This systematic review includes the following key points: (1) hip and knee implant materials and their debris formation mechanism; (2) debris from different hip and knee artificial joints; (3) particle isolation methods; (4) quantitative analysis of wear debris; (5) morphology of particles retrieved from hip and knee joints; (6) human periprosthetic cells and mediators and (7) in vitro inflammatory response to foreign particles.

Articles or the part of articles focused on any of the following criteria were considered beyond the scope of this review: (1) biological responses that are only limited to animal cells (murine/rats); (2) implant debris from shoulder, ankle, spinal joints; (3) prediction by numerical or computational analysis and (4) modeling of prosthetic joints and lubrication characteristics (except from those related to wear debris).

We also identified four partially-relevant review publications in this area, particularly on metal wear debris [13], particle isolation methods [14], and the biological response of orthopedic wear debris [15,16].

UHMWPE was first introduced as an implant material in the early 1960s by Sir John Charnley [17] when he developed the concept of low-friction arthroplasty. He was probably the first to identify polyethylene and cement debris in infected reconstructed joints [18]. Nevertheless, the initial success of UHMWPE as the cup material [19] has prevailed for 30 years, UHMWPE being the dominant orthopaedic material in total joint replacements (TJRs) [20]. Subsequently, macrophage and giant cells showed adverse response to the particles of polyethylene together with metal and acrylic cement debris [21]. Recently, the historic UHMWPE was replaced by the newer cross-linked polyethylene [22,23], which possesses superior mechanical properties with developed wear resistant characteristics [24,25]. Heiner et al. [26] investigated third-body scratches on both conventional UHMWPE and highly cross-linked polyethylene. They concluded that there was no significant difference between the two materials with respect to protection against severe scratching induced by large embedded third-body particles.

Two distinct wear mechanisms of UHMWPE, based on the scale of intimate asperity interactions, were reported by Wang et al. [27] that are operational in both total hip and total knee replacements. They revealed that the wear rate is strongly affected by the ultimate tensile strength and breaking elongation of the UHMWPE material. Particle detachment from bearing surfaces can be induced mechanically (repeated cyclic stress leads to fatigue) or chemically (changes microstructure in contacting surface) [28]. However, pitting and delamination were identified as the most common form of knee wear that can produce wear debris of a much larger scale [29,30].

UHMWPE with ceramic or metallic counter face causes stretching and reorientation on the crystalline and amorphous polymer phases. Often, a transfer of a thin film of UHMWPE on ceramic or metal counter face can result in lumpy shaped wear particles or granules, splinters, and flakes [31]. Adhesion between the liner and metal counter face generates fibrils on the surface that are later torn off by mechanical action, resulting in loose micro wear particles [32].

Surface roughness of implant surfaces were found to increase the propensity of wear and were associated with increased loosening rates [33]. Lately, in vitro and in vivo wear debris morphology was compared with associated wear mechanism for the same friction pair of UHMWPE and CoCr alloy [34]. Different shape and sizes of UHMWPE were defined as the consequences of different wear mechanisms. The larger particles are the outcome of adhesive wear, whereas the smaller particles are usually formed by the fragmentation of large wear debris or the exfoliation of surface micro-convex-bodies of friction pairs. Flat block shape and sheet/flake wear debris are found to be the results of adhesive and fatigue wear, respectively, whereas tearing wear debris (most irregular) is found to be the product of composite motion of friction pairs. Multi-directional motion imposes a higher wear rate of UHMWPE than reciprocating linear motion [35]. The crosslinking of UHMWPE reduces the degree of molecular orientation during sliding [36] and shows better wear resistance compared to conventional UHMWPE [37].

The first MoM hip prosthesis components were originally made of stainless steel [38], which was replaced by CoCr alloy to mitigate the excessive friction of the original sliding pair [39]. The second generation MoM THRs was introduced in the early 1990s to reduce polyethylene wear and to resist the rapid initiation of osteolysis [40]. Uses of CrCo alloy in MoM pair were shown to exhibit much less linear wear than MoP [41]. Even CoCr alloys were found to have less damage on UHMWPE than Ti-6Al-4V alloys [19,42] in MoP coupling. A study on MoP bearing with different metal couplings against polyethylene counterpart demonstrated different kinds of metal release rates. The linear wear rate of CoCr alloy was about 0.1 micron per year (10⁶ cycles), whereas the wear rate of 316L stainless steel and Ti-6Al-4V were in the order of 0.2 microns and 1 micron per year (10⁶ cycles), respectively [43].

Understanding the tribological mechanisms of metal components in TJRs is always important to improve the mechanical properties of sliding pairs. It is reported that the tribo-material formed in a nano-crystalline structure (having a thickness of less than 300 nm) when MoM hip joints articulate under ultra-mild sliding wear conditions incorporated with corrosion and fretting [44]. These tribo-materials have different chemical and mechanical properties than the bulk materials [45,46]. In addition, changes of surface wettability, oxidative wear of metal surfaces, micro-abrasion of metal surfaces from oxide film damage, and surface abrasion from third-body bone/PMMA debris affect wear rate and metal ion release from the metal surfaces in TJRs [43]. Recently, Wimmer et al. [47] reported that the nano-crystalline tribolayers of MoM components incorporate organic material stemming from the synovial fluid, termed as “mechanical mixing”. This mechanical mixing changes the bearing surface of the uppermost 50 to 200 nm from pure metallic to an organic composite material. It hinders direct metal contact (thus preventing adhesion) and limits wear. This finding of a mechanically mixed zone and organic constituents provides basic understanding of particle release from MoM arthroplasty.

In addition to material properties, geometry plays an important role in the tribology of MoM hip joints. For example, Leslie et al. [48] concluded that larger diameter MoM hip joints have lower wear rate compared to smaller diameter hip joints after a certain period of rubbing. Even, size of cobalt level was found to be higher in the smaller diameter hip joints after half-a-million cycles. Similarly, clearance was found be an influencing factor in MoM hip joints—a mean diametrical clearance of 94 μm had significantly lower friction and wear rate, followed by 53 and 150 μm diametrical clearances [49]. However, recent report showed that the number of complaints against the larger diameter hip joints is increasing in the UK [50], which indicates that the outcomes of in vitro tests do not always match those in vivo.

Orthopedic surgery employed ceramics for the first time in artificial hip and knee replacements in the early 1970s [51,52]. Recent trends indicate that CoC implants are likely to replace MoP because of their reduced risk of osteolysis, chemical inertness, and resistance to corrosion, low wear rates and non-allergic properties [53–55]. First generation ceramics implant used mechanically weaker Alumina (Al₂O₃) [56] and comparatively strengthened Zirconia (ZrO₂) [57,58]; however, ceramic composites [59,60] are being studied intensively to improve their mechanical performances in TJRs by reducing their brittleness and slow crack growth [61] that led to joint failures [62] associated with variably-described sounds, namely, squeaking, pop, and click [63,64].

Grain pull-out is reported to damage the ceramic-bearing sliding surface which leads to higher surface roughness and increased friction in this area [65]. Macroscopic stripe wear is another common form of ceramic wear caused by edge loading. Grains are fractured out of the surface when the stripe wear appears to have resulted from the direct contact of the femoral head with the acetabular shell [62]. This contact is also the probable cause for the sudden onset of squeaking in the previously ‘silent’ hip articulation. Multiple smaller fragments are generated and, hence, the surface roughness increases, leading to a higher wear rate [66,67]. Grain pull-out occurred in ceramic prostheses, despite their better surface wettability properties than the conventional MoP bearing materials [68]. Squeaking of ceramic materials is found to influence wear mechanism of CoC hip joints. Currier et al. [69] found that the ceramic ball-in-socket bearing couple alone, without any metal devices incorporated, can be made to vibrate at an audible frequency when articulated. Consideration of the geometry of current generation CoC hip bearings led to a hypothesis of a rolling/sliding mechanism causing vibration and squeaking. In addition, a new mechanism of failure of a CoC THRs is reported by Bonnaig et al. [70], due to fretting corrosion and failure of the Morse taper. Failure of the Morse taper led to metal debris, which rubbed with the ceramic and caused dramatic third-body wear. The malfunction of the Morse taper, as reported in this case, represents a possible failure mechanism of a CoC THR.

Hard coating on bearing surfaces is another option to fabricate a mechanically superior and highly biocompatible surface for implanted bearing [71]. Diamond-like-carbon (DLC) is a coating material with good wear resistance and chemical inertness properties [72]; such ideal materials were proposed for protecting implants more than a decade ago [73]. In addition, the improved biocompatibility and reduced ion release with better wear-resistant properties of titanium and chromium nitride coatings [74–76], tantalum-based multilayer coating [77,78], carbon ion implantation (CII) coating [79], and amorphous diamond coatings [80], on conventional metal bearing have been investigated. Other surface engineering techniques are found to be effective to reduce friction and wear properties in local contact area of sliding pair—patterning concave dimples on polyethylene [81]; wavy, square grid and simple dimpling on metal and ceramics bearing surface [82], and modeling circular pattern [83] are found to be significantly effective to improve boundary lubrication and wear resistance of bearing surfaces.

The wear mechanism of such hard coatings have been studied and revealed different results. Sliding-induced heat accumulating on local contact areas of DLC can possibly cause a gradual destabilization of the carbon-hydrogen bond in the sp³ tetrahedral structure of DLC [84]. The movement of hydrogen atoms can thus trigger the transformation of the sp³ structure in to a graphite-like sp² structure. Such graphitization of DLC is promoted by thermal and strain effects under higher load. The repeated cyclic wear then damages the secondary film formed on DLC [85]. Tribo-oxidation is discussed as another mechanism of such hard coatings under different tribological environment [86].

The collected and preserved (freeze-dried) biopsy samples were harvested into small segments and were washed with chloroform/methanol for lipid or lipid group removal. The extracted tissue samples were then dried. This pre-stage of tissue digestion (delipidation) is reported in several studies [91,92].

The digestion of tissues and sera free the particles from a sticky host. The chemical methods, namely, alkaline and acidic digestion, as well as enzymatic digestion, have been employed for the last 15 years to digest organic tissues and sera to isolate metal, ceramic, and polyethylene particles. The application of the different available digestion processes for different materials are shown in Table 1.

The suitable approach of debris isolation from periprosthetic tissue or simulated body fluids depends on the material and medium of the debris. In fact, all three approaches (alkaline, acidic, and enzymatic digestion) can be applied to ceramics (metal oxides or carbides), which are chemically inert. The acidic protocol remains popular [88,107] for isolating ceramic particles from periprosthetic tissues. However, the detrimental effects of aggressive alkaline solution on CrCo alloy particles were reported [93,94] as metals are prone to being ionized and oxidized. On the other hand, the enzymatic protocol allows the superior isolation and characterization of metal particles without affecting the shape and size of particles [92–94,98]. Comparisons of these three approaches were individually studied [89,91,100] and the relevant discrepancies were reported. Niedzwiecki et al. [89] reported that the enzyme method generated the least amount of hazardous waste compared to chemical (alkaline and acidic) protocols; thus, an optimized enzyme method was suggested as a practical standard for debris isolation and analysis.

Slouf et al. [91] found that the acid method was the most convenient, given the time needed for isolation, the cost of chemicals, and the final purity of the isolated particles. Baxter et al. [100] showed that 5 M NaOH, 5 M KOH, and 15.8 M HNO₃ enabled the most complete digestion of human hip tissues and highlighted the enzymatic protocol for perfect digestion.

The sample was aspirated, heated, and then diluted by chloroform and methanol before centrifugation to separate the remaining contaminating proteins and lipids after digestion. In fact, the three digesting methods more or less applied centrifugation to separate the particles from the digested tissue solution [92–96]. Such centrifugation process enables separation of different particles, based on their density level. A method was developed to avoid centrifugation, based on the digestion of paraffin-embedded tissue samples, because of the possibility of morphological changes in the particles during centrifugation [101]. However, centrifugation speeds up to 105,000× g were later found to have no effect on the morphology and quantitative image analysis parameters, such as equivalent diameter, circularity, and elongation [102].

Particles with excessive contamination were made agglomeration-free and were uniformly dispersed into the solution through ultrasonic action [88,92]. The dispersion solution was subjected to vacuum filtration [90,91] at different nanometer to micrometer pore sizes after the confirmation of quality. The use of different filtration sizes limited the particle sizes in the same cohort. The filter paper with particles and the solution with particles of limited sizes were then dried.

A few articles were identified on histological analysis of particle characterization, which does not require the isolation protocol. Solis-Arrieta et al. [108] determined the composition of the debris materials, using energy dispersive X-Ray analysis (EDX), following the conventional histological technique. Laser capture micro-dissection into periprosthetic tissue [103] and transmission electron microscopy (TEM) [104] were also employed to characterize the intercellular particles.

The particles isolated from the simulators and periprosthetic tissues appeared to be predominantly submicron in size [109] and had both regular and irregular shapes, as shown in Figures 3 and 4. The sizes and shapes of these particles were found to vary between in vivo and in vitro analysis. Nevertheless, the evaluation of in vitro tribological studies is justified as they reproduce in vivo results.

However, Catelas et al. [111] concluded with partial uncertainty that CoCr particles retrieved from MoM joint simulator were very similar in composition, length and shape to the particles retrieved from MoM joint of patients. The common shapes of the particles retrieved from joint prosthetics were found spherical, flake, and fibril (Figure 4), whereas the joint simulator generated cylindrical, radial broken, block, fibril/twig, spherical sheet, and flake [34,110], as shown in Figure 3. Hongtao et al. [34] reported the in vivo and in vitro difference of particle sizes from UHMWPE and CoCr alloy friction pairs. They found that UHMWPE particles from joint simulator were larger in size (average diameter of 6.89 μm) than the particles isolated from the periprosthetic tissues (average diameter of 1.33 μm, which is about18% the size of the debris from the joint simulator). Buscher et al. [44] found that the majority of the CoCr wear particles in vitro were globular with a diameter <100 nm, whereas the mean diameter of the in vivo particles were <80 nm and had a minority of particles that were needle-shaped in both of the cases identified by a scanning electron microscope (SEM).

The differences in wear mechanisms and wear outcomes between hip and knee should be attributed to the difference in loading and sliding configurations with different degree of freedoms influencing debris morphology. Knee prostheses were found to produce larger UHMWPE particles with the mode of particle size in the 0.1–1.0 μm size range, compared to <0.1 μm size range for hip prostheses [96]; however, there was no significant difference in wear rate between these two joints. In addition, Benz et al. [104] reported that more than 75% of the UHMWPE particles retrieved from the hip joint had a length <0.5 μm, but only 43% of the UHMWPE particles from the knee joints were <0.5 μm in length. Similar results were found by Mabrey et al. [99] who reported that the particles from the hip joint had an equivalent circular diameter (ECD) of 0.694 ± 0.005 μm, which is relatively smaller than those retrieved from the knee joints (ECD of 1.190 ± 0.009 μm). Furthermore, the debris sizes were found to be influenced by the bearing type and bearing size. Some investigators suggested that mobile bearings were [109,113] advantageous over fixed bearings, based on their wear behavior and improved kinematics. However, no significant difference was found in wear rate and debris size between the mobile and fixed bearings of knee prostheses, using knee simulators [114,115]. Therefore, the previous suggestion was rejected. Leslie et al. [48] reported that debris size, wear rate, and ion levels were not influenced by bearing sizes. They conducted an in vitro investigation on 39- and 55-mm diameter MoM bearings. The investigation showed no significant differences in mean particle size (ranging from 8 nm to 108 nm and having round/globular shape) derived from both bearings. No needle-shaped particles were observed. The ion levels measured suggested both bearing sizes had similar initial wear rate; and the 55-mm diameter bearing reached steady state wear more rapidly than the bearing of 39 mm. However, a previous study on MoM bearings reported that 56-mm bearings produce reduced-sized particles compared with 28 mm bearings [116].

The aforementioned findings from the different studies show different results on debris morphology and wear behavior derived from different sizes and types of weight-bearing joints. However, it is accepted that debris characterization can be a parameter to optimize bearing sizes for different weight-bearing joints.

UHMWPE, metals, and ceramics were found to be the predominantly-studied materials for debris characterization (Tables 2 and 3). The cross-linking of UHMWPE definitely improved wear resistance [23,117], indicating successful material development. Therefore, highly cross-linked UHMWPE was found to produce >90% fewer wear particles in large size ranges and smaller-sized particles than the conventional UHMWPE [24]. However, a counter finding showed [118] no significant difference between cross-linked and non-cross-linked UHMWPE in the percentage number and percentage volume of particles in the size ranges tested in a multi-directional pin on a plate wear simulator.

Most of the studies on UHMWPE debris report material characterization of a large range of sizes (0.1 to 10 μm) [101,104,105,119–121] with irregular-shaped particles of higher aspect ratio [114]. Nano-sized (18.5–21.2 nm) UHMWPE wear debris were recently investigated in vivo for the first time by Lapcikova et al. [92]. These UHMWPE particles were found to have the most irregular shapes compared to those from MoC bearing surfaces, such as, fibril, flake, cylindrical, globular, twig, and, sometimes, spherical shapes, as summarized in Tables 2 and 3.

Metal particles retrieved from MoM implants were found to be smaller in size than polyethylene debris from MoP joints. The hip simulator for CoCr alloys with different carbon contents generated metal particles in the range of 25 nm to 36 nm [122]. A similar outcome was reported in vivo by Brown et al. [93], who indicated that most of the generated debris retrieved from hard-on-hard (MoM and CoM) hip prosthesis were less than 50 nm with round and irregular morphology. Wear debris with sizes ranging from 30 nm to 100 nm were also found in vitro [98,123] with mostly round to oval shapes and some needle shapes. Milosev and Remskar [106] also identified needle-shaped particles, ranging from 40 nm to 120 nm and containing both Co and Cr, isolated from the periprosthetic tissue of the MoM bearing. The globular particles reached 90 nm and contained high levels of Cr and no Co.

Wear debris concentrations from CoC hip joints in vivo were two to 22 times lower than those of MoP and CoP [107]. An earlier study, conducted by Mochida et al. [88], reported that no significant difference in the average size exists among the different types of particles retrieved from either CoC or CoP hip prostheses. The nanometer-sized ceramic wear particles in retrieved tissues were first reported to [103] range from 5 nm to 90 nm in size, measured by TEM. However, studies using SEMs, which have lower resolution than TEMs, revealed ceramic wear particle sizes ranging from 0.05nm to 3.2 mm. The presence of very small alumina wear debris (2 nm to 27.5 nm) was noticed during the micro-separation of the prosthesis components of the CoC joint [56].

The type of lubricants used in the joint simulator influenced the shape and size (length) of the debris. Serum produced smaller and thinner particles in size than the particles produced in water as lubricants for metal [122] and polyethylene [124] materials. Wear particle size was found to remain unchanged with changes of head-cup pair material, despite being considerably affected by wear rate. The change in the head materials in a hip joint simulator did not show any effect on debris size distribution [123]. The influence of head roughness on wear particles was evident and showed an increase in minimum particle size and surface roughness. Atomic force microscopy (AFM), along with SEM and TEM imaging techniques, added a new dimension in the debris characterization. A 3D size and shape characterization of UHMWPE wear debris was recently presented [125], although Scott et al. [109] previously introduced the AFM to improve the estimation of UHMWPE volumetric wear rate in vitro. A MiaoXAM2.5X-50X ultra-precision contourgraph was used to investigate the 3D morphology and thickness of the wear debris [34]. Gladkis et al. [126] subsequently showed the quantification of the size and shape of UHMWPE wear debris in all three spatial dimensions (Figure 5). The investigation clearly defined the length L, width W, and height H measurements. The approach was a sensible compromise between the practical considerations of the AFM technique and the correct determination of the particle dimensions.

Wear debris distribution was not homogeneous throughout the tissues because of clumping and clearing of the debris through drainage. The number of particles collected per unit of wet tissue was highly dependent on the biological variations of the tissue [112]. Therefore, randomizing the harvested tissue samples became a general practice before digestion. In addition debris morphology may largely be influenced by the lack of experimental precision because of different types of quantitative methods.

The common parameters in defining each particle employed by most of the researchers were ECD, roundness (R), form factor (FF), aspect ratio (AR), and elongation factor (E) [92,99,101,112,115,127]. Most of the studies used 2D SEMs and TEMs as the input to obtain quantitative statistics. The American Society for Testing and Materials (ASTM) F1877-98 [113] outlines that fractal dimension was sometimes accounted for the characterization of the morphology, number, size, and size distribution of the particles. Tipper et al. [112] determined the total number of particles, using a mean thickness value, the mean area of the particles, the density of the material, and the mass of the debris on the filter. SEM was later used to develop an automated quantification method (SEMq) with errors less than 10%, as verified with several sets of experiments [91]. The SEMq methods indicated that the distribution of UHWMPE particles around the total joint replacements was non-homogenous. Slouf et al. [128,129] later introduced infrared spectroscopy (IR) and LSC (Light Scattering with Calibration spheres) methods to determine the total volume of the UHMWPE and number of wear debris produced. The results showed good correlation with the radiographic appearance and indicated that extended tissue damage in a particular zone around the total joint that was proportional to the volume of the wear debris in that zone. Another extraction method of analysis is laser diffraction particle analysis [87], which has advantages in retaining the particles in the solution produced by the purification technique that avoids agglomeration and contamination [123]. Three-dimensional imaging approaches for particle quantification were reported in several studies [34,109,125,126]. Figure 5 outlines the 3D quantification of the particle measurement of different shapes and sizes described by Gladkis et al. [126].

From the debris morphology, data were extracted, organized, and interpreted to create a graphical presentation. Normal distribution was commonly reported, using the mean and standard deviation of length and width [105,111].Numerous studies represented the particle area, maximum dimension (length) [102,112], and volume distribution [118,123,129], using particle size and number.

Very recently, the impact of different methodologies was compared by Schröder et al. [130]. They concluded that particle characterization is a complex analytical method with a multiplicity of influencing factors. It becomes apparent that a comparison of results of wear particles among different research groups is challenging.

Chronic inflammatory response, initiated by particulate debris at the implant-bone interface in a wide array of cell types, limited the longevity of joint reconstruction. These cells include macrophages, fibroblasts, giant cells, neutrophils, lymphocytes, and, most importantly, osteoclasts [132] studied in vitro.

Cells, cultured with particles of different materials [133,134] with different sizes [135–137], shapes, and doses [138], secreted different types of functional inflammatory mediators that acted locally at the site of cell damage and infection [136,139,140]. After activation by the wear particles, the phagocytes produced inflammatory mediators/secreted factors such as TNF-α, RANKL, IL-6, PGE₂, and IL-1b, which are implicated in osteoclast activation and bone resorption [140,141]. The expression of bcl-2, bax, and caspase-3 was studied to understand the mechanisms that lead to apoptosis in macrophages. Bcl-2 is considered a death-regulating gene. Bax has a powerful death-promoting ability for cells. Caspase-3 is probably the most correlated with apoptosis among the different proteases [141,142]. Human osteoblast and fibroblast with metal alloy and ions were also studied [133,143–146] to investigate cytotoxicity and genotoxicity. The tests were related to cell viability, proliferation, alkaline phosphatase activity (APL), DNA damage, and chromosome aberrations. The formation of mineral nodules into cells was also identified [145]. Zymography analysis [141] was conducted to reveal the protein expression of cells affected by ions or debris.

Phagocytosis of the particles was found to be correlated with changes in particle morphology. Cell proliferation, differentiation, and prostanoid production were affected by size, shape, and dose of wear particles. In addition, the chemical composition of particles from metal alloy and polyethyleneare found to affect alkaline phosphatase and PGE₂ [133]. Previous studies have suggested that small polyethylene particles (less than 1 μm) can be more easily phagocytized than larger particles, and elongated particles may induce a stronger cellular reaction than round particles [5,6]. However, no statistically significant differences in (in vitro) biologic responses were noted between highly cross-linked and conventional polyethylene debris at low and intermediate doses. Only at the highest dose tested, highly cross-linked polyethylene was significantly more inflammatory than conventional polyethylene, based on relative TNF-α and vascular endothelial growth factor secretion levels [156]. In vivo analysis by Illgen et al. [157] also showed that cross-linking increases the inflammatory response to similar-sized conventional polyethylene debris. Polyethylene particles with mean sizes of 0.21, 0.49, 4.3, 7.2, and 88 μm were co-cultured with cells for 24 h prior to the assessment of the cell viability and production of the osteolytic mediators, such as IL-1b, IL-6, TNF-α, GM-CSF, and PGE₂ [136]. Cell viability was unaffected by UHMWPE particle sizes. Only particle sizes between 0.21 and 0.49 μm produced significantly enhanced cytokine secretion.

The afore-mentioned study on particle characterization (Section 5) implied that MoC implants generate significantly smaller particles than MoP. Clinically-relevant CrCo alloy nanoparticles from MoM joints appeared to disintegrate within the cells faster than micro-particles. These nanoparticles (Figure 6) induced more DNA damage, aneuploidy, and cytotoxicity than micron-sized particles of an equivalent volumetric dose [139,158]. Metal nanoparticles from MoM hip joints [93,98,122,158] vastly increase the total surface area of the metal, which increases the propensity of releasing metal ions in vivo. The variation of cellular damage with different Cr (III) complexes ([Cr(en)₃]³⁺) [146] inhibited cell proliferation of human dermal fibroblasts and causes intracellular damage through the formation of apoptotic bodies and chromatin condensation, all of which indicate cell death. The Co²⁺ and Cr³⁺ ions inhibited bcl-2 expression but stimulated bax and caspase-3 expression [141] at different periods of incubation with the macrophage. The release of soluble ions from CoCr particles was identified as the most likely cause for DNA damage within the first hour [144]. The overall level of DNA damage and structural aberrations caused by the CoCr alloy is approximately the same for both young and older cells. Older cells showed a greater loss of viability, induction of senescence, and a lower rate of mitosis and cell growth than young cells [143].The effect of the micro-sized particles of Ni-free Fe-based alloys resulted in mineralization into osteoblasts after 21 days (Figure 7), where the cells were overloaded with small particles in the cytoplasm [145]. Mineral nodules were observed all over the surface of the multi-layered cells, despite being fewer in number than the unexposed cells. The viability and proliferation of the osteoblast were found substantially unaffected by the presence of the particles of the FeAlCr alloys, which were phagocytized, based on size. The ion release rate from the aluminum and chromium particles in the culture medium increased with higher doses.

Nano-toxicity is now highly linked with osteolysis. Knee prostheses are thought to have lower osteolytic risks compared to the hip prostheses [96] as comparatively smaller particles are found in hip prosthesis. MoM implants are found to reduce the potential for the induction of osteolysis [147] as MoM implants generate comparatively smaller particles. In addition, the small size of the wear particles may facilitate their dispersal via the lymphatic system to sites distant from the implant and it has been reported that cobalt-chrome particles can accumulate in the liver, spleen, lymph nodes, and bone marrow of patients [148,149]. Moreover, titanium nitride (TiN), chromium nitride (CrN), and chromium carbon nitride (CrCN) coatings applied on cobalt–chrome alloy (CoCr) substrate produces nano-size debris less than 40 nm. These wear particles showed reduced cytotoxic effect compared to the CoCr alloy debris cultured with U937 macrophages [75]. A dose-dependent reduction in bone resorption was achieved using human peripheral blood monocytes, cultured with osteoblast-like UMR 106 cells exposed to metal wear particles. This decrease in resorption was greater after exposure to CoCr and 316L-SS particles than to TiAlV and commercially pure Ti particles [135]. However, Sabokbar et al. [153] concluded that osteoclast formation is not significantly induced by particle characteristics (size, shape, and dose). Macrophage involvement in periprosthetic osteolysis also did not depend on particle phagocytosis. Zhang et al. [150] demonstrated that nano-sized ceramic particles were bioactive to cells, despite the significant secretion of inflammatory mediators from cells shown by nanoparticles of other materials (Figure 6). The aluminum nanoparticles significantly promoted the alkaline phosphatase (ALP) activity of the MG63 cells at a low concentration and did not show irritation to the macrophages. However, ALP activity of those treated with Ti microparticles was lower than that treated by ceramic nanoparticles. A larger volume of alumina particles (5 nm to 20 nm and a few >0.2 μm), from the hip joint simulator under micro-separation conditions, was required to activate the human peripheral blood mononuclear than the commercial alumina particle at 0.5 μm [151]. The critical particle size range to stimulate cell response was defined from 0.1 to 1 μm [136,151].

Debris from bone cements (CMW original, CMW1RO and Palacos R, CMW calcium phosphate, CMW copolymer bone cement) with sizes from 0.1 to 0.5 μm [134,140] were studied, and no statistical differences between the levels of bone resorption were induced by these cement types. Cements that contained pure CMW1 and CMW with calcium phosphate failed to induce the macrophages to express bone resorption activity, even at a high debris concentration (100:1 ratio). However, a major cytokine (TNF-α) was produced at the 100:1 ratio [134]. A similar study [140] demonstrated that bone cement particles are capable of inducing increase in TNF-α production in vitro, based on cement particle size, volume and cement particle type. Cement particles that contained radio-opaque additives were the most active. However, Baets et al. [152] demonstrated that metal debris occupied only 1.5% of total volume of wear debris retrieved from a cemented implant in vivo with 56.5% bony fragments and 42% cement fragments. The study prompts a rethink on the contribution of metal debris in bone resorption.

The inflammation and loosening of joint implants, incorporated with wear debris, result in the need for revision surgery. Deep infection results in severe complications and high economic burden. Demand for primary and revision joint replacements is expected to increase exponentially in the next two decades [155]. In addition, adjustable MoM bearings were found to be associated with a higher risk of periprosthetic joint infection when compared with CoC bearings [7]. Revision TJRs are associated with lower success rate, more complicated surgery and higher healthcare costs (by one third) [159] compared to initial TJRs surgery, which may induce additional damage to the surrounding tissues.