Massive endoprostheses using a cemented intramedullary stem are widely used to allow early resumption of activity after surgery for tumours. The survival of the prosthesis varies with the anatomical site, the type of prosthesis and the mode of fixation. Revision surgery is required in many cases because of aseptic loosening. Insertion of a second cemented endoprosthesis may be difficult because of the poor quality of the remaining bone, and loosening recurs quickly.

We describe a series of 14 patients with triplate fixation in difficult revision or joint-sparing tumour surgery with a minimum follow-up of four years. The triplate design incorporated well within a remodelled cortex to achieve osseomechanical integration with all patients regaining their original level of function within five months.

Our preliminary results suggest that this technique may provide an easy, biomechanically friendly alternative to insertion of a further device with an intramedullary stem, which has a shorter lifespan in revision or joint-sparing tumour surgery. A short segment of bone remaining after resection of a tumour will not accept an intramedullary stem, but may be soundly fixed using this method.

Massive cemented endoprosthesis are used to enable early resumption of activity after tumour surgery. The longevity of the prosthesis varies with anatomical site, pros-thesis type, and mode of þxation. Revision surgery will be required in approximately 50% of cases of endopros-thetic replacements around the knee by 10 years because of aseptic loosening. Insertion of a second cemented endoprosthesis is a challenge because of the poor quality of the remaining bone and loosening recurs quickly. The use of extracortical plate þxation in joint sparing surgery where the remaining bone after tumour resection will not accept an intramedullary stem is also described.

The þrst series of 14 patients with extracortical plate þxation in difþcult revision or joint-sparing tumour surgery with a mean follow-up of 5 years are described. The three-plate design incorporates well within a remodelled cortex to achieve osseomechanical integration with all patients regaining their premorbid level of function within 5 months. At 5 years the Enneking scores averaged 27.3.

One revision was required in a femoral replacement because of loosening. It was possible to insert a new endo-prosthesis as the intramedullary bone had reconstituted.

The preliminary results suggest that this technique may provide an easy, biomechanically friendly alternative to a device with an intramedullary stem, which has a shorter lifespan in revision tumour surgery.