Purpose: We carried out a biomechanical study by finite element analysis to compare treatment with a plate and treatment with a nail in pseudoarthrosis of the humeral shaft.

Materials and methods: We used a cadaver humerus and the two fixation devices to generate the geometry with design software (CATIA® v4.2). We then modelled the shapes with finite element analysis software (MSC.Patran®) and created three experimental models: healthy humerus, humerus with shaft pseudoarthrosis stabilised with AO plate and humerus with shaft pseudoarthrosis stabilised with locking nail. Both implants were titanium. The three models were subjected to nine different load conditions and the results compared.

Results: The nail model is stiffer than the plate in compression (3002.80 vs 789.68 N/mm), traction (6576.73 vs 1559.90 N/mm) and torsion (4.67 vs 2.73 N/mm). However, the plate model is biomechanically superior to the nail under other load conditions (mediolateral flexion, anteroposterior flexion, anteroposterior shear and mediolateral shear).

Conclusions: Although we can understand and compare the stability of the plate model with the nail, joint clinical and biomechanical studies are needed to determine the minimum stiffness required so that it will not interfere with the process of union under different load conditions.

Aim: Investigate the influence of various types of allograft (from the tibia, femur, and fibula) through finite element analysis to evaluate the best clinical configuration.

Methods: A non-linear 3D finite element model of a lumbar spine L3–L5 was used as a physiologic model (Noailly, 2003). The model was modified with the insertion of a transpedicular instrumentation (Surgival SA, Spain) and the removal of the L4 body and two adjacent discs. CT scans of a femur, tibia and fibula from the same patient were performed. Fragments of each bone were reconstructed and inserted within the model. Four configurations of allografts were investigated: one femur fragment, one tibial fragment, three fragments of fibula, six fragments of fibula. Four types of loadings were applied: compression (1000N), flexion, extension, and rotation (15Nm). Strain and stresses were calculated in large displacement (MARC, MSC Software).

Results: Von Mises stresses within the internal fixator are well below the Yield stress and the fatigue limit and therefore no fracture of the fixator is foreseen. The use of a fixator to create fusion of the two vertebras makes the lumbar spine much stiffer. The geometry and configuration of the allografts have a large influence on the strain and stresses within the adjacent vertebrae with a reduction of strains and stresses. The use of fragments of fibula gives the most stable configuration. However, this is also the configuration that changes most the maximal principal strains within the vertebrae. Results obtained with the femur or the tibia are very similar between each other. However, due to its ellipsoidal geometry, the allograft in tibia gives more asymmetric deformations than the femur.

Conclusion: Allografts harvested from the femur seems to be more reliable and change least the strain and stress distributions within the lumbar spine compared to allografts from the tibia or fibula.

Aim: Investigate the influence of end-plate preparation in a model of corporectomy to evaluate the best biomechanical configuration.

Methods: A non-linear 3D finite element model of a lumbar spine L3–L5 was used as a physiologic model (Noailly, 2003). The model was modified with the insertion of a transpedicular instrumentation (Surgival SA, Spain) and the removal of the L4 vertebral body and two adjacent discs. A femur allograft was inserted anteriorly. Four configurations were investigated: with allograft supported on the entire end-plate, with allograft supported on the half of cartilage endplate thickness, with allograft supported on the subcondral cortical shell and, finally, with allograft supported on the trabecular bone. Four types of loadings were applied: compression (1000N), flexion, extension, and rotation (15Nm). Strain and stresses were calculated in large displacement (MARC, MSC Software).

Results: Results indicate that the preparation of the end-plates has a minor influence on the strain and stresses within the adjacent vertebrae when rigid transpedicular instrumentation was placed. The use of a fixator to create fusion of the two vertebras makes the lumbar spine much stiffer. The resection of the cartilage and support the allograft in the cortical shell changes most the maximal principal strains in the remaining end-plate, and creates a peak stress in the contact area. On the other hand, complete resection of cartilage and subcondral cortical end-plate is the configuration that changes least the maximal principal strains within the adjacent vertebrae.

Conclusion: Preservation of the cortical end-plate may not offer a significant biomechanical advantage in reconstructing the anterior column when rigid transpedicular instrumentation was used.

Aims: The study we present compares quantitatively the bone regeneration in experimental animals obtained with autologus and homologus grafts against a calcium phosphate cement. Methods: We performed cavitary defects o 6 mm of diameter in the metaphiseal region of the distal femur of 48 rabbits of albine race. They were divided in 4 groups, and received respectively autologous grafts, homologous freezed graft, calcium phosphate cement or the absence of any implant (control group). Results: The results are valued by radiological, histological and histomorphometrical studies (with digitalysed images). Histological study shows a correct integration of the calcium phosphate cement, without þbrous interphase, and a bone regeneration which is progressive and centripetal. Statistical analysis of the histomorphometrical data shows that bone regeneration obtained with the calcium phosphate cement its similar to the one obtained with the grafts. Conclusions: Calcium phosphate cement is a biocompatible material, biodegradable and conductor.