A parametric axisymmetric model study on the interface motions in porous-surfaced tibial implants

THE INFLUENCE OF TIBIAL PROSTHESIS DESIGN FEATURES ON STRESSES RELATED TO ASEPTIC LOOSENING AND STRESS SHIELDING

Aseptic loosening caused by mechanical factors is a recognized failure mode for tibial components of knee prostheses. This parametric study investigated the effects of prosthesis fixation design changes, which included the presence, length and diameter of a central stem, the use of fixation pegs beneath the tray, all-polyethylene versus metal-backed tray, prosthesis material stiffness, and cement mantle thickness. The cancellous bone compressive stresses and bone–cement interfacial shear stresses, plus the reduction of strain energy density in the epiphyseal cancellous bone, an indication of the likelihood of component loosening, and bone resorption secondary to stress shielding, were examined. Design features such as longer stems reduced bone and bone–cement interfacial stresses thus the risk of loosening is potentially minimized, but at the expense of an increased tendency for bone resorption. The conflicting trend suggested that bone quality and fixation stability have to be considered mutually for the optimization of prosthesis designs. By comparing the bone stresses and bone–cement shear stresses to reported fatigue strength, it was noted that fatigue of both the cancellous bone and bone–cement interface could be the driving factor for long-term aseptic loosening for metal-backed tibial trays.

Analysis of bone-prosthesis interface micromotion for cementless tibial prosthesis fixation and the influence of loading conditions

A lack of initial stability of the fixation is associated with aseptic loosening of the tibial components of cementless knee prostheses. With sufficient stability after surgery, minimal relative motion between the prosthesis and bone interfaces allows osseointegation to occur thereby providing a strong prosthesis-to-bone biological attachment. Finite element modelling was used to investigate the bone-prosthesis interface micromotion and the relative risk of aseptic loosening. It was anticipated that by prescribing different joint loads representing gait and other activities, and the consideration of varying tibial-femoral contact points during knee flexion, it would influence the computational prediction of the interface micromotion. In this study, three-dimensional finite element models were set up with applied loads representing walking and stair climbing, and the relative micromotions were predicted. These results were correlated to in-vitro measurements and to the results of prior retrieval studies. Two load conditions, (i) a generic vertical joint load of 3 x body weight with 70%/30%M/L load share and antero-posterior/medial-lateral shear forces, acted at the centres of the medial and lateral compartments of the tibial tray, and (ii) a peak vertical joint load at 25% of the stair climbing cycle with corresponding antero-posterior shear force applied at the tibial-femoral contact points of the specific knee flexion angle, were found to generate interface micromotion responses which corresponded to in-vivo observations. The study also found that different loads altered the interface micromotion predicted, so caution is needed when comparing the fixation performance of various reported cementless tibial prosthetic designs if each design was evaluated with a different loading condition.

Addition of a short central extension to surface cemented tibial trays in primary TKA: an in vitro study of the effect on initial fixation stability and its relationship to supporting bone density

BACKGROUND

Short central extensions which do not enter the tibial medullary canal are incorporated to cemented tibial components to increase initial stability in primary total knee arthroplasty. Their role in tibiae of differing preoperative mechanical quality has been little studied.

METHODS

Twelve embalmed cadaveric tibiae were paired and divided into two groups, receiving a similar cemented tibial component with or without a non-cemented short central extension (10 mm diameter, 35 mm length). The specimens were subjected to 6000 cycles of a medially applied 1350 N load. Relative bone-tray displacements were measured and the evolution of inducible and permanent micromotions were computed. The apparent density of the cancellous bone under the tibial tray and at the area to support the extension was computed from computed tomography images of each specimen.

FINDINGS

No significant differences between groups were detected for any parameters. For the group with extension, a significant negative linear correlation (P = 0.009, r(2) = 0.849) was found between the inducible tilt of the tray and the bone density at the zone of the extension. Also a trend towards a negative linear relation (P = 0.07, r(2) = 0.59) was observed for the same group between maximum subsidence and apparent density at the zone of the extension.

INTERPRETATION

The study did not find that the addition of a non-cemented short central extension provides any overall improvement of the initial fixation stability. However, it was found that short extensions may enhance tilting stiffness of the bone-implant construct if bone of sufficient mechanical quality is located around its supporting area.

A Circumferentially Flanged Tibial Tray Minimizes Bone-Tray Shear Micromotion

Aseptic loosening of the tibial component is the major complication of total knee arthroplasty. There is an association between early excessive shear micromotion between the bone and the tray of the tibial component and late aseptic loosening. Using non-linear finite element analysis, whether a tibial tray with a circumferentially flanged rim and a mating cut in the proximal tibia could minimize bone-tray shear micromotion was considered. fifteen competing tray designs with various degrees of flange curvature were assessed with the aim of minimizing bone-tray shear micromotion. A trade-off was found between reducing micromotion and increasing peripheral cancellous bone stresses. It was found that, within the limitations of the study, there was a theoretical design that could virtually eliminate micromotion due to axial loads, with minimal bone removal and without the use of screws or pegs.

Long stemmed total knee arthroplasty with interlocking screws: a computational bone adaptation study

The ability of an interlocking screw fixation technique to minimize bone loss related to stress shielding in the tibia was investigated and compared to the abilities of cement and press-fit fixation. Full bony ingrowth has been associated with greater stress shielding than partial ingrowth; therefore, the effect of intimate bonding of the stem to bone on subsequent bone loss was also studied. A damage- and disuse-based remodeling theory was coupled with a two-dimensional finite element model of the tibia to predict changes in bone remodeling following long stemmed total knee arthroplasty (TKA) for four different fixation techniques (cement, press-fit, interlock with bony ingrowth, and interlock without bony ingrowth). Remodeling changes commenced with the model state variables--bone area fraction, mechanical stimulus, damage, and remodeling activity--at steady-state values predicted by the intact tibia simulation. After TKA and irrespective of fixation technique, the model predicted elevated remodeling due to disuse, in which more bone was removed than replenished. In regions below the tibial tray and along the cortices, the interlocking stem with full bony ingrowth and the cemented stem caused the least amount of bone loss. An interlocking stem with a smooth, matted finish did not reduce the bone loss associated with interlocking fixation.

Finite Element Analysis of Tibial Implants - Effect of Fixation Design and Friction Model

A three dimensional nonlinear finite element model was developed to investigate tibial fixation designs and friction models (Coulomb's vs nonlinear) in total knee arthroplasty in the immediate postoperative period with no biological attachment. Bi-directional measurement-based nonlinear friction constitutive equations were used for the bone-porous coated implant interface. Friction properties between the polyethylene and femoral components were measured for this study. Linear elastic isotropic but heterogeneous mechanical properties taken from literature were considered for the bone. The Tensile behaviour of polyethylene was measured and subsequently modeled by an elasto-plastic model. Based on the earlier finite element and experimental pull-out studies, pegs and screws were also realistically modeled. The geometry of every component was obtained through measurement. The PCA tibial baseplate with three different configurations was considered; one with three screws, one with one screw and two short inclined porous-coated pegs, and a third one with no fixation for the sake of comparison. The axial load of 2000N was applied through the femoral component on the medial plateau of articular insert. It was found that Coulomb's friction significantly underestimates the relative micromotion at the bone-implant interface. The lowest micromotion and lift-off were found for the design with screws. Relative micromotion and stress transfer at the bone-implant interface depended significantly on the friction model and on the baseplate anchorage configuration. Cortical and cancellous bones carried, respectively, 10-13% and 65-86% of the axial load depending on the fixation configuration used. The remaining portion was transmitted as shear force by screws and pegs. Normal and Mises stresses as well as contact area in the polyethylene insert were nearly independent of the baseplate fixation design. The Maximum Mises stress in the polyethylene exceeded yield and was found 1-2 mm below the contact surface for all designs.

Analysis of cementless implants using interface nonlinear friction--experimental and finite element studies

Measured interface nonlinear friction properties are used to develop models to study the short-term fixation response of smooth- and porous-surfaced posts, bone screws, and plates fixed with and without posts/screws. Experimental studies are carried out to validate the model predictions and identify the relative role of posts and screws in fixation of a plate on a polyurethane block under symmetric/eccentric axial compression loads. The idealized Coulomb's friction is also used for the sake of comparison. The incorporation of measured nonlinear, rather than the idealized Coulomb, friction is essential to compute realistic results. For plate fixation, the experimental and finite element results show that the screw fixation yields the stiffest response followed by the smooth- and then porous-coated post fixation. For example, under 1000 N eccentric axial compression, the edge of the plate opposite the loaded edge is measured to lift by 1147 +/- 72, 244 +/- 38, or 112 +/- 28 microns, respectively, for the cases with no fixation, with smooth-surfaced posts, or with screws. The corresponding models predict, respectively, values of 1538, 347, or 259 microns and also 556 microns for the plate fixed with porous coated posts. The satisfactory agreement between numerical and experimental results confirms the importance of proper interface modelling for the analysis of posts, screws, and complex fixation systems. This becomes further evident when considering cementless implants in which the bone-implant interface exhibits relatively large displacements as the maximum resistance force is reached. The developed models can be used to investigate the post-operative short-term stability of various cementless implant designs.

Bidirectional friction study of cancellous bone-porous coated metal interface

Bidirectional friction tests between cancellous bone cubes and a porous-coated metal plate were performed to determine the mechanical properties of the interface required in 3-dimensional (3-D) finite element model studies of cementless implants. Bone specimens were obtained from different proximal regions of four resurfaced cadaveric tibiae. A beaded porous-surfaced plate similar to those used in implants was used. Tangential loads in perpendicular directions with different magnitudes were applied at the interface in the presence of constant normal pressure, and the displacements were monitored in the same directions. Measured results showed that the interface load-displacement curve is highly nonlinear with significant coupling between two perpendicular directions. The interface friction coefficient (defined as the ratio of the maximum resultant tangential force divided by the normal load) was found to remain nearly unchanged with the relative magnitude of tangential stresses and the bone location. Moreover, bidirectional tests suggested that the load-displacement relation when evaluated for resultant values is similar to that obtained in a unidirectional testing condition. Constitutive equations that account for the cross-stiffness coupling terms between perpendicular directions were also developed. These relations were used in a 3-D finite element model study of preceding bidirectional friction tests. The influence of the coupling terms on results was investigated by comparison of predictions with measurement results. A satisfactory agreement was found between the results of the experiments with those of finite element studies confirming the constitutive relations as well as the importance of coupling terms.

Relative motion at the bone-prosthesis interface

Bone ingrowth in porous surfaces of human joint implants is a desired condition for long-term fixation in patients who are physically active (such as in sport or work). It is generally recognized that little actual bone ingrowth occurs. The best clinical results report between 10 and 20% of the total prosthetic surface in contact with bone will feature good bone ingrowth. One inhibiting factor is the relative motion of the bone with respect to the implant during load-bearing. This study investigated mathematically the interface micromotion (transverse reversible relative motion) between a flat metal tibial prosthetic surface of a prototype implant, and the bone at the resection site. The aim was to assess the effect of perimeter fixation versus midcondylar pin fixation and the effect of plate thickness and plate stiffness. Results showed that in the prototype design the largest reversible relative bone motion occurred at the tibial eminence. By design, the skirt fixation at the perimeter would prevent bone motion. A PCA (Howmedica Inc.) prosthesis has been widely used clinically and was chosen for a control because its fixation by two pegs beneath the condyles is a common variation on the general design of a relatively thick and stiff metal tibial support tray with pegs in each condylar area. The PCA tibial prosthesis showed the largest bone motion at the perimeter along the midcondylar mediolateral line, while being zero at the pegs. Maximum relative bone motion for the prototype was 37 μm and for the control was 101 μm. Averaged values showed the prototype to have 38% of the relative reversible bone motion of the control (PCA).

Interface fixation analysis of artificial joints

Implant/cement interface stresses, which affect the quality of implant fixation to the surrounding materials and therefore the long-term performance of the total joint replacement, are addressed. It is noted that previous modeling approaches result in stresses that are discontinuous at the interface between adjacent elements. Three distinct formulations that yield accurate continuous stresses as well as displacements are presented. These are a displacement-based interface element with an extra node on the interface, a penalty-modified compatible formulation, and a mixed stress-displacement formulation. Results for a tibial fixation model subjected to a nonaxisymmetric compression load are presented. The results are predicted using the penalty-based continuous stress formulation. The model has been analyzed for two types of loading conditions. One condition consists of two loaded areas, simulating a situation when the femur is exerting pressure evenly on the two tibial condyles. The second condition consists of a single loaded area simulating the loading of only one condyle. Results for the latter case are reported. The proposed formulations have performed satisfactorily in the test examples and are, therefore, considered as reliable tools to predict the interface mechanics in the total joint replacement.