Expandable prostheses were designed to allow progressive growth after tumour resection in children. The aim of this study was to report the late results of the non-invasive growing prostheses designed by A Soubeyran (Phenix prosthesis or Wright Repiphysis).

From 1994 to January 2006, 27 children aged 4 to 12 (mean 8.5), underwent a resection of the knee for a bone tumour, with reconstruction by a non invasive expandable prosthesis. There were 16 boys and 11 girls. The tumours were 25 osteosarcomas and 2 Ewing tumours. All patients received pre and post-operative chemotherapy. There were 18 distal femur, 7 proximal tibia, and 2 femur + tibia resections.

There were different successive designs based on the same electro-magnetic growing mechanism using a pre bent spring, released by eating in an induction coil.

After, 7.2 years mean follow-up (4 months to 15 years), 20 patients had no evidence of disease and 7 were deceased. Two with a local recurrence were amputated. Mean lengthening was 5.1 centimeters (0 to 8), after 3 to 11 lengthening procedures. Mean limb-length discrepancy was 1.8 cm. Two patients had a secondary infection. Eleven had a revision for arthrofibrosis. All surviving patients were revised to a conventional hinged prosthesis. The mean MSTS functional score of the definitive prosthesis was 82% (63 to 96%).

Theses prostheses showed many mechanical complications as loosening, fracture of the growing mechanism, and arthrofibrosis. The positive outcome was the possibility to perform a progressive lengthening, without surgery limiting the risk of infection. Theses prostheses should be considered as temporary until reconstruction with a conventional hinged prosthesis. Patients with multiple revisions had a tendency to show less favourable late functional results than with primary implanted hinged prosthesis.

Failure of massive knee endoprostheses implanted for malignant tumours of the distal femur in children presents a difficult problem.

We present the results of rotationplasty undertaken under these circumstances in four boys. They had been treated initially at a mean age of 9.5 years for a stage-IIB malignant tumour of the distal femur by resection and implantation of a massive knee endoprosthesis. After a mean period of eight years and a mean of four operative procedures, there was failure of the endoprosthesis because of aseptic loosening in two and infection in two. Function was poor with a mean Musculoskeletal Tumor Society score of 7.5/30, and considerable associated psychological problems.

At a mean follow-up of 4.5 years after rotationplasty there was excellent function with a mean score of 27.5/30 and resolution of the psychological problems.