CHAPTER OUTLINE
Tribology 468
Clinical Performance 471
Metal-on-Metal Hip Resurfacing 471
Metal-on-Metal Primary and Revision Hip Surgery 472
Metal Ions 472
Biological Response 473
Inflammatory versus Immunological Responses 474
Genotoxic Potential 475
Concluding Remarks on Biological Response 475
During the last two decades there has been revived interest in alternative bearings such as metal-on-metal (MM) because of their lower volumetric wear rates than conventional metal-on-polyethylene (MPE) bearings and, as a consequence, the hypothesized reduction of osteolysis and aseptic loosening. However, although the volumetric wear rates associated with MM bearings are much lower than those associated with MPE bearings, wear does still occur, and concerns have been raised regarding the release of corrosion products, mostly from the wear particles, with possible risks of hypersensitivity and, although not proven, risks related to genotoxicity and carcinogenicity. Therefore understanding the tribologic properties of MM bearings is critical, because the wear performance will influence the clinical function of the implant. This chapter reviews the key factors determining the tribology of MM bearings, the clinical results, current applications, the effects of metal particles, and the role of metal ions and metal hypersensitivity in tissue response and implant performance.
TRIBOLOGY
The three pillars of tribology—friction, lubrication, and wear—are inevitable consequences of allowing load to be transmitted in vivo through two interacting surfaces in relative motion. Cobalt-based alloys are used exclusively for MM hip implants, and the MM pairing is used exclusively in hip resurfacing arthroplasty ( Fig. 62-1 ). The principle constituents of these alloys are cobalt, chromium, and molybdenum, but they can be differentiated as being either high carbon (0.20% to 0.25%) or low carbon (0.05% to 0.08%), with the former usually demonstrating lower in vitro wear rates. MM hip bearings produce wear particles by some combination of the four classic mechanisms of adhesion, abrasion, corrosion, and surface fatigue, with an emphasis on abrasion and surface fatigue. The in vivo volumetric wear has been estimated from implant retrieval studies, and it increases with implantation time. The type of lubrication influences the severity of friction and wear. In simulator studies, the steady-state wear rate and the amount of running-in wear have been related to elastohydrodynamic film thickness for mean operating conditions of load, speed, and viscosity, along with elastic properties and combined surface roughness of the metals employed in each prosthesis. Using the data of other investigators, Paré and colleagues have suggested that in vivo run-in wear is higher than the steady-state wear that is established after about 18 months, providing dislocation and loosening do not occur.
Recent concerns with MM hip bearings are related to wear, including the elevated metal ion levels in the blood of patients, mostly caused by in situ corrosion of the wear particles, as well as the possibility of biologic responses to these ions and wear particles. There is further concern regarding fixation issues that are related to load and friction, alloy metallurgy, and elevated wear from microseparation. The purpose of this section is to identify the essential features of MM hip bearings that significantly influence tribologic performance and to suggest strategies regarding implant selection.
Lubrication and Geometry
Ultimately the clinical performance must provide the proof that an arthroplasty procedure is effective and that the implant involved is optimal. However, implant performance can be examined using wear simulators, with some measure of reliability and long before the clinical performance is ascertained. It is important to remember, however, that a wear simulator evaluates the hip bearing under conditions that represent some aspects of the clinical situation, and therefore a wear simulator does not necessarily provide good fidelity with all clinical situations. In particular, the details of the surgery including implant alignment, ligament tensions, and fixation are not usually included in the simulation. However, microseparation damage, which is presumably related to ligament tensions, has been examined in simulator testing and has been shown to increase wear significantly.
Having qualified the expectations of wear simulators, they do remain an important investigative approach. For example, the simulator wear of MM hip implants has been shown by Chan and colleagues and later by Paré and colleagues to decrease with an increasing ratio of the theoretical elastohydrodynamic lubricant film thickness to the measured initial combined root mean square surface roughness, known as the lambda (λ) ratio. Such behavior is consistent with the usual tribologic interpretation of the lambda ratio. For λ ≤ 1, boundary lubrication involving the protection afforded by lubricant molecules that have weak chemical attachments to the surface tends to dominate, whereas for λ ≥ 3, fluid film lubrication tends to dominate with a physical entrainment of a fluid between the surfaces that depends on fluid rheological (or flow) properties. For MM hip implants (and other applications) the fluid film lubrication also depends on surface deformation and thus is called elastohydrodynamic lubrication. In simulator testing, the fluid film lubrication regimen (λ ≥ 3) has been more effective than boundary lubrication in reducing wear.
To link this fluid film lubrication behavior to clinical reality, a simple geometric quantity for the contact, known as the effective radius ( R ), was specified by Medley and colleagues as R = R H (R H + C)/C ≅ R H 2 / C where C = R C − R H , R H = radius of curvature of the head, and R C = radius of curvature of the cup. The effective radius combined the size of the implant (R H ) and the radial clearance (C) into a single parameter. Because λ increased in an almost direct proportion with R, it was possible to correlate the steady-state volumetric wear rate ( V R in cubic millimeters per year) both in simulators and as determined from retrieved implants with the effective radius ( Fig. 62-2 ), where a value of 2 million cycles (Mc) in the simulator was assumed to be equivalent to 1 year in vivo. Dowson presented correlations of similar scope to those in Figure 62-2 but involving the relationship of both run-in and steady-state simulator wear with theoretical lubricant film thickness. The present study simplified this approach (by not differentiating between run-in and steady-state wear and by correlating wear with R and not theoretical film thickness) but still had a somewhat similar rationale.
In Figure 62-2 , the wear was assumed to be 1.5 times the wear of the head for the data of McKellop and colleagues, who had measured the volumetric wear of the heads (V H ) of mostly McKee-Farrar implants of 34.9 to 41.3 mm in diameter that had been in vivo for 1 to 24 years. Rieker and colleagues measured linear wear (sum of maximum depth of wear penetration in both head and cup) in mostly 28-mm–diameter implants that had been in vivo for up to 10 years. In order to include the data of Rieker and colleagues in Figure 62-2 , linear wear (L) in micrometers was converted to volumetric wear (V) in cubic millimeters using the formula V = 0.0466 L 1.47 ( Fig. 62-3 ), which was derived from the following:
- ▪
Calculations of L from simulator V from Medley and colleagues
- ▪
Calculations of V from measured retrieval L from Willert and colleagues
- ▪
Measured retrieval wear, both L and volumetric wear of the head only (V H ) , with the assumption that V = 1.5 V H , from McKellop and colleagues
Then V was divided by implantation time to get the volumetric wear rate (V R ).
Data values from wear simulators were also included in Figure 62-2 . Using a MATCO hip simulator, Medley and colleagues measured simulator wear with progressively improved precision, and eventually Chan and colleagues measured simulator wear of 28-mm–diameter implants with high R values; data from all of these studies were included in Figure 62-2 . In addition, simulator data from Vassiliou and colleagues was included. They measured the wear of 50-mm–diameter Birmingham Hip Resurfacing implants (supplied by Midland Medical Technologies [Smith and Nephew], UK), in a Durham hip simulator for 5 Mc.
Effective radius (R) is not the only factor involved in lubricant film generation and wear reduction, but, as shown in Figure 62-2 , it is a powerful one, for both simulator wear rates and estimated in vivo wear rates. The R value can be optimized for design purposes, especially with hip resurfacing implants. A high R value ensures that lubricant can be entrained into a gradually convergent gap between the surfaces, and it ensures that when there is direct contact the normal stresses remain relatively low. Implant components can be surface ground to specific radii of curvature with little waviness to produce low radial clearances and thus high R. The clearance must not be too low because a gap in the equatorial region is needed to avoid a combination of lubricant starvation and direct surface contact that can lead to high frictional torque and implant seizure, causing extensive surface damage and loss of fixation. Cup deformation also has to be considered when one decides how low a clearance to design. However, assuming seizure can be avoided, a low clearance and thus high effective radius can usually give thicker lubricant films and, when direct contact does occur, lower contact stress. Low levels of waviness are also likely to be important in promoting thick films and low contact stress. The R value can be set by the manufacturer and remains relatively constant in vivo and therefore is one of the most useful ways to control wear in MM hip implants. A beneficial larger R can be achieved by a large head and as low a clearance as permitted by manufacturing tolerances.
If large-diameter, low-clearance precision ground implants are used, Figure 62-2 suggests that volumetric wear rates can be driven to virtually zero, thus reducing most wear-related problems. Furthermore, subtle changes in surface geometry may help maintain beneficial lubrication effects in the long-term performance of MM hip implants. However, as extra insurance—for instance, for highly active patients or in the event of microseparation damage —the addition of a protective surface layer may be very beneficial if it remains intact over the long term, especially in reducing blood ion levels.
Alloy Microstructure
Lubrication and geometry, even with the help of a protective surface coating, cannot solve all MM wear problems because some direct contact must occur during stop-dwell-start motions in vivo. Under conditions of direct contact, boundary lubrication and surface microstructure become important in influencing wear. The boundary lubrication is not very well understood and is difficult to influence in any case. The influence of surface microstructure is under the control of the manufacturer, but the relationship to long-term wear is not well understood either. It is important to mention that MM bearings can be manufactured as wrought or cast materials. Although these two materials display similar chemical compositions, their microstructures are quite different. The wrought material has a small grain size with a fine homogenous distribution of small carbides, whereas the cast material exhibits large grains with blocky carbides. In order to produce more uniformly distributed carbides along grain boundaries, some implant manufacturers submit their implants to a high-temperature heat treatment process.
Wear studies for cobalt-based alloys (used exclusively for MM hip implants) have been conducted in a 25% serum solution with a linear-tracking, reciprocating pin on plate apparatus to examine the influence of microstructure on wear. Varano and colleagues found that dissolved carbon (in solid solution, not carbide form) was effective in stabilizing the face-centered-cubic phase of the cobalt-based alloy and thus inhibited strain-induced transformation (SIT) to the hexagonal close-packed phase to give lower wear ( Fig. 62-4 ). Varano and colleagues also found a somewhat lower wear level for as-cast compared with solution-annealed cast cobalt-based alloy (a topic of some industrial and commercial interest). This finding was probably related to the performance in the pin on plate apparatus of the different carbide morphologies and might not be clinically significant because the simulator studies of Bowsher and colleagues did not find this material-processing effect. However, McMinn and Daniel remained convinced that solution annealing can result in elevated wear and clinical problems in hip resurfacing arthroplasty.
Concluding Remarks on the Tribology
Taken together, the results presented (see Figs. 62-2 and 62-4 ) suggest that an implant with large effective radius (R) and high dissolved carbon would benefit from improved lubrication and reduced SIT to achieve lower wear. The lower contact stress associated with the large effective radius would further discourage SIT. The effective radius and dissolved carbon content are both parameters that are under manufacturers’ control and remain relatively constant in vivo. Therefore a simple yet potentially effective strategy for manufacturers has been identified, and some key tribologic issues discussed along the way. Recently, Medley has provided a detailed account of MM hip implant tribology.
Specifically, with good manufacturing precision a larger-diameter implant provides an opportunity to achieve a high R value and thus lower wear. Surface microstructure may be important in reducing long-term wear, and higher levels of dissolved carbon may be beneficial. The efforts to introduce a protective surface layer may be very effective in reducing wear. Problems may still exist with frictional torque and long-term fixation, but the revision rates reported by Rieker and colleagues are not alarming.
CLINICAL PERFORMANCE
Although MM total hips started to be implanted around the same time as low frictional torque MPE prostheses, early failures of poorly designed MM total hips such as the Stanmore contributed to the popularity of the Charnley concept of the MPE bearing. However, long-term follow-up looking at other implants such as the McKee-Farrar showed a cumulative survivorship at 20 years of 77% compared with 73% for the Charnley implants. These results, combined with their own observations of the good performance of the McKee-Farrar prostheses, motivated Müller and Weber to re-explore MM bearings in 1984. With the assistance of Sulzer Orthopaedics (now part of Zimmer), they developed the Metasul bearing, which is a MM hip implant made from wrought cobalt chromium molybdenum alloy with high carbon content (0.20% to 0.30%).
Although many of the initial designs were all-cemented implants, long-term results with total hip arthroplasty (THA) and even more so with hip resurfacing showed a high incidence of radiolucencies and a high failure rate of the cemented acetabular component, respectively, leading the majority of implant manufacturers to opt for cementless acetabular fixation. Currently it is estimated that more than 300,000 of these second- or modern-generation components have been implanted in patients worldwide, with cementless and/or hybrid designs showing excellent survivorship for both hip resurfacing and stem-type hip replacements.
In terms of actual in vivo wear, early data come from retrieved McKee-Farrar prostheses, which had an average linear wear rate of 0.003 mm/year and 0.004 mm/year for the femoral head and cup, respectively. It is interesting to note that the larger-diameter femoral heads (42 mm versus 35 mm) had a twofold lower mean volumetric wear rate, 0.7 mm and 1.4 mm per year. In a key article, Rieker and colleagues reported on the largest retrieval analysis of failed MM bearings of the current generation (high carbon, wrought cobalt chromium molybdenum alloy). They analyzed 608 individual retrievals (femoral heads and/or cups from 337 revisions), with the vast majority of implants having a diameter of 28 mm. The linear wear rate for the first year in vivo averaged 27.8 µm/year and was mainly influenced by the prostheses that were revised because of dislocation. The wear rate after the second year in vivo averaged 6.2 µm/year, with femoral heads generally exhibiting a higher wear rate compared with the cups. There was also a positive correlation between clearance and wear rate. In contrast, Reinisch and colleagues reporting on the retrieval of a low carbon cobalt chromium alloy found a higher mean linear wear rate of 7.6 µm/year, even though the bearing had the same diameter as, and clearances comparable to, those reported by Rieker and colleagues.
The favorable wear properties of large-diameter MM bearings have fostered the reintroduction of hip resurfacing and the use of larger-diameter femoral heads in stem-type hip replacements for both primary and revision situations.
Metal-on-Metal Hip Resurfacing
Hip resurfacing is no longer at the stage of evaluation and is regularly being offered as an alternative to THA for the young and active adult. As in other forms of conservative hip surgery, patient selection has helped minimize complications and risk of failure. The reintroduction of hip resurfacing as an MM bearing went through a variety of designs including all cemented as well as all cementless designs. Both short-term and long-term results have shown a poor survivorship of cemented acetabular fixation, with Beaule and colleagues reporting 66% at 7 years. Currently, hybrid fixation has the greatest amount of clinical data, with four major clinical series reporting survivorships of 97% to 99% at 4 to 5 years. All but one study prospectively applied some form of patient selection, which included but was not limited to diagnosis of osteoarthritis, absence of osteopenia, and absence of large femoral head cysts. The two main modes of failures were femoral component loosening and fracture of the femoral neck. Based on a retrieval analysis of modern MM hip resurfacings, Campbell and colleagues noted that cement penetration within the femoral head varied tremendously, occupying in some cases 89% of the femoral head and being associated with osteonecrotic lesions. As with the introduction of any new technology, surgical technique with respect to cementation technique and implant positioning has affected the survivorship and clinical function of hip resurfacing. One aspect of the surgical technique unique to hip resurfacing is the optimization of the femoral head-neck offset to minimize the risk of impingement, which commonly occurred in the original MPE hip resurfacings. This is especially relevant to MM bearings because impingement can lead to abnormal wear patterns as well as persistent groin pain. Therefore, although the application of MM bearings has permitted hip resurfacing to preserve bone on both the femoral and the acetabular sides and minimize wear-related failures, it remains a technically more demanding arthroplasty owing to the absence of modularity on the femoral side.
Metal-on-Metal Primary and Revision Hip Surgery
As stated, the midterm performance of current generation MM bearings has been excellent, with minimal complications. However, impingement has been reported as a cause of premature failure of MM bearings in primary THA. Because of the differential hardness between titanium alloy stem and cobalt alloy bearing, impingement leads to abnormal wear patterns and osteolysis. Reports highlight the different causes of osteolysis that are not directly related to metal debris generated from the articular surface. Several centers have reported osteolysis-related failures of MM THA secondary to a delayed hypersensitivity reaction to the metal debris. One of the most detailed reports is that by Milosev and colleagues, in which 25 hips out of 640 underwent revision surgery for aseptic loosening or pain; 16 of the 25 (64%) had an osteolytic lesion. Histologic analysis was performed on 17 hips, and 13 of those had evidence of a hypersensitivity-like reaction with aseptic inflammatory changes accompanied by moderate to extensive diffuse and perivascular infiltration of lymphocytes. Based on these numbers, the incidence of a hypersensitive reaction leading to revision surgery is 2%. Another manuscript, by Park and colleagues, reported osteolysis in 9 of 169 hips with an MM bearing. They concluded that this osteolytic lesion was secondary to a delayed-type hypersensitivity reaction based on histologic specimens available in only two hips and positive skin patch test results in eight of the nine hips. These conclusions, however, need to be put in perspective, because only two patients had histologic analyses and the remainder had skin patch tests, which have significant limitations. However, both of these studies reported on implants with a probable higher number of wear particles, which affects the immune response and risk of osteolysis. For example, in Park and colleagues the femoral implant was modular (S-ROM, DePuy Johnson & Johnson, Warsaw, IN), and in both Milosev and colleagues (low carbon on low carbon) and Park and colleagues (low carbon on high carbon), the bearing couplings were shown in hip simulator studies to have higher wear rates than high carbon combinations. Further research is required in this field because hypersensitivity reactions have also been reported in high carbon alloy implants. Also, we do not know their prevalence relative to the other bearing combinations.
Owing to the relative thinness of the all-metal acetabular liner (e.g., 40-mm head with a 46-mm socket) and the favorable wear properties of large femoral heads, most implant companies now offer this reconstructive option to minimize the risk of dislocation as well as impingement. This is especially relevant for revision hip surgery, where dislocation still remains a significant problem and the thin liner can be cemented within a well-fixed cementless socket or cage ( Figs. 62-5 and 62-6 ). In addition, the risk of fatigue fracture of the insert is nonexistent compared with a highly cross-linked polyethylene liner. There are few reports on large-diameter MM primary hip replacements. When Cuckler and colleagues compared large- to small-diameter MM primary hip replacements, they found a decreased incidence of dislocation with the large head size (38 mm) compared with the 28-mm size (0% versus 2.5%, respectively). They did not report any adverse outcomes with the large head sizes, but this was only at a short-term follow-up.
