Purpose: Implant retrieval analysis provides valuable information to researchers, clinicians and manufactures about the in vivo wear characteristics. Wear analysis can quantify visible damage on retrieved UHMWPE bearing surfaces used in total knee arthroplasty (TKA). Various researchers have developed wear scores to help provide insight into the modes of prosthetic failure. Four scoring methods are compared in order to determine their inter-rater reliability.

Method: A cohort of 60 retrieved G2, Smith & Nephew/Richards, TKA implants were analyzed by two observers using four scoring techniques. The scoring methods used had been developed by Hood, Wasielewski, Currier, and Brandt. The intraclass correlation coefficient (ICC) was used in assessing the inter-rater reliability.

Results: The ICC demonstrated significant correlation between raters (P< 0.05). Hood’s method produced a correlation of 0.89, Wasielewski’s method 0.83, Currier’s method 0.58, and Brandt’s method 0.82. All but Currier’s method had excellent correlation between raters.

Conclusion: The analysis techniques for articular surface wear for total knee bearings by Hood, Brandt, and Wasielewski showed excellent inter rater reliability; however currier’s method was only fair. One common issue among all these methods is that the scoring systems do not identify or assign differentiating weights to clinically relevant wear modalities to capture inferior implants designs. A new wear analysis technique that is efficient and assigns clinical severity weights to wear modes in TKA bearings should be investigated.

Purpose: The accuracy and precision for shoulder radio-stereometric analysis (RSA) is not as well documented as for hip and knee replacement implants. Shoulder replacement glenoid component have a relatively high rate of aseptic loosening when compared to primary hip and knee replacement components. The purpose of this study is to validate our marker based RSA system for a shoulder phantom using computed radiography.

Method: A Sawbones humerus was surgically prepared with a total shoulder implant by an experienced orthopaedic surgeon. A pegged glenoid component (3 pegs) previously marked with 7 tantalum beads was cemented into a Sawbones scapula. The glenoid component was mounted to a 32mm thick acrylic plate. The simulated humerus with implant was fixed to a linear translation stage. The stage was able to move in 0.010 mm increments with an accuracy of 0.002 mm. The Humeral component was then incrementally moved along the x, y, and z axis from 0 to 0.050, 0.100, 0.150, 0.200, 1.000 mm with duplicates taken at each increment. This examination was performed a total of 3 times. From these 9 RSA exams, the accuracy and precision of the UmRSA Digital Measure V6.0 RSA system was determined from 90 pairs of linear displacements.

Results: The standard deviation of the total average error for the X, Y, Z axis were 0.023, 0.022, and 0.070 mm respectively. The accuracy for phantom shoulder model using computer radiography was 0.008 mm in the medial direction, 0.007 mm in the superior direction and 0.019 mm in the anterior direction. The corresponding precision measurements were 0.005, 0.005, 0.015 mm.

Conclusion: This preliminary assessment of accuracy and precision of a shoulder phantom model illustrates that marker based RSA is a useful system to monitor the micro-motion of total shoulder designs.