Abstract
Background: Idiopathic adhesive capsulitis
is a commonly recognized but poorly understood cause of a painful and
stiff shoulder. Although most orthopaedic literature supports treatment
with physical therapy and stretching exercises, some studies have
demonstrated late pain and functional deficits. The purpose of this
study was to evaluate the outcome of patients with idiopathic adhesive
capsulitis who were treated with a stretching-exercise program.
Methods: Seventy-five consecutive patients (seventy-seven shoulders)
with phase-II idiopathic adhesive capsulitis were treated with use
of a specific four-direction shoulder-stretching exercise program
and evaluated prospectively. The initial evaluation included the
recording of a detailed medical and orthopaedic history and assessment
of pain, range of motion, and function. The outcome evaluation included
assessment of pain, range of motion, and function; completion of
the Disabilities of the Arm, Shoulder, and Hand (DASH) Questionnaire; and
completion of the Short Form-36 (SF-36) Health Survey. The mean
duration of follow-up was twenty-two months (range, twelve to forty-one
months). One patient died prior to the final evaluation, and three
patients were lost to follow-up.
Results: Sixty-four (90 percent) of the patients
reported a satisfactory outcome. Seven (10 percent) were not satisfied
with the outcome, and five (7 percent) underwent manipulation and/or
arthroscopic capsular release. The outcomes of the patients who
did not have manipulation or capsular release were evaluated. There
were significant improvements in the scores for pain at rest (from
a mean of 1.57 points before treatment to a mean of 1.16 points
at the final evaluation; p < 0.001) and pain with activity (from
a mean of 4.12 points before treatment to a mean of 1.33 points
at the final evaluation; p < 0.0001). On the average, active
forward elevation increased 43 degrees, active external rotation increased
25 degrees, passive internal rotation increased eight vertebral
levels, and the glenohumeral rotation arc at 90 degrees of abduction increased
72 degrees (p < 0.00001). The number of "yes" responses to the
Simple Shoulder Test increased from a mean of 4.1 (of a possible
twelve) to a mean of 10.75 (p < 0.00001). Despite the significant
improvements and the high rate of patient satisfaction, there were
still significant differences in the pain and motion of the affected
shoulder when compared with those of the unaffected, contralateral
shoulder (p < 0.00001).
At the final outcome evaluation, the DASH scores demonstrated
limitations when compared with known population norms, whereas the
profiles of the SF-36 were comparable with those of age and gender-matched
control populations.
Prior treatment with physical therapy and a Workers' Compensation
claim or pending litigation were the only variables that were associated
with the eventual need for manipulation or capsular release. Male
gender and diabetes mellitus were associated with worse motion at
the final evaluation. Patients with a greater severity of pain with
activity at the initial evaluation had significantly lower DASH
scores at the final evaluation, and patients with lower initial
scores on the Simple Shoulder Test had comparatively lower scores
on the Simple Shoulder Test at the outcome evaluation.
Conclusions: The vast majority of patients who
have phase-II idiopathic adhesive capsulitis can be successfully treated
with a specific four-direction shoulder-stretching exercise program.
Although measurable limitations and deficiencies were noted at the
outcome evaluation, these appeared to be acceptable to most of the
patients and did not affect their general health status. Patients
with more severe pain and functional limitations before treatment
had relatively worse outcomes. More aggressive treatment such as
manipulation or capsular release was rarely necessary, and the efficacy
of early use of these treatments should be further studied.
Adhesive capsulitis is a common but poorly understood syndrome
of painful shoulder stiffness. Frozen shoulder syndrome was first
described by Duplay in 18726.
He used the term peri-arthritis scapulo-humerale and believed that
manipulation under anesthesia had a role in its treatment. In 1934,
Codman used the term frozen shoulder to describe this condition4. He stated that most cases resolved
in about two years without treatment. In 1945, Neviaser coined the
term adhesive capsulitis to reflect his findings at surgery and
autopsy in patients treated for a painful, stiff shoulder17. More recently, Zuckerman and Cuomo
defined frozen shoulder, or idiopathic adhesive capsulitis, as a
condition of uncertain etiology characterized by substantial restriction
of both active and passive shoulder motion that occurs in the absence
of a known intrinsic shoulder disorder33.
Although adhesive capsulitis is generally considered to be a
self-limiting condition that can be treated with physical therapy,
the best treatment has been the subject of extensive investigation10,25,28,30. A variety of different
treatments have been recommended, and numerous studies have demonstrated
successful results. The types of treatment have included benign
neglect10, chiropractic manipulation22, oral corticosteroids3,24, injection of corticosteroids26,28, physical therapy exercises and
modalities11,13,15,18, brisement7,8,27, manipulation under anesthesia12,18,21,24, and arthroscopic19,23,32 and open18,20 releases
of the contracture. Recent studies have emphasized the surgical
management of recalcitrant shoulder stiffness32.
Many of these studies have been flawed because they have lacked
objective and subjective outcome criteria.
The purpose of the present study was to prospectively evaluate
the outcome of treatment of idiopathic adhesive capsulitis with
a specific, supervised stretching-exercise program. Outcomes were
determined by the assessment of subjective and objective parameters.
Inclusion Criteria
Between September 1993 and September 1996, seventy-five consecutive
patients with phase-II idiopathic adhesive capsulitis16 were evaluated prospectively by
the senior author (A. G.). Phase-II idiopathic adhesive capsulitis was
defined as the stage of involvement in which the patient had pain
and globally limited shoulder motion. No attempt was made to specifically
recruit patients for inclusion in this study.
The criteria for inclusion in the study were (1) no or only trivial
shoulder trauma, (2) marked loss of active and passive shoulder
motion (greater than a 50 percent loss of external rotation) especially with
the shoulder abducted 90 degrees, (3) pain at the extremes of all
motions, (4) globally limited glenohumeral translation, and (5)
normal findings on true anteroposterior and axillary lateral radiographs
of the glenohumeral joint. Patients with adhesive capsulitis and
extrinsic disorders such as cervical spine pathology, cardiopulmonary
conditions, and systemic disorders (for example, diabetes mellitus
and thyroid dysfunction) were included in the study (Table III). We included
all patients with idiopathic atraumatic adhesive capsulitis so that
we could analyze the effect of extrinsic conditions on the outcome
of nonoperative treatment. Just as the pathophysiology of idiopathic
adhesive capsulitis is poorly understood, so are the effects of
extrinsic conditions. Few prior studies have addressed this issue,
and we thought that our study would afford us the opportunity to
do this.
The exclusion criteria included (1) intrinsic glenohumeral pathology
such as glenohumeral arthritis, (2) a history of substantial shoulder
trauma, (3) previous shoulder surgery, (4) initiation of successful
treatment within six weeks before the initial evaluation by the
senior author, and (5) reflex sympathetic dystrophy.
Patient Demographics
Fifty-eight (77 percent) of the patients were female, and seventeen
(23 percent) were male. The mean age was fifty-three years (range,
thirty-five to seventy-six years). There was no significant difference
between the ages of the male and female patients (p = 0.64). The
dominant shoulder was involved in thirty (40 percent) of the patients,
and the nondominant shoulder was involved in forty-five (60 percent).
The left shoulder was involved in forty-four patients (59 percent).
Two patients (3 percent) had concurrent bilateral involvement, and consequently
seventy-seven shoulders were included in the study. Four patients
(5 percent) had had contralateral idiopathic adhesive capsulitis previously.
Twenty-two patients (29 percent) had a history of trivial shoulder
trauma that they related to the onset of the symptoms. Five patients
(7 percent) had a Workers' Compensation claim or pending litigation.
The mean duration of shoulder pain prior to the initial evaluation
by the senior author was 9.2 months (range, 1.3 to forty-seven months).
The duration of shoulder stiffness could not be determined because
the findings of previous physical examinations were not available
and it was impossible to determine when stiffness began in the course
of the adhesive capsulitis.
Forty-six shoulders (60 percent) had had previous unsuccessful
treatment. Twenty-five shoulders (32 percent) had undergone only
a course of physical therapy, and seven (9 percent) had had only
a corticosteroid injection. Fourteen shoulders (18 percent) had
had both physical therapy and a corticosteroid injection. Although
we were unable to precisely determine the details of the prior physical
therapy, it usually included ultrasound, heat, massage, and active
exercises rather than passive stretching. Similarly, we were unable
to definitively determine whether the injections were intra-articular
or subacromial. One of the patients had had two manipulations under
anesthesia prior to our evaluation. Thirty-one shoulders (40 percent) had
not had previous treatment.
Eight patients (11 percent) had diabetes mellitus, and seven
(9 percent) had hypothyroidism. Six (8 percent) had coronary artery
disease, and three of them had undergone coronary artery bypass
grafting. One patient had myasthenia gravis. Forty-two (72 percent)
of the women were postmenopausal.
Fourteen patients (19 percent) had another condition or had had
surgery that involved the affected upper extremity or shoulder girdle.
One patient had had recent carpal tunnel surgery. Two other patients
had symptoms of carpal tunnel syndrome. One patient had had an ipsilateral
anterior subcutaneous ulnar nerve transposition immediately before
the onset of the shoulder stiffness. Four patients related a history
of previous "frozen shoulder" involving the same extremity. Five
patients had undergone ipsilateral breast surgery just prior to
the onset of symptoms of frozen shoulder. One had a history of poliomyelitis
involving the affected side.
Advanced shoulder imaging had been performed by another physician
for twelve patients prior to our initial evaluation. One patient
had a computerized tomographic arthrogram that revealed normal findings.
Eleven patients had magnetic resonance imaging of the shoulder;
six of these studies revealed normal findings, three suggested a
partial-thickness supraspinatus tear, one showed a small full-thickness
rotator cuff tear, and one suggested an anterior labral tear. The
patients with a partial-thickness tear did not have relief of shoulder
pain or stiffness after subacromial injection with ten centiliters
of 1 percent plain lidocaine. The patient with a full-thickness
rotator cuff tear did not require rotator cuff repair. The patient
with a labral tear had eventual failure of our treatment protocol and
underwent shoulder arthroscopy. Arthroscopic evaluation demonstrated
only mild anterior labral fraying. Thus, none of the pathological findings
on these five magnetic resonance imaging scans were clinically relevant.
It was for this reason that we included these patients, despite
the pathological findings identified by the magnetic resonance images.
All of these patients had markedly limited passive glenohumeral
motion that otherwise fit the inclusion criteria for this study.
Evaluation
The initial evaluation included the recording of a detailed medical
history with attention directed at the identification of relevant
comorbidities, the recording of a history of the shoulder disorder,
physical examination, completion of a subjective pain questionnaire,
and subjective functional evaluation with use of the Simple Shoulder
Test. Plain radiographs were made in the true anteroposterior and
axillary lateral projections. Follow-up evaluations were performed
at six-week intervals until either the patient clearly had enough
improvement for routine follow-up to be discontinued or the treatment
protocol had failed.
At the initial evaluation and at each follow-up visit, the patient
was asked to rate pain at rest and with activity according to a
scale in which no pain was assigned 1 point; mild pain, 2 points;
moderate pain, 3 points; marked pain, 4 points; and disabling pain,
5 points.
Active and passive shoulder motion was measured with a large
handheld goniometer. Bilateral active motions were assessed simultaneously
with the patient in the seated position. Active forward elevation
in the scapular plane was evaluated by measuring the angle formed
by the arm and thorax. Active external rotation was evaluated with
the arm adducted and the elbow at the side and flexed to 90 degrees.
Passive internal rotation of the arm behind the back was assessed
by determining the vertebral level that could be reached by the
tip of the thumb. All other passive shoulder motions were assessed
with the patient supine. This position limits the contribution of
scapulothoracic motion to overall shoulder motion. Passive forward elevation
was determined by measuring the angle formed by the arm and thorax.
Passive external rotation was measured with the arm at the side
and the elbow flexed 90 degrees. Passive shoulder rotation was measured
with the shoulder abducted to 90 degrees. The total passive arc
of rotation in this position was calculated. When the arm could
not be abducted to 90 degrees the arc was considered to be 0 degrees.
Shoulder strength was evaluated with manual muscle-testing. A
more detailed or precise strength evaluation was not carried out
because adhesive capsulitis is not generally considered to be a
cause of shoulder weakness. In addition, it is impossible to assess
rotator cuff strength accurately when glenohumeral motion is limited.
The Simple Shoulder Test was used to assess the patient's subjective
comfort and the ability to perform twelve activities at the initial
and final evaluations. Matsen et al. demonstrated the ease of use and
reproducibility of this test14.
They found that patients between the ages of sixty and seventy years
who had normal shoulders responded "yes" to all of the questions
nearly all of the time.
Treatment Protocol and Determination of Failure
All patients were treated with the same rehabilitation exercise
protocol. This included pendulum circumduction and passive shoulder-stretching
exercises in forward elevation, external rotation, horizontal adduction,
and internal rotation (FIg. 1-A, FIg. 1-B, FIg. 2-A, FIg. 2-B, FIg. 3-A, FIg. 3-B, FIg. 4-A, and FIg. 4-B). The patient was provided with
instructions for these exercises and was referred for formal physical
therapy. The patient was instructed to stretch the shoulder to the
point of tolerable discomfort five times each day. On the average,
the patients reported performing home exercises twice each day.
Sixty-eight patients (91 percent) participated in a formal physical
therapy program that emphasized these passive stretching exercises.
On the average, they attended physical therapy sessions twice each
week. We recommended formal physical therapy sessions to teach the
exercises. Unfortunately, we had little control over the actual number
of formal sessions, which tended to be determined by the individual
therapists.
We considered the treatment to have failed when the patient was
dissatisfied with the status of the shoulder. This determination
was based upon the patient's report and not on our impression of
the shoulder's status. Most patients reported minimal reduction
of shoulder pain at three months after the initial evaluation. The
final determination of failure of treatment was made at the time
of the data analysis.
We did not specifically prescribe the use of oral nonsteroidal
anti-inflammatory medications or discourage or prohibit patients
from taking them. Specific strengthening exercises were deferred
until the shoulder pain was reduced and the shoulder motion had
improved. More aggressive treatment (manipulation or capsular release)
was not recommended until the protocol had failed (the patient was
dissatisfied with the outcome) after it had been tried for three
months or longer.
Outcome Assessment
At the final evaluation, the subjective outcome was assessed
by querying the patients about whether they were satisfied with
the outcome of the treatment of their shoulder and whether they thought
that they were much better, better, the same, or worse compared
with their status before treatment. The Disabilities of the Arm,
Shoulder, and Hand (DASH) Questionnaire1 and
the Short Form-36 (SF-36) Health Survey31 were
used at the final outcome evaluation to compare the patients with
each other and with established population norms. The patients who
were lost to follow-up or had undergone a procedure to increase
shoulder motion (manipulation under anesthesia or arthroscopic capsular
release) were not evaluated with these questionnaires.
All of the patients had their final evaluation as part of the
outcome study. The duration of follow-up was dependent upon our
ability to coordinate the evaluations of the majority of patients
who had stopped returning for routine follow-up evaluations.
For the purpose of the outcome analysis, two groups of patients
were considered. Group I included all of the patients who were managed
with the treatment protocol except for those who had been managed
with manipulation under anesthesia or capsular release. This group
included patients who were not satisfied with the outcome of the treatment
protocol if they had not had additional treatment. Group II consisted
of the patients who were not satisfied with the outcome and went
on to have manipulation under anesthesia or capsular release. The
combination of Groups I and II included all of the patients.
The results were analyzed with two major goals in mind. The first
was to assess the outcomes in Group I (the patients who did not
have manipulation or capsular release). Pain, range of motion, and
functional ability before and after treatment were compared, and
the associations between demographic variables and the findings
on the initial evaluation, the subjective outcomes, and the DASH
scores were analyzed. Additionally, the DASH and SF-36 scores of
these patients (Group I) were compared with published population norms.
All of the patients in Group I who were eventually satisfied with
the outcome had been satisfied with their progress after three months
of the treatment protocol. We did not determine the duration to
the final result, as most of the patients did not continue returning
for routine follow-up once they began to have resolution of their
symptoms.
The second goal was to analyze the demographic variables and
the findings, at the initial evaluation, with respect to pain, range
of motion, and the results of the Simple Shoulder Test for the patients in
Group II in order to identify predictors of unsatisfactory outcome.
Statistical Analysis
The paired t test was used to analyze shoulder motion and responses
to the Simple Shoulder Test. The Wilcoxon signed-rank test was use
to analyze pain at the various evaluation intervals. Associations
between various initial and outcome parameters were analyzed with
use of the Pearson product-moment correlation, Spearman, Mann-Whitney
U, and Kruskal-Wallis tests. P values of less than 0.05 were considered
significant.
The mean duration of follow-up was twenty-two months (range,
twelve to forty-one months). One patient died prior to the final
evaluation, and three patients were lost to follow-up. Fifty-nine
(77 percent) of the shoulders (fifty-seven patients) had complete
follow-up, including physical examination and outcome assessment.
Fourteen (19 percent) of the patients who did not return for final physical
examination participated in an evaluation of pain and satisfaction
and completed the outcome evaluation questionnaire over the telephone. The
range of motion was not determined for the patients who were evaluated
by telephone.
Group I, the patients who did not undergo manipulation under
anesthesia or capsular release, included sixty-six patients. Group
II, the patients who had manipulation under anesthesia or capsular
release, included five patients.
Pain
The patients' evaluations of pain at rest and pain with activity
are summarized in Table I. In Group I, the mean score for
pain at rest decreased from 1.58 points at the initial evaluation
to 1.16 points at the final evaluation (p < 0.001), and the
mean score for pain with activity decreased from 4.12 to 1.33 points
(p < 0.0001).
The pain scores at the midterm evaluation (six to twelve weeks
after the initial evaluation) were available for forty-six patients
(Table I). These
data included the scores for the patients who eventually underwent
manipulation or arthroscopic capsular release. The mean score for
pain at rest was 1.2 points, and the mean score for pain with activity
was 2.1 points. The scores for pain at rest (p < 0.0001) and
with activity (p < 0.001) at the midterm evaluation were significantly
lower than those at the initial evaluation.
In Group I, the severity of the initial pain at rest and with
activity was not associated with the final subjective outcome. Similarly,
the severity of the initial pain at rest was not associated with
the final DASH score (p = 0.083). In contrast, the initial pain
with activity in Group I was associated with a greater (worse) final
DASH score (p = 0.0235).
Shoulder Motion
The active and passive ranges of motion of the affected shoulder
in Group I at the initial and final evaluations are summarized in Table II. There was
a significant increase in all ranges of motion at the final evaluation
(p < 0.00001). Nevertheless, the final range of motion of the
affected shoulder was significantly less than the motion of the
unaffected shoulder (p < 0.00001) (Table II). Despite this difference, the
ratio of the initial motion to the final motion was not significantly associated
with the final subjective assessment of pain, the subjective outcome,
or the final DASH score. Similarly, there was no significant association
between the improvement in range of motion and the outcome with
respect to pain, function (Simple Shoulder Test), or the DASH or
SF-36 score.
Diabetes mellitus and male gender were the only variables associated
with greater limitations of final motion (p < 0.01).
Functional Evaluation
In Group I, sixty-three patients responded to the Simple Shoulder
Test. The mean number of "yes" responses to the Simple Shoulder
Test increased from 4.1 at the initial evaluation to 10.75 at the outcome
evaluation (p < 0.00001). Even at the midterm evaluation, there
was significant improvement, to 7.0, in the mean number of "yes"
responses (p < 0.00001).
At the final outcome evaluation, all of the patients were comfortable
with the arm at the side and could reach the small of the back to
tuck in their shirt. Only two patients were unable to return to their
usual work, one patient was unable to sleep comfortably, and one
was unable to place a coin on a shelf at the level of the shoulder
without bending the elbow.
At the final evaluation, twenty-seven (43 percent) of the sixty-three
patients who responded had less than twelve "yes" responses to the
Simple Shoulder Test. These patients had had a mean of 3.15 "yes"
responses before treatment, whereas the patients who had all "yes"
responses at the final outcome evaluation had had a mean of 4.67
"yes" responses before treatment. There was a positive trend for
an association between the numbers of "yes" responses at the initial
and final evaluations (p > 0.05 and < 0.10, respectively).
Assessment of the negative responses to the Simple Shoulder Test
at the final outcome evaluation revealed that, of the sixty-three
patients, twenty (32 percent) were unable to throw a ball overhead a
distance of thirty yards (27.4 meters), eleven (17 percent) were
unable to lift eight pounds (3.6 kilograms) to the level of the
shoulder without bending the elbow, eleven were unable to carry
twenty pounds (9.1 kilograms) at the side, eight (13 percent) were
unable to toss a softball underhand ten yards (9.1 meters), and
eight were unable to wash the back of the contralateral shoulder.
In Group I, there was no significant association between the
responses to the final Simple Shoulder Test and the final range
of motion of the affected shoulder. Similarly, there was no association
between the improvement in the range of motion and the responses
to the Simple Shoulder Test.
Outcome Evaluation
Satisfaction with the outcomes was determined by questioning
the patients. Ninety percent (sixty-four) of the patients (sixty-six
shoulders) considered the outcome of the nonoperative treatment protocol
to be satisfactory. Forty-nine (67 percent) of the seventy-three
shoulders were considered to be much better, twenty-one (29 percent)
were thought to be better, and three (4 percent) were considered
to be unchanged. No shoulder was worse after the nonoperative treatment.
Two patients who were not satisfied elected not to have additional
treatment. Interestingly, contralateral idiopathic adhesive capsulitis
developed in three patients (4 percent) during the follow-up period.
In total, nine (13 percent) of the patients in this study had bilateral
idiopathic adhesive capsulitis.
Five (7 percent) of the original seventy-one patients who were
available for outcome evaluation were not satisfied with the result
of the treatment protocol and elected to undergo a procedure to
increase shoulder motion at a mean of six months (range, three to
thirteen months) after our initial evaluation. Two of these patients
had had improvement after nonoperative treatment but still elected
to have arthroscopic capsular release. Both had additional improvement
after this intervention. One patient who did not have improvement after
nonoperative treatment had a successful manipulation under anesthesia.
Two patients who had manipulation and arthroscopic capsular release
did not have improvement after this intervention.
Both of the patients for whom the nonoperative and operative
treatment failed were diabetic. One did not seek additional treatment.
The other went to another physician and had another manipulation under
anesthesia, which decreased the symptoms.
Demographic variables were specifically analyzed for any association
with subjective dissatisfaction and the eventual need for manipulation
or capsular release (Group-II patients). Previous treatment with
physical therapy (p = 0.028) and a Workers' Compensation claim or
litigation (p < 0.001) were the only variables associated with
failure or eventual manipulation or capsular release. Treatment failure
and the need for manipulation or capsular release were not associated
with age, menopause, duration of symptoms, the shoulder that was
affected, trivial trauma, medical comorbidities, initial pain, responses
to the Simple Shoulder Test, or range of motion.
The effect of several extrinsic factors on the outcome was assessed
(Table III).
We analyzed the effect of coronary artery disease, previous surgery,
diabetes mellitus, and hypothyroidism as well as the combined group
of extrinsic factors.
The mean DASH score at the final evaluation of the patients managed
conservatively (Group I) was 9.7 ± 13.6 points.
The established mean normal values for the DASH score are between
3 and 6 points (Dorcas Beaton, personal communication, 1997). Fifty-two
percent of the patients in Group I had a DASH score of less than
3 points, and 61 percent had a score of less than 6 points. Thus,
nearly 40 percent of the Group-I patients had an abnormally high
DASH score (more than 6 points). The mean DASH score for patients
with a rotator cuff tear (about 27 points [Dorcas Beaton, personal
communication, 1997]) was used as a relevant comparison. Twelve
percent of the patients in Group I who considered the shoulder to
be improved had a DASH score of greater than 27 points.
There was no significant association between the final score
for pain at rest and the DASH score (p = 0.082). However, there
was a significant association between the initial score for pain
with activity and the DASH score (p = 0.0235) and a highly significant
association between the final score for pain with activity and the
DASH score (p < 0.0001).
The eight profiles of the SF-36 Health Survey completed by the
patients in Group I were compared with values in the normal population
(both the general population and appropriate age and gender-matched
populations). The mean scores in Group I were never lower than the
published population norms. When compared with the general population,
the patients in Group I had a significantly higher mental health
score (p = 0.04). All of the other profiles were not significantly
greater than those for the general population. The scores for the
men in Group I were not significantly different from those in the
general male population. The women in Group I had significantly
better scores for vitality (p = 0.012) and mental health (p = 0.049),
as well as a trend toward greater scores on all of the other profiles,
compared with the general female population.
In this study, we prospectively evaluated the subjective and
objective outcomes of the treatment of idiopathic adhesive capsulitis
with a specific shoulder-stretching exercise program. The vast majority
of patients with this condition were successfully treated. The measurable
limitations and deficiencies at the outcome evaluation were acceptable
to most of the patients and did not appear to affect their general
health status. We concluded that most cases of idiopathic adhesive
capsulitis can be treated with shoulder-stretching exercises. In
this study, this was true even in many patients who had had prior
treatment attempts with physical therapy and corticosteroid injections.
Although previous, retrospective studies15,29 have
also demonstrated that patients who recover from adhesive capsulitis
have residual limitations, those studies failed to document the
extent of improvement and patient satisfaction that occurred. Our
prospective study provided data that enabled us to analyze the reduction
in pain and improvement in function as well as to document the limitations
and general health status at the time of final follow-up. Consequently,
the somewhat pessimistic view expressed by authors of retrospective studies
about the outcome of idiopathic adhesive capsulitis can be compared
with the marked improvements and overall satisfaction that we observed.
Despite the success of treatment with a specific stretching exercise
program, there were measurable differences between the affected
and unaffected shoulders. Even among the patients who were satisfied,
a substantial number were not pain-free. Ten percent had mild pain
at rest, and 27 percent had mild or moderate pain with activity.
We also found significant differences between the involved and uninvolved
shoulders with regard to all ranges of active and passive shoulder
motion at the final evaluation. Additionally, the responses to the Simple
Shoulder Test were not normal. However, there was no association
between the final range of motion and shoulder function as measured
by the Simple Shoulder Test. This latter finding is probably valid
because small differences in ranges of motion are not likely to
substantially affect functional capacity.
Furthermore, despite the high rate of satisfaction, the results
of the DASH Questionnaire demonstrated that nearly 40 percent of
the satisfied patients had abnormal shoulder function. This finding
indicates that completely normal function of the shoulder and upper
extremity is not a prerequisite for patient satisfaction. Most importantly,
a patient's perception of outcome is relative not only to the normal
shoulder but also to the change in pain and function as a result
of the treatment. The amount that the affected shoulder is improved compared
with the preoperative status probably affects the subjective perception
of outcome. This is particularly true if "normal" is not the expected outcome.
Review of the previous literature on adhesive capsulitis demonstrates
controversy about which of the many available treatments is best3,7,8,11-13,15,16,18,19,21-24,26,27,29,32.
This controversy is due, in part, to a failure of many authors to
precisely define and accurately identify idiopathic adhesive capsulitis
among other causes of shoulder pain and stiffness. Many studies
have emphasized some form of invasive therapy, including corticosteroid
injection, brisement, manipulation under anesthesia, and arthroscopic
and open releases. Additionally, many of the studies have addressed
the treatment of recalcitrant adhesive capsulitis. It is difficult
to determine when adhesive capsulitis is recalcitrant, especially
if the initial treatment modality is not efficacious. In our study,
we clearly defined our patient population and we prospectively evaluated the
patients.
Our results are consistent with the findings reported by Shaffer
et al. in a long-term retrospective study29.
Fifty percent of their sixty-two patients reported having either
mild pain or stiffness at the time of follow-up. Miller et al. also
reported the long-term results of an orthopaedist-directed home
rehabilitation program15. Although
they stated that all of their patients regained significant motion
and returned to activities of daily living without pain, they did
not compare the results with the status of the unaffected shoulder.
The results of the SF-36 Health Survey did not demonstrate significantly
lower scores for the satisfied patients compared with the general
population. This finding suggests that this survey is not sensitive
enough to detect some major abnormalities of the shoulder and upper
extremity. It also suggests that patients who overcome adhesive
capsulitis do not have an intrinsic emotional, psychological, or
personality disorder as has been suggested by some authors5,9.
We attempted to identify the variables that have prognostic importance
for the treatment of idiopathic adhesive capsulitis. However, the
high rate of satisfaction rendered a statistical analysis of variance
difficult. Nevertheless, we identified variables that seemed to
be associated with some of the outcome parameters. There was a trend
toward worse results among men and among patients with diabetes.
In addition, we found that patients with adhesive capsulitis were
generally comfortable at rest and that pain with activity was associated
with function. This is supported by our finding that the level of
pain with activity at the outcome evaluation was associated with
the functional outcome while the final range of motion and improvement
in the range of motion were not.
This study also demonstrated that the level of self-reported
pain with activity and function during the most severe phase of
adhesive capsulitis is associated with the outcome of treatment.
Even the patients who were satisfied had consistent perceptions
of self-reported pain with activity and function between the initial
and final evaluations - that is, the patients who had the worst
perceptions of the shoulder before the treatment tended to have
the worst outcomes. This is a crucial finding that has not been
previously identified in studies of adhesive capsulitis, to our
knowledge.
We made no attempt to compare the effectiveness of the various
available treatment options, and the prescribed protocol was followed
unless the patient requested an alternative treatment. The decision
to perform a manipulation under anesthesia or an arthroscopic capsular
release was delayed until the shoulder had failed to improve despite
at least three months of the treatment protocol.
On the basis of the findings of our study, we recommend that
patients with phase-II idiopathic adhesive capsulitis should be
treated with a four-direction shoulder-stretching exercise program that
includes passive forward elevation, passive external rotation, passive
internal rotation, and passive horizontal adduction. This treatment should
be continued for at least three months before more aggressive or
invasive management is considered. Our experience indicates that
a significant reduction in pain and improvement in function should
be expected by three months and the vast majority of shoulders should
have a satisfactory outcome. Consequently, we do not recommend a
more accelerated treatment protocol. Although our patients agreed
to this protocol, we recognize that the treating physician can have
substantial influence on the patient's expectations and treatment
preference. We caution against inappropriate early manipulation
or surgical intervention. When our treatment protocol failed, the success
of manipulation or arthroscopic capsular release was limited.
To our knowledge, this is the first prospective study of idiopathic
adhesive capsulitis to provide detailed information about the level
of pain, active and passive range of motion, and functional outcome
of nonoperative treatment. The results clearly demonstrate that
the vast majority of patients with phase-II idiopathic adhesive
capsulitis can be effectively treated with a specific shoulder-stretching
exercise program. When the final outcome was assessed, all of the
parameters evaluated were found to be significantly improved and
90 percent of the patients were satisfied.
This study defined outcomes in terms of patient self-assessment
rather than categorical ranking. This type of information is critically
important when determining the management of any musculoskeletal
condition. The results of this study provide a basis for treatment
of phase-II idiopathic adhesive capsulitis as well as an estimate
of the expected outcome and identify significant prognostic factors.
We believe that the efficacy of early use of more aggressive or
invasive treatments should be studied further before widespread
utilization is advocated.
American Academy of Orthopaedic
Surgeons: Disabilities of the Arm, Shoulder, and Hand. Rosemont,
Illinois, American Academy of Orthopaedic Surgeons, 1995.
Barrett, W. P.; Franklin, J. L.; Jackins,
S. E.; Wyss, C. R.; and Matsen, F. A., III: Total shoulder arthroplasty.
J. Bone and Joint Surg., 69-A: 865-872, July 1987.
Binder, A.; Hazleman, B. L.; Parr,
G.; and Roberts, S.: A controlled study of oral prednisolone in frozen
shoulder. British J. Rheumatol., 25: 288-292, 1986.
Codman, E. A.: The Shoulder. Rupture
of the Supraspinatus Tendon and Other Lesions in or about the Subacromial
Bursa, pp. 216-224. Boston, privately printed, 1934.
Coventry, M. B.: Problem of painful
shoulder. J. Am. Med. Assn., 151: 177-185, 1953.
Duplay, S.: De la p豩-arthrite scapulo-hum象le et
des raideurs de l'诡ule qui en sont la consequè£e. Arch. gen.
m袮, 20: 513-542, 1872.
Ekelund, A. L., and Rydell, N.: Combination treatment
for adhesive capsulitis of the shoulder. Clin. Orthop., 282: 105-109,
1992.
Fareed, D. O., and Gallivan, W. R.,
Jr.: Office management of frozen shoulder. Treatment with hydraulic
distension under local anesthesia. Clin. Orthop., 242: 177-183,
1989.
Fleming, A.; Dodman, S.; Beer, T.
C.; and Crown, S.: Personality in frozen shoulder. Ann. Rheumat.
Dis., 35: 456-457, 1975.
Grey, R. G.: Brief note. The natural
history of "idiopathic" frozen shoulder. J. Bone and Joint Surg.,
60-A: 564, June 1978.
Hamer, J., and Kirk, J. A.: Physiotherapy
and the frozen shoulder: a comparative trial of ice and ultrasonic
therapy. New Zealand Med. J., 83: 191-192, 1976.
Hill, J. J., Jr., and Bogumill, H.:
Manipulation in the treatment of frozen shoulder. Orthopedics, 11:
1255-1260, 1988.
Lee, M.; Hag, A. M.; Wright, V.;
and Longton, E. G.: Periarthritis of the shoulder: a controlled
trial of physiotherapy. Physiotherapy, 59: 312-315, 1973.
Matsen, F. A., III; Lippitt, S. B.;
Sidles, J. A.; and Harryman, D. T., II: Evaluating the shoulder. In
Practical Evaluation and Management of the Shoulder, pp. 1-17. Philadelphia,
W. B. Saunders, 1994.
Miller, M. D.; Wirth, M. A.; and
Rockwood, C. A., Jr.: Thawing the frozen shoulder: the "patient" patient.
Orthopedics, 19: 849-853, 1996.
Murnaghan, J. P.: Adhesive capsulitis
of the shoulder: current concepts and treatment. Orthopedics, 11:
153-158, 1988.
Neviaser, J. S.: Adhesive capsulitis
of the shoulder. A study of the pathological findings in periarthritis
of the shoulder. J. Bone and Joint Surg., 27: 211-222, April 1945.
Neviaser, R. J., and Neviaser, T.
J.: The frozen shoulder. Diagnosis and management. Clin. Orthop.,
223: 59-64, 1987.
Ogilvie-Harris, D. J.; Biggs, D.
J.; Fitsialos, D. P.; and MacKay, M.: The resistant frozen shoulder:
manipulation versus arthroscopic release. Clin. Orthop., 319: 238-248,
1995.
Ozaki, J.; Nakagawa, Y.; Sakurai,
G.; and Tamai, S.: Recalcitrant chronic adhesive capsulitis of the
shoulder. Role of contracture of the coracohumeral ligament and
rotator interval in pathogenesis and treatment. J. Bone and Joint
Surg., 71-A: 1511-1515, Dec. 1989.
Parker, R. D.; Froimson, A. I.;
Winsberg, D. D.; and Arsham, N. Z.: Frozen shoulder. Part II: Treatment
by manipulation under anesthesia. Orthopedics, 12: 989-990, 1989.
Polkinghorn, B. S.: Chiropractic
treatment of frozen shoulder (adhesive capsulitis) utilizing mechanical
force, manually assisted short lever adjusting procedures. J. Manip.
and Physiol. Therapeut., 18: 105-115, 1995.
Pollock, R. G.; Duralde, X. A.;
Flatow, E. L.; and Bigliani, L. U.: The use of arthroscopy in the treatment
of resistant frozen shoulder. Clin. Orthop., 304: 30-36, 1994.
Quigley, T. B.: Indications for manipulation and
corticosteroids in the treatment of stiff shoulders. Surg. Clin.
North America, 43: 1715-1720, 1963.
Reeves, B.: The natural history of
frozen shoulder syndrome. Scandinavian J. Rheumatol., 4: 193-196,
1975.
Rizk, T. E.; Pinals, R. S.; and Talaiver,
A. S.: Corticosteroid injections in adhesive capsulitis: an investigation
of their value and site. Arch. Phys. Med. and Rehab., 72: 20-22,
1991.
Rizk, T. E.; Gavant, M. L.; and Pinals,
R. S.: Treatment of adhesive capsulitis (frozen shoulder) with arthrographic
capsular distention and rupture. Arch. Phys. Med. and Rehab., 75:
803-807, 1994.
Roy, S., and Oldham, R.: Management
of painful shoulder. Lancet, 1: 1322-1324, 1976.
Shaffer, B.; Tibone, J. E.; and Kerlan,
R. K.: Frozen shoulder. A long-term follow-up. J. Bone and Joint
Surg., 74-A: 738-746, June 1992.
Simmonds, F. A.: Shoulder pain. With
particular reference to the "frozen" shoulder. J. Bone and Joint
Surg., 31-B(3): 426-432, 1949.
Ware, J. E., Jr.: SF-36 Health Survey:
Manual and Interpretation Guide. Boston, The Health Institute, New
England Medical Center, 1993.
Warner, J. J. P.; Allen, A.; Marks,
P. H.; and Wong, P.: Arthroscopic release of chronic, refractory
adhesive capsulitis of the shoulder. J. Bone and Joint Surg., 78-A:
1808-1816, Dec. 1996.
Zuckerman, J. D., and Cuomo, F.: Frozen shoulder.
In The Shoulder: A Balance of Mobility and Stability, pp. 253-267.
Edited by F. A. Matsen, III, F. H. Fu, and R. J. Hawkins. Rosemont,
Illinois, American Academy of Orthopaedic Surgeons, 1993.