ARTICLE IN PRESS

J Shoulder Elbow Surg (2017) ■■, ■■–■■

www.elsevier.com/locate/ymse

ORIGINAL ARTICLE

Randomized controlled trial of supervised

physiotherapy versus a home exercise program
after hydrodilatation for the management of
primary frozen shoulder
Paul M. Robinson, FRCS(Tr&Orth)a,*, Jennie Norris, BSc(Hons)b,
Christopher P. Roberts, FRCS(Tr&Orth)b
a
Peterborough and Stamford Hospitals NHS Foundation Trust, Peterborough, UK
b
The Ipswich Hospital NHS Trust, Ipswich, UK

Hypothesis and background: Hydrodilatation and physiotherapy are commonly used treatments for primary
frozen shoulder. Little is known about the optimal form of physiotherapy. This study reports a random-
ized controlled trial comparing 2 forms of physiotherapy after hydrodilatation. The null hypothesis was
that there would be no difference between the 2 groups at 1 year as measured by the Oxford Shoulder
Score (OSS).
Methods: We randomized 41 patients undergoing hydrodilatation for primary frozen shoulder into 2 treat-
ment groups: group 1 (n = 20) underwent supervised physiotherapy in addition to a home exercise program,
and group 2 (n = 21) followed a self-directed home exercise program in isolation. Assessment was carried out
by a blinded research nurse at baseline, 4 weeks, 3 months, 6 months, and 1 year. The primary outcome measure
was the OSS. Other measures were range of movement, visual analog scale pain score, and EQ-5D index.
Results: There was no significant difference between the treatment groups at any time point as measured by
the OSS or EQ-5D index. In group 1, the OSS improved significantly from 25.00 (95% confidence interval
[CI], 21.92-28.08) at baseline to 38.29 (95% CI, 34.01-42.58; P < .0001) at 4 weeks and 43.71 (95% CI,
41.61-45.80; P < .0001) at 1 year. In group 2, the OSS improved significantly from 26.60 at baseline (95%
CI, 22.50-30.70) to 40.07 (95% CI, 36.77-43.36; P < .0001) at 4 weeks and 43.00 (95% CI, 39.69-46.31;
P < .0001) at 1 year. All outcome measures improved significantly from baseline to 4 weeks.
Conclusion: In this group of patients, after a hydrodilatation procedure for the treatment of primary frozen
shoulder, there was no significant difference in clinical outcomes between supervised physiotherapy in
addition to a home exercise program and a self-directed home exercise program in isolation.
Level of evidence: Level II; Randomized Controlled Trial; Treatment Study
© 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
Keywords: Shoulder; adhesive capsulitis; frozen shoulder; hydrodilatation; hydrodistension; physiotherapy

This study was performed at The Ipswich Hospital NHS Trust, Ipswich, UK. Regional ethics committee (REC) approval was obtained (National Research Ethics
Service (NRES) Committee East of England—Cambridge East, REC reference 12/EE/0403, Integrated Research Application System (IRAS) project number
103304).
*Reprint requests: Paul M. Robinson, FRCS(Tr&Orth), Peterborough and Stamford Hospitals NHS Foundation Trust, Department of Trauma and Orthopaedics,
Peterborough City Hospital, Edith Cavell Campus, Bretton Gate, Peterborough PE3 9GZ, UK.
E-mail address: paulrobinson79@doctors.org.uk (P.M. Robinson).

1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
http://dx.doi.org/10.1016/j.jse.2017.01.012
 ARTICLE IN PRESS
2 P.M. Robinson et al.
Frozen shoulder is a common and debilitating disorder,
accounting for around 5% of all shoulder disease, with an
estimated incidence of between 0.75% and 5.0% in the
general population.5,13 It presents with a characteristic gradual
onset of shoulder pain and stiffness.18 The patient typically
has severe shoulder pain with night pain in the initial
stages, causing them to seek treatment from medical profes-
sionals. Frozen shoulder is not a benign condition, and
long-term clinical outcome studies have documented ongoing
pain and disability several years after the diagnosis.22 There
is no gold-standard initial treatment for frozen shoulder,
and many different management strategies are used, with
hydrodilatation and physiotherapy being two of the most
common.17 In the senior author’s unit, the first-line treat-
ment for primary frozen shoulder usually consists of a
combination of fluoroscopically guided hydrodilatation with
a mixture of local anesthetic and steroid, followed by
supervised one-to-one physiotherapy. Hydrodilatation
(arthrographic distension with steroid and saline solution)
has been reported to provide short-term benefits regarding
pain, range of movement (ROM), and function for frozen
shoulder.2 The use of physiotherapy (manual therapy and
exercise) is routine for many shoulder conditions, including
frozen shoulder. However, high-quality evidence for the
efficacy of physiotherapy in patients with frozen shoulder is
currently lacking.17 Furthermore, the efficacy of supervised
physiotherapy compared with that of a home exercise
program has not been previously established after
hydrodilatation. Unsupervised, home-based exercise pro-
grams have provided equivalent results to traditional supervised
physiotherapy after other musculoskeletal interventions, such
as rotator cuff repair12 and total knee replacement.11 In the
shoulder, they have also been used after breast and axillary
surgery.15 We hypothesized that a self-directed home exer-
cise program would produce equivalent clinical outcomes
to a supervised face-to-face physiotherapy program after
hydrodilatation for primary frozen shoulder.
The aim of this research was to establish if there is a dif-
ference in the clinical outcome as measured by the Oxford
Shoulder Score (OSS) between supervised physiotherapy ac-
companied by a home exercise program and a self-directed
home exercise program in isolation after patients under-
went hydrodilatation for primary frozen shoulder. The null
hypothesis of the study was that there would be no differ-
ence between the mean OSS of the 2 groups of patients at
12 months after treatment.
Materials and methods
The study was a randomized controlled trial comparing super-
vised physiotherapy in addition to a home exercise program
(group 1) with a self-directed home exercise program in isolation
(group 2). The study conformed to the CONSORT (CONsolidated
Standards Of Reporting Trials) 2010 statement.6 All participants
underwent an informed consent process before taking part in the
study.
Table I Criteria for diagnosis of frozen shoulder4 used in study
Shoulder pain for at least 1 month
Sleep disturbance: night pain or inability to lie on affected
side
Restriction of all active and passive movements
Restriction of passive glenohumeral external rotation by at
least 50% compared with normal side
Table II Study inclusion and exclusion criteria
Inclusion criteria
Primary frozen shoulder treated by hydrodilatation
Ability to understand and participate in study
Exclusion criteria
Glenohumeral osteoarthritis on radiographs
Previous shoulder procedures
Steroid medication
Other shoulder pathology
Inflammatory joint disease
Significant shoulder injury within 6 months
Neurologic or referred pain
Diagnosis
Cases of frozen shoulder were diagnosed in the orthopedic clinic
and by the musculoskeletal triage service at our hospital. We used
the diagnostic criteria described by Bulgen et al4 (Table I) for di-
agnosing frozen shoulder. The diagnosis was made by 2 experienced
musculoskeletal physiotherapists and a consultant shoulder surgeon
(C.P.R.). The inclusion and exclusion criteria are listed in Table II.
Screening and recruitment
Potential participants were screened by a research nurse for their
suitability for the study by reviewing referrals made for hydrodilatation
for primary frozen shoulder. Potential participants were also ap-
proached by the extended scope physiotherapist (J.N.) and consultant
shoulder surgeon (C.P.R.) in their respective clinics. If a patient was
thought to be suitable for the study, a letter of invitation and a par-
ticipant information sheet were issued and follow-up contact was
made by telephone. Participants underwent the consent process and
were recruited into the trial at a further hospital visit.
Randomization
Participants were randomized to 1 of 2 treatment groups by
drawing sealed opaque envelopes. For randomization purposes, 40
envelopes were created: 20 contained an allocation card to group
1 and 20 contained an allocation card to group 2. The cards were
shuffled by a blinded research nurse, placed into opaque sealed
envelopes, and shuffled again. The envelopes were then numbered
1 to 40. Once a participant was recruited, the next envelope in
numerical order was drawn and the participant was allocated to
the appropriate study group.
One participant who was randomized to group 2 was not
contactable for any of the assessments after the hydrodilatation
 ARTICLE IN PRESS
Physiotherapy versus home exercises after hydrodilatation
procedure. This allocation card was placed into a new envelope,
number 41 in the sequence, so that another participant could be ran-
domized in the original participant’s place. Therefore, 20 participants
were randomized to group 1 and 21 participants were randomized
to group 2.
All participants received the allocated treatment apart from the
following: 1 participant allocated to group 1 and 1 participant al-
located to group 2 declined the hydrodilatation procedure because
their symptoms had improved before the date it was scheduled to
be administered. In addition, 1 participant allocated to group 2 had
a full-thickness rotator cuff tear diagnosed on the arthrogram and
therefore did not undergo hydrodilatation.
Hydrodilatation procedure
Hydrodilatation was performed by 2 musculoskeletal radiologists
in the radiology department using an aseptic technique and
under fluoroscopic guidance. Needle position in the shoulder
joint was confirmed with radiopaque contrast. For the injections,
one radiologist used 1 mL of triamcinolone acetonide, 40 mg/mL
(Kenalog; Bristol-Myers Squibb Pharmaceuticals, Uxbridge,
UK), in a mixture of 9 mL of 0.25% bupivacaine (Marcain;
AstraZeneca UK, Cambridge, UK), 9 mL of iohexol (Omnipaque;
GE Healthcare AS, Oslo, Norway), and 20 mL of saline solution.
The other radiologist used 1 mL of triamcinolone acetonide,
40 mg/mL (Kenalog; Bristol-Myers Squibb Pharmaceuticals); 5 mL
of 1% lidocaine; 5 mL of iohexol (Omnipaque); and saline solu-
tion (usually 20-30 mL in total). Dilation of the joint continued
until capsular rupture was achieved or until pain stopped the
procedure.
Interventions
The home exercise program was composed by a specialist shoul-
der physiotherapist and included wall slides (into flexion and
abduction); active-assisted exercises with a stick into flexion, ab-
duction, internal rotation, and external rotation (ER); table-lean passive
stretches; scapula setting; and isometric strengthening. The use of
a hot pack was also advised.
The leaflet was reviewed by the hospital communications de-
partment to ensure readability and clarity for all levels of reading
ability. Diagrams of the exercises were also included in the leaflet
for clarity (Appendix S1).
Immediately after hydrodilatation, all participants (groups 1 and
2) were given the home exercise program leaflet and instructed to
begin the exercises on the same day. They were instructed to com-
plete these 5 times per day, beginning with 5 repetitions and increasing
up to 10 repetitions at each setting.
In addition, those participants in the supervised physiotherapy
group (group 1) received once-weekly appointments with a quali-
fied physiotherapist for 4 weeks, beginning the day after the
hydrodilatation procedure. Each appointment lasted approximate-
ly 20 minutes. It consisted of advice, exercise therapy, manual therapy,
and therapist-applied passive stretches, glenohumeral joint acces-
sory mobilizations, glenohumeral joint passive physiological
mobilizations, and cervical and thoracic spine accessory mobiliza-
tions. The physiotherapy treatment received during the sessions and
the advice and home exercise program were pragmatic and depen-
dent on patient need.
3
Assessment
Participants were assessed prospectively in a clinic by an indepen-
dent blinded research nurse who was not involved in the patients’
treatment. These assessments took place at baseline, 4 weeks, and
12 months. Additional telephone follow-up was conducted at 3 months
and 6 months.
The clinic assessments consisted of shoulder ROM measure-
ments and the completion of patient-reported outcome measures.
The patient-reported outcome measures consisted of the OSS,7 a visual
analog scale (VAS) for shoulder pain, and the EQ-5D-3L health ques-
tionnaire (English version for the United Kingdom).1,10 Telephone
assessment consisted of the OSS and EQ-5D.
The OSS is a subjective patient-reported questionnaire that con-
tains 12 questions derived from 2 parameters, pain and function.
Scores from each of the questions are added to produce a single score
ranging from 0 to 48, with 48 representing the best outcome.7 The
minimal clinically important difference (MCID) of the OSS is 5
points.
The EQ-5D-3L consists of 2 parts: the EQ-5D descriptive system
(EQ-5D) and the EQ visual analog scale (EQ VAS). The EQ-5D-
3L descriptive system comprises questions covering the following
5 dimensions: mobility, self-care, usual activities, pain/discomfort,
and anxiety/depression. Each dimension has 3 levels: no prob-
lems, some problems, and extreme problems. The EQ-5D descriptive
system score is converted into a single summary index by apply-
ing a formula that attaches weights to each of the levels in each
dimension. This formula is based on the valuation of EQ-5D health
states from general-population samples. A score of 1.0 represents
perfect health, and a score of 0.0 is equal to death. For the UK pop-
ulation, the EQ-5D index ranges from 1.0 to –0.594, based on the
time trade-off technique.
The EQ VAS records the respondent’s self-rated health on a ver-
tical VAS on which the endpoints are labeled “best imaginable health
state” and “worst imaginable health state.” The EQ VAS score ranges
from 0 (representing worst imaginable health state) to 100 (repre-
senting best imaginable health state).
Shoulder ROM measurements were made using a goniometer,
according to a predefined protocol. The movements measured were
active flexion, abduction, and ER, as well as passive glenohumeral
joint abduction and passive glenohumeral joint ER. Internal rota-
tion was measured as the highest anatomic level that the thumb could
reach when the participant actively placed his or her hand behind
his or her back. Data including patient demographic characteris-
tics, duration of symptoms, previous steroid injections, and medical
comorbidities were also collected.
Statistical analysis
We aimed to recruit 40 patients in this study on the basis of previ-
ous referral rates for hydrodilatation at our hospital and a realistic
time frame for recruiting participants and concluding the study. This
would form the basis for a future powered study.
At the time of the trial design, there were no robust pre-
existing data for the OSS after hydrodilatation and physiotherapy
for frozen shoulder. Therefore, a power analysis could not be
performed.
Statistical analysis was performed using GraphPad InStat
(GraphPad Software, La Jolla, CA, USA). Data were analyzed
within the groups to assess the effects of each intervention on the
 ARTICLE IN PRESS
4 P.M. Robinson et al.

Enrollment
Assessed for eligibility (n = 226)

Excluded (n = 187)
Not meeting inclusion criteria (n = 91)
Declined to participate (n = 64)
Other reasons (n = 32)

Randomized (n = 41)

Allocation
Allocated to Group 1 (n = 20) Allocated to Group 2 (n = 21)
Received allocated intervention (n = 19) Received allocated intervention (n = 19)
Did not receive allocated intervention (n = 1) Did not receive allocated intervention (n = 2)
• (Condition improved and declined hydrodilatation • (Condition improved and declined hydrodilatation
n = 1) n=1, hydrodilatation not performed as rotator cuff
tear noted on arthrogram n = 1)

Follow-Up

Lost to follow-up (n = 0) Lost to follow-up (n = 1)

Discontinued intervention (n = 0) • (Unable to contact n = 1)
Discontinued intervention (n = 0)

Analysis

Analyzed (n = 17) Analyzed (n = 15)

Excluded from analysis (n = 2) Excluded from analysis (n = 3)
• (Required second hydrodilatation n = 1) • (Required second hydrodilatation n = 2)
• (Required arthroscopic procedure n = 1) • (Previous shoulder injury n = 1)

Figure 1 CONSORT (CONsolidated Standards Of Reporting Trials) flow diagram.

outcome measures and analyzed between groups to compare the
effects of the intervention. Discrete data were analyzed using the
Fisher exact test. Continuous data were tested for normality of
distribution using the Kolmogorov-Smirnov method. Unpaired
parametric data were analyzed using a 2-tailed Student t test, and
nonparametric data were analyzed with a 2-tailed Mann-Whitney
test. Paired parametric data were analyzed using a 2-tailed paired t
test, and nonparametric data were analyzed with a 2-tailed Wilcoxon
matched-pairs signed rank test. P < .05 was considered statistical-
ly significant. Graphs were constructed using Microsoft Excel
(Microsoft, Redmond, WA, USA).
Results
In total, 226 patients were screened (Fig. 1). Forty-one pa-
tients were recruited and randomized. The flow of participants
through the study is represented in Fig. 1. Baseline data for
the groups were similar in all respects apart from the dura-
tion of shoulder symptoms (Table III).
Oxford Shoulder Score
At 4 weeks, the mean OSS in group 1 (n = 17) had improved
from 25.00 (95% confidence interval [CI], 21.92-28.08) to 38.29
(95% CI, 34.01-42.58; P < .0001, paired t test). The mean im-
provement in the OSS was 13.29 (95% CI, 8.96-17.63).
At 4 weeks, the mean OSS in group 2 (n = 15) had im-
proved from 26.60 (95% CI, 22.50-30.70) to 40.07 (95% CI,
36.77-43.36; P < .0001, paired t test). The mean improve-
ment in the OSS was 13.47 (95% CI, 8.89-18.05).
At 1 year, the mean OSS in group 1 (n = 17) had im-
proved from 25.00 (95% CI, 21.92-28.08) to 43.71 (95% CI,
41.61-45.80; P < .0001, Wilcoxon matched-pairs signed rank
test). The mean improvement in the OSS was 18.71 (95% CI,
15.60-21.82).
At 1 year, the mean OSS in group 2 (n = 15) had im-
proved from 26.60 (95% CI, 22.50-30.70) to 43.00 (95% CI,
39.69-46.31; P < .0001, paired t test). The mean improve-
ment in the OSS was 16.40 (95% CI, 11.34-21.46) (Fig. 2).
 ARTICLE IN PRESS
Physiotherapy versus home exercises after hydrodilatation 5

Table III Baseline participant characteristics

Home exercises Physiotherapy P value
Mean age (95% CI), y 55.2 (52.5-58.0) 57.9 (53.2-62.5) .3119*
Female (%) 15 (71) 13 (65) .7442†
Mean duration of symptoms (95% CI), mo 6.5 (5.5-7.5) 8.5 (7.2-9.7) .0145*
Thyroid disease (%) 2 (9.5) 0 (0) .4878†
Diabetes (%) 4 (19) 1 (5) .3433†
Previous steroid injection (%) 5 (24) 4 (20) > .99†
CI, confidence interval.
* Student t test.
†
Fisher exact test.

48

40
Oxford Shoulder Score

32

24

16

0
Baseline 4 week 3 month 6 month 12 month

Home Exercises Physiotherapy

Figure 2 Oxford Shoulder Score at baseline and follow-up time points.

There was no significant difference between the OSS of
the 2 groups at any time point (See Supplementary Table I).
The mean OSS of both groups improved by greater than the
MCID (5 points) at both 4 weeks and 1 year.
VAS pain score
At 4 weeks, the mean VAS pain score in group 1 (n = 17)
had improved from 7.06 (95% CI, 6.30-7.82) to 2.12 (95%
CI, 1.21-3.02; P < .0001, Wilcoxon matched-pairs signed rank
test). The mean improvement was 4.94 (95% CI, 3.94-5.95).
At 4 weeks, the mean VAS pain score in group 2 (n = 15)
had improved from 5.13 (95% CI, 3.81-6.46) to 1.53 (95%
CI, 0.65-2.42; P = .0017, Wilcoxon matched-pairs signed rank
test). The mean improvement was 3.60 (95% CI, 1.75-5.45).
At 1 year, the mean VAS pain score in group 1 (n = 17)
had improved from 7.06 (95% CI, 6.30-7.82) to 2.12 (95%
CI, 1.12-3.11; P < .0001, Wilcoxon matched-pairs signed rank
test). The mean improvement was 4.94 (95% CI, 3.81-6.07).
At 1 year, the mean VAS pain score in group 2 (n = 15)
had improved from 5.13 (95% CI, 3.81-6.46) to 2.13 (95%
CI, 0.50-3.77; P = .0353, Wilcoxon matched-pairs signed rank
test). The mean improvement was 3.00 (95% CI, 0.43-5.57)
(Fig. 3).
At baseline, the VAS pain score was significantly higher
in group 1. There was no significant difference at all other
time points (Supplementary Table I).
Passive glenohumeral joint external rotation
At 4 weeks, the mean ER in group 1 (n = 16) had improved
from 20° (95% CI, 16°-25°) to 33° (95% CI, 22°-45°;
P = .0084, Wilcoxon matched-pairs signed rank test). The mean
improvement was 13° (95% CI, 2°-24°).
At 4 weeks, the mean ER in group 2 (n = 14) had im-
proved from 24° (95% CI, 13°-34°) to 40° (95% CI, 27°-
53°; P = .0040, Wilcoxon matched-pairs signed rank test). The
mean improvement was 16° (95% CI, 7°-26°).
At 1 year, the mean ER in group 1 (n = 17) was 26° (95%
CI, 21°-31°). This was not significantly different from base-
line (P = .0507, Wilcoxon matched-pairs signed rank test).
The mean improvement was 6° (95% CI, 0°-12°). At 1 year,
the mean ER in group 2 (n = 15) was 25° (95% CI, 19°-
30°). This was not significantly different from baseline
(P = .5417, Wilcoxon matched-pairs signed rank test). The
mean improvement was 1° (95% CI, 9°-11°) (Fig. 4). There
was no significant difference between the ER of the 2 groups
at any time point (Supplementary Table I).
 ARTICLE IN PRESS
6 P.M. Robinson et al.

9
8
7

VAS pain score

6
5
4
3
2
1
0
Baseline 4 week 12 month

Home Exercises Physiotherapy

Figure 3 Visual analog scale (VAS) pain score at baseline and follow-up time points.

60
Passive GHJ External Rotation (⁰)

50

40

30

20

10

0
Baseline 4 week 12 month

Home Exercises Physiotherapy

Figure 4 Passive glenohumeral joint (GHJ) external rotation at baseline and follow-up time points.

EQ-5D index Other parameters and complications

At 4 weeks, the mean EQ-5D index in group 1 (n = 17) had All other clinical parameters improved for each group over
improved from 0.591 (95% CI, 0.462-0.721) to 0.824 (95% the course of the study (Supplementary Table I). There were
CI, 0.767-0.882; P = .0002, Wilcoxon matched-pairs signed no complications experienced in relation to either of the treat-
rank test). At 4 weeks, the mean EQ-5D index in group 2 ment regimens.
(n = 15) had improved from 0.645 (95% CI, 0.519-0.770) to
0.864 (95% CI, 0.786-0.942; P = .0007, Wilcoxon matched-
pairs signed rank test). Discussion
At 1 year, the mean EQ-5D index in group 1 (n = 17) had
improved from baseline to 0.684 (95% CI, 0.492-0.876). This The main finding of our study is that there is no significant
did not quite reach statistical significance (P = .0577, Wilcoxon difference in the OSS of the 2 treatment groups at any time
matched-pairs signed rank test). At 1 year, the mean EQ-5D point after hydrodilatation for primary frozen shoulder.
index in group 2 (n = 15) had improved from baseline to 0.882 Neither the primary outcome measure nor any of the
(95% CI, 0.806-0.958; P = .0017, Wilcoxon matched-pairs secondary outcome measures indicate that there is any
signed rank test) (Fig. 5). There was no significant differ- difference in the outcomes of supervised physiotherapy and
ence between the EQ-5D indexes of the 2 groups at any time a home exercise program when compared with the out-
point (Supplementary Table I). comes of a self-directed home exercise program. Kelley
 ARTICLE IN PRESS
Physiotherapy versus home exercises after hydrodilatation
1.200
1.000
0.800
EQ-5D Index
0.600
0.400
0.200
0.000
Baseline
Figure 5
et al14 have suggested that there is no clear evidence to
guide clinicians on which patients will benefit from super-
vised physiotherapy and who will benefit from home exercises
alone. We are unable to clarify this point any further from
the results of this study.
We have shown that hydrodilatation followed by either
of the studied treatments results in a significant improve-
ment in the OSS, VAS pain score, passive glenohumeral
ER, and EQ-5D index at both 4 weeks and 1 year (apart
from a nonsignificant improvement in the EQ-5D index in
group 1 at 1 year). We are unable to deduce exactly what
proportion of the observed effect was related to the
hydrodilatation and steroid, what was related to the study
interventions, and what was attributable to the natural
history of the condition.
Other studies have found home exercise programs to be
effective in the management of frozen shoulder. Kivimäki
et al16 compared patients treated with a home exercise
program (after 2 physiotherapy advice sessions) with pa-
tients treated with manipulation under anesthesia and a
home exercise program (after the same advice sessions).
Other than a slight increase in range of motion at 6 weeks
and 3 months, the group performing just a home exercise
program did not differ at any follow-up point in terms of
pain or working ability.
The effectiveness of traditional physiotherapy after
hydrodilatation has previously been questioned. Buchbinder
et al3 reported that physiotherapy (manual therapy and di-
rected exercise) after hydrodilatation provided no additional
benefits in terms of pain, function, and quality of life when
compared with placebo (sham ultrasound). Physiotherapy did
result in sustained greater active range of shoulder move-
ment and participant-perceived improvement up to 6 months.3
Ryans et al20 conducted a randomized trial comparing intra-
articular steroid injection alone versus steroid injection with
standardized physiotherapy, placebo injection, and placebo
injection with standardized physiotherapy. They reported that
at 6 weeks, the shoulder disability questionnaire scores had
improved significantly more in patients who had received a
steroid injection compared with those who had not. They also
7
4 week 3 month 6 month 12 month
Home Exercises Physiotherapy
EQ-5D index at baseline and follow-up time points.
found that physiotherapy improved passive ER at 6 weeks.
However, by 16 weeks, all groups were similar in terms of
their outcome scores. We found no significant difference in
ER (passive or active) between the 2 groups in our study. It
is possible that the main treatment benefit in our study was
caused by the steroid injections, but we are unable to draw
any conclusions on this.
Diercks and Stevens8 performed a prospective cohort study
comparing intensive physiotherapy and passive stretching with
supervised neglect for the treatment of frozen shoulder. Pa-
tients in the supervised-neglect group were instructed to
perform pendulum exercises and active exercises within the
painless range and to resume all activities that were toler-
ated. Despite there being no difference in the mean Constant
score between the 2 groups at baseline, the supervised-
neglect group had a significantly better Constant score at all
other time points, up to and including 24 months. The authors
hypothesized that the physiotherapy and stretching place more
stress on the soft tissues and could therefore stimulate the in-
flammatory response and fibroblast activity within the shoulder
joint capsule, leading to worse outcomes.
Russell et al19 conducted a randomized controlled trial com-
paring the efficacy of 3 treatment regimens—exercise class
plus home exercises, individual multimodal physiotherapy plus
home exercises, and home exercises alone—for the treat-
ment of primary frozen shoulder (with at least a 3-month
history to minimize the number of patients in the pain-
predominant phase). All groups showed a significant
improvement in their Constant score and OSS over the 1-year
trial period and at each follow-up point. The Constant score
and the OSS of the exercise class group improved signifi-
cantly more than those of both the home exercise group and
the individual physiotherapy group. Elevation and ER im-
proved significantly in all groups, but the improvement was
significantly greater in both of the physiotherapy interven-
tion groups at all time points compared with the home exercise
group.
There has been one similar randomized trial to ours, con-
ducted in patients with impingement syndrome. Senbursa et al21
compared a home exercise program (after instruction) with
 ARTICLE IN PRESS
8 P.M. Robinson et al.
physiotherapy (mobilization, stretching, ice application, and
strengthening) for the treatment of shoulder impingement. In
their randomized controlled trial, they reported a signifi-
cantly greater improvement in outcomes (ROM, Neer
questionnaire, and shoulder satisfaction score) for those pa-
tients in the physiotherapy and manipulation group.
An interesting finding of our study was that some of the
outcome measures such as passive glenohumeral joint ER and
EQ-5D improved greatly over the initial 4 weeks but in some
cases deteriorated by 12 months. This could be due to several
factors, such as an initial beneficial effect of the steroid that
was seen at 4 weeks but had worn off by 12 months. Another
explanation could be the initial beneficial effect of
hydrodilatation or capsular rupture. Alternatively, it could rep-
resent disease progression or a decrease in compliance with
the home exercise program that each group had been given.
One of the strengths of our study is that we had a very
low rate of loss to follow-up (1 participant); moreover, we
were able to review patients both soon after the interven-
tions (at 4 weeks) and at 1 year. This is also the first
randomized trial to compare physiotherapy with a home ex-
ercise program after hydrodilatation.
The study has several limitations. We are unable to
comment on the compliance of either of the groups regard-
ing the frequency with which the home exercises were
performed or their completeness. As both groups were issued
the home exercise program, it is likely that both groups would
have experienced compliance issues. It is uncertain how much
this could have influenced the results of the study.
The 2 radiologists used different types of local anesthet-
ic for their hydrodilatation procedures. However, the steroid
type and amount used were identical.
The study lacks a no-treatment control group to repre-
sent the true natural history of the condition. We feel that it
would be clinically and ethically difficult to recruit partici-
pants to a trial that included such a group.
Four participants were excluded from the study (two
from each group) because they received additional interven-
tions. The three additional hydrodilatations were performed
because the patients returned early via referral from their
general practitioners because of continued pain. After as-
sessment and counseling about treatment options, they
underwent further injection. One participant, who had type
2 diabetes, underwent an arthroscopic capsular release after
being referred by the participant’s general practitioner because
of continued pain. After assessment and counseling, this
patient underwent surgical treatment. We felt that it would
confound the study results if these participants were in-
cluded in the final analysis, as their further treatments
would affect the outcome of their physiotherapy and home
exercise programs.
One participant was excluded after a history of recent shoul-
der injury came to light. This was an exclusion criterion of
the study.
A sample size calculation was performed using the
data generated by this study with PS Power and Sample Size
Calculations, version 3.1.2 (http://biostat.mc.vanderbilt.edu/
PowerSampleSize).9 To achieve a power of 0.8 with α set at
.05 and with estimation of the MCID of the OSS to be 5 points
with an SD of 6.88, the study would require 31 participants
per group. The lack of good-quality pre-existing outcome data
in the literature in this group of patients prevented a mean-
ingful sample size calculation. This study has, however,
provided useful data for a future larger study. We can also
conclude that within this particular cohort of patients, there
was no significant difference in outcomes between the 2
groups.
Conclusion
This study has shown that in this particular group of pa-
tients, after a hydrodilatation procedure for the treatment
of primary frozen shoulder, there was no significant dif-
ference in the clinical outcomes between supervised
physiotherapy in addition to a home exercise program and
a self-directed home exercise program in isolation. There
was a marked clinically and statistically significant im-
provement in the clinical outcome measures in both
treatment groups 4 weeks after hydrodilatation that was
maintained at 1 year after the intervention.
Acknowledgments
The authors acknowledge the contribution of Maggie
Dawson, The Ipswich Hospital NHS Trust Research and
Development Department and The Ipswich Hospital NHS
Trust Musculoskeletal Clinical Assessment Service.
Disclaimer
The authors, their immediate families, and any research
foundations with which they are affiliated have not re-
ceived any financial payments or other benefits from any
commercial entity related to the subject of this article.
Appendix
Supplementary data
Supplementary data to this article can be found online at
http://dx.doi.org/10.1016/j.jse.2017.01.012
References
1. Brooks R. EuroQol: the current state of play. Health Policy 1996;37:53-
72.
2. Buchbinder R, Green S, Youd JM, Johnston RV, Cumpston M.
Arthrographic distension for adhesive capsulitis (frozen shoulder).
Cochrane Database Syst Rev 2008;(1):CD007005. http://dx.doi.org/10
.1002/14651858.CD007005
 ARTICLE IN PRESS
Physiotherapy versus home exercises after hydrodilatation
3. Buchbinder R, Youd JM, Green S, Stein A, Forbes A, Harris A, et al.
Efficacy and cost-effectiveness of physiotherapy following glenohumeral
joint distension for adhesive capsulitis: a randomized trial. Arthritis
Rheum 2007;57:1027-37. http://dx.doi.org/10.1002/art.22892
4. Bulgen DY, Binder AI, Hazleman BL, Dutton J, Roberts S. Frozen
shoulder: prospective clinical study with an evaluation of three treatment
regimens. Ann Rheum Dis 1984;43:353-60.
5. Bunker T. Time for a new name for frozen shoulder—contracture of
the shoulder. Shoulder Elbow 2009;1:4-9.
6. CONSORT Group. CONSORT statement. <http://www.consort
-statement.org/>; 2010 [accessed 23/08/2016].
7. Dawson J, Rogers K, Fitzpatrick R, Carr A. The Oxford shoulder score
revisited. Arch Orthop Trauma Surg 2009;129:119-23. http://dx.doi
.org/10.1007/s00402-007-0549-7
8. Diercks RL, Stevens M. Gentle thawing of the frozen shoulder: a
prospective study of supervised neglect versus intensive physical therapy
in seventy-seven patients with frozen shoulder syndrome followed up
for two years. J Shoulder Elbow Surg 2004;13:499-502. http://dx.doi
.org/10.1016/j.jse.2004.03.002
9. Dupont WD, Plummer WD. Power and sample size calculations. A
review and computer program. Control Clin Trials 1990;11:116-28.
10. EuroQol Group. EuroQol—a new facility for the measurement of
health-related quality of life. Health Policy 1990;16:199-208.
11. Florez-García M, García-Pérez F, Curbelo R, Pérez-Porta I, Nishishinya
B, Rosario Lozano MP, et al. Efficacy and safety of home-based exercises
versus individualized supervised outpatient physical therapy programs
after total knee arthroplasty: a systematic review and meta-analysis. Knee
Surg Sports Traumatol Arthrosc 2016;http://dx.doi.org/10.1007/s00167
-016-4231-x. Epub ahead of print.
12. Hayes K, Ginn KA, Walton JR, Szomor ZL, Murrell GA. A randomized
clinical trial evaluating the efficacy of physiotherapy after rotator cuff
repair. Aust J Physiother 2004;50:77-83. http://dx.doi.org/10.1016/
S0004-9514(14)60099-4
13. Hsu JE, Anakwenze OA, Warrender WJ, Abboud JA. Current review
of adhesive capsulitis. J Shoulder Elbow Surg 2011;20:502-14.
http://dx.doi.org/10.1016/j.jse.2010.08.023
9
14. Kelley MJ, McClure PW, Leggin BG. Frozen shoulder: evidence and
a proposed model guiding rehabilitation. J Orthop Sports Phys Ther
2009;39:135-48. http://dx.doi.org/10.2519/jospt.2009.2916
15. Kilgour RD, Jones DH, Keyserlingk JR. Effectiveness of a self-
administered, home-based exercise rehabilitation program for women
following a modified radical mastectomy and axillary node dissection:
a preliminary study. Breast Cancer Res Treat 2008;109:285-95.
http://dx.doi.org/10.1007/s10549-007-9649-x
16. Kivimäki J, Pohjolainen T, Malmivaara A, Kannisto M, Guillaume J,
Seitsalo S, et al. Manipulation under anesthesia with home
exercises versus home exercises alone in the treatment of frozen
shoulder: a randomized, controlled trial with 125 patients. J Shoulder
Elbow Surg 2007;16:722-6. http://dx.doi.org/10.1016/j.jse.2007.02
.125
17. Page MJ, Green S, Kramer S, Johnston RV, McBain B, Chau M, et al.
Manual therapy and exercise for adhesive capsulitis (frozen shoulder).
Cochrane Database Syst Rev 2014;(8):CD011275. http://dx.doi.org/10
.1002/14651858.CD011275
18. Reeves B. The natural history of the frozen shoulder syndrome. Scand
J Rheumatol 1975;4:193-6.
19. Russell S, Jariwala A, Conlon R, Selfe J, Richards J, Walton M. A
blinded, randomized, controlled trial assessing conservative management
strategies for frozen shoulder. J Shoulder Elbow Surg 2014;23:500-7.
http://dx.doi.org/10.1016/j.jse.2013.12.026
20. Ryans I, Montgomery A, Galway R, Kernohan WG, McKane R. A
randomized controlled trial of intra-articular triamcinolone
and/or physiotherapy in shoulder capsulitis. Rheumatology
(Oxford) 2005;44:529-35. http://dx.doi.org/10.1093/rheumatology/
keh535
21. Senbursa G, Baltaci G, Atay A. Comparison of conservative treatment
with and without manual physical therapy for patients with shoulder
impingement syndrome: a prospective, randomized clinical trial. Knee
Surg Sports Traumatol Arthrosc 2007;15:915-21. http://dx.doi.org/
10.1007/s00167-007-0288-x
22. Shaffer B, Tibone JE, Kerlan RK. Frozen shoulder. A long-term
follow-up. J Bone Joint Surg Am 1992;74:738-46.