Ultra high molecular weight polyethylene is the bearing surface of choice for implants, but because of wear, it is often a limiting factor for the longevity of the implant. Understanding the wear potential of new implants such as the total disc replacement (TDR) is critical due to their rise in clinical implementation. Unidirectional motion shows dramatically lower wear rates then multidirectional motion due to a phenomenon known as cross-shear. Previously, numerical wear models did not incorporate cross-shear, limiting the accuracy of the model for prediction of wear over multiple loading conditions. The objective of this study was to develop an adaptive finite element model to study the effects of cross-shear wear of TDR over multiple loading conditions. The framework for the calculation of wear was developed from a previous wear model used in the hip and knee, and updated to include cross-shear based upon the A/(A+B) theory of Turell et al. Based on the sliding distances and directions, the model computed a wear factor for each node which was used to adaptively compute the wear at each update interval representing ten million wear cycles. The platform was applied to the lumbar ProDisc implant (ProDisc-L) TDR, and run for 10 million wear cycles with multiple loading conditions to find how the varying cross-path motions affected wear. Next, a cervical ProDisc (ProDisc-C) implant was run with three different loading conditions and compared to experimental results to validate the cross-shear based model using the A/(A+B) method and a modified A/(A+B) method.

As expected with unidirectional motion, the lumbar model predicted the wear to be essentially zero. With multidirectional motion, the more cross path motion experienced, the higher the wear. Using the A/(A+B) method, the cervical model showed good agreement in trend to the experimental wear results, while the modified A/(A+B) method showed good agreement in trend and value to the experimental results. Even though this model uses a simplification of the wear process, the results prove that by incorporating cross-shear into a numerical wear analysis, the wear rates can be accurately predicted.