Computer simulation of subchondral bone adaptation to mechanical loading in an incongruous joint

We hypothesized that the typically bicentric distribution of subchondral bone density (i.e., two maxima) in incongruous joints with deeper sockets could be predicted by a computer simulation employing a concavely incongruous finite-element model and current bone remodeling theory. Additional objectives were to assess the uniqueness of the solution with respect to assumed model parameters and initial conditions and to determine the relationship between contact areas, subchondral bone stress, and subchondral bone density patterns.

An idealized model of the humeroulnar joint was constructed with a quantitatively realistic representation of its natural incongruity. A currently accepted remodeling theory was implemented with a finite-element code using a node-based approach.

The simulation predicted a dense subchondral bone plate after application of 3,000 daily load cycles for 300 days. The pronounced bicentric distribution of subchondral mineralization emerged. The solution was virtually independent of the initial density distribution and other assumed model parameters. Furthermore, the model predicted high tensile stresses in the subchondral bone, when the joint socket was spread apart during loading. Therefore, the locations of maximal strain energy density in the subchondral bone did not correspond with areas of joint contact.

The results of the bone remodeling simulations are consistent with patterns of subchondral bone density determined experimentally. Furthermore, the solutions exhibited a high degree of uniqueness, were rather insensitive to changes in cartilage stiffness, moderately sensitive to number of applied loading cycles, and highly sensitive to loading magnitude. Tensile stresses seem to play a dominant role in subchondral bone remodeling due to bending in the subchondral bone plate. Thus, we conclude that, in the case of an incongruous joints with deeper sockets, the density of the subchondral bone cannot be regarded as a direct measure of the adjacent articular pressure.