Producing and avoiding stress shielding. Laboratory and clinical observations of noncemented total hip arthroplasty.

Experimental canine model studies of stiff versus flexible, fully porous-coated, metallic femoral stems (differing by three- to fivefold in stiffness characteristics) revealed markedly different resorptive bone remodeling patterns. The flexible stem resulted in about 30% more cortical bone retention adjacent to the implant at one-year postimplantation and larger differences in dogs killed two and three years after surgery. Strain-gauge studies confirmed that there are differences in cortical bone strains with the two stem designs, the flexible stem producing a more uniform and more nearly normal strain distribution medially. Differences in cortical bone remodeling were quantified using dual energy X-ray absorptiometry (DEXA). The bone mineral content in femora with the flexible stem decreased less than 20%, compared to normal. At three years postimplantation, the bone mineral content of the femora with the stiff stem was about 50% that of the femora with the flexible stem. Clinically, DEXA revealed that 5%-15% changes in bone mineral density at various periimplant sites were common within the first two years after surgery; these changes were not usually evident roentgenographically. Serial roentgenographically distinct bone resorption was usually associated with bone mineral density changes of 20%-50%. Five- to 13-year roentgenographic follow-up observations of 213 cases with the Anatomic Medullary Locking prosthesis showed that pronounced bone resorption occurred in 33% of patients. Larger stems (greater than 13 mm in diameter) and stems with extensive porous coating had a significantly higher incidence of pronounced bone resorption than smaller stems and those with proximal coating. The stiffness characteristics of the human femur were established as a function of canal size and compared with those of noncemented hip prostheses. Increased mechanical compatibility was found for stems made of titanium alloy and with design features that reduce cross-sectional area and moment of inertia. Clinical data suggest that to reduce the likelihood of pronounced bone resorption, it would be beneficial for the implant to possess a bending stiffness of about one half to one third that of the human femur.