Finite element models of total shoulder replacement: application of boundary conditions

The widespread use of FEA within orthopaedics is often prohibited by the limits of available computational power, with simplifications to the model often necessary in order to permit solution. An example of this includes the use of osseous models that exclude muscular loading, and may consist of only a partial or truncated region of the anatomy. However, is it possible to make such simplifications without affecting the predictive quality of the model? This issue has been considered using the specific example of the total shoulder reconstruction, where the effects of including the entire osseous region and/or the muscle loadings, has been evaluated. The effect of including the muscle loadings and the entire osseous structure was seen to increase with distance from the articular surface of the glenoid prosthesis. Stresses in the cement mantle were reduced in the absence of either the entire scapula bone or the muscle loading. The study suggests that the use of a fully defined scapula (hard- and soft-tissue) is particularly important when investigating fixation, whilst less comprehensive models should be appropriate for studies of the prosthesis exclusively.