Corticosteroid therapy versus physiotherapy on pain, mobility and function in shoulder impingement

Key points

1. 01A corticosteroid injection showed a short-term advantage for shoulder function at 6 to 7 weeks, but not for pain or range of motion.
2. 02At 6 months and 12 months there was no meaningful difference between injection and physiotherapy for any outcome.
3. 03Both treatments improved pain, range of motion and function, so patient preference should guide the choice.
4. 04Physiotherapy is favoured as the lower-risk option given limited long-term safety evidence for repeated injections.
5. 05The conclusion rests on only three trials with substantial differences between them, so confidence is moderate, not high.

How it was conducted

Design

Evidence statement based on a published systematic review and meta-analysis of 3 randomised controlled trials (Burger et al. 2016)

Participants

452 adults aged 18 to 65 with moderate or severe unilateral shoulder pain from subacromial impingement syndrome

Groups

Subacromial corticosteroid injection versus physiotherapy combining passive and active joint and soft tissue mobilisation, exercise and electrotherapy

Outcome measures

Pain (Visual Analogue Scale, Numeric Pain Rating Scale), glenohumeral range of motion, and shoulder function (global rating of change, shoulder disability questionnaire)

Follow-up

Short term 1 to 3 months, medium term 6 months, long term 12 months

What they found

• Shoulder function improved in favour of corticosteroid injection at 6 to 7 week follow-up (p < 0.0001).
• No evidence of superiority of injection over physiotherapy for pain or range of motion over the short term of 4 to 12 weeks.
• At 24 and 48 weeks, no evidence of superiority of injection over physiotherapy for shoulder function, pain or range of motion.
• Heterogeneity was high, with an I2 statistic of 84%, so pooling for pain and range of motion was judged inappropriate and results were summarised narratively.
• Of 14 full-text studies reviewed, only 3 met the inclusion criteria, with sample sizes of 136, 207 and 109 for a total of 452 participants.
• The 3 trials were NHMRC evidence Level II and scored an average of 7.3 out of 10 on the PEDro scale.
• Overall grade was a weak recommendation with moderate quality of evidence based on three RCTs (2B).

Limitations

• Only three trials met inclusion criteria, which the authors acknowledge as a limitation.
• Substantial statistical heterogeneity (I2 of 84%) prevented pooling for pain and range of motion, so those findings are narrative only.
• The composition and site of the injection and the physiotherapy protocols differed across the studies, introducing indirectness.
• Long-term safety and effectiveness of corticosteroid injection in this condition remain poorly evidenced.

Why it matters

For patients

You can reasonably choose either an injection or physiotherapy based on your preference, since long-term results are similar and physiotherapy carries less risk.

For clinicians

Use this to discuss options with patients, weighing the short-term functional gain from injection against the lower-risk profile of physiotherapy over the longer term.

For readers

A short, honest synthesis showing that for shoulder impingement, injection and physiotherapy converge over time despite an early functional edge for injection.