed through most cases
are additional boxes with supportive academic content
pertaining to aspects of the case and very good red flag
boxes highlighting clinical features that warrant im-
mediate concern and medical consultation.
The intention of this book is to assist readers’
integration of neuromuscular techniques and comple-
mentary medicine (e.g. diet) into their practice. This is
achieved through inclusion of posture, lifestyle, biome-
chanical and psychosocial considerations through all
cases with corresponding management suggestions.
Consistent with the book’s focus, physical impairment
of muscle, typically in the form of trigger points, but
also with respect to posture, habits of movement, muscle
length and to some respect motor control, are thor-
oughly considered. While joint impairment is included,
this is considerably more superficial in both assessment
doi:10.1016/j.math.2007.01.003
www.elsevier.com/locate/math
and management with a notable lack of contemporary
content and referencing. The case authors’ clinical
reasoning is essentially a summary of management
suggestions, not explicitly linked to supporting or
negating evidence, with no discussion of recommended
outcome measures to monitor or thoughts on progres-
sion of treatment. While listing of management options
in this manner is very useful, case reasoning could be
developed and articulated better. An interesting, but
also frustrating feature of this book is its extensive use of
highlighting key words in red throughout its cases.
Highlighted words can then be looked up in the index of
the two companion books ‘‘Clinical Application of
Neuromuscular Techniques, Volumes 1 and 2’’ for
definition and explanation. Many of the words high-
lighted refer to assessment or management procedures
that are somewhat specialised to ‘‘Neuromuscular
Therapists’’ and their philosophy of practice, making
the cases less accessible to manual therapists without
this background. Musculoskeletal physiotherapy does
not feature very positively, with the majority of the cases
either highlighting the poor results of previous phy-
siotherapy management received or making recommen-
dations for osteopathic, chiropractic but not
physiotherapy referral. As such, I anticipate this book
is likely to be of greatest interest to ‘‘Neuromuscular
Therapists’’.
Mark Jones
Postgraduate Coursework Masters
Programs in Physiotherapy,
School of Health Sciences,
University of South Australia, Australia
E-mail address: mark.jones@unisa.edu.au
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e6
Book review
Rehabilitation of the Spine. A Practitioners Manual, 2nd
ed. Liebenson C. Elsevier/Churchill Livingstone, Amster-
dam/New York (2007) (972pp). ISBN:0781729971.
With over 900 pages and 39 chapters, the size of this
text is a reminder of the diversity of opinion, and
methods of management, of spinal pain disorders. The
author, Craig Liebenson has written or co-authored 14
chapters, and the remainder have been written by
invited contributors. Much of the text is focused on
the lumbar spine with relatively less attention given to
cervical spine pain. Surprisingly, not one chapter is
devoted to disorders of the thoracic spine and their
management. The introductory chapters provide an
overview of some fundamental concepts in spinal pain
including the anatomical sources of pain, mechanisms of
injury, pain physiology and the biopsychosocial model.
These chapters are well written and supported by
relevant literature.
The chapters on assessment were somewhat disap-
pointing as the focus was on the diagnostic triage of
spinal pain, including screening for psychosocial ‘yellow
flags’. The only chapter on physical examination from a
physical therapy perspective was in relation to muscle
imbalance, and the remaining 3 chapters in this section
reviewed physical performance testing. What appears to
be missing here is an overview of the clinical examina-
tion process for patients with spinal pain, and the
clinical reasoning process, which would guide the
interpretation of the information obtained. The absence
of this information makes the application of the
following chapters on physical treatment difficult to
interpret. This would be particularly the case for
students of the physical therapy professions, or clin-
icians looking for a stronger framework to support their
approach to patient management.
In the next section of the text there are 9 chapters
which describe ‘acute care management’ which include
treatment approaches such as massage, muscle release
doi:10.1016/j.math.2007.01.004
www.elsevier.com/locate/math
techniques, rehabilitation of breathing patterns, the
McKenzie method, neural mobilisation and joint
manipulation. Specifically how these treatment methods
relate to the management of acute pain is not stated, and
it seems that a number would be equally or better suited
to the management of non-acute conditions. The
following 10 chapters on ‘recovery care management’
mostly present various approaches to trunk muscle
strengthening around the ‘stabilisation’ theme. While
there is a lot of information presented and some useful
ideas for exercise prescription, these chapters are some-
what repetitive, and again, in many cases do not provide
a good framework in relation to the types of patient
presentation where these approaches may be most
useful. Equally disappointing were the chapters on
‘integrated management’ which surprisingly presented
similar material to the preceding chapters, and with case
studies linked to anatomical diagnoses such as ‘facet
syndrome’ and ‘discogenic radiculopathy’. This ap-
proach was clearly inconsistent with the introductory
chapters which highlight the problems with anatomical
diagnoses, and emphasise a broader approach to the
evaluation of ‘non-specific’ spinal pain.
For clinicians with a sound approach to clinical
reasoning and treatment prescription, this text may
provide a range of new ideas or treatment techniques
which could be integrated into patient management.
However, for the inexperienced clinician or manual
therapy student the range of treatment approaches
presented may be somewhat overwhelming, and of less
value in relation to the common problem of deciding
which approach to treatment is likely to be most
useful.
Steve Edmondston
School of Physiotherapy,
Curtin University of Technology, Perth, Australia
E-mail address: S.Edmondston@curtin.edu.au
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e7
Book review
Physiotherapeutic Management of Lumbar Spine Pathol-
ogy, MacDonald David, Jemmett Rick. first ed. Novont
Health (2005). (199pp.), ISBN: 9024234344.
I was pleased to be asked to review this book as I am
always on the lookout for books that underpin and
extend my clinical practice using an up to date evidence
base. However, the title was a little misleading as rather
than a concise text on spinal pathology, it focused on the
area of segmental dysfunction. Nevertheless, I found it
an easy book to read and as the authors’ state, it is a
‘‘useful reference tool for evidence-based management
of patients with segmental dysfunction’’.
The first three chapters deal with an overview of
relevant anatomy, biomechanics and spinal pathology
including motor control dysfunction with an up to date
reference list to back them up. The subsequent chapters
deal with the assessment and management of segmental
dysfunction with the authors proposing their own model
doi:10.1016/j.math.2007.01.005
www.elsevier.com/locate/math
of segmental dysfunction. These chapters are reinforced
by clear photos to demonstrate points and techniques,
although there is less research evidence presented in this
part of the book to support the techniques. Chapter
seven presents case studies, which are of benefit to
clinicians and students new to this area, although there
is a lack of objective measurement tools used through-
out this section. Chapter 8 consists of invited peer
review commentaries, which is uncommon in texts, but
provides a balanced view of the text as a whole.
Overall, the authors have summarised lumbar dys-
function and present a useful reference for clinicians
and/or students in clinical practice.
Lyndsay Alexander
School of Health Sciences, Robert Gordon University,
Aberdeen, UK
E-mail address: l.a.alexander@rgu.ac.uk.
 ARTICLE IN PRESS
Manual Therapy 12 (2007) e8
Book review
Principles of Neuromusculoskeletal Treatment and Man-
agement: A Guide for Therapist, N.J. Petty. Churchill
Livingstone, New York, NY (2004). h47.99, (368pp.),
ISBN: 0443070628.
The aim of this book is to make explicit the
underlying principles behind treatment and manage-
ment of patients with neuromusculoskeletal disorders.
In this extensive book, the author has succeeded in her
aim. It is well written, easy to follow and it gives a lot of
information about the state of the art in assessment and
treatment.
This book is divided into nine chapters in which items
as assessment, function and dysfunction of joint, muscle
and nerve, principles of treatment strategies and finally a
chapter on principles of patient management by A.
Moore. Each chapter is completed with an extensive
reference list.
Although most of the content is used within curricula
of education in physical therapy and or manipulative
therapy, this book brings all the knowledge together,
with clear figures and illustrations. The content is based
doi:10.1016/j.math.2007.01.006
www.elsevier.com/locate/math
on fundamental or basic research, and on that basis it
tries to add knowledge to an evidence-based approach
of patient management. Unfortunately, it does not
answer the question whether a chosen approach in
treatment of muscle, joint or nerve is really effective.
The author states that ‘the best treatment is the one that
improves the patient’s sign and symptoms in the shortest
period of time.’ The reader will be curious about
effectiveness and efficacy of chosen treatment strategies,
but that was a priori, not the primary aim. However, in
evidence-based management, this knowledge could
make the clinical reasoning process more clear.
This book is highly recommended for physical
therapists and manipulative therapists as a reference
book and for students as a textbook.
Jan M. Pool
EMGO Institute, VU University Medical Center,
Amsterdam, The Netherlands
E-mail address: j.pool@vumc.nl