Materials and Their Failure Mechanisms in Total Disc Replacement

Adults suffering from lower back pain often find the cause of pain is degenerative disc disease. While non-surgical treatment is preferred, spinal fusion and total disc replacement remain surgical options for the patient. Total disc replacement is an emerging and improving treatment for degenerative discs. This paper provides a review of lumbar disc replacement for treatment of lower back pain. The mechanics and configuration of the natural disc are first discussed, followed by an introduction of treatment methods that attempt to mimic these mechanics. Total disc replacement types, materials, and failure mechanisms are discussed. Failure mechanisms primarily involve biochemical reactions to implant wear, as well as mechanical incompatibility of the device with natural spine motion. Failure mechanisms include: osteolysis, plastic deformation of polymer components, pitting, fretting, and adjacent level facet and disc degeneration.


Introduction
Lower back pain (LBP) is among the most common ailments for adults in the United States.According to a national survey performed by the Center for Disease Control and Prevention (CDC), 28.8% of adults have complaints of lower back pain [1].Many cases of LBP are caused by intervertebral OPEN ACCESS disc (IVD) injury or degenerative disc disease (DDD), or other disc injury.The IVD relies on natural flexibility of its materials, as well as its pseudo-pneumatic structure to transmit loads down the spine even in awkward loading positions [2].In addition to transmitting loads and maintaining structure, the natural IVD behaves as a "hydrodynamic ball bearing" [2] allowing for smooth rotational motion.Natural IVD permits angular movement in the three planes of motion (sagittal, lateral, and transverse) giving it six degrees of freedom (DOF) [2][3][4].It is in this capacity that the IVD is unlike any other joint.The balance of motion and motion resistance allow the body to maneuver in a complex manner while minimizing the stress induced by other parts of the body.Unfortunately, its duty to simultaneously undergo torsion, shear stress, localized compressive stress, and localized tensile stress both causes DDD and creates problems for treatment thereof.
A simple model of the IVD is shown in Figure 1.It is composed of a nucleus pulposus (inner structure) and the annulus fibrosus (outer structure).The nucleus can absorb some impact between adjacent vertebral bodies, thus protecting the rigid vertebral bodies from each other.The pulposus also acts to preserve the disc itself by helping withstand compression [4].Surrounding the nucleus is the lamellar [2,4] annulus fibrosus.The annulus is composed of layers of oriented collagen fibers [4,5].The collagen fibers in the annulus are arranged such that every other layer's fibers have the same orientation.The adjacent layers, however have the opposite orientation [4,[6][7][8].These fibers are generally positioned at a 65° angle from the y axis in Figure 1 [4].One treatment for DDD is spinal fusion.This treatment method alleviates pain, but limits range of motion (ROM) and mobility for the patient.Total disc replacement (TDR) is another treatment method for DDD that aims to preserve motion and limit complications related to spinal fusion, such as adjacent level wear and disc degeneration [8][9][10][11][12].This motion preservation is critical to preventing wear and injury in the area surrounding the treated disc.Through the inhibition of motion fusion creates stress concentrations at adjacent level facets and discs [9,[12][13][14].TDR is an effort to solve these issues and help maintain normal range or motion for patients [11].
Although the preservation of motion through TDR alleviates some problems associated with spinal fusion, TDR introduces different types of problems and failures.TDR issues and problems not seen in other implants vary based on the design of the TDR implant.These issues are primarily derived from the following: hyper-mobility, hypo-mobility, material wear, and wear debris particles.This paper will review failure mechanisms of TDR implants.

Lumbar Total Disc Replacements
To be effective, TDR implants must fulfill three requirements.The implant must have a solid, nondestructive interface with the adjacent vertebral bodies, provide mobility, and resist wear.More is required of lumbar than cervical TDR due to the extra loads it must bear.The lumbar spine supports more weight and encounters moments of greater magnitude than the cervical spine.

Types of Lumbar TDR
Lumbar total disc replacement can be classified according to their configuration as well as materials.The configurations of these devices are dependent upon the type of modules involved in the working disc.To allow maximum range of motion and permit the most freedom, current designs are built around a bearing.Bearings are one-piece (1P), Metal-on-Metal (MoM), or Metal-on-Polymer (MoP).MoM and MoP bearings use a ball and socket design to permit motion in all directions.Table 1 shows the different materials, classifications, and bearing designs for current TDRs.Descriptions in Table 1 also indicate bearing design as mobile, fixed, constrained, or 1P.Except for 1P bearings, the designs involve a ball and socket.A fixed bearing involves no moving parts except the sliding of the socket over the ball, but mobile and constrained bearings permit motion of the ball component.

Ball-and-Socket
When designing the ball and socket TDR, a major source of inspiration is the total hip replacement.Hip replacements use a ball and socket design but also experience higher stress and load concentrations than what is needed for TDR.Material selection, therefore, is often inspired by materials used in hip replacement prostheses.The sliding surfaces for the ball and socket bearings are composed of CoCr-CoCr for MoM designs and CoCr-UHMWPE for MoP designs.CoCr alloys are used because they have been found to produce less wear debris in knee and hip replacements [10].Figure 2 is a drawing of the ball and socket mechanism in these devices to roughly illustrate how they work.

One Piece
The one piece bearing design is a recent development for TDR.The aim of this design is to more adequately mimic the natural disc behavior through the implant.One piece designs reduce the number of surfaces on which wear can occur and they reduce hypermobility of the joint.Additionally, since these are made of a softer elastomer, they also mimic the cushioning provided by the natural intervertebral disc.

Implant Materials
Biomaterials being used in current implant designs are: cobalt-chromium alloy (CoCr), titanium (Ti), diamond-like carbon (DLC) coatings, ultrahigh-molecular-weight-polyethylene (UHMWPE), silicone, and a polyurethane-polycarbonate elastomer (PU-PC) [15,25].Ti is typically used as the material to interface between implant and bone; the other materials listed are used as bearing materials for each design.Material selection is not entirely random; most of the early design ideas for TDR implants are a result of current designs in other total joint replacements.Bearing material selections are typically inspired by hip or knee replacement devices.

Device Degradation
Orthopedic implant failure can often be traced to the degradation of an implant.There are two mechanisms for the degradation of TDR implants: wear and corrosion [27].This is to be expected with articulating bearings in harsh environments.

Metal-on-Polymer
When comparing mobile and fixed MoP devices, one can compare Prodisc-L with CHARITE TDR devices [28].The mobile design in the CHARITE allows the UHMWPE core to shift as the back bends, which reduces stress experienced by posterior facets [28].While this implies that the mobile bearing is a better design due to smaller loadings on facets and adjacent bodies, four degree of freedom (DOF) and 5DOF tests indicate higher UHMWPE wear mass loss in the CHARITE design [29,30].
Removed Mobidisc implants revealed that large UHMWPE particles surrounding the implant that caused an inflammatory response [21].Particles were also found in patients with the Activ-L implant, but they were smaller and still triggered an inflammatory response [21].Wear particles from UHMWPE tend to be spherical and ranging between 0.5 and 10 microns in diameter [31,32].This means that even though UHMWPE is not chemically reactive and spherical particles are typically nonreactive as well, the large particle size would induce inflammatory and osteolytic responses [31][32][33].
In studies that have retrieved UHMWPE cores from used implants show measurable wear on the surface of the bearing [18,34].As in hip replacements, multidirectional scratches and some penetration into the polymer were observed.Surface wear and damage are observed near the center of the core as a result of adhesive/abrasive mechanisms [18,34,35].Some long term retrieval studies of the Prodisc-L also found third body wear and end plate impingement, often coupled with burnishing of the metallic endplates [36,37].Impingement was also demonstrated in both Activ-L, CHARITE, and Mobidisc implant designs [21,38].Retrieval studies also observed wear and damage that was similar to that of hip and knee implants.Central dome regions of the UHMWPE core exhibited microscopic scratches resembling the wear patterns of knee and hip MoP devices [35,36,39].Scratch penetration into the polymer core increases at a decreasing rate with implantation time [34,40].This behavior is similar to wear seen in hip implants.These studies, however, found this damage to be insignificant.The rim of the core was subject to extensive plastic deformation and fracture, which is sometimes observed in total knee replacements [34,41].

Metal-on-Metal
In hip arthroplasty it was noticed that MoP bearing surfaces produced high volume wear rates when compared to MoM devices.UHMWPE joints can wear at a rate over 100 microns per year [33].MoM bearings were introduced to hip replacements to produce less wear debris and reduce particle-driven osteolysis [15,42,43].However, in spite of the decrease in wear volume, the number of wear particles actually increased [44].This is due to particle size differences in the materials.While MoP devices tend to wear at faster rates yielding more volume loss, MoM devices produce a large number of particles which range in size from 10 to 119 nm [45].When compared to the particle size of UHMWPE particles on the order of microns in diameter, this means a substantial increase in the amount of reactive particles and reactive surface area of debris.This was especially damaging when the bearing was made of a reactive cobalt-based alloy, which is a major concern for the application of MoM bearings in TDR [15,44,46].
Although limiting volumetric wear is desirable, that is not all that must be considered.MoM bearings tend to produce more particles, which are much smaller than those produced by MoP bearings [44,46].CoCr is used for these bearings because it has good mechanical properties [47,48] hard and passive in the biological environment [47,[49][50][51].It is not, however, inert.CoCr wear particles are chemically reactive within the body.This, combined with the large surface area of the debris, make the particles susceptible to electrochemical processes, which can lead to corrosion of the material and implant within the body [50,52,53].Unlike the UHMWPE debris mentioned above, MoM bearing wear to produce fine, needle and fiber-shaped particles.This particle shape is chemically more reactive in nature, thus contributing to corrosion, tribocorrosion, and toxic and biological responses, such as metallosis, pseudotumors, biological reactions, osteolysis, and inflammation [15,44,46,[54][55].
MoM implants reduce corrosion and negative biological response through passive oxide films [56,57].Cyclic loading of these implants can lead to fretting which can disrupt oxide films and limit their effectiveness [52,53,[58][59][60].Corrosion in these implants is often a result of this fretting wear, which deteriorates the passive oxide layer.Corrosive wear usually produces particles of cobalt or chromium oxides or metal ions, which can build up or bind to proteins within the body causing severely negative biological response [27,42,43,52,60].This response includes toxic reactions as well as metal build up and bone decay which are absolutely detrimental to patient health.

Polymer-on-Polymer
In addition to TDR, some Nubac, a PEEK-on-PEEK device proposes a nucleus replacement solution to minimize some of the biomechanical incompatibilities of TDR.This type of devices utilizes a PEEK-on-PEEK (PoP) bearing design, which replaces the nucleus pulposus.This design is not usually pursued because nucleus replacements have a tendency to be expelled by the body naturally.The Nubac device, however, showed a low risk of expulsion [22].This device shows promise as, in addition to resisting expulsion, the fatigue resistance and wear resistance of PEEK.Preclinical cadaver and animal tests show that the device does not have detectible wear debris or any negative biological response to implantation [22].There are issues, however, with device rejection and biological reactions.Long-term clinical trials showed an unacceptable rate of device migration with Nubac, which, in preclinical trials, showed little to no risk [61].This migration was severe enough to have the device move into surrounding muscle tissue [61].

Diamond-Like Carbon
One disc design, the Baguera-L (See Table 1), moves to reduce the amount of metal wear debris by utilizing a DLC coating on the Ti endplate.DLC is known for its wear resistance and having low friction [62].These properties make it an ideal candidate for ball and socket bearings in TDR.In spite of its superior wear resistance DLC has issues with long-term adhesion to the substrate.DLC films, however, are prone to failure and delamination.Delamination primarily occurs due to corrosion of the substrate by body fluids that pass through pinholes in the DLC film.Delamination can also be driven by pitting corrosion under pinholes over time.A corrosive electrolyte may form and begin to corrode the substrate to fuel local debonding [62,63].This local debonding can induce delamination through corrosion cracking (CC) and stress corrosion cracking (SCC) [62,64].

One Piece
An early design for a singular piece bearing in a TDR implant is the Acroflex disc design by Acromed Corporation.This design uses a polyolefin rubber to mimic the mechanical behavior of a natural disc.Although the device was tested for biological and biomechanical compatibility prior to clinical trials [25,[65][66][67][68], patients suffered core material tears and failure mechanisms associated with fatigue [25].Recent designs have employed PUPC as a core material which has a longer fatigue life.
Even with the improved elastomeric core, these devices could still fail due to PUPC tears or loss of adhesion between the different materials (PUPC-Ti, Ti-bone, or PUPC-bone).Since these designs are still fairly young, evaluating wear and corrosion resistance, as well as effectiveness, for these next-generation PUPC TDR designs require more long-term clinical testing [3,15,25].

Biological Response
In TDR implants the biological response to wear debris is based on number of particles, particle size, particle shape, and chemical composition of the debris [27].This indicates that biological responses are TDR-design and material dependent.Material-independent responses include inflammation and osteolysis.Osteolysis is often a result of the inflammatory response.
Osteolysis is a mode of degradation which involves the destruction of bone [69].In the case of orthopedic implants and arthroplasty, osteolysis occurs at the interface between bone and implant.Osteolysis in TDR implants is primarily driven by micro-motion of the implant and the body's biological response to wear debris [33].It is easy to picture how motion would cause bone degeneration: just like the periodic maneuvering of a shovel in soft clay will loosen and remove some clay.The osteolytic response to debris particles, however, is not so straight forward.
Bone growth and maintenance are the results of equilibrium responses of bone growth and resorption, which are driven by osteoblasts and osteoclasts, respectively.Debris particles disrupt bone homeostasis [27] through an inflammatory response which in turn stimulates the maturation of osteoclasts increasing bone resorption [27,33].The combination of these wear processes leads to increased wear rates over time as resorption loosens the device, thus, creating more space for osteolysis-causing motion and debris [70].
Osteolysis is a prominent mode of failure for total joint replacements including TDR [71].According to some studies osteolysis is not as prevalent TDR due to it having a smaller range of motion than hip and knee replacements, which would indicate smaller wear tracks and fewer debris particles.These sources reported that wear particles do remain a concern in spinal implants because they indirectly induce osteoclastogenesis which can lead to bone resorption and inflammation [15,59,72,73].This limited wear of TDR, however, is not necessarily the case for MoM implants.A review of spine implant debris by Hallab reveals that though MoP hip implants have a substantially higher wear volumerate than TDR, MoM hip and intervertebral disc replacements have comparable wear rates [27].
Problematic responses to CoCr MoM wear particles include metal hypersensitivity, metallosis, the formation of pseudotumors, and vasculitis [15,44,46,[74][75][76].Metal-specific biological responses are a result of wear particle chemistry and shape.Co and Cr ions are chemically reactive and therefore are pro-inflammatory [27,45].Aside from the chemical nature of CrCo debris, shape is an important factor dictating biological response.As seen historically with asbestos fibers, fiber shaped particles from TDR wear cause greater inflammatory response than round particles through a natural response called "danger signaling" [32,[77][78][79].

Mechanical Response
A problem that separates TDR from other joint replacements lies in the complex duty of the intervertebral disc.It is not merely a joint which connects two moving parts.The intervertebral disc, especially in the lumbar region, serves to resist motion and reduce stress concentrations in the adjacent level tissue [4,9].
(a) Ball-and-Socket Bearing Ball and socket bearings to not completely replicate the functions of the natural IVD.These bearings tend to be axially rigid and are not designed to resist bending or rotational moments, allowing motion to occur in excess of that permitted by the natural disc [9,80].These qualities can lead to changes in range of motion (ROM), segmental lordosis, or facet stressing [3,15,26,[81][82][83].
Among limited number of available reports, the Spine Institute at The Ohio State University (OSU) used a hybrid biomechanical model to analyze lumbar function after implantation of Prodisc-L TDR under different external loading conditions.This model was made after an individual male and uses a flexible multi-body dynamic analysis system [9].The model looked at TDR at all lumbar levels, and examined vertebral body loading, ROM, and ligament and facet joint forces.These measurements were taken when the subject was not exerting at all and when he was asked to lift 9.5 and 19 kg separately.The figures below are drawn based on the OSU study.In the figures, the abbreviations SUP and INF represent the interface at the superior and inferior levels of the implant respectively.Figure 3 [9] shows mean values of data from a virtual model simulation of vertebral loadings from the OSU study at different implantation levels.The original data can be found in reference [9].Before reviewing these data some terminology should be reviewed.Numbers such as L5 and S1 describe specific vertebrae.L5 is the fifth and most inferior lumbar (L) vertebrae.S1 is representative of the sacrum which is inferior to the spine and is immobile.
(b) Vertebral Body Stresses Figure 3 below shows approximate mean value data for different simulated loadings based on disc type and implantation level.The information is organized such that data for a given implantation level are reported in the same color.The study found that the substantial differences in performance between TDR and the natural disc lied closer to L5/S1 level.Therefore, the information below focuses on that region.After examining Figure 3 (a) it can be noted that in general at these lower lumbar levels (L4/L5 to L5/S1) that compressive loads apply on the interfacing vertebral bodies.These loading changes are primarily due to high stiffness of the implant and discontinuity at the bone-implant interface.Similar trends of excess loadings due to TDR can be noted for vertebral body shear loadings as well.In the case of AP shear loading, seen in Figure 3 (c) the direction and magnitude of the shear are dependent upon the shape of the spine.Since the resistance to motion is less with TDR, less shear loading is exhibited with TDR than an intact disc in regions of spinal curvature.This, however would indicate higher loading of the nearby ligaments.Figure 4 shows mean adjacent facet loads for TDR at each level according to the OSU study.Once again, facet loading most substantially increases with TDR implantation at L5/S1 level.Additionally, the OSU study shows increased ligament loading and ROM all directions specifically at L4/L5 and L5/S1 levels of implantation [9].The most drastic ROM difference occurred in the transverse plane while the motion was twisting about the spine's axis [9].
Cadaveric study performed has also examined lumbar implantation.Similarly, those concluded that lateral bending and axial rotation result in the most significant increases in facet loading [38,84,85].
This increase in facet loading and ROM can cause stress fractures of the pars and bilateral pedicle, as well as adjacent level disc degeneration [25,[86][87][88][89].This indicates much need for improvement of TDR implants in the mechanical regime.The articulating ball and socket joints, while they soothe pain, tend to be a source of mechanical issues and injury risk for patients.Two possible solutions for the mechanical incompatibilities as shown above are: 1P bearings and partial replacements like Nubac and the extreme lateral TDR (XL-TDR) by NuVasive.The XL-TDR is designed to limit the amount of natural tissue removed.This allows the body to retain some of its natural cushioning and resistance to motion.Figure 5a shows how the lateral insertion can eliminate some need for tissue removal.Figure 5b is an illustration of the nucleus replacement and how it helps preserve natural motion.
The XL-TDR design improves on other MoM implants by preventing hypermobility, providing some natural biomechanical support, and provides a unique look at modern TDR implant designs [90].

One Piece
In order to correct problems seen with ball and socket TDR implants, the one piece bearing employs an elastomer to mimic the body's natural motion [3,15,83].The current implant designs using the PUPC bearing passed fatigue and endurance bending tests prior to introduction into clinical trials [25].These cores were found to resist tearing and wear and were found to be more biomechanically compatible with natural motions [25].As mentioned before, these devices must be studied further in more long term trials to effectively understand the body's response.
Table 2 below gives a brief summary of failure mechanisms based on implant materials and design.[18] Increased ROM Facet and ligament loading [16] No shock absorption Compressive stresses on vertebral bodies [9] PEEK Prosthesis migration Biomechanical incompatibility, stress on remaining annulus, total ejection of device [61] Endplate reaction Severe biological rejection [61] 1P PUPC More studies necessary [25] 5. Future Prospects TDR looks to be a suitable treatment for lumbar DDD.More specifically, studies show that the elastomeric 1P bearing TDR are the future of DDD treatment [94,95].These 1P bearings provide a solution to common problems seen with ball and socket TDR implants [20].A hurdle for these designs is finding suitable core materials for biocompatibility and adhesion to the vertebral bodies [25].This problem seems to have been solved by the use of the PUPC cores in the one piece (1P) bearing implant design, but more research still needs to be done for these designs [15].Although the XL-TDR provides a solution to some of the problems with the traditional ball and socket design, they still involve articulating metal surfaces which can be the site of wear and corrosion.For this reason, research indicates that the 1P bearing would be the ideal design for the future of TDR.Perhaps even a 1P design placed laterally could further improve the TDR implant design.

Conclusions
Total disc replacement (TDR) are an effective solution to degenerative disc disease (DDD), but currents designs still require improvement to be better substitutes for a healthy disc.Ball and socket bearing designs, or any implant involving articulating surfaces increase the risk of failure due to an introduction of more wear and corrosion surfaces.These articulating bearings also do not adequately resist motion and provide cushioning or stability for the lumbar spine (with the exception of the XL-TDR).This would imply that the future of TDR employs a flexible elastomer bearing which can better mimic a natural intervertebral disc (IVD).While this seems like a natural solution, this design type must be looked into further and studied more closely before any conclusions can be drawn.

Figure 1 .
Figure 1.Drawing of the structure of an intervertebral disk.

Figure 2 .
Figure 2. Illustration of working mechanisms of ball-socket devices.
(c) ROM; Facet and Ligament StressesA study was conducted by Burger et al. to examine lumbar facet forces on cadavers after Prodisc-L TDR[84].A facet is another joint at which adjacent vertebrae interface.The purpose of the facet joint is to aid in the inhibition of rotation and excess motion in order to keep vertebrae aligned.Each level has 2 facet joints, one on the left and one on the right side of the spinous process.

Figure 4 .
Figure 4. Selected data from simulated facet load comparison of natural disc to TDR implantation based on implantation level and exterior loading [9].

Table 1 .
Summary of current total disc replacement (TDR) classification, materials, bearing type and regulatory status.

Table 2 .
Summary of common problems of different implant materials and their effects leading to device failure.