The following studies have demonstrated that the theoretical expectation of a low polyethylene wear rate has been fulfilled in practice.
Retrieval studies
Twenty-three bearings were retrieved from 18 failed bicompartmental Oxford arthroplasties, 1–9 years after implantation (Argenson and O’Connor, 1992). The minimum thickness of each was measured with a dial gauge (Fig. 2.10) and compared with the mean thickness of 25 unused bearings. The mean penetration rate was very low; calculated by two methods, it was either 0.043 or 0.026 mm/year. There was no correlation between the initial minimum thickness of the bearings (range 3.5–10.5 mm) and their rate of wear.
Figure 2.10 Dial gauge used to measure the thickness of a bearing at the bottom of the spherical socket.
Figure 2.11 Retrieval studies. Measured penetration into the bearing plotted against duration of implantation. Bearings without impingement shown in blue and with impingement shown in red. The vertical line through each data point represents one standard deviation of the mean of the measured thickness of unused bearings obtained direct from the manufacturer. The regression lines through the data give the penetration rates (mm/year). (Reproduced with permission and copyright © of the British Editorial Society of Bone and Joint Surgery. [Psychoyios V, Crawford RW, O’Connor JJ, Murray DW. Wear of congruent meniscal bearings in unicompartmental knee arthroplasty. A retrieval study of 16 specimens. J Bone Joint Surg [Br] 1998; 80-B: 876–82].)
The same method was used to study 16 bearings retrieved from failed OUKA (Phase 2) medial arthroplasties, 0.8–12.8 years after implantation (Psychoyios et al., 1998). The mean penetration rate was 0.036 mm/year (maximum 0.08 mm/year). Again, there was no correlation between the rate of wear and the initial bearing thickness (range 3.5–11.5 mm).
Ten bearings had erosion of their non-articular surfaces, caused by impingement against bone or cement. The most common site was anterior, produced by impingement in extension against bone in front of the femoral component (Fig. 2.12). The six bearings without impingement had a mean penetration rate of 0.01 mm/year compared with 0.054 mm/year for the 10 bearings with impingement (P < 0.0001) (Fig. 2.11). It was observed that there was a strong correlation between impingement and mechanical causes of failure such as loosening or lateral OA, suggesting that impingement may cause failure. It was also found that, in three quarters of cases revised for pain with no mechanical problem, the pain did not improve.
Figure 2.12 Retrieved bearing showing damage due to anterior impingement. (Reproduced with permission and copyright © of the British Editorial Society of Bone and Joint Surgery [Psychoyios et al., 1998].)
Kendrick et al. (2010) used the same method to study a further 47 Phase 1 and Phase 2 bearings retrieved after OUKA at a mean time to revision of 8.4 years (SD 4.1). Twenty had been implanted for more than 10 years (maximum 17 years). Thirty-one of the 47 bearings showed evidence of impingement, and the mean penetration rate in these was 0.07 mm/year. The rate for the 16 bearings without impingement was 0.01 mm/year, the same as that found by Psychoyios et al. (1998). The penetration rate of Phase 1 bearings (machined from blocks of Hostulen RCH1000 polyethylene) was about double that of Phase 2 bearings (individually compression moulded from Montel Hifax 1900H powder). However, the impingement rate in Phase 1 implants (91%) was also much higher than in Phase 2 implants (58%).
Kendrick et al. (2010) further stratified the impinged bearings into a group showing evidence only of extra-articular impingement damage and those also showing intra-articular surface damage (Fig. 2.13). Those showing intra-articular impingement had a penetration rate 2.5 times that of the group with extra-articular surface damage alone, while the latter had a penetration rate five times higher than those (0.01 mm/year) free of impingement damage.
Figure 2.13. The penetration (mm) for different subgroups: no impingement; and a normal articular surface; abnormal macroscopic wear and normal articular surfaces with extra-articular impingement; abnormal macroscopic wear and abnormal articular surfaces with intra-articular impingement ± signs of non-congruous articulation (mm/year). (Reproduced with permission and copyright © of the British Editorial Society of Bone and Joint Surgery [Kendrick BJ, Longino D, Pandit H, Svard U, Gill HS, Dodd CA, Murray DW, Price AJ. Polyethylene wear in Oxford unicompartmental knee replacement: a retrieval study of 47 bearings. J Bone Joint Surg ]Br] 2010; 92(3): 367-73].)
In vivo penetration studies using Röntgen stereometric analysis
We have developed a method of measuring wear in vivo using Röntgen stereometric analysis (RSA), and have applied it to patients following OUKA (Price et al., 2005; Kendrick et al., 2015). The method does not require markers attached to the components or implanted in the patient’s bones and therefore can be used retrospectively. Penetration of the bearings was measured in eight controls (three weeks after OUKA) and in seven patients in whom the prosthesis had been implanted about 10 years previously (Price et al., 2005). The mean penetration for the control group was 0.1 mm, demonstrating the accuracy of the method. The mean penetration rate for the 10-year group was 0.02 mm/year, similar to that observed in retrieved bearings without impingement (0.01 mm/year).
Kendrick et al. applied the same RSA technique to 13 knees in nine patients treated by JWG for AMOA (Kendrick et al., 2011; Murray et al., 1998). The mean follow-up at the time of examination was 20.9 years (range 17.2 to 25.9). The range of penetration rates for the six Phase 1 knees in the cohort was 0.023 to 0.099 mm/year (Fig 2.14), whereas that for the seven Phase 2 knees was 0.016 to 0.030 mm/year, with a mean value of 0.022 mm/year, a value similar to those observed in non-impinging bearings in the retrieval studies. We infer that the Phase 2 design was more successful in minimising impingement and its effects and that any oxidation of the polymer after 20 years in vivo had not accelerated the wear rate.
Figure 2.14. Scatterplot showing the distribution of rates of linear wear by Phase. The horizontal dashed lines indicate the mean wear rate for each phase (Kendrick et al., 2011).
Wear simulation studies
Retrieval and in vivo studies are the ‘gold standard’, providing evidence of actual performance of components in the infinitely varied circumstances of real life. Simulation studies have the disadvantage that the tests may be so far from lifelike as to invalidate the results. However, they are useful for longitudinal measurement of wear, for comparison of competing designs and materials under similarly controlled conditions, and for simulating the effects of years of natural wear in a few months.
A Stanmore knee simulator (Walker et al., 1997) was used to test a group of OUKA bearings over 3 million cycles (Scott and Schroeder, 20004). During the first million cycles (thought to represent one year of normal activity), the bearings had a measured penetration of 0.05 mm (Fig. 2.15). Thereafter, the penetration rate was steady at 0.019 mm/year. The early higher rate of penetration was attributed to creep of the viscoelastic polyethylene which ceased once the bearings had bedded in after about a million cycles. The wear rate thereafter was similar to that observed in non-impinging retrieved bearings and in the in vivo studies of Phase 2 bearings.
Figure 2.15 Simulator studies. Measured penetration plotted against millions of cycles of movement and load, with a regression line through the >1 million cycle data. (Reproduced, with permission, from Scott R, Schroeder D. Correlation of knee simulation to in-vivo use: Evaluating the Oxford Unicompartmental knee. Transactions of the Orthopaedic Research Society vol. 25, Orlando, Florida, 2000; 434.)
Finite-element analysis
Morra and Greenwald (2003) carried out a finite-element stress analysis of four unicompartmental prostheses, two fixed-bearing and two mobile-bearing (the OUKA and an implant with a polycentric femoral component). The shapes of the surfaces of the components were determined using a coordinate measuring machine. They modelled three instants in the normal walking cycle with the knee near extension, and calculated surface contact areas and pressures and the maximum value of the von Mises stress, said to be a measure of the tendency for delamination.
As expected, the fixed-bearing designs both had small contact areas, high contact stresses, and von Mises stresses significantly in excess of the material damage threshold (9 MPa). Both mobile-bearing knees had contact areas at least three times larger than the fixed-bearing designs but smaller than their nominal values because of manufacturing tolerances. The contact areas of the OUKA implant varied from 284 to 346 mm2, compared with the nominal value of 580 mm2 for ideally shaped components. Both mobile-bearing designs exhibited ‘… very low contact stress and an absence of von Mises stress above the material damage threshold’ (Fig. 2.16(b)). No calculations were performed for knees flexed to 90°–100°, when the loads can be larger and the contact areas of the polycentric mobile bearing prosthesis would have been smaller than those of the congruent OUKA.
Figure 2.16 Contact stresses (a) fixed flat bearing, (b) mobile fully congruous bearing (Morra & Greenwald, 2003).
Discussion
The gradual diminution of bearing thickness over time is due to a combination of creep (bulk cold-flow of the viscoelastic polymer) and loss of material due to wear at its two articular surfaces. Penetration measurements of retrieved bearings do not allow an estimate of the relative contributions of these two processes, but the simulator studies suggest that creep only occurs initially and that, thereafter, penetration can be attributed entirely to surface wear.