39 patients with active knee osteoarthrosis, chosen according to ACR criteria, were assigned to receive flurbiprofen (n 12, 2 x 100 mg), tiaprofenic acid (n 14, 2 x 300 mg) and placebo (n 13) in a 3-week, placebo-controlled study. All patients completed the study, and both medications were found to be effective: improvement occurred in the clinical signs. These drugs reduced the TNF-alpha levels. Flurbiprofen especially affected the IL-6 levels. Our findings indicate that NSAIDs may be effective in the etiopathogenesis of osteoarthrosis.