Characterization of Material Loss from Femoral Stem Taper Surfaces through Development of a Responsive Morphological Filtering Technique

The presence of taper junctions in hip arthroplasty designs has been suggested to contribute to premature clinical failure, in part due to corrosion-assisted material loss. Characterization of the material loss at the taper junction is an important factor in determining not only taper performance but also the requirement for revision of primary stems. It is therefore clear that a robust metrological solution to this issue is required to understand the connection between damage to the stem surface and explicit criteria to initiate retrieval. Previously detailed studies have focused on the characterization of tapers with plain-machined surfaces, with most research concentrated on femoral head tapers. This study details a measurement and analysis method to characterize threaded-type microgroove trunnion surfaces that make up a large proportion of current implanted stem taper designs. Femoral stem designs from two manufacturers were measured using a Talyrond 365 (Ametek, Inc., Berwyn, PA) out-of-roundness measurement machine. Submicron measurement resolution was achieved, and the use of a 5-µm radius diamond stylus meant that any mechanical filtering effect was insignificant compared to conventional contact measurement techniques. The grooved topography of the trunnion surface presents challenges in ascertaining an “unworn” surface from which surface damage data are then separated. A novel filter was developed to delineate form and material loss. The filter is responsive such that the wavelength and amplitude of an individual component’s microgrooves can be removed from the analysis. Application of the filter to leveled data allowed bearing area analysis to be utilized to determine an accurate volume of material loss based on the residual form-removed data. A cohort of as-manufactured samples (n = 21) were measured and used to validate the filtering methodology. A component exhibiting induced surface wear (n = 1) was measured using the same instrumentation and processed using the filtering technique.