AN INVESTIGATION into the FRICTIONAL PROPERTIES between BONE and VARIOUS ORTHOPEDIC IMPLANT SURFACES - IMPLANT STABILITY

Background: The frictional properties between bone and uncemented orthopedic implants are important for initial stability. Accurate frictional data are used in computer modeling to improve implant design. Unfortunately, friction data derived from standardized testing methods is rare. We aimed to validate two testing methods and to provide friction data. Method: Two experimental designs for testing friction (Sled and Rotation) were investigated. One (Sled) employed a friction testing standard (ASTM D4518) while the other (Rotation) was a novel design. Both designs were initially validated using metal and polymer surfaces of known friction coefficients. The effect of variables, speed and normal force was evaluated. Finally trabecular bone from fresh frozen cadaveric femurs (n=4) was tested against four commonly used orthopedic surfaces. Porosity of each bone sample was obtained using micro-CT and we reviewed the correlation between porosity and friction coefficient. Results: Based on the validation experiments, the Rotation method delivered more accurate friction estimates, and was less affected by normal force and speed, than the Sled method. The testing of bone against implant surfaces produced a variety of different force displacement curves and a wide range of friction coefficients (range 0.19-0.78). Bone excised from the lateral femoral condyle produced a higher friction coefficient than bone from the medial femoral condyle (p-value <0.05). Trabecular bone samples yielded a range of porosities (91.30-79.03%) as determined by micro-CT. A weak positive correlation was observed between bone porosity and friction coefficient. Conclusion: This study demonstrated the importance of a validated testing method when determining frictional properties between bone and orthopedic implants. Specific testing standards should therefore be created, to ensure accurate and reproducible data. The data produced in this study will be utilized for further experimental studies and computer modeling.