Retrieval, experimental, and computational assessment of the performance of total knee replacements

Wear mechanisms in polyethylene components for total knee replacements are inherently mechanical; the local stresses or strains exceed some material limit. Retrieval analysis and knee simulators have provided the means to quantify the damage observed in vivo or in vitro. These results have been circumstantially linked to the material stresses obtained from computational simulations using finite element analysis, knee simulator tests, and computational simulations of two condylar knee designs. We hypothesize that if an equivalent loading environment is produced in the computational simulation, we can correlate the distribution of computed stresses with observed damage of simulator specimens and further relate design differences to in vivo performance from retrieval analyses. The finite element model agreed with the knee simulator kinematics and kinetics within 2-13%, and composite FEA contact areas matched 66-90% of the damage areas due to burnishing on the simulator specimens. Burnishing was the primary mode of damage for both the simulator and retrieval specimens corresponding with the relatively low magnitudes of contact stress observed. Both the computational and experimental techniques underpredicted the amount determined from retrieval analysis, but the differences between the two designs were consistent for all three methods. Combining these techniques strengthens the applicability of the computational simulation while highlighting the complementary approach of these methods for preclinical testing and assessing the link between material state and damage.