Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Steroids in lateral epicondylitis.

Three Part Question

In [patients presenting with an acute lateral epicondylitis] does [a local injection of corticosteroid] reduce [pain and improve time to recovery]?

Clinical Scenario

A 45 year old man presents to A&E with a 2/7 history of painful left elbow. He admits to undertaking a lot of gardening in the previous 3 weeks. Clinical examination reveals tenderness over the lateral humeral epicondyle and pain on resisted extension of the wrist. A clinical diagnosis of lateral epicondylitis is made. The patient is anxious to return to work and has heard that an injection can cure him.

Search Strategy

Medline 1966-04/98 using the OVID interface.
({[exp tennis elbow OR epicondylitis ti,ab,sh] AND [exp steroids OR steroid$ ti,ab,sh OR exp adrenal cortex hormones OR corticosteroids ti,ab,sh]} LIMIT to [english AND human])

Search Outcome

37 papers identified of which 2 were metaanalyses. The paper by Assendelft et al (1) incorporates ll of the PRCT's from the Labelle paper (2). The paper by Labelle concluded that the available trials were of insufficient quality to conduct a formal metaanalysis; trial quality was taken into account in the paper by Assendelft. The Labelle paper was therefore discarded following critical appraisal.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Assendelft WJ
12 randomised controlled trialsMetaanalysis-Further analysis of the results, stratifying for methodological quality revealed similar results in high and low quality trials. In the 5 placebo controlled trials the pooled odds ratio was 0.20 (95% CI 0.11-0.36) in favour of therapy.Several RCT's analysed are methodologically weak. Methodological scores for reviewed papers varied from 29 to 63 out of 100. Corticosteroid was compared against a variety of different treatments including placebo. No attempt to identify unpublished (grey) literature.
Pain FunctionPooled analysis indicated short term effectiveness (2-6 weeks) of local corticosteroid injection against control: pooled odds ratio (OR) = 0.15 (95% CI 0.10-0-23).
Time to functional recoveryLonger term follow up (>6 weeks) showed no statistical difference between corticosteroid injection and other treatments including placebo (pooled odds ration 0.73 (95% CI0.37-1.44).


This is a well-written paper that looks at PRCT's relating to the treatment of lateral epicondylitis. The authors have drawn tentative conclusions from studies, which they have identified as being methodologically weak. However, they take this into account in their analysis and show that both high and low quality papers draw similar conclusions. There was no attempt to look at unpublished data in this paper and it is therefore subject to publication bias. The papers examined are quite heterogeneous. 5 studies compare steroids against placebo, 2 compare different regimens of corticosteroid and the rest compare against the local normal practice (eg physiotherapy/ultrasound). The authors are at pains to point out these potential flaws in their work and do not draw unreasonable conclusions from the papers analysed.

Clinical Bottom Line

At the present time corticosteroid injections in lateral epicondylitis appear to be of benefit in the short term (2-6 weeks). Any long term benefit is unclear at the present time.


  1. Assendelft WJ, Hay EM, Adshead R, Bouter LM. Corticosteroid injections for lateral epicondylitis; a systematic overview. Br J Gen Pract 1996;46(405):209-216