Manipulation under anaesthesia versus conservative treatment in phase 2 of a frozen shoulder: a randomized controlled trial

Adhesive capsulitis (frozen shoulder) is a common cause of shoulder pain and
affects approximately 2-4% of the general population (vanderWindt 1995). The
peak incidence is between the fifth and sixth decade, slightly more frequent in
women than in men, and the most important associated condition is diabetes
(White 2011, Wang 2013). The pathophysiology of adhesive capsulitis is still
poorly understood. Idiopathic adhesive capsulitis is characterised by a
spontaneous onset of pain and stiffness of the shoulder, especially a loss in
external rotation, without a prior traumatic event (Robinson 2012). Earlier
studies considered it to be a self-limiting condition with a duration of 1-3
years (Reeves 1975, Jayson 1981). However, persistent symptoms or restrictions
can occur. (Binder 1984, Shaffer 1992). Even if the natural course is mostly
self-limiting, patients experience a prolonged period with a considerable
amount of pain and disability in daily life.
The condition is traditionally divided in three phases (Reeves). Phase one, a
freezing phase with severe pain and increasing stiffness. Phase 2, a frozen
phase with established stiffness but reduced pain. And a third phase, gradual
improvement of motion occurs in the thawing phase. However, clear definitions
of these separate phases are lacking, and there is a large variation in the
duration of each phase. The most important factors that are used to describe
the each phase are: time since onset of complaints, intensity of the pain and
the amount of restriction in range of motion. But convenient clear cut off
values that define each phase are not available.
Traditionally, manipulation under anaesthesia is a well-established treatment
procedure for a frozen shoulder if conservative treatment fails (Dodenhoff
2000). However, MUA still is a controversial procedure because it might lead to
serious complications as a humeral fracture, glenohumeral dislocation, brachial
plexus traction injury or intra-articular damage to the cartilage or cuff (Loew
2005). Yet, the reported complication rate of 0.5% is rather low (Grant 2013).
In the current situation in our hospital, a different strategy is followed by
the individual orthopedic surgeons in the treatment of frozen shoulders in
phase 2. One of the orthopaedic surgeons has a much lower threshold to decide
for manipulation under anesthesia compared with the other orthopedic surgeons,
who prefer a more conservative treatment. Therefore, the received treatment
depends at least partially on the fact, by chance, at which doctor the patient
gets an appointment. This variability was confirmed in a survey among Dutch and
Belgian orthopaedic surgeons. In addition, we found that the most procedures of
manipulation under anesthesia were conducted by one orthopaedic surgeon in our
hospital. (The results of the survey will be published in 2016 in Acta
Orthopaedica Belgica). Secondly we reviewed our own results after manipulation
in a retrospective cohort study, which showed none of the above mentioned
complications and rather satisfied patients (article will be submitted soon).
Therefore, we think that the present situation in our hospital offers an ideal
setting for a randomized controlled trial to compare manipulation versus
conservative treatment. We deliberately chose to keep the design of the study
as close as possible to the current usual care.

There are almost no good quality randomized controlled trials in which
manipulation under anesthesia is compared to conservative treatment. Kivimäki
et al conducted a randomized trial in which manipulation under anesthesia and a
home exercise program was not superior to a home exercise program alone. This
is interesting, and does not advocate manipulation under anesthesia. However, a
home exercise program is not part of our usual care of patients with a frozen
shoulder. In addition, Kivimäki et al report a loss to follow up of 34% at six
months. In addition, we think timing of the internvention is a crucial
parameter in a condition that is mainly self limiting, because shortening of
the duration of complaints and disability is probably more interesting than the
final result. A lot of studies lack a good explanation of the timing of the
procedure. Theoretically, the biggest advantage of manipulation is obtained
when manipulation is done at an early time point, but after the initial phase
one which is characterized by pain and inflammation. However, the side effect
of early manipulation can be an increased number of overtreatment in patients
with a mild and relative short natural course of the disease. So, the timing of
manipulation seems to be a very important issue when the profit of manipulation
is debated. The time interval from onset of complaints to the point where
manipulation is initiated varies highly in several studies. To our knowledge,
there is no consensus in the literature about the optimal timing. It is known
that too early manipulation in phase one (the painful inflammatory phase) is
even counterproductive and can lead to a prolonged course of the disease. In a
retrospective study, Vastamäki et al considered between 6 and 9 months as the
most optimal timepoint for manipulation. However, only a general comparison
between group A (between 6 and 9 months) and group B (the others, including
before six months and after 9 months) was presented. We think this does is
study is not detailed enough to draw firm conclusions about the optimal timing.

We hypothesize that the course of the disease can be shortened with early
manipulation under anaesthesia in phase 2 of a frozen shoulder, but after phase
one of the disease, potentially leading to a quicker functional recovery and
gain in range of motion than following conservative treatment.

Primary Objective:
The primary objective of this study is to evaluate the difference in functional
outcome, measured by the SPADI, after treatment of a frozen shoulder with or
without manipulation under anaesthesia.

Primary outcome parameter:
Shoulder Pain and Disability Index (SPADI) at one month

Secondary Objective(s):
The secondary objective is to evaluate the difference in the secondary outcome
measures after treatment of the two different treatment groups (A and B).

Secondary outcome measures:
Oxford Shoulder Score (OSS)
Duration of symptoms / disabilities
Pain: Numeric Pain Rating Scale (NPRS).
Health related quality of life: EQ-5D
Passive Range of Motion (ROM)
Two PROMS anchor questions regarding the change in pain and in daily functioning
Working ability: WORQ-UP, single item work ability score and absenteeism last
month
Usage of analgesics
Number of repeated corticosteroid infiltrations

Single center, prospective, randomized controlled trial.

MUA: Manipulation under anaesthesia is performed by the orthopaedic surgeon at
the recovery room under locoregional anesthesia. If necessary or desired by the
patient, general anesthesia can be added. The scapula indirectly stabilized by
the supine position, a short lever arm and 90 degrees of elbow flexion is used
to prevent fractures and brachial plexus traction injuries. The glenohumeral
joint is forced through a full range of motion in a strict pattern: anteflexion
-> abduction, external rotation in 90 degrees abduction-> internal rotation in
90 degrees abduction -> horizontal adduction with dorsal compression and
external rotation in neutral. A recognizable tearing sound is always present
when dealing with a frozen shoulder. This sequence can be repeated until full
range of motion is acquired. Postoperative physiotherapy is started directly
the same day to maintain the full range of motion that is obtained. People stay
at the ward of the orthopaedic department for one night. The first two weeks
the patient has to visit a physiotherapist as much as possible, the aim is
daily. Paracetamol and meloxicam (if not contra indicated) are prescribed for
ten days.

Conservative treatment: The conservative treatment group is given explanation
and advise about the natural course of the disease. A corticosteroid injection
in the glenohumeral joint of kenacort 40mg (1ml) and chirocaine (4ml) is mostly
given in the first three months of the complaints, before inclusion in the
study is possible. When the pain is not sufficiently diminishing, this can be
repeated. Physiotherapy is discussed on an individual base and is not part of
the study protocol in the conservative treatment arm. When patients are
referred to a physiotherapist, a guideline for the goals and methods is
offered.

It is important to note that the accessory risks of this study are not
increased compared to the current usual treatment methods in the Amphia
ziekenhuis. Both treatment modalities (MUA and conservative treatment) are
already used for a prolonged period by different orthopaedic surgeons. However,
this randomized study is designed to reliably compare the outcome in both
groups.

The burden associated with participation is an increased number of subjective
scores to fill out, and increased number of (non-invasive) measurements of the
function of the shoulder in addition to a further normal treatment and follow
up for both methods.

There is a risk of overtreatment in the manipulation group. It is difficult to
predict which patients will have a mild self-limiting natural course of the
condition. However, even in these patients, manipulation is meant to shorten
the duration of symptoms. Manipulation under anaesthesia is associated with a
minimal risk of fracture, glenohumeral dislocation, brachial plexus traction
injury or intra-articular damage to the cartilage or cuff. The reported
clinical significant complication rate is 0.5% (Grant 2013). These serious
complications are mainly because of technical errors of the procedure. We
recently reviewed a retrospective cohort of patients after manipulation, and
none of these complications had occurred in our center.
The effect of a possible corticosteroid is mainly local but also has a small
systemic effect which could potential lead to dysregulation of diabetes
mellitus. We consider that there are no risks of the conservative treatment
program. The benefit of the study is to provide an answer to the question if
manipulation under anaesthesia can indeed shorten the duration of symptoms and
the advantages outweigh the possible disadvantages.