Secondary Logo

Journal Logo

Hard-on-Hard Total Hip Impingement Causes Extreme Contact Stress Concentrations

Elkins, Jacob, M., MS1, 2; O’Brien, Megan, K., BS1, 2; Stroud, Nicholas, J., BSE1, 2; Pedersen, Douglas, R., PhD1, 2; Callaghan, John, J., MD1, 2, 3; Brown, Thomas, D., PhD1, 2, 4, a

Clinical Orthopaedics and Related Research: February 2011 - Volume 469 - Issue 2 - p 454–463
doi: 10.1007/s11999-010-1632-x
Symposium: Papers Presented at the Hip Society Meetings 2010
Free
SDC

Background Impingement events, in addition to their role immediately proximate to frank dislocation, hold the potential to damage new-generation hard-on-hard bearings as a result of the relatively unforgiving nature of the materials and designs. Because of the higher stiffness and tighter design tolerances of metal-on-metal and ceramic implants, surgical positioning plausibly has become even more important.

Questions/purposes We asked (1) whether, and under what cup orientation conditions, hard-on-hard impingements might challenge implant material failure strength; and (2) whether particle generation propensity at impingement and egress sites would show similar dependence on cup orientation.

Methods Realistic computational simulations were enabled by multistage finite element analyses, addressing both global construct motion and loading, and focal stress concentrations at neck impingement and rim egress sites. The global model, validated by a cadaveric simulation in a servohydraulic hip simulator, included both hardware components and advanced anisotropic capsule characterization. Parametric computational runs explored the effect of cup orientation for both ceramic-on-ceramic and metal-on-metal bearing couples for two distinct motion sequences associated with dislocation.

Results Stress concentrations from impingement increased nearly linearly with increased cup tilt and with cup anteversion. In some situations, peak values of stress approached or exceeded 1 GPa, levels challenging the yield strength of cobalt-chromium implants, and potentially the fracture strength of ceramics. The tendency for impingement events to generate debris, indexed in terms of a new scraping severity metric, showed orientation dependences similar to that for bulk material failure.

Conclusions Damage propensity arising from impingement events in hard total hip bearings is highly orientation-dependent.

1 Department of Orthopaedics and Rehabilitation, University of Iowa, Iowa City, IA, USA

2 Department of Biomedical Engineering, University of Iowa, Iowa City, IA, USA

3 Iowa City Veterans Administration Medical Center, Iowa City, IA, USA

4 Orthopaedic Biomechanics Laboratory, University of Iowa, 2181 Westlawn, 52242, Iowa City, IA, USA

a e-mail; tom-brown@uiowa.edu

One or more of the authors is a paid consultant for DePuy Orthopaedics, Inc (JJC) and Smith and Nephew Orthopaedics, Inc (TDB). One or more of the authors (TDB, JJC) received funding from the National Institutes of Health (AR46601 and AR53553) and the Veterans Administration (JJC, TDB).

This work was performed at the University of Iowa, Iowa City, IA, USA.

Back to Top | Article Outline

Introduction

Although concerns about impingement usually focus primarily on its role in dislocation, considerations related to impingement damage to the implant components themselves also merit attention. Impingement-associated rim damage has of course been habitually observed in retrieval studies of metal-on-(conventional) polyethylene (Mo[C]P) implants typically involving a majority (approximately 70%) of specimens [11, 16]. Concerns noted in that context include accelerated debris liberation [3], reduced stability, and liner crack/fracture initiation [7]. Because Mo(C)P designs are currently implanted in many millions of patients worldwide, such liner rim damage will be a matter of legacy clinical relevance for the foreseeable future.

The shift to advanced bearing couples, along with the increased head sizes they often afford, raises new considerations regarding implant impingement damage. Modern-era bearing couple advancements have been driven dominantly by the goal of reducing wear. Although successful in that regard, all three of the major classes of new alternative bearings (metal-metal, ceramic-ceramic, and highly crosslinked polyethylene) share the common feature that should impingement occur, the surfaces coming into unintended contact would be seemingly much less “forgiving” than is the case for impingement of a conventional polyethylene liner. Clinical concerns for ceramic components include chips or particles entering the bearing surface as third bodies and catastrophic fracture [2]. For impingement of metal-metal implants, impingement and rim egress sites obviously constitute a plausible source of accelerated particle and ion liberation [18], and there is a possibility for local material failure (yield) to compromise the highly precise bearing surface clearances necessary for successful function of fluid film lubrication. For traditional Mo(C)P designs, the stress concentrations accompanying THA impingement events have been extensively documented [12, 15].

Finite element (FE) analysis allows large number of parametric variants of implant design and component surgical positioning to be straightforwardly considered and provides the ability to input patient motions and joint loadings realistically associated with impingement events [13]. However, similarly conceived studies of hard-on-hard (HoH) impingement events face a major new challenge biomechanically. For HoH implants, the very localized mechanical contact at the site of neck impingement on the liner rim, and for the acetabular rim line loading at the egress site, involves much smaller physical areas and much more severe spatial gradients of stress concentration than occur for Mo(C)P impingements. Whole-implant finite element zonings with spatial resolution sufficient to realistically capture stresses at HoH impingement/egress sites unfortunately are logistically prohibitive from a computational resource viewpoint. We used a multistage approach to address this limitation.

The specific questions addressed were (1) whether, and under what cup orientation conditions, hard-on-hard impingements might approach or exceed implant material failure strength; and (2) whether the tendency for particle generation resulting from scraping at impingement and egress sites would show similar dependence on cup orientation.

Back to Top | Article Outline

Materials and Methods

Realistic analysis of impingement, subluxation, and dislocation of total hip impingement requires accurate model construction and input of accurate motion kinematics at the whole-construct level. This includes accounting for the large-scale deformations of the periarticular soft tissue and for the large sliding distances between abnormally contacting surfaces (Fig. 1). These considerations, in addition to the need for computational economy, place a premium on innovative planning of the global meshing structure. At the same time, quantifying contact phenomena occurring during short time periods at highly localized impingement and rim egress sites necessitates a distinctly different approach to finite element modeling: very high meshing densities at specific locations that can only be known a posteriori from global analysis. To simultaneously address both of these objectives, a multitiered approach to component meshing was developed in which the first stage required whole-construct analysis at a zoning resolution computationally tractable for analyzing the global impingement event. A subsequent stage involved a refined (coupled) submodel of the sites of stress concentration zoned at sufficiently high resolution (analytically validated) to capture the very high stress gradients and stress concentrations involved for HoH contact. This multistage formulation allowed an appreciation for how severely the local stress concentrations challenge the bulk failure strengths of the respective constituent materials (yield strength for metals, fracture strength for ceramics). We also developed a new scraping severity metric to enable assessment of the relative local debris generation propensities of individual impingement events based on “wear-like” interplay [1] of local mechanical stresses and local sliding distances at the stress concentration sites.

Fig. 1

Fig. 1

Given the highly nonlinear material behavior of the hip capsule and the existence of its multiple contact surfaces, an explicit/dynamic FE solution was required to avoid the numeric instabilities that typically arise in quasistatic (implicit) equilibrium FE solutions of such systems. Time steps used for the (explicit) kinetic and kinematic inputs were physiologically realistic as determined from motion capture from subjects performing dislocation-prone activities [13]. Output metrics of interest for the global solutions included impingement-free range of motion (ROM), ROM to frank dislocation, and the development of resisting moment. The latter is the internal resistance resulting from mechanical equilibrium of moment arms developed from all component-on-component contact interactions and from capsule distension during impingement.

The global FE model (Fig. 1) consisted of four parts: the femoral component (28-mm head), the liner (outer diameter 46 mm), the metal backing, and the native tissue bed (periarticular bone and the capsule). The capsule was modeled from CT-derived geometry of a cadaveric hemipelvis, in which major fiber bundle orientations had been identified on the exposed intact surface and which had been imported into an anisotropic hyperelastic constitutive material model. The FE zoning for the implant hardware was developed from manufacturer-provided surface geometry (IGES) files with True Grid (Version 2.3, XYZ, Scientific Applications, Inc, Livermore, CA) as the mesh preprocessor. Each global model was accompanied by a corresponding separate submodel, which was preprocessed in a fashion similar to that for the global components. The metal-on-metal (MoM) implants were modeled as linearly elastic (cobalt-chromium [CoCr], elastic modulus E = 210 GPa, Poisson’s ratio ν = 0.3, density ρ = 9.2 g/cc) with radial clearance δ of 0.029 mm and with a friction coefficient μ of 0.1 [4]. The ceramic portions of the ceramic-on-ceramic (CoC) constructs were also assigned linear elastic material coefficients (alumina, E = 380 GPa, ν = 0.23, ρ = 3.98 m/cc), δ = 0.034 mm, and μ = 0.04 [4].

For each global analysis, the metal backing was rigidly fixed in space by constraining nodal rotation and translation. Frictional interactions maintained the acetabular liner within the metal shell. The baseline orientation of the femoral component was at 5° of anteversion. All rotations and loads associated with each dislocation motion sequence were prescribed at the center of the femoral head, and each FE analysis was run either until completion of the maneuver or until frank dislocation occurred. Two separate dislocation-prone motion sequences were considered in this investigation using motion data previously collected for 10 subjects performing (1) a sit-to-stand from a normal height and (2) a stooping motion [13]. These entire motion sequences were discretized into multiple incremental steps (54 for sit-to-stand, 43 for stooping), each with prescribed triaxial rotations and time-variant three-dimensional joint reaction force. ROM before impingement, resisting moment, and the distribution of von Mises stress at the impingement and egress sites were registered for each run. For those analyses that resulted in frank dislocation, the output data were truncated after 4 mm of femoral head subluxation.

Besides the risk of damage to the bulk implant material (ie, MoM yield, CoC fracture), another concern with HoH impingement is the potential for liberation of particulate debris resulting from the localized scraping motion(s) occurring at very highly stressed impingement and/or egress sites. Classically [6], impingement sites have been conceptualized as (fixed) fulcra for lever-out dislocation. Contemporary three-dimensional analysis [12] has shown, however, that appreciable relative sliding/scraping motion typically occurs at impingement and egress sites during the additionally still stable hip motion, often up into the 10° to 20° range, that remains available after initial impingement. Of course, the incidence of impingement [11] is far higher than that of frank dislocation [5], so only a small fraction of impingement events would logically involve the full amount of impingement/egress site scraping potentially possible during impingement-initiated head subluxation.

The micromechanics of debris liberation at local scraping sites is obviously complex, and its definitive analysis lies beyond the scope of the present study. To index relative severity, however, this scraping phenomenon might be viewed as a very short-term abrasive/adhesive wear event, for which debris production would depend on the interaction of local mechanical stress, local sliding distance, and mutual tribologic abrasiveness of the two contacting surfaces. A basis for quantifying the instantaneous local wear rate, Symbol, from such interaction is provided by the Archard-Lancaster formulation: Symbol, in which σ is local mechanical stress, Symbol is the local relative sliding speed of the two contacting surfaces, and k is a wear factor that reflects contacting surface abrasiveness [1]. Relative deleteriousness of specific scraping events, in terms of propensity to liberate debris at impingement sites and/or egress sites, was assessed by temporospatially integrating this instantaneous local “wear” over all acetabular liner elements that experienced either neck contact (impingement sites) or line contact with the subluxating femoral head (egress sites).

Symbol

Symbol

Symbol

Symbol

Symbol

Symbol

To strike a balance between solution accuracy and computational logistic cost, mesh sensitivity and convergence studies were conducted to identify a mesh structure appropriate for the large-scale, large-deformation analysis. However, mesh structures deemed suitable for such purpose are far too coarse for accurate stress concentration assessment at local sites of interest during a dislocation event. Additional convergence studies as well as comparisons with analytical Hertzian [17] contact solutions identified appropriate meshing densities for these localized analyses (Fig. 2). Attempting to run the explicit global model with such small element sizes would have required computational runtimes measured in months, wholly intractable for considering a wide parametric range of cup orientations. For this reason, a multitiered modeling approach using submodeling was pursued (Fig. 3) allowing for high-mesh-density component models to be driven by solutions passed from the large-scale global model. The global FE model behavior was physically validated using novel transpelvic implantation of THA-replicating specialty hardware in a cadaveric pelvis. This transpelvic procedure allowed for surgical placement of hardware without violating the capsule integrity, thus allowing the full contribution of a healed hip capsule to be modeled. A cadaveric hemipelvis, with THA-analog hardware components having identical bearing geometry as those in the FE model, was fixed to a purpose-built four degree-of-freedom servohydraulic hip simulator. Under axial loading, the resultant moment was recorded for a sit-to-stand simulation. The measured resultant moment behavior agreed very favorably with that for an FE analysis for identical loading and motion sequence inputs (Fig. 4). The veracity of the computed contact mechanics from the FE model was further substantiated using a Hertzian analytical solution for a situation of edge loading between the head and liner during head subluxation using a partially nested sphere-in-torus formulation (Fig. 5). Given the occurrence of line contact, computed contact pressure as predicted by the FE model was highly sensitive to mesh structure. Also, this analytical Hertz contact solution constituted a gold standard for identifying necessary resolution for the submodeling analyses.

Fig. 2

Fig. 2

Fig. 3A-C

Fig. 3A-C

Fig. 4

Fig. 4

Fig. 5

Fig. 5

One hundred forty-eight separate global FE simulations were executed (Table 1) using Abaqus/Explicit, each one followed by a separate submodel analysis executed using Abaqus/Standard. Fifty-four distinct cup orientations were considered with cup tilt (lateral opening or inclination) and anteversion being defined relative to a pelvic reference frame [14]. Each FE simulation required approximately 120 processor-hours of computer time on a dual quad-core Intel ® Xeon platform configured with 24 GB of RAM running a 64-bit Linux operating system.

Table 1

Table 1

Table 1

Table 1

Back to Top | Article Outline

Results

Values for peak surface effective vM stresses at the egress and impingement sites were seen to increase approximately linearly with cup tilt and cup anteversion over most of the range of cup orientations modeled (Fig. 6). However, at some extreme conditions of cup orientation, impingement either was relatively short in duration or even failed to occur (Table 1). In such circumstances, appreciable egress-site stresses did not develop (Fig. 7). Egress-site stress, because of the highly localized nature of rim line-loading, involved stress concentrations that were higher than those at the impingement site (Figs. 6-8). Peak stresses developed for the stoop dislocation sequence were consistently higher than those for the sit-to-stand challenge for any given cup orientation (Figs. 6-8). For the ceramic bearing, peak stresses at the egress site were consistently greater than those for the MoM bearing owing to the higher modulus for the ceramic versus metal liner (Fig. 7). However, this dependency was less sensitive for stresses occurring at the impingement site (Fig. 8). Frank dislocation occurred in 88 of the 148 cases run (Table 1). Impingement-free ROM, similar to the trend for contact stress, also increased approximately linearly with increasing cup tilt and anteversion (Fig. 9). ROM before component impingement was consistently greater for the sit-to-stand challenges (111° versus 105°). The kinematics for the stooping sequence, although still flexion-dominated, involved significantly more internal rotation and abduction than those for the sit-to-stand motions, thus tending to reduce the flexion ROM (Fig. 10).

Fig. 6

Fig. 6

Fig. 7A-D

Fig. 7A-D

Fig. 8A-D

Fig. 8A-D

Fig. 9A-B

Fig. 9A-B

Fig. 10A-B

Fig. 10A-B

Computed values for volumetric “wear” from the scraping events for the MoM impingement simulations show that metal-surface vM stresses were highly coupled (concordance coefficient = 0.833) with scraping severity (Table 1) regarding their respective cup orientation dependencies. Computed scraping wear at the egress site generally exceeded that at the impingement site (Fig. 11). Across the range of tilt-abduction combinations studied, the egress sites accounted for 57% of the overall amount of computed volumetric scraping wear.

Fig. 11

Fig. 11

Back to Top | Article Outline

Discussion

At present, almost nothing is known about the extreme local mechanical conditions that would likely prevail for impingement of new-generation advanced bearing implants. Because experience has amply taught that potentially adverse phenomena in “new” THA designs merit proactive scrutiny, a study was designed to address the currently least explored of such conditions: impingement of hard-on-hard bearings. We investigated the role of surgical component positioning as it relates to propensity for (1) bulk material damage arising from high localized stress concentrations; and (2) particle/debris liberation from local scraping sites associated with THA impingement.

Our study is subject to a number of assumptions and limitations. First, at present, information regarding the micromechanics of debris creation resulting from impingement is largely lacking. The present quantification of such debris in terms of “scraping” wear through the Archard formula is obviously a substantial simplification of what are undoubtedly very complex local tribologic interactions at these sites. Moreover, the specific k factor used in this analysis was inferred from articulation of fluid-film-lubricated polished CoCr surfaces [10] and therefore probably substantially underestimates the abrasiveness of nonlubricated, nonpolished contacting metal surfaces. For these reasons, the absolute values of computed volumetric scraping wear need to be interpreted with caution. Second, the dislocation motion sequences investigated represent only a small fraction of possible impingement/dislocation scenarios. In vivo, implanted THA components obviously encounter a seemingly indeterminable variety of impingement-consequential motions, and such impingements as do occur no doubt involve a wide range of severities. The stress results presently reported, which involve motion inputs for two of seven dislocation-associated maneuvers that previously had been studied under controlled laboratory conditions [13], therefore represent only a very sparse sampling of what is no doubt a very wide variety of impingement events spontaneously occurring in the THA population. Third, a somewhat related limitation of this study is that the motion inputs were kinematically driven. The femoral head center underwent a prescribed (triaxial) rotation relative to the acetabular component center regardless of the rotation resistance that developed if/when neck impingement occurred. (Appropriately, the FEA formulation allowed the femoral head to then subluxate or even to dislocate in response to that rotation resistance, exactly as would occur physically for a patient single-mindedly executing the specified hip motion maneuver regardless of consequences.) Clinically, however, proprioception and/or pain onset could intervene to cause the patient to instead reflexively terminate the maneuver, thus developing less intense impingement stresses than would otherwise occur. Fourth, clearly, a great many factors besides cup orientation and motion challenge also influence THA impingement. The list includes femoral component malalignment, implant design (eg, femoral head diameter, neck cross-section, head/neck ratio, etc), fixation/ingrowth integrity, capsule compromise, and muscle activation. Many such factors are straightforwardly amenable to study with the present FEA model, and others can be addressed by augmentations of the formulation.

Surface stresses developed at hard-on-hard THA impingement/egress sites are highly sensitive to component orientation. The extremely high stress levels computed here are among the most severe mechanical environments yet quantified in the orthopaedic literature. For many cup orientations in this series, peak computed von Mises stress approached or in some cases even broached into the GPa range. Situations of high cup anteversion involved particularly high stresses. In the case of MoM, the material failure modality of primary concern is that the yield strength of the CoCrMo alloy would be exceeded, leading to localized permanent (plastic) deformation. Specific assumptions, simplifications, and limitations in the present FE formulation (see previously) might of course be questioned. By prevailing standards, however, most features of the present model are well enough grounded both physiologically and biomechanically, that computed stresses in the near-GPa range are disturbing. The yield strengths for cast and wrought CoCrMo alloy are 450 and 827 MPa, respectively [8]. For ceramic bearings, the concern of course is not ductile (yield) failure, but rather catastrophic fracture. The mechanical considerations bearing on crack nucleation and crack propagation in brittle solids such as ceramics are much more complex than those for yield failure in ductile materials such as metals. Quantifying ceramic fracture propensity requires a formal fracture mechanics analysis, an endeavor beyond the scope of the present numeric formulation. As a point of reference, however, in ongoing work we have developed a formal fracture mechanics FE model of ceramic liner impingement from which a pilot case has shown a computed stress intensity factor of 1.90 MPa*m1/2, a value not insignificantly below the ISO specifications for alumina ceramic [9] (critical stress intensity factor = 4.0 MPa*m1/2, fatigue stress intensity factor = 1.0 MPa*m1/2). Taken as a whole, therefore, there seems good cause for concern that impingement events, if and when they should occur, pose a significant threat to the refined engineering integrity of HoH contemporary THA bearings.

Creation of particulate debris at impingement and egress sites is a second concern. Such particles would have only a very short path to migrate into the bearing surface to become third bodies, thus compromising otherwise excellent wear performance. Additionally, such debris would constitute direct particulate burden in the periarticular tissue bed and/or in distant organs. In the case of metal particles, the surface-to-volume ratio is enormously higher than for the bulk implant members, favoring dramatic acceleration of the metal ion release rate, possibly contributing to the formation of aseptic lymphocytic vasculitis-associated lesions or other immunologic reactions [19].

The present parametric results convincingly demonstrate that (1) some impingement events are far more deleterious than others in terms of propensity to generate such debris; (2) the egress site is a substantially greater cause of concern in that regard than is the impingement site; and (3) the cup orientations that are especially problematic in terms of bulk-level implant material failure generally correspond to the cup orientations that are also especially problematic in terms of scraping wear. The multistage formulation (global construct analysis followed by mesh-refined submodels in the impingement and egress regions) identified severe stress concentrations at those critical sites, plausibly to levels of concern from the standpoint of implant material failure (yield or fracture). The local stress data were also used to index the propensity for generation of particulate debris resulting from scraping at the impingement and egress sites. Because HoH bearings represent a very compelling path forward from the standpoint of minimizing bearing surface wear, it is important that both the orthopaedic surgical community and the orthopaedic industry be positioned to systematically address potential problems that may occasionally arise with these new designs.

Back to Top | Article Outline

Acknowledgments

We thank DePuy, Inc. for providing technical data regarding design parameters of the implant components; Dr Darryl D’Lima for helpful suggestions regarding the scraping severity metric; and Mr. Tony Sanders for generously giving valuable advice and direction regarding the development of the Hertzian edge-loading contact model.

Back to Top | Article Outline

References

1. Archard, JF. Contact and rubbing of flat surfaces. J Appl Phys 1953; 24: 981-988. 10.1063/1.1721448
2. Barrack, RL., Burak, C. and Skinner, HB. Concerns about ceramics in THA. Clin Orthop Relat Res 2004; 429: 73-79. 10.1097/01.blo.0000150132.11142.d2
3. Barrack, RL. and Schmalzried, TP. Impingement and rim wear associated with early osteolysis after a total hip replacement—a case report. J Bone Joint Surg Am 2002; 84: 1218-1220.
4. Brockett, C., Williams, S., Jin, Z., Isaac, G. and Fisher, J. Friction of total hip replacements with different bearings and loading conditions. J Biomed Mater Res Appl Biomater 2007; 81: 508-515.
5. Brown, TD. and Callaghan, JJ. Impingement in total hip replacement: mechanisms and consequences. Curr Orthop 2008; 22: 376-391. 10.1016/j.cuor.2008.10.009
6. Coventry, MB. Late dislocations in patients with Charnley total hip arthroplasty. J Bone Joint Surg Am 1985; 67: 832-841.
7. Holley, KG., Furman, BD., Babalola, OM., Lipman, JD., Padgett, DE. and Wright, TM. Impingement of acetabular cups in a hip simulator—comparison of highly cross-linked and conventional polyethylene. J Arthroplasty 2005; 20: (Suppl 3):77-86. 10.1016/j.arth.2005.04.026
8. ISO 5832. Implants for Surgery—Metallic Materials. Geneva: International Organization for Standardization; 1996.
9. ISO 6474. Implants for Surgery—Ceramic Materials Based on High Purity Alumina. Geneva: International Organization for Standardization; 1994.
10. Liu, F., Leslie, I., Williams, S., Fisher, J. and Jin, Z. Development of computational wear simulations of metal-on-metal hip resurfacing replacements. J Biomech 2008; 41: 686-694. 10.1016/j.jbiomech.2007.09.020
11. Lundberg, HJ., Liu, SS., Callaghan, JJ., Pedersen, DR., O’Rourke, MR., Goetz, DD., Vittetoe, DA., Clohisy, JC. and Brown, TD. Association of third body embedment with rim damage in retrieved acetabular liners. Clin Orthop Relat Res 2007; 465: 133-139.
12. Nadzadi, ME., Pedersen, DR., Callaghan, JJ. and Brown, TD. Effects of acetabular component orientation on dislocation propensity for small-head-size total hip arthroplasty. Clin Biomech (Bristol, Avon) 2002; 17: 32-40. 10.1016/S0268-0033(01)00096-1
13. Nadzadi, ME., Pedersen, DR., Yack, HJ., Callaghan, JJ. and Brown, TD. Kinematics, kinetics, and finite element analysis of commonplace maneuvers at risk for total hip dislocation. J Biomech 2003; 36: 577-591. 10.1016/S0021-9290(02)00232-4
14. Pedersen, DR., Callaghan, JJ. and Brown, TD. Activity-dependence of the ‘safe zone’ for impingement versus dislocation avoidance. Med Engr Phys 2005; 27: 323-328. 10.1016/j.medengphy.2004.09.004
15. Scifert, CF., Noble, PC., Brown, TD., Bartz, RL., Kadakia, N., Sugano, N., Johnston, RC., Pedersen, DR. and Callaghan, JJ. Experimental and computational simulation of total hip arthroplasty dislocation. Orthop Clin North Am 2001; 32: 552-567. 10.1016/S0030-5898(05)70226-1
16. Shon, WY., Baldini, T., Peterson, MG., Wright, TM. and Salvati, EA. Impingement in total hip arthroplasty: a study of retrieved acetabular components. J Arthroplasty 2005; 20: 427-435. 10.1016/j.arth.2004.09.058
17. Ugural, AC. and Fenster, SK. Advanced Strength and Applied Elasticity, 4th ed. Upper Saddle River, NJ: Prentice Hall; 2008.
18. Urban, RM., Tomlinson, MJ., Hall, DJ. and Jacobs, JJ. Accumulation in liver and spleen of metal particles generated at nonbearing surfaces in hip arthroplasty. J Arthroplasty 2004; 19: (Suppl 3):94-101. 10.1016/j.arth.2004.09.013
19. Willert, HG., Buchhorn, GH., Fayyazi, A., Flury, R., Windler, M., Koster, G. and Lohmann, CH. Metal-on-metal bearings and hypersensitivity in patients with artificial hip joints. A clinical and histomorphological study. J Bone Joint Surg Am 2005; 87: 28-36. 10.2106/JBJS.A.02039pp
© 2011 Lippincott Williams & Wilkins LWW