Investigation of how natural joints functioned seemed closely dependent on the lubrication mechanisms involved. This was equally relevant to artificial joints where, if fluid-film lubrication could be generated, the rubbing surfaces would be completely separated by the liquid film which would have the advantage of reducing friction, since this depends only on the shearing of the lubricant film, and reducing wear since the two surfaces would not be touching.

In 1969 when I first entered this research area, hip joints were mainly small diameter (<32mm) made from ultra high molecular polyethylene (UHMWPE) rubbing against stainless steel or cobalt chromium molybdenum alloy (CoCrMo), metal-on-metal joints and alumina-on-alumina joints. A few calculations showed that the small diameter metal-on-metal hips and the UHMWPE acetabular components in combination with any type of head material were unable to produce fluid-film lubrication. Insufficient film-thicknesses could be generated to separate the rough surfaces of the joints so surface-to-surface contact prevented full separation.

Ceramic-on-ceramic was different. This could be polished very smoothly and was hydrophilic so it could draw the water based lubricants (synovial fluid), into the contact region, which in turn generated fluid-films. This meant that with alumina-on alumina, wear was not a problem-but fracture was in some circumstances.

As more was learned about lubrication, large diameter CoCrMo hip resurfacing devices became possible. Whilst small diameter metal-on-metal hip joints were unable to generate fluid-film lubrication, larger diameter hips could, provided the lubricant contained serum (similar to synovial fluid). This was interesting since water based lubricants of similar rheology to synovial fluid (carboxy methyl cellulose, CMC) could not produce fluid-films (Figure 1) even though theory suggested that they should. Thus it was assumed that the proteins present in the natural lubricant were important, but the reason was unknown.

Returning to the original assumptions of Osborne Reynolds in deriving the theory of hydrodynamic lubrication, we see that in order to draw fluid into the contact area, the fluid adjacent to the solid boundary was assumed to be travelling at the same speed as the boundary itself. To do this the lubricant must ‘wet’ the surface or attach to it- but what if the surfaces are hydrophobic? The speed of drawing the fluid into the contact will be lower than the surface speed and so less fluid will be drawn in and the pressure generated will be lower than predicted by theory. So a simple experiment was tried using a large diameter CoCrMo alloy hip resurfacing device where a water-based lubricant (CMC) first had bovine serum added, then a simple detergent to reduce the surface tension. Figure 1 shows clearly that the lubrication improves markedly with the detergent – even more so than the bovine serum. This suggests that the wettability of the surfaces is important.

Another approach to enhancing fluid-film lubrication stems from the concept of elasto-hydrodynamic theory. Here, lower modulus, more compliant surfaces, produce thicker fluid-films for similar entraining velocities and applied loads. Thus we developed compliant hip and knee joints using hydrophilic poly carbonate urethane (PCU) acetabulae against metals or ceramics. These produced phenomenally low coefficients of friction (circa 0.001) and in the knee, wear rates of only 0.06 mm³/million cycles (two orders of magnitude lower than metal on UHMWPE).

Another interesting biomaterial is carbon-fibre reinforced poly-ether-ether-ketone (CFR-PEEK). Very long term hip simulator wear experiments (25 million cycles), showed wear rates which were lower than cross-linked polyethylene (circa 1.5 mm³/million cycles), yet friction was very high (µ=0.2–0.3). Clearly this was not fluid-film lubrication but improvements are being investigated.

Introduction

Recent concerns over adverse effects of metal ion release, have led to the development of alternative hip joint replacements. This study reports the performance of new hemispherical MOTIS® (milled pitch-carbon fibre reinforced polyetheretherketone) acetabular cups articulating against Biolox Delta® femoral heads with the aim of producing lower wear and more biologically compatible bearings.

With the boom in metal-on-metal hip resurfacing and hard-on-hard total hip replacements (THRs) with extremely low wear, accurate tribological measurements become difficult. Characterizing THR friction can help in this, especially if the progress of such friction can be tracked during wear tests. Friction measurement can also be used as a tool to study the effects of acetabular-liner deformation during insertion, and possible femoral head “clamping”. This study presents estimates of friction during extended wear testing on THRs of the same size but with different material combinations, using a technique (previously introduced) based on equilibrium of forces and moments measured in the simulator.

All tests were based on five million cycles (Mc) and samples of size-44mm (head diameter). Samples included 6 metal-on-UHMWPE (MOP) (3 with conventional UHMWPE and 3 with highly-cross-linked (HXL) UHWMPE liners), 6 metal-on-metal (MOM) (3 TiN-coated and 3 uncoated), 6 MOM resurfacing (3 standard and 3 with small pockets for lubrication transport), and 3 ceramic-on-UHMWPE (COP) THRs (MOM resurfacing and COPs for 2Mc only). All were lubricated with diluted bovine serum with 20g/l protein concentration at 37°C, and subjected to the loading and rotations of the walking cycle in ISO-14242-1 on a twelve-station hip simulator (AMTI, Boston).

The conventional and HXL MOPs had steady friction factors of 0.045±0.009 and 0.046±0.003 over 5Mc, explained by the stability of wear rates of both these MOP types (72.0±2.81mg/Mc and 14.2±3.57mg/Mc, respectively). However, during the “bedding-in” period (first 0.5Mc), the conventional MOP friction factor rose from 0.047±0.004 to 0.057±0.004 while high wear was occurring (147.1±10.08mg/Mc). The TiN-coated and uncoated MOMs displayed initial friction factors of 0.124±0.117 and 0.039±0.003 respectively. The high standard deviation for the coated THRs was due to coating removal on one specimen which caused scratches and scuffs on its articulating surfaces. This specimen had a friction factor of 0.260 at 0.028Mc. By 1Mc, the TiN coating wore away on the other two coated specimens (friction factors at 1Mc: coated 0.081±0.036, uncoated 0.050±0.014). Over the 5Mc test, average friction factors for the coated and uncoated THRs were 0.097±0.020 and 0.049±0.014 respectively. The 44mm standard and “pocketed” MOM resurfacing THRs displayed initial friction factors of 0.038±0.009 and 0.059±0.026 respectively that increased to the same level at 2Mc (0.094±0.020 and 0.094±0.029, respectively). No difference in wear was detected between the two resurfacing head types (wear rates over 2Mc: standard 3.32±0.25mg/ Mc, pocketed 2.22±1.76mg/Mc), but curiously, both types exhibited an equal level of scratching and scuffing on their articular surface. Finally, the three COP THRs exhibited high liner wear over 2Mc (97.44±3.08mg/Mc), which slowed after the “bedding-in” period. The friction factor also decreased from 0.091±0.005 to 0.070±0.008 over the same period as the UHMWPE liner conformed to the ceramic head.

The method utilized here facilitates on-line sampling throughout the progress of a prolonged wear test, and therefore allows predictions on THR performance/wear to be made. When high friction factors were observed, a high wear rate was occurring and measured on the THR specimens, or damage to articulating surfaces was seen.

Artificial hip joints have been in use for a number of years; various combinations of metals and polymers have been tested both in vitro and in vivo. Modern ceramics have found application as bearings in hip replacement due to the enhanced wear and friction that they offer. It has been hypothesised that during the swing phase of gait it is possible for the Femoral head and the Acetabular cup to dislocate, before relocating during heel contact.

Severe loading such as this could cause greater levels of wear to occur in artificial hip joints.

This study provides comparative analysis between ceramic-on-ceramic hip joint pairings under both severe and standard loading profiles.

Five zirconia-toughened alumina (ZTA) 28mm diameter bearing pairs were tested on a ProSim Hip Simulator for 5.3 million cycles (MC), two under severe loading and three under standard loading conditions. Additionally a Loaded Soak Control, Soak Control and Environmental Control were used. Wear was recorded every 0.5 MC by gravimetric measurement. Surface microscopy images from a Zygo New View 100 and an Atomic Force Microscope (AFM) were taken before testing and then at, 0.5 MC. 2.5 MC, and 5.3 MC.

The standard loading profile followed ISO14242-1 standard with 2650±50N maximum force, ±10° internal-external rotation and −15–30° flexion-extension. To simulate aggressive wear condition, microseparation inferiorly and micro-lateralisation laterally were applied during the swing phase. Dual acting cylinders were used to apply a constant force of 350±50N in opposition to the standard loading profile to enable separation between the Femoral Head and the Acetabular Cup during the swing phase. This microseparation was measured by means of a Linear Variable Differential Transformer (LVDT) and the setting gave a reading of 1.2mm ± 0.1mm at the start of each 0.5 million run. The value for microlateralisation was 0.9mm whilst the inferior separation was 1.2mm.

Wear rates for the ceramic cups under severe wear condition were found to be 0.0356±0.0059mm3/ MC and for the standard wear condition to be 0.0178±0.0049mm3/MC.

The femoral heads had wear rates of 0.0164±0.0046mm3/MC for severe wear condition and no wear was detected for the standard wear condition.

The results of the present study showed almost no wear under standard gait condition and only a modest increase in wear occurred when using severe wear condition. Thus the resulting wear rates are still significantly lower than those found for alumina-alumina total hip joints [1, 2].

Particulate debris created during a fiber-filled PEEK material (MOTISTM) rubbing on a ceramic femoral head in a hip wear simulation study was characterized. The particles were cleaved from the protein lubricant with a double enzymatic protocol and then sized using two different techniques. The sizes obtained were verified using an AFM imaging technique.

Many metal-on-UHMWPE joints ultimately fail due to late aseptic loosening. This occurs due to the particulate debris built up in the periprosthetic area. The body’s natural immunological response leads to bone resorption, the prosthesis becomes loose and severe pain can then necessitate revision. It is therefore important to characterize the wear particles of novel materials in order to understand their biological impact.

Particles were generated in a Durham hip wear simulator from a MOTISTM acetabular cup articulating against a ceramic femoral head for 25 million cycles1. The samples were generated in 500 ml of bovine serum lubricant (17 g/l protein) and a 10 ml sample of this lubricant was analyzed.

A double enzymatic protein cleavage protocol was used as it was shown to be the least harmful to the particles.

A bi-modal distribution of sizes was seen with a large number of particles of 100 nm and a large number at the two micron size range. AFM results verified the size of the particle distribution and also showed that the smaller particles were round to oval and the larger particles were long and thin. No carbon fibers were evident in the AFM images. Although the wear rate over the 25 million cycles1 remained low and linear, the average particle size tended to increase over the 25 million cycles whilst the volume of the particles decreases over the period.

Howling² studied particle debris from a pin-on-plate carbon fiber reinforce PEEK against ceramic test using a 6M KOH protocol and resin embedded TEM analysis.

This method only allowed around 100 particles to be imaged at a time, no size distribution was given. Ctyotoxicity was also tested using U937 monocytic cells indicating that MOTISTM has no cytotoxic effects such as necrosis.

In the majority of cases, failure of conventional metal-on-ultra-high molecular weight polyethylene (UHMWPE) artificial joints is due to wear particle induced osteolysis. Therefore, new materials have been introduced in an attempt to produce bearing surfaces that create lower, more biologically compatible wear. Polyetheretherketone (PEEK-OPTIMA) has been successfully used in a number of implant applications due to its combination of mechanical strength and biocompatibility.

Multi-directional pin-on-plate wear tests were performed on carbon fibre reinforced PEEK-OPTIMA (CFR-PEEK) against CoCrMo. PAN-based CFR-PEEK was tested against both low carbon and high carbon CoCrMo and Pitch-based CFR-PEEK was tested against high carbon CoCrMo only. The multi-directional motion of the pin-on-plate machine replicated the crossing of the wear paths that would be expected in vivo. For each test, four pin and plate samples were tested for two million cycles at a cycle frequency of 1 Hz under a 40 N load (which resulted in a contact stress of about 2 MPa). The lubricant used was bovine serum diluted with distilled water to a protein content of 15 gl-1. This was maintained at 37 °C. The wear was determined gravimetrically. Soak control specimens were used to account for any weight changes due to lubricant absorption.

The average steady state wear for the CFR-PEEK pins was found to be 0.144, 0.176 and 0.123 × 10-6 mm 3N-1m-1 for the CFR-PEEK PAN pins against low carbon CoCrMo, CFR-PEEK PAN pins against high carbon CoCrMo and CFR-PEEK Pitch-based pins against high carbon CoCrMo. A comparison of the results from the low and high carbon plates articulating against the PAN-based pins shows that the high carbon CoCrMo produced slightly higher wear than the low carbon CoCrMo. The protruding carbides on the high carbon CoCrMo plates may have caused this increase in wear. The lowest wearing material combination in this study was CFR-PEEK Pitch against high carbon CoCrMo. Published papers on the wear of UHMWPE against stainless steel [1] have shown higher wear factors (1.1 × 10-6 mm3N-1m-1).

Pitch and PAN-based CFR-PEEK against CoCrMo (low carbon or high carbon) provided low wear rates. On average, the Pitch-based material against high carbon CoCrMo provided the lowest wear in these tests. All the material combinations gave lower wear than found for UHMWPE-on-stainless steel tested under similar conditions. This gives confidence in the likelihood of this material combination performing well in orthopaedic applications.

The authors wish to thank INVIBIO Ltd for funding this research.

Aseptic loosening caused by UHMWPE wear debris induced osteolysis is a major cause of revision in total hip arthroplasty (THA)¹. While second generation hard-on-hard bearings, metal-on-metal (MOM) and ceramic-on-ceramic (COC), have been shown clinically to address the wear issues associated with conventional UHMWPE bearings, there remain some concerns over the potential effects of metal ions produced by MOM and the risk of liner fracture in COC. Recently, hybrid ceramic-on-metal (COM) articulation has received a great deal of attention as a promising alternative bearing.

Advantages include reduced wear and metal ion release compared with MOM. In addition, it is thought that there may be a reduced tendency for fracture of the ceramic component due to the softer metallic cup.

In this study a 5 million cycle wear test was carried out on the Mark II Durham Hip Wear Simulator. A set of six, 38mm diameter HIPed alumina heads and as-cast CoCr alloy cups were tested in bovine serum. Surface topography analysis was carried out at 0, 2, 3 and 5 million cycles. Additionally imaging of the bearing surfaces using ESEM and AFM was undertaken on the final bearing surface. Friction testing, using the Durham Hip Friction Simulator was carried out on one of the joints worn to 5 million cycles and the results were compared with theoretical calculations.

Wear of the ceramic heads was virtually undetectable using the conventional gravimetric methods. However, minor surface damage in the form of grain pull out and abrasive scratches was observed in the wear patch when the bearing surfaces were analysed using ESEM and AFM. The grains were not visible in the unworn sections of the head. The average surface roughness remained constant throughout the test. The CoCr cups showed a decrease in roughness between 0 and 2 million cycles, after which it remained relatively constant. This was consistent with the wear results in which a biphasic wear rate was found. The more frequently obtained wear results showed running in wear rate of 1.02±0.078 mm3/million cycles between 0–0.5 million cycles, followed by a steady state wear rate of 0.030±0.011 mm3/million cycles. These results are consistent with those of a recent study undertaken elsewhere².

Friction testing produced a Stribeck curve which was indicative of full fluid film lubrication with a friction factor of 0.027±0.002 for 25% bovine serum (η=0.0014 Pa s-1). Other tests were also carried out using carboxy methyl cellulose fluid as the lubricant to investigate the effect of proteins. This showed that there was a small decrease in friction factor when proteins were absent from the lubricant. It is thought that the difference in friction factors is due to adsorption of the proteins onto the bearing surfaces, when lubricated in bovine serum. This introduces large proteins between the bearing surfaces, which penetrate the lubricant film, causing protein on protein interactions, in addition to the friction caused by shearing of the lubricant film.

The introduction of unicondylar knee prostheses has allowed the preservation of the non-diseased compartment of the knee whilst replacing the diseased or damaged compartment. However, as is well known, there is concern that the body’s biological reaction to ultra-high molecular weight polyethylene (UHMWPE) wear particles leads to bone resorption and subsequent loosening and failure of the joint. Also, in some cases, delamination of the UHMWPE tibial bearing surface has been found to occur leading to failure of these conventional joints. Therefore new material combinations have been investigated within the laboratory.

The unicondylar knee that was tested consisted of CoCrMo tibial and femoral components between which a mobile Pitch-based carbon fibre reinforced polyetheretherketone (CFR-PEEK OPTIMA®) meniscal bearing was mounted. The joints were supplied by INVIBIO Ltd. Tribological tests were performed on these knees using the Durham six station knee wear simulator and the Durham friction simulator II. In both cases the loading and motion were similar to the standard walking cycle. On the six station wear machine five stations applied both the loading and motion and were the active stations and one applied loading only as it was used as the loaded soak control station. Approximately every 500,000 cycles, the wear of the CFR-PEEK meniscal bearings was assessed gravimetrically (using a Mettler Toledo AX 205 balance, accurate to 0.01 mg) and the loaded soak control was used to take account of any change in weight due to lubricant absorption. The joints were tested to five million cycles (equivalent to approximately five years in vivo) with diluted new-born calf serum as the lubricant which gave a protein content of 17 gl-1. At periods throughout the wear test the surface topography was measured on the Zygo NewView 100 non-contacting profilometer. Friction tests were performed at the beginning and the end of the wear test.

The average volumetric wear rate of the medial and lateral components was found to be 1.70 and 1.02 mm3/million cycles respectively (range 0.66 – 2.73 and 0.59 – 2.45 mm3/million cycles respectively). This is lower than the reported wear rate of metal-on-UHMWPE unicondylar knee joints (6.69 and 2.98 mm3/million cycles for the medial and lateral components respectively) [1]. The surface topographical analysis of the CFR-PEEK bearings showed a reduction in surface roughness and also a change to more negative skewness (i.e. more valleys than peaks) which may aid in lubrication. Before and after wear testing the joints were found to be operating in the boundary/mixed lubrication regime.

The Pitch-based CFR-PEEK unicondylar knee joints performed well in these wear tests. They gave lower volumetric wear rates than metal-on-UHMWPE uni-condylar knee prostheses. The friction tests showed that at physiological viscosities, these joints operate in the boundary/mixed lubrication regime. These results show that this novel joint couple may potentially be an alternative solution for the reduction of wear and osteolysis.

The authors wish to thank INVIBIO Ltd for funding this research.

In designing artificial joints the main criteria are to reduce the wear-rate of the material and the reaction of the body to the wear particles produced. These can be achieved by using harder materials (metals, ceramics) or by reducing the chances of producing a wear particle by the choice of material and design. This paper will look at combinations of PEEK-OPTIMA against different counterfaces with the aim of reducing the wear-rate of artificial hips and knees.

Pin-on-Plate Studies Twenty-six different sets of experiments combining PEEK-OPTIMA in different formulations and against different counterfaces were conducted to evaluate the lowest wear combination.

The lowest wear-rate combination was CFR-PEEK PAN against low carbon CoCrMo alloy (K=0.144×10-6 mm3/Nm) which is only about 1/8th of the wear of UHMWPE against stainless-steel (1.1×10-6 mm3/Nm). Gamma radiation sterilisation did not seem to affect the PEEK wear-rate.

Hip Simulator Studies A 25 million cycle wear study has been conducted on the Durham Hip Simulator using 54 mm diameter alumina heads against CFR-PEEK thin-walled acetabular cups (MITCH). Five joints were in active stations and one acted as a loaded control. Wear was measured gravimetrically.

Particles were analysed using a NanoSight LM10 instrument at 0.5, 10 and 25 million cycles. Also an Atomic Force Microscope (AFM) was used to look at particles above 2μm which is the limit of the NanoSight instrument.

The wear-rate was linear over the whole 25 million cycle test at 1.16 mm3/ million cycles (range 0.811–1.392 mm3/million cycles). As the test progressed, the number of particles reduced and the dominant particle size increased from about 40nm to circa 200 nm. The AFM showed some particles as large as 3μm to exist also. No fluid film lubrication was observed to be generated in these joints so the low wear-rate was due to the inherent low-wear properties of the material combination.

Knee Simulator Studies CFR-PEEK was moulded into the interpositional bearings for experimental lateral and medial unicompartmental knee designs and tested for 5 million cycles using 5 pairs of active joints and one pair of loaded controls in the Durham Knee Simulator. Wear was measured gravimetrically.

Whilst the CFR-PEEK components gave a total wear-rate (both medial and lateral) of 2.72 mm3/million cycles, UHMWPE inserts in a similar application¹, gave 9.67mm3/million cycles. This represents a reduction of 72 per cent for the wear of the CFR-PEEK components.

Conclusions CFR-PEEK used in the correct combination and application can give a much reduced wear-rate compared with UHMWPE. It does not have the problem of metal ion release and has been shown by others not to exhibit cytotoxicity.²

Carbon fibre reinforced polyetheretherketone (CFR-PEEK) has been introduced recently as an alternative material to be used in joint prostheses. During injection moulding of the CFR-PEEK the carbon fibres tend to become orientated in the direction of the plastic flow. The direction of these fibres may affect the wear produced by these materials.

Reciprocation only and reciprocation plus rotation (multi-directional) pin-on-plate wear tests were performed on PAN-based CFR-PEEK against itself. The plates were manufactured with the carbon fibres mainly orientated either longitudinally (in the direction of reciprocation motion) or mainly transversally (perpendicular to the direction of motion) to determine the effect of carbon fibre orientation on the wear of these materials. For each test, four pin and plate samples were tested (two reciprocation only and two reciprocation plus rotation) for three and a half million cycles at a cycle frequency of 1 Hz under a 40 N load (which resulted in a contact stress of about 2 MPa). The lubricant used was bovine serum diluted with de-ionised water to a protein content of 17 gl-1. This was maintained at 37 °C. The wear was determined gravimetrically. Soak control specimens were used to account for any weight changes due to lubricant absorption.

The average steady-state wear for the CFR-PEEK samples that underwent reciprocation motion only was found to be 5.41 and 18.7 × 10-8 mm3N-1m-1 for the longitudinal carbon fibres and the transverse fibres respectively. For the multi-directional tests, the average steady-state wear was 5.88 and 19.9 × 10-7 mm3N-1m-1 for the longitudinal and transverse fibres respectively. It is clear from these results that for both reciprocation motion only and reciprocation plus rotation the wear was considerably lower with the fibres orientated in the longitudinal direction than the transverse direction. Also, these tests show that reciprocation only gives approximately an order of magnitude lower wear than multi-directional motion.

It can be concluded that the wear rate of CFR-PEEK is lower when the sliding motion occurs in the same direction as the carbon fibre orientation. Also, in these pin-on-plate tests, the wear produced using reciprocation motion only was an order of magnitude lower than that for the tests using multi-directional motion.

The authors wish to thank INVIBIO Ltd for funding this research.

The generation of particle debris from ultra high molecular weight polyethylene (UHMWPE) against metal hip joints has been shown to cause osteolysis leading to joint loosening in the medium term. This is known as late aseptic loosening since infection is absent¹.

In an attempt to reduce the volume of wear debris, attention has moved to metal-on-metal prostheses as the total volume of wear debris is less. However, the size, shape and number of the particles are important as well as the total volume as these affect the biological response of the body leading to aseptic loosening.

The Durham Mk I Hip Joint Simulator was used to generate CoCrMo wear particles in a series of tests. Four simulator tests took place in succession, initially 50 mm Birmingham hip replacements were tested where both the head and the cup were as-cast CoCrMo alloy. A second test was conducted where the components were 38 mm and both head and cup were as-cast CoCrMo. A third test using 50 mm components was completed where both head and cup were double heat treated CoCrMo alloy and a final test took place where both components were 50 mm the head was as-cast and the cup was as-cast which had been pre-worn to 5 million cycles. Bovine serum with a concentration of 17 g/l of protein was used as a lubricant and particles were sampled every half million cycles. The volumetric wear was also obtained gravimetrically.

A double enzymatic protocol was used to cleave the proteins from the particles taking great care to minimise any effect on the particles. Finally the particles were suspended in distilled, de-ionised water to enable them to be analysed using a NanoSight LM10 particle analyser. This yielded the size distribution of the particles. This was then confirmed by placing an aliquot of the suspended particles onto a carbon coated copper grid and drying them under a lamp. These particles were then imaged in the TEM. Energy Dispersive X-ray analysis was also used to obtain the chemical composition of the particles.

The results indicated a strong correlation between the gravimetric wear and the number of particles. All the as-cast CoCrMo alloy tests showed a consistent particle modal average size. The double heat treated particles were shown to be smaller, with occasional large flake like particulates which were identified under the TEM. This particle data correlates extremely well with previous data from simple material testing pin on plate experiments.

Previous studies have used microscopy to investigate the size and morphology of the particulate debris², however these studies are limited due to the time taken to image the particles individually. This current method allows many more particles to be analysed, thus the data accumulated is more statistically significant and may be compared with the wear volumes calculated gravimetrically.

The first metatarsophalangeal (MTP) joint is the key joint of the foot in terms of function during gait. Various replacement toe joint prostheses are commercially available but unlike other replacement joints such as the hip or knee, few simulator based studies have been conducted to evaluate the performance and reliability of these prostheses. Presented are results obtained using a newly developed and validated multi-station MTP joint test-rig that is able to simulate the natural biomechanics of the toe joint. The developed simulator is a multi-station computer controlled electro-pneumatic metataso-phalangeal joint simulator that applies dynamic loading and motions commensurate with the walking gait cycle. This involves the combination of plantar-dorsi flexion range of 32 degrees, 5 degrees of inversion/eversion and toe-off dynamic loading peaking at up to 820 N. Presented are the validation and in vitro test results of MTP joint simulations carried out on silastic and articulating metal and polymer designs of MTP prostheses. Despite being subjected to a reduced loading regime of 300 N peak force, the silastic prostheses were found to perform poorly in the simulator, ultimately failing due to a combination of fatigue crack growth and joint collapse. This behaviour and failure mode was consistent with that of ex vivo origin silastic MTP prostheses examined and provides confidence in the validity of the simulator.

Introduction: The Birmingham Hip Resurfacing (BHR) system comprises both a BHR femoral head and a large modular femoral head for use should a total hip replacement be required. The modular femoral head has identical material chemistry, microstructure, spherical form, and surface roughness of the bearing surfaces of resurfacing femoral head and both BHR and THR devices share the same acetabular components. Hence, if the femoral component of a BHR needs revision surgery, the Birmingham hip system provides the potential of converting it to a THR without the need to also revise the well fixed cup. Although it stands to reason that the wear behaviour of the BHR and Birmingham THR will be similar, it is important to investigate the wear behaviour of new THR modular heads against worn BHR cups, representing revision of BHR to Birmingham THR without cup revision. The aim of this study is to assess the viability of the femoral component revision for BHR devices whilst leaving the acetabular components in situ in the pelvis.

Materials and Methods: The wear and friction tests were conducted with pristine modular heads paired with BHR cups which had already undergone 5 million cycles (Mc) of wear in a hip simulator against BHR heads.

Results and Discussions: The average wear rate of the new Birmingham THR modular heads against worn cups was 0.42 mm³/Mc whilst the new BHR heads against new cups generated wear rate of 0.67 mm³/Mc. Supported by the friction test results, it indicated that the new femoral heads paired with worn cup did not negatively affect the substantial amount of fluid-film lubrication that had developed over the course of the original test. Therefore, it is acceptable to use new femoral heads against worn cups, if the cups are not damaged, well fixed and correctly orientated.

Introduction: In an attempt to prolong the lives of implantable devices, several ‘new’ materials are undergoing examination to determine their suitability as joint couplings. As part of a series of tests, polyetherether-ketone (PEEK) against cobalt chrome molybdenum (CoCrMo) and carbon fibre reinforced-PEEK against CoCrMo were tested on a multidirectional pin-on-plate machine.

Materials and methods: The two four station pin-on-plate machines used in this study applied both reciprocation and rotational motion. Each material combination was tested individually on separate machines. Four samples of PEEK pins against CoCrMo plates were tested and eight samples (two tests) of CFR-PEEK pins against CoCrMo plates were tested. The pins were supplied by Invibio Ltd. A 40 N load was provided to each station. The lubricant used was 24.5 % bovine serum (protein content: 15 g/l) and this was heated to 37 degrees C. The wear was assessed gravimetrically and the tests each completed 2 million cycles.

Results: On average, the pin and plate wear factors were 7.37 and 0.010 x 10 -6 mm3/Nm for PEEK-CoCrMo and 0.144 and 0.011 x 10 -6 mm3/Nm for the CFR-PEEK against CoCrMo specimens respectively. These results show the wear of the components corrected relative to the control specimens that therefore took into account the weight gain due to lubricant absorption.

Discussion: The CFR-PEEK pins gave considerably lower wear against CoCrMo than the PEEK pins. It is interesting to note that the total wear factor provided by high carbon CoCrMo pins articulating against high carbon CoCrMo plates (which is known as a low wearing material combination in hip implants) was found to be 0.84 x 10 -6 mm3/Nm (1) which is actually higher than that found in these studies for CFR-PEEK against CoCrMo tested under the same conditions.

Conclusions: CFR-PEEK articulating against CoCrMo provided much lower wear than the PEEK-CoCrMo samples. This material combination also gave lower wear than metal-on-metal samples. This, therefore, indicates that this material combination may perform well in joint applications.

Introduction: The wear of orthopaedic biopolymers is recognised as a major factor in the failure of total joint replacements. Clinical wear data exists for acetabular cups manufactured from three biopolymers: ultra high molecular weight polyethylene (UHMWPE); poly tetra fluoro ethylene (PTFE); and polyacetal. The aim of this paper was to wear test these biopolymers and compare the results with clinical data.

Method and Materials: The biopolymers were tested using a modified, four-station, pin-on-plate wear rig [1]. In the tests, two of the four stations applied reciprocating motion and two applied multi-directional motion. Biopolymer pins articulated against stainless steel plates under a load of 40N. The lubricant consisted of 25% bovine serum and 75% distilled water. A standardised cleaning and weighing protocol was followed, and the biopolymer wear factors were calculated by dividing the volume lost by the product of the load and the sliding distance.

Discussion and Conclusions: Failed and retrieved UHMWPE acetabular cups have been reported as having a clinical wear factor of 2.1 x 10⁻⁶mm³/Nm [2]. However, UHMWPE cups which have been functioning well until removal at post-mortem have been said to show 45 to 69% less wear than revised UHMWPE cups [3]. Combining these values suggests clinical wear factors for functional UHMWPE in the range of 0.95 to 1.45 x 10⁻⁶mm³/Nm. This range fits well with the value of 1.1 x 10⁻⁶mm³/Nm shown in table 1 for UHMWPE under multi-directional motion. A clinical wear factor of 37 x10⁻⁶mm³/Nm has been calculated for PTFE acetabular cups [4]. When compared with the mean wear factor for PTFE pins under multi-directional motion obtained from the pin-on-plate rig, the match is remarkable. For polyacetal cups a mean volumetric wear of 136mm³/ year has been reported [5] and it has been calculated that explanted hip prostheses averaged 1.54 million cycles/year [2]. In polyacetal acetabular cups of 37mm diameter, an average sliding distance of 25mm/cycle can be calculated [6] and it has been said that an equivalent static load of 1000N applies [7]. Taking these four values permits a clinical wear factor for polyacetal cups of 3.5 x 10⁻⁶mm³/Nm to be calculated. This number compares well with the value of 3.8 x 10⁻⁶mm³/Nm seen for the polyacetal test pins under multi-directional motion. In summary, all three biopolymers subject to multi-directional motion exhibited clinically relevant values of wear.

Introduction The clinical use of an all-polymer knee which articulated a polyacetal femoral component against an ultra high molecular weight polyethylene (UHMWPE) tibial component has been reported [1]. A ‘polyacetal group’ of 63 total knee replacements were followed for at least ten years and no instances of femoral component fracture or failure due to wear occurred [1]. Such results are remarkable for an all-polymer pros-thesis in such a heavily loaded joint as the knee. Recently a wear screening device has been described which reproduced in vitro the clinical wear rates reported for three biopolymers which have been employed as the acetabular cup material in hip prostheses [2]. Given this validated rig, the objective of the work reported here was to undertake wear tests of polyacetal against UHMWPE.

Materials and Methods The polyacetal and UHMWPE couples were tested using a modified, four-station, pin-on-plate wear test rig [2]. The modification entailed the addition of rotational motion to the test pins, in addition to the standard reciprocating motion, to give multi-directional motion. In the wear tests, two stations had reciprocation-only and two applied multi-directional motion. Investigating the influence of both types of motion permitted a fuller tribological analysis to be undertaken. Control pins and control plates were included to account for any weight change due to lubricant uptake. A load of 40N was employed and reciprocating and rotating speeds of 1Hz were chosen. The lubricant consisted of 25% bovine calf serum and 75% distilled water, which was heated to 37°C during testing. A standardised cleaning and weighing protocol was followed, and the pins and plates were weighed on a balance sensitive to 0.1mg.

Results and Discussion After an average of 1.4 million cycles of sliding, the mean wear factors were: UHMWPE pins rubbing against polyacetal plates, 1.5 x 10-6mm3/ Nm under reciprocation, and 4.1 x 10-6mm3/Nm under multi-directional motion. For polyacetal pins rubbing against UHMWPE plates they were 0.7 x 10-6mm3/ Nm under reciprocation, and 2.8 x 10-6mm3/Nm under multi-directional motion. As can be seen, the wear factors depended on both the orientation of the material, whether it was a pin or a plate, and the motion it was subjected to. The increase in weight of the polyacetal control components due to lubricant uptake was many times that of the UHMWPE components. For example the UHMWPE control plate showed an increase of 0.2mg compared with 33.4mg for the polyacetal control plate. Using the same wear screening rig, the wear factors for UHMWPE articulating against stainless steel were measured to be 0.1 x10-6mm3/Nm under reciprocating motion and 1.1 x10-6mm3/Nm under multi-directional motion [2]. Though greater than this latter value, the all-polymer wear factors were not excessively high and were less under reciprocation-only. How much multi-directional motion, or cross-shear, it is appropriate to apply to a wear simulation of an artificial knee joint is worth further investigation, as it may be much less than in the hip joint.

New generation alumina-on-alumina (A-A) prostheses have been introduced to try and overcome the problem of osteolysis often attributed to polyethylene wear particles liberated within conventional metal-on-ultra high molecular weight polyethylene (UHMWPE) joints. This study uses a hip simulator to compare the volumetric wear rates of five different radial clearances of A-A joints. Atomic force microscopy (AFM) provided topographic characterisation of the prosthesis surfaces throughout the wear test.

Materials and methods: The wear test was performed on the Durham hip joint wear simulator. The 28 mm diameter alumina ceramic couples investigated were manufactured by Morgan Matroc Ltd., in accordance with ISO 6474. Four radial clearances (33, 40, 48 and 74 μm) of A-A joints were tested to one million cycles with 25% new-born calf serum as the lubricant. Contact mode AFM (TopoMetrix Explorer SPM) was used to produce a topographical map of the poles of the four alumina heads every 0.1 million cycles. Every 0.2 million cycles the wear was assessed gravimetrically using a Mettler AE 200 balance (accurate to 0.1 mg).

Results and discussion: There was no measurable wear of either the heads or cups during this one million cycle wear period. Throughout the wear test the alumina equiaxed grain structure became apparent on the AFM images, the mean alumina grain size was 2 μm. The grain surfaces were below the mean femoral head surface height. Such topography was not observed on an as-received head. Some granular pull-out also took place. As the wear test proceeded, the average area surface roughness increased from 2.33 nm to 4.42 nm for the heads but stayed relatively constant for the cups (from 2.75 nm to 2.97 nm).

Conclusions: The very low wear produced by these A-A hip joints is very difficult to measure gravimetrically as it is close to the limits of resolution of the weighing equipment. The surface topography analysis, however, shows changes to the ceramic surfaces during the wear test and gives an indication of the wear processes and lubrication regimes acting within such joints. The authors wish to thank EPSRC for funding this research and Morgan Matroc Ltd. for supplying the joints.

Modular knee bearings typically consist of a femoral component, a tibial base-plate and a polyethylene insert, which is located in the tibial base-plate using some sort of locking mechanism. Although modular knee bearings offer many advantages there is the potential for micro-motion between the tibial insert and the base-plate.

Tests were performed on six large Kinemax Plus knee bearings (snap-fit design) to evaluate the amount of movement between the tibial inserts and the tibial base-plates. The knee bearings were tested up to one million cycles on the Durham Six-Station Knee Wear Simulator which subjected the bearings to similar motion and loading profiles that would be experienced by the natural knee during walking. The movement of the tibial inserts was measured with dial gauges (accuracy of ±0.01 mm) before and after the bearings were tested on the simulator, when unloaded, and throughout the tests whilst the bearings were being dynamically loaded in the simulator. Movement occurred between the tibial insert and the tibial base-plate after initial assembly due to the snap-fit mechanism used to locate the tibial insert within the tibial base-plate. However, this decreased appreciably when the bearings were loaded in the simulator. The amount of movement did not change with time when the bearings were continuously loaded in the simulator. However, after each test the amount of movement of the tibial inserts, when unloaded, was less than before the test. This was thought to be due to creep of the UHMWPE inserts. The movement between the tibial insert and tibial base-plate in-situ is likely to be much less than that observed by a surgeon at the time of assembly due to loading of the knee bearing in the body. However, the amount of movement when the tibial inserts are loaded may still be great enough to produce a second interface where wear of the tibial insert may take place.

Metacarpophalangeal (MCP) arthroplasty usually involves the fitting of a silicone spacer, commonly Swanson prosthesis, but more recently the Sutter prosthesis has been introduced.

Four Sutter MCP prostheses, two each sized 30 and 40, were removed from the right hand of a female patient. The patient aged 61 years ate revision, had longstanding rheumatoid arthritis. Using a single station finger stimulator¹ two Sutter size 50 MCP prostheses were tested. This stimulator ran at a speed of 100 cycles per minute. During each of these cycles, which flexed the test prosthesis through a 90° arc of motion, the load across the test prosthesis varied between 10N and 15N after 3000 cycles, the stimulator applied a static ‘pinch’ load and the whole combined load cycle began again. Ringer solution heated at 37°C was used as a lubricant. Clinically, the prostheses had been implanted for 53 months. All four had fractured at the junction of the hinge and distal stem. In the simulator tests the Sutter size 50 prosthesis managed just over 10 million cycles of flexion-extension, including over 3300 ‘pinch’ loads before fracture occurred, at the junction of the distal stem and hinge. The second prosthesis fractured in the same manner after 5.3 million cycles of flexion-extension.

These are the first reported in vitro results of fracture of Sutter prosthesis as well as the first paper to state the site of ex vivo fractures of Sutter prostheses. A computer model described in a recent paper ² indicated that failure of the Sutter prosthesis should occur at the central hinge region. Clearly the in vitro results and the ex vivo experience disagree with the computer model. McArthur and Milner ³ have shown clinically that the Swanson joint appears to be superior to the Sutter implant, a result confirmed elsewhere4. The finger stimulator has previously caused fracture of Swanson pros-thesis in a time and a manner comparable with surgical experience¹. Therefore another correlation with ex vivo results, but testing the Sutter prostheses has further validated the finger simulator.

The direction of wear in the acetabular socket has implications for the amount of debris that is generated during movement, for the magnitude of eccentric loading and for the incidence of impingement of the neck. We observed the direction of penetration with respect to a global co-ordinate system in 84 acetabular components retrieved at reoperation.

The mean direction of wear relative to the open face of the sockets was found to be 37° with a range from 0° to 87°. For those values determined using the inclination of the socket on the prerevision radiograph, the mean direction of penetration in the coronal plane had a lateral, rather than a medial, component. The mean angle was 84° (SD 17°) with respect to the horizontal. The angle of penetration was found to correlate significantly with the depth, in that the lateral component became larger as the wear progressed. There was also a significant correlation between the rate of penetration and the direction of wear. Despite the theoretical advantage of penetration in the superolateral direction, i.e., along the margin of the socket, in reducing the probability of impingement of the neck, no significant correlation was seen between the angle of penetration and the period of use in vivo. This may suggest that impingement of the femoral neck on the rim of the socket may not be the dominant factor in loosening of the socket but can still be important in a few cases.