Effect of counterface material on the characteristics of retrieved uncemented cobalt-chromium and titanium alloy total hip replacements

Transcription

1 191 Effect of counterface material on the characteristics of retrieved uncemented cobalt-chromium and titanium alloy total hip replacements Eur Ing K Brummitt, BSc, CEng, MIM, C S Hardaker, BSc, P J J McCullagh, PhD, CEng, FIMechE, K J Drabu, FRCS and R A Smith DePuy International, Beeston, Leeds A number of total hip components explanted at revision with bearing surfaces in either cobalt-chromium-molybdenum alloy or titanium4% aluminium4% vanadium alloy were examined and compared to contemporaneously manufactured but unused items; particular attention was paid to the bearing surfaces which were examined visually, by low-power microscopy, scanning electron microscopy (SEM), confocal microscopy, white light interferometry, laser profilometry and conventional stylus profilometry. The cobalt alloy heads maintained their surface finish weif over periods up to 12 years. The titanium implants became badly damaged over much shorter periods although even badly scratched heads continued to meet the current standards for titanium alloy heads. Analysis showed that the damage to the titanium alloy heads was not a random but a well-defined process of scarring of a consistent size created by abrasion with small particles of bone. These damaged heads had the potential to wear the matching UHMWPE Components rapidly creating large amounts of polymer debris. The finding that measurement of these damaged heads is within current standards raises concerns as to whether current standards incorporate fully the requirements for clinical performance. Key words: revision, total hip prosthesis, surface characterization, interferometer, titanium alloy, cobalt alloy, wear 1 INTRODUCTION The wear behaviour of the bearing materials in total joint replacement has been an issue since the early days of joint replacement when the rate of wear and debris produced from polytetrafluoroethene (PTFE) hip components was found to be unacceptable (1). It has been found that general wear rates in total hip replacement (THR) are low (2-9, of the order of mm/year, and that THR, with a metal/polyethylene articulation, can be satisfactory over a period of 30 years. It is equally true that there is a limited patient population which shows significantly higher wear rates than this. These higher wear rates have been associated with damage to the metal half of the bearing, roughening the hard articulating surface and thereby inducing more rapid wear of the soft ultra-high molecular weight polyethylene (UHMWPE) part. It has long been known that for an evenly textured surface, the wear rate in metal/uhmwpe bearings is proportional to the surface roughness of the metal part (6). Dowson et al. (7) demonstrated that a single positive feature on the surface can markedly accelerate the wear rate of UHMWPE sliding against metal, The creation and maintenance of a good surface finish on the hard bearing surface of a total joint prosthesis is, therefore, of paramount importance, both in reducing the wear of the joint parts and also in the release of polyethylene debris which has been associated with osteolytic reactions in bony tissue adjacent to the joint and consequent loosening of both femoral and acetabular components. Current British and International standards specify the surface finish requirement for a THR by placing an upper limit on the roughness measured with a mechani- The MS was received on 5 December 1995 and was accepted for publication on 27 February 19%. H06495 Q IMechE 1996 cal stylus profilometer using R, as a quantitive indication of surface quality. The figures quoted are 0.05 pm for a stainless steel or cobalt-chromium component but 0.10 pm R, for titanium alloy components. The longterm performance of a THR depends not only on these parameters having been achieved but also the maintenance of quality throughout the life of the joint. The present work represents a study of the fate of bearings produced in cobalt-chromium-molybdenum alloy and titanium-6% aluminium-4% vanadium alloy following implantation. 2 MATERIALS AND METHODS Forty matching cup and stem components were available from revision operations at the joint replacement unit of the East Surrey Hospital, the majority by one of the authors (KJD). These were explanted over the period September 1989 to December The selection of prostheses from the total available was made to give a balance of titanium alloy and cobalt alloy bearing surfaces but was otherwise random. All the components were of the Ring design in various stages of its evolution, the cobalt alloy stems were of a traditional form with an essentially smooth matt finish while the titanium stems were of a later cementless design comprising a titanium-6% aluminium-4% vanadium stem with pads of coarse commercially pure titanium expanded metal mesh bonded to the proximal anterior and posterior faces. All the acetabular components and the majority of stems were loose at revision, this facilitated removal without further damage to the component. In each case the bearing surface of the stem component was protected prior to surgical removal avoiding damage from metallic surgical instrumentation. Prior to examination, metal parts were sterilized by gamma irradiation and polymer parts by cold chemical sterilization. Three unused titanium alloy prostheses were Proc Instn Mech Engrs Vol 210

2 192 K BRUMMITT, C S HARDAKER, P J I McCULLAGH, K J DRABU AND R A SMITH Table 1 Comparison of R. value with instrument Patient Mechanical Laser identitv Interferometer Drofilometer Drofilometer Worn Ti alloy MB Worn Ti alloy Ed Ba Worn Co-Cr Ka Be Worn Co-Cr AC available for comparative examination together with unused cobalt alloy prostheses. Explanted components had been in place 3-12 years from primary operation. All parts were subjected to an initial visual assessment and study under a low-power binocular microscope over a magnification range 1-4Ox. Notes were made of any physical damage to the components, surface damage, corrosion, abrasion, bone-like adhesion or other change of state. Features of interest, especially the bearing surfaces, were examined at magnifications of up to x using a Cambridge S600 scanning electron microscope fitted with a link 860 energy dispersive X-ray analysis system (EDAX). Surface roughness was measured by a conventional contacting profilometer with a cut-off of 0.8 mm and a stylus tip radius 1.5 pm (Rank Taylor Hobson Tallysurf 4) following the current British Standard. Three measurements were taken in different positions on each head and the mean was reported. A number of samples were also examined using noncontacting methods including laser profilometry (Rodenstock) confocal microscopy where results were reported as the mean of at least three measurements and a limited number of samples were examined by white light interferometry (Zygo). 3 RESULTS Visual assessment of the acetabular components showed two distinct regions similar to those reported by other investigators (8) corresponding to areas of bearing contact (high wear) and occasional or no contact (low wear). These areas had the appearance and features reported frequently enough to be regarded as typical of Fig. 1 Comparison of (a) new and (b) explanted titanium-6% A1-4% V heads using interferometer measurement Part H : Journal of Engineering in Medicine Q IMechE 1996

3 EFFECT OF COLJNTERFACE MATERIAL ON CHARACTERISTICS OF RETRIEVED THRs 193 explanted UHMWPE acetabular components. There were no essential differences in the surface features of acetabular components articulating against different metal counterfaces. The metal bearing surfaces, however, showed considerable differences in appearance. Both cobalt-chromium-molybdenum alloy and titanium-6% aluminium-4% vanadium heads had a similar bright smooth polished appearance when new. The cobalt alloy components retained this appearance when explanted with a small number showing isolated scratches. The titanium components, by contrast, showed a distinct dulling or tarnishing with two zones readily distinguished by the extent of dulling, to which the terms zone 1 and zone 2 were applied. Zone 1 was present on all explanted titanium femoral heads and covered at least 50 per cent of the surface area examined under the low-power microscope; this could be seen as an area of fine multidirectional scratches of some depth and there was an apparent correspondence of the scratch directions with the arc of hip movement expected in a walking cycle. Zone 2 showed much finer scratches in the periphery of the head without a distinct pattern. Surface finish measurements on cobalt alloy heads showed an essentially unchanged surface roughness when assessed by R, using a conventional stylus profilometer. When the titanium heads were measured, distinct differences were found between the new prostheses and explanted heads; this was especially so in the zone 1 area. However, the measured roughness surprisingly remained generally within the current standards. The Rodenstock laser profilometer has essentially a zero radius probe and invariably produces a higher R, value for any given head as can be seen from Table 1. The quantitative parameters obtained from this technology, however, were not essentially helpful to understanding the wear processes compared to the use of a stylus instrument since the effect was simply to multiply the R, by an approximately constant factor. The white light interferometer gave roughness readings estimated over a small area which yielded R, values much closer to those obtained by conventional measurement, but the advantage of this instrument was the powerful analytical facility which enabled the three-dimensional surface data to be broken down by feature size (Figs 1 and 2). This analysis on new heads of either metal showed a tight Gaussian distribution of scratches created by controlled polishing with tightly graded abrasive media. Explanted heads in titanium alloy again showed the central peak from the original polishing but also showed two additional discrete peaks representing the surface damage centred nm below and above the polishing peak. Cobalt alloy heads did not show any additional peaks but did show a broadening of the peak and a skewing of the distribution to create a tail of higher surface features. Scanning electron microscopy was used to examine the bearing surfaces. This confirmed the overall impression that the surface of the titanium components had suffered extensive scratching especially in the active bearing area and that these were both significantly more frequent and deeper than seen in corresponding heads in cobalt alloy which showed only occasional scarring and some fine pitting associated with eutectic carbide. In both cases chips of bone in the size range 1-5 pm could be observed on both acetabular and femoral components, these were much less frequent on cobalt alloy components than on titanium alloy components. As in an earlier paper (9) there was strong evidence of bone chips scarring the surface of titanium components by cutting grooves approximately the width of the chip. This was not observed with cobalt alloy parts ) 1m lmm 2 8ooo 0 6ooo am 2ooo X.ZS dl S s Displacement from nominal surface (nm) (a) New cobalt-chromium-molybdenum alloy head Displacement from nominal surface (nm) (b) Explanted cobalt-chromiunwnolybdenum alloy head Fig. 2 Comparison of (a) new and (b) explanted cobalt alloy heads using interferometer measurement (6 IMcchE 1996 Proc Instn Mech Engrs Vol210

4 194 K BRUMMITT, C S HARDAKER, P J J McCULLAGH, K J DRABU AND R A SMITH 4 DISCUSSION In the past the emphasis has been placed by some authors on wear as a mechanism affecting the function of the joint by restricting articulation (10) and thus limiting the useful lifetime, indeed one motivation for the original development of metal-backed components was to enable an easy interchange of worn liners in revision surgery (1 1). Over the last decade, however, the focus of concern has gradually shifted towards the effect of the polymer wear debris on the host tissue and especially the possibility of implant loosening being the result of polymer debris induced osteolysis. It has been demonstrated that, within the range of surface roughnesses used for total joint replacements, wear of a sliding UHMWPE/ metal couple is directly proportional to the surface roughness of the metal part (6, 12). The surface finish of the femoral component and the maintenance of that finish is therefore of paramount importance. It is generally held that titanium, with its two phase structure, is harder to polish than the other alloys used for bearing surfaces and this, rather than the functional requirement, is reflected in the current standards. Explanted cobalt alloy parts generally had an as new finish with occasional quite shallow scratches which appeared more prominent to the eye against a polished surface than they proved to be when examined analytically. Titanium alloy heads, despite having a shorter average implantation time, were invariably badly scratched. This can be attributed in part to the alloy being softer and its oxide more friable and easily removed. The large amount of bone debris present which was not observed in the cobalt prostheses may, however, have contributed to this finding. The tightly defined size range of damage to the head observed by optical interferometry was surprising and demands a source of hard (bone) particles of consistent size. These were present but the source was unclear. In a previous paper (9), it was shown that cups of this design were almost invariably loose and produced debris in their motion against the acetabular bone; also many of the titanium stems were loose at revision, as were virtually all the cups. It is probable therefore that the movement of the loose prostheses against bone is the source, if this is the case then a further question to be addressed is whether the prosthesis design actively contributes to wear performance. With the loose stems, the expanded mesh pads might be a contributor in this case with their rasp-like surface. However, these would be expected to cut a larger chip in the first instance; possibly larger particles are produced which have to be crushed to less than 5 pm before they can readily migrate. In view of the small particle size, the source may simply be motion of the general prosthesis surfaces against the bone. The head of a femoral prosthesis would commonly be polished with abrasive with a modal particle size of 1 pm or less. The difference between these features and the larger new features generated in vivo is consistent with the size of the bone chips observed. The damage to the cobalt alloy heads is not only less severe but also has produced damage much closer to the existing texture of the polished surface. This situation is clearly preferable so far as the wear of the UHMWPE cup is concerned. The increased Part H: Journal of Engineering in Medicine hardness of cobalt-chromium alloy compared to titanium alloy may be the explanation for the finer scratches. The smaller number of measured scratches might be due to the difficulty in distinguishing new scratches of this size from the surface. However, less bone chips were noted at post revision examination. This may indicate that there is a contribution of the production of bone fragments from stem movement with a loose Timesha prosthesis. It is also possible that these are created by a different third body; other authors have identified various sources of third body damage (13-15), however, in this study the authors were unable to detect any other source in the bearings of these joints despite actively searching for the relevant elements by EDAX mapping. It is clear from the study of these retrieved heads that the current approach to standardization of surface finish of total hip prosthesis is flawed, either in test methodology or numerical description of surfaces. Some of the explanted titanium heads in this study were scratched to the extent that they no longer had the appearance of being polished. Despite this, the increase in numerical surface roughness was insufficient to take the parts outside the current standards for such bearings. Although it is doubtful whether any surgeon would implant a bearing with this appearance, the upper limit of acceptable roughness for titanium alloy surfaces in standards should be reviewed. Reducing this to the figure of 0.05 p R, in line with cobalt components would seem sensible. The method of calculating surface roughness, however, remains a problem for this application and is well illustrated by these results. R, is well suited to the description of an even texture, however, the texture of the damaged surfaces is no longer uniform and the deeper scratches are lost in the averaging process. Dowson et al. (7) demonstrated that a single scratch on a polished surface can cause significant increases in wear rate. Conventional profilometers are approaching the limits of their resolution with prosthetic bearing surface which may have typically p R, for metal components and 0.01 p R, for ceramic parts. Limitations are created by the noise associated with the necessary mechanical device, by the radius of the stylus tip and especially in softer metals by the stylus cutting its own groove (scratch) and smoothing out delicate features. The use of R, as a measurement parameter must also be questioned since even with short cut-off lengths, a single scratch or small group of scratches which might significantly increase wear (7) will be diluted in the averaging process. It also seems reasonable to accept that Dowson et al. (7) are correct in stating that negative features are much less important than positive ones, yet R, gives equal weight to both. Ideally, standards should be based around non-contacting methods with a measure giving weight to the largest positive features on an area of surface. Some authors have suggested the use of R, which is the total vertical displacement in the measured section, but although this would be an improvement, this should be regarded as an interim step. 5 CONCLUSIONS Several authors, such as Agins et al. (16), have suggested that untreated titanium alloy does not make a IMechE 1996

5 EFFECT OF COUNTERFACE MATERIAL ON CHARACTERISTICS OF RETRIEVED THRs 195 bearing surface with UHMWPE; the present work supports the view that as a bearing material it should be reserved for cases of metal sensitivity when its use is unavoidable. The third body damage to implanted metal surfaces was not a random process but systematic and capable of close definition. The softer titanium alloy is an ideal model material for the experimental study of these processes, but it is inferior in clinical practice. It is possible that prosthesis design can affect wear rate by the production of third body particles. The current British and IS0 Standards do not currently define surface finish requirements in a way which completely reflects clinical requirements, in particular the use of R, as a specifying parameter should be reconsidered. REFERENCES Chnmley, J. Low friction arthroplasty of the hip, 1979, pp. 6-7 (Springer-Verlag). charaley, J. and Cupic, Z. The nine and ten year results of low friction arthroplasty of the hip. Clin. Orthop. Rel. Res., 1973, 95, Wroblewski, B. M. Direction and rate of socket wear in Charnley low friction arthroplasty. J. Bone Jt Surg., 1985,67B, GrEth, M. J., Scide.ostein, M. K, Willipms, D. and Chunley, J. Socket wear in Charnley low friction arthroplasty of the hip. Clin Orthop. Rel. Rex, 1978, 137, Joshi, A. B., Mnrkovic, L., Noble, P. C and Murphy, J. C. Long term wear analysis of the acetabular cup, analysis of risk factors. 21st Society of Biomaterials, Trans. SOC. Biomater., Vol. XVIII, 1995, p. 44. Dowson, D., Challen, J. M., Holmes, K. and Atkinson, J. R. The influence of counterface roughness on the wear rate of polyethylene. In The wear of non metallic materials (Us D. Dowson, M. Godet and C. M. Taylor) 1978, pp Proceedings of the 3rd Leeds-Lyon Symposium on Tribology, Leeds, 1976 (Mechanical Engineering Publications, London). 7 Dowson, D, Taberi, S. and Wdbridge, N. C. The role of counterface imperfections in the wear of polyethylene. Wear, 1987, 119, Atkinson, J. R., Dowson, D, Jsaac, G. H. and WroMewski, B. M. Laboratory wear tests and clinical observations of the penetration of femoral heads into acetabular cups in total hip joints 11. A microsurgical study of the surfaces of Charnley polyethylene acetabular sockets. Wear, 1985,104(3), Drabu, K., Michrud, R. J, McCulhgh, P. J. J., Bruinmitt, K. and Smith, R A. Assessment of titanium alloy on polyethylene bearing surfaces in retrieved uncemented total hip replacements. Proc. Instn Mech. Engrs, Part H, l994,208(h2), Wroblewski, B. M. Direction and rate of socket wear in Charnley low friction arthroplasty. J. Bone Jt Surg., 1985,67B, Mel, D. L, Wright, T. M. and Edwards, P. The effect of metal backing on stresses in polyethylene acetabular components. In The hip (Ed. D. S. Hungerford), Proceedings of the 11th Meeting of the Hip Society, 1983, pp (C V Mosby and Company). 12 Weightman, B. and Light, D. The effect of surface finish of alumina and stainless steel on the wear rate of UHMW polyethylene. Biomaterials, 1986,7, Atkinson, J. R., Isaac, G. H., Dowson, D. and Wroblewski, B. M. Laboratory wear tests and clinical observations of the penetration of femoral heads into acetabular cups in total joint replacements 111. The measurement of internal volume changes in explanted Charnley sockets after 2-16 years in viuo and the determination of wear factors. Wear, 1985,104, Cnravin, L., Dowson, D., Fisher, J. and Jobbins, B. The influence of bone and bone cement debris on the counterface roughness in sliding wear tests of ultra-high molecular weight polyethylene on stainless steel. Prm. Instn Mech. Engrs, Part H, 1990, 204(H1), Isaac, G. H., Atkinson, J. R., Dowson, D., Kennedy, P. D. and Smith, M. R The causes of head roughening in explanted Charnley hip prostheses. Engng in Medicine, 1987, Agins, H. J, Akock, N. W, Bamal, M., Sahti, E, Wilson, P. D., Pellki, P. M. and Bdlough, P. G. Metallic wear in failed titanium alloy total hip replacement. J. Bone Jt Surg., 1988,70A, IMechE 1996 Roc Iosto Mcch Eogrs Vol210