You are on page 1of 19

Physical Therapy Reviews 2006; 11: 125–142

EXERCISE AND MANUAL THERAPY FOR THE


TREATMENT OF IMPINGEMENT SYNDROME OF THE
SHOULDER: A SYSTEMATIC REVIEW

ATHANASIOS TRAMPAS1 AND ATHANASIOS KITSIOS2

1
Thessaloniki, Greece
2
Department of Physical Education and Sports Science, Aristotelio
University of Thessaloniki, Thessaloniki, Greece

‘Impingement syndrome’ is a generic term for rotator cuff lesions encompassing all stages of
tendon disease. The fundamental aim of this review was to assess whether the quality of
randomised controlled trials (RCTs) on this specific topic has improved during the last 3 years.
A computer-aided search of databases was carried out from 2003 to 2005, using a combination
of key words. Five RCTs were critically appraised using the PEDro scale and the American
College of Sports and Medicine guidelines. The mean quality score of the included trials was
6.2, ranging from 4 to 8 points out of a possible 10. There is moderate evidence that supports
the use of therapeutic exercises alone in reducing pain and function, whereas there is limited
evidence to support the effectiveness of both manual and exercise therapy in combination.
Future studies should better define impingement syndrome, based upon well-designed
research methodology and the use of validated measurement tools.

Keywords: Subacromial impingement syndrome, therapeutic exercises, manual therapy,


randomised controlled trials (RCTs)

INTRODUCTION individuals younger than 25 years but may occur at


any age in those who are engaged in excessive use of
‘Impingement syndrome’ is a generic term for rotator upper limbs. The pathological changes are oedema
cuff lesions encompassing all stages of tendon disease and haemorrhage of the subacromial bursa. Stage 2 is
from early inflammation through degeneration and characterised by thickening and fibrosis; in stage 3,
eventually partial or complete tears. Anatomical rotator cuff tears, biceps’ ruptures, and bone changes
anomalies, particularly type III hook-shaped are observed. If impingement continues, the condi-
acromions or congenital subacromial stenosis, may tion becomes chronic, producing thickening and
predispose to impingement syndrome.1 Acromial fibrosis of the bursa and tendinitis of the cuff. Stage 2
morphology appears to have a predictive value in disease is commonly seen in patients between 25–40
determining the success of conservative measures and years of age. Stage 3 results from further impinge-
the need for surgery in patients with impingement ment, producing degeneration or complete or incom-
syndrome.2 The rotator cuff, especially supraspinatus, plete tears of the rotator cuff. These more advanced
maintains the subacromial space by depressing the changes are usually seen in patients over 40 years old.
head of the humerus to prevent superior translation Although some clinical criteria for impingement
during abduction and elevation movements.3 syndrome have been proposed,5 none has received the
Neer4 has classified three progressive pathological level of acceptance required to make it useful as a
stages of impingement: stage 1 usually occurs in study selection criterion for the present review. This
© W. S. Maney & Son Ltd 2006 DOI 10.1179/108331906X99065
126 TRAMPAS AND KITSIOS

was one of the main outcomes of the last literature direct effect on the quality of life for the chronic pain
review conducted by Desmeules et al.6 for impinge- patient by means of operant conditioning or of pain
ment syndrome treated with exercises and manual behaviours, that may be increased or decreased.
therapy. The authors concluded that future trials Therapeutic exercise is the most extensively investi-
should better define shoulder impingement syn- gated form of rehabilitation, and may be defined as a
drome, presenting the inclusion and exclusion criteria series of specific movements for the purpose of train-
of the included studies in detail. It was, therefore, ing or developing the body through systemic practice,
decided that it would be appropriate for this review to or as a bodily exertion for the promotion of physical
assess studies dealing with any of the impingement- health.13 Therapeutic exercise encompasses many
related diagnoses (e.g. tendinitis or bursitis), and to well-known exercise strategies such as rehabilitating
compare the inclusion and exclusion criteria used to the functional demands of the muscle system, enhanc-
the findings of previous recent systematic reviews6,7 as ing cardiovascular fitness, or improving joint and
a criterion to assess whether the degree to which the muscle flexibility.14 The therapeutic exercise pro-
quality of definition for impingement syndrome has grammes have generally consisted of stretching the
improved or not during the last 2.5 years to provide anterior and posterior shoulder girdle, muscle relax-
more homogeneous study populations. ation techniques, motor learning to normalise dys-
Van der Windt et al.8 investigated the incidence and functional patterns of motion, and strengthening the
management of intrinsic shoulder disorders in Dutch rotator cuff and scapular muscles. Improvements in
general practice, and evaluated which patient charac- pain, patient satisfaction, levels of disability and func-
teristics were associated with specific diagnostic cate- tional loss, strength, shoulder range of motion, pain
gories. In 11 general practices (35,150 registered with subacromial compression, and overall shoulder
patients), all consultations concerning shoulder com- use have been demonstrated with therapeutic exercise
plaints were registered during a period of 1 year. programmes.7
Patients with an intrinsic shoulder disorder who had There is a small number of literature reviews evalu-
not consulted their general practitioner for the com- ating the effectiveness of exercise and manual therapy
plaint during the preceding year (incident cases) were for impingement syndrome. Van der Heijden15 con-
asked to participate in an observational study. cluded that knowledge of the effect of all methods in
Participants completed a questionnaire regarding the the field of physiotherapy on long-term outcome or
nature and severity of their complaints. The general recurrence of shoulder disorder is lacking. Based on
practitioners recorded data on diagnosis and therapy. three of the studies listed in the recent review by
The cumulative incidence of shoulder complaints in Desmeules et al.,6 van der Heijden15 also concluded
general practice was estimated to be 11.2/1000 that very limited evidence supports the effectiveness
patients/year. Rotator cuff tendinitis was the most fre- of therapeutic exercises or manual therapy. An earlier
quently recorded disorder and it was found to be review by van der Heijden et al.16 reached similar con-
around 3.2–4.2 per 1000 persons/year (29%). clusions. According to Philadelphia Panel evidence-
Maitland indicated that the word ‘manipulation’ can based clinical practice guidelines,17 the investigators
be used in two ways. First, as a general term to refer to were unable to make any conclusions regarding the
all passive movements used in the treatment of muscu- effectiveness of therapeutic exercises for shoulder
loskeletal disorders including ‘manipulation technique’. pain because poorly defined inclusion criteria and
Maitland then went on to define the term ‘mobilisa- non-validated outcomes were used. Only one study18
tion’, providing a more detailed and specific definition related to therapeutic exercises was included and
than that provided earlier by Kaltenborn.9 Manual reviewed. In both systematic reviews by van der
therapy requires that the therapist apply an external Heijden,15,16 there was an attempt to evaluate physio-
force to the patient to produce a desirable amount of therapy interventions on shoulder pain or non-specific
deformation of the targeted joint connective tissues, shoulder pathology but with inconsistent findings.
and a relative displacement of the bones at either end of Therefore, it is strongly believed that future reviews
the joint in accordance with the load–displacement on the effectiveness of physiotherapy should focus on
relationship of the targeted tissues. According to Le-T specific treatment methods and properly defined clin-
et al.10 manipulation techniques are mainly used for ical syndromes; otherwise, the results have limited use
adhesive capsulitis (under anaesthesia) and, for patients for clinical practice, because without specific diagnoses
with impingement syndrome grade I manual therapy there is insufficient guidance for the development of
must be the primary choice for management.11 treatment programmes.
It has long been recognised that exercise can have a Following the same pattern, Green et al.19 reviewed
beneficial effect in treating pain. Fordyce et al.12 common physiotherapy interventions for shoulder
showed that the benefits of exercise could have a pain. The authors tried to sub group trials into type
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 127

of disorder for analysis and the methods sections of tions that can be distinguished from it, such as the
all identified trials were reviewed independently ‘frozen shoulder’ or neurological conditions. Also,
according to predetermined criteria by two reviewers. further validation of outcome measures applied to
When the two reviewers disagreed, discussion was patients with shoulder impingement syndrome is nec-
facilitated in order to reach consensus; if this failed, essary. Even though there is no gold standard for
the trial was sent to a third reviewer for arbitration, as evaluating shoulder disorders such as shoulder
suggested by the Cochrane Collaboration guidelines. impingement syndrome, the use of validated func-
Twelve trials used exercise and/or manual therapy tional outcomes is necessary to evaluate fully the
(mobilisation, manipulation) as therapeutic modali- patient’s condition and progression. The use of pain,
ties. Trial populations were generally small, with range of motion, and strength as secondary outcome
many trials underpowered to demonstrate a differ- measures seems to be acceptable, and their basic con-
ence between groups if one was present; hence, the struct validity for the evaluation of a musculoskeletal
results may be biased by Type II error (the failure to disorder is acceptable. However, there is a need to val-
demonstrate a difference which is in truth present, or idate the qualities of these outcome measures fully.
false negatives). There was substantial clinical hetero- The most recent systematic review on impingement
geneity with respect to the interventions tested; hence, syndrome was conducted by Michener et al.7 The pur-
few trials could be combined in meta-analysis to pose of this systematic review was to examine the evi-
reach an overall conclusion about the effect of physio- dence for rehabilitation strategies for patients with
therapy interventions for shoulder disorders. A super- shoulder impingement syndrome – specifically, the
vised exercise regimen has been demonstrated to be of effectiveness of non-surgical, non-pharmacological
significant benefit in both the short and longer term. treatment procedures. Two examiners used a quality
Due to differences in outcome scales, and use of a checklist developed according to the guidelines of
combination of change scores and final values, results Sackett et al.22 Agreement between the two examiners
for pain could not be combined; however, the two tri- was assessed to determine the presence of a discrepancy
als conducted in participants with rotator cuff disease of greater than one quality point on any single quality
both demonstrated a significant difference in reduc- checklist item. If a discrepancy existed, the single item
tion of pain at 3–4 weeks for the exercise plus mobili- was discussed to reach consensus. Therapeutic exercise
sation group, over the group performing exercise was indicated as an effective intervention for patients
alone.20,21 The difference between groups with respect with shoulder impingement syndrome as opposed to no
to range of motion, strength and function are based treatment or placebo treatment. Furthermore, manual
on only one of the three trials, but demonstrate bene- therapy combined with therapeutic exercise appeared to
fit of adding mobilisation to exercise. provide better outcomes than therapeutic exercise
Another published systematic review was con- alone, but the evidence was unclear as to the specific
ducted by Desmeules et al.6 Studies were included if joint mobilisation techniques that may provide benefit.
they were randomised, controlled trials, related to The authors also investigated the use of acupuncture,
impingement syndrome, rotator cuff tendinitis, or laser therapy and ultrasound, and concluded that the
bursitis, and one of the treatments included therapeu- evidence was limited and that the results of the studies
tic exercise or manual therapy. Two independent were conflicting when compared to exercise or manual
observers reviewed the methodological quality of the therapy. Therefore, it is considered appropriate to con-
studies using an assessment tool developed by the duct systematic reviews focused on specific treatment
Cochrane Musculoskeletal Injuries Group, but this methods with the aim of developing treatment guide-
tool has not been yet formally validated. There was lines and the aforementioned combinations would also
limited evidence to support the effectiveness of man- be an objective for the present review.
ual therapy and therapeutic exercise to treat the Despite the fact there is a positive physiological
impingement syndrome. Some of the included stud- effect of therapeutic exercise and orthopaedic manual
ies: (i) did not clarify the existence of shoulder therapy, there is insufficient clinical information on
impingement syndrome in the sample; (ii) contained a the effectiveness of these interventions for shoulder
proportion of patients without impingement syn- pain and, more specifically, for shoulder impingement
drome; or (iii) used a sample of patients who had syndrome. These results concur with recent system-
undergone subacromial decompression surgery. More atic reviews on physical rehabilitation interventions
methodologically sound studies are needed to evalu- for painful shoulders.6,7,15,17 The reasons for equivocal
ate further these interventions, since the quality of the results are varied; authors reporting positive findings
included studies was low to very low. Future trials have been criticised for not employing proper controls
need to define shoulder impingement syndrome bet- or for the absence of an appropriate statistical analy-
ter, and to exclude the main clinical shoulder condi- sis. In contrast, literature producing negative results
128 TRAMPAS AND KITSIOS

has been rebutted by proponents of manual therapy 1. The study was concerned with the effectiveness of
on the basis of improper patient selection or inade- exercise therapy and manual therapy (as defined) and
quate skill on the part of the therapist. in the treatment of impingement syndrome only.
Although the shoulder pain literature is extensive, there
2. Subjects should have been diagnosed with impinge-
is an urgent need to conduct well-designed studies. Special
ment-related pain; studies with shoulder pain as the
attention to the characteristics of the therapeutic applica-
only inclusion criterion that stated a significant
tion is needed in the field of rehabilitation. For example,
number of subjects had shoulder impingement
the types of exercise used, adequate exercise intensity, and
syndrome were included.
progression, need to be clarified according to patient-spe-
cific classification of physical dysfunction, needs and 3. At least one of the treatments was therapeutic exercise
treatment goals.17 The effectiveness of manual therapy is or manual therapy supervised or given by a
influenced by the types of manoeuvres used, the approach physiotherapist.
adopted, years of experience of the therapist, number and
4. Other interventions typical of physiotherapy (thermo-
size of the muscles involved, the patient position used,
therapy, cryotherapy, electrotherapy, massage, trans-
pressure exerted, rhythm and progression, and frequency
verse friction), as well as medical interventions (e.g.
and duration of the treatment sessions.23
arthroscopic subacromial decompression) could have
Due to the wide clinical importance of impingement
been used, but therapeutic exercises and/or manual
syndrome incidence, and the extensive use of manual
therapy had to be identified as the main treatment.
therapy techniques and exercises for its treatment, it
was felt necessary by the authors to conduct a system- 5. Outcome measures included not only pain, strength
atic review on the effects of these interventions on and range of movement (ROM), but also functional
shoulder impingement syndrome. Furthermore, there is tests, functional questionnaires and self-perceived
no systematic review identified in the literature focused change in symptoms. These outcome measures have
specifically on the effectiveness of manual and exercise been evaluated and used in previous studies.20,21,25
therapy over the past 2.5 years. This paper provides an
6. Each article was published in the English language.
update on the review by Desmeules et al.6 in order to
Since the number of journals published in languages
assess whether the quality of the most recent studies has
other than English indexed in electronic databases
improved and, if appropriate, to revise recommenda-
such as MEDLINE and EMBASE is limited, there is
tions on evidence-based practice in this area.
a small possibility that relevant trials might have been
Although RCTs are considered to be the best design
missed.
for control of validity and precision, flaws in their
design and conduct can result in overestimation or 7. Only RCTs in full text were included.
underestimation of treatment effects, and conse-
8. Time limits were included concerning the year of pub-
quently can lead to false positive or false negative
lication: from 2003 up to date, since the last review on
results and conclusions.24 To minimise such problems,
this topic was published in 2003.
this review placed special attention on the quality of
the methods of the selected studies.
Diagnostic criteria
The following diagnostic criteria were identified, which
METHODOLOGY
were relevant to the clinical features of shoulder
impingement syndrome:
The main objectives of this paper were to: (i) deter-
mine if manual therapy and therapeutic exercises are 1. Age: 18–66 years old, working age.
effective approaches for the treatment of the impinge-
2. Symptom duration: more than 1 month.
ment syndrome of the shoulder; (ii) determine the
characteristics of the therapeutic intervention used in 3. Active or passive range of motion: often normal,
the field of rehabilitation (e.g. the types of interven- with the exception of movements that stretch the
tion used, adequate intensity and progression) in posterior glenohumeral capsule, which is
order to address specific protocols; and (iii) investi- sometimes limited.
gate how treatment outcomes.
4. Resisted tests: positive resisted tests of the rotator-
cuff muscles.
Criteria for selecting studies for this review
5. Painful arc: Jobe test26/Speed test.27
Clinical studies were selected for inclusion if they met
6. Impingement test: Lidocaine injection/Neer test.4
the following criteria:
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 129

7. Other tests: Yocum test.27 techniques might be used individually, in tandem, or in


combination with other treatment agents in what is called
8. Impingement signs: (i) passive overpressure at full
multimode care. For example, mobilisation plus exercise
shoulder flexion with scapula stabilised; (ii) passive
can be used in combination with heat therapy and ultra-
internal rotation at 90° flexion; (iii) Neer sign;
sound. All studies used comparison groups that could be
(iv) passive abduction 90° and horizontal adduction–
either a control group (placebo control, active control, or
abduction; (v) forced elevation and internal rotation;
no-treatment control) or various other treatment groups.
and (vi) Kennedy-Hawkins sign.28
The control interventions were placebo mobilisation,
9. Other signs: (i) resistant to physiotherapy; (ii) resist- exercise, other therapeutic interventions, or no treatment.
ant to NSAIDs and steroids; (iii) pain at the Studies describing manipulation while the patient
shoulder; (iv) primary impingement syndrome; (v) was under anaesthesia, or manual therapies that were
limited functional movement patterns; (vi) ‘impinge- not directed at influencing joint function (e.g. trac-
ment pain’; (vii) no or disappointing results of > 6 tion) were excluded, as were studies using interven-
months of conservative treatment; (viii) shoulder tions in a field other than physiotherapy, except if
complaints of any source; (ix) shoulder pain therapeutic exercise or manual therapy were one of
reproduced by active shoulder movements; (x) pain the main treatment applications.
at rest for more than 1 year; and (xi) ‘isolated
shoulder disease’.
Types of outcome measure
These diagnostic criteria, as well as the exclusion criteria
below, were selected based on the two most recent
As outlined by van der Heijden,15 there is no gold stan-
published systematic reviews.6,7 According to the
dard that provides a valid and reliable estimate for clini-
literature, the most reliable clinical test for impingement
cally relevant changes in patients with shoulder disorders.
syndrome was found to be the Kennedy-Hawkins test,28
For this reason, RCTs were considered as having relevant
although none of the previous diagnostic methods is
outcome measures if they included the following most
adequately validated to define this pathology.
important outcomes: pain, strength/endurance, func-
tional loss or disability. Disability was found to be a sig-
Criteria for excluding studies from this review nificant outcome for shoulder disorders,29 and so was
used for this systematic review. Physiological outcomes
Instability, rotator-cuff tears or rupture, adhesive cap- of physical examination (e.g. range of motion and mus-
sulitis, acromioclavicular joint pathology, cervical cle flexibility), generic health status and other symptoms
pathology (e.g. herniated disc), radiological/imaging such as medication use and side-effects were consid-
findings, workmen’s compensation claim were the ered secondary outcomes. They were also extracted
main exclusion criteria for this review. Other exclusion since their correlation with the clinical status of the
criteria specified were: subjects with shoulder pain with patient is controversial but not poor.30 There were no
serious medical problems (e.g. advanced cancer, heart restrictions on the type of tool used in the studies to
failure), definable neurological abnormalities (e.g. measure these outcomes, as there were no universally
peripheral neuropathy, hemiplegia), disorders with bony accepted tools available; however, a number of studies
lesions (e.g. osteoporosis, fracture), and significant men- used validated tools.
tal disorders (e.g. psychosis, mania, major depression). Although data on all reported outcomes were
Finally, subjects with rheumatic disease or absence of extracted, pain relief and strength were the primary out-
pain aggravation on active, functional movement tests (i.e. comes commonly available across most of the reports
indicating non-organ symptoms) were also excluded. and these had to be explicitly patient-oriented. Details
Non-randomised, controlled, clinical trials and of the timing of the outcome assessments as well as data
controlled before and after studies, studies in abstract on side-effects of treatments were recorded.
form, articles on technique descriptions, diagnostic
tests, outcome reliability, or other topics that did not
describe manual and exercise therapy of the shoulder Methodological quality
as an intervention were excluded.
The best way of selecting high-quality physical ther-
apy trials for a systematic review has not yet been
Types of intervention determined. Most of the available scales for assessing
the validity of randomised, controlled trials derive a
Studies using manual techniques and therapeutic exer- summary score by adding the scores (with or without
cises were all included in the review. In the studies, these differential weights) for each item. While this
130 TRAMPAS AND KITSIOS

Table 1. Evaluation of the validity of physiotherapy interventions according to the PEDro scale

1 Eligibility criteria were specified


2 Subjects were randomly allocated to groups (in a crossover study, subjects were randomly allocated an order in which
treatments were received)
3 Allocation was concealed
4 The groups were similar at baseline regarding the most important prognostic indicators
5 There was blinding of all subjects
6 There was blinding of all therapists who administered the therapy
7 There was blinding of all assessors who measured at least one key outcome
8 Measures of at least one key outcome were obtained from more than 85% of the subjects initially allocated to groups
9 Subjects for whom outcome measures were available received the treatment or control condition as allocated or,
where this was not the case, data for at least one key outcome were analysed by ‘intention to treat’
10 The results of between-group statistical comparisons are reported for at least one key outcome
11 The study provides both point measures and measures of variability for at least one key outcome

Slightly more detail on some of these items is provided in the PEDro tutorial and can be assessed at
<http://www.pedro.fhs.usyd.edu.au/scale_item.html#scale_2>.

approach offers appealing simplicity, it is not sup- have face validity36 but are yet to be validated by other
ported by empirical evidence.31,32 Notably, scales with means. According to Maher et al.37 the reliability of
multiple items and complex scoring systems take some PEDro scale items was only ‘fair’ or ‘moderate’,
more time to complete than simple approaches. They but can be improved when consensus ratings were used.
have not been shown to provide more reliable assess- Furthermore, the precision of the PEDro scale also
ments of validity;33 they may carry a greater risk of should be considered by users of the PEDro database.
confusing the quality of reporting with the validity of None of the scale items had perfect reliability for the
the study; they are more likely to include criteria that consensus ratings (consensus ratings are displayed on
do not directly measure internal validity; and they are the PEDro database); thus, users need to understand
less likely to be transparent to users of the review. For that the PEDro scores contain some error. Although it is
these reasons, it is preferable to use simple approaches not possible to blind patients to exercise and manual
for assessing validity that can be fully reported (i.e. therapy, the authors chose to award a point to trials
how each trial scored on each criterion). clearly stating that patients were given minimal infor-
A simple approach to evaluate the validity of phys- mation about the differences between interventions,
iotherapy interventions is the PEDro scale (Table 1). and this was the only adjustment made in accordance
The scale is called the PEDro scale because it was ini- with the methods used by Busch et al.,38 who reviewed
tially developed to rate quality of RCTs on the exercise training in patients with fibromyalgia.
Physiotherapy Evidence Database, and is considered In order to take a more comprehensive look at the
to assess physiotherapy interventions adequately. methodological quality, the trials were categorised
Therefore, these criteria are also considered suitable according to the following criteria:
for this review, because in the assessment of method-
1. High quality trials. At least 6 of the methodological
ological quality of intervention research for shoulder
criteria met plus internal validity of ≥4/7.
impingement pain, they determine the shortcomings
in the design and reporting of trials that evaluated 2. Moderate quality trials. At least 5 of the method-
common interventions. ological criteria met plus internal validity of ≥3/7.
The PEDro scale is an 11-item scale designed for rat-
3. Low quality trials. Five or less of the methodological
ing methodological quality of RCTs. Each satisfied item
criteria met plus internal validity of < 3/7.
(except for item 1, which, unlike other scale items, per-
tains to external validity) contributes one point to the Items 2, 3, 5, 6, 7, 8, 9 assess internal validity; items 1
total PEDro score (range, 0–10 points). The scale has and 4 assess descriptive validity; items 10 and 11
been used to rate the quality of over 3000 RCTs in the assess statistical validity. Each trial was assessed
PEDro database.34 The scale is based on the Delphi con- independently by two examiners (AT and AK). If a
sensus technique. There is evidence for discriminative discrepancy existed, the single item was discussed to
validity for three of the scale items – randomisation,35 reach consensus. The researchers were not blinded to
concealed allocation.32,35 The other items are reported to
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 131

Table 2. American College of Sports Medicine guidelines39

MUSCLE STRENGTHENING
Frequency of 2–3 days per week, at least one set of 8–12 repetitions at the 8–12 repetitions maximum of each exercise using any type of
exercise that can be progressed over time performing both the concentric and the eccentric phase of the exercises in a controlled manner

FLEXIBILITY TRAINING
Controlled static stretch held for a given duration, intensity to a position of mild discomfort for 10–30 s; 6-s contraction followed by 10–30 s
assisted stretch for PNF, 3–4 repetitions of each stretch on two (minimum) or more days of the week

the title, authors, or journal title of the trials. If the sion) in order to address specific protocols.
disagreements persisted, a third reviewer (Carol
Shacklady) arbitrated. It is worth noting, that despite
Search strategy
some differences on methodological score, there were no
disagreements between the reviewers on which study was
The search strategy was based on combination of the
of low, moderate or high quality.
major key words: impingement syndrome, rotator
cuff, tendonitis, bursitis, exercise, manual therapy,
mobilisation, randomised controlled trial. Keywords
Quality of treatment protocols
and medical subject headings related to the condition
and potential treatment were identified prior to initi-
According to Philadelphia Panel evidence-based clin-
ating the search. Citation indexing was used to track
ical practice guidelines,17 the investigators were unable
referencing and key authors in the field conference
to make any conclusions regarding the effectiveness
abstracts were reviewed.
of therapeutic exercise for shoulder pain because
Extensive literature searches in texts and via com-
poorly defined inclusion criteria and non-validated
puter databases provided the material for review. The
outcomes were used. Furthermore, there were no
databases chosen were AMED, CINAHL, EMBASE
studies evaluating the effectiveness of manual therapy
and MEDLINE.
for non-specific shoulder pain. In this review, the
Subject headings (MeSH) and key words included
quality of exercise used in trials was assessed using the
anatomical terms, disorder or syndrome terms, treat-
American College of Sports Medicine guidelines39 for
ment terms, and methodological terms consistent
exercise dose (Table 2). This was complicated due to
with those advised by the Cochrane Back Group.
the variety of exercises used and the insufficiently
Specifically the following search-strategies were used
detailed programmes. Therefore, the exercise quality
for MEDLINE, EMBASE, CINAHL and AMED:
for each trial was determined based on the predomi-
nant exercise used. Those fulfilling the American 1. Shoulder Pain/
College of Sports Medicine39 criteria for muscle 2. Shoulder Impingement Syndrome/
strengthening and flexibility were classified as being 3. Rotator Cuff/
‘high quality exercises’. Those not meeting the criteria 4. exp Bursitis/
were ‘low quality exercises’. Exercise quality was con- 5. [(shoulder$ or rotator cuff) adj5 (bursitis
sidered ‘unclear’ if the dose was not mentioned or or frozen or impinge$ or tendinitis or
insufficient information was provided. tendonitis or pain$)].mp.
The quality of stabilisation exercises is not
6. rotator cuff.mp.
accounted for in the American College of Sports
7. or/1–6
Medicine guidelines. For this reason, it was elected to
include those trials using dynamic stabilisation within 8. exp Rehabilitation/
the strengthening category (Table 3). A similar 9. exp Physical Therapy Techniques/
approach to evaluate the quality of exercises for low- 10. exp Musculoskeletal Manipulations/
back pain has been previously reported by Liddle et 11. exp Exercise Movement Techniques/
al.40 Finally, since no guidelines were found in the lit- 12. (rehabilitat$ or physiotherap$ or physical
erature regarding the quality and dose of manual therap$ or manual therap$ or exercis$
therapy, there is an attempt in this review to deter- or mobili$).mp.
mine the characteristics of this therapeutic applica- 13. or/9–12
tion in the field of rehabilitation (e.g. the types of 14. Clinical trial.pt
manoeuvres used, adequate intensity and progres- 15. random$.mp.
132 TRAMPAS AND KITSIOS

Table 3. Treatment programmes for each of the included studies

Study LUDEWIG & BORSTAD (2003)46


Type of intervention (1) Symptomatic subjects with impingement syndrome (intervention group). (2) Symptomatic control
(no treatment). (3) Asymptomatic control (no treatment)
Setting Home exercise programme
Frequency Stretches: 5 repetitions. Exercise protocol: 3 times per week
Intensity Stretches: 30 s. Exercise protocol: 3 SETS of 10 repetitions 1st week, 15 repetitions 2nd week, 20 repetitions
3rd week, then increased resistance
Duration 3 weeks in total
Quality of exercise High quality for both strengthening and stretching exercises

Study HAAHr et al. (2005)45


Type of intervention (1) Control (training) group: Heat, cold packs, soft-tissue treatments and exercises. Exercise protocol: active
training of the periscapular muscles and strengthening of the stabilising muscles of the shoulder joint (the
rotator cuff). This was done within the limits of pain. (2) Surgery group: bursectomy with partial resection of
the antero-inferior part of the acromion and the coracoacromial ligament
Setting Hospital setting and home exercise programme afterwards
Frequency 3 times/week (first 2 weeks), twice/week (next 3 weeks), once/week (last 7 weeks). Then instructed to follow the
programme twice or three times weekly
Intensity Not stated
Duration 19 sessions, 60 min each. 12 weeks in total
Quality of exercise Unclear quality

Study JOHANSSON et al. (2005)49


Type of intervention (1) A group that received acupuncture (n = 44) and (2) a group that received ultrasound (n = 41). Both
interventions were given by physical therapists twice a week for 5 weeks in addition to a home exercise
programme (2-step programme)
Setting Home exercise programme (2 steps): first step, restore motion; second step, strengthen the rotator cuff muscles
Frequency First step: every day (2 exercises). Second step: every other day (6 exercises – 5 isometric, 1 concentric)
Intensity First step: 1 SET of 20 and 30 repetitions, respectively. Second step: 1 SET of 10 repetitions for the
5 stabilising exercises and 2 SETS of 15 repetitions for the concentric strengthening exercise
Duration First step: weeks 1–5. Second step: weeks 4–5
Quality of exercise Low quality exercises

Study CITAKER et al. (2005)47


Type of intervention (1) Hot-packs, manual mobilisation, and Theraband exercises were performed in the mobilisation group
patients and (2) hot-packs, PNF and Theraband exercises were performed in the PNF group
Setting Not stated
Frequency No details given
Intensity No details given
Duration 20-session treatment followed by 3 weeks of Theraband exercises
Quality of exercise Unclear quality

Study WALTHER et al. (2004)48


Type of intervention (1) Standardised self-training: centring and stretching exercises (elastic Theraband). Guidance by a
physiotherapist for a maximum of 4 sessions. (2) Conventional physiotherapy: 10 sessions of physiotherapy
consisting of centring training for the rotator cuff (stretching if necessary). (3) Functional brace: functional
shoulder brace
Setting Standardised self-training: at home. Conventional physiotherapy: in a physiotherapy department
Frequency Standardised self-training: at least 5 times per week, 10–15 min. Conventional physiotherapy: 2–3 times per week
Intensity Standardised self-training: 7 strengthening exercises (the majority of them 1 SET of 10 repetitions) and
1 stretching exercise of the neck (twice for 15 s). Conventional physiotherapy: no details given
Duration 12 weeks
Quality of exercise The quality of the exercises described was high

Each programme was evaluated according to the ACSM guidelines.39 The programmes fulfilling the criteria of the ACSM were scored
as ‘high quality’ exercises in contrast to those not meeting the criteria and characterised as ‘low quality’ exercises. Exercise quality was
considered ‘unclear’ if insufficient information was presented. Manual therapy programmes were used to address specific protocols,
since no guidelines were found in the literature for manual therapy and shoulder disorders.
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 133

16. [(single or double) adj (blind$ or 6, Evidence of adverse effect was used for trials that
mask$)].mp. showed lasting negative changes.
17. placebo$.mp. Sensitivity analysis exploring the results when high
18. or/14-17 quality was redefined using other thresholds44 was
19. 7 and 13 and 18 also performed.
A similar search strategy was followed in another
review, according to the principles of the Cochrane
Musculoskeletal Review Group.41 Abstracts from RESULTS
proceedings and unpublished trials were not included.
The computerised literature search strategies located
302 citations. Of these, 192 were excluded immedi-
Data presentation and analysis ately from their abstract because they were considered
irrelevant to this review and 105 of the remainder
The following data were extracted from each study: were also excluded because they did not fulfil the
methodological quality assessment, methods of inter- selection criteria – either they defined manual and
nal validity, number of patients included in each arm, exercise therapy differently from the definition given
causes of impingement, types of intervention, number in this systematic review, or they did not define the
of sessions, types of outcome measures, timing of out- interventions at all. The mean quality score of the
come assessment, the authors’ conclusions about the included trials was 6.2, ranging from 4 to 8 points out
effectiveness of the interventions, and the reviewers’ of a total of a possible 10 points (Table 5).
score. Also data on the characteristics of the study Five randomised, controlled trials were included in this
population (such as type, location and duration of systematic review; all were published in English. Three of
pain, age and gender) and on the manual therapy and the studies were undertaken in Europe, one in the US and
exercise regimen (such as frequency, intensity, dura- one in Asia. One RCT was identified from AMED, one
tion and setting) were extracted. The protocol con- from CINAHL, two from EMBASE, one from MED-
tained a special data extraction form designed by the LINE. One study included subjects from both genders.45
authors and based on the guidelines of the Cochrane Ludewig and Borstad46 included only males (one female
Library.42 This form listed the data items to be volunteered but did not meet the clinical inclusion crite-
extracted from each of the primary studies and was ria). Construction workers as used in this study are gener-
used to facilitate the comparison process (Table 4). ally 98% male, and it is uncertain if response to exercises
According to the reviewers, the studies were clini- would be similar in female workers. In the rest of the stud-
cally heterogeneous with respect to the type of the ies, gender was not specified and there was no reference to
interventions and the outcomes. Furthermore, the out- the male:female ratio.47–49
comes were poorly presented in three of the five studies Only one study investigated the effectiveness of both
so that pooling was not possible. In these instances, a manual and exercise therapy,47 one investigated the
qualitative review was performed where possible (best effectiveness of exercise to control (placebo) treatment,46
evidence synthesis) by attributing various levels of two examined the effectiveness of exercise to other
evidence of the effectiveness of mobilisation and exercise treatments,45,48 and one investigated the effectiveness of
therapy, taking into account the participants, interven- exercises in combination with other conventional
tions, outcomes, and methodological quality of the origi- treatments.49
nal studies.42 In addition, qualitative analysis was carried None of the studies included patients with acute symp-
out, using the levels of evidence listed below which have toms (duration < 3 weeks). One study49 included patients
been used in a previous systematic review:43 with subacute symptoms (duration at least 2 months) and
1. Strong evidence denoted consistent findings in two studies45,46 included patients with chronic symptoms
multiple high quality RCTs. (5 years of intermittent symptoms and symptoms
between 6 months and 3 years, respectively). In the study
2. Moderate evidence denoted findings in a single, by Citaker et al.47 the duration of symptoms for the
high quality RCT or consistent findings in included subjects was not mentioned.
multiple low-quality trials. Three studies adequately defined shoulder impinge-
3. Limited evidence indicated a single low-quality RCT. ment syndrome, according to the diagnostic criteria
reported by Desmeules et al.6 and Michener et al.7
4. Conflicting evidence denoted inconsistent results Two of these studies had also the best methodological
in multiple RCTs. score,46,49 whilst the third study by Haahr et al.45 was
5. No evidence meant no studies were identified. of moderate methodological quality. The other two
134 TRAMPAS AND KITSIOS

Table 4. Summary data for RCTs evaluated

Study HAAHR et al. (2005)45


Participants Ninety consecutive patients with subacromial impingement agreed to participate. Forty-five cases were randomised to
conservative treatment and 45 to surgical treatment. In the physiotherapy group, 42 persons (93%) were followed for
12 months with the main outcome measure. (Constant-Murley score). In the surgery group, 40 persons (89%) had complete
follow-up data. The two groups were very similar with respect to demographic characteristics, though a slightly
greater proportion within the surgery group had been on sick leave owing to shoulder pain within the past 3 years
Inclusion criteria Diagnostic criteria: the presence of shoulder pain, pain on abduction of the shoulder with painful arch, a positive
impingement sign (Hawkins sign) and a positive impingement test (relief of pain within 15 min after injection of local
anaesthetic [bupivacaine 5 ml] into the subacromial space). Eligibility criteria: fulfilment of all diagnostic criteria,
report of shoulder symptoms between 6 months and three years (because surgery in general was not offered to cases
with symptoms of shorter duration), and aged 18–55 years. Previous treatment with rest, non-steroidal anti-
inflammatory drugs, subacromial injection, and physiotherapy was allowed. Normal passive glenohumeral movement
was a requirement
Exclusion criteria Impaired rotation in the glenohumeral joint, a history of acute trauma, previous surgery or previous fracture in the
proximity of the affected shoulder, known osteoarthritis in the acromioclavicular or glenohumeral joints,
calcifications exceeding 2 cm in the rotator cuff tendons, or signs of a rupture of the cuff or cervical root syndromes
Follow-up periods At baseline and at 3, 6 and 12 months
Outcomes Pain, disability, strength, ROM (Constant-Murley score) – Another set of questions (PRIM) and information on job
title and workplace (measurement tool?)
Authors’ conclusions Surgical treatment of rotator cuff syndrome with subacromial impingement was not superior to physiotherapy
with training. There is a need for larger scale studies with sufficient numbers of participants to allow for stratification
into subgroups with different baseline levels of disability

Study JOHANSSON et al. (2005)49


Participants Eighty-five patients with clinical signs of impingement syndrome. The subjects were recruited from 3 urban primary
health care centres in the county of Ostergotland, Sweden, from March 1997 to June 2000. Aged 30–65 years
Inclusion criteria Typical history: pain located in the proximal lateral aspect of the upper arm (C5 dermatome), especially during arm
elevation; a positive Neer impingement test (subacromial injection of anaesthetic); at least 2 months’ duration of the
current episode; additionally three of the following four inclusion criteria should be positive: (i) Hawkins-Kennedy
impingement sign; (ii) Jobe supraspinatus muscle test (in 90° of abduction in the scapular plane); (iii) Neer
impingement sign; and (iv) painful arc between 60° and 120° of active abduction
Exclusion criteria Radiological findings: malignancy, osteoarthritis of the glenohumeral joint, skeletal abnormalities decreasing the
subacromial space (bony spurs, osteophytes), known or suspected polyarthritis, rheumatoid arthritis, or diagnosed
fibromyalgia. Previous fractures. Dislocation of the glenohumeral joint or the clavicular joints. History or current
clinical findings of instability. Suspicion of frozen shoulder and pain during intra-articular mobilisation. Problems
from the cervical spine. Having received any of the treatment alternatives in the study earlier for the current problem.
Having received a corticosteroid injection during the last 2 months for the current problem. A clinical picture of
ruptured rotator cuff. Acute subacromial bursitis. Communication problems
Follow-up periods Baseline, after treatment and at 3, 6 and 12 months
Outcomes Three disease-specific shoulder assessment scales: the CM Score, the Adolfsson-Lysholm Shoulder Score (AL Score)
and the University of California at Los Angeles End-Result Score (UCLA Score)
Authors’ conclusions The results suggest that acupuncture is more efficacious than ultrasound when applied in addition to home exercises,
but the influence of psychosocial factors is unknown, because no instrument covering this area was used

Study WALTHER et al. (2004)48


Participants Sixty consecutive patients (aged 25–66 years) with painful disabling impingement syndrome of the shoulder (Neer I
and II) were recruited
Inclusion criteria Diagnosis of subacromial impingement was established by clinical examination, radiographs of the shoulder in three
planes, and ultrasound. The Neer test (subacromial injection of 10 ml pure bupivacaine) was positive in all patients
Exclusion criteria Concomitant cervical radiculopathy, frozen shoulder, full-thickness tear of the rotator cuff, disorders of the
acromioclavicular joint, degenerative arthritis of the glenohumeral joint, calcifying tendinitis, shoulder instability,
post-traumatic disorders, and involvement in workers’ compensation claims
Follow-up periods At 6 and 12 weeks after the initial visit
Outcomes Constant-Murley score, which includes pain, activities of daily living, pain-free range of motion, and muscle power.
Also pain intensity was recorded (VAS) and a diary was used to document the frequency of therapy, the frequency
of self-training, and the usage of the brace, as well as the inability to work and additional medication
Authors’ conclusions An analysis of the sub-items of the Constant-Murley score showed that the functional brace led to a significant
improvement in strength compared with groups 1 and 2. All other sub-items did not show any significant difference
among the three groups. Initial concerns that the brace may compromise shoulder mobility did not prove to be true.
A reason might be the effects on neuromuscular mechanisms, such as proprioceptive feedback transmitted by
cutaneous receptors and the reduction in the levels of pain

(continued on next page)


TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 135

Table 4. (continued) Summary data for RCTs evaluated

Study CITAKER et al. (2005)47


Participants Forty patients were equally randomised. Mean age: 52.8 ± 9.86 years, mobilisation group; 55.5 ± 8.55 years, PNF
group. Impingement-related, diagnosed clinically and radiologically documented. None of the patients in the study
underwent surgical procedures, physical treatments or rehabilitation programmes before the study. Patients were not
treated with local steroid injections previously. The patients did not to take oral analgesics during the experimental
period. No further details were stated about the patients, such as demographic characteristics, but it was mentioned
that they were in accordance with those in the literature
Inclusion criteria Not defined
Exclusion criteria Not defined
Follow-up periods Not clarified
Outcomes Pain (Visual Analogue Scale), range of motion (goniometer) and University of California at Los Angeles (UCLA)
criteria, before and after treatment at night and day
Authors’ conclusions Mobilisation and PNF methods are both effective, but mobilisation is a painless technique and, therefore, was better
tolerated than PNF; this was the main difference between the techniques

Study LUDEWIG & BORSTAD (2003)46


Participants Ninety-two patients(all male): intervention (n = 34), control-symptomatic (n = 33), control-asymptomatic groups
(n = 28). Groups had similar demographic variables (age, height, weight, overhead work, years in trade)
Inclusion criteria Occupational exposure to overhead work more than 1 year, minimum of 130° active abduction, referred pain to
C5–C6 dermatome above the deltoid insertion, two positive impingement tests (Neer, Hawkins/Kennedy, Yocum,
Jobe and/or Speeds tests) and two positive out of three additional tests: (i) a painful arc on active scapular plane
abduction of the arm; (ii) tenderness to palpation of the biceps or rotator cuff tendons; and (iii) pain with one or
more resisted glenohumeral joint motions (flexion, abduction, internal rotation, or external rotation)
Exclusion criteria History of rotator cuff surgery. History of glenohumeral dislocation, or other traumatic injury to the shoulder.
Periscapular or cervical pain during arm elevation. Shoulder symptoms by a cervical assessment
Follow-up periods In an average of 10 weeks. Seven subjects were lost to follow-up, four (11.8 %) in the exercise intervention group,
and three control subjects (one symptomatic [3%] and two asymptomatic [8%])
Outcomes Shoulder Rating Questionnaire and 10 modified questions using the format of Shoulder Pain and Disability Index
(all self-reported outcome measures)
Authors’ conclusions A home exercise programme could be effective in reducing symptoms and improving the self-reported functional
status in construction workers. High compliance subjects were also those who showed the most consistent
improvement and increases in repetitions and resistance. Further study assessing more objective measures of the
physiological effects of the exercise intervention would be beneficial since physiological variables are less likely
impacted by possible placebo effects

The table lists the data items to be extracted from each of the primary studies and used to facilitate the comparison process.

studies47,48 used only one and three diagnostic criteria, scores remained significantly below those of the asympto-
respectively (Table 4). In both of these studies, it was matic group. In another study by Walther et al.45 the effec-
mentioned that the diagnosis of impingement syndrome tiveness of therapeutic exercises versus physiotherapy and
was based on radiographs and clinical examination bracing was evaluated. All three groups showed a signifi-
without stating any further details. cant improvement in shoulder function and reduction in
One study46 compared the effectiveness of exercises pain, with no significant differences among the three
with two control groups receiving no treatments (one groups. This confirms the effect of muscular strengthen-
symptomatic and one asymptomatic group). The purpose ing of the rotator cuff, either by physiotherapy or by
of this investigation was to compare the results with guided self-training. The authors concluded that from the
healthy subjects from the same population. There was evi- economic point of view, self-training with supervision to a
dence that a home exercise programme can be effective in maximum of 4 sessions is superior to conventional phys-
reducing symptoms and improving function in this popu- iotherapy with about 30 sessions. However, for this to be
lation, but subjects in the intervention group did not successful, competent instruction as well as good compli-
obtain similar scores to the asymptomatic group. Subjects ance by the patient, is mandatory. If this cannot be
in the intervention group showed significant improve- secured, conventional physiotherapy should follow the
ments from pretest to post-test in their functional scores principles of centring the humeral head.
(Shoulder Rating Questionnaire and Shoulder Pain and Another study47 compared mobilisation and
Disability Index), while control subjects remained essen- Theraband exercises versus proprioceptive neuromuscu-
tially stable, resulting in an interaction of group and time lar facilitation (PNF) and Theraband exercises again. In
(F = 9.12; P < 0.001). Despite this, at post-test, all symp- both groups, subjects were also treated with the applica-
tomatic subjects’ average Shoulder Rating Questionnaire tion of hot packs. The authors sought to evaluate whether
136 TRAMPAS AND KITSIOS

Table 5. The PEDro scoring of the included RCTs.

Study Citaker et al. Haahr et al. Johansson et al. Ludewig & Borstad Walther et al.
(2005)47 (2005)45 (2005)49 (2003)46 (2004)48

1 Yes Yes Yes Yes Yes


2 ? No Yes Yes ?
3 Yes Yes Yes Yes Yes
4 No No No No No
5 No No No No No
6 No No Yes Yes Yes
7 ? Yes Yes Yes Yes
8 ? ? Yes Yes ?
9 Yes Yes Yes Yes Yes
10 Yes Yes Yes Yes Yes
Quality of the trial Low (4/11) Moderate (6/11) High (9/11) High (9/11) Moderate (7/11)

The PEDro scoring of the included RCTs. Item 1(listed in Table 1) is not included, since this item pertains to external validity.
Consequently, item 1 from Table 5 corresponds to item 2 (randomisation) from Table 1, etc.

mobilisation techniques were more effective and clinically acupuncture plus exercises should be advocated over
useful than PNF; both were found to be effective, but ultrasound when treating patients with impingement
mobilisation is a painless technique and, therefore, was syndrome.
better tolerated than PNF. This was the only difference
between the two groups. The authors came to this conclu-
sion as the decrease in pain complaints was found to be Outcomes
more significant (after treatment, during night and day, at
rest and after activity) in the mobilisation rather than the Pain intensity was the primary outcome in four of the
PNF group. However, since no statistically significant dif- five studies. One of these studies was of high method-
ference between the groups was found with regard to sub- ological quality,49 two were of moderate quality,45,48
jective pain intensity when pre- and post-treatment values and one was of low quality.47 All used the Visual
were compared. Analogue Scale (VAS) as the pain measurement
Haahr et al.45 compared the effects of exercises aimed at instrument. In two of the previous studies, pain was
strengthening of the stabilisers and decompressors of the assessed as a separate outcome47,48 as well as part of
shoulder (in conjunction with hot packs, cold packs, and functional questionnaires. Citaker et al.47 recorded
soft tissue treatments of the shoulder) to the effects of sub- pain intensity during the day and night, whether the
acromial arthroscopic surgery, and found surgery for patients were active or motionless, as well as before
rotator cuff syndrome with subacromial impingement and after treatment. They reported that there was a
was not superior to physiotherapy. Even though the significant reduction in pain intensity in both treat-
authors found better outcome scores in the surgery group, ment groups for patients suffering from stage II and
they were reluctant to recommend surgery in cases with stage III impingement syndrome, but there was no sig-
stage II impingement. There is a need for trials with suffi- nificant difference between the groups (pain scores post-
cient numbers of participants to be stratified into sub- treatment: mobilisation, 1.50 ± 2.30, P < 0.002; PNF
groups before more rigorous definitive recommendations 1.85 ± 2.64, P < 0.001). Mobilisation is a painless tech-
are made. Johansson et al.49 compared manual acupunc- nique and the authors concluded that it was better tol-
ture and continuous ultrasound, both applied in addition erated than PNF. This was the only positive finding
to a home exercise programme. At the first treatment visit, with regards to pain intensity measured with VAS,
patients received instructions from the physical therapist and it is not supported by the other measures for pain
and practiced the exercises in part one of the programme. in the same study. Walther et al.48 compared a home
They were instructed to perform the programme daily for exercise programme with physiotherapy and bracing.
5 weeks. After the first half of the treatment period, the They investigated pain intensity at rest, at night and
patients received instruction and practiced the second during load, and at 6 and 12 weeks after the initial
part of the exercise programme. Compliance with the visit. All three treatment groups showed a significant
exercise programme was monitored by a home exercise reduction in pain intensity, at night as well as during
adherence log. The results of this study suggested that rest periods, and while under stress (P < 0.05) during
acupuncture is more effective than ultrasound when used the study but again there were no statistically signifi-
in conjunction with home exercises; for this reason, cant differences between groups at any time.
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 137

Furthermore, in the other two studies by Johansson et ferences between the subjects in the intervention and
al.49 and Haahr et al.45 and in the others described the symptomatic control groups showed the signifi-
above, pain was also measured as an outcome as an cantly improved Shoulder Rating Questionnaire
integral component of the outcome measurement. scores in the intervention group (F = 7.66; P < 0.01).
The Constant-Murley questionnaire used in three Satisfaction scores also showed a significant interac-
studies, two of moderate45,48 and one of high method- tion of group and time with significant improvements
ological quality,49 measures pain intensity on a VAS (as in satisfaction in the intervention group (F = 8.97; P
an item for the questionnaire); these results showed sig- < 0.001). Similar patterns of responses across groups
nificant reduction in the treatment groups, but no dif- were also noted for work-related pain and disability
ferences between them. Johansson et al.49 did not scores, with significant interactions of group and time
provide sufficient information with regards to pain (F = 13.95, P < 0.001; F = 9.16, P < 0.001, respec-
intensity, only the mean score from all three question- tively). In a second high quality study by Johansson et
naires, reporting that both treatment groups improved al.,49 the authors used the Adolfsson-Lysholm
during the study (P < 0.001). Haahr et al.45 presented Shoulder Score with high intra-observer reliability for
the subscores of the different variables measured with impingement syndrome; this scale is purely patient
the Constant-Murley questionnaire. The mean self-assessment. This trial also used the UCLA crite-
change for pain in both groups ranged from 2.8 to 3.8 ria (as described above) including both objective and
(VAS: 0, maximum; 15, no pain), but no statistically subjective evaluations (function, satisfaction, pain in
significant differences were found. daily activities); the results indicated that both treat-
Strength was another important outcome used in ments were equally effective in reducing symptoms.
these studies as a correlate of pain intensity; in all UCLA criteria were also used in a low quality study
cases, strength was not assessed as a separate variable by Citaker et al.47 which reported that 95% of ‘good’
but measured as a subscore of a questionnaire also results were obtained by performing mobilisation and
assessing other outcome variables. The University of 65% by performing PNF (P < 0.001). Finally, in a
California at Los Angeles (UCLA) criteria were used study with moderate methodological quality,45 the
in a low47 and a high quality study.49 In both studies, patients completed a follow-up questionnaire after 1
the authors reported that the total UCLA score was year of the initial visit; this included a self-assessment
significantly better for the treatment groups, but no measure of pain and discomfort in daily activities and
further details were given for the subscore. Haahr et in work-related activities and found no group differ-
al.45 used the Constant-Murley score to assess strength ences. The scale had been used previously in a Danish
at 3, 6 and 12 months follow-up (P < 0.71, P < 0.88, project on research and intervention in monotonous work
P < 0.96, respectively) but changes were not statisti- (PRIM).50 The Constant-Murley score used in three stud-
cally significant between the two groups. In a moderate ies45,48,49 is a joint measure of four subscores: (i) pain mea-
quality study by Walther et al.48 Constant-Murley sured on a visual analogue scale (VAS); (ii) limitations in
scores showed that the group treated with bracing had activities of daily living; (iii) active range of motion in four
significantly better strength gains than the exercise and directions in the shoulder joint; and (iv) isometric shoul-
the physiotherapy groups at 12 weeks’ follow-up. der strength measured in kilograms with a portable
The functional status of patients was measured in muscle strength analyser. In the study by Haahr et
all of these studies using a variety of questionnaires as al.45 the mean improvement in Constant-Murley score
measurement tools. Results indicated that manual in the physiotherapy group was 23.0 (95% confidence
and exercise therapy were more effective (but not sig- interval [CI], 16.9–29.1), and in the surgery group the
nificantly) in increasing the functional status of the improvement was 18.8 (95% CI, 11.5–26.1). Two
patients compared to other treatment modalities. In patients in the physiotherapy group and eight in the
four studies, self-perceived changes in symptoms were surgery group had a reduction in the Constant-Murley
measured using different questionnaires as measure- score. However, there were no differences at any follow-
ment tools. The high quality study by Ludewig and up point, neither did the results suggest any trends dur-
Borstad46 measured the functional levels of construc- ing the study period. Johansson et al.49 gave no specific
tion workers using two types of questionnaires with details on the Constant-Murley score, but Walther et
self-reported measures of pain and disability; the al.48 found that there was a significant improvement in
Shoulder Rating Questionnaire and 10 modified the Constant-Murley score within 12 weeks in all three
questions of the Shoulder Pain and Disability Index. groups (P < 0.05); however, there were no differences in
The results showed significant improvements for the improvements between the three groups.
treatment group over the course of treatment, but Range of motion as a separate outcome was evalu-
scores remained significantly below the asymptomatic ated only in a low quality study by Citaker et al.47
controls at all follow-up periods. The analysis of dif- using a goniometer as a measurement tool. In the
138 TRAMPAS AND KITSIOS

mobilisation group, the changes of ROM values were in reducing pain in patients with impingement syn-
highly significant for flexion (P < 0.001) and moderately drome. There is no evidence to support the effective-
significant (P < 0.01) for abduction, external rotation ness of manual therapy alone for the treatment of
and internal rotation in stage 2 impingement. In con- impingement syndrome. However, there is moderate
trast, in the PNF group, the changes were highly signifi- evidence that supports the use of therapeutic exercises
cant for flexion (P < 0.001) and moderately significant alone in reducing pain and function. Finally, there is
(P < 0.01) for abduction, external rotation, internal rota- limited evidence to support the effectiveness of com-
tion and hyperextension in stage 2 impingement. Finally, bined manual and exercise therapy for the treatment
in stage 3 impingement flexion, abduction and external of shoulder impingement syndrome. These results are
rotation were statistically significant in both groups (P < in keeping with two previous systematic reviews.6,7
0.05). The active range of motion for flexion was also However, because the trials investigated were clini-
evaluated in two moderate quality studies.45,48 Both stud- cally too heterogeneous with regard to patients’ char-
ies measured ROM as subscore for the Constant- acteristics, control treatments and co-interventions, it
Murley score, indicating no statistically significant is difficult to determine definitively the effectiveness
differences among the groups. of manual and exercise therapy alone in patients with
shoulder impingement syndrome.
There is limited evidence to support the use of exer-
Quality of exercise and manual therapy cise and manual therapy to improve strength in
patients with impingement syndrome. None of the
Two trials, one of high methodological quality46 and studies evaluated strength as a separate outcome, but
one of moderate quality48 adequately described the only as part of functional questionnaires. Four stud-
type of exercises used. Ludewig and Borstad46 ies, one of high,49 two of moderate,45,48 and one of
reported that a home programme of strengthening low47 methodological quality showed a significant
and stretching exercises could be effective in reducing improvement in several types of questionnaire which
pain and disability of construction workers. These included strength as an item. Only one study of mod-
authors reported details for setting, frequency, inten- erate quality by Walther et al.48 presented the sub-
sity and duration of their programme, and quality scores of the included variables, and showed no
based on the ACSM39 criteria proved to be high. A significantly better results for bracing compared to
high quality of exercise programme was also used by exercises with regards to power/strength.
Walther et al.,48 who concluded that a home exercise None of the studies evaluated interventions using
programme (fully described duration, intensity and fre- subjects with acute impingement syndrome indicating
quency) could be equally as effective as conventional a need for further research in this area. One study49
physiotherapy and bracing in improving pain and func- used subjects with subacute symptoms and showed
tion. Johansson et al.49 also gave a full description of that exercises were not superior to bracing or physio-
their treatment programme, but the quality compared to therapy. In chronic impingement syndrome, there is
the ACSM39 guidelines was considered to be low. Haahr moderate evidence to support the use of exercises
et al.45 showed one basic limitation in describing their alone, since two studies, one of high46 and one of
programmes, since no details were provided with regards moderate45 quality showed that exercises significantly
to the intensity of the exercise protocol; thus, the quality improved the symptoms compared to no treatment;
of the exercise programme was unclear. Finally, Citaker however, exercise was not superior to surgery. Haahr
et al.47 had several limitations in presenting the treat- et al.45 reported adverse effects for both exercises and
ment protocols; no details were given for frequency, surgery, indicating some degree of caution may be
quality and intensity of the PNF patterns, and the required with these treatments.
Godman pendulum home exercises used in the trial. For The use of functional outcomes is necessary to evalu-
these reasons the quality of the exercises was unclear. ate the patient’s condition and progress fully. The only
Furthermore, the authors gave no information on the functional questionnaire to have been approved by the
type of mobilisation manoeuvres used; thus, no conclu- executive committee of the European Society for
sions could be extrapolated from this review regarding Surgery of the Shoulder and Elbow51 is the Constant-
the quality of manual therapy. Murley score, which has been used in three of the stud-
ies in the current review.45,48,49 The UCLA criteria were
used in two studies.47,49 Although these questionnaires
DISCUSSION seemed appropriate, data on their construct, content
and criterion validity are currently lacking. Other self-
The results of the current review showed that exercise perceived changes in symptoms were evaluated by the
therapy in combination with other treatments is effective Shoulder Rating Questionnaire and 10 modified ques-
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 139

tions of the Shoulder Pain and Disability Index.46 The moderate methodological quality48 would have been
first of these questionnaires has good reliability but its redesignated as of higher quality. It could be assumed
validity and responsiveness have not been previously that Walther et al.48 had adequate allocation conceal-
reported, whilst the modified questions were not vali- ment, and that intention-to-treat analysis was used.
dated at all. Beyond this, the PRIM score used by The quality of the other studies by Haahr et al.45 and
Haahr et al.45 and the Adolfsson-Lysholm shoulder Citaker et al.47 with criteria scored as ‘don’t know’
score used by Johansson et al.,49 as secondary measure- would remain the same. The overall conclusions of
ment tools, did not appear to be sufficiently validated. this review are, therefore, sensitive to change in favour
Moreover, the strengthening and stretching pro- of exercise therapy, since there would then be strong
gramme studies were too heterogeneous with regards evidence (in contrast of moderate evidence) to sup-
to intensity, frequency and duration, and no general port its effectiveness. Similarly, if only high quality
recommendations can be made. Two studies, one of studies were included in this analysis, the same con-
high46 and one of moderate48 quality described clusion would have been reached, but there would be
stretching and strengthening programmes based on the no evidence to support the effectiveness of manual
ACSM guidelines,39 whilst the quality of the pro- therapy.
grammes used in the rest of the studies was unclear.
Furthermore, it is still not possible to make any recom-
mendations concerning the type of manual therapy Methodological flaws
which is effective in treating impingement syndrome of
the shoulder, since the only study including this inter- Methodological flaws were seen in both high and low
vention was of low quality,47 and the manoeuvres used quality trials. The most important was the lack of
were not described at all. blinding of patients.45–49 If participants are aware of
Finally, this review confirmed the lack of uniformity in which intervention they are to receive, they may
defining and evaluating the shoulder impingement syn- respond differently according to their own prefer-
drome. Based on these findings, impingement syndrome ences and expectations. They may have more faith in a
typically affects people between the age of 18–65 years,45–49 treatment that appears to be active than in one that
and the Hawkins sign45,49 and Neer test46,48,49 are the appears to be passive. Similarly, if the outcome asses-
most widely used diagnostic tests; usually, a painful sors have a preference for one treatment over another,
arc is observed, and the duration of symptoms, in and know which subjects have received which treat-
most cases, is longer than 2 months.45–48 Future stud- ment, they may consciously or subconsciously over-
ies should include more complete details of the clini- rate the improvement in the group receiving the
cal examination used resulting in a better-defined ‘preferred’ treatment. Thus, another important
clinical pathology. methodological flaw was the lack of blinded outcome
assessors detected in all studies except the one trial by
Common limitations in the studies reviewed included: Ludewig and Borstad.46 Furthermore, blinding of ther-
apists performing the treatment involves ensuring that
1. Outcome measures demonstrated poor or unmeasured
therapists were unable to discriminate whether individ-
reliability and/or validity.
ual subjects have or have not received the treatment. In
2. Many trials had poor design and potential type-2 physiotherapy, however, blinding of the care provider
errors in both data acquisition and analysis. is not feasible in most cases and for this reason no
specific attention was paid to this methodological
3. Randomisation procedures were poor or inadequate.
flaw (although obvious in all trials).
4. The majority of clinical trials do not provide sufficient The lack of description of drop-outs was another
details to allow the studies to be replicated. methodological flaw and it is related to attrition bias.
Two studies47,48 did not mention any drop-outs. Haahr
5. Descisions on the effectiveness of manual therapy and
et al.45 identified drop-outs, but did not clarify
exercises are confounded by interventions which
whether these subjects were included in an intention-
include other elements of physical therapy treatment,
to-treat analysis or not. Only two studies stated the
such as ultrasound, acupuncture and hot/cold packs.
exact number of patients who withdrew, and included
these in an intention-to-treat analysis, although in
each case the results of the study were not affected.46,49
Sensitivity analysis Patients sometimes withdraw from studies because they
have recovered or they follow other interventions dur-
After re-scoring positively those validity items, which ing the treatment period; others might drop-out
had been scored as ‘don’t know’, only one study of because they deteriorate, or are not satisfied with the
140 TRAMPAS AND KITSIOS

treatment. Given this, it is necessary to report the num- alone, or in combination with various other types of
bers and the reasons for drop-outs in each study group. treatment, for pain, function and disability; in con-
Another problem appears to be the randomisation trast, exercises were effective in combination with
procedure. Most publications lack information about other treatments, such as ultrasound, acupuncture
the allocation of patients to study groups: only stating and hot/cold packs, for pain and functional status. It
that the patients were divided ‘randomly’, or similar was not possible to determine which type and dosage
expressions. Thus, readers have no information on of exercise or technique and dosage of manual ther-
which to determine the degree of potential bias. For apy was more beneficial, or if certain subgroups bene-
example, the use of concealed allocation,49 the selec- fited more from one form of care than another.
tion of a slip of paper indicating group assignment,46 Although the Visual Analogue Scale was the pri-
or allocating patients on an alternating basis are sus- mary measurement tool for pain intensity, the use of
ceptible to (selection) bias. Unfortunately, it is not functional outcomes was also necessary to evaluate
known whether this problem is limited to reporting fully the patient’s condition and progression. The
only, or whether less valid procedures were used. Constant-Murley score appeared to be the most
Finally, power analysis was undertaken in the two stud- appropriate, and it is the only one to have been
ies of high methodological quality,46,49 whereas in the rest approved by the executive committee of the European
of the studies it was not mentioned. Moreover, in the Society for Surgery of the Shoulder and Elbow.51
study by Ludewig and Borstad,46 only male workers Despite this, current knowledge on construct, content
were included, and it is not certain if response to and criterion validity is not sufficient.
exercise would be similar in female workers.
Additionally, the majority of subjects in the study
experienced chronic, intermittent shoulder symp- Implications for research
toms, with 3 years as the average time since initial
onset of symptoms. More acute subjects may The emphasis of future randomised, controlled trials
respond differently to a similar exercise programme. should be on specific therapeutic methods, such as well-
defined therapeutic exercise programmes or a manual
therapy regimen. This would assist in establishing the
CONCLUSIONS effectiveness of these techniques and allow clinicians to
take an evidence-based practice approach when treating
This systematic review identified five studies that patients.
assessed the effectiveness of manual and exercise therapy Most methodological problems such as sample size of
in the treatment of shoulder impingement syndrome. study populations, blinding of assessors, can be improved
Although there were methodological flaws which pre- with better design, but some appear unavoidable or diffi-
clude strong conclusions, it appears that manual and cult to prevent, such as blinding of patients. Thus, the pre-
exercise therapy may be more effective than other inter- sent studies do not offer sufficient evidence for strong
ventions in the treatment of shoulder impingement syn- conclusions to be drawn. Future research should ade-
drome, especially in respect of short-term effects. This quate design and control in all trials.
finding is based on limited data at this time. Future trials need to define shoulder impingement
Specifically, analysis of the existing data indicates: syndrome better, and further validation of outcome
(i) strengthening and stretching exercises are effective measures applied to patients suffering from this
approaches in subacute and chronic shoulder pathology is necessary. More studies are needed in
impingement syndrome patients; (ii) there is no order to evaluate the effectiveness of manual therapy
strong evidence of the effectiveness of manual therapy for shoulder impingement syndrome, since there was
in acute, subacute and chronic shoulder impingement only one report of low methodological quality. It is
syndrome; and (iii) there is no evidence establishing also necessary that future trials should compare the
the effectiveness of exercises and manual therapy in effectiveness of manual and exercise techniques in
acute shoulder impingement syndrome patients. In shoulder impingement syndrome with other com-
addition, further research on the long-term effective- monly accepted treatments for the same condition.
ness of manual and exercise therapy techniques in
acute, subacute or chronic impingement syndrome of
the shoulder is still warranted. REFERENCES
Multimodal care, including mobilisation and exer-
cises, appears to be beneficial for pain relief and func- The most important references are denoted with an asterisk.
tional improvement, but the evidence is limited. 1 Burkhart SS. Congenital subacromial stenosis. Arthroscopy
Moreover, the evidence did not favour manual therapy 1995;11:63–8
TREATMENT OF IMPINGEMENT SYNDROME OF THE SHOULDER: A SYSTEMATIC REVIEW 141

2 Wang JC, Horner G, Brown ED, Shapiro MS. The relationship 24 Van Der Heijden CJMG, Beurskens AJMH, Koes BW,
between acromial morphology and conservative treatment of Assendelft WJJ, De Ve HCW, Bouter LM. The efficacy of
patients with impingement syndrome. Orthopedics traction for back and neck pain: a systematic blinded review of
2000;23:557–9 randomized clinical trial methods. Phys Ther 1995;75:93–104
3 Copeland S. Throwing injuries of the shoulder. Br J Sports 25 Winters JC, Sobel JS, Groenier KH, Arendzen HJ, Meyboom-
Med 1993;27:221–7 de Jong B. Comparison of physiotherapy, manipulation, and
4 Neer CS. Impingement lesions. Clin Orthop 1983;173:70–7 corticosteroid injection for treating shoulder complaints in
5 McClure PW, Bialker J, Neff N, Williams G, Karduna A. general practice: randomised, single blind study. BMJ
Shoulder function and 3-dimensional kinematics in people 1997;314:1320–5
with shoulder impingement syndrome before and after a 6- 26 Jobe FW, Moynes DR. Delineation of diagnostic criteria and a
week exercise program. Phys Ther 2004;84:832–48 rehabilitation program for rotator cuff injuries. Am J Sports
6* Desmeules F, Cote C, Fremont P. Therapeutic exercise and Med 1982;10:336–9
orthopedic manual therapy for impingement syndrome: a 27 Magee DJ. Orthopedic Physical Assessment, 3rd edn.
systematic review. Clin J Sports Med 2003;13:176–82 Philadelphia, PA: WB Saunders, 1987
7* Michener L, Walsworth M, Burnet EN. Effectiveness of 28 Hawkins RJ, Kennedy JC. Impingement syndrome in athletes.
rehabilitation for patients with subacromial impingement Am J Sports Med 1980;8:151–8
syndrome: a systematic review. J Hand Ther 2004;17:152–64 29 Beaton DE, Katz JN, Fossel AH, Wright JG, Tarasuk V,
8 Van der Windt DA, Koes BW, de Jong BA, Bouter LM. Shoulder Bombardier C. Measuring the whole or the parts? Validity,
disorders in general practice: incidence, patient characteristics and reliability, and responsiveness of the disabilities of the arm,
management. Ann Rheum Dis 1995;54:959–64 shoulder and hand outcome measure in different regions of the
9 Kaltenborn F. Manual mobilization of the extremity joints. upper extremity. J Hand Ther 2001;14:128–46
Oslo: Olaf Norlis, 1989 30 Hay EM, Thomas E, Paterson SM, Dziedzic K, Croft PR. A
10 Le-T TV, Sheperd TM, Eathorte SW. Common shoulder pragmatic randomised controlled trial of local corticosteroid
problems: A hands-on approach to adhesive capsulitis. injection and physiotherapy for the treatment of new episodes
Consultant 2004;44:1481–5 of unilateral shoulder pain in primary care. Ann Rheum Dis
11 Baltaci G, Besler A, Bayrakcitunav V, Ergun N. The effect of 2003;62:394–9
manual therapy in the conservative management of 31 Emerson JD, Burdick E, Hoaglin DC, Mosteller F, Chalmers
impingement syndrome in shoulder. Joint Dis Relat Surg TC. An empirical study of the possible relation of treatment
2002;13:27–33 differences to quality scores in controlled randomized clinical
12 Fordyce WE, Fowler RSJ, Lehmann JF. Operant conditioning trials. Control Clin Trials 1990;11:339–52
in the treatment of chronic pain. Arch Phys Med Rehabil 32 Schulz KF, Chalmers I, Hayes RJ, Altman D. Empirical
1973;54:339–408 evidence of bias. JAMA 1995;273:408–12
13 Camplello M, Nordin M, Weiser S. Physical exercise and low- 33 Juni P, Witschi A, Bloch R, Egger M. The hazards of scoring
back pain: a review. Scand J Med Sci Sports 1996;6:63–72 the quality of clinical trials for meta-analysis. JAMA
14 Richardson CA, Jull GA. Muscle control – pain control. What 1999;282:1054–60
exercise would you prescribe? Man Ther 1995;1:2–10 34 Sherrington C, Herbert RD, Maher CG, Moseley AM.
15 Van der Heijden GJ. Shoulder disorders: a state of the art PEDro: a database of randomised trials and systematic reviews
review. Baillière Clin Rheumatol 1999;13:287–309 in physiotherapy. Man Ther 2000;5:223–6
16 Van der Heijden GJ, van der Windt DA, de Winter AF. 35 Kunz R, Oxman A. The unpredictability paradox: review of
Physiotherapy for patients with soft tissue shoulder disorders: empirical comparisons of randomised and non-randomised
a systematic review of randomised clinical trial. BMJ clinical trials. BMJ 1998;317:1185–90
1997;315:25–30 36 Verhagen AP, De Vet HCW, De Bie RA. Balneotherapy and
17 Philadelphia Panel. Evidence-based clinical practice guidelines quality assessment: interobserver reliability of the Maastricht
on selected rehabilitation interventions for shoulder pain. Phys criteria list for blinded quality assessment. J Clin Epidemiol
Ther 2001;81:1719–30 1998;51:335–41
18 Brox JI, Staff PH, Ljunggren AE, Brevik JI. Arthroscopic 37 Maher CG, Sherrington C, Herbert RD, Moseley AM, Elkins
surgery compared with supervised exercises in patients with M. Reliability of the PEDro scale for rating quality of
rotator cuff disease (stage II impingement syndrome). BMJ randomized controlled trials. Phys Ther 2003;83:713–21
1993;307:899–903 38 Busch AJ, Schachter CL, Peloso PM. Fibromyalgia and
19 Green S, Buchbinder R, Hetrick S. Physiotherapy interventions exercise training: a systematic review of randomized clinical
for shoulder pain. Cochrane Library. Issue 3, 2005 trials. Phys Ther Rev 2001;6:287–306
20 Bang MD, Deyle GD. Comparison of supervised exercise with 39 American College of Sports Medicine. ACSM’s guidelines for
and without manual physical therapy for patients with exercise testing and prescription, 6th edn. Philadelphia, PA:
shoulder impingement syndrome. J Orthop Sports Phys Ther Lippincot Williams and Wilkins, 2000
2000;30:126–37 40 Liddle SD, Baxter GD, Gracey JH. Exercise and chronic low-
21 Conroy DE, Hayes KW. The effect of joint mobilization as a back pain: what works? Pain 2004;107:176–90
component of comprehensive treatment of primary shoulder 41 Tugwell P, Brooks P, Wells G, Davies J, Shea B, Judd M et al.
impingement syndrome. J Orthop Sports Phys Ther Musculoskeletal Group. About The Cochrane Collaboration
1998;28:3–14 (Collaborative Review Groups (CRGs)) 2003, Issue 3. Art.
22 Sackett DL, Straus SE, Richardson WS, Rosenberg W, Haynes No.: MUSKEL
RB. Evidence-based Medicine. New York: Churchill Livingstone, 42 Higgins JPT, Green S. (eds) Cochrane Handbook for
2000 Systematic Reviews of Interventions 4.2.5 [updated May 2005].
23 Hsu AT, Ho L, Chang JH, Chang GL, Hedman T. In: The Cochrane Library. Issue 3, 2005. Chichester, UK: John
Characterization of tissue resistance during a dorsally directed Wiley
translational mobilization of the glenohumeral joint. Arch 43 Gross AR, Kay T, Hondras M, Goldsmith C, Haines T, Peloso
Phys Med Rehabil 2002;83:360–6 P et al. Manual therapy for mechanical neck disorders: a
142 TRAMPAS AND KITSIOS

systematic review. Man Ther 2002;7:131–49 48* Walther M, Werner A, Stahlschmidt T, Woelfel R, Gohlke F.
44 Van Tulder MW, Assendelft WJJ, Koes BW, Bouter LM, The subacromial impingement syndrome of the shoulder
Bombardier C, Nachemson AL et al. Method guidelines for treated by conventional physiotherapy, self-training and a
systematic reviews in the Cochrane Collaboration Back shoulder-brace: Results of a prospective, randomised study. J
Review Group for spinal disorders. Spine 1997;22:2323–30 Shoulder Elbow Surg 2004;17:417–23
45* Haahr JP, Ostergaard S, Dalsgaard J, Norup K, Frost P, 49* Johansson KM, Adolfsson LA, Foldevi MOM. Effects of
Lausen S et al. Exercises versus arthroscopic decompression in acupuncture versus ultrasound in patients with impingement
patients with subacromial impingement: a randomised, syndrome. Randomised clinical trial. Phys Ther
controlled study in 90 cases with a one year follow-up. Ann 2005;85:490–501
Rheum Dis 2005;64:760–4 50 Kaergaard A, Andersen JH, Rasmussen K, Mikelsen S.
46* Ludewig PM, Borstad JD. Effects of a home exercise program Identification of neck-shoulder disorders in an one year
on shoulder pain and functional status in construction follow-up study. Validation of a questionnaire-based method.
workers. Occup Environ Med 2003;60:841–9 Pain 2000;86:305–10
47* Citaker S, Taskiran H, Akdur H, Arabaci UO, Ekici G. 51 European Society for Surgery of the Shoulder and the Elbow
Comparison of the mobilization and PNF methods in the (2005) Assessment systems. Rome: European Society for
treatment of shoulder impingement syndrome. Pain Clin Surgery of the Shoulder and the Elbow. [updated 2005 Oct 24]
2005;17:197–202 Available from: <http://www.secec.org>

ATHANASIOS TRAMPAS MSc (for correspondence)


Physiotherapist, 25 Sakelaridi str 54248, Thessaloniki, Greece
Tel: +302310320804; E-mail: glossa2004@yahoo.com

ATHANASIOS KITSIOS PhD


Postgraduate Programme Leader, Department of Physical Education and Sports Science, Aristotelio University of Thessaloniki,
8 Tseliou str 54454, Thessaloniki, Greece
Tel: +302310992244; E-mail: a.kitsios@phed.auth.gr

You might also like