Intra-articular steroid therapy for osteoarthritis of the knee has been reassessed by two placebo-controlled trials of 20 mg of triamcinalone hexacetonide in 48 joints. Steroid injections caused a significantly greater reduction in pain and tenderness than placebo, and were preferred by patients. However, the benefits were small and transient. Maximum pain reduction occurred one week after injection, and was accompanied by a fall in the thermographic index suggesting an anti-inflammatory mode of action. Synovial fluid was unaffected by injections, and there was no correlation between synovial fluid cell counts or the radiological grading, and the degree of pain reduction.