Purpose: The rate of failure knee arthrodesis can be high when performed after an infectious complication of a total knee arthroplasty. We evaluated, in 14 patients at a mean 19 months (3–44 months) follow-up, a technique for knee arthrodesis using a custom-made endomedullary implant.

Material and methods: Mean age of these eight women and six men was 68 years. There were 11 patients with an infected prosthesis, one with post-trauma arthritis, one with aseptic loosening of a hinge prosthesis, and one with pseudarthrodesis. For the 12 patients with ongoing infection, surgery was performed in two times with insertion of a spacer (with antibiotics) between the operations. Mean delay between the two operations was 18 weeks. The surgical procedure was associated with a two-drug antibiotic regimen given for a mean three months after arthrodesis. All arthrodeses were stabilised with a custom-made femora-tibial implant with reaming and fixation with two screws. A graft was always used, composed of the reaming products and powder bone substitute in seven cases, reaming products alone in three, and bone-bank heads in four.

Results: Morbitiy: There was one misalignment of the tibial insertion that was not revised. One skin cover problem was treated with a vastus medius flap. One nonunion evolved favourably after a new graft. There were two recurrent infections: chronic fistulae that were controlled by local care and adapted antibiotic therapy. Weight-bearing was started during the first postoperative week in 13 patients. Bone healing (assessed radiographically with resolution of pain) was achieved at a mean three months (2–6 months) in 13 cases. At last follow-up all patients had achieved a satisfactory level of independence.

Discussion: According to the literature, intramedullary devices are superior for consolidating knee arthrodeses, with a lower rate of complications. Use of custom-made endo-medullary implants facilitates the operation and assures better stabilisation of the arthrodesis, allowing rapid weight-bearing. The infection must be controlled before using these implants. The results in our patients are in agreement with the most recent series reported (Barry, Stephen, Kuoan).

Conclusion: In our hands, this type of implant provides an effective means of attaining bony fusion, including in patients who require arthrodesis for an infected prosthesis.

Purpose: We were confronted with four cases of major loosening with migration of the metal-backed implant and acetabular osteolysis at seven years follow-up. We thus conducted a retrospective clinical and radiological analysis of our 192 prosthesis of this type implanted in 1993.

Material and methods: The clinical results were analysed using the Postel Merle d’Aubigné score (PMA). Radiographic wear was measured on the AP views without loading according to the Charnley method. Osteolysis was determined on the AP view using the De Lee and Charnley criteria. Certain files also had a scintigraphy or a CT scan performed to search for bone lysis. Osteolysis and femoral loosening were studied on the AP and lateral views. In agreement with the literature, wear of 0.1 mm/year was considered normal for these cemented prostheses and 0.15 mm/year for metal-backed cups.

Results: For the femoral component, there were two cases of loosening with massive osteolysis with a sanded titanium inserted with cement. None of the patients, with a cemented or non-cemented prosthesis underwent revision for pain. At the acetabular level, there were no failures for primary instability. Wear was normal (mean 0.08) in 89 cases (46.5%) and excessive (mean 0.26) in 103 cases (53.5%). Significant factors for wear were: patient age, gender, and level of activity. Factors without a significant influence were: size of the acetabular cup, type of bead, patient overweight. Osteolysis was the consequence of abnormal wear since in 42% of the abnormal wear cases showed osteolysis compared with 20% when wear was less than 0.15. The same observations were made for the clinical impact as 15.5% of the worn cups were symptomatic (PMA < 4) and 18 of the 103 patients (17.5%) underwent or will undergo revision.

Conclusion: Metal-backed cups present excessive wear at mid term causing early failure by osteolysis and implant migration. There are two options: remove the polyethylene sing a new metal-on-metal combination or a ceramic-ceramic combination, or returning to the metal-polyethylene combination with a cemented cup.