Methods

105 femoral heads with a Goldberg score 4 were evaluated. Coordinate measuring machine data was used to compute material loss and visualize damage features. Time in situ and stem alloy were identified. Metallographic samples were produced for each case. Grain size and banding were identified using light-microscopy. Mann-Whitney tests were conducted to compare material loss between groups.

Methods

Finite element models of generic cobalt-chromium (CoCr) heads on a titanium stem taper were developed and driven using dynamic assembly loads collected from clinicians. To verify contact mechanics at the taper interface, comparisons of deformed microgroove characteristics (height and width of microgrooves) were made between model estimates with those measured from five retrieved implants. Additionally, these models were used to assess the role of assembly technique—one-hit versus three-hits—on the taper interlock mechanical behaviour.

Methods

In total, 11 retrieved polyethylene tibial inserts were matched with gait analysis testing performed on those patients. Volumetric wear on the articular surfaces was measured using a laser coordinate measure machine and autonomous reconstruction. Knee kinematics and kinetics from individual gait trials drove computational models to calculate medial and lateral tibiofemoral contact paths and forces. Sliding distance along the contact path, normal forces and implantation time were used as inputs to Archard’s equation of wear to predict volumetric wear from gait mechanics. Measured and modelled wear were compared for each component.

Methods

A two-dimensional, axisymmetric FEA model representative of a CoCrMo femoral head taper and Ti6Al4V stem taper was created using median geometrical measurements taken from over 100 retrieved implants.³ Surface finish—micro-grooves—of the head/stem taper were modeled using a sinusoidal function with amplitude and period corresponding to retrieval measurements of micro-groove height and spacing, respectively. Two stem taper micro-groove geometries— “rough” and “smooth”—were modeled corresponding to the median and 5^th percentile height and spacing measurements from retrievals. All models had a 3' (0.05°), proximal-locked angular mismatch between the tapers.

To simulate implant assembly during surgery, multiple dynamic loads (4kN, 8kN, and 12kN) were applied to the femoral head taper in a sequence of one or three strikes. The input load profile (Figure 1) used for both cases was collected from surgeons assembling an experimental setup with a three-dimensional load sensor. Models were assembled and meshed in ABAQUS Standard (v 6.17) using four-node linear hexahedral, reduced integration elements. Friction was modeled between the stem and head taper using surface-to-surface formulation with penalty contact (µ=0.2). A total of 12 implicit, dynamic simulations (3 loads × 2 assembly sequences × 2 stem taper surface finishes) were run, with 2 static simulations at 4kN for evaluating inertial effects. Outcome variables included contact area, contact pressure, equivalent plastic strain, and pull-off force.

Methods

Eleven retrieved ultra-high molecular weight polyethylene (UHMWPE) cruciate-retaining tibial liner components from ten separate patients (implantation time = 8.6±5.6 years) had matching gait trials of normal level walking for each knee. Volume loss on retrieved components was calculated using a coordinate measuring machine and autonomous reconstruction.³ Motion analysis of normal level walking gait had been conducted between 1986 and 2005 for various previous studies and stored in a consented Human Mechanics Repository, ranging from pre-operative to long-term post-operative testing. Contact location between the femoral component and the tibial component on the medial and lateral plateaus were calculated throughout stance.⁴ A previously validated and fine-tuned parametric numerical model was used to calculate TKR contact forces for each gait trial.⁵ Vertical contact forces and contact paths on the medial and lateral plateaus were input as normal force and sliding distance to a simplified Archard equation for wear with material wear constant = 2.42 × 10⁻⁷ mm³/Nm^2,6 to compute average wear per gait cycle. Wear rates were calculated using linear regression, and Pearson correlation examined correlations between modeled and measured wear.

Methods

We developed a computational framework for predicting TKA contact mechanics and wear. The framework accepts multiple forms of input data: patient-specific, population-specific, or standardized motions and forces. CAD models are used to create the FEA mesh. An analytical wear model, calibrated from materials testing (wheel-on-flat) experiments, is fully integrated into the FEA process. Isight execution engine runs a design of experiments (DOE) analysis with an outcome variable, such as volumetric wear, to guide statistical model output. We report two DOE applications to test the utility of the computational framework for performing large variable studies in an efficient manner: one to test the sensitivity of TKA wear to the femoral center of rotation, and the second to test the sensitivity of TKA wear to gait input perturbations.