Introduction: The management of a hip destroyed by infection during growth is difficult and controversial. Two clinical cases illustrate differing management options and question what might be the best solution for this problem.

Case 1: A 12-year-old girl presented with a Staphylococcus aureus septic arthritis of the left hip. She underwent an arthrodesis of the hip one-year later because of a poor position of the hip. She had 20mm of shortening of the left lower limb. After 6 years of “normal” life she received a total hip replacement when aged 19 years. Since then she has no pain and has 12 mm of shortening of her left lower limb and a mobile left hip (100/0, 40/30, 30/30) and works as a cashier.

Case 2 A 2-year-old boy presented with a Salmonella septicaemia and septic arthritis of the right hip. The hip deteriorated but the parents refused a hip arthrodesis. He underwent a Colonna procedure in 1994. Five years later he had developed 50mm of shortening in the right lower limb and he had an epiphysiodesis of the left knee. He is now 17 years old, has a painless hip, satisfactory range of movement (90/0, 0/15, 5/5), a slight limp, equal leg length and enjoys normal teenager activities. Radiographs suggest that he would benefit from a hip replacement.

Immobilization of septic arthritis is an ancient and always recommended notion. Before discovery of antibiotics, immobilization allowed an articular ankylosis in functional position. Since discovery of antibiotic chemotherapy, immobilization is justified for its antalgic and anti-inflammatory qualities. However, Salter demonstrated experimentally the interest of continuous passive mobilization during septic arthritis. The authors also demonstrated the deleterious effects of immobilization on articular cartilage during Staphylococcus aureus induced arthritis in a rabbit model. The authors compared two series of children treated for septic arthritis. All children were treated by articular lavage, and by intravenous antibiotic chemotherapy during 10 days, then by enteral antibiotic chemotherapy for 6 supplementary weeks. 14 children were immobilized during 1 month, while 14 others were mobilized from the first days. Consumption and class of antalgic chemotherapy, inflammatory balances (Blood Count, C Reactiv Protein), articular range motion during of the first and sixth month clinical review, were compared. Only articular range motion noted during the clinical review of the first month were significantly different in the two groups. Articular range motion of the not immobilized children were close of normal, while the other children suffered from articular stiffness. Immobilization had no beneficial effects either in pain or in correction of the inflam-matory process.