Embed Size (px)
Citation preview
Computational Fracture Analysis of Ceramic-on-Ceramic Total Hip Liners
+1Elkins, JM; 2Liu X; 1Pedersen, DR; 1,3Callaghan, JJ; 1Brown, TD +1Department of Orthopaedics and Rehabilitation, University of Iowa, Iowa City, IA, 2SIMULIA, Providence, RI, 3VAMC, Iowa City, IA.
[email protected]@uiowa.edu INTRODUCTION: Alumina ceramics for total hip arthroplasty (THA) were introduced nearly four decades ago, to address concerns over polyethylene-particle-induced osteolysis and to improve long-term results in younger and more active THA patients. While even more popular abroad, usage of ceramic-on-ceramic (CoC) bearings in the United States is increasing, and currently represents 14% of all THAs performed annually [1]. However, due to the brittle nature of ceramic materials, concerns persist regarding implant failure due to catastrophic fracture. In general, both components of a CoC implant are prone to fracture. Fracture of the ceramic head is a well recognized problem historically, and extensive investigation has led to several design-specific improvements. As a result, fracture rates of ceramic heads have decreased from 13% in 1st generation alumina to less than 0.004% for contemporary ceramics [2]. In contrast, for ceramic liners systematic analysis regarding fracture risk mitigation has been much more limited. Consequently, current fracture rates for alumina liners are approximately 400-fold higher than for ceramic heads. While it is well established clinically that impingement between the femoral neck and liner can predispose to fracture, to date, little quantitative information exists regarding fracture propensity for liners. To help close this knowledge gap, an eXtended Finite Element Model (XFEM) of THA impingement was developed to investigate fracture risk and crack propagation for various implant designs, surgical orientations and patient factors.
METHODS: A previously developed and physically validated [3] non-linear dynamic FE model of THA impingement was used to determine stresses developed during impingement scenarios for a two CoC implants (Fig. 1). For the 28mm implant (Fig. 1a), the effect of cup edge chamfer radius was investigated for four separate cup edge profile geometries (Fig. 2a). Four additional models were generated to investigate surgical cup orientation by varying the cup inclination between 30° and 60°, each with a constant 10° of acetabular anteversion. For these models, a lateral stooping impingement challenge was modeled, using previously reported opto-electronic data [4].
Figure 1: Global THA impingement models demonstrating the development of tensile stresses during an impingement event, for both 28mm (a) and 36mm (b) CoC implants. These stresses are then passed to a separate XFEM submodel to analyze fracture imitation/propagation.
For the 36mm implant (Fig. 1b), fracture initiation and propagation was investigated for two fracture-prone challenges (squatting and stooping) for 25 variations in cup orientation (45° ± 15° inclination, 15° ± 15° anteversion). Since obesity has been identified as a risk factor for liner fracture [ref], two distinct BMIs were considered: normal (25) and morbidly obese (50). These impingement models were executed in Abaqus/Explicit. Stresses occurring during the Explicit simulations were passed (node-based) to the separate (Abaqus/Standard) XFEM model of liner fracture. The XFEM submodel contained two separate fracture-enhanced enrichment regions, enabling fracture initiation at both the impingement and egress sites. Damage initiation criteria were specified at 500MPa maximum principal stress, with literature-based mixed mode (power-law) damage evolution. Since microscopic imperfections (which are often present in sintered ceramics) decrease tensile stresses necessary for
fracture, material properties of alumina were varied to simulate both with and without micro-imperfections for both models.
Figure 2. (a) Four separate edge-profile geometries were investigated for 28mm implants. (b) A propagated fracture across the edge of a 0mm radiused cup edge
RESULTS: For the 28mm implant, fracture initiation was demonstrated to be sensitive to both cup orientation and cup edge radius, with fracture risk increased for sharper edges at higher values of cup inclination. Substantially higher occurrence of fracture was observed for the (assumed flawed) reduced fracture criteria analyses. Fractures occurred predominately at the cup egress region. For the 36mm implant, in the normal BMI simulations, 17 out of 50 fractures occurred in the alumina bearings with micro-imperfections. Spatially, cracks occurred at an intermediate location between the liner edge (Fig. 3) and pole, at the inner edge, at the outer edge, and at the impingement site, in 41%, 41%, 18% and 6% of these fracture instances, respectively. No fractures occurred in the absence of alumina imperfections. In the high BMI group, fracture occurred in 39/50 simulations with micro-imperfections, with fracture occurring in nearly 87% of these instances at the intermediate location, and only 13% at the cup edge. Fracture occurred in three simulations without imperfections, with cracks initiating at the cup edge in all three, which were for cups positioned in 0° of anteversion.
Figure 3: Fracture propagation in the 36mm implant.
DISCUSSION: Fracture of ceramic liners remains a serious concern. As opposed to fracture of the head, quantitative analyses for liner fracture risk are greatly lacking. At present, only a single computational study of ceramic liner fracture mechanics exists in the orthopaedic literature [5], although the methodology it used is not conducive to rapid parametric study of multiple patient-, surgical-, and implant-specific risk factors. The current XFEM methodology, by contrast, facilitates rapid systematic analysis of ceramic liner fracture.
SIGNIFICANCE: Ceramic liner fracture risk was found to be higher for the 28mm implant versus the 36mm implant. Fracture risk was shown to increase at increased cup inclination and for sharp cup edge profiles.
REFERENCES: [1] Bozic K et al. JBJS Am. 2009; [2] Willmann G et al. CORR 2000. [3] Elkins JM et al. JOR 2011; [4] Nadzadi M et al. J. Biomech 2003 ; [5] Elkins JM et al. J. Arthroplasy 2011.
ACKNOWLEDGEMENTS: Support provided by the VAMC and the NIH AR46601 and AR53553
Poster No. 2006 • ORS 2012 Annual Meeting
