o

CL

Effects of Acupuncture Versus

U
O
Ultrasound in Patients With
0)
0
Impingement Syndrome:
Randomized Clinical Trial
Background and Purpose. Theri- i.s no definitive evidence for the
efficacy of tiie physical therapy interventions used for patients with
impingement syndrome. The purpose of thi.s study was to compare
manual acupuncture and continuous ultrasound, both applied in
addition to home exercises, for patients diagnosed with impingemeni
syndrome. Subjects and Methods. Eighty-five patients with clinical signs
of impingement syndrome were randomly assigned to either a grouj)
that received acupuncture (n = 44) ora group that received ultrasound
(n=41). Both intenentions were given by physical therapists twice a
week for 5 weeks in addition to a home exercise program. Scores from
3 shoulder disability measures, combined in the analysis, measured
chauí^e during a period of 12 months. Results. Both groups improved,
but the acupuncture group had a larger improvement in the comhined
score. Discussion and Conclusion. The results suggest that acuputu turc
is more eftlcacious than ultrasound when applied in addition to home
exercises. [Johansson KM, Adolfsson l.E, Foldevi MOM. Effects of
acupuncuire vei-sus ultixisound in patients with im]jiugc-nicut syndrome:
randomized clinical trial. Phys '¡'her. 20Ü5;85:49Ü-r)ül.j

Key W o r d s : Family paclke. Home exercise program. Physical therapy. Rotator cuff, Subacromial pain.

Kajsa M Johansson, Lars E Adolfsson, Mats OM Foldevi

490 Physical Therapy . Volume 85 . Number 6 . June 2005

 mpingement syndrome is one of the most common Furthermore, there is a strong need for sttidies on

I diagnoses of patients with shottlder problems. A

prevalence of 7% has been reported in a Swedish
population.' In Dutch general practice, about 48%
of patients who consitlted a genei-al pi-actitioner for shoul-
der problems were diagnosed witli impingement syn-
dtome,^ and this condition is reported to be persistent.''
In the current stndy, the term "impingement syndrome"
is used. Patients with this syndrome experience pain in
the delloid tiuiscle area, especially dtiring arm elevation.
Différent nianetivers cotnpressing the subacrotïiial btusa
and the supraspinatits muscle between the acromion
and the htimeral head can be used to reproduce this
pain. ' ''
(^ontrovetsy exists about the pain-generating tuecha-
nism.s ill patients with impingenietit syndrotne. In
(heoiy, pain from the subacromial structures can occur
IVom extrinsic mechanical wear or compression from the
coracoacrotuial arch, bta there also may be intrinsic
causes such as degenerative changes in the rotator cuff.''
These patients often receive different kinds of physical
therapy interventions.'^ btit there is no definitive evi-
dence that physical therapy intenentions are efficacious
for patients with impingement .syndrome.^'' The inter-
ventions chosen in our study—acupuncture, ultrasound,
and home exercises—are commonly used intenentions
among physical therapists in Swedish primaiy care.""*
Acuptmclure has been used by physical therapists in
Sweden since the mid-1980s," a more recent treatment
alternative tiian tiltrasomid. Befbie the start of this sttidy
in 1997, two reviews had expressed doubt about the
efficacy of therapetitic tiltrasound for musculoskeletal
ilisorders.'-'-'-^ The most comniiin sttategy of physical
tlîcrapists is to use a combination of interventions.
existing physical therapy intenentions." Because bodi
acupuncture and ultrasound are common interventions,
often used in cotnbination with exercises, and have a
similar treatment setup, it seemed reasonable to com-
pare them in this clinical trial.
The purpose of ottr study was to evaluate and compare
the efficacy of 2 physical therapy strategies for patienLs
with impingement syndrome: (1) actiptmcture applied
in additioti to home exercises and (2) coutintiotis ttltra-
soimd therapy applied in addition to home exercises.
Using previotisly pitblished Ínstr\mients, the otttcomes
were measttred dtiring a period of 12 months.
Method
A prosjjective, obser\er-bllnd, tandomi/cd clinical trial
was conducted.
Subjects
The subjects were recruited from 3 urban primar)' health
care centers in the cotmt\- of Östergolland, Sweden,
from March 1997 to J tine 2000. Patients with slioulder
pain who contacted the geneial practitioners or physical
therapists at these primary health care centers were
offered an encoitnter with the research physical thera-
pist (K]) if they were between 30 and 65 years of age. The
general practitioners and physical therapists were
instrttcted to recruit patients with clinical signs of a
probable impingement syndrome, described as pain
during abduction and pain located in the proximal
lateial aspect of the upper arm.
Potential participants underwent a statidaidized clinical
examination perfotmed by the research physical theia-
pist. At the inchtsion visit, backgrotind data on age. sex,
duration, occupation related to arm load, leisut-e activi-

KM fohiinssoii. PT. Pliti, is Lecturer. Physical Tlierapy Program. Department of Healtli and Sue icty. Primary Care, Linkopings Universitet, S-581
8:i. l.inköpiiig, Sweden (Kajsa.Johansson@ihs.liii.se). Address M roncspondeiire to Dr Johansson.

LE Adolfs.son. MD, PhD, is Associate Professor. Department of Neuroscieiite and Locomotion. Orthopfdics and Sports Medicine, Linköpings
Univereitet.

MOM Foldevi. MI>. PhD. is Assi^ciate Professor, Department of Health and Society, Primary Care, Linköpings Universitet.

All authors provided concept/idea/research design, writing, and project management. Dr Johansson provided data collection, and Dr Johansson
and Dr FoUkvi provifk-d data analysis and fund procurement. Dr Adolfsson and Dr Fi>lde\i proxnded suhjects. The authors acknowledge
statisticians John Ciii^u-nsen. MaLs Kredrikson. and Oile Eriksson for statistical consultiuion. Mi Jan Brandingei. Ms Karin Lindgren, Mr C-hrisier
Nilsson, and Mrs Gunilla StAlmarck are acknowledged for their work as physical therapists throughout the suidy. The authors also thank all
participating patients and staíí at the involved priman' health care centers.

This study was approved by the Ethics Committee of the Faculty of Health Sciences at Linköpings lUiiversitet.

This study was supported by funding and facilities provided by the Coiiniy C;ouncil of Östergötland and Linköpings Universitet. Sweden.

This article ii'os recäved Febntary 5. 2004, and loas accepted November 19, 2004.

Physical Therapy . Volume 85 . Number 6 . June 2005 Johansson et al . 491

 tit's, .stiHíkiiig. atid medical histoiT were doctmietiifd.
The histuiT iiuitided clescripiioii olsytnptonis atid pain
Iniation, duration of cttrreiit episode, circtimstances at
utiset of paitt. piiin related to rest, iiiglit sleep and
activities, tecttrtence or a fhst-titne probletn, medica-
tioii. atid sick leave. The complete set of inclttsion atid
excltisioii critetia is sh<nvn in Figttre 1.
Patiettts WIKÍ were diagnosed as Iiaving probable
iiiipiii-ieinent .svtidrome had a ftiial incltision lest, an
iinpingenicnt test as described by Necr atid Welsli."' The
procedtire was first to perform tbe Neer impingetîient
sign test'' atid tbe Hawkins-Kennedy inipitigetnent sign
test.' Tben tbe itnpingemetit test, where a local anes-
thetic, lOtnLof piilocaitic {10 ing/inL), was injected by
a general piactilioiier with tbe patient seated, ttsing a
posterolatetal injectioti approach with tbe needle enter-
ing tbe subacioniial space.''* If tbe patient reported
relief of pain wben the impingement sigti tests were
repeated 10 minntes aíter tbe injectioti, the test was
Jtidged as positive. If not, tbe tnanettvei-s were repeated
after another 20 mintttes. If none of the excbtsioii
criteria (Fig. 1) were ptesent, tbe patient was asked to
give itiformed consent to participate in the sttidv. All
patients sigtied informed consent statements at the etui
of the incbtsioti visit.
Of tbe 173 patients who visited the research physical
therapist, 88 were diagno.sed as having inipitigenieni
syndrome and fttiftUed tbe otbcr inclttsion criteria.
Three of these patients did not enter tbe study, one dtte
to working conditions, one becatise of a fear of needles,
and one becatise of a myocardial iniantioii that
occurred between the inclusion visit and tbe start oi'
intenentions. The affected shoulder of tbe retnaiiiing 85
patients was radiologically examined to exclude malig-
nancy, t)steoarthritis of tbe glenohumeral joint, and
skeletal ahnormalities. Standard antetoposterior and
lateral ptqjecllons wete taketi as well as a special projec-
tion of tbe act omioclavictilar joint. None of the patients
were exclttded on the hasis of the radiographs. Accord-
ingly. 85 patients etuered tbe stttdy and were landonity
a.ssigned to eitber a group that received acupuncture
combitied witb bome exerci.ses (n=44) or a grotip tbat
teceived contintiotis ttltrasottnd combined witb home
exercises {n = 41).
Concealed randomization, based on a random list, witb
tbe treattuetit alternative in envelopes was carried oitt
beforehand. Tbe inienention was tben ititrodticed and
perlbrmed by 4 pinsical therapists al the same priman
health care center. All of the physical therapist-s were
experienced and bad worked in pritnaiy cate for at least
12 years. In Sweden, acupuncture was approved in U)84
by tbe Swedisb National Board of Healtb and Welfare to
492 . Johansson el al
Compulsory inclusion criteria:
• 3 0 - 6 5 years of age
• 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 anesthetic]
• At least 2 months' duration of the current episode
Three of the following 4 inclusion criteria
must be positive:
• Hawkins-Kennedy impingement sign
• Jabe supraspinotus muscle test (in 90° of abduction
in the scapular plane)
• Neer impingement sign
• Painful arc between 60° and 1 20° of active abduction
Exclusion criteria:
• Radiolagical findings: malignancy, osteoarthritis of
the glenohumeral joint, skeletal abnormalities
decreasing the subacromial space (bony spurs,
osteophytes)
• Known or suspected polyarthritis, rheumatoid arthri-
tis, or diagnosed fibromyalgia
• Previous fractures of any bone in the shoulder
complex or shoulder surgery on the affected side
• Dislocation of the glenahumeral ¡aint or the clavic-
ular joints on the affected side
• History or current clinical findings of instability in
any joint of the shoulder complex (negative appre-
hension sign-relocation test for exclusion af ventral
instability of the glenohumeral joint)
• Suspicion of frozen shoulder: time-dependent
decreased range of movements following the cap-
sular pattern (external ratation-abductian-internal
rotation) and pain during intra-articular mabilization
• Problems from the cervical spine: shoulder symp-
toms reproduced with neck movements or a positive
test for the foramina intervertebrolia (pain or neuro-
logical symptoms during manual extension com-
bined with manual lateral flexion and rotation
toward the tested side)
• 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 clinicai picture of ruptured rotator cuff (trauma,
pronounced weakness, atrophy)
» Acute subacromial bursitis, making a clinical exam-
ination impossible due to pain
• Difficulty participating in data collection due to
communication problem
Figure 1.
Inclusion and exclusion criteria.
Physical Therapy . Volume 85 . Number 6 . June 2005
be used by registered medical professions after addi-
tional training,"
Procedure
Ibc atiipiinctiire group received 10 ireatinenl sessions.
The physical ibcrapists used standardized needle place-
ment at 4 local points (LI 14 [Binao], LI 15 [|ianyai].
LU 1 [Zbongfti]. and TE 14 [jianliao]) and one distal
point (LI 4 [Ht'gu]) (Appendix I). All points were
cbosen in accordance witb current practice, and. before
startiîig tbe study, all physical iberapists were trained to
locate these points. Tbe depib and angle of needle
insertion were those described in a Swedish mantial."'
Tbe type of needle used was a HECiU* sterile and
single-packaged one-lime needle no. 8 (30 nnn long and
0.30 mm in diameter). Tbe treatment was repeated twice
a week for 5 weeks, and each treatment session lasted 30
miinites. The patients lay on a treatment table on tbeir
unaffected side. After insertion into tbe defined points,
tbe needle was rotated a few seconds until "de qi" was
exjjerienced Ijy the patient. De qi is often described by
patients as a sensation of bea\nness, numbness, and
radiating parestbesia." It is believed to be a sign of tbe
activation of the descending pain inhibiton systems, and
opioid peptides are released, especially by the midbrain
¡H-riatiuaductal grey."'** In total, 3 stimulations were
perloinied (ie, at insertion and after 15 and 30 minutes).
De qi was to be experienced at every stimulation at each
acupuncture point, if not tbe needle was adjusted nntil
ibis wiis tbe case.
The ultra.sound group received continuotis ultrasound
twice a week Ihr 5 weeks (10 treatment sessions). Eacli
session lasted 10 minutes, and a standardized mode
(frequency=I MHz, spatial-average inten.sity= 1 W/cm'-^,
gel coupling) was used. The size of the transducer was
4 cm-, and the skin area treated was twice this size,
covering an area of about 8 to 10 cm- inferior to the
anterior and lateral part of the acromion. Tbe trans-
dticer bead was moved in small circles covering tbe
area."' Tbe patients were .seated with tbe glenobumeral
joint extended and medially rotated in order to make
tbe muscle insertion of tbe supraspinattis muscle appear
beneatb and anterior to tbe acromion.-" Tbis joint
position was maintained by placing tbe arm behind the
back of the chair. Tbe equipment tised was a Pbyaction
190^ ultrasound device. The same equipment was used
for all patients, and it was tested by an independent
medical tecbnician before starting tbe study and tben
once every 12 months. No recalibration was needed
during tbe study.
* S\i-iiska .\B. PO BoK 89. SK-57(1 12 l.;inclsbro. Swcrit-n.
' tinipliy, PO Biix S.'iS, NL-3(i()l) KiiiiUinvfii. ilic NrllK-rlancls.
Phy5ical Therapy . Volume 85 . Number ó . June 2005
Both intei-ventions were combined witb a 2-step bome
exercise program developed in and based on clinical
practice as well as supported by lesearcb.-' Tbe chosen
exercises in tbe first part of the exercise program were
targeted to maintain or restore motion as well as to
stimulate circulation in tbe rotator cufl" using uiany
repetitions of low-intensity exercises. witbotU provoking
pain from tbe involved tissttes. In the second pait of tlie
exercise program, the target was to strengthen the
rotator cuff muscles witb the upper arm in a neutral
position to avoid impingement. In all exercises, tbe posi-
tion of a retracted shoulder was empbasized, in line witli
tbe findings of Solem-Bertoft et al,-- where a protracted
shotilder resulted in a narrowing of the anterior aspects t)f
tbe stibacromial space. Appendix 2 gives detailed descrip-
tions of all of tbe exercises in ibe program.
At tbe first treatment visit, the patients received instruc-
tions from the physical therapist and practiced tbe
exercises in part one of tbe program. They were
insüucted to perform tbe program daily for 5 weeks.
Alter the first half of the treatment period, tbe patients
received instruction and practiced tbe second part ol die
exercise program. All rotations were performed witb a
pillow in tbe axilla to decrease the activity in tlie di-ltoid
muscle. The exercises were to be done eveiy tHher da)
during the fourth and fiftb weeks. Pain dining tbe
exerci.ses was not to last more than 10 to 15 minutes after
the ptogram. If pain persisted longer than that, tbe
patients were instructed to decrease either tbe resistance
or tbe force produced. Adberence to the exercise i)ro-
gram was monitored by a bome exercise adbeience log,
and tbe use of additional medications was reported.
Outcotne Measures
Tbe researcb physical therapist, who performed tbe
examinations and all a.ssessments, was uninformed of
treatment grotip assignments tbrougbout the study. The
same clinical examination (tbe Neer impingement test
excluded), witb the same assessment instruments, was
repeated tbe week after tbe period of acupuncttne or
ultrasound was completed and 3, 6, and 12 montbs from
the date of the initial visit. At each visit after tbe
inclusion, current symptoms and differences from base-
line were documented.
During the planning of tbe study in 1996, tbere was no
consensus about which instrument should be used when
assessitig patients witb impingement syndrome. This
uncertainty and tbe decision of tbe Emopean Society for
Surgeiy of tbe Sbouldei" and Elbow-" tbat the Clonstant-
Murley Shoulder Assessment (CM Score)-'' should be
used in all reseaich involving patients witb sboulder
problems led to tbe cboice of tising 3 disease-specific
shoulder assessment scales: tbe CM Score, tbe AdoUsson-
Lysbolm Sboulder Score (AL Score),--' and tbe Utiiver-
Johansson et al . 493
 Table 1.
Outcome Measures" and Variables
Variable
Pain at rest
Pain during activity
Ache, constant or intermittent, requiring
painkillers or related to rest or activities
Disturbed sleep due to pain
Active moments in the glenohumeral ¡aint
Abduction
Flexion
Inward/outward rotation
Level of activity
Muscle force (glenohumeral ¡oint):
Abduction
Flexion
Instability (glenohumeral joint)
Disability in leisure titne/sports
Disability af work
Expectation of satisfaction with results
[baseline)
Sotisfaction with results (at follow-ups)
.\l.=.-\dol!ss(in-l,yslnjlm Shouldtir Si-orc, C:M = Const:iiu-Murlcv Shoulder
-weiMiment, UCl A^l.'nivei-sily ol Caliliirnia at Lus /Vngtl cs tud-R<isuli Score.
sity of California at L,os Angeles End-Result Score (UCLA
Score).-'' All 3 scales were used al baseline and at each
assessment visit. Their variables are described in more
detail in Table 1.
The maximum score for the VCIA Score was 35 points;
for the other 2 scales, the maximum score was 100
point-s. The AL Score is a pure patient self-assessment,
and the other 2 scales also include clinical measures. For
all instruments, the construct, content, and criterion
\alidity and the knowledge about reliability, in our
opinion, is insufficient. The CM Score and UCLA Score
seemed appiopriate ba.sed on our clinical experience,
itud both scales have been widely used.
The AL Score was chosen because it was developed for
patients with impingement syndrome.-"' With respect to
tcst-retest factors, we recently evaluated intraobser\er
reliability for the AL Score and found it to be stable ovei
time for patients with impingement syndrome {unpub-
lished data). Tfiirty-five patients with impingement syn-
drome of at least 2 months' duration completed the
score twice. The inteival was 3 to 7 days, and the score
was repeated at the same time of the day. None of the
patients received any intervention during the study, and,
if they used medication for their symptoms, they were
instructed to maintain the same level. The results were
analyzed with a repeated-measures analysis of variance
494 , Johansson et al
(ANOVA). The intraclass correlation coefficient for the
total score was .91.
Outcome Data Analysis
Measure
.W\ patients were adherent to the study protocol (no
AL missed or additional intenentions) during tlie 5 weeks
AL, CM of acupuncture oi ultrasoimd. At the 3-, 6-, and
UCLA 12-month visits, the number of patients who were adher-
ent to the study protocol changed, as showLi in Figure 2.
AL, CM
In total, 64 patients were adherent to the study protocol
throughout the study. The data were analyzed both lot
CM
the group adliering to the study protocol and with an
CM, UCLA "intention-to-treat" (ITT) application model for analysis
CM of data for clinical trials.-" The latter analysis included all
AL, CM, UCLA patients who were randomly assigned lo gixnips. The
principle of last observation carried forward (LOCF) was
CM
used in both analyses, using the scores recorded just
UCLA prior to the missing scores in case of missing postircat-
AL
ment values.^^ The number of patients where LOCF was
used is illustrated in Figure 2.
AL. CM
AL, CM
A sample-size estimation resulted in a requirement ol a
UCLA
minimimi of 40 patients in each gionp. if the expected
rate of improved patients was to be 30% better in one
UCIA
group than in the other group (ß=.8O, a=.05).=^« This
level of difference was a compromise of whai we ¡udgerl
to be a relevant clinical effect and an assumption ol how
many patients we expected to be possible lo included in
a 3-year period.
In the data analysis, we have chosen lo combine the
scores for the 3 outcome measures, using the moan of
the 3 outcome measures" total scores. The maximum of
35 points for the UCLA Score and of 100 for the CM
Score and the AL Score \\"as corrected by tintltiplying the
UCLA Score by 100 and then dividing by 35.
To compare background variables between the 2 treat-
ment grotips, we used a Student / test lor continuous
data, a chi-square test for categorical data, and a Mann-
Whitney U test for ordinal data. A repeated-tneasmt-s
ANOVA was used to analyze the change in the combined
shoulder disability score over time within the treatment
groups, To compare the outcome between the groups,
we used a general linear model analysis of covariance
(ANCOVA) for repeated measures. The combined
sboulder disability score at all 4 visits after the 5 weeks of
acupuncture or ultrasound sened as the dependent
variable, and the stardng score was the covariate. This
analysis was chosen to adjust for the difference in
baseline score between the groups. The level of statistical
significance for all testing was P<.05.
Results
Before treatment, there were no differences in the
background variables between treatment groups
Physical Therapy . Volume 85 . Number 6 . June 2005
 Before Acupuncture Group Ultrasound Group
treatment n=41
n=44
r
Visit after 5
weeks of n=44 1 additional
n=41 7 additional
treatment treatment treatments
r 1 withdrawal r

3-mo visit n=42 (2) n=34(1)

5 additional 1 withdrawal
treatments 1 surgery

6-mo visit n=37 n=32 (2)

3 additional 1 additional
treatments treatment
n=30 (4) 1 surgery
12-mo visit n=34 (2)

Figure 2«
Flow chart of the sample during the study (with the number where last observations carried farward has been used in parentheses).

Table 2.
Background Variables for the 2 Treatment Graups and Statistical Analysis Used

Acupuncture Group Ultrasound Group

(n=44) (n=41) Statistical Analysis

Sex: female/male (n) 32/12 27/14 NS,'" Yales' corrected chi-square test

Age (y)
X 49 49 NS, Student f test
SD 7 8
Duration of current episode [n]
2-3 mo 13 11 '\
4-6 mo 8 10 1 NS, Mann-V>/hitney U test
7-12 mo 10 11 1
>12 mo 13 9 J
Occupation (n)
Repetitive arm lifting, at least moderate load 18 17 1
Static arm boding {ie, hair dresser) 4 3 1
Computer work 16 15 > NS, Fisher exact test
Similar to octivities of daily living 4 4 1
Retired 2 2 J
Sick leave at start (n) 5 2 NS, Fisher exact test

Used painkillers during treatment or follow-up (n) 5 7 NS. Yates' corrected chi-square test

Exercise regularly or leisure octivities looding 34 36 NS. Yotes' corrected chi-squore test
the arm, at least once a week (n)
Smoking ?10 cigarettes o day (n) 5 2 NS. Fisher exact test

(suiiisliail U-vrl

Physical Therapy . Volume 85 . Number 6 . June 2005 Johansson et al . 495

 Table 3.
Outcome Measures for the Combined Score at Baseline and ol Each Assessment Visit for Both (he Group Adhering to the Study Protocol and the
Intention-to-Treaf Group (ITT) Using Lost Observotion Carried Forword for Missing Values

Adhering to Study Protocol (n=64) i n (n-85)

Acupuncture Ultrasound Acupuncture Ultrasound
X SD X SD X SD X SD
Baseline (before treatment) 6] 7 63 6 61 7 63 6
After treatment 79 9 76 11 79 9 76 1]
3-mo assessment 84 9 83 10 81 12 78 13
6-mo assessment 90 7 88 11 83 17 83 15
1 2-mo assessment 93 4 89 10 88 13 85 14

(Tab. 2). No adverse effects or side effects were reported

in eillicr group during or after the treatment period. 100

Nine patients in the acupuncture group and 8 patients 95

in the ultrasound group received additional treatment. 90

These 17 patients were consequently not adhering to the 85

study protocol, and their data were included in the ITT BO

analyses. 75
70

Two patients undemenl surgical suhacromial decom- 65 -I

— * — Acupuncture Group
pression, and 2 padents withdrew from the study. One 60
- - • - - Ultrasound Group
woman found the participation tinic-consuniing and 55 -I

conflicting witli her work, and one man declined furlhei- 50

participation because he felt no improvement (Fig. 2).
These 4 patients did not appear to differ in background
characteristics or in scores in comparison with the other
patients.
There were no differences in adherence to home exer-
cises or in use of additional pain medication between the
treatment groups. The number where LOCF was used
was similar in both groups. The scores of the patients for
whom LOCF was used did not appear lo differ from the
other patients.
Both tteatment groups' mean scores at baseline and at
each assessment visit are presented in Table 3. There
were no differences between the treatment groups'
mean scores at baseline. Both treatment groups
improved during the study (P<.0001, ANOVA).
The between-group analysis, including the mean scores
from all 4 assessment visits (after 5 weeks of acupuncture
or ultra.sound and at 3. 6, and 12 months), showed a
larger change (P=.O45, ANCOVA) in the combined
scoie for the acupuncture group, analyzed with those
adhering to die study protocol (Fig. 3). This effect was
seen already at the first assessment visit and was main-
tained over time. In the ITT analyses, no differences
were found across the 4 data collection periods.
Discussion and Conclusion
We set out to evaluate and compare 2 common treat-
ment strategies in physical therapy for patients with
Baseline Atter 3 mo 6 mo
Treatment
12 mo
Figure 3.
Presentation of the mean in the cambined score at each visit for the 2
treatment groups, including all potienfs who adhered to the study
protocol (n = 64). Between-group difference anolyzed wilh analysis of
covarionce (P=.O45).
impingement syndrome. The results showed that treat-
tiient with acupuncture in aiklition lo home exercises
was more efficacious than adding continuous ultta-
sound. This efficacy of acupuncture supports the
earlier findings by Kleinhenz et al,-" who compared
acupuncture and a placebo needle and concluded
that acupuncture was superior.
The magnitude of the treatment effect is unknown in
the absence of a true control group. To our knowledge,
no earlier study has dealt with the natural course oí
impingement syndrome. For unspecified shoulder pain,
however. Ginn et al^" reported no imptoveinent after 1
month without treatment, and Macfarlane et al'"
described persistent disabling problems after 3 years.^"
The outcome illustrated in Figure 3 is probably a com-
bination of a treatment eflect atid the natutal course.
In our view, the patients selected for our study were
representative of the general population seeking care for
this type of shoulder problem. With a few exceptions, all
referrals came from general piactitioncrs and physical
therapists in pi imaiy care, and this should avoid the risk

496 . Johansson et ol Physical Therapy , Volume 85 - Number 6 . June 2005

of studying a highly selective group (eg, patients from a
sports or surgical clinic).
The analysis of the group adhering to the protocol in
this study was used to evaluate the efficacy of the studied
interventions, and the ITT analyses were used to exam-
ine tiie overall benefits of intcr\'entions in primary care.
When interpreting the results, physical therapists should
be aware that, although the patients included in this
study had a more specific diagnosis than a group of
patients with shoulder pain, there still could be other
reasons than subacromial reasons for their pain and
disability.
The maneuvers used for inclusion in our study have
been reported to compress the structures of interest.'^^
Some authors''-^ have reported how the subacromial
pressure increases during the impingement sign test.
Higli sensiti\it)^ has been reported for Necr impinge-
ment sign test (75%-89%^'*-'^-') as well as for Hawkins- procedure decreased the variabilit) and tbus increased
Kennedy impingement sign test (88%-92%^'*'^'^), but
their specificity is lower (Neer impingement sign test:
31%-51%; Hawkins-Kennedy impingement sign test:
2b%-44%'^'*-^--), which lessens their discriminative abil- tainty about the instruments' qualities. The differences
ity. The sensiti\'ity for the Neer impingement sign test
has been reported to be 70% to 83%.^''••'" This diagnosdc overall effect over time, but the- differences thai we
injection test has been used in earlier research as the
gold standard for idendfying impingement syndrome.-^''
In our study, it was tised as a compulsory criterion, but in
combination with other findings to increase its positive
predictive value. In a recent review,^** the inclusion
criteria used in our study were reported as proper and
ihe exclusion criteria as adequate because they control
for conditions interfering with a successful otitcome of
treatment for patients with impingement syndrome.
Despite a lack of certainty abotit what diagnostic tests
should be used, we believe tliat the chosen combination of
inchision and exclusion criteria was sufficient for identify- both combined witb a home exercise program. The
ing a grotip of patients with impingement syndrome. Still,
it is difficult to state whether or not there is a partial
rupture of tlie rotator cuff, which could explain why some
patients had less improvement than otlier patients.
Our choice to emphasize the importance of selecting
interventions with similar setups as well as the use of a
standardized treatment protocol is supported by the
lecenl published CONSORT statement.^'^ The conclti-
sion in ottr earlier review that ultrasound is ineffective.'" therapeutic
together with Ktirtaiç Gürsel and colleagues' conclusion
tbat ultrasound bad no effect as additional treatment
to physical therapy interventions,"' may imply that this
study compared home exercises with and without
acuptincture.
The exercises used in our study were similar to those
reported as efficacious in earlier studies.^'^^ The chosen
Physical Therapy . Volume 85 . Number 6 . June 2005
exercise involving external rotation, with fixed elbows
using a tube, has been reported to result in the highest
activation of the infraspinatus muscle, a muscle that is
important to strengthen in patients v^ith impingement
syndrome.'^ Ginn et al'*" reported tbat strengthening
exercises and motor retraining were s\iperior to no
intervention for patients with shotilder pain. The
repeated clinical examination at each assessment visit
has A methodological advantage over follow-ups with
scores mailed to the patients because bias from changing
diagnoses over time can be avoided."
iimitotions of the Study
The major limitation concerns the instruments tised for
the outcome measure. Because all 3 measures have
indeterminate measiuement ]irt)perties. we chose to use
the mean of tbe 3 total scores, mainly to make the
reporting of results less complicated than if all scores
from all visits should be prcst-ntt-d sejiarately. This
the power in the statistical analysis, but the sensitivit)* for
change probably decreased, with a possible underestima-
tion of the real efíect. However, there is still an uncer-
between llie grotips were small compared with the
found corresponded to 1 to 2 steps in the outcome scales
(eg, .5 points corresponds to having or not having
disturbing nightly pain that interfeies with sleep, a
difference we regard as clinically significant).
Another aspect when interpreting the resulu fiom tliis
study is that the influence of psychosocial factors is
unknown, because no instrument covering this area was
used. To our knowledge, this is the first randomized
clinical trial involving patients with impingement syn-
drome and comparing acupuncture and ultrasound,
larger improvement in the acupuncture group, which
was seen at the first assessmeni visit and maintained over
rime, indicates Uiat a physical therapy strategy with a
combination of acupuncture and home exercises is
more beneficial for most patients \\ith impingement
.syndrome. This conclusion is supported by our recent
review,'" where tentative evidence was found for the
short-term efficacy of acupuncture- and strengihoning
exercises. Furthermore, we concluded in that review that
ultrasound was ineffective in these
patients.^" In conclusion, acupuncture is advocated
before ultrasotmd, in addition to home exercises, for
patients witli impingement syndrome.
References
1 Jatobssaii I-, Lindgarde F. Manihorpe R. The commonest rhfuinalic
complaint over a six-week diiralion in a twelve-month peiiüd in a
defined Swedish population: provalence and relaiionships, Smnd
JRhnivmUil.
Johansson et al . 497
 2 Van der Wiiidt DAWM, Kocs B\V. Bot-ke AJP. ei al. Shoulder disor-
ders in general practire: prognostic iiiditatnrs of oulcdmc. Br J Gen
Pract. 19
3 Chard MD, Satelle LM. Hazlcinan BI.. Tlie long-temi oiucome of
rotator curt" tendiniiis: a review study, lirf liheumntol. 1988:2Tim'i-SSy.
4 Hawkins RJ. Kennedy j C Impingemeni syndrome in athletes. Am f
Sfmts MM. i;)H();8:151-158.
5 Neer CS. Welsh R?. The shoulder in sports. Orihop Can North Am.
1977;8:583-591.
6 Fii FM. Hanier CD, Klein AH. Shoulder impingement syndrome: a
crilical leview. CJni Oilhop. I9yi:^(i9:lf»2-i73.
7Johans.son KM, Adolfsson LE, Foldevi MOM. Atliuides toward man-
agement of paiienls wiLh siibacromi;il pain in Swedish primaiy care.
¡'am hart. l999;
8 Green S. Buchbinder R, Glazier R. Forhes A. Systematir review ol
randomi.sfd conlrolled trials ol' intcnentions for painful shouldeis:
selection iriteria, oiitrome assessment, iiiid efHcacy. HMJ- 19il8;316:
354-360.
9 van der Heijden G|Mr,, Van tier Windl DAWM, De Winter AF.
Physiotherapy for patients with soft us.sue sliouldei- disorders: a system-
atic review of randomised clinieal trials. BMJ. 1997:315:25-30.
10Johan.s.son KM, Öberg B, Adolfsson LE. Foldevi MOM. A combina-
tion oi sysiemaiic review and clinicians' helicis in inteiventions for
snbaeromial pain. BrJ Gen Pract. 2002:52:145-.'i2.
11 Andefss<jn S. The functional backgrotind in acuptmctiire effects.
SiintdJ Hehtihil Med. l993:29:'iI-r>0.
12 Holmes MAM. RiidlandJR. Clinical trials of ultrasomid treatment
in soft tissue injury: a review and critique. l'h%útilhempy Theiny and
ñmiice. 1991:7:163-17.').
l3Nyholm Clam A.Johann.sen F. Ultrasound therapy in musculoskel-
ital disorders: a meta-analysis. Pain. 199.5:ti3:85-i)l.
14 Edwards S. Partrid)>:e C. Mee R. Treatment schedules lor research:
,1 model for |)h\siotherapy. Physiotherapy. 1990;76:605-607.
L") Miclu-o VVF. Rodrigue/ R,A., .\myE. Joint andsolt-ussue injections of
till' uppei- extieinity. Physical Medicine ami Hehahditatiori Clinics of Nin1h shoulder: an arthrographic evaluation. Artliniscofiy. 2002:18:887-891.
America. 1995;6:823-840.
16 Andersson S. Lundeberg T, Lund 1. et al. Knmpmdium i Akiinpnktm
I in Swedisii|. Göteborg. Sweden: Vasastadeiis Bokbinderi .AB: 1993.
17 lakeshige C. Sato T, .Mera T. étal. Descending pain inhihitor-v
systems involved in iuupuncttue analge.sia. Brain firs Bull. 1992:29:
rtl7-6:t4.
18 Dehreceni L. Chemical releases associated with acupuncltire and
elc-ctric stiiTUilatioii. Critical Rnúm's in Physical and RehahiUtatiim Meili- management ol soft tisstie disorders of the slioul<ler: a landoriii/i-d
19Ntisshaum EL. Ultrasound: to heat or not ui biat—tbat i.s the
(jiiestion. Physical Therajry Rexneuis. l997;2:.'J9-72.
20 Mattingly G, Mackarcy i». Optimal methods for shoulder tendon
palpation: a cadaver study. Phy.s Ther. 1996:76:166-174.
21 McCann PD, W.xnicn MF. K:idaba MP. Bigliaiii LL'. A kinematic
;md i'le( tromyograpliif study of shoulder lehabilitation exercises. Clin
Í993:288:179-l«8.
22 Solem-Bertolt F. Thuoonias K-Â. Westetberg t^E. The inlhience of
scapular renaciimi and protraction on the width of the suhacromial
space: au MRI sliidy. Clin Ortlmp. I993;296:99-IO3.
23 .\s.sessni< nt systems. European Society For Surgen' of the Shoulder
and the Elbow. .Available at: http://wvvw.secec.org. Acte.ssed Februaiy
8. 2005.
498 . Johansson et al
24 Constant CR, Murley AHG. A <lini(al method of functional assess-
ment of the shoulder. CUn (hihop. 19H7;2l4:160-164.
25 Adolfsson L, Lysholm j . Results of arthroscopic acromioplasty
related to rotator cuff lesions, ¡nt Orthop. 1993:17:228-231.
26 Elltiian H, Hanker G. Bayer M. Repair of the roiatorcufr.7/Jí>Hí'//i/>¡/
SurgAm. 1986:68:1136-1144.
27 i'.illings D. Koch G. The application of the ptinciple of intention-
to-trcat to the analysis ol clinical trials. DruglnfJ. 1991:25:411-424.
28Kirkwood BR. Essentials of Medical Statistics. Oxford, United King-
dom: Blarkwell Scientiiu Publications: 1994:191-200.
29 Kieinhen/ |. Streithetger K. Windeli-r |. et al. Randomised clinical
trial comparing the eliects of acupuncture and a newly designed
placebo needle in rotator cuff tendinitis. Pain. 1999:83:235-241.
SO C.iini K/\, Herben RD, KJioiiw W, Lee R. Randomi/ed. controlled
clinical nial ol a treatment for shoulder pain. Pliys Ther. 1997:77:
802-811.
31 Macfariaiie (;|. Hunt IM, Silman AL. Predictoi-s of chronic sboulder
pain: a population-based prospective study, journal oj Hheumatohgy.
1998:25:1612-1610.
32 \aladie A III. jobe C. Pink M, et al. Anaiomv of provocative tests for
impingemeut syndrome of the sbotilder. J Shoulder ¡•M»,ii' Siirg. 2000:9:
36-46.
33 Sigholm G, StyfJ. Stibacioniial pressure during diaguostic sluiuldri
tests. Clin Hiomech. 198H::i:l87-189.
34 Calis M, Akgtm K. Binane M, et al. Diagnostic valtie of clinical
diagnostic tests in subacromial impingement syndrome. Ann Winim
Dis. 2000:59:44-4 7.
35 MacDonald P. Clark P, Sutherland K. An analysis of the diagnostic
accuracy of the Hawkins and Neer subacromiai impingement signs.
/ Shouldfr EUmi< Surg. 2000:9:299-301,
.tfi Partington P, Broome C Diagnostic itijection arotnid the shotilder:
hit and miss? A cadaveric study of injecUon accnnicy. / Shoulder Elhom
Sttrg. 1998:7:147-1.50.
37Van]akad(i K. Tbe target acctiracy of subacn»mial injcition to the
38 Desnieules F. Côté CH. Fremont P. Tbeiapentic exercise and
orthopetiic manual therapy for iuipingement syndrome: a sysiematic
n-view. Clin I Sptnu Mrd. 2003:1 :i: 176-182.
39 Moher D. Scbul/ KF, Altuian 1). Tlu- CONSORT statement: revised
recommendation loi impnninií the qualiivoi rcporis oJ parallel-group
r;uidotni/e(l trials. /,1A/.1. 200! :2S5:1987-1991.
40 Kurtai5 tiüm-l VK, Ll|us^', Bilgiv .\. et al. Adding uliiasouiid in tbe
placehiMTontrulled trial. Phys Thn\ 2004:84:336-34:1
41 lirox JI, Stall PH. I.junggren .\E. Brevik JI. Arthroscopic surgery
compaicd with supemsed exercises in patients with rotator cuff disease
(stage II impingement syndrome). HMJ. 1993:307:899-903.
42 R0e C. Brox [I. Iliihmcr .VS, Vollestad NK. Mtiscle activation after
supemsed exercises in patients with cotatiir tendinosis. Arch Ph\.\ Med
¡irhahil. 2(tO():8]:(i7-72.
43 Rcddy AS. Mohr KJ. Pink MM,Jobe FW. Electromyograpliic analy.sis
ol the deltoid and rotator cuff miisdi-s in persons with subaciotuial
inipingfiuent. /.S/zim/i/iTWAoH' Surg. 2OOit:9:r) 19-523.
44 Wimen* JC. Sobel JS, Grocnier KH, et al. The long-term courae of
shoulder complaints: a prospective study iti general pra( tice. Hheiima-
tology. 1999:38:160-163.
Physical Therapy . Volume 85 . Number 6 . June 2005
Appendix 1.
Description ot Placement of the Acupuncture Needles

Acupuncture
Point Location Depth Angle

U 4 Hegu Befween the first and second metacarpal bones. On the 0.5-0.8 cun" 90"
radial side, in level with the middle of the second
metacarpal bone and at the highest point of the
interosseus dorsalis muscle when the thumb is
adducted.
LI 14 Binao Lateral side of the humérus, in front of the insertion of 0.5-0.7 cun 90"
deltoid muscle (in line with LI 15)
LI ^5 Jianyu Distal to the anterior part of acromion, in the anterior 0.7-1 cun 45° distal, longitudinal to the humérus
hollow in the deltoid muscle, appearing during
abduction of the glenohumeral ¡oint
LU 1 Zhongfu Anterior of the processus coracoideus 0.3-0.5 cun 90" toward ihe palpable part of the
processus coracoideus
TE 14 Jianliao Distal to the posterior part of acromion, in the dorsal 0.7-1 cun 45" distal, longitudinal to the humérus
hollow in the deltoid muscle, appearing during
abduction of the glenohumeral ¡oint

t iiin=llic width ut ÚU' thimib.

Physical Therapy . Volume 85 . Number 6 . June 2005 Johansson et al . 499

Appendix 2.
Home Exercise Program

Home exercise program, part I:

Perform the program once 0 day between weeks 1 and 5.
Note each time In your home exercise adherence log.

1. Seoted under, for example, a hat rack, eievate the arms

alternately by pulling in the sling [use a skipping rope or similar
item).

20 repetitions

2. Lie on the side; rest your upper arm along the side of the trunk. Put a small pillow in the axilla ond bend your elbow to about 90', rotate
externally and then Tower it slowly.

30 repetitions

(continued}

500 . Johansson et al Physical Theropy . Volume 85 . Number 6 . June 2005

Appendix 2.
Continued

Home exercise program, part II:

Perform the program once every other day between weeks 4 and 5.
Note each time in your home exercise adherence log.

Strengthening exercises:
3. Bend your elbow to 90°. Stand in a doorway
and press your fist ogainst doorpost in the fallowing manner: use a pillow in the
axilla, no movement shall occur.

10 repetitions in each direction

Foieward Backward

Internal rotation External To the side (abduction)

4, Standing with both elbows bent to 90°.

Externafrotation of the shoulder, using a section of tubing as
resistance, and then return slowly to starting position. Put pillows
or fowels in the axilla.

15 repetitions x 2

Pain during these exercises should not remain more than 10 to 15

minutes after the end of program. If a longer duration is
experienced, decrease either the resistance or the force produced.

Physical Therapy , Volume 85 . Number 6 , June 2005 Johansson et al . 501

Copyright of Physical Therapy is the property of American Physical Therapy Association and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.