Corticosteroid injections are the mainstay of treating tennis elbow even though their effectiveness has not been well established by controlled studies. A survey of consultant rheumatologists confirmed a widespread preference for this treatment but they varied in their choice of steroid dose and preparation. We examined the value of some practices by comparing local injections of 2 ml 1% lignocaine with either 10 mg triamcinolone or 25 mg hydrocortisone made up to 2 ml with 1% lignocaine (Study 1). The investigation was conducted double blind. Within the first 8 weeks, pain relief was greater for triamcinolone than hydrocortisone although the differences were not statistically significant. The response to both steroid preparations was significantly better than for lignocaine up to this point but at 24 weeks, the degrees of improvement were similar for all three groups and many patients still had pain. Relapse was common. In a separate but similarly designed study, triamcinolone 10 mg was compared with 20 mg of the same agent. Improvements of pain were similar and followed the same time scale. Post-injection worsening of pain occurred in approximately half of all steroid treated patients in both studies and this was sometimes severe and persistent. It was less frequent amongst those given lignocaine alone. Skin atrophy was reported in all groups but was more frequent amongst those given triamcinolone in Study 1. In conclusion, more rapid relief of symptoms was achieved with 10 mg triamcinolone than with 25 mg hydrocortisone or lignocaine alone and there was less needed to repeat injections. Results obtained with 20 mg triamcinolone were similar to those of the smaller dose.(ABSTRACT TRUNCATED AT 250 WORDS)