Objective. - To retrospectively evaluate the benefits of knee joint lavage with intraarticular glucocorticoid injection in patients who have juvenile idiopathic arthritis with knee involvement unresponsive to repeated intraarticular glucocorticoid injections.

Patients. - Seventeen knees in 10 children (eight girls and two boys) were treated from 1997 to 2000. Mean age was 14 years 9 months and mean disease duration was 7.2 years. The diagnoses were juvenile oligoarthritis (n = 6, including two with extended disease), systemic arthritis (n = 2), juvenile spondyloarthropathy (n = 1), and juvenile dermatomyositis (n = 1). Repeated intraarticular triamcinolone hexacetonide injections had been performed in all the patients, the mean number of injections being 2.2 per patient within the last 30 months. Plain radiographs were normal in six of the eight patients. Mean erythrocyte sedimentation rate was 21.7 mm/h and mean C-reactive protein level was 20.6 mg/l. Joint fluid was obtained from 10 knees and had a mean cell count of 12 660 mm^-3. Second-line therapy was with methotrexate alone or combined with cyclosporine or azathioprine. Oral glucocorticoids and/or nonsteroidal antiinflammatory drugs were used for symptom relief.

Treatment procedure. - Lavage was performed under strict aseptic conditions with simple analgesia, on a day-hospital basis. After aspiration of the joint, lavage was performed with saline, and a delayed-action glucocorticoid was injected. The knee joint was immobilized in the extended position for 48 h. Efficacy criteria were presence of effusion, presence of pain, and presence of a systemic treatment-sparing effect.

Results. - Freedom from effusion and pain was noted in all 17 knees after 1 month, in eight (47%) knees after 6 months, and in seven (41%) knees after 12 months. The patients with the longest lasting improvements had systemic polyarthritis. After joint lavage, second-line treatment was reduced in two patients and oral glucocorticoid therapy was stopped in two others. None of the variables studied (age, sex, disease duration, inflammatory syndrome, or joint fluid cytology) predicted a good response. No adverse effects were recorded.

Conclusion. - These preliminary results show that joint lavage with glucocorticoid injection is safe in children. The improvements were modest, but the patients had a history of arthritis refractory to multiple triamcinolone hexacetonide injections. Thus, joint lavage may have a place in the treatment pyramid just before synovectomy.