Effectiveness of corticosteroid injections versus physiotherapy for treatment of painful stiff shoulder in primary care: randomised trial

Editor,

The recent paper by van der Windt et al1 helps to highlight some of
the challenges which exist in clinical studies of shoulder disorders.
These include the use of appropriate terminology and diagnostic criteria,
the need to establish a well designed study with comparable study groups
and adequate duration of follow-up and drawing appropriate conclusions
from the results obtained.

The study involved cases with ‘painful stiff shoulder', a term rarely
used now, as it represents a set of symptoms rather than a diagnostic
entity (somewhat paradoxically, ‘frozen shoulder' remains widely
accepted). It appears that the authors aimed to study individuals with
adhesive capsulitis. If this were the case then the diagnostic criteria
used should have included the global limitation of passive and active
movement in the absence of degeneration of the glenohumeral joint2.
Indeed, some workers have suggested that, particularly in clinical
studies, the diagnosis should be confirmed by arthrography 3 . Some cases
of adhesive capsulitis, particularly those in the later stages, exhibit
restriction in the absence of pain; such cases were excluded in this
study. Hence only a subset of the possible wide spectrum of cases - those
who are more likely to respond to anti-inflammatory approaches such as
corticosteroids - were included. Failure to exclude subjects with
glenohumeral arthritis, in addition to subjects who are known to have more
severe forms of adhesive capsulitis (in particular, those with diabetes
mellitus) may also have significantly affected the outcome. Administration
of the injections by such a large number of practitioners is another
avoidable source of error.

The important differences between the study groups which were noted
by the authors make interpretation of the results difficult. The
conclusions, that the benefits of corticosteroids ‘are superior to
physiotherapy', should be qualified. The results implied a short term
benefit with corticosteroids, but little benefit over physiotherapy in the
long term, as has been shown by other workers 4. Perhaps a control (no
treatment) group would have helped to indicate whether either treatment
makes any difference in the long term; other studies have implied that
they do not 4. Finally, the relatively high incidence of side-effects
noted with corticosteroid injections may have been related to the high
doses administered over a short period of time, doses which many would not
use in clinical practice.

Cathy Speed Clinical Fellow
Rheumatology Department,
Addenbrooke's Hospital,
Cambridge. CB2 2QQ.

1. van der Windt DAWM, Koes BW, Deville W, Boeke AJP, de Jong BA,
Bouter LM. Effectiveness of corticosteroid injections versus physiotherapy
for treatment of painful stiff shoulder in primary care: randomised trial.
Br Med J 1998; 317: 1292-6.