We used a new postprocessing method with the results from a three-dimensional finite element analysis to describe the general load transfer patterns for a cementless hip arthroplasty in the early post-operative situation, and to determine the effects of porous coating [full, partial (2/3), and none] and calcar-collar support (ideal initial contact with separation allowed upon loading, no collar) on this early load transfer. No-tension interfaces were modeled over the entire bone-prosthesis interface, with an upper bound on the Coulomb-friction over coated surfaces, and zero friction over smooth surfaces to accentuate the frictional effects of the coating.

The results indicate that the anteroposterior, mediolateral, and axial forces acting on each cross section of the bone were substantially different from the corresponding homeostatic (no prosthesis) forces for the fully coated device with collar support. The frontal bending moments acting on the bone were substantially less than the homeostatic values all along the prosthesis, while the sagittal bending and torsional loads were relatively similar to the homeostatic values. By far, the largest change in these loading patterns occurred with the loss of collar support, where axial loads acting on the bone were so low that over half the bone was in net tension because appreciable transfer of the compressive head force did not occur until well below the lesser trochanter. Both axial and torsional loads were transferred more distally for devices with more coating, and torsional loading of the bone was also sensitive to the degree of collar support. The frontal bending moments acting over most of the bone were insensitive to the coating or collar support. The strain energy density in the endosteal bone was most sensitive to these design variables in the proximal region, and the largest values occurred without a collar and without coating.

These findings indicate that all load components acting on the proximal bone in the early postoperative situation (no bone-ingrowth or fibrous tissue at the interface) are altered by the frictional coefficient of the bone-prosthesis interface (i.e. the presence of porous coating or some other surface treatment) and the degree of collar support, while only the axial and torsional loads are altered in the distal bone. From a prosthesis design perspective this implies that surface treatments and collar support can be used to control the axial forces and the torsional moments acting all along the bone. By contrast, the distal frontal bending moment, which dominates stresses in the diaphysis, cannot be altered by these design variables.