Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=iptp20

Physiotherapy Theory and PracticeAn International Journal of Physical Therapy

ISSN: 0959-3985 (Print) 1532-5040 (Online) Journal homepage: https://www.tandfonline.com/loi/iptp20

Scapular focused interventions to improveshoulder pain and function in adults withsubacromial pain: A systematic review and meta-analysis

Hiroki Saito, Meg E. Harrold, Vinicius Cavalheri & Leanda McKenna

To cite this article: Hiroki Saito, Meg E. Harrold, Vinicius Cavalheri & Leanda McKenna (2018)Scapular focused interventions to improve shoulder pain and function in adults with subacromialpain: A systematic review and meta-analysis, Physiotherapy Theory and Practice, 34:9, 653-670,DOI: 10.1080/09593985.2018.1423656

To link to this article: https://doi.org/10.1080/09593985.2018.1423656

View supplementary material

Published online: 19 Jan 2018.

Submit your article to this journal

Article views: 1622

View related articles

View Crossmark data

SYSTEMATIC REVIEW & META-ANALYSIS

Scapular focused interventions to improve shoulder pain and function in adultswith subacromial pain: A systematic review and meta-analysisHiroki Saito, BSc Physio, MClin Physio (Manips), Meg E. Harrold, BSc Physio, PhD, Vinicius Cavalheri, BSc Physio,MSc, PhD, and Leanda McKenna, B.App.Sc (Physio), Master of Sports Physiotherapy, PhD

School of Physiotherapy and Exercise Science, Faculty of Health Sciences, Curtin University of Technology, Perth, WA, Australia

ABSTRACTThe relationship between subacromial pain syndrome (SAPS) and altered scapular movement hasbeen previously reported. The purpose of this review was to determine the effect of interventionsthat focus on addressing scapular components to improve shoulder pain, function, shoulder rangeof motion (ROM), and muscle strength in adults with SAPS. Databases searched in September2016 were: PubMed, the Cochrane Central Register of Controlled Trials [Central], EMBASE [viaOvid] and PEDro. All studies selected for this review were randomized controlled trials. In total, sixstudies met the inclusion criteria and were included in the meta-analyses. In adults with SAPS,scapular focused interventions significantly improved pain with activities (MD [95% CI] = −0.88[−1.19 to −0.58], I2 43%) and shoulder function (−11.31 [−17.20 to −5.41] I2 65%) in the short term.No between-group difference in shoulder pain and function were found at follow up (4 weeks). Abetween-group difference in shoulder abduction ROM in the short term only was found (12.71[7.15 to 18.26]°, I2 36%). No between-group difference in flexion ROM, supraspinatus musclestrength, pectoralis minor length or forward shoulder posture were found. In conclusion, in adultswith SAPS, scapular focused interventions can improve short-term shoulder pain and function.

ARTICLE HISTORYReceived 12 December 2016Accepted 6 May 2017Revised 2 April 2017

KEYWORDSImpingement; randomizedcontrolled trials; scapula

Introduction

Shoulder pain is a symptom experienced by 7%–27% ofthe general population (Luime et al., 2004). This issuesubstantially contributes to lost working days and dis-ability (Virta, Joranger, Brox, and Eriksson, 2012). InSweden, the mean health care cost was €326 per patientper six months in 2009, and 60% of people withshoulder pain required physical therapy intervention(Virta, Joranger, Brox, and Eriksson, 2012). Of theseveral shoulder conditions that include shoulderpain, the most common is subacromial impingementsyndrome, accounting for 44%–65% of all complaintsof shoulder pain (Chard et al., 1991; Van der Windt,Koes, de Jong, and Bouter, 1995; Vecchio, Kavanagh,Hazleman, and King, 1995). This syndrome is tradi-tionally characterized by a narrowing of the subacro-mial space that causes encroachment of thesubacromial tissues such as the rotator cuff and sub-acromial bursa (Lewis, Green, and Wright, 2005; Neer,1972). However, “subacromial pain syndrome (SAPS)”may be considered more accurate terminology, giventhe inconsistencies in the traditional model (Diercks

et al., 2014; Lewis, 2011; Savoie et al., 2015). The tradi-tional term subacromial impingement suggests that thesurrounding bony architecture of the subacromial spaceis responsible for the development of the painful syn-drome. However, there does not appear to be a consis-tent relationship between acromial shape or acromio-humeral distance and symptoms (Balke et al., 2013; Gillet al., 2002; Michener et al., 2015; Moor et al., 2014;Worland et al., 2003; Yi et al., 2015) and surgicalapproaches to remove encroaching bone (i.e., to reduce“impingement”) are not more effective than conserva-tive approaches to manage symptoms (Dong et al.,2015; Page et al., 2016), suggesting that “impingement”is not the defining characteristic of this syndrome.Therefore, SAPS may include bursitis, biceps tendinitis,calcific tendinitis, supraspinatus tendinopathy, partialtear of the rotator cuff, and tendon cuff degeneration(Diercks et al., 2014; Lewis, 2011).

Altered scapular movement (dyskinesis) may be afactor associated with SAPS (Kibler et al., 2013). Priorresearch has shown that these altered movementsinclude decreased scapular external rotation, either

CONTACT Leanda McKenna, B.App.Sc (Physio), Master of Sports Physiotherapy, PhD L.McKenna@curtin.edu.au School of Physiotherapy andExercise Science, Faculty of Health Sciences, Curtin University of Technology, GPO Box U1987, Perth, WA 6845, Australia.Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/iptp.

The supplemental data for this article can be accessed here

PHYSIOTHERAPY THEORY AND PRACTICE2018, VOL. 34, NO. 9, 653–670https://doi.org/10.1080/09593985.2018.1423656

© 2018 Taylor & Francis

increased or decreased scapular upward rotation,decreased scapular posterior tilting, with alteredmotor control (i.e., decreased activity of serratus ante-rior, middle and lower trapezius and increased activityof the upper trapezius or levator scapuli muscle) (Coolset al., 2004; Lawrence, Braman, Laprade, and Ludewig,2014). Recently, two prospective studies (Reuther et al.,2014, 2015) on biomechanical properties of tendons inrats indicated that scapular motor control contributedto the alteration of tendon mechanical properties inrotator cuff muscles, which was consistent with tendonpathology. Therefore, it is postulated that scapular dys-kinesia could be a factor to consider with SAPS.

Physical therapy interventions for people with SAPScould aim to address pain, loss of range of motion, andmuscle function (Ellenbecker and Cools, 2010).Techniques can include exercise therapy, manual ther-apy, massage, electrotherapy, ultrasound treatment,laser treatment, extracorporeal shockwave therapy,and acupuncture (Desjardins-Charbonneau et al.,2015; Diercks et al., 2014; Kromer et al., 2009; Kuhn,2009). Systematic reviews have found that the evidenceto support the effectiveness of the majority of theseinterventions on the symptoms and impairments asso-ciated with SAPS is limited (Diercks et al., 2014;Kromer et al., 2009). It is worthwhile examining theeffect that scapular focused interventions may have inthose with SAPS, to provide information on whether itcould be used as an adjunct to non-scapular focusedinterventions. Two systematic reviews on the effective-ness of physical therapy interventions for people withSAPS have specifically compared whether interventionsthat focus on addressing scapular dysfunction are moreeffective than interventions that do not focus onaddressing scapular dysfunction to improve shoulderpain and function (Bury, West, Chamorro-Moriana,and Littlewood, 2016; Reijneveld et al., 2017).However, one review did not conduct a meta-analysis(Reijneveld et al., 2017), the reviews differed in whichclinical trials were included and did not examine short-and long-term time effectiveness, nor did they generatea pooled estimate of effect for scapular related out-comes (Bury, West, Chamorro-Moriana, andLittlewood, 2016; Reijneveld et al., 2017). Therefore,the aim of this systematic review was to investigatethe effectiveness of interventions that focus on addres-sing scapular components compared to interventionsthat do not focus on addressing scapular componentsto improve shoulder pain, function, shoulder range ofmotion (ROM) muscle strength and scapular outcomesin adults with SAPS. In particular, this review aims toconduct a prospectively registered systematic review,using the recommended databases (Michaleff et al.,

2011), while examining short- and medium-term timepoints and scapular outcomes.

Methods

Data sources and searches

Studies were identified by searching four main data-bases (PubMed, the Cochrane Central Register ofControlled Trials [Central], EMBASE [via Ovid] andPEDro). The search was conducted in September 2016,and databases were searched from their inception toSeptember 2016. The search terms that were used arepresented in the online Supplement 1. In addition,hand searching of abstracts from scientific meetings ofthe American Shoulder and Elbow Surgeons (ASES),European Society for Surgery of the Shoulder and theElbow (ESSSE), Shoulder and Elbow Society ofAustralia (SESA) was performed (from 2013 to 2015).This systematic review was prospectively registered atPROSPERO. Trial registration number wasCRD42015025949.

Study selection

This systematic review included randomized controlledtrials (RCTs) in which participants with shoulder paindue to SAPS were allocated to receive a physical ther-apy intervention that focused on addressing scapularcomponents and either an intervention that did notfocus on addressing scapular components or no inter-vention (online Supplement 2). Studies published inlanguages other than English, Japanese, Portuguese, orSpanish were excluded.

After removing duplicates, two reviewers (HS andLM) independently screened the titles, and abstracts ofall studies were identified. After this, disagreementsbetween the two reviewers (HS and LM) were resolvedby discussion with the other two reviewers (MH andVC). Two reviewers (HS, LM) then independentlyexamined the full text of each study to be included.Authors of potential studies were contacted to clarifydetails regarding inclusion criteria and interventions.

Inclusion criteria comprised studies that includedparticipants who were 18 years and older who hadSAPS. Studies had to indicate that the clinical diagnosiswas determined by: 1) shoulder pain; and 2) at least onepositive finding in the Jobe, Hawkins, Neer, full can orempty can tests (Cools, Cambier, and Witvrouw, 2008;Cools et al., 2003, 2004; Moraes, Faria, and Teixeira-Salmela, 2008).

Studies in participants who had any other primarydiagnosis (i.e., adhesive capsulitis, acromioclavicular

654 H. SAITO ET AL.

joint pathology, and cervical pathology) were excluded.Other exclusion criteria were studies that included par-ticipants who had: pain below the elbow; previouslyundergone surgery for SAPS; cardiac surgery includingpacemakers and device implantations; shoulder paindue to other serious medical conditions (e.g., cancer);neurological abnormalities (e.g., hemiplegia and neuro-pathy); bony lesions (e.g., osteoporosis and fracture);rheumatic diseases; and other specific reasons thatwould significantly impact on shoulder function (e.g.,use of manual wheelchairs).

Although there are a number of factors that influencescapular kinematics, scapular focused intervention wasdefined as interventions that focused on the scapu-lothoracic joint or muscles that move the scapula relativeto the thorax such as serratus anterior, trapezius, rhom-boids and levator scapulae. Studies had to include scap-ular focused interventions with ≥4 occasions of servicefor longer than 2 weeks. Interventions that targeted therotator cuff were excluded. Accepted interventions con-sisted of manual therapy, stretching, motor controltraining, ultrasound, laser, extracorporeal shockwavetherapy, and acupuncture. Comparison interventionswere either physical therapy interventions that did notfocus on addressing the scapulothoracic joint (i.e., gle-nohumeral mobilization and stretching, cervical inter-vention) or no intervention.

The primary outcome of this systematic review wasshoulder pain as measured by any scale or questionnaire(e.g., Visual analogue scale [VAS], Numerical Rating Scale[NRS], and pain subscale of the Shoulder Pain andDisability Index [SPADI]). Secondary outcomes com-prised: 1) shoulder function: Disabilities of the arm,shoulder, and hand (DASH) (Hudak, Amadio, andBombardier, 1996); SPADI score (Roach, Budiman-Mak, Songsiridej, and Lertranakul, 1991); the ShoulderDisability Questionnaire (SDQ) (Van der Windt et al.,1998); and the Western Ontario Rotator Cuff (WORC)index (Kirkley, Griffin, and Dainty, 2003); 2) shoulderrange of motion (ROM); and 3) shouldermuscle strength.Scapular associated outcomes such as pectoralis musclelength and forward shoulder posture were also examined.

Data extraction and quality assessment

Two reviewers (HS, VC) extracted data, and data waschecked for accuracy by two independent reviewers(MH, LM). After this, meta-analyses were conductedby the three reviewers (LM, VC, MH) for the primaryand secondary outcomes using Review Manager 5.3(Nordic Cochrane Centre, Copenhagen).

The Cochrane Collaboration’s “seven evidence-baseddomains” tables (i.e., random sequence generation,

allocation concealment, blinding of participants, blind-ing of outcome assessment, incomplete outcome data,selective reporting and other possible sources of bias)were used to assess risk of bias as low, high, or unclearfor each of the seven domains (Higgins and Green,2011). Quality or risk of bias was cross-checked byfour reviewers (HS, VC, MH, LM). Reporting bias wasassessed by searching protocols of RCTs published inclinical trials registries to determine whether all theoutcome measures described in the protocol werereported in the final study.

Data synthesis and analysis

The mean difference (MD) or standardized mean differ-ence (SMD) with their respective 95% confidence inter-vals (CI) were used in the meta-analyses of continuousoutcomes; MD for outcomes assessed using the samemeasurement scales and SMD for outcomes assessedusing different measurement scales across the includedstudies. Effect sizes with a 95% CI for between-groupdifferences of the outcome measures used were calcu-lated for each of the included studies. If median andinterquartile range (IQR) were provided, they were con-verted to mean and standard deviation (SD) (Hozo,Djulbegovic, and Hozo, 2005). When intervention andcontrol groups had similar measures at baseline, post-intervention values were used for the meta-analysis. Ifthere was a significant difference between groups atbaseline, and CIs were provided, changes from baselineto post-intervention were used for the meta-analysis.Where the SD of the change from baseline was notgiven, it was estimated based on 95% CIs. For standar-dization of pain scales to the VAS that ranges between 0and 10, the SPADI pain subscale was converted from 0and 100 to a scale between 0 and 10. For standardizationof shoulder function, the total score of the WORC wassubtracted from 100 in order to change the direction ofthe scale so that theWORC had the same direction of theother shoulder functional scales (DASH, SPADI, andSDQ). That is, for shoulder function scales, 0 denotednormal function and 100 indicated severe limitation.

Clinical heterogeneity within all included trials wasexamined and reported. Statistical heterogeneity wasinvestigated by visual assessment of forest plots andcalculation of the I2 statistic. The I2 provided an esti-mate of the percentage of variability due to heteroge-neity rather than to chance alone. Where the I2 estimate50% or more, this was treated as “substantial hetero-geneity” (Higgins and Green, 2011).

Meta-analyses were conducted in Review Manager5.3 using the inverse variance DerSimonian and Lairdmethod (DerSimonian and Laird, 1986) for the primary

PHYSIOTHERAPY THEORY AND PRACTICE 655

and secondary outcome measures. A random-effectmodel was utilized in order to calculate summary esti-mates. If the studies were shown to be homogeneous, afixed-effect model was applied.

Sensitivity analysis was performed to assess whetherstudy design affected the overall result of the meta-analysis. Two studies utilized a slightly different RCTdesign (Moezy, Sepehrifar, and Solaymani Dodaran,2014; Struyf et al., 2013); a direct comparison betweenscapular focused interventions versus non-scapularfocused interventions, whereas the remainder of thestudies investigated the addition of scapular focusedintervention to usual care (Aytar et al, 2015; Başkurt,Başkurt, Gelecek, and Özkan, 2011; Miller andOsmotherly, 2009; Shah, Sutaria, and Khant, 2014).Sensitivity analysis was performed to assess whetherintervention type (i.e., passive versus active interven-tion) affected the overall result of the meta-analysis.

Results

Flow of studies through the review

Figure 1 displays the study flow diagram. The initialsearch identified 655 records from all databases, and no

records were found in abstracts from scientific meet-ings. After removing duplicates, 479 studies were iden-tified for further assessment. Of these, 404 studies wereexcluded on the basis of title and abstract. Seventy-fivefull-texts were assessed for eligibility. Of these, 69 stu-dies were excluded and six RCTs, which were written inEnglish, were included in this review.

Characteristics of studies

All six studies (Aytar et al, 2015; Başkurt, Başkurt,Gelecek, and Özkan, 2011; Miller and Osmotherly,2009; Moezy, Sepehrifar, and Solaymani Dodaran,2014; Shah, Sutaria, and Khant, 2014; Struyf et al.,2013) only included participants with a diagnosis ofSAPS according to the inclusion criteria of this review.The total number of participants across the six RCTswas 250, ranging from 18 to 68 in each study. Themean or median age and gender distribution of eachstudy are described in Table 1.

The interventions utilized by the six included studiesvaried and included scapular mobilization, taping, mus-cle retraining and stretching (Table 1). Four studiesinvestigated the addition of scapular focused

Figure 1. Flow of studies through the review.

656 H. SAITO ET AL.

Table1.

Characteristicsof

includ

edstud

ies.

Stud

yParticipants

Interventio

nsOutcomemeasure

Resultmean

Between-grou

pMeanDifference

(MD)

(95%

CI)

Effect

size

(95%

CI)

Aytaret

al.

(2015)

Additive

stud

y

N=44

(Interventio

ngrou

p=22;C

ontrol

grou

p=22)

Age(yr)mean:

all=

52.0

(SD

4.0);Interventiongrou

p=52.0

(SD3.0);C

ontrol

grou

p=52.0

(SD4.0)

M/F:

All=

15/51(22.7/77.3);

Interventio

ngrou

p=4/18

(18.2/81.8);Co

ntrolg

roup

=8/

14(36.4/63.6)

Totalinterventiondu

ratio

nforallg

roup

swas

3weeks

with

atotalo

f9treatm

ent

sessions.

Interventio

ngrou

p=scapular

mob

ilizatio

n3

times

for10

repetitions

andarate

of1cycle

per6second

s,with

a30-secon

dinterval

betweensets.

Controlg

roup

=sham

scapular

mob

ilizatio

n3tim

esfor10

repetitions

andarate

of1

cycleper6second

swith

30-secon

dintervalsbetweensets.

VASwith

activity,Turkish

versionof

theQuick

DAS

H,sho

ulderRO

Mflexion

VASwith

activity

-Atthecompletion

Interventio

n:3.8(2.8)

Control:4.8(2.3)

-At4-weekfollow

upInterventio

n:3.0(2.8)C

ontrol:3.8(2.4)

DAS

H-Atthecompletion

Interventio

n:29.7

(18.6)

Control:25.6

(17.1)

-At4-weekfollow

upInterventio

n:19.5

(14.0)

Control:18.8

(13.5)

Shou

lder

flexion

ROM

-AtthecompletionInterventio

n:173.0

(10.0)

Control:171.0(9.0)

-At4-weekfollow

upInterventio

n:178.0(8.0)

Control:171.0(8.0)

VASwith

activity

-Atthecompletion:

MD:−

1.0(−2.5to

0.5)

Effect

size:

0.4(−1.0to

0.2)

-At4-weekfollow

up:−

0.8(−2.4to

0.8)

Effect

size:

-0.3

(−0.9to

0.3)

DAS

H-Atthecompletion:

4.1(−6.5to

14.7)

Effect

size:

0.2(−0.4to

0.8)

-At4-weekfollow

up:0

.7(−7.8to

9.2)

Effect

size:

0.1(−0.6to

0.7)

Shou

lder

flexion

ROM

-Atthecompletion:MD:2.0(−3.6to

7.6)

Effect

size:

0.2(−0.4to

0.8)

-At4-weekfollow

up:

MD:7

.0(2.0

to12.0)

Effect

size:

0.9(0.2

to1.5)

Baskurtet

al.

(2011)

Additive

stud

y

N=40

(Interventio

ngrou

p=20;C

ontrol

grou

p=20)

Age(yr):Intervention

grou

p=51.5

(SD8.4);C

ontrol

grou

p=20

51.3

(SD11.6)

Totalinterventiondu

ratio

nforallg

roup

swas

6weeks

with

atotalo

f18

treatm

ent

sessions

Interventio

ngrou

p=astandardized

flexibility,streng

theningandCo

dman

exercises+scapular

stabilizatio

nexercise

Controlg

roup

=astandardized

flexibility,

streng

theningandCo

dman

exercises

VASwith

activity,W

ORC

,sho

ulder

flexion

andabdu

ctionRO

M,

supraspinatusmusclestreng

th(kg)

VASwith

activity

-AtthecompletionInterventio

n:3.0

(1.6)

Control:3.2(2.1)

WORC

#

-AtthecompletionInterventio

n:17.4

(10.3)

Control:29.2

(19.7)

Shou

lder

flexion

ROM

-Atthecompletion

Interventio

n:179.8(1.1)C

ontrol:178.5

(4.6)

Shou

lder

abdu

ctionRO

M-Atthecompletion

Interventio

n:179.8(11.1)

Control:

177.0(13.4)

Supraspinatusmusclestreng

th-Atthecompletion

Interventio

n:11.6

(1.3)Co

ntrol:10.8

(1.6)

#=theWORC

was

subtracted

from

100

VASwith

activity

-Atthecompletion:

MD:−

0.2(−1.4to

1.0)

Effect

size:

-0.1

(−0.7to

0.5)

WORC

-Atthecompletion:

MD:-11.9(−21.6

to−2.2)

Effect

size:

-0.7

(−1.4to

−0.1)

Shou

lder

flexion

ROM

-Atthecompletion:MD:1.3(−0.8to

3.4)

Effect

size:

0.4(−0.3to

1.0)

Shou

lder

abdu

ctionRO

M-Atthecompletion:

MD:2

.8(−11.4

to17.0)

Effect

size:

0.1(−0.5to

0.7)

Supraspinatusmusclestreng

th-Atthecompletion:

MD:0

.80(−0.1to

1.7)

Effect

size:

0.5(−0.1to

1.2)

(Con

tinued)

PHYSIOTHERAPY THEORY AND PRACTICE 657

Table1.

(Con

tinued).

Stud

yParticipants

Interventio

nsOutcomemeasure

Resultmean

Between-grou

pMeanDifference

(MD)

(95%

CI)

Effect

size

(95%

CI)

Miller

and

Osm

otherly

(2009)

Additive

stud

y

N=22

(interventio

ngrou

p=12;C

ontrol

grou

p=10)

Age(yr)median:

interventio

ngrou

p=62.0

(IQR51.0–67.0);

controlg

roup

=54.5

(IQR

45.5–62.5)

M/F:Intervention=3/7;

Control=

7/5

Totalinterventiondu

ratio

nforallg

roup

swas

2weeks

with

atotalo

f6treatm

ent

sessions

Interventio

n=routineph

ysical

therapy

treatm

ent+taping

Control=

routineph

ysicaltherapy

interventio

ns

SPAD

Itotal,SPA

DIp

ain,

Shou

lder

flexion

andabdu

ctionRO

M(ASdata

wereprovided

inmedian

andIQR,

they

wereconvertedto

meanandSD

)

SPAD

Ipain&

-AtthecompletionInterventio

n:2.3

(0.5)Co

ntrol:4.05

(1.4)

-At4-weekfollow

upinterventio

n:1.6(0.7)C

ontrol:2.5(1.7)

SPAD

Itotal

-Atthecompletion

Interventio

n:21.0

(4.4)Co

ntrol:38.0

(12.0)

-At4-weekfollow

upinterventio

n:15.5

(9.1)Co

ntrol:24.6

(17.0)

Shou

lder

flexion

ROM

-Atthecompletion

Interventio

n:128.5(3.4)C

ontrol:130.8

(8.4)

-At4-weekfollow

upInterventio

n:142.3(6.3)C

ontrol:147.5

(5.5)

Shou

lder

abdu

ctionRO

M-Atthecompletion

Interventio

n:123.0(6.1)C

ontrol:114.0

(20.8)

-At4-weekfollow

upInterventio

n:127.5(5.76)

Control:119.5(20.2)

&=SPAD

Ipaindata

was

convertedto

scaleof

0to

10

SPAD

Ipain

-Atthecompletion:

MD:−

1.8(−2.7to

−0.9)

Effect

size:

-1.4

(−2.6to

−0.3)

-At4-weekfollow

up:

MD:−

0.9(−2.1to

0.3)

Effect

size:

-0.6

(−1.6to

0.4)

SPAD

Itotal

-Atthecompletion:

MD:−

16.5(−24.2to

−8.9)

Effect

size:

-1.5

(−2.7to

0.4)

-At4-weekfollow

up:

MD:−

9.1(−21.5

to3.3)

Effect

size:

-0.6

(−1.6to

0.4)

Shou

lder

flexion

ROM

-Atthecompletion:

MD:−

2.3(−7.8to

3.2)

Effect

size:

-0.3

(−1.3to

0.7)

-At4-weekfollow

up:

MD:−

5.2(−11.2

to0.8)

Effect

size:

-0.9

(−1.9to

0.2)

Shou

lder

abdu

ctionRO

M-Atthecompletion:

MD:9

.0(−3.7to

21.7)

Effect

size:

0.5(−0.5to

1.5)

-At4-weekfollow

upMD:8

.0(−0.5to

20.8)

Effect

size:

0.5(−0.6to

1.5)

(Con

tinued)

658 H. SAITO ET AL.

Table1.

(Con

tinued).

Stud

yParticipants

Interventio

nsOutcomemeasure

Resultmean

Between-grou

pMeanDifference

(MD)

(95%

CI)

Effect

size

(95%

CI)

Moezy,

Sepehrifar,

and

Solaym

ani

Dod

aran

(2014)

Direct

comparison

stud

y

N=72

(Interventio

ngrou

p=36)

Controlg

roup

=36)

Age(yr)mean:

Interventio

ngrou

p=48.2

(SD13.8);

Controlg

roup

=47.8

(SD7.8)

Totalinterventiondu

ratio

nforallg

roup

swas

6weeks

with

atotalo

f18

treatm

ent

sessions

Interventio

n=scapular

stabilizatio

nbased

exercise

Control=

conventio

nalp

hysiotherapy

(physicalm

odalities

andrang

eof

motion

exercise)

VASwith

activity,abd

uctio

nRO

M,

pectoralisminor

muscleleng

th,

forwardshou

lder

translation

VASwith

activity

-AtthecompletionInterventio

n:2.8

(2.2)

Control:3.1(2.1)

Shou

lder

abdu

ctionRO

M-Atthecompletion

Interventio

n:171.8(12.4)

Control:

155.7(16.3)

Pectoralisminor

muscleleng

th-AtthecompletionInterventio

n:12.0

(1.1)

Control:11.8

(1.2)

Forwardshou

lder

translation

-AtthecompletionInterventio

n:14.5

(1.3)

Control:14.7

(0.9)

VASwith

activity

-Atthecompletion

MD:−

0.30

(−1.32

to0.72)

Effect

size:

-0.1

(−0.6to

0.3)

Shou

lder

abdu

ctionRO

M-Atthecompletion

MD:1

6.1(9.24to

22.96)

Effect

size:

1.1(0.6

to1.6)

Pectoralisminor

muscleleng

th-Atthecompletion

MD:0

.2(−0.4to

0.7)

Effect

size:

0.2(−0.3to

0.6)

Forwardshou

lder

translation

-Atthecompletion

MD:−

0.3(−0.8to

0.3)

Effect

size:

-0.2

(−0.7to

0.2)

Shah,Sutaria,

andKh

ant

(2014)

Additive

stud

y

N=60

(Interventio

ngrou

p=30)

Controlg

roup

=30)

Age(yr)mean:

Interventio

ngrou

p=46.9;C

ontrol

grou

p=47.0

Totalinterventiondu

ratio

nforallg

roup

swas

4weeks

with

atotalo

f24

treatm

ent

sessions

Interventio

n=conventio

nalp

hysiotherapy

(strengthening

andstretching

prog

ram)+

scapular

stability

exercise

Control=

conventio

nalp

hysiotherapy

VASwith

activity,SPA

DI

VASwith

activity

-Atthecompletion

Interventio

n:4.0(0.7)

Control:4.8(0.8)

SPAD

I-Atthecompletion

Interventio

n:31.8

(6.2)Co

ntrol:46.1

(6.1)

VASwith

activity

-Atthecompletion

MD:−

0.8(−1.2to

−0.4)

Effect

size:

-1.1

(−1.6to

−0.5)

SPAD

I-Atthecompletion

MD:−

14.3

(−17.4

to−11.2)

Effect

size:

-2.3

(−3.0to

−1.6)

(Con

tinued)

PHYSIOTHERAPY THEORY AND PRACTICE 659

Table1.

(Con

tinued).

Stud

yParticipants

Interventio

nsOutcomemeasure

Resultmean

Between-grou

pMeanDifference

(MD)

(95%

CI)

Effect

size

(95%

CI)

Struyf

etal.

(2013)

Direct

comparison

stud

y

N=22

(Interventio

ngrou

p=12)

Controlg

roup

=10)

Age(yr)mean:

Interventio

ngrou

p=46.2

(SD13.5);

Controlg

roup

=45.4(SD15.1)

M/F:Interventiongrou

p=5/7;

Controlg

roup

=5/5

Totalinterventiondu

ratio

nforallg

roup

swas

4–8weeks

with

atotalo

f9treatm

ent

sessions

Interventio

n=scapular

dynamicstability

andpo

sitio

ning

Control=

exercise

therapy(rotator

cuff

streng

th)and

manualtherapy

(GHmob

ilizatio

nand

ultrasou

nd)

VASwith

activity,SDQtotal,

isom

etric

elevationstreng

th(N),

pectoralisminor

muscleleng

th,

forwardshou

lder

posture

VASwith

activity

-AtthecompletionInterventio

n:3.0

(1.9)

Control:5.1(2.0)

SDQ

-AtthecompletionInterventio

n:35.0

(14)

Control:48.7

(11.3)

Supraspinatusmuscle

€

-Chang

efrom

baselineInterventio

n:4.4(10.5)

Control:11.2

(34.3)

Pectoralisminor

muscleleng

th-AtthecompletionInterventio

n:10.3

(0.7)

Control:9.2(0.5)

Forwardshou

lder

posture

-AtthecompletionInterventio

n:0.40

(0.05)

Control:0.46

(0.05)

VASwith

activity

-Atthecompletion:

MD:−

2.1(−3.8to

−0.4)

Effect

size:

-1.0

(−2.0to

−0.1)

SDQ

-Atthecompletion:

MD:−

13.7(−24.9to

−2.6)

Effect

size:

-1.0

(−2.0to

−0.1)

Supraspinatusmusclestreng

th-Atcompletion:

MD:−

6.8(−29.0

to15.4)

Effect

size:

-0.3

(−1.1to

0.6)

Pectoralisminor

muscleleng

thMD:1

.1(−3.3to

5.5)

Effect

size:

0.2(−0.7to

1.1)

Forwardshou

lder

posture

MD:−

0.1(−0.1to

−0.0)

Effect

size:

-1.2

(−2.1to

−0.2)

€:data

ofisom

etric

elevationstreng

thwas

changesfrom

baseline;DAS

H:disability

ofthearm,sho

ulderandhand

;ER:externalrotatio

n;F:female;IR:internalrotation;M:m

ale;RO

M:range

ofmotion;SD

Q:the

shou

lder

disabilityqu

estio

nnaire

SE:scapu

larexercise;SM:scapu

larmob

ilizatio

n;SPAD

I:shou

lder

pain

anddisabilityindex;SSM:sham

scapular

mob

ilizatio

n;WORC

:Western

Ontario

RotatorCu

ff

660 H. SAITO ET AL.

intervention to usual care, that is, an additive design(Aytar et al., 2015; Başkurt, Başkurt, Gelecek, andÖzkan, 2011; Miller and Osmotherly, 2009; Shah,Sutaria, and Khant, 2014), whereas two studies directlycompared scapular focused intervention with non-scap-ular focused (i.e., a direct comparison design) (Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Struyf et al.,2013). Usual care provided to both groups in fourstudies included: hot packs (Aytar et al, 2015); transcu-taneous electrical nerve stimulation (TENS) (Aytaret al, 2015); glenohumeral capsular stretching(Başkurt, Başkurt, Gelecek, and Özkan, 2011; Struyfet al., 2013); rotator cuff strengthening (Başkurt,Başkurt, Gelecek, and Özkan, 2011; Shah, Sutaria, andKhant, 2014; Struyf et al., 2013); advice (Başkurt,Başkurt, Gelecek, and Özkan, 2011); education(Başkurt, Başkurt, Gelecek, and Özkan, 2011); soft tis-sue massage (Miller and Osmotherly, 2009); and jointmobilization techniques (Miller and Osmotherly, 2009).The control group in one of the additional designstudies (Aytar et al, 2015) was also allocated to receivesham scapular mobilization. Non-scapular interventionprovided to the control group of the two direct com-parison studies included: glenohumeral joint range ofmotion exercises (Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Struyf et al., 2013); infrared therapy(Moezy, Sepehrifar, and Solaymani Dodaran, 2014);ultrasound therapy (Struyf et al., 2013); TENS(Moezy, Sepehrifar, and Solaymani Dodaran, 2014);rotator cuff strengthening (Struyf et al., 2013); frictionalmassage (Struyf et al., 2013); and a home exerciseprogram (Struyf et al., 2013). Intervention durationvaried (Table 1) from 2 weeks (Miller andOsmotherly, 2009) to 8 weeks (Struyf et al., 2013) andtotal intervention dosage from 6 sessions (Miller andOsmotherly, 2009) to 24 sessions (Shah, Sutaria, andKhant, 2014). There appears to be no relationshipbetween intervention duration or dosage in comparisonto mean differences and effect sizes, at the cessation ofintervention.

Pain with activity was assessed in five studies (Aytaret al, 2015; Başkurt, Başkurt, Gelecek, and Özkan, 2011;Miller and Osmotherly, 2009; Moezy, Sepehrifar, andSolaymani Dodaran, 2014; Struyf et al., 2013), with onestudy unclearly reporting whether pain was assessedduring activity (Shah, Sutaria, and Khant, 2014).Shoulder function (Quick DASH scale, SPADI, SDQ,and WORC) was assessed in five studies (Aytar et al,2015; Başkurt, Başkurt, Gelecek, and Özkan, 2011;Miller and Osmotherly, 2009; Shah, Sutaria, andKhant, 2014; Struyf et al., 2013). One study (Millerand Osmotherly, 2009), presented median and inter-quartile range for the SPADI, and the data was

subsequently transformed to mean and standard devia-tion. Shoulder flexion ROM was assessed in three stu-dies (Aytar et al, 2015; Başkurt, Başkurt, Gelecek, andÖzkan, 2011; Miller and Osmotherly, 2009). Shoulderabduction ROM was assessed in three studies (Başkurt,Başkurt, Gelecek, and Özkan, 2011; Miller andOsmotherly, 2009; Moezy, Sepehrifar, and SolaymaniDodaran, 2014). Supraspinatus muscle strength wasassessed in two studies (Başkurt, Başkurt, Gelecek,and Özkan, 2011; Struyf et al., 2013). Several differentscapular outcomes were assessed in the studies andincluded: serratus anterior strength (Başkurt, Başkurt,Gelecek, and Özkan, 2011); pectoralis muscle length(Moezy, Sepehrifar, and Solaymani Dodaran, 2014;Struyf et al., 2013); static positional symmetry (Moezy,Sepehrifar, and Solaymani Dodaran, 2014); acromionposition (Moezy, Sepehrifar, and Solaymani Dodaran,2014; Struyf et al., 2013); scapular motor control (Struyfet al., 2013); habitual scapular posture (Başkurt,Başkurt, Gelecek, and Özkan, 2011; Moezy, Sepehrifar,and Solaymani Dodaran, 2014; Shah, Sutaria, andKhant, 2014; Struyf et al., 2013); rotation (Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Struyfet al., 2013); and protraction (Moezy, Sepehrifar, andSolaymani Dodaran, 2014). Pectoralis minor length wasnormalized to height in one study (Struyf et al., 2013)and given as raw data in the other study (Moezy,Sepehrifar, and Solaymani Dodaran, 2014). Acromionposition was measured from the posterior aspect of theacromion in supine and normalized to height in onestudy (Struyf et al., 2013) and from the anterior acro-mion in standing in the other study (Moezy, Sepehrifar,and Solaymani Dodaran, 2014). Habitual scapular posi-tion was determined by visual evaluation (Struyf et al.,2013), distance of the spine to the scapula at the level ofthe inferior angle (Shah, Sutaria, and Khant, 2014), thedifference between sides for the spine to scapula dis-tance (Başkurt, Başkurt, Gelecek, and Özkan, 2011),and the spine to scapular distance, normalized forscapular size (Moezy, Sepehrifar, and SolaymaniDodaran, 2014). These measurements were not consid-ered to be clinically heterogeneous, and thus a pooledestimate of effect was not generated. Scapular rotationwas measured with an inclinometer in one study(Struyf et al., 2013), and with bony landmarks inanother (Moezy, Sepehrifar, and Solaymani Dodaran,2014). While measurement of scapular rotation with aninclinometer demonstrates acceptable clinometricproperties, measurement with bony landmarks doesnot (Larsen, Juul-Kristensen, Lund, and Sogaard,2014), and thus a pooled estimate of effect was notgenerated. A pooled estimate of effect was generatedfor pectoralis minor length and acromion position.

PHYSIOTHERAPY THEORY AND PRACTICE 661

Data were extracted at two time points: 1) immediatelyafter completion of intervention (that ranged between 2and 8 weeks across studies) (Aytar et al, 2015; Başkurt,Başkurt, Gelecek, and Özkan, 2011; Miller andOsmotherly, 2009; Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Shah, Sutaria, and Khant, 2014; Struyfet al., 2013); and 2) 4 weeks following completion of theintervention period (Aytar et al, 2015; Miller andOsmotherly, 2009).

Risk of bias in included studies

A summary of risk of bias across the included studies isshown in Figure 2. Four studies (Aytar et al, 2015;Başkurt, Başkurt, Gelecek, and Özkan, 2011; Millerand Osmotherly, 2009; Moezy, Sepehrifar, andSolaymani Dodaran, 2014) were assessed as havinglow risk of bias on randomization (low risk on selectionbias), while two studies (Shah, Sutaria, and Khant,2014; Struyf et al., 2013) did not report sufficient detailon this procedure (unclear risk on selection bias). Two

studies (Miller and Osmotherly, 2009; Struyf et al.,2013) reported adequate allocation concealment (lowrisk on selection bias). In one study (Aytar et al, 2015),despite the use of the online random allocation soft-ware, the treating physical therapist was responsible forallocation, which suggests concealment was notadhered to (high risk on selection bias). Three otherstudies (Başkurt, Başkurt, Gelecek, and Özkan, 2011;Moezy, Sepehrifar, and Solaymani Dodaran, 2014;Shah, Sutaria, and Khant, 2014) did not report suffi-cient detail on allocation concealment (unclear risk onselection bias).

Five studies (Aytar et al, 2015; Miller andOsmotherly, 2009; Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Shah, Sutaria, and Khant, 2014; Struyfet al., 2013) were assessed as having a high risk bias onblinding of personnel delivering the intervention (highrisk on performance bias), while one study (Başkurt,Başkurt, Gelecek, and Özkan, 2011) did not reportsufficient information (unclear risk on performancebias). In four studies (Aytar et al, 2015; Miller andOsmotherly, 2009; Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Struyf et al., 2013), blinding of outcomeassessors occurred and thus was considered as low risk(low risk on performance bias). Two studies (Başkurt,Başkurt, Gelecek, and Özkan, 2011; Shah, Sutaria, andKhant, 2014) did not report sufficient information onblinding outcome assessors (unclear risk on perfor-mance bias).

Three studies (Aytar et al, 2015; Moezy, Sepehrifar,and Solaymani Dodaran, 2014; Struyf et al., 2013) wereassessed as having low risk of bias for incomplete out-come data (low risk on attrition bias), while one study(Miller and Osmotherly, 2009) was assessed as havinghigh risk of bias, due to high dropout rates in theintervention group (high risk on attrition bias). Twostudies (Başkurt, Başkurt, Gelecek, and Özkan, 2011;Shah, Sutaria, and Khant, 2014) did not report suffi-cient information (unclear risk of attrition bias).

Outcome: Shoulder pain

Five studies (Aytar et al, 2015; Başkurt, Başkurt, Gelecek,and Özkan, 2011; Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Shah, Sutaria, and Khant, 2014; Struyfet al., 2013) measured pain with activities using a VAS,while one study (Miller and Osmotherly, 2009) utilizedthe SPADI pain scale. Data are presented as mean differ-ence on a 0 to 10 scale. Based on the pooled data from thesix studies (Aytar et al, 2015; Başkurt, Başkurt, Gelecek,and Özkan, 2011; Miller and Osmotherly, 2009; Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Shah, Sutaria,and Khant, 2014; Struyf et al., 2013), on completion ofFigure 2. Risk of bias of included studies.

662 H. SAITO ET AL.

intervention, compared to the control group, pain duringactivity was significantly less in the group that receivedscapular focused interventions (MD [95% CI] = −0.88[−1.19 to −0.58], I2 43%) with no substantial heterogeneity(Figure 3A). Sensitivity analysis demonstrated that theexclusion of the two direct comparison studies (Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Struyf et al.,2013) did not change the significance of the overall result(MD [95% CI] = -.90 [−1.22 to −0.57], I2 = 47%).Sensitivity analysis demonstrated that a meta-analysisconducted at 0 weeks using passive intervention studies(Aytar et al, 2015; Miller and Osmotherly, 2009) yielded amore favorable point estimate for shoulder pain (MD[95% CI] = −1.60 [−2.36 to −0.83], I2 = 0%) than ameta-analysis using only active intervention studies(Başkurt, Başkurt, Gelecek, and Özkan, 2011; Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Shah, Sutaria,and Khant, 2014; Struyf et al., 2013) (MD [95% CI] = −.75[−1.08 to −0.41], I2 = 26%). Pooled data from two studies(Aytar et al, 2015; Miller and Osmotherly, 2009) demon-strated no significant difference in pain at 4 weeks follow-ing completion of intervention (MD [95% CI] = −0.87[−1.80 to 0.07], I2 = 0%), with no substantial heterogeneity(Figure 3B).

Outcome: Shoulder function

Each study measured shoulder function using differentmeasurement scales. However, data are presented asmean difference on a 0–100 scale. Based on the pooleddata from the five studies (Aytar et al, 2015; Başkurt,Başkurt, Gelecek, and Özkan, 2011; Miller andOsmotherly, 2009; Shah, Sutaria, and Khant, 2014;Struyf et al., 2013), on the completion of intervention,compared to the control group, shoulder function was

significantly better in the group that received scapularfocused interventions (MD [95% CI] = −11.31 [−17.20to −5.41] I2 65%) with substantial heterogeneity(Figure 4A). Sensitivity analysis demonstrated that theexclusion of the direct comparison study (Struyf et al.,2013) did not change the significance of the overall result(MD [95% CI] = −10.66 [- 17.79 to −3.54], I2 = 74%).Sensitivity analysis demonstrated that a meta-analysisconducted at 0 weeks using passive intervention studies(Aytar et al, 2015; Miller and Osmotherly, 2009) yielded aless favorable point estimate for shoulder function (MD[95% CI] = −6.53 [−26.21 to 13.64], I2 = 90%) than ameta-analysis using only active intervention studies(Başkurt, Başkurt, Gelecek, and Özkan, 2011; Shah,Sutaria, and Khant, 2014; Struyf et al., 2013) (MD [95%CI] = −14.05 [−16.92 to −11.19], I2 = 0%). Pooled datafrom two studies (Aytar et al, 2015; Miller andOsmotherly, 2009) demonstrated no significant differencein shoulder function at 4 weeks following completion ofintervention (MD [95% CI] = −3.12 [−12.49 to 6.25],I2 = 39%), with no substantial heterogeneity (Figure 4B).

Outcome: Shoulder flexion ROM

Based on pooled data from three studies (Aytar et al,2015; Başkurt, Başkurt, Gelecek, and Özkan, 2011;Miller and Osmotherly, 2009), on the completion ofintervention, compared to the control group, shoulderflexion ROM was not significantly different in thegroup that received scapular focused interventions(MD [95% CI] = 0.97 [−0.86 to 2.80], I2 = 0%), withno substantial heterogeneity. Pooled data from twostudies (Aytar et al, 2015; Miller and Osmotherly,2009), demonstrated no significant difference inshoulder flexion ROM at 4 weeks following completion

Figure 3. Mean difference in shoulder pain with activities A) on completion of intervention and B) at 4-week follow up.

PHYSIOTHERAPY THEORY AND PRACTICE 663

of intervention (MD [95% CI] = 1.02 [−10.93 to 12.97],I2 = 89%), with substantial heterogeneity (onlineSupplement 3; Figures 3.1 and 3.2).

Outcome: Shoulder abduction ROM

Based on the pooled data from three studies (Başkurt,Başkurt, Gelecek, and Özkan, 2011; Miller andOsmotherly, 2009; Moezy, Sepehrifar, and SolaymaniDodaran, 2014), on the completion of intervention,compared to the control group, shoulder abductionROM was significantly different to the group thatreceived scapular focused interventions (MD [95%

CI] = 12.71 [7.15–18.26], I2 = 36%), with no substantialheterogeneity (Figure 5A). Sensitivity analysis demon-strated that the exclusion of the direct comparisonstudy (Moezy, Sepehrifar, and Solaymani Dodaran,2014) changed the significance but did not change thedirection of the overall result (MD [95% CI] = 6.23[−3.24 to 15.71], I2 = 0%). Data at 4 weeks follow upwas not available in included studies.

Outcome: Supraspinatus muscle strength

Based on the pooled data from two studies (Başkurt,Başkurt, Gelecek, and Özkan, 2011; Struyf et al., 2013),

Figure 4. Mean difference in shoulder function A) on completion of intervention and B) at 4-week follow up.

Figure 5. Mean difference on completion of intervention in A) shoulder abduction range of motion and B) supraspinatus musclestrength.

664 H. SAITO ET AL.

on the completion of intervention, compared to thecontrol group, supraspinatus muscle strength was notsignificantly difference to the group that received scap-ular focused interventions (MD [95% CI] = −0.20[−0.57 to 0.97], I2 = 52%), with substantial heterogene-ity (Figure 5B). Data at 4 weeks follow up was notavailable in included studies.

Outcome: Pectoralis minor muscle length

Based on the pooled data from two studies (Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Struyfet al., 2013), on the completion of intervention, com-pared to the control group, Pectoralis Minor musclelength was not significantly different to the group thatreceived scapular focused interventions (MD [95%CI] = 0.18 [−0.24 to 0.59], I2 = 0%), with no substantialheterogeneity (online Supplement 3; Figure 3.3). Dataat 4 weeks follow up was not available in includedstudies.

Outcome: Forward shoulder posture

Based on the pooled data from two studies (Moezy,Sepehrifar, and Solaymani Dodaran, 2014; Struyfet al., 2013), on the completion of intervention, com-pared to the control group, forward shoulder posturewas not significantly different to the group thatreceived scapular focused interventions (MD [95%CI] = −0.60 [−1.46 to 0.27], I2 = 64%), with substantialheterogeneity (online Supplement 3; Figure 3.4). Dataat 4 weeks follow up was not available in includedstudies.

Discussion

This systematic review suggests that scapular focusedinterventions are a beneficial adjunct to usual interven-tions that do not focus on addressing scapular compo-nents to improve pain with activities (MD [95%CI] = −0.88 [−1.19 to −0.58]), shoulder function (MD[95% CI] = −11.31 [−17.20 to −0.54]), and abductionROM (MD [95% CI] = 12.71 [7.15 to 18.26] º), inadults with SAPS, in the short term. No between-group differences were found after 4 weeks of followup. No between-group differences were found inshoulder flexion ROM, supraspinatus muscle strength,pectoralis minor length or forward shoulder posture atboth completion of interventions and 4 weeks followup. The scapular focused interventions varied acrossthe six studies included in this review (Table 1).Intervention groups were provided with: scapularmobilization (Aytar et al., 2015); stabilization including

strengthening, or retraining (Başkurt, Başkurt, Gelecek,and Özkan, 2011; Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Shah, Sutaria, and Khant, 2014; Struyfet al., 2013); and taping (Miller and Osmotherly, 2009);the control groups received: electrophysical modalities(Aytar et al, 2015; Moezy, Sepehrifar, and SolaymaniDodaran, 2014; Struyf et al., 2013); glenohumeralmobilization techniques (Başkurt, Başkurt, Gelecek,and Özkan, 2011; Miller and Osmotherly, 2009;Struyf et al., 2013); massage (Struyf et al., 2013);sham scapular mobilization (Aytar et al, 2015); rotatorcuff strengthening (Başkurt, Başkurt, Gelecek, andÖzkan, 2011; Shah, Sutaria, and Khant, 2014; Struyfet al., 2013); glenohumeral joint ROM exercises(Moezy, Sepehrifar, and Solaymani Dodaran, 2014;Struyf et al., 2013); advice (Başkurt, Başkurt, Gelecek,and Özkan, 2011); education (Başkurt, Başkurt,Gelecek, and Özkan, 2011); and a home exercise pro-gram (Struyf et al., 2013). Duration and dosage ofscapular intervention appeared to have no influenceon mean differences or effect size (Table 1) in theshort term.

The study (Miller and Osmotherly, 2009) thatreported the largest effect size for decrease in pain(−1.4) utilized the simple intervention of adding scap-ular taping for 2 weeks to a program of conventionalphysical therapy that included soft tissue massage, jointmobilization, stretches, rotator cuff exercises and gen-eralized strengthening. More involved treatments thatincluded more than seven different scapular exercisesto increase isometric and multiplanar lower trapeziusand serratus anterior function (Shah, Sutaria, andKhant, 2014), and with scapular mobilization and leva-tor scapulae, rhomboids and pectoralis minor stretches(Struyf et al., 2013) yielded slightly smaller effect sizesfor pain of −1.1 and −1.0, respectively. However, insome studies, involved treatments such as thosedescribed above also yield small effect sizes as low as−0.1 (Başkurt, Başkurt, Gelecek, and Özkan, 2011;Moezy, Sepehrifar, and Solaymani Dodaran, 2014),indicating that perhaps scapular taping may be themost valuable scapular focused addition to an interven-tion program to address subacromial pain syndrome.With regard to function, the intervention that yieldedthe largest effect size of 2.3 (Shah, Sutaria, and Khant,2014) utilized the more involved scapular focused inter-vention that aimed to increase isometric and multi-planar lower trapezius and serratus anterior function.Taping yielded the second best improvement in func-tion, with an effect size of 1.5 (Miller and Osmotherly,2009). The more involved scapular focused interventionthat aimed to increase isometric and multiplanar lowertrapezius and serratus anterior function (Başkurt,

PHYSIOTHERAPY THEORY AND PRACTICE 665

Başkurt, Gelecek, and Özkan, 2011), and scapularmobilization and levator scapulae, rhomboids and pec-toralis minor stretches (Struyf et al., 2013) providedeffect sizes of 0.7 and 1.0, respectively, for function.The study that added only scapular mobilization tothe interventions yielded the poorest effect size of 0.2for function (Aytar et al, 2015).Thus, it would seemthat for improvement of SAPS function in the shortterm, the addition of scapular mobilization to interven-tion is not worthwhile.

Two previous systematic reviews have provided cau-tious support for the use of scapular focused interven-tions for SAPS but were unable to reach a firmconclusion (Bury, West, Chamorro-Moriana, andLittlewood, 2016; Reijneveld et al., 2017). This is inpart because one review did not conduct a meta-analy-sis (Reijneveld et al., 2017), and the other included only3–4 studies in their meta-analysis (Bury, West,Chamorro-Moriana, and Littlewood, 2016). The cur-rent review and a previous review (Bury, West,Chamorro-Moriana, and Littlewood, 2016) both founda significant decrease in pain and shoulder functionwith scapular focused interventions. No pooled esti-mates for other outcomes are available for comparisonfrom the previous reviews, as this current review is thefirst to conduct a meta-analysis on the effect of scapularfocused interventions on shoulder ROM, rotator cuffstrength, or scapular associated outcomes.

The key strengths of the current review were that acomprehensive searching method that used the recom-mended multiple databases (Michaleff et al., 2011) wasutilized, and various scapular outcomes at differenttime points were examined. Additionally, fourreviewers were involved in the process of independentstudy identification, data extraction, and quality assess-ment, minimizing possible bias and error in the sys-tematic review findings. However, the findings of thecurrent systematic review should be interpreted withcaution, as included studies had several limitations.First, meta-analyses could only be conducted for out-come measures assessed on completion of interventionand at 4 weeks follow up, which were the most com-mon time points across the included studies. Second, at4 weeks follow up, meta-analyses could only be con-ducted on three outcome measures (Aytar et al, 2015;Miller and Osmotherly, 2009) (i.e., pain with activities;shoulder function; and flexion range of motion) asthese were the common outcomes for that time pointacross the included studies. However, this currentreview remains the only review that has examined afollow-up time point. Third, two studies were excludedfrom the review despite meeting inclusion criteria asthe original authors did not have access to the missing

data required or did not reply (Polimeni et al., 2003;Surenkok, Aytar, and Baltaci, 2009). Another study wasexcluded from review when communication from theauthors indicated that it was not a randomized trial andthus did not meet inclusion criteria (Yuksel andYesilyaprak, 2015). Last, the majority of included stu-dies had evidence of biases that may have impacted onthe results. Although most of the included studies hadlow risk of bias on randomization, selection bias due toallocation concealment was considered as high orunclear in four of six studies.

The findings of this review indicate that scapularfocused interventions provide a statistically significantdifference; however, they may not offer a clinical dif-ference when compared to the minimal clinicallyimportance difference (MCID). A previous study hasfound that the MCID for improvement in shoulderpain with activities using the VAS was −1.4 cm(Tashjian, Deloach, Porucznik, and Powell, 2009).This MCID for shoulder pain is marginally largerthan the difference found by this current systematicreview. Additionally, previous studies have found thatthe MCIDs for improvement in shoulder function inthe four questionnaires of interest in this systematicreview range between 10.2 and 24.5 (Angst et al.,2011; Smith et al., 2012). These MCIDs for shoulderfunction range from marginally less than to substan-tially larger than the difference found by this currentsystematic review. Therefore, the current systematicreview suggests that in adults with SAPS, short-termimprovements in shoulder pain with activities andshoulder function following scapular focused interven-tions are statistically significant but may not be clini-cally important. There may be two reasons why thesummarized effect difference was small. First, thiscould be due to differences in study design. However,sensitivity analysis that excluded the two studies withdiffering design did not impact on the significance ofthe meta-analysis results. Second, the majority of stu-dies commonly applied an intervention without initiallydetermining that it was a required intervention; forexample, scapular stabilization exercises were providedwithout initially determining if scapular stability wasdeficient. Therefore, the actual effect size of scapularfocused interventions may be larger than found in thissystematic review.

Although on completion of the intervention, painwith activities and shoulder function improved signifi-cantly in the group that received scapular focused inter-vention compared to the control group, at 4 weeksfollow up, there was no between-group difference inthese outcomes. There are two possible reasons for atemporal difference. Firstly, the interventions (Table 1)

666 H. SAITO ET AL.

provided to participants of these two studies werescapular mobilization (Aytar et al., 2015) or scapulartaping (Miller and Osmotherly, 2009). It is suggestedthat these interventions are passive interventions ratherthan exercise-based interventions, and their effective-ness is not maintained in the long term (Desjardins-Charbonneau et al., 2015). The sensitivity analysis thatexamined passive versus active interventions suggestedthat active interventions provide a more favorable out-come than passive interventions for shoulder function(although not for shoulder pain). Previous systematicreviews have demonstrated that maintenance ofimprovements following treatment favored exercise-based physical therapy interventions compared to con-trol groups that received either inactive ultrasound orshockwave therapy (Dong et al., 2015; Hanratty et al.,2012). Secondly, the duration of interventions acrossthe two studies that assessed participants at 4 weeksfollow up was short, ranging between 2 and 3 weeks.Prior studies have demonstrated that improvements inpain and shoulder function in people with SAPS weremaintained in the long term, provided physical therapyinterventions were at least 8–12 weeks long (Holmgrenet al., 2014; Lombardi et al., 2008). A dose responseeffect has been demonstrated by a prior RCT thatshowed significant differences in shoulder pain andfunction in participants who received a high dose ofexercise, compared to those that received a low dose(Østerås, Torstensen, and Østerås, 2010). For clinicians,this may indicate that active rather than passive inter-vention of long duration and high dose scapularfocused interventions are preferable to maintainimprovements long term, and this should be confirmedwith further research.

There was no significant between-group differencein shoulder flexion ROM and supraspinatus musclestrength at any time point. These findings are consis-tent with previous systematic reviews that found non-specific shoulder exercise programs do not demonstrateimprovements in ROM (Marinko, Chacko, Dalton, andChacko, 2011) and only small improvements in musclestrength (Hanratty et al., 2012) for people with SAPS.Findings of the current review and that of previousreviews (Hanratty et al., 2012; Marinko, Chacko,Dalton, and Chacko, 2011) therefore suggest that theeffectiveness of scapular focused interventions toreduce pain and improve shoulder function is unlikelyto be due to increased shoulder ROM or improvedsupraspinatus muscle strength. It is unknown whetherglenohumeral ROM and strength outcomes are specificenough to detect the effectiveness of scapular focusedinterventions (De Mey, Danneels, Cagnie, and Cools,2012; Hsu et al., 2009; Worsley et al., 2013). For

instance, two recent case studies (De Mey, Danneels,Cagnie, and Cools, 2012; Worsley et al., 2013) that usedsurface electromyography (EMG) and a motion capturesystem demonstrated that scapular focused interven-tions improved pain and shoulder function in peoplewith SAPS, and these improvements were related tochanges in scapular muscle recruitment (i.e., normal-izing trapezius and serratus anterior activities) andscapular kinematics (i.e., increasing scapular upwardrotation and posterior tilt) during dynamic tasks.These effects may not translate to a large effect onglenohumeral ROM or muscle strength.

Although four of the included studies utilizedscapular outcomes, only two of these outcomeswere considered similar enough to pool data, witha large variety of outcomes collected by the studies.A recent systematic review (Larsen, Juul-Kristensen,Lund, and Sogaard, 2014) identified 54 positionaland functional scapular outcome measures thathave been used in prior research, highlighting thelack of consistency of methodologies for measuringscapular positioning and function. This may indicatethe underlying multifactorial complex role that thescapula plays in the development of shoulder painand subsequent lack of consensus regarding whatscapular outcomes should be measured in RCTs.None of the studies included in the current reviewinvestigated muscle recruitment using EMG orassessed scapular kinematics using a motion capturesystem. Future RCTs should closely consider whichscapular measurement outcomes to use that areclinically reliable, valid, and responsive, in additionto investigating the mechanisms by which scapularfocused interventions may improve pain andshoulder function in adults with SAPS.

In conclusion, this systematic review suggests thatphysical therapy interventions for adults with SAPSshould consider adding interventions that focus onaddressing scapular components in order to improveshoulder pain, function, and abduction ROM. Thisreview suggests that maintenance of symptom relief islimited after cessation of scapular focused interventionsand that future research that examines the influences oftype (active versus passive) and dose of scapular inter-vention is necessary, to determine if symptom relief canbe maintained. This review also suggests that on com-pletion of intervention, shoulder flexion ROM andsupraspinatus muscle strength are similar between peo-ple who undergo interventions that do and do notaddress scapular components. There is a need for lar-ger, high-quality RCTs that adequately examine scapu-lar outcomes that are of interest to the clinician andstrengthen these findings.

PHYSIOTHERAPY THEORY AND PRACTICE 667

Declaration of interest

The authors report no conflicts of interest.

References

Angst F, Schwyzer HK, Aeschlimann A, Simmen BR, GoldhahnJ 2011Measures of adult shoulder function: Disabilities of theArm, Shoulder, and Hand Questionnaire (DASH) and itsshort version (QuickDASH), Shoulder Pain and DisabilityIndex (SPADI), American Shoulder and Elbow Surgeons(ASES) Society standardized shoulder assessment form,Constant (Murley) Score (CS), Simple Shoulder Test (SST),Oxford Shoulder Score (OSS), Shoulder DisabilityQuestionnaire (SDQ), and Western Ontario ShoulderInstability Index (WOSI). Arthritis Care and Research 63(Suppl 11): S174–S188.

Aytar A, Baltaci G, Uhl T, Tuzun H, Oztop P, Karatas M2015 The effects of scapular mobilization in patients withsubacromial impingement syndrome: A randomized, dou-ble-blind, placebo-controlled clinical trial. Journal of SportRehabilitation 24: 116–129.

Balke M, Schmidt C, Dedy N, Banerjee M, Bouillon B, LiemD 2013 Correlation of acromial morphology with impinge-ment syndrome and rotator cuff tears. Acta Orthopaedica84: 178–183.

Başkurt Z, Başkurt F, Gelecek N, Özkan MH 2011 The effec-tiveness of scapular stabilization exercise in the patientswith subacromial impingement syndrome. Journal of Backand Musculoskeletal Rehabilitation 24: 173–179.

Bury J, West M, Chamorro-Moriana G, Littlewood C 2016Effectiveness of scapula-focused approaches in patientswith rotator cuff related shoulder pain: A systematicreview and meta-analysis. Manual Therapy 25: 35–42.

Chard MD, Hazleman R, Hazleman BL, King RH, Reiss BB1991 Shoulder disorders in the elderly: A community sur-vey. Arthritis and Rheumatism 34: 766–769.

Cools AM, Cambier D, Witvrouw EE 2008 Screening theathlete’s shoulder for impingement symptoms: A clinicalreasoning algorithm for early detection of shoulder pathol-ogy. British Journal of Sports Medicine 42: 628–635.

Cools AM, Witvrouw EE, Declercq GA, Danneels LA, CambierDC 2003 Scapular muscle recruitment patterns: trapeziusmuscle latency with and without impingement symptoms.American Journal of Sports Medicine 31: 542–549.

Cools AM, Witvrouw EE, Declercq GA, Vanderstraeten GG,Cambier DC 2004 Evaluation of isokinetic force produc-tion and associated muscle activity in the scapular rotatorsduring a protraction-retraction movement in overheadathletes with impingement symptoms. British Journal ofSports Medicine 38: 64–68.

De Mey K, Danneels L, Cagnie B, Cools AM 2012 Scapularmuscle rehabilitation exercises in overhead athletes withimpingement symptoms: Effect of a 6-week training pro-gram on muscle recruitment and functional outcome.American Journal of Sports Medicine 40: 1906–1915.

DerSimonian R, Laird N 1986 Meta-analysis in clinical trials.Controlled Clinical Trials 7: 177–188.

Desjardins-Charbonneau A, Roy JS, Dionne CE, Fremont P,MacDermid JC, Desmeules F 2015 The efficacy of manualtherapy for rotator cuff tendinopathy: A systematic review

and meta-analysis. Journal of Orthopaedic and SportsPhysical Therapy 45: 330–350.

Diercks R, Bron C, Dorrestijn O, Meskers C, Naber R, DeRuiter T, Willems J, Winters J, Van Der Woude HJ 2014Guideline for diagnosis and treatment of subacromial painsyndrome. Acta Orthopaedica 85: 314–322.

Dong W, Goost H, Lin XB, Burger C, Paul C, Wang ZL,Zhang TY, Jiang ZC, Welle K, Kabir K 2015 Treatmentsfor shoulder impingement syndrome: A PRISMA systema-tic review and network meta-analysis. Medicine 94: e510.

Ellenbecker TS, Cools A 2010 Rehabilitation of shoulderimpingement syndrome and rotator cuff injuries: An evi-dence-based review. British Journal of Sports Medicine 44:319–327.

Gill TJ, McIrvin E, Kocher MS, Homa K, Mair SD, HawkinsRJ 2002 The relative importance of acromial morphologyand age with respect to rotator cuff pathology. Journal ofShoulder and Elbow Surgery 11: 327–330.

Hanratty CE, McVeigh JG, Kerr DP, Basford JR, Finch MB,Pendleton A, Sim J 2012 The effectiveness of physiother-apy exercises in subacromial impingement syndrome: Asystematic review and meta-analysis. Seminars in Arthritisand Rheumatism 42: 297–316.

Higgins J, Green S 2011 Cochrane Handbook for SystematicReviews of Interventions. Chichester, England, Wiley-Blackwell.

Holmgren T, Hallgren HB, Öberg B, Adolfsson L, JohanssonK 2014 Effect of specific exercise strategy on need forsurgery in patients with subacromial impingement syn-drome: Randomised controlled study. British Journal ofSports Medicine 48: 1456–1457.

Hozo SP, Djulbegovic B, Hozo I 2005 Estimating the meanand variance from the median, range, and the size of asample. BMC Medical Research Methodology 5: 13.

Hsu YH, Chen WY, Lin HC, Wang WTJ, Shih YF 2009 Theeffects of taping on scapular kinematics and muscle per-formance in baseball players with shoulder impingementsyndrome. Journal of Electromyography and Kinesiology19: 1092–1099.

Hudak PL, Amadio PC, Bombardier C 1996 Development ofan upper extremity outcome measure: The DASH (disabil-ities of the arm, shoulder, and head). American Journal ofIndustrial Medicine 29: 602–608.

Kibler WB, Ludewig PM, McClure PW, Michener LA, Bak K,Sciascia AD 2013 Clinical implications of scapular dyskin-esis in shoulder injury: The 2013 consensus statementfrom the ‘Scapular Summit’. British Journal of SportsMedicine 47: 877–885.

Kirkley A, Griffin S, Dainty K 2003 Scoring systems for thefunctional assessment of the shoulder. Arthroscopy 19:1109–1120.

Kromer TO, Tautenhahn UG, De Bie RA, Staal JB,Bastiaenen CHG 2009 Effects of physiotherapy in patientswith shoulder impingement syndrome: A systematicreview of the literature. Journal of RehabilitationMedicine 41: 870–880.

Kuhn JE 2009 Exercise in the treatment of rotator cuffimpingement: A systematic review and a synthesized evi-dence-based rehabilitation protocol. Journal of Shoulderand Elbow Surgery 18: 138–160.

Larsen CM, Juul-Kristensen B, Lund H, Sogaard K 2014Measurement properties of existing clinical assessment

668 H. SAITO ET AL.

methods evaluating scapular positioning and function. Asystematic review. Physiotherapy Theory and Practice 30:453–482.

Lawrence RL, Braman JP, Laprade RF, Ludewig PM 2014Comparison of 3-dimensional shoulder complex kine-matics in individuals with and without shoulder pain,Part 1: sternoclavicular, acromioclavicular, and scapu-lothoracic joints. Journal of Orthopaedic and SportsPhysical Therapy 44: 636–645.

Lewis JS 2011 Subacromial impingement syndrome: A mus-culoskeletal condition or a clinical illusion? PhysicalTherapy Reviews 16: 388–398.

Lewis JS, Green A, Wright C 2005 Subacromial impingementsyndrome: the role of posture and muscle imbalance.Journal of Shoulder and Elbow Surgery 14: 385–392.

Lombardi I, Magri ÂG, Fleury AM, Da Silva AC, Natour J2008 Progressive resistance training in patients withshoulder impingement syndrome: A randomized con-trolled trial. Arthritis Care and Research 59: 615–622.

Luime JJ, Koes BW, Hendriksen IJ, Burdorf A, Verhagen AP,Miedema HS, Verhaar J 2004 Prevalence and incidence ofshoulder pain in the general population: A systematicreview. Scandinavian Journal of Rheumatology 33: 73–81.

Marinko LN, Chacko JM, Dalton D, Chacko CC 2011 Theeffectiveness of therapeutic exercise for painful shoulderconditions: A meta-analysis. Journal of Shoulder andElbow Surgery 20: 1351–1359.

Michaleff ZA, Costa LO, Moseley AM, Maher CG, ElkinsMR, Herbert RD, Sherrington C 2011 CENTRAL, PEDro,PubMed, and EMBASE are the most comprehensive data-bases indexing randomized controlled trials of physicaltherapy interventions. Physical Therapy 91: 190–197.

Michener LA, Subasi Yesilyaprak SS, Seitz AL, Timmons MK,Walsworth MK 2015 Supraspinatus tendon and subacro-mial space parameters measured on ultrasonographic ima-ging in subacromial impingement syndrome. KneeSurgery, Sports Traumatology, Arthroscopy 23: 363–369.

Miller P, Osmotherly P 2009 Does scapula taping facilitaterecovery for shoulder impingement symptoms? A pilotrandomized controlled trial. Journal of Manual andManipulative Therapy 17: E6–E13.

Moezy A, Sepehrifar S, Solaymani Dodaran M 2014 Theeffects of scapular stabilization based exercise therapy onpain, posture, flexibility and shoulder mobility in patientswith shoulder impingement syndrome: a controlled rando-mized clinical trial. Medical Journal of the IslamicRepublic of Iran 28: 87.

Moor BK, Wieser K, Slankamenac K, Gerber C, Bouaicha S2014 Relationship of individual scapular anatomy anddegenerative rotator cuff tears. Journal of Shoulder andElbow Surgery 23: 536–541.

Moraes GF, Faria CD, Teixeira-Salmela LF 2008 Scapularmuscle recruitment patterns and isokinetic strength ratiosof the shoulder rotator muscles in individuals with andwithout impingement syndrome. Journal of Shoulder andElbow Surgery 17: 48S–53S.

Neer C 1972 Anterior acromioplasty for the chronic impin-gement syndrome in the shoulder. Journal of Bone andJoint Surgery (Am) 54-A: 42–50.

Østerås H, Torstensen TA, Østerås B 2010 High-dosagemedical exercise therapy in patients with long-term

subacromial shoulder pain: A randomized controlledtrial. Physiotherapy Research International 15: 232–242.

Page MJ, Green S, McBain B, Surace SJ, Deitch J, Lyttle N,Mrocki MA, Buchbinder R 2016 Manual therapy andexercise for rotator cuff disease. Cochrane Database ofSystematic Reviews 6: CD012224.

Polimeni V, Panuccio A, Furfari P, Crupi D, Barreca G,Forgione C, Serranò R, Africa E, Africa A 2003Preliminary study on the efficacy of various rehabilitationtherapies for shoulder pain. Europa Medicophysica 39:59–63.

Reijneveld EA, Noten S, Michener LA, Cools A, Struyf F 2017Clinical outcomes of a scapular-focused treatment inpatients with subacromial pain syndrome: A systematicreview. British Journal of Sports Medicine 51: 436–441.

Reuther KE, Thomas SJ, Tucker JJ, Yannascoli SM, Caro AC,Vafa RP, Liu SS, Gordon JA, Bhatt PR, Kuntz AF, et al2014 Scapular dyskinesis is detrimental to shoulder tendonproperties and joint mechanics in a rat model. Journal ofOrthopaedic Research 32: 1436–1443.

Reuther KE, Tucker JJ, Thomas SJ, Vafa RP, Liu SS, GordonJA, Caro AC, Yannascoli SM, Kuntz AF, Soslowsky LJ2015 Effect of scapular dyskinesis on supraspinatus repairhealing in a rat model. Journal of Shoulder and ElbowSurgery 24: 1235–1242.

Roach KE, Budiman-Mak E, Songsiridej N, Lertranakul Y1991 Development of a shoulder pain and disabilityindex. Arthritis Care and Research 4: 143–149.

Savoie A, Mercier C, Desmeules F, Frémont P, Roy JS 2015Effects of a movement training oriented rehabilitationprogram on symptoms, functional limitations and acro-miohumeral distance in individuals with subacromial painsyndrome. Manual Therapy 20: 703–708.

Shah M, Sutaria J, Khant A 2014 Effectiveness of scapularstability exercises in the patient with the shoulder impin-gement syndrome. Indian Journal of Physical Therapy 2:79–84.

Smith MV, Calfee RP, Baumgarten KM, Brophy RH, WrightRW 2012 Upper extremity-specific measures of disabilityand outcomes in orthopaedic surgery. Journal of Bone andJoint Surgery (Am) 94 A: 277–285.

Struyf F, Nijs J, Mollekens S, Jeurissen I, Truijen S, MottramS, Meeusen R 2013 Scapular-focused treatment in patientswith shoulder impingement syndrome: a randomized clin-ical trial. Clinical Rheumatology 32: 73–85.

Surenkok O, Aytar A, Baltaci G 2009 Acute effects of scapularmobilization in shoulder dysfunction: A double-blind ran-domized placebo-controlled trial. Journal of SportRehabilitation 18: 493–501.

Tashjian RZ, Deloach J, Porucznik CA, Powell AP 2009Minimal clinically important differences (MCID) andpatient acceptable symptomatic state (PASS) for visualanalog scales (VAS) measuring pain in patients treatedfor rotator cuff disease. Journal of Shoulder and ElbowSurgery 18: 927–932.

Van der Windt DA, Koes BW, de Jong BA, Bouter LM 1995Shoulder disorders in general practice: incidence, patientcharacteristics, and management. Annals of the RheumaticDiseases 54: 959–964.

Van der Windt DA, Van der Heijden GJ, de Winter AF, KoesBW, Devillé W, Bouter LM 1998 The responsiveness of the

PHYSIOTHERAPY THEORY AND PRACTICE 669

Shoulder Disability Questionnaire. Annals of theRheumatic Diseases 57: 82–87.

Vecchio P, Kavanagh R, Hazleman BL, King RH 1995Shoulder pain in a community-based rheumatology clinic.Rheumatology 34: 440–442.

Virta L, Joranger P, Brox JI, Eriksson R 2012 Costs ofshoulder pain and resource use in primary health care: Acost-of-illness study in Sweden. BMC MusculoskeletalDisorders 13: 17.

Worland RL, Lee D, Orozco CG, SozaRex F, Keenan J 2003Correlation of age, acromial morphology, and rotator cufftear pathology diagnosed by ultrasound in asymptomaticpatients. Journal of the Southern Orthopaedic Association12: 23–26.

Worsley P, Warner M, Mottram S, Gadola S, Veeger HEJ,Hermens H, Morrissey D, Little P, Cooper C, Carr A, et al2013 Motor control retraining exercises for shoulderimpingement: Effects on function, muscle activation, andbiomechanics in young adults. Journal of Shoulder andElbow Surgery 22: e11–e19.

Yi A, Avramis IA, Argintar EH, White ER, Villacis DC,Hatch GF 2015 Subacromial volume and rotator cufftears: Does an association exist? Indian Journal ofOrthopaedics 49: 300–303.

Yuksel E, Yesilyaprak SS 2015 The effectiveness of scapularstabilization exercises in patients with subacromial impin-gement syndrome and scapular dyskinesis. Annals of theRheumatic Diseases 74: 1311–1316.

670 H. SAITO ET AL.