INTRODUCTION

The clinical use of an all-polymer knee which artic-ulated a polyacetal femoral component against anultra high molecular weight polyethylene (UHMW-PE) tibial component has been reported (1). A“polyacetal group” of 63 total knee replacementswere followed for at least ten years and no instancesof femoral component fracture or failure due towear occurred (1). Such results are remarkable foran all-polymer prosthesis in such a heavily loadedjoint as the knee. In addition, an earlier paper re-ported an in vitro comparison between all-polymerand conventional metal-on-polymer hip prostheses.Using a hip simulator it was found that hip pros-theses having a polyacetal ball articulating with anUHMWPE cup showed 23% less wear than when a

cobalt chrome ball was used with an UHMWPE ac-etabular component (2). It is widely accepted that the wear of prostheticcomponents can eventually lead to their failurethrough a negative cascade of reactions caused bythe body’s response to the wear debris, eventuallyleading to osteolysis around the implant (3, 4).Therefore it is crucial both to understand and tominimise the wear processes taking place in vivo.Part of this learning process involves reproducingin vitro the wear factors occurring in vivo. Recently a wear screening device was describedwhich reproduced in vitro the clinical wear rates re-ported for three biopolymers (UHMWPE, polyac-etal and poly tetra fluoro ethylene) which havebeen employed as the acetabular cup material inhip prostheses (5). Given this validated pin-on-plate

Journal of Applied Biomaterials & Biomechanics 2005; Vol. 3 no. 3: 141-146

The wear of two orthopaedic biopolymers against each other

T.J. JOYCE

School of Mechanical and Systems Engineering, University of Newcastle upon Tyne, Newcastle upon Tyne - UK

ABSTRACT: The potential for all-polymer prostheses has not been widely investigated. It might be expected that the wear of suchbiomaterial combinations would be excessive, but an in vivo study of all polymer knee prostheses reported that there were no fail-ures due to wear, even after ten years of clinical use. This design of knee prosthesis used polyacetal and ultra high molecularweight polyethylene (UHMWPE) as the biopolymers. Similarly, an earlier in vitro study of polyacetal and UHMWPE hip pros-theses indicated lower wear than for a cobalt chrome and UHMWPE combination. Therefore this study set out to test the poly-acetal and UHMWPE combination in a wear screening rig which had previously been validated against clinical data for ar-tificial hip joints. Two different motion conditions were applied to the test samples and each biopolymer was tested as both pinand plate. Interestingly it was found that, whatever the contribution from pin or plate, the total mean wear factors were 1.5 �

10–6mm3/Nm under reciprocation-only, and 4.1 � 10–6mm3/Nm under multi-directional motion. These wear factors weregreater than those found when a conventional metal-on-UHMWPE couple was tested under the same loading, motion and lu-bricant conditions. A comparison was also undertaken with the wear of other orthopaedic biopolymer combinations, namelycross-linked polyethylene (XLPE) against itself, and UHMWPE against itself. The XLPE pairing showed somewhat lower wearthan the polyacetal and UHMWPE couple, while the UHMWPE pairing showed the highest wear of all, approximately an or-der of magnitude greater than the polyacetal and UHMWPE combination. (Journal of Applied Biomaterials & Biomechanics2005; 3: 141-6)

KEY WORDS: Wear, UHMWPE, Polyacetal, Biopolymer, XLPE

Received 20/09/05; Revised 14/11/05; Accepted 17/11/05

1722-6899/141-06$15.00/0© Società Italiana Biomateriali

rig, the objective of the work reported here was toundertake wear tests of polyacetal against UHMW-PE. The aims were to measure the wear factors; toinvestigate whether the orientation of the polymer-ic material, as pin or plate, would influence thewear factors found; and to compare the wear fac-tors with those of a “standard” biomaterial wearcouple of UHMWPE and stainless steel. If thewear of an all-polymer prosthesis is less than thatof a conventional metal-on-polymer prosthesis,then this could move the design of prostheses ina new direction and the cost of total joint re-placements could be markedly reduced.

MATERIALS AND METHODS

The polyacetal and UHMWPE couples were testedusing a modified, four-station, pin-on-plate weartest rig (6). The rig is shown schematically in Fig-ure 1. The modification entailed the addition ofrotational motion to the test pins, in addition tothe standard reciprocating motion, to give multi-directional motion. Such a combined motion re-sulted in each point on the wear face of the testpins following elliptical or quasi-elliptical weartracks, similar to those motions seen on implantedhip prostheses (7). In the wear tests reported here,two stations had reciprocation-only and two ap-plied multi-directional motion. Investigating theinfluence of both types of motion permitted afuller tribological investigation to be undertaken.A load of 40N was employed and reciprocating androtating speeds of 1Hz were chosen. The lubricantconsisted of 25% bovine calf serum and 75% dis-tilled water, to which was added 0.1% sodiumazide. During testing the lubricant was heated to37 °C and it was changed at approximately 60 hourintervals. A standardised cleaning and weighingprotocol was followed, and the pins and plateswere weighed on a balance sensitive to 0.1mg. Con-trol pins and control plates were included to ac-count for any weight change due to lubricant up-take. The wear factors (k) were calculated usingthe wear equation, i.e. by dividing the correctedvolume lost by the product of the load and the slid-ing distance. The density of UHMWPE was takento be 950 kg/m3 and that of polyacetal as 1410kg/m3. Tests were undertaken with UHMWPEplates and polyacetal pins, then with polyacetalplates and UHMWPE pins. After testing had beencompleted, the UHMWPE test plates were washedin acetone, gold-coated in an Emitech K550 coaterand examined in a Hitachi S-4700 Scanning Elec-tron Microscope (SEM).

RESULTS

After an average of 1.4 million cycles of sliding,equivalent to 48 km, the mean wear factors calcu-lated for the UHMWPE and polyacetal pins andplates are given in Table I. As can be seen, thebiopolymer wear factors depended on both the ori-entation of the material, whether it was a pin or aplate, and the motion it was subjected to. However,the mean wear factor for each biopolymer and foreach component, i.e. pin or plate, was less underreciprocation-only than when multi-directional mo-tion was applied to it. Under reciprocation-only thepolyacetal components showed the greater contri-bution to the total wear factor measured, but undermulti-directional motion it was the UHMWPE com-ponents which contributed the greater wear. Themean wear of the UHMWPE plates and that of thepolyacetal pins is shown in Figures 2 and 3 respec-tively. The increase in weight of the polyacetal controlcomponents due to lubricant uptake was manytimes that of the UHMWPE components. For ex-ample the UHMWPE control plate showed an in-crease of 0.2 mg compared with 33.4 mg for thepolyacetal control plate. This greater lubricant up-take by the polyacetal components can in part ex-plain the fluctuations seen in the graphical data inFigure 3. In contrast, the UHMWPE componentsshowed less lubricant uptake and so the wear datashown in Figure 2 appears far more linear. FromFigure 2 it can be seen that UHMWPE wear oc-curred at a relatively steady rate and there was nosignificant “bedding-in” wear. The topography of the worn pins and plates was re-lated to the motion they had been subjected to.The pins and plates under multi-directional motionhad a much more uniform surface finish than thosesubject to reciprocation-only. These latter samples

142

The wear of two orthopaedic biopolymers against each other

TABLE I - MEAN WEAR FACTORS (K � 10–6mm3/Nm) OF

POLYACETAL AND UHMWPE BIOPOLYMERS

Material k reciprocation- k multi-only directional

Polyacetal plate 1.4 1.7

UHMWPE pin 0.1 2.4

Total 1.5 4.1

Polyacetal pin 0.8 1.4

UHMWPE plate 0.7 2.7

Total 1.5 4.1

had a macroscopic topography which appeared asparallel scratches in the direction of sliding. SEMimages of the wear track of the UHMWPE plates aregiven in Figures 4 to 7. Figures 4 and 5 show imagesfrom the plates subject to reciprocation-only, whileFigures 6 and 7 show images from the plates undermulti-directional motion. Figure 4 shows the weartrack of an UHMWPE plate subject to reciproca-tion-only at 30� magnification, and the contrast be-tween the unworn area of the plate on the upperright hand side with the worn area below is clear.On the wear track itself, the parallel scratches in thedirection of sliding, approximately from left toright, can also be discerned. Figure 5 focuses onone of these “scratches” to reveal that at 1000� mag-nification they appear to consist of irregular inden-tations perpendicular to the direction of motion,with a width across the wear track of approximately10 µm. Also shown is a wear particle or piece of de-bris, approximately 30 µm long. Figure 6 shows thewear track of an UHMWPE plate subject to multi-di-

rectional motion at 30� magnification, and againthe contrast between the unworn area of the plate,this time on the bottom right hand side with theworn area above, is clear. Figure 7 focuses on part

143

Joyce

Fig. 1 - Schematic diagram of the modified pin-on-plate rig. Fig. 2 - Mean wear of UHMWPE plates (when rubbing againstpolyacetal pins).

Fig. 3 - Mean wear of polyacetal pins (when rubbing againstUHMWPE plates).

Fig. 4 - SEM image of UHMWPE plate under reciprocation-only(30� magnification).

Fig. 5 - SEM image of UHMWPE plate under reciprocation-only(1000� magnification).

Volu

me

loss

(m

m3 )

Volu

me

loss

(m

m3 )

of the wear track at a higher magnification to reveallong indentations in several different directions. Ascan be seen these indentations also cross each oth-er and help to show that the wear track of the platewas subject to multi-directional motion. A particlecan also be seen which appears to have producedone of these indentations.

DISCUSSION

Using the same wear screening rig, and the same lu-bricant, loading and motion conditions, the wearfactors for UHMWPE articulating against stainlesssteel were measured to be 0.1 � 10–6mm3/Nm un-der reciprocation-only and 1.1 � 10–6mm3/Nm un-der multi-directional motion (5). At a total wear forthe polyacetal and UHMWPE combination of 4.1 �

10–6mm3/Nm under multi-directional motion and1.5 � 10–6mm3/Nm under reciprocation-only, these

all-biopolymer values are clearly larger than thosefor the standard metal-on-polymer combinationrepresented by stainless steel rubbing againstUHMWPE. However, the polyacetal rubbing against UHMWPEwear factors were not excessively high and were lessunder reciprocation-only. How much multi-direc-tional motion, or cross-shear, it is appropriate toapply to a wear simulation of an artificial knee jointis worth further consideration. Wang et al statedthat the motion in the knee joint is much more lin-ear than in the hip, and suggested that cross-shearmotion is a maximum of 10° in the knee joint,compared with 90° for the hip joint (8). A recentpaper offered maximum values of 4° and 9° of bi-di-rectional crossing motion for gait and stair-climb-ing respectively in relation to an artificial knee joint(9). Furthermore, the ASTM standard relating tothe in vitro wear testing of biopolymers includes adelamination wear application, such as has been

144

The wear of two orthopaedic biopolymers against each other

Fig. 6 - SEM image of UHMWPE plate under multi-directionalmotion (30� magnification).

Fig. 7 - SEM image of UHMWPE plate under multi-directionalmotion (300� magnification).

Fig. 8 - Comparison of wear of all-biopolymer couples under rec-iprocation-only.

Fig. 9 - Comparison of wear of all-biopolymer couples under mul-ti-directional motion.

Wea

r fa

cto

r (x

10–6

mm

3 /N

m)

Wea

r fa

cto

r (x

10–6

mm

3 /N

m)

observed in polymer tibial components of kneeprostheses, which has no cross-shear motion (10).In addition, retrieval studies of artificial knee pros-theses mention that parallel scratching in the di-rection of the sliding motion was the principal dam-age observed (11) and that parallel scratches in thedirection of the articulation were predominant(12). Therefore, further work may be useful in thisarea, for example reducing the speed of the rota-tional motors on the test rig, so that less cross-shearmotion is produced at the pin/plate interface. Onechallenge to all of this work is that, from a review ofthe literature, it appears that to date no in vivo wearfactors have been offered for total knee prostheses,an impediment which has been noted previously(13). One paper reported on the in vitro wear of polyac-etal against UHMWPE prostheses (2). Here, 41 mmdiameter polyacetal femoral heads were articulatedagainst UHMWPE acetabular cups in a hip simula-tor which employed bovine serum as the lubricant.Testing ran to over one million cycles, after which atotal wear of approximately 80 mm3/million cycleswas reported for polyacetal against UHMWPE. Ithas been suggested that the average sliding dis-tance per cycle in an artificial hip joint can be cal-culated from 0.67 times the diameter of the femoralhead (14). Therefore, in the 41 mm diameter hipjoints used by McKellop the average sliding dis-tance would be 27.5 mm per cycle. It has also beensuggested that an equivalent static load of 1000Napplies across an artificial hip joint (15). Takingthese three values of wear volume, sliding distanceand load, and applying the wear equation, suggeststhat the wear factor for the polyacetal and UHMW-PE hip prosthesis was 2.9 � 10–6mm3/Nm. This isslightly lower than the value of 4.1 � 10–6mm3/Nmfor the polyacetal and UHMWPE couple undermulti-directional motion reported here, but it is ofa similar order. The McKellop et al paper noted that the weight lossof the polyacetal components was more erratic thanthat of the UHMWPE components. This commentreflects the experience with the polyacetal andUHMWPE components used here, as can be seenby a comparison of Figure 2 and Figure 3. Unfortu-nately the McKellop et al paper (2) did not specifythe weight increase of the control polyacetal com-ponents. Such information would have provideduseful comparative data for the work reported here.If pin-on-plate tests involving polyacetal were to berepeated, then it is recommended that a greaternumber of control components be used so that theinfluence of lubricant uptake on the calculation ofthe wear factors can be minimised.

The wear testing of other all-polymer couples usingthe same pin-on-plate rig, and with the same testconditions in terms of loading, speed and lubricanthas been reported recently (16). The materialsused were UHMWPE rubbing against itself, andsilane cross-linked polyethylene (XLPE) rubbingagainst itself. UHMWPE was chosen as it is the mostwidely used orthopaedic biopolymer. Silane XLPEwas chosen as it has been used in an all-polymer fin-ger joint (17) and acetabular cups manufacturedfrom this material have been successfully used inconjunction with alumina femoral heads since thelate 1980s (18). A comparison of the mean overallwear factors of the various biopolymer couples isshown in Figures 8 and 9. These figures also showthe mean contribution from pin and plate sepa-rately. Figure 8 shows the wear of biopolymer cou-ples under reciprocation-only. The total wear factorfor XLPE against itself is 0.6 � 10–6mm3/Nm, forthe polyacetal and UHMWPE combinations it is 1.5� 10–6mm3/Nm, and for the all UHMWPE couple itis 34 � 10–6mm3/Nm. In three of the four cases thebulk of the wear originates from the plate ratherthan the pin. Figure 9 shows the wear of the fourbiopolymer couples when subjected to multi-direc-tional motion. The total wear factor for XLPEagainst itself is 3.4 � 10–6mm3/Nm, for the polyac-etal and UHMWPE combinations it is 4.1 � 10–6mm3/Nm, and for the all-UHMWPE couple it is 41 �

10–6mm3/Nm.

ACKNOWLEDGEMENTS

The author would like to gratefully acknowledge the as-sistance of Dean Riddell and Professor Tony Unsworth atthe School of Engineering at the University of Durham,UK. The SEM was used at the National Centre for Bio-medical Engineering Sciences at the National Universityof Ireland, Galway, and its use is greatly appreciated.

Address for correspondence:T.J. Joyce, PhDSchool of Mechanical and Systems EngineeringStephenson BuildingUniversity of Newcastle upon TyneClaremont RoadNewcastle upon TyneNE1 7RU - UKt.j.joyce@ncl.ac.uk

145

Joyce

REFERENCES

1. Moore DJ, Freeman MAR, Revell PA, Bradley GW,Tuke M. Can a total knee replacement prosthesis bemade entirely of polymers? J Arthroplasty 1998; 13:388-95.

2. McKellop HA, Rostlund T, Bradley G. Evaluation ofwear in an all polymer total knee replacement. Part1: laboratory testing of polyethylene on polyacetalbearing surfaces. Clin Mat 1993; 14: 117-26.

3. Ingham E, Fisher J. Biological reactions to wear de-bris in total joint replacement. J Engng Med 2000;214: 21-37.

4. Ingham E, Fisher J. The role of macrophages in os-teolysis of total joint replacement. Biomaterials 2005;26: 1271-86.

5. Joyce TJ, Monk D, Thompson P, Chiu P, Unsworth A,Green SM. Polymeric biomaterial wear test rig vali-dated to ASTM F732-00 and against clinical data.Seventh World Biomaterials Congress, Sydney, Aus-tralia, 537.

6. Joyce TJ, Unsworth A. A multi-directional wearscreening device and preliminary results of UHMW-PE articulating against stainless steel. Biomed MatEngng 2000; 10: 241-9.

7. Ramamurti BS, Estok DM, Jasty M, Harris WH.Analysis of the kinematics of different hip simulatorsused to study wear of candidate materials for the ar-ticulation of total hip arthroplasties. J Ortho Res1998; 16: 365-9.

8. Wang A, Essner A, Polineni VK, Stark C, DumbletonJH. Lubrication and wear of ultra-high molecularweight polyethylene in total joint replacements.1998; 31: 17-33.

9. Hamilton MA, Sucec MC, Fregly BJ, Banks SA,Sawyer WG. Quantifying multidirectional sliding

motions in total knee replacements. J Tribology2005; 127: 280-6.

10. ASTM-F732-00. (2000). Standard test method forwear testing of polymeric materials used in totaljoint prostheses.

11. Lakdawala A, Todo S, Scott G. The significance ofsurface changes on retrieved femoral componentsafter total knee replacement. J Bone Joint Surg Br2005; 87-B: 796-9.

12. Que L, Topoleski LDT, Parks NL. Surface rough-ness of retrieved CoCrMo alloy femoral compo-nents from PCA artificial total knee joints. J Bio-med Material Res 2000; 53: 111-8.

13. Saikko V, Calonius O. Simulation of wear rates andmechanisms in total knee prostheses by ball-on-flatcontact in a five-station, three-axis test rig. Wear2002; 253: 424-9.

14. Wang A, Sun DC, Yau SS, et al. Orientation soften-ing in the deformation and wear of ultra-high mol-ecular weight polyethylene. Wear 1997; 203-204:230-41.

15. Saikko V, Ahlroos T. Type of motion and lubricantin wear simulation of polyethylene acetabular cup.J Engng Med 1999; 213: 301-10.

16. Joyce TJ, Unsworth A. Wear studies of all UHMW-PE couples under various bio-tribological condi-tions. Journal of Applied Biomaterials & Biome-chanics 2004; 2: 29-34.

17. Joyce TJ, Unsworth A. The influence of bovineserum lubricant on the wear of cross-linked poly-ethylene finger prostheses. Journal of Applied Bio-materials & Biomechanics 2004; 2: 136-42.

18. Wroblewski BM, Siney PD, Fleming PA. Low fric-tion arthroplasty of the hip using alumina ceramicand cross-linked polyethylene. A 17-year follow-upreport. J Bone Jt Surg 2005; 87-B: 1220-21.

146

The wear of two orthopaedic biopolymers against each other