Page 207 - Reconstructive Techniques in Musculoskeletal Tumor Surgery

proximal tibia (OR 2.2). Infections occurred more often in the proximal tibia (OR 2.2) and less often in the distal radius (OR 0.1). In our retrospective study, we included 38 patients. With allograft-related failure as the end-point, graft survival rates at two, ve and ten years were 91%, 54% and 42%. We concluded that osteoarticular allograft reconstruction of major joints is associated with an unacceptably high rate of complications, and cautioned others against the routine employment of segmental osteoarticular allografts for reconstructions around the knee or shoulder.
In the studies on allograft reconstructions, nonunion has been identi ed as one of the major complications and failure mechanisms. In chapter 8, we present an evaluation on risk factors for nonunion of allograft-host junctions in intercalary allograft reconstructions of the femur and tibia, and assessed the in uence of cortical contact. To that end, we assessed the degree of contact in 96 transverse osteotomies that were xed using plates, on two orthogonal views. We introduced a novel classi cation system, in which we classi ed the degree of cortical contact into grades 1 (full contact over the entire length of the osteotomy), 2A (≥50% contact), 2B (<50% contact), and 3 (lack of cortical contact). We found that nonunion did not occur in grade 1 osteotomies. The risk of nonunion was 7% for grade 2A, 18% for grade 2B, and 50% for grade 3 junctions. Reconstruction site, patient age >16 years, localization within the bone or chemotherapy use did not signi cantly in uence nonunion risk. We concluded that, although future, larger studies will have to con rm our ndings, care should be taken to obtain rigid xation with rm contact at the junction site to minimize or even eliminate the risk of allograft-host nonunion.
Modular endoprostheses have largely replaced allografts as the method of choice for reconstruction of joints after tumor resection. In chapter 9, we reported the long-term results of knee replacement with MUTARS modular endoprostheses from two centers. A total of 110 consecutive reconstructions in 101 patients were evaluated with a minimum follow-up of ve years. Eighty-nine reconstructions (81%) were distal femoral replacements (78 uncemented, 87%; 42 of which hydroxyapatite-coated, 54%), 21 (19%) were reconstructions of the proximal tibia. In 26 reconstructions (24%), surgery was preceded by a failed previous reconstruction. Complications of soft-tissue or instability occurred in 6%. Loosening occurred in 28% of the implants used for revision of a failed reconstruction and
10
General summary
205