Materials in Hip Surgery
Metals as a Bearing Material
Sophie Williams and John Fisher
• Metal (usually >0.2% carbon cobalt chromium molybdenum) is used as a bearing material in MOM hip replacements and surface replacements; manufacturing processes have little effect on the alloy’s wear characteristics.
• Design variables (such as diameter, clearance, and inclusion angle) will influence the wear performance of metal-on-metal (MOM) prostheses in optimal conditions.
• Wear behavior will be affected by nonoptimal component positioning (e.g., steeply inclined cups).
• Concern exists regarding some clinical failures of MOM bearings.
• Aspheric bearings, ceramic-on-metal articulations, and the use of coatings may offer alternatives to current MOM designs in the future.
Metal-on-metal (MOM) bearings are used worldwide in conventional hip replacements and hip resurfacing designs. In Australia, approximately 19% of all hip prostheses implanted in 2008 were MOM (of which approximately 12% were conventional designs).1 In the United Kingdom in 2008, 6% of implanted hip prostheses were surface replacements (data for numbers of conventional MOM hip replacements implanted were not available).2
Metal-on-metal hip replacements gained early prominence in the 1960s; usage then declined following reports of early failures in initial series and the success of metal-on-polyethylene bearings. With these early bearings, a number of configurations were used; in some cases, they were existing components intended for use in hemiarthroplasty such as the McBride/Moore, Urist/Moore, and Urist Thompson systems. Observed problems were primarily impingement, loss of range of motion, and the presence of a stress riser in the stem portion of the acetabular component. More widely used was the McKee-Farrar design, which used a standard Thompson femoral component (later modified to reduce impingement). Clinical experience with early devices yielded less than satisfactory results in many cases; however, there were exceptions.3
The observation that a small number of patients with first-generation MOM prostheses exhibited good clinical and radiologic results after 20 years in vivo led to the development of second-generation MOM hip prostheses.4 In 1988, the Metasul prosthesis was introduced into clinical practice; early experience demonstrated low wear rates, and few prostheses required revision. More recently, MOM hip resurfacing has been offered as an alternative, in particular to young, active patients. Clinical results of MOM resurfacing are generally favorable; however, some variation in outcome is dependent on a number of factors. Clinical wear rates of MOM vary up to 40-fold.5,6 Factors effecting wear have been cited as design, component geometry (diameter and clearance), metallurgy of the alloy, component positioning, and prosthesis use. There is a drive to reduce MOM wear and ion release, following observations that some patients have increased cobalt and chromium blood/serum and/or urine levels. Long-term consequences of elevated levels of ions and effects of metal particles are not known.
Cobalt-based alloys dominate the material selection for bearing surfaces of MOM prostheses, because of their high wear resistance and corrosion resistance. The composition is specified by American Society for Testing and Materials (ASTM) F-1537 (Table 8-1); carbon content can vary (carbon is responsible for the generation of carbides, which strengthen the material and affect the wear resistance7). Additionally, processing (wrought or cast; with or without heat treatment) can affect the microstructure of the alloy. This has generated much debate in terms of effects on wear rate, production of wear particles, and ultimately the release of metal ions, all of which will be affected by the altered distribution of carbides.
Composition of CoCr Alloy as Specified by ASTM F-1537 Low and High Carbon
|ASTM F-1537 (low carbon) Forged||ASTM F-1537 (high carbon) Forged|
|Nickel||1 max||1 max|
|Iron||0.75 max||0.75 max|
|Manganese||1 max||1 max|
|Silicon||1 max||1 max|
|Nitrogen||0.25 max||0.25 max|
ASTM, American Society for Testing and Materials; max, maximum; n/s, not significant.
High-carbon (>0.2% w/w) CoCr alloy has a biphasic structure; small grains of CoCr are surrounded by embedded, hard, scratch-resistant carbides, which restrict grain size. Low-carbon (<0.05% w/w) CoCr alloys are softer than high-carbon alloys (because of the lack of carbides) and comprise a single-phase structure of larger grain size. Low carbon content alloys produce significantly higher wear rates than high carbon content alloys in both simple configuration wear tests and hip joint wear simulator tests.3,8-10 Hence, the pairing of low carbon cups with low carbon femoral heads is not recommended. High carbon/high carbon pairings show the lowest wear rates in hip joint simulator tests.10
The wear rates of cast and wrought CoCrMo alloys with and without various heat treatments have been compared and are the subject of debate. Dowson and associates11 reported no significant differences between wear volumes of wrought and cast high carbon CoCrMo materials. Heat treatments and hot isostatic pressing have been shown to have little effect on the wear rate of MOM hip prostheses. The effect of the method of manufacture on the wear resistance of MOM bearings has been further studied under adverse wear conditions in hip simulator studies. Bowsher and colleagues12 investigated the wear of double–heat-treated and as cast large diameter MOM hip bearings using standard and “severe” gait simulations. High carbon MOM bearings (40 mm diameter) were manufactured and were subjected to hot isostatic pressing and solution annealing, or to no heat treatment, after casting. No differences between the two groups under running-in and steady-state conditions were observed, and the authors concluded that changes in alloy microstructure (due to manufacturing route) do not appear to influence the wear behavior of high carbon cast MOM articulations with similar chemical compositions.
The low wear rates recorded for metal-on-metal articulations are surprising in the context of traditional engineering terms, which presume that like-on-like materials do not produce low wearing surfaces. In recent years, several mechanisms have been suggested to explain this observation.13 Abrasion is commonly suggested as a wear mechanism, because scratches and grooves are obvious on in vitro tested samples and MOM retrievals.14–16 Abrasion may be induced by foreign particles (contaminants from outside the system) or most likely by inherent particles in the system, such as fractured carbides, compacted wear debris, and plastically deformed parts of the metal matrix.
In theory, fluid film lubrication is a potential mechanism for generating low wear in like-on-like bearings.17 However, hydrodynamic lubrication is unlikely to be achieved in practice, because surfaces generally are roughened through the effects of third-body particles, and the articulations are subjected to conditions ranging from loaded static to cyclical motion, with frequent changes in load, velocity, and direction of relative motion. Wear mechanisms previously discussed include boundary lubrication by proteins, lipids, and even calcium phosphate deposits, and high carbon content carbides acting as ceramic/metal composites.8 Following pin-on-plate testing of CoCr on CoCr articulations, Tipper and co-workers8 suggested further alternative mechanisms: that multidirectional motion and its polishing action may act as a mechanism for reducing wear, and that nanometer-sized spherical wear particles may act as self-lubricating ball bearings, acting as third bodies between bearing surfaces, rolling, deforming, and acting as sites for motion and velocity accommodation, thereby minimizing the wear of the actual bearing surfaces. Later, Wimmer and associates7 carried out in vitro studies to assess the acting wear mechanisms. It was concluded that tribolayers (also seen on ex vivo samples13) are derived from protein buildup on surfaces due to a combination of mechanical and thermal contact stresses generated between the surfaces. These layers act as solid lubricants and act to reduce wear.
The wear mechanism of MOM bearings has been further considered with investigation of bio-tribocorrosion processes. A series of studies have demonstrated that depassivation of CoCr materials occurs as a result of contact between metallic counterfaces,18 and that ion release is dominated by the production of Co ions, but not in the ratio of the base alloy.19 In tribometer studies, corrosion can contribute up to 44% of the total damage,19,20 as reported by other authors.7,13 Yan and colleagues21,22 reported on the production of a protein-assisted tribofilm; it is believed that this is responsible for the wear-induced passivation seen in polarization studies. Corrosion also plays a significant role in ion release; corrosion enhanced by wear and wear debris dissolution are the two main sources, each having very different kinetics.23
MOM prostheses have been estimated to have 40 to 100 times less wear than metal-on-polyethylene bearings24; this is critical in extending the life of MOM bearings. However, much in vitro evidence suggests that the wear of MOM prostheses is highly dependent on the materials, the tribological design, and the finishing technique. Clinical studies of retrieved first- and second-generation MOM hip prostheses have shown linear penetrations of approximately 5 µm/yr25 and volumetric measures of approximately 0.33 mm3/yr.15 However, large levels of variation have been observed.
The wear of hard-on-hard bearings such as MOM hip prostheses has two distinct phases: (1) a period of initially elevated bedding-in wear that lasts approximately 1 million cycles, or the first year in vivo, followed by (2) a lower steady-state wear period, once the bearing surfaces have been subjected to the self-polishing action of metal wear particles, which may act as a solid-phase lubricant. This phenomenon is reported in numerous in vitro hip simulator tests9-12,26,27 and has been studied in greater detail than the clinical situation described by Heisal and co-workers.28 In vitro hip simulator testing of MOM implants and a parallel study assessing clinical serum metal ion concentration were conducted with the aim of characterizing the early running-in period in vivo and in vitro by assessing metal ion levels. Hip resurfacing prostheses were implanted in 15 consecutive patients, and serum metal ion concentrations were determined preoperatively and at 1, 6, 12, 24, and 52 weeks; also, the number of walking cycles was measured. In vitro, five similar components were investigated for three million cycles in a hip simulator; wear was assessed by quantifying wear particles and ions in serum samples. Serum chromium and cobalt levels of patients continuously increased during the first 6 months and showed an insignificant decrease thereafter. In contrast, simulator measurements showed a different wear pattern with a high-wear running-in period and a low-wear steady-state phase. The running-in period was delayed by 300,000 cycles and lasted up to 1 million cycles. In contrast, clinical data showed a slow increase in measured ion concentrations. The difference in wear patterns was attributed to the effects of distribution, accumulation, and excretion of particles and ions in vivo.
Implant Design Factors
The head diameter of total hip replacements has long been recognized as a factor affecting the stability and range of motion of the articulation because of the basic premise that the larger the head, the larger the distance must be displaced to dislocate from the cup.29
In terms of MOM bearings, the diameter of the head and cup and the clearance between them have been cited as design factors affecting the tribological performance of the bearing and so will be considered in this section. The premise that head diameter will affect wear is driven by theoretical predictions of lubrication conditions at the bearing surfaces. These analyses suggest that increasing the diameter will lead to reductions in wear rates caused by increased entrainment velocity of the surrounding fluid for a given angular velocity of the extremity, which, in turn, is predicted to improve lubrication and reduce friction.17
The effect of diameter has become increasingly important with resurfacing prostheses, as these cover the reamed femoral head (rather than replacing it) and therefore are of large diameter (average, approximately 54 mm). In hip simulator testing of MOM (CoCrMo on CoCrMo) prostheses with femoral heads of 16, 22.225, and 28 mm diameter, increasing the head size from 16 mm to 22.225 mm increased the mean volumetric wear rate (4.85 mm3/million cycles for 16-mm-diameter bearings and 6.30 mm3/million cycles for 22.225-mm bearings). When the diameter was further increased to 28 mm, it was observed that the average wear rate dropped to 1.6 mm3/million cycles.30 Dowson and colleagues31 further considered 36-mm total hip replacements and 54-mm resurfacing prostheses in a hip simulator study; steady-state wear rates were quickly established as the head diameter increased from 28 to 36 mm and then to 54 mm. In agreement with previous studies, as head diameter increased, wear volume decreased markedly, with steady-state values of 0.17 mm3/106 cycles for the 54-mm-diameter bearings.
Direct comparison has been made of surface replacements of different diameters (approximately 39 mm and 55 mm).32 Again, two distinct phases of wear were observed for both bearing sizes: bedding-in (up to 1 million cycles), during which the wear rate was elevated, and steady state (beyond 1 million cycles), where the wear rate was reduced. The bedding-in wear rate of the 39-mm bearings was significantly greater (123%) than that of the 55-mm bearings. It is interesting to note that this difference ceased to be significant between 1 and 15 million cycles, again showing the wear of surface replacements to be biphasic with bedding-in and steady-state wear phases, consistent with previous findings for MOM total hip replacements.9–12,27,28,30–32
A theoretical study by Jin and associates17 and previously discussed experimental studies all confirm that increasing head diameter wear in MOM bearings decreases overall wear rate.30,31 However, a study by Leslie and colleagues,32 comparing 39-mm and 55-mm bearings of the same type, was the first to report that the bedding-in period (as demonstrated by measurements of ion levels from the lubricating serum, in addition to gravimetric wear assessment) was shorter for the larger bearing. This suggests the possibility that the 55-mm bearings had a similar initial wear rate to the 39-mm bearings but a shorter bedding-in period, resulting in reduced wear in the first million cycles—a conclusion that is consistent with the geometric analysis of Hu and co-workers.33 As the volume of material that must be removed for bedding-in decreases with head diameter, the duration of the bedding-in period and the total wear volume generated are less with larger bearings, even if the actual rate of wear remains constant.
The work of Leslie and associates32 has also demonstrated that bearing size has no influence on the steady-state wear of larger bearings. Previous theoretical studies of lubrication have predicted differences in wear rates on the assumption that the wear process itself would not change the geometry of contact between counterfaces. However, as bedding-in occurs, the contact area increases and contact pressures decrease. Theoretical analysis indicates that the worn contact area (and therefore contact pressure) following bedding-in (after 1 million cycles) is similar for 39- and 55-mm bearings, despite the fact that the initial contact area is less (and contact pressures higher) for the smaller, 39-mm bearing. At the end of 15 million cycles of simulator testing, contact pressures and contact areas were similar for the 39-mm- and 55-mm-diameter bearings. So the importance of the conventional lubrication theory in determining wear of MOM bearings is mainly evident during the initial bedding-in stage. However, after the bedding-in stage, it appears that wear is determined largely by improved conformity of the bearing surfaces generated by bedding-in wear, as well as by the corresponding contact mechanics. The fact that little difference is observed in the measured wear volume of 39-mm and 55-mm bearings appears to be the result of two competing effects: the higher entraining velocity of the larger size, leading to improved fluid film lubrication, versus the shorter sliding distance of the smaller size.
The effect of the bearing diameter of MOM prostheses has also been studied clinically. Antoniou and associates34 compared blood ion levels (cobalt, chromium, and molybdenum) of patients with metal-on-metal total hip prostheses versus a 28- or 36-mm-diameter femoral head, and patients with hip resurfacing prostheses. Variations between groups with MOM bearings of different diameter were noted 6 months postoperatively (e.g., the median cobalt level was significantly lower in the 36-mm hip replacement group than in the 28-mm hip replacement group). However, neither median cobalt levels nor median chromium levels were significantly different among the three MOM groups at 12 months. These findings reflect in vitro findings32 where the most significant differences in wear were observed in the bedding-in period.
Langton and colleagues35 considered a series of 76 consecutive patients after resurfacing arthroplasty and measured chromium and cobalt ion concentrations in whole blood. They found that patients with smaller (≤51 mm) femoral components had ion levels that were significantly higher than those with larger (≥53 mm) components at a mean of 26 months postoperatively. These findings contrast with those from the study by Antoniou and co-workers.34 Langton and associates35 also reported the effects of variations in cup positioning on ion levels. Cup position is important because it affects bedding-in wear and may possibly explain the differences between published studies.
The trend widely observed with MOM bearings where wear decreases with increasing diameter contrasts with that reported for conventional ultra-high-molecular-weight polyethylene (UHMWPE)-on-metal hip prostheses, where the wear of the UHMWPE acetabular cups was shown to be proportional to the sliding distance,36 as predicted by basic engineering principles.37 Therefore, reducing the femoral head diameter in polyethylene bearings should lead to a reduction in wear volume and extension of prosthesis life. Charnley demonstrated the validity of this relationship and showed that the maximum wear life of hip replacements could be achieved by making the head diameter half the acetabular socket diameter.38 The Charnley low-friction arthroplasty, appropriately regarded as the “gold standard” of hip replacement, falls within this range, with a standard femoral head diameter of 22.225 mm.