Aim: To determine whether moderate bone loss in revision total knee arthroplasty can be corrected using an uncemented prosthesis combined with cancellous bone grafting.

Methods and Patients: 40 revision total knee replacements were undertaken by the senior author between May 1999 and June 2004. 27 one stage revisions for aseptic loosening and 13 two stage revisions for infection. All cases involved bone loss of grades F1/2 and or T1/2 according to the Anderson Orthopaedic Research Institute Classification (Engh 1998). Bone loss was treated with a mixture of morselized autograft, morselized allograft and bone reamings loosely packed into any contained or uncontained defects following the technique of Whiteside (1992). Uncemented prostheses with long contact bearing stems were then inserted. Patients were followed up prospectively with Oxford and HSS knee scores.

Results: All 40 cases were able to partially weight bear immediately postoperatively, indicating satisfactory early press fit. No cases of loosening or cases suspicious of loosening have been noted. Mean follow up of 37 months with no patients requiring re revision, no persistent stem pain and no infection in the one stage revisions. 2 cases of infection in the 2 stage group are discussed, neither have required implant removal. Intraoperative and postoperative complications are discussed as well as range of motion, pain and patient satisfaction.

In 39/40 cases bone stock has been restored. In 1 case there was significant bone resorption under the tibial base plate due to stress shielding.

Conclusions: This technique is successful in building up moderate bone loss in revision total knee arthroplasty, therefore avoiding the need for excessive bone resection, large metal augments, mass allografts or custom made prostheses.