There have been no large comparative studies of the blood levels of metal ions after implantation of commercially available hip resurfacing devices which have taken into account the effects of femoral size and inclination and anteversion of the acetabular component. We present the results in 90 patients with unilateral articular surface replacement (ASR) hip resurfacings (mean time to blood sampling 26 months) and 70 patients with unilateral Birmingham Hip Resurfacing (BHR) implants (mean time 47 months).
The whole blood and serum chromium (Cr) and cobalt (Co) concentrations were inversely related to the size of the femoral component in both groups (p < 0.05). Cr and Co were more strongly influenced by the position of the acetabular component in the case of the ASR, with an increase in metal ions observed at inclinations > 45° and anteversion angles of < 10° and > 20°. These levels were only increased in the BHR group when the acetabular component was implanted with an inclination > 55°.
A significant relationship was identified between the anteversion of the BHR acetabular component and the levels of Cr and Co (p < 0.05 for Co), with an increase observed at anteversion angles < 10° and > 20°. The median whole blood and serum Cr concentrations of the male ASR patients were significantly lower than those of the BHR men (p < 0.001). This indicates that reduced diametral clearance may equate to a reduction in metal ion concentrations in larger joints with satisfactory orientation of the acetabular component.
Hip resurfacing using cementless acetabular components with metal-on-metal bearings for the treatment of second-stage hip osteoarthritis was introduced approximately ten years ago.1,2 The implants in use currently are popularly referred to as the ‘third-generation’ and include the Birmingham Hip Resurfacing (BHR) (Smith and Nephew, Warwick, United Kingdom). Independent intermediate clinical results have shown the BHR to be a highly successful procedure when performed to a satisfactory technical standard after appropriate patient selection.3
De Smet et al4 recently demonstrated that serum metal ion concentrations can be used as a surrogate marker of articular wear, owing to the positive correlation with metal ion concentrations in the joint fluid. Increased wear has been implicated in the early failure of metal-on-metal joints secondary to metal sensitivity and the development of pseudotumours.5,6 With the added theoretical concern of carcinogenicity, low levels of metal ions are desirable in resurfacing arthroplasty.
It has not yet been determined in vivo whether the cumulative effects of the macro- and microscopic differences in the design of resurfacing devices from competing manufacturers will lead to an increase or reduction in wear rates or, more importantly, a change in the clinical outcome of the patient receiving the implant. No large-scale independent comparative study of the different types of hip resurfacing has been carried out using the same laboratory for metal ion analysis and taking into account the anteversion and inclination of the acetabular component and the implant size. We sought to determine whether there is a significant difference in the chromium (Cr) and cobalt (Co) concentrations in the blood of patients surfaced with two common types of resurfacing arthroplasty. We also sought to identify the orientations of the component most strongly associated with the lowest blood metal ion levels and, in so doing, test our previous recommendations for the orientation of the acetabular component7 of the articular surface replacement (ASR) (DePuy, Johnson and Johnson, Leeds, United Kingdom) using a larger group of patients.
Patients and Methods
In vitro studies8,9 have suggested that full fluid film lubrication can be promoted by reducing the radial clearance between the articulating surfaces of the resurfacing components. It was on this basis that the ‘fourth-generation’ ASR was introduced with a radial clearance of 50 μm. This nominal clearance is significantly lower than the 100 μm radial clearance of the BHR.
Failure analyses and metal ion studies have shown that edge loading is an important factor leading to increased wear secondary to increased angles of acetabular component inclination.10–12 The ASR acetabular component is sub-hemispherical in design, as opposed to the near full hemisphere of the BHR implant. De Haan et al13 recently showed that reduced component cover can render a joint more susceptible to the negative effects of edge loading at lower angles of inclination than expected.
Although all manufacturers use high-carbon-containing cobalt-chromium alloy, the processing of the alloy differs. Cast components may undergo post-casting heat treatments such as hot isostatic pressing and/or solution heat treatment. The significance of these treatments has been debated over the last decade. Annealing results in depletion of the surface carbides, but hip simulator studies do not demonstrate significant differences between the wear behaviour of as-cast arthroplasty and heat-treated alloys.14 Table I⇓ summarises the important similarities and differences of the ASR and the BHR.
From April 2004 a prospective trial of the ASR was undertaken. The clinical results of the first 200 patients involved in this independent trial have already been published.15 Metal ion analysis has been carried out at our unit on a routine basis since June 2007 for patients under the care of the senior author (AVFN). This paper presents data from patients attending for routine follow-up during the period from June 2007 to June 2008. The results of the first 76 of the 90 ASR patients included in this study have been published previously.7
Owing to the disparity in patient numbers between the ASR and the BHR groups, 12 BHR patients attending review clinics at a second local centre who met the inclusion criteria gave consent to recruitment into this study. Each operating surgeon at this centre is a lower limb arthroplasty specialist. The posterior approach was used in each case. At this centre, all blood samples were obtained by the same orthopaedic registrar involved in the main centre study, using identical equipment and the same laboratory for analysis. For both groups of patients, the only exclusion criteria were the presence of another joint replacement and the resurfacing having been performed within 12 months of blood sampling. Patient demographics are shown in Table II⇓.
Blood samples are obtained via Venflon, with the first 5 ml being discarded before the definitive sample is drawn. All samples are frozen and sent to the same laboratory for blinded whole blood and serum Cr and Co analysis using inductively coupled plasma mass spectrometry.
The same technique of EBRA (Einzel-Bild-Roentgen-Analyse)7,16 (University of Innsbruck, Innsbruck, Austria) analysis of standing radiographs was used to assess the BHR group and the larger ASR patient group. As well as the orientation of the acetabular component, we measured the femoral stem/shaft angle relative to the anatomical axis of the femur as previously described.7 The orientation of the components was assessed and recorded prior to the metal ion results becoming available. For all measured parameters, all technically satisfactory weight-bearing radiographs available for each patient were analysed and the mean values calculated and used in the final analysis. An overview of joint sizes and orientations of the acetabular component is shown in Table II⇑.
Spearman’s rank correlation was used to identify any significant relationships between the independent variables (as above, as well as the Harris Hip Score17 and the University of California, Los Angeles activity score18 at the time of blood sampling, the time from surgery and age) and the dependent variables (whole blood and serum values of Cr and Co) of the ASR and BHR patient groups. Each variable was also plotted individually against the metal ion values in order to identify non-linear relationships.
In order to determine the joint orientations most strongly associated with the lowest blood metal ion levels, the angles of inclination and anteversion were divided into subgroups. Mann-Whitney U tests were then used to identify significant differences (p < 0.05) between these subgroups. Two-sample t-tests were used to compare values from groups with normal distributions. The same methods were used to determine significant differences in independent and dependent variables between the ASR and BHR groups. Windows SPSS version 16.0 (SPSS Inc., Chicago, Illinois) was used for statistical analysis.
An overview of the Cr ion data in whole blood can be seen in the box and whisker plots in Figure 1⇓. There was a highly significant positive correlation between whole blood/serum Cr and Co concentrations (Tables III⇓ and IV⇓), and we therefore feel it appropriate to use Cr and Co interchangeably for graphical representation.
The results from this larger group of patients showed a significant inverse relationship between femoral size and Cr and Co concentrations (Fig. 2⇓ and Tables II⇑ and III⇑). In order to investigate the effect of the interaction between size and orientation of the acetabular component on the ion levels, the ASR patients were split into groups according to individual femoral sizes. A multiple regression model was then constructed for each group size, with Co as the dependent variable and acetabular component inclination and anteversion as the explanatory variables. As illustrated in Figure 3⇓, as the size of the ASR implants increased the corresponding R2 value decreased (i.e. acetabular component orientation had less influence on Co as the femoral size increased). The relationship between acetabular component inclination/anteversion and metal ions was negligible in ASRs with femoral components > 51 mm. For the purposes of this study, ASRs with femoral components > 51 mm are referred to as large ASRs. Below this size there was a significant increase in metal ions when acetabular components were positioned at > 45° of inclination and/or > 20° of anteversion. In this larger series of patients it also became apparent that acetabular component anteversion < 10° was also associated with an increase in metal ion concentrations (Fig. 4⇓). ASRs with femoral components ≤ 51 mm are henceforth referred to as small.
There was a significant inverse relationship between femoral size and metal ion concentrations in the BHR patients (Spearman’s rank correlation (r) = − 0.265, p = 0.038, Cr) (Fig. 2⇑). The same multiple regression model described above was used to examine the interaction between femoral size and orientation of the acetabular component on metal ion levels. As with the ASR, increasing femoral size reduced the effect of orientation of the acetabular component on metal ion concentrations (Fig. 3⇑). Acetabular component orientation had no significant effect on BHRs with femoral components in the fourth quartile of the size range (≥54 mm). For the rest of the paper, the implants in the fourth quartile of the size range are referred to as large and the remaining implants are referred to as small. Table V⇓ shows the significant differences in Cr and Co concentrations between the large and small BHR groups.
Small BHR implants
Inclination of the acetabular component.
We identified no linear correlation between the inclination of the acetabular component and the concentration of Cr or Co ions. However, the largest three Co and Cr readings were found in the blood of patients with acetabular components implanted with > 55° of inclination. For acetabular components placed below 55° (32.5° to 54.9°), a non-significant inverse relationship to Cr and Co was observed (p = 0.08 and 0.120, respectively) (see Fig. 5⇓).
Anteversion of the acetabular component.
There was a significant positive correlation between anteversion of the acetabular component and Co (r = 0.309, p = 0.047). A similar, but slightly weaker relationship was identified between Cr and anteversion (r = 0.274, p = 0.079). The lowest median values of whole blood Cr and Co were associated with acetabular components positioned in 10° to 20° of anteversion (3.46 and 1.44 μg/l, respectively). Acetabular components placed in versions above or beyond these boundaries were associated with higher Co and Cr concentrations (p = 0.004 and 0.063, respectively) (Fig. 4⇑).
Comparison of devices by acetabular component orientation.
Each implant group was divided into sub-groups according to angles of acetabular component inclination and anteversion. Median ion values for each sub group were then calculated and plotted (Figs 4⇑ and 5⇑). Table VI⇓ shows the overall relative effects of suboptimal acetabular component position on the two resurfacing systems using patients from the small ASR and BHR implant groups.
Comparison of devices by gender.
Men were implanted with significantly larger components than the women and acetabular components with significantly larger angles of anteversion in both the ASR and BHR groups (p < 0.001 for both variables).
There were 52 men with ASR and 44 men with BHR. Whole blood Cr was found to be significantly lower in the ASR men than in the BHR patients (p = 0.012), although Co concentrations were comparable (Table VII⇓).
There were 38 women with ASR and 26 with BHR implants. The mean femoral component size was larger in the ASR group, as were the mean angles of inclination and anteversion of the acetabular component. The median Co concentration in the women with ASR was three times that of the median Co in the women with BHR (p = 0.010) (Table VII⇑).
Our results substantiate our previous conclusions7 that generation of metal ions following ASR resurfacing is linked to both the size of the implant and the three-dimensional orientation of the acetabular component. The present, larger group of results provides further evidence of the importance of the anteversion of the acetabular component with increased metal ion loads observed in components with both excessive (> 20°) and insufficient (< 10°) radiological anteversion.
As with the ASR, there was a significant inverse relationship between the size of the femoral component and whole blood Cr in the BHR patients. In the fourth quartile of ASR and BHR implant sizes, the blood metal ion concentrations were significantly lower than those found in the patients with smaller implants. Larger ASR and BHR acetabular components appear more resistant to suboptimal positioning in terms of metal ion generation. This is probably due to the thicker fluid film achieved by the larger implants8 as well as the greater arc of cover. Both of these factors protect against the increased rates of wear associated with edge loading.
A further variable that has been shown to affect lubrication in vitro is clearance. Rieker et al9 showed that wear rates can be reduced by reducing clearance as much as manufacturing processes will allow. The improved lubricating film that the reduced clearance of the ASR theoretically provides may be the reason for the significantly lower whole blood Cr values in the male ASR patients. These clearance values are, however, nominal. When the components are implanted they are vulnerable to the effects of distortion,19 the extent of which is impossible to measure in vivo routinely. By definition, a low-clearance device is at greater risk of accelerated wear should the combined distortion of the components be greater than expected. Whereas the median Cr value of the male ASR patients is lower than that of the male BHR group, the range of the male ASR Cr and Co concentrations is much broader, with several extremely high values.
In order to conserve bone and increase impingement-free range of movement, the ASR acetabular component is sub-hemispherical in design. For the smallest ASR acetabular component, the arc of cover (the subtended angle to the articular surface) is 148°, increasing to 160° for the largest. This is in sharp contrast to the cover provided by the BHR acetabular component with 158° for the smallest, increasing to 166° for the largest acetabular component. The reduced cover provides a potential explanation for the increased vulnerability of the ASR device to angles of inclination greater than only 45°. This also accounts for the increased sensitivity to suboptimal position of small BHRs compared to their larger counterparts. This is consistent with the results of De Haan et al,13 who found a highly significant relationship between the arc of cover of the acetabular component and serum metal ions. The fact that the larger acetabular component of the BHR is associated with lower ion levels suggests that the mechanism of metal ion release is more commonly due to the effects of edge loading rather than impingement.
Anteversion may cause an increase in metal ion levels owing to a combination of factors. First, increased anteversion will reduce component cover, potentially impairing the generation of a sufficient fluid film and causing wear to occur nearer to the rim of the acetabular component. Excessive anteversion may cause femoral neck impingement on the posteroinferior portion of the acetabular component when the hip is extended and externally rotated. The reverse process, with reduced anteversion causing anterior impingement during deep flexion, could explain the slight increase in metal ions observed in near-neutral ASR and BHR acetabular components. However, we believe that this effect is more likely to be explained by posterior edge loading in deep flexion and stair climbing,20 as near-neutral ASR acetabular components are associated with a greater increase in metal ions than the hemispherical acetabular component of the BHR.
Taking the value of 10 μg/l (which is double the United Kingdom workplace exposure limit21) as an arbitrary unacceptable threshold for whole blood Co concentrations, Fisher’s exact test showed a significant difference between the ASR and BHR data groups, with a p-value of 0.014. There were only two BHR patients with Co levels above 10 μg/l compared to 14 in the ASR group. The mean inclination and anteversion angles of these two BHR joints were higher than those of the 14 ASR joints (BHR inclination/anteversion = 59.9°/33.4°; ASR inclination/anteversion = 54.8°/26.9°). The BHR joints were also smaller, with a mean femoral size of 44.0 mm compared to 47.8 mm for the ASRs. This is compatible with the idea of the BHR acetabular component being more forgiving of suboptimal positioning.
A theoretical model to explain the variation in ion levels associated with the two resurfacing systems.
Bergmann et al20 showed that, in the standing position, the average hip joint contact force is directed 14° medially from the longitudinal axis in the anteroposterior plane and 16° anteriorly in the transverse plane. Using these data and the orientations of the acetabular components of the patients in this study, we calculated the distance from the centre of the theoretical contact patch to the superior anterior edge of the articular surface of the acetabular component (or its rim) for each patient in the standing position. We assumed the contact patch to be centred over the hip contact force vector. The resultant distance between the contact patch and the acetabular component rim (CPR distance) is therefore dependent on the diameter and the arc of cover provided by the acetabular component, and also the radiological angles of inclination and anteversion. We found a highly significant inverse correlation between the CPR distance and blood and serum Cr and Co levels in both implant groups, with no large values observed when the contact patch was calculated to be > 10 mm from the rim (Fig. 6⇓). This is consistent with retrieval studies which have demonstrated increased wear secondary to edge-loaded implants. The ‘wear scar’ on retrieved components is commonly located at 10° to 15° from the vertical, consistent with Bergmann’s in vivo joint contact area measurements.22,23 In this series a number of patients had acceptable levels of CrCo even with a calculated CPR distance < 10 mm. However, as the time from surgery to blood sampling increased, so did the probability that patients with a CPR distance < 10 mm were found to have grossly increased blood metal ion concentrations (Fig. 7⇓). This observation may represent the temporary tolerance of the hard metal-on-metal surfaces to a suboptimal position until the joint reaches a state of accelerated wear.22
We believe that the calculated CPR distance is a reliable indicator of the vulnerability of a joint to the effects of anterior edge loading and anterior subluxation. If subluxation does occur we speculate that the effects of the sharp ASR rim are likely to be particularly damaging to the femoral component. CPR distance does not, however, guage the vulnerability of the resurfaced joint to the risk of anterior impingement and posterior edge loading/subluxation. The surgeon must strike a balance between attempting to reduce the negative effects of edge loading by implanting an acetabular component with reduced inclination and anteversion, and attempting to maintain a satisfactory range of movement of the hip. From the data available to us, we feel it reasonable to recommend that the ASR acetabular component to be placed in 40° (sd 5) of inclination and 15° (± 5) of weight-bearing radiological anteversion in order to reduce levels of metal ions in the blood. We have a smaller amount of ion data for the BHR device. However, the metal ion increase in our series only appears above 55°. When considering the optimal inclination for the BHR, the lowest ion values are associated with acetabular components positioned between 45° and 55°, rather than around the lower limits (< 40°). We believe that low inclination angles of near-hemispherical acetabular components with reduced anteversion, coupled with the reduced head to neck ratio inherent to surface arthroplasty, may lead to anterior impingement and posterior edge loading/subluxation during common activities such as stair climbing and rising from a chair.20,24,25 Reduced inclination angles combined with high anteversion may increase the risk of anterior edge wear with or without posterior impingement.26,27 For this reason, we feel that the BHR acetabular component should be placed with a target inclination of 45° in mind, allowing a 5° margin of error on either side and remaining within the limits conducive to a satisfactory outcome in terms of function and metal ion levels. With regard to standing radiological anteversion, we recommend 10° to 20°, with the same justification and allowance for a margin of error. In the senior author’s patient group, the minimum Harris Hip Score of the 14 patients with acetabular components in this target area was 97 (mean score 99.0, mean femoral size 48.0 mm (42 to 54)).
There are two main limitations to this study, the first being the difference in time post-operatively at which blood samples were collected in the two groups. We do, however, believe that this factor is relatively insignificant compared to implant size and orientation. We believe that comparison between groups is justified as the mean time to sample collection in both groups was well beyond the bedding-in phase, where accelerated wear is said to occur.28,29 Our data suggest that suboptimally placed acetabular components will produce increasing amounts of metal debris with time, and therefore if the difference in follow-up between the ASR and BHR groups were truly significant we would expect the BHR metal ion levels to be higher. Furthermore, the Birmingham group themselves have published convincing evidence that in well-positioned joints, metal ion levels tend to decrease over time.30 This supports the idea that the reduced clearance of the ASR is indeed beneficial, given that Cr levels are lower in the male ASR patients despite the shorter follow-up period.
The second limitation is that we did not collect pre-operative blood samples for metal ion analysis, a variable which Vendittoli et al11 showed to correlate with post-operative ion levels. We believe this weakness is offset by the large number of patients and the fact that all those involved in the study came from the same geographical region.
Increased blood Cr amd Co concentrations following the Birmingham Hip Resurfacing procedure are associated with smaller components, acetabular component anteversion > 20° and < 10° and inclination > 55°. The female ASR patients in this series had a median Co level three times that of the female BHR group. The male ASR group had a slight but significantly reduced Cr level compared to the BHR group. The overall results suggest that the reduced clearance of the ASR promotes an improved joint lubrication regimen but this positive effect may be overridden by the increased vulnerability of the ASR acetabular component to suboptimal positioning. The BHR acetabular component is tolerant of a wider range of orientations. We have found no evidence in this study to support the idea that inclination angles < 40° equate to a reduction in metal ion concentrations.
Further information relating to these findings and the measurement of the contact force and contact patch distance is available with the electronic version of this article on our website at www.jbjs.org.uk
Although none of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received but will be directed solely to a research fund, foundation, educational institution, or other nonprofit organisation with which one or more of the authors are associated.
- Received January 15, 2009.
- Accepted June 2, 2009.
- © 2009 British Editorial Society of Bone and Joint Surgery