Skip navigation
Acute Pain | Chronic Pain | General

Non-steroidal anti-inflammatory drugs for painful shoulder

Clinical bottom line:

based on a small number of small trials of moderate to poor quality, NSAIDs appear to offer significant benefit compared with placebo. It is unclear whether symptom relief is as effective as with corticosteroid injection, and whether any of these treatments offer relief of clinical relevance. It was not possible to establish whether specific patients benefit more than others from NSAID treatment. There does not appear to be any difference between the different NSAIDs studied.

Conservative treatments for shoulder disorders include non-steroidal anti-inflammatory drugs (NSAIDs), corticosteroid injections and physiotherapy. Shoulder disorders include a number of different conditions including periarthritis, adhesive capsulitis, frozen shoulder, tendinitis and tendonitis. Approximately half of patients consult a physician, although shoulder complaints are usually self-limiting. However, for some patients pain will persist for several years.

Systematic review

Van Der Windt DAWM, Van Der Heijden GJMG, Scholten RJPM, Koes BW, Bouter LM. The efficacy of non-steroidal anti-inflammatory drugs (NSAIDS) for shoulder complaints. A systematic review. J Clin Epidemiol. 1995; 48:691-704.

Date review completed: 1993

Number of trials included: 19

Control group: active or placebo

Main outcomes: success rates (predominantly recovery or substantial improvement from baseline based on patient ratings).

Inclusion criteria were randomised, controlled trials of NSAIDs for shoulder complaints; full journal publication in English, German, French or Dutch.

Reviewers provided a descriptive summary of trials together with a quality evaluation and extracted information on success. Information on success rates was extracted on an intention-to-treat basis using data from patient ratings. Group differences in success rates were calculated with 95% confidence intervals.


Included trials were of moderate to poor quality. Drugs and dosing regimes, outcomes and diagnosis varied between trials. Patient diagnosis was not reported by reviewers, and blinding status was only reported for double-blind trials making it difficult to interpret information from remaining trials. Timing of outcome measures was not always reported, but it appears that no longer term assessments were available. Trials were generally small, and reviewers note that only three trials were of sufficient power to detect a clinically meaningful difference (defined as at least 25% difference in success rates between groups).

NSAID versus placebo

One single- or non-blind trial of 69 patients with acute tendinitis or bursitis of the shoulder compared daily doses of flurbiprofen (decreasing from 300 mg to 200 mg over a maximum of 14 days) with placebo. Flurbiprofen was significantly better than placebo (time not stated) for pain, tenderness, swelling and range of movement.

NSAID versus local anaesthetic and/or corticosteroid injection

Double-blind trials. Two of three trials showed corticosteroid injection to be better than NSAID. Two of two showed NSAID plus lidocaine injection to be better than lidocaine alone. One trial of 100 patients compared naproxen 1000 mg daily for four weeks with and without a single intrabursal injection of triamcinolone with placebo with and without injection. Steroid injection was significantly better than naproxen, and naproxen was significantly better than placebo (all patients received intrabursal lidocaine injection). A second trial of 60 patients compared diclofenac 150 mg daily for 4 weeks with subacromial corticosteroid injection or placebo. Similarly, corticosteroid was more effective than NSAID which was more effective than placebo (all patients received subacromial lidocaine injection).The third trial of 40 patients showed no benefit of indomethacin 100 mg daily for 25 days plus a single intrabursal injection of lidocaine compared with an injection of lidocaine and triamcinolone.

Single- or non-blind trials. One small trial showed no benefit of tolmetin sodium 1200 mg daily for four weeks plus one intra-articular injection of prednisolone and lidocaine compared with injection given alone.

Comparisons of different NSAIDs

Fourteen trials compared different NSAIDs. In general, trials revealed few differences between NSAIDs . Three low quality trials reported some differences, but in all cases these findings were questionable.

Adverse effects

Reviewers extracted this information where possible, and report that rates varied between 8% and 76% for NSAIDs. These were mainly gastrointestinal, skin rash, headache or dizziness, and were usually considered to be moderate to mild. In general, less than 10% of participants withdrew due to adverse effects, but three trials reported rates higher than this (12% to 20%)

Further reading

Related topics