- Email: [email protected]
The effect of counterface roughness on the wear of UHMWPE for rectangular wear paths
Wear 259 (2005) 984–991
The effect of counterface roughness on the wear of UHMWPE for rectangular wear paths M.E. Turell a,b , G.E. Friedlaender b , A. Wang c , T.S. Thornhill a , A. Bellare a,∗ a
Department of Orthopaedic Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115, USA b Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, New Haven, CT 06510, USA c Stryker Howmedica Osteonics, 300 Commerce Court, Mahwah, NJ 07430, USA Received 1 August 2004; received in revised form 30 December 2004; accepted 31 January 2005 Available online 14 March 2005
Abstract Ultra-high molecular weight polyethylene (UHMWPE) is used worldwide as a bearing material in total joint replacement prostheses. Multi-directional motion has been identified as a major factor affecting the wear rate of UHMWPE in total hip replacement prostheses. The trajectory of relative motion between a femoral head and an acetabular cup takes a general quasi-elliptical or rectangular shape during the patient’s gait cycle. Differences in motion pattern can affect the in vivo wear rates of UHMWPE cups in patients when all other factors are equal. In a previous study that utilized smooth Co–Cr counterfaces, we compared wear factors, k, for UHMWPE articulated in a series of rectangular wear paths (width = A, length = B) with systematically increasing aspect ratios (B/A) and linear tracking (A = 0), all with identical path lengths per cycle. The results showed that the wear factor significantly decreased in the rectangular wear path with the highest aspect ratio and in linear tracking. The goal of our current study was to quantify the effect of a roughened counterface on the cross-path wear of UHMWPE. UHMWPE pins were articulated against both smooth and rough Co–Cr disks in diluted calf serum using a multi-directional wear tester under physiological loading conditions. Five different rectangular wear path geometries and linear tracking, all with identical path lengths per cycle, were employed for each wear test. Gravimetric weight loss was converted into volumetric wear rates and wear factors, k. The results showed that roughened counterfaces produced a larger increase in the wear factor in rectangular wear paths with higher aspect ratios. The ratio of krough /ksmooth decreased monotonically as a function of increasing width of rectangles, normalized by total path length, or A/(A + B). © 2005 Elsevier B.V. All rights reserved. Keywords: Ultra-high molecular weight polyethylene; Multi-directional wear; Total joint replacement; Joint arthroplasty
1. Introduction Ultra-high molecular weight polyethylene (UHMWPE) has been used worldwide as a bearing surface of total joint replacement prostheses since 1962. The high entanglement density associated with high molecular weight imparts superior fracture toughness [1,2] and high wear resistance to UHMWPE, making it an attractive choice as a bearing surface ∗ Corresponding author. Present address: Orthopaedic Research Laboratory, MRB 106, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115, USA. Tel.: +1 617 732 5864; fax: +1 617 732 6705. E-mail address: [email protected] (A. Bellare).
0043-1648/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2005.01.050
in both total knee and total hip joint replacement prostheses. Despite its superior mechanical and tribological properties, wear damage of UHMWPE continues to limit the clinical lifespan of implanted knee and hip joint prostheses. Volumetric wear rate of an acetabular cup articulating against a 32 mm femoral head is approximately 80 mm3 /year, which translates into an average linear wear of 0.1 mm annually [3–6]. The thickness of UHMWPE components is usually 1 cm or more. Therefore, it would take 100 years for UHMWPE components to wear through. While this estimate eliminates the likelihood that UHMWPE implant components are in any danger of wearing through over the course of a patient’s lifetime, billions of sub-micron wear debris particles are pro-
M.E. Turell et al. / Wear 259 (2005) 984–991
duced annually [7,8], which continues to be the primary concern in total hip replacement prostheses. The accumulation of particulate debris in the peri-prosthetic tissue can elicit a biological response leading to bone loss (osteolysis) and implant loosening, necessitating complicated revision surgery to replace the implant [9–11]. Osteolysis associated with particulate wear debris has prompted numerous investigations, which study the size, shape and morphology of particulate debris and the biological pathway that leads to particle-induced osteolysis. Wear debris collected from peri-prosthetic tissue retrieval studies showed that the majority of debris particles associated with hip replacement prostheses were micron or sub-micron in size and that particles could generally be categorized as fibrous or particulate in shape [12–16]. Particulate wear also occurs in total knee replacement (TKR) components, although the particle size distribution differs from that of total hip replacement (THR) wear particles, and TKR particles have been shown to be generally larger than THR wear particles [17]. Computational and experimental studies have shown that multi-directional motion or “cross-shear” motion can affect wear rate of UHMWPE in total hip and total knee replacements [18,19]. Linear tracking motion, whether unidirectional or reciprocating produces an extremely low wear rate, and in fact produces two to three orders of magnitude less wear than is observed clinically in total knee replacement prostheses [19–24]. The higher clinical wear rates of total knee replacement prostheses can be attributed to both counterface scratching [25,26] as well as the effects of cross-shear motion. Wang et al. have proposed that in a wear environment, UHMWPE macromolecules orient preferentially along the principal direction of sliding [19,22–24]. Unlike in linear tracking, where orientation results in strain hardening of surface material and ultimately increases wear resistance as sliding progresses, in multi-directional motion, the wear surface experiences both compressive and shear forces in multiple directions. As sliding proceeds, the UHMWPE wear surface may strengthen along the direction of sliding, while it weakens in the perpendicular direction. It has been demonstrated that the loci or trajectory of motion at the contact point between a femoral head of an orthopaedic implant and an acetabular cup is a quasi-ellipse or an approximate rectangle during a gait cycle [18,19,27]. Some patients have either more elongated (approximately rectangular) or more closed (approximately more square-like) motion patterns than others [27]. Examples of elongated and closed motion patterns traced by the contact point between femoral and acetabular components are illustrated in the examples shown in Fig. 1. Bennett et al. have postulated that the differences in motion patterns affect the in vivo wear rates of UHMWPE acetabular cups in patients where factors such as age, weight and body proportion were similar, but gait patterns varied widely [27]. The multi-directional sliding, or orientation-softening wear model, proposed by Wang has been termed the unified theory of wear for UHMWPE [28]. This theory proposed that when a femoral head slides against an acetabular cup
985
Fig. 1. Elongated motion pattern (left) and closed motion pattern (right) traced by the contact point between the formal and acetabular components of orthopaedic implants.
along the wear path defined by the rectangular loop, frictional energy is dissipated in both the A and B directions (see Fig. 1). Since A < B, B defines the principal direction of sliding motion while A is secondary. Previous studies indicated that motion in the principal sliding direction, B, leads to plastic deformation or macromolecular orientation whereas motion in the secondary direction, A, leads to material removal by fracture [19,22–24,28]. Therefore, only energy released in the A direction is directly responsible for wear [28]. The expression for wear rate can then be expressed as: V ∝ µmP(2A)
(1)
where V is volumetric wear rate, µ is the coefficient of friction, P is the applied normal load and 2A is the sliding distance in the secondary direction per cycle. Since the wear coefficient or wear factor, k, is defined as (volumetric wear rate)/(load × total sliding distance), in the case of a rectangular wear path illustrated above, k can be written as: k=
V µP(2A) A ∝ ∝ P(2A + 2B) P(2A + 2B) A+B
(2)
Eq. (2) theoretically quantifies the effect of cross-path motion on the wear factor of UHMWPE. In a previous set of experiments, we compared wear factors for UHMWPE articulated in a series of graded wear paths in which the aspect ratio of the wear path was systematically increased. These wear tests were conducted using a cobalt–chromium articulation surface with an implant-grade smooth finish. The results showed that the wear factors from these experiments were in agreement with the proposed model only at high aspect ratios, i.e. wear factors were found to significantly decrease in the rectangular wear path with the highest aspect ratio and in linear tracking [29]. A question that remains to be answered is whether the proposed model remains valid under conditions of abrasive wear. Studies have shown that the addition of bone cement (with zirconium and barium sulfate additives) and bone particles to test serum lubricant produces significantly greater surface damage to stainless steel articulation counterfaces which in turn results in surface roughening and increased wear rates [25]. Similarly, other authors have found that roughening of the femoral head, to a degree seen typically in retrieval specimens, can increase the observed variability of volumetric wear rates approximately seven-fold. This fact may explain why random femoral head scratching in vivo accounts for otherwise difficult to explain variations in wear rates as abrasive wear may be a key factor causing excessive wear in the most problematic subset of the patients with to-
986
M.E. Turell et al. / Wear 259 (2005) 984–991
tal joint replacements [26]. The purpose of the current study is to determine whether the unified theory of wear is applicable to conditions of abrasive wear, such as the case of a roughened counterface articulating against UHMWPE along rectangles of various aspect ratios. It was hypothesized that under conditions of abrasive wear, the model would in general describe the cross-path wear of UHMWPE as accurately as it had under conditions in which a smooth counterface was employed.
Commercially available, ram-extruded GUR 1050 (Hoechst-Ticona, Bayport, TX) rod stock (PolyHi Solidur, Ft. Wayne, IN) was used as the starting material for all wear tests. Rod stock with a diameter of 7.6 cm was machined into cylindrical pins with overall dimensions of 20 mm in length and 9 mm in diameter for all wear tests. All pins were tested in the as-machined condition (no sterilization).
set-up in Fig. 2. All wear tests were conducted at a cycle frequency of 1 Hz (constant sliding speed of 20 mm/s along the wear track) and with a constant applied load of 192 N or an applied stress of 3 MPa, well within the physiological range of 2–5 MPa for the hip joint. A bovine serum (JRH Biosciences) lubricant was used for all wear tests, and was diluted with distilled water to contain 23 g/L protein, 20 mM EDTA, and 0.2% sodium azide. The lubricating bovine serum temperature was maintained at 37 ◦ C using a re-circulating water bath. Cobalt–chromium disks with dimensions of 25 mm diameter and 3 mm thickness were used as the articulation counterface. Two series of wear tests using each of the six experimental articulation patterns were performed. In one series of tests, cobalt–chromium disks were polished to implant-grade surface smoothness with a centerline roughness of 0.015 m (Ra). In a second set of experiments, the cobalt–chromium disks were scratched along random directions using 320 grit emery paper in accordance with the previously established method of Wang et al., resulting in an average roughness, Ra, of 0.45 m [30]. A total of six pins plus an additional pin serving as a soak control were used for all wear tests.
2.2. Wear testing protocol and apparatus
2.3. Data analysis
A series of six different articulation patterns including a 5 mm × 5 mm square, 4 mm × 6 mm, 3 mm × 7 mm, 2 mm × 8 mm, and 1 mm × 9 mm rectangles, and a 0 mm × 10 mm linear tracking pattern were digitized into an OrthoPODTM (Advanced Mechanical Technology Inc.) multi-directional wear tester. For each wear test, the OrthoPODTM multi-directional wear tester was loaded with six UHMWPE pins as shown in the experimental
Wear rates for each of the six articulation patterns were determined in experiments in which both smooth and rough counterfaces were employed. Gravimetric weight loss per pin was determined approximately every 200,000 cycles. The soak control consistently revealed that the amount of absorption of bovine serum by the UHMWPE specimen was undetectable and therefore any corrections to compensate for fluid absorption were unnecessary. For each group, wear tests were
2. Materials and methods 2.1. UHMWPE starting material
Fig. 2. AMTI OrthoPODTM multi-directional wear tester.
M.E. Turell et al. / Wear 259 (2005) 984–991
continued until at least one million cycles had been reached. The gravimetric weight loss per pin was determined by taking the average weight loss of the six wear tested pins. Prior to being weighed, UHMWPE pins were first washed twice with distilled water followed by an alcohol and acetone rinse, respectively. The pins were dried in air at 25 ◦ C for 15 min prior to determination of weight loss. Gravimetric weight loss was further converted into volumetric wear data by using a density value of 0.943 g/cm3 for UHMWPE. Wear factor values, k, were also calculated for each sample by dividing the volumetric wear rate by the product of the load and the sliding distance per cycle. 2.4. Statistics Statistical analysis was conducted using ANOVA with Fisher’s protected least significant difference (PLSD) posthoc test in which a p-value of less than 0.05 was used to define statistical significance. In addition, single-factor ANOVA using a 95% confidence interval (α = 0.05) was also used to more rigorously assess the statistical significance between groups. Wear factor values from each wear path geometry using both smooth and rough counterfaces were compared against one another in order to establish the level of significance between different groups.
3. Results 3.1. Wear testing Wear tests showed that the wear rates increased incrementally as the wear path decreased in aspect ratio and approached the 3 mm × 7 mm configuration. Interestingly, for both experiments in which smooth counterfaces and rough counterfaces were used, wear reached a maximum when a 3 mm × 7 mm wear path was employed (Fig. 3). In the smooth counterface test series, the wear factor for the 5 mm × 5 mm square path was found to be 2.5 times greater in comparison to the 2 mm × 8 mm path (p < 0.05, ANOVA with Fisher’s PLSD post-hoc test), which agrees remarkably well with the
987
prediction of Eq. (2). For the same comparison in the rough counterface test series, the wear factor for the 5 mm × 5 mm square path was found to be only 1.1 times greater than the wear factor for the 2 mm × 8 mm path. Although this trend was observed, in the case of the rough counterface test, the differences between the 5 mm × 5 mm path and the 2 mm × 8 mm path was not statistically significant. 3.2. Effect of counterface roughness As predicted, counterface roughness had a substantial effect on wear rate. Although standard deviations between groups were large, the differences in wear rates between rough and smooth tests for each wear path geometry were statistically significant (p < 0.05) according to ANOVA (using a 95% confidence interval). This is due to the large number of data points included in each group. An exception is that the differences in the wear factor for the 4 mm × 6 mm rough versus smooth tests were not statistically significant. Using ANOVA (with the 95% confidence interval criteria) within the smooth counterface series, the wear factor of the 1 mm × 9 mm wear path was significantly greater than that of the 0 mm × 10 mm path as was the wear factor of the 3 mm × 7 mm versus the 2 mm × 8 mm. However, the wear factors of 2 mm × 8 mm versus the 1 mm × 9 mm, the 4 mm × 6 mm versus the 3 mm × 7 mm, and the 5 mm × 5 mm versus the 4 mm × 6 mm were not significantly different from one another. Using this same analysis, in the rough counterface set of wear tests, the wear factor associated with the 1 mm × 9 mm path was significantly greater than that of the 0 mm × 10 mm path as was the 3 mm × 7 mm path compared to the 4 mm × 6 mm path. The wear factor of the 2 mm × 8 mm versus the 1 mm × 9 mm path, the 3 mm × 7 mm versus the 2 mm × 8 mm path, and the 5 mm × 5 mm versus the 4 mm × 6 mm path were not statistically different. An important finding of this study was that the counterface roughness affected wear rates to a greater extent in experiments in which more linear rectangular paths were used as compared to the more square-like wear paths. In comparison to the smooth counterface series, the rough counterface tests demonstrated wear rates that were significantly greater than predicted when more linear wear paths were tested. There was a general decreasing trend in the ratio of krough /ksmooth as the aspect ratio of the rectangles decreased (approached the square configuration), as shown in Fig. 4. 3.3. Unified theory of wear
Fig. 3. Comparison of wear factors for both smooth and rough cobalt–chromium counterfaces.
In order to assess the accuracy of the unified theory of wear in predicting the wear rate of UHMWPE based on a particular wear path geometry, the experimentally observed wear factor values for both the smooth and rough counterface tests were plotted as a function of the ratio A/(A + B) (Fig. 5). These curves were then compared to predicted wear factor trends, in which the maximum predicted wear factor was taken to be the value of the wear factor obtained from the
988
M.E. Turell et al. / Wear 259 (2005) 984–991
Fig. 6. Liner fit (with maximum A/(A + B) = 0.3) for wear factor, k, vs. A/(A + B) ratio.
Fig. 4. Comparison of wear factor ratios (rough/smooth).
5 mm × 5 mm square wear path. Analysis revealed that the orientation-softening theory of wear was unable to accurately predict wear over the entire range of rectangles for both the smooth and rough counterface series. In smooth counterface tests, the parameter A/(A + B) was particularly accurate in the prediction of wear rates in cases where the aspect ratio of the rectangular path was highest, such as in the 0 mm × 10 mm to 3 mm × 7 mm wear paths. The monotonic increase in wear factors in the A/(A + B) range of 0.0 and 0.3 for UHMWPE in both rough and smooth counterfaces suggested a linear correlation between wear factor, k, and A/(A + B). Therefore, a linear curve fit was conducted in a separate plot (Fig. 6) for the wear factors in the A/(A + B) range of 0.0–0.3. Fig. 6 shows that the wear factor values from both the smooth and rough counterface tests show a strong linear relationship in this range. In both the smooth and rough counterface experiments, the wear factor increased with increasing A/(A + B) ratio, but only up to a value of 0.3. There were significant deviations in the experimental data from the theoretical linear correlation predicted by the orientational softening theory (Fig. 5). These differences are particularly apparent in the set of experiments
Fig. 5. Wear factor, k, vs. A/(A + B) ratio.
employing rough counterfaces. According to the theoretical model of Wang [28], the maximum wear factor should occur at A = B or A/(A + B) = 0.5. Instead the experimentally observed wear factor peaked at A/(A + B) = 0.3 for both smooth and rough counterface experiments. In fact, the 3 mm × 7 mm showed a 16% increase over the 5 mm × 5 mm square path in the series where smooth counterfaces were used (p < 0.05, ANOVA with Fisher’s PLSD post-hoc test).
4. Discussion 4.1. Wear tests This study was conducted as an extension of a previous study that sought to quantify the effect of the cross-path motion on wear rate of UHMWPE. The aim of this study was to address the effects of counterface roughness on wear rates for different cross-path motions. For a rectangular wear path with a width A and a length B (A < B), the numerical parameter A/(A + B) is a convenient measure to relate wear factor, k, to cross-path motion for rectangles with various aspect ratios, B/A. This numerical parameter, defined by the unified wear model, was relatively accurate in predicting the general trend that wear rates would increase as a function of decreasing aspect ratio (i.e. as a function of the wear path approaching a square configuration) but only in the range of the 1 mm × 9 mm wear path to the 3 mm × 7 mm configuration. For both smooth and rough counterface tests, a linear trend was particularly evident in cases where the aspect ratio of the rectangular path was highest, such as in the 0 mm × 10 mm, 1 mm × 9 mm, and 2 mm × 8 mm wear paths. In both counterface groups, wear factors reached a maximum when a 3 mm × 7 mm wear path was employed, and the trend was no longer observed as the aspect ratio of the wear path continued to decrease beyond this point. The failure of the unified wear model to predict wear rates in some cases for both rough and smooth counterface tests raises ques-
M.E. Turell et al. / Wear 259 (2005) 984–991
tions about some of the underlying assumptions upon which the model is based. For example, for any rectangular wear path with sides A and B, where A < B, the model assumes that orientation on the molecular level occurs in the B direction, while fracture, or the actual wearing of UHMWPE, occurs in the A direction. Based on the results of this study, in which wear reached a maximum for the 3 mm × 7 mm wear path, this assumption may not be a valid description of the orientation and wear processes. It is possible that complete orientation and strain hardening of UHMWPE in the B direction is required for wear to occur solely in the A direction, and that this is only achieved at 7 mm of sliding. It is therefore possible that for the 5 mm × 5 mm and 4 mm × 6 mm wear paths, there is significant reorientation in the A direction as well rather than fracture of the fully oriented, strain hardened UHMWPE fibrils. Reorientation would imply that a different wear mechanism is operative since biaxial orientation prior to wear would suggest that a more sheet-like material wears rather than splitting of fibrillar UHMWPE. These wear and orientation mechanisms appeared to occur in two discrete A/(A + B) ranges, irrespective of the roughness of counterface. Fig. 6 shows that there was a linear correlation between wear factor, k, and the numerical parameter A/(A + B), regardless of counterface roughness. The linear equations for these correlations as well as the R-values (to measure the degree of fit) were obtained for both the smooth and rough counterface tests. In the equation for the line, k = m[A/(A + B)] + c, the constant c (y-intercept) represents the fraction of wear factor that is due to linear, abrasive wear and the constant m (slope) is a measure of the dependence of cross-path wear on the wear path geometry. The following equations were obtained for the linear correlations: A k = 7E–06 − 6E–08 (R = 0.8821) A+B (smooth counterface) k = 1E–05
A A+B
(3)
+ 4E–07 (R = 0.9922) (rough counterface)
(4)
It is evident that the constant c cannot assume a negative value, and the low, negative wear factor value of 6E−08 for linear tracking for smooth heads obtained from the linear fit is merely a consequence of experimental error, which can be taken to be zero or replaced by the experimentally measured positive value of 5.06E−08. The value of m was higher in the rough counterface series of wear tests. The ratio of mrough /msmooth was 1.4, revealing that there was a steeper dependence of wear factor, k, on the numerical parameter, A/(A + B), in the case of rough counterface. For both the smooth and rough counterface, the high R-values indicated a strong linear correlation between the wear factor and A/(A + B). The good fit of the wear factors in the series of
989
wear paths in the range of 0 mm × 10 mm to 3 mm × 7 mm paths suggests that complete orientation followed by fracture of fibrils was the primary mechanism of wear in both of these cases. Independent of the accuracy with which the model investigated in this paper predicted wear rate, the overall experimental results support the hypothesis that wear rate is in fact dependent upon the wear path geometry, counterface roughness and the interplay of these two variables. 4.2. Clinical relevance The finding that wear rate is dependent upon both wear path pattern and counterface roughness has a number of clinically relevant implications. The observation that differences in wear rates, between tests employing rough versus smooth counterfaces, are greater in more linear motion path patterns and that these differences systematically decrease as the wear approaches a square pattern is an important finding. In total joint replacement applications where linear wear is known to be operative and to have a significant impact upon the lifetime of the joint replacement, for example, as is the case in the knee joint, the effects of abrasive wear (simulated by a roughened counterface in this study) are of greater concern. It should be noted that the results of this study reveal a somewhat oversimplification of the wear mechanisms as they occur in clinical application since the effects of third body wear and fatigue-related wear mechanisms were not investigated. A question which remains to be answered is whether it may be possible to predict motion path patterns by conducting gait analysis and if so, whether motion path pattern can be therapeutically manipulated using methods of gait training or by making improvements to implant design. A challenge in addressing these questions lies in the difficulty in assessing gait in patients who require total joint replacement surgery as gait in these individuals may be altered from their normal baseline gait due to the orthopaedic complications, which necessitate surgery in the first place. Similarly, gait patterns in these individuals may be dramatically altered following total joint replacement surgery, making it difficult to predict the effects of implant design in advance of surgery. 4.3. Limitations The results of this study represent an attempt to quantify the effect of the motion path pattern and counterface roughness on wear rate of UHMWPE. It should be noted that the data, in particular the wear factor values, that have been analyzed in this paper represent the results of preliminary wear tests conducted for each wear path geometry to a period of at least one million cycles. A more comprehensive study, and one that would employ more rigorous wear testing would generate a larger number of samples for tests encompassing a broader range of motion path patterns. From such experiments, trends in the values of UHMWPE wear rates would be more reliably generated and would be of a greater level of clinical significance. A study of the morphology of wear
990
M.E. Turell et al. / Wear 259 (2005) 984–991
particles generated from the various motion path patterns for both smooth and rough counterface tests is also necessary (and forthcoming) to obtain a more accurate understanding of the wear mechanisms that are operative during the various wear paths. In addition, the wear tests in this paper employed a constant applied load and it is important for future studies to include wear test loading parameters which more closely model the loading conditions found in knee and hip joints in vivo. The relationship between wear path and wear rates established in this study applies only to uncrosslinked medical grade UHMWPE. Further wear testing on the various crosslinked UHMWPEs currently in clinical use is required for such relationships to be established.
5. Conclusions Wear of UHMWPE as it articulates against a metallic counterface in a particular wear path and under conditions that mimic wear in a total hip replacement prosthesis, likely occurs via two discrete steps for rectangles with an aspect ratio greater than 2.33 (3 mm × 7 mm rectangular path). For such rectangles, the wear tests of this study support the hypothesis that there is orientation or texturing of UHMWPE on parallel edges of the rectangle followed by wear of the textured UHMWPE on the other two parallel edges. However, for rectangles in the aspect ratio range of 1.0–2.33 (5 mm × 5 mm to 4 mm × 6 mm paths), the decline in wear rates for both smooth and rough counterface experiments provide evidence that wear and orientation processes may not occur in discrete stages. In addition, the unified wear model predicts zero wear for linear tracking, which is not the case, especially when more abrasive conditions of wear occur such as the case of a roughened counterface. A more robust model is required to predict wear of UHMWPE during articulation against a metallic counterface along a rectangular path covering the entire range of aspect ratios of rectangles.
Acknowledgments This project was funded through a biomedical engineering grant provided by the Whitaker Foundation and by a fellowship provided by the Orthopaedic Research and Education Foundation.
References [1] A. Gomoll, T. Wanich, A. Bellare, J-integral fracture toughness and tearing modulus measurement of radiation cross-linked UHMWPE, J. Orthop. Res. 20 (6) (2002) 1152–1156. [2] L.C. Duus, H.A. Walsh, A.M. Gillis, E. Noisiez, S. Li, The effect of resin grade, manufacturing method and cross linking on the fracture toughness of commercially available UHMWPE, Trans. Orthop. Res. Soc. 25 (2000) 544.
[3] M. Griffith, M. Seidenstein, D. Williams, J. Charnley, Socket wear in Charnley low friction arthroplasty of the hip, Clin. Orthop. 137 (1978) 37–47. [4] J. Livermore, D. Ilstrupand, B. Morrey, Effect of femoral head size on wear of the polyethylene acetabular component, J. Bone Joint Surg. 72A (4) (1990) 518–528. [5] R. Hall, A. Unsworth, P. Siney, B. Wroblewski, Wear in retrieved Charnley acetabular sockets, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 197–207. [6] H. Oonishi, E. Tsuji, Y. Kim, Retrieved total hip prostheses. Part I. The effects of cup thickness, head size and fusion defects on wear, J. Mater. Sci.: Mater. Med. 9 (1998) 394–401. [7] I.C. Clarke, J.M. Kabo, Wear in total hip replacement, in: H.C. Amstutz (Ed.), Hip Arthroplasty, Churchill Livingstone, New York, 1991. [8] G. Lewis, Polyethylene wear in total hip and knee arthroplasties, J. Biomed. Mater. Res. (Appl. Biomater.) 38 (1997) 55–75. [9] H. Amstutz, P. Campbell, N. Kossovsky, I. Clarke, Mechanisms and clinical significance of wear debris-induced osteolysis, Clin. Orthop. Relat. Res. 276 (1992) 7–18. [10] W. Harris, The problem is osteolysis, Clin. Orthop. Relat. Res. 311 (1995) 46–53. [11] T. Schmalzried, L. Kwong, M. Jasty, R. Sedlacek, T. Haire, D. O’Connor, C. Bragdon, J. Kabo, A. Malcolm, W. Harris, The mechanism of loosening of cemented acetabular components in total hip arthroplasty: analysis of specimens retrieved at autopsy, Clin. Orthop. Relat. Res. 274 (1992) 60–78. [12] P. Campbell, S. Ma, B. Yeom, H. McKellop, T. Schmalzried, H. Amstrutz, Isolation of predominantly submicron-sized UHMWPE wear particles from periprosthetic tissues, J. Biomed. Mater. Res. 29 (1995) 127–131. [13] A. Shanbhag, J. Jacobs, T. Glant, J. Gilbert, J. Black, J. Galante, Composition and morphology of wear debris in failed uncemented total hip replacement, J. Bone Joint Surg. 76B (1994) 60–67. [14] J. Savio, L. Overcamp, J. Black, Size and shape of biomaterial wear debris, Clin. Mater. 15 (1994) 101–119. [15] K. Margevicius, T. Bauer, J. McMahon, S. Brown, K. Merritt, Isolation and characterization of debris in membranes around total joint prostheses, J. Bone Joint Surg. 76A (1994) 1664–1676. [16] P. Campbell, P. Doorn, F. Dorey, H. Amstutz, Wear and morphology of ultra-high molecular weight polyethylene wear particles from total hip replacements, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 167–175. [17] A. Shanbhag, H. Bailey, D. Hwang, C. Cha, N. Eror, H. Rubash, Quantitative analysis of ultrahigh molecular weight polyethylene (UHMWPE) wear debris associated with total knee replacements, J. Biomed. Mater. Res. 53 (1) (2000) 100–110. [18] B. Ramamurti, C. Bragdon, D. O’Connor, J. Lowenstein, M. Jasty, D. Estok, W. Harris, Loci of movement of selected points on the femoral head during normal gait, J. Arthroplasty 11 (7) (1996) 852–855. [19] A. Wang, D. Sun, S.-S. Yau, B. Edwards, M. Sokol, A. Essner, V. Polineni, C. Stark, J. Dumbelton, Orientation softening in the deformation and wear of ultra-high polyethylene, Wear 203–204 (1997) 230–241. [20] V. Saikko, T. Ahlroos, Type of motion and lubricant in wear simulation of polyethylene acetabular cup, Proc. Inst. Mech. Eng. H: J. Eng. Med. 213 (1999) 301–310. [21] C. Bragdon, D. O’Connor, J. Lowenstein, M. Jasty, W. Syniuta, The importance of multidirectional motion on the wear of polyethylene, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 157–166. [22] A. Wang, C. Stark, J. Dumbleton, Mechanistic and morphological origins of ultra-high molecular weight polyethylene wear debris in total joint replacement prostheses, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 141–155. [23] A. Wang, V. Polineni, A. Essner, M. Sokol, D. Sun, C. Stark, J. Dumbleton, The significance of non-linear motion in the wear screening of orthopaedic implant materials, Test. Eval. 25 (2) (1997) 239–245.
M.E. Turell et al. / Wear 259 (2005) 984–991 [24] A. Wang, A. Essner, V. Polineni, C. Stark, J. Dumbleton, Lubrication and wear of ultra-high molecular weight polyethylene in total joint replacements, Tribol. Int. 31 (1–3) (1998) 17–33. [25] L. Caravia, D. Dowson, J. Fisher, B. Jobbins, The influence of bone and bone cement debris on counterface roughness in sliding wear tests of ultra-high molecular weight polyethylene on stainless steel, Proc. Inst. Mech. Eng. H: J. Eng. Med. 204 (1) (1990) 65–70. [26] T.D. Brown, K.J. Stewart, J.C. Nieman, D.R. Pedersen, J.J. Callaghan, Local head roughening as a factor contributing to variability of total hip wear: a finite element analysis, J. Biomech. Eng. 124 (6) (2002) 691–698.
991
[27] D. Bennett, J. Orr, R. Baker, Movement loci of selected points on the femoral head for individual total hip arthroplasty patients using three-dimensional computer simulation, J. Arthroplasty 15 (7) (2000) 909–915. [28] A. Wang, A unified theory of wear for ultra-high molecular weight polyethylene in multidirectional sliding, Wear 248 (2001) 38–47. [29] M. Turell, A. Wang, A. Bellare, Quantification of the effect of cross-path motion on the wear rate of ultra-high molecular weight polyethylene, Wear 255 (2003) 1034–1039. [30] A. Wang, V.K. Polineni, C. Stark, J.H. Dumbleton, Effect of femoral head surface roughness on the wear of ultrahigh molecular weight polyethylene acetabular cups, J. Arthroplasty 6 (1998) 615–620.
The effect of counterface roughness on the wear of UHMWPE for rectangular wear paths M.E. Turell a,b , G.E. Friedlaender b , A. Wang c , T.S. Thornhill a , A. Bellare a,∗ a
Department of Orthopaedic Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115, USA b Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, New Haven, CT 06510, USA c Stryker Howmedica Osteonics, 300 Commerce Court, Mahwah, NJ 07430, USA Received 1 August 2004; received in revised form 30 December 2004; accepted 31 January 2005 Available online 14 March 2005
Abstract Ultra-high molecular weight polyethylene (UHMWPE) is used worldwide as a bearing material in total joint replacement prostheses. Multi-directional motion has been identified as a major factor affecting the wear rate of UHMWPE in total hip replacement prostheses. The trajectory of relative motion between a femoral head and an acetabular cup takes a general quasi-elliptical or rectangular shape during the patient’s gait cycle. Differences in motion pattern can affect the in vivo wear rates of UHMWPE cups in patients when all other factors are equal. In a previous study that utilized smooth Co–Cr counterfaces, we compared wear factors, k, for UHMWPE articulated in a series of rectangular wear paths (width = A, length = B) with systematically increasing aspect ratios (B/A) and linear tracking (A = 0), all with identical path lengths per cycle. The results showed that the wear factor significantly decreased in the rectangular wear path with the highest aspect ratio and in linear tracking. The goal of our current study was to quantify the effect of a roughened counterface on the cross-path wear of UHMWPE. UHMWPE pins were articulated against both smooth and rough Co–Cr disks in diluted calf serum using a multi-directional wear tester under physiological loading conditions. Five different rectangular wear path geometries and linear tracking, all with identical path lengths per cycle, were employed for each wear test. Gravimetric weight loss was converted into volumetric wear rates and wear factors, k. The results showed that roughened counterfaces produced a larger increase in the wear factor in rectangular wear paths with higher aspect ratios. The ratio of krough /ksmooth decreased monotonically as a function of increasing width of rectangles, normalized by total path length, or A/(A + B). © 2005 Elsevier B.V. All rights reserved. Keywords: Ultra-high molecular weight polyethylene; Multi-directional wear; Total joint replacement; Joint arthroplasty
1. Introduction Ultra-high molecular weight polyethylene (UHMWPE) has been used worldwide as a bearing surface of total joint replacement prostheses since 1962. The high entanglement density associated with high molecular weight imparts superior fracture toughness [1,2] and high wear resistance to UHMWPE, making it an attractive choice as a bearing surface ∗ Corresponding author. Present address: Orthopaedic Research Laboratory, MRB 106, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115, USA. Tel.: +1 617 732 5864; fax: +1 617 732 6705. E-mail address: [email protected] (A. Bellare).
0043-1648/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2005.01.050
in both total knee and total hip joint replacement prostheses. Despite its superior mechanical and tribological properties, wear damage of UHMWPE continues to limit the clinical lifespan of implanted knee and hip joint prostheses. Volumetric wear rate of an acetabular cup articulating against a 32 mm femoral head is approximately 80 mm3 /year, which translates into an average linear wear of 0.1 mm annually [3–6]. The thickness of UHMWPE components is usually 1 cm or more. Therefore, it would take 100 years for UHMWPE components to wear through. While this estimate eliminates the likelihood that UHMWPE implant components are in any danger of wearing through over the course of a patient’s lifetime, billions of sub-micron wear debris particles are pro-
M.E. Turell et al. / Wear 259 (2005) 984–991
duced annually [7,8], which continues to be the primary concern in total hip replacement prostheses. The accumulation of particulate debris in the peri-prosthetic tissue can elicit a biological response leading to bone loss (osteolysis) and implant loosening, necessitating complicated revision surgery to replace the implant [9–11]. Osteolysis associated with particulate wear debris has prompted numerous investigations, which study the size, shape and morphology of particulate debris and the biological pathway that leads to particle-induced osteolysis. Wear debris collected from peri-prosthetic tissue retrieval studies showed that the majority of debris particles associated with hip replacement prostheses were micron or sub-micron in size and that particles could generally be categorized as fibrous or particulate in shape [12–16]. Particulate wear also occurs in total knee replacement (TKR) components, although the particle size distribution differs from that of total hip replacement (THR) wear particles, and TKR particles have been shown to be generally larger than THR wear particles [17]. Computational and experimental studies have shown that multi-directional motion or “cross-shear” motion can affect wear rate of UHMWPE in total hip and total knee replacements [18,19]. Linear tracking motion, whether unidirectional or reciprocating produces an extremely low wear rate, and in fact produces two to three orders of magnitude less wear than is observed clinically in total knee replacement prostheses [19–24]. The higher clinical wear rates of total knee replacement prostheses can be attributed to both counterface scratching [25,26] as well as the effects of cross-shear motion. Wang et al. have proposed that in a wear environment, UHMWPE macromolecules orient preferentially along the principal direction of sliding [19,22–24]. Unlike in linear tracking, where orientation results in strain hardening of surface material and ultimately increases wear resistance as sliding progresses, in multi-directional motion, the wear surface experiences both compressive and shear forces in multiple directions. As sliding proceeds, the UHMWPE wear surface may strengthen along the direction of sliding, while it weakens in the perpendicular direction. It has been demonstrated that the loci or trajectory of motion at the contact point between a femoral head of an orthopaedic implant and an acetabular cup is a quasi-ellipse or an approximate rectangle during a gait cycle [18,19,27]. Some patients have either more elongated (approximately rectangular) or more closed (approximately more square-like) motion patterns than others [27]. Examples of elongated and closed motion patterns traced by the contact point between femoral and acetabular components are illustrated in the examples shown in Fig. 1. Bennett et al. have postulated that the differences in motion patterns affect the in vivo wear rates of UHMWPE acetabular cups in patients where factors such as age, weight and body proportion were similar, but gait patterns varied widely [27]. The multi-directional sliding, or orientation-softening wear model, proposed by Wang has been termed the unified theory of wear for UHMWPE [28]. This theory proposed that when a femoral head slides against an acetabular cup
985
Fig. 1. Elongated motion pattern (left) and closed motion pattern (right) traced by the contact point between the formal and acetabular components of orthopaedic implants.
along the wear path defined by the rectangular loop, frictional energy is dissipated in both the A and B directions (see Fig. 1). Since A < B, B defines the principal direction of sliding motion while A is secondary. Previous studies indicated that motion in the principal sliding direction, B, leads to plastic deformation or macromolecular orientation whereas motion in the secondary direction, A, leads to material removal by fracture [19,22–24,28]. Therefore, only energy released in the A direction is directly responsible for wear [28]. The expression for wear rate can then be expressed as: V ∝ µmP(2A)
(1)
where V is volumetric wear rate, µ is the coefficient of friction, P is the applied normal load and 2A is the sliding distance in the secondary direction per cycle. Since the wear coefficient or wear factor, k, is defined as (volumetric wear rate)/(load × total sliding distance), in the case of a rectangular wear path illustrated above, k can be written as: k=
V µP(2A) A ∝ ∝ P(2A + 2B) P(2A + 2B) A+B
(2)
Eq. (2) theoretically quantifies the effect of cross-path motion on the wear factor of UHMWPE. In a previous set of experiments, we compared wear factors for UHMWPE articulated in a series of graded wear paths in which the aspect ratio of the wear path was systematically increased. These wear tests were conducted using a cobalt–chromium articulation surface with an implant-grade smooth finish. The results showed that the wear factors from these experiments were in agreement with the proposed model only at high aspect ratios, i.e. wear factors were found to significantly decrease in the rectangular wear path with the highest aspect ratio and in linear tracking [29]. A question that remains to be answered is whether the proposed model remains valid under conditions of abrasive wear. Studies have shown that the addition of bone cement (with zirconium and barium sulfate additives) and bone particles to test serum lubricant produces significantly greater surface damage to stainless steel articulation counterfaces which in turn results in surface roughening and increased wear rates [25]. Similarly, other authors have found that roughening of the femoral head, to a degree seen typically in retrieval specimens, can increase the observed variability of volumetric wear rates approximately seven-fold. This fact may explain why random femoral head scratching in vivo accounts for otherwise difficult to explain variations in wear rates as abrasive wear may be a key factor causing excessive wear in the most problematic subset of the patients with to-
986
M.E. Turell et al. / Wear 259 (2005) 984–991
tal joint replacements [26]. The purpose of the current study is to determine whether the unified theory of wear is applicable to conditions of abrasive wear, such as the case of a roughened counterface articulating against UHMWPE along rectangles of various aspect ratios. It was hypothesized that under conditions of abrasive wear, the model would in general describe the cross-path wear of UHMWPE as accurately as it had under conditions in which a smooth counterface was employed.
Commercially available, ram-extruded GUR 1050 (Hoechst-Ticona, Bayport, TX) rod stock (PolyHi Solidur, Ft. Wayne, IN) was used as the starting material for all wear tests. Rod stock with a diameter of 7.6 cm was machined into cylindrical pins with overall dimensions of 20 mm in length and 9 mm in diameter for all wear tests. All pins were tested in the as-machined condition (no sterilization).
set-up in Fig. 2. All wear tests were conducted at a cycle frequency of 1 Hz (constant sliding speed of 20 mm/s along the wear track) and with a constant applied load of 192 N or an applied stress of 3 MPa, well within the physiological range of 2–5 MPa for the hip joint. A bovine serum (JRH Biosciences) lubricant was used for all wear tests, and was diluted with distilled water to contain 23 g/L protein, 20 mM EDTA, and 0.2% sodium azide. The lubricating bovine serum temperature was maintained at 37 ◦ C using a re-circulating water bath. Cobalt–chromium disks with dimensions of 25 mm diameter and 3 mm thickness were used as the articulation counterface. Two series of wear tests using each of the six experimental articulation patterns were performed. In one series of tests, cobalt–chromium disks were polished to implant-grade surface smoothness with a centerline roughness of 0.015 m (Ra). In a second set of experiments, the cobalt–chromium disks were scratched along random directions using 320 grit emery paper in accordance with the previously established method of Wang et al., resulting in an average roughness, Ra, of 0.45 m [30]. A total of six pins plus an additional pin serving as a soak control were used for all wear tests.
2.2. Wear testing protocol and apparatus
2.3. Data analysis
A series of six different articulation patterns including a 5 mm × 5 mm square, 4 mm × 6 mm, 3 mm × 7 mm, 2 mm × 8 mm, and 1 mm × 9 mm rectangles, and a 0 mm × 10 mm linear tracking pattern were digitized into an OrthoPODTM (Advanced Mechanical Technology Inc.) multi-directional wear tester. For each wear test, the OrthoPODTM multi-directional wear tester was loaded with six UHMWPE pins as shown in the experimental
Wear rates for each of the six articulation patterns were determined in experiments in which both smooth and rough counterfaces were employed. Gravimetric weight loss per pin was determined approximately every 200,000 cycles. The soak control consistently revealed that the amount of absorption of bovine serum by the UHMWPE specimen was undetectable and therefore any corrections to compensate for fluid absorption were unnecessary. For each group, wear tests were
2. Materials and methods 2.1. UHMWPE starting material
Fig. 2. AMTI OrthoPODTM multi-directional wear tester.
M.E. Turell et al. / Wear 259 (2005) 984–991
continued until at least one million cycles had been reached. The gravimetric weight loss per pin was determined by taking the average weight loss of the six wear tested pins. Prior to being weighed, UHMWPE pins were first washed twice with distilled water followed by an alcohol and acetone rinse, respectively. The pins were dried in air at 25 ◦ C for 15 min prior to determination of weight loss. Gravimetric weight loss was further converted into volumetric wear data by using a density value of 0.943 g/cm3 for UHMWPE. Wear factor values, k, were also calculated for each sample by dividing the volumetric wear rate by the product of the load and the sliding distance per cycle. 2.4. Statistics Statistical analysis was conducted using ANOVA with Fisher’s protected least significant difference (PLSD) posthoc test in which a p-value of less than 0.05 was used to define statistical significance. In addition, single-factor ANOVA using a 95% confidence interval (α = 0.05) was also used to more rigorously assess the statistical significance between groups. Wear factor values from each wear path geometry using both smooth and rough counterfaces were compared against one another in order to establish the level of significance between different groups.
3. Results 3.1. Wear testing Wear tests showed that the wear rates increased incrementally as the wear path decreased in aspect ratio and approached the 3 mm × 7 mm configuration. Interestingly, for both experiments in which smooth counterfaces and rough counterfaces were used, wear reached a maximum when a 3 mm × 7 mm wear path was employed (Fig. 3). In the smooth counterface test series, the wear factor for the 5 mm × 5 mm square path was found to be 2.5 times greater in comparison to the 2 mm × 8 mm path (p < 0.05, ANOVA with Fisher’s PLSD post-hoc test), which agrees remarkably well with the
987
prediction of Eq. (2). For the same comparison in the rough counterface test series, the wear factor for the 5 mm × 5 mm square path was found to be only 1.1 times greater than the wear factor for the 2 mm × 8 mm path. Although this trend was observed, in the case of the rough counterface test, the differences between the 5 mm × 5 mm path and the 2 mm × 8 mm path was not statistically significant. 3.2. Effect of counterface roughness As predicted, counterface roughness had a substantial effect on wear rate. Although standard deviations between groups were large, the differences in wear rates between rough and smooth tests for each wear path geometry were statistically significant (p < 0.05) according to ANOVA (using a 95% confidence interval). This is due to the large number of data points included in each group. An exception is that the differences in the wear factor for the 4 mm × 6 mm rough versus smooth tests were not statistically significant. Using ANOVA (with the 95% confidence interval criteria) within the smooth counterface series, the wear factor of the 1 mm × 9 mm wear path was significantly greater than that of the 0 mm × 10 mm path as was the wear factor of the 3 mm × 7 mm versus the 2 mm × 8 mm. However, the wear factors of 2 mm × 8 mm versus the 1 mm × 9 mm, the 4 mm × 6 mm versus the 3 mm × 7 mm, and the 5 mm × 5 mm versus the 4 mm × 6 mm were not significantly different from one another. Using this same analysis, in the rough counterface set of wear tests, the wear factor associated with the 1 mm × 9 mm path was significantly greater than that of the 0 mm × 10 mm path as was the 3 mm × 7 mm path compared to the 4 mm × 6 mm path. The wear factor of the 2 mm × 8 mm versus the 1 mm × 9 mm path, the 3 mm × 7 mm versus the 2 mm × 8 mm path, and the 5 mm × 5 mm versus the 4 mm × 6 mm path were not statistically different. An important finding of this study was that the counterface roughness affected wear rates to a greater extent in experiments in which more linear rectangular paths were used as compared to the more square-like wear paths. In comparison to the smooth counterface series, the rough counterface tests demonstrated wear rates that were significantly greater than predicted when more linear wear paths were tested. There was a general decreasing trend in the ratio of krough /ksmooth as the aspect ratio of the rectangles decreased (approached the square configuration), as shown in Fig. 4. 3.3. Unified theory of wear
Fig. 3. Comparison of wear factors for both smooth and rough cobalt–chromium counterfaces.
In order to assess the accuracy of the unified theory of wear in predicting the wear rate of UHMWPE based on a particular wear path geometry, the experimentally observed wear factor values for both the smooth and rough counterface tests were plotted as a function of the ratio A/(A + B) (Fig. 5). These curves were then compared to predicted wear factor trends, in which the maximum predicted wear factor was taken to be the value of the wear factor obtained from the
988
M.E. Turell et al. / Wear 259 (2005) 984–991
Fig. 6. Liner fit (with maximum A/(A + B) = 0.3) for wear factor, k, vs. A/(A + B) ratio.
Fig. 4. Comparison of wear factor ratios (rough/smooth).
5 mm × 5 mm square wear path. Analysis revealed that the orientation-softening theory of wear was unable to accurately predict wear over the entire range of rectangles for both the smooth and rough counterface series. In smooth counterface tests, the parameter A/(A + B) was particularly accurate in the prediction of wear rates in cases where the aspect ratio of the rectangular path was highest, such as in the 0 mm × 10 mm to 3 mm × 7 mm wear paths. The monotonic increase in wear factors in the A/(A + B) range of 0.0 and 0.3 for UHMWPE in both rough and smooth counterfaces suggested a linear correlation between wear factor, k, and A/(A + B). Therefore, a linear curve fit was conducted in a separate plot (Fig. 6) for the wear factors in the A/(A + B) range of 0.0–0.3. Fig. 6 shows that the wear factor values from both the smooth and rough counterface tests show a strong linear relationship in this range. In both the smooth and rough counterface experiments, the wear factor increased with increasing A/(A + B) ratio, but only up to a value of 0.3. There were significant deviations in the experimental data from the theoretical linear correlation predicted by the orientational softening theory (Fig. 5). These differences are particularly apparent in the set of experiments
Fig. 5. Wear factor, k, vs. A/(A + B) ratio.
employing rough counterfaces. According to the theoretical model of Wang [28], the maximum wear factor should occur at A = B or A/(A + B) = 0.5. Instead the experimentally observed wear factor peaked at A/(A + B) = 0.3 for both smooth and rough counterface experiments. In fact, the 3 mm × 7 mm showed a 16% increase over the 5 mm × 5 mm square path in the series where smooth counterfaces were used (p < 0.05, ANOVA with Fisher’s PLSD post-hoc test).
4. Discussion 4.1. Wear tests This study was conducted as an extension of a previous study that sought to quantify the effect of the cross-path motion on wear rate of UHMWPE. The aim of this study was to address the effects of counterface roughness on wear rates for different cross-path motions. For a rectangular wear path with a width A and a length B (A < B), the numerical parameter A/(A + B) is a convenient measure to relate wear factor, k, to cross-path motion for rectangles with various aspect ratios, B/A. This numerical parameter, defined by the unified wear model, was relatively accurate in predicting the general trend that wear rates would increase as a function of decreasing aspect ratio (i.e. as a function of the wear path approaching a square configuration) but only in the range of the 1 mm × 9 mm wear path to the 3 mm × 7 mm configuration. For both smooth and rough counterface tests, a linear trend was particularly evident in cases where the aspect ratio of the rectangular path was highest, such as in the 0 mm × 10 mm, 1 mm × 9 mm, and 2 mm × 8 mm wear paths. In both counterface groups, wear factors reached a maximum when a 3 mm × 7 mm wear path was employed, and the trend was no longer observed as the aspect ratio of the wear path continued to decrease beyond this point. The failure of the unified wear model to predict wear rates in some cases for both rough and smooth counterface tests raises ques-
M.E. Turell et al. / Wear 259 (2005) 984–991
tions about some of the underlying assumptions upon which the model is based. For example, for any rectangular wear path with sides A and B, where A < B, the model assumes that orientation on the molecular level occurs in the B direction, while fracture, or the actual wearing of UHMWPE, occurs in the A direction. Based on the results of this study, in which wear reached a maximum for the 3 mm × 7 mm wear path, this assumption may not be a valid description of the orientation and wear processes. It is possible that complete orientation and strain hardening of UHMWPE in the B direction is required for wear to occur solely in the A direction, and that this is only achieved at 7 mm of sliding. It is therefore possible that for the 5 mm × 5 mm and 4 mm × 6 mm wear paths, there is significant reorientation in the A direction as well rather than fracture of the fully oriented, strain hardened UHMWPE fibrils. Reorientation would imply that a different wear mechanism is operative since biaxial orientation prior to wear would suggest that a more sheet-like material wears rather than splitting of fibrillar UHMWPE. These wear and orientation mechanisms appeared to occur in two discrete A/(A + B) ranges, irrespective of the roughness of counterface. Fig. 6 shows that there was a linear correlation between wear factor, k, and the numerical parameter A/(A + B), regardless of counterface roughness. The linear equations for these correlations as well as the R-values (to measure the degree of fit) were obtained for both the smooth and rough counterface tests. In the equation for the line, k = m[A/(A + B)] + c, the constant c (y-intercept) represents the fraction of wear factor that is due to linear, abrasive wear and the constant m (slope) is a measure of the dependence of cross-path wear on the wear path geometry. The following equations were obtained for the linear correlations: A k = 7E–06 − 6E–08 (R = 0.8821) A+B (smooth counterface) k = 1E–05
A A+B
(3)
+ 4E–07 (R = 0.9922) (rough counterface)
(4)
It is evident that the constant c cannot assume a negative value, and the low, negative wear factor value of 6E−08 for linear tracking for smooth heads obtained from the linear fit is merely a consequence of experimental error, which can be taken to be zero or replaced by the experimentally measured positive value of 5.06E−08. The value of m was higher in the rough counterface series of wear tests. The ratio of mrough /msmooth was 1.4, revealing that there was a steeper dependence of wear factor, k, on the numerical parameter, A/(A + B), in the case of rough counterface. For both the smooth and rough counterface, the high R-values indicated a strong linear correlation between the wear factor and A/(A + B). The good fit of the wear factors in the series of
989
wear paths in the range of 0 mm × 10 mm to 3 mm × 7 mm paths suggests that complete orientation followed by fracture of fibrils was the primary mechanism of wear in both of these cases. Independent of the accuracy with which the model investigated in this paper predicted wear rate, the overall experimental results support the hypothesis that wear rate is in fact dependent upon the wear path geometry, counterface roughness and the interplay of these two variables. 4.2. Clinical relevance The finding that wear rate is dependent upon both wear path pattern and counterface roughness has a number of clinically relevant implications. The observation that differences in wear rates, between tests employing rough versus smooth counterfaces, are greater in more linear motion path patterns and that these differences systematically decrease as the wear approaches a square pattern is an important finding. In total joint replacement applications where linear wear is known to be operative and to have a significant impact upon the lifetime of the joint replacement, for example, as is the case in the knee joint, the effects of abrasive wear (simulated by a roughened counterface in this study) are of greater concern. It should be noted that the results of this study reveal a somewhat oversimplification of the wear mechanisms as they occur in clinical application since the effects of third body wear and fatigue-related wear mechanisms were not investigated. A question which remains to be answered is whether it may be possible to predict motion path patterns by conducting gait analysis and if so, whether motion path pattern can be therapeutically manipulated using methods of gait training or by making improvements to implant design. A challenge in addressing these questions lies in the difficulty in assessing gait in patients who require total joint replacement surgery as gait in these individuals may be altered from their normal baseline gait due to the orthopaedic complications, which necessitate surgery in the first place. Similarly, gait patterns in these individuals may be dramatically altered following total joint replacement surgery, making it difficult to predict the effects of implant design in advance of surgery. 4.3. Limitations The results of this study represent an attempt to quantify the effect of the motion path pattern and counterface roughness on wear rate of UHMWPE. It should be noted that the data, in particular the wear factor values, that have been analyzed in this paper represent the results of preliminary wear tests conducted for each wear path geometry to a period of at least one million cycles. A more comprehensive study, and one that would employ more rigorous wear testing would generate a larger number of samples for tests encompassing a broader range of motion path patterns. From such experiments, trends in the values of UHMWPE wear rates would be more reliably generated and would be of a greater level of clinical significance. A study of the morphology of wear
990
M.E. Turell et al. / Wear 259 (2005) 984–991
particles generated from the various motion path patterns for both smooth and rough counterface tests is also necessary (and forthcoming) to obtain a more accurate understanding of the wear mechanisms that are operative during the various wear paths. In addition, the wear tests in this paper employed a constant applied load and it is important for future studies to include wear test loading parameters which more closely model the loading conditions found in knee and hip joints in vivo. The relationship between wear path and wear rates established in this study applies only to uncrosslinked medical grade UHMWPE. Further wear testing on the various crosslinked UHMWPEs currently in clinical use is required for such relationships to be established.
5. Conclusions Wear of UHMWPE as it articulates against a metallic counterface in a particular wear path and under conditions that mimic wear in a total hip replacement prosthesis, likely occurs via two discrete steps for rectangles with an aspect ratio greater than 2.33 (3 mm × 7 mm rectangular path). For such rectangles, the wear tests of this study support the hypothesis that there is orientation or texturing of UHMWPE on parallel edges of the rectangle followed by wear of the textured UHMWPE on the other two parallel edges. However, for rectangles in the aspect ratio range of 1.0–2.33 (5 mm × 5 mm to 4 mm × 6 mm paths), the decline in wear rates for both smooth and rough counterface experiments provide evidence that wear and orientation processes may not occur in discrete stages. In addition, the unified wear model predicts zero wear for linear tracking, which is not the case, especially when more abrasive conditions of wear occur such as the case of a roughened counterface. A more robust model is required to predict wear of UHMWPE during articulation against a metallic counterface along a rectangular path covering the entire range of aspect ratios of rectangles.
Acknowledgments This project was funded through a biomedical engineering grant provided by the Whitaker Foundation and by a fellowship provided by the Orthopaedic Research and Education Foundation.
References [1] A. Gomoll, T. Wanich, A. Bellare, J-integral fracture toughness and tearing modulus measurement of radiation cross-linked UHMWPE, J. Orthop. Res. 20 (6) (2002) 1152–1156. [2] L.C. Duus, H.A. Walsh, A.M. Gillis, E. Noisiez, S. Li, The effect of resin grade, manufacturing method and cross linking on the fracture toughness of commercially available UHMWPE, Trans. Orthop. Res. Soc. 25 (2000) 544.
[3] M. Griffith, M. Seidenstein, D. Williams, J. Charnley, Socket wear in Charnley low friction arthroplasty of the hip, Clin. Orthop. 137 (1978) 37–47. [4] J. Livermore, D. Ilstrupand, B. Morrey, Effect of femoral head size on wear of the polyethylene acetabular component, J. Bone Joint Surg. 72A (4) (1990) 518–528. [5] R. Hall, A. Unsworth, P. Siney, B. Wroblewski, Wear in retrieved Charnley acetabular sockets, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 197–207. [6] H. Oonishi, E. Tsuji, Y. Kim, Retrieved total hip prostheses. Part I. The effects of cup thickness, head size and fusion defects on wear, J. Mater. Sci.: Mater. Med. 9 (1998) 394–401. [7] I.C. Clarke, J.M. Kabo, Wear in total hip replacement, in: H.C. Amstutz (Ed.), Hip Arthroplasty, Churchill Livingstone, New York, 1991. [8] G. Lewis, Polyethylene wear in total hip and knee arthroplasties, J. Biomed. Mater. Res. (Appl. Biomater.) 38 (1997) 55–75. [9] H. Amstutz, P. Campbell, N. Kossovsky, I. Clarke, Mechanisms and clinical significance of wear debris-induced osteolysis, Clin. Orthop. Relat. Res. 276 (1992) 7–18. [10] W. Harris, The problem is osteolysis, Clin. Orthop. Relat. Res. 311 (1995) 46–53. [11] T. Schmalzried, L. Kwong, M. Jasty, R. Sedlacek, T. Haire, D. O’Connor, C. Bragdon, J. Kabo, A. Malcolm, W. Harris, The mechanism of loosening of cemented acetabular components in total hip arthroplasty: analysis of specimens retrieved at autopsy, Clin. Orthop. Relat. Res. 274 (1992) 60–78. [12] P. Campbell, S. Ma, B. Yeom, H. McKellop, T. Schmalzried, H. Amstrutz, Isolation of predominantly submicron-sized UHMWPE wear particles from periprosthetic tissues, J. Biomed. Mater. Res. 29 (1995) 127–131. [13] A. Shanbhag, J. Jacobs, T. Glant, J. Gilbert, J. Black, J. Galante, Composition and morphology of wear debris in failed uncemented total hip replacement, J. Bone Joint Surg. 76B (1994) 60–67. [14] J. Savio, L. Overcamp, J. Black, Size and shape of biomaterial wear debris, Clin. Mater. 15 (1994) 101–119. [15] K. Margevicius, T. Bauer, J. McMahon, S. Brown, K. Merritt, Isolation and characterization of debris in membranes around total joint prostheses, J. Bone Joint Surg. 76A (1994) 1664–1676. [16] P. Campbell, P. Doorn, F. Dorey, H. Amstutz, Wear and morphology of ultra-high molecular weight polyethylene wear particles from total hip replacements, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 167–175. [17] A. Shanbhag, H. Bailey, D. Hwang, C. Cha, N. Eror, H. Rubash, Quantitative analysis of ultrahigh molecular weight polyethylene (UHMWPE) wear debris associated with total knee replacements, J. Biomed. Mater. Res. 53 (1) (2000) 100–110. [18] B. Ramamurti, C. Bragdon, D. O’Connor, J. Lowenstein, M. Jasty, D. Estok, W. Harris, Loci of movement of selected points on the femoral head during normal gait, J. Arthroplasty 11 (7) (1996) 852–855. [19] A. Wang, D. Sun, S.-S. Yau, B. Edwards, M. Sokol, A. Essner, V. Polineni, C. Stark, J. Dumbelton, Orientation softening in the deformation and wear of ultra-high polyethylene, Wear 203–204 (1997) 230–241. [20] V. Saikko, T. Ahlroos, Type of motion and lubricant in wear simulation of polyethylene acetabular cup, Proc. Inst. Mech. Eng. H: J. Eng. Med. 213 (1999) 301–310. [21] C. Bragdon, D. O’Connor, J. Lowenstein, M. Jasty, W. Syniuta, The importance of multidirectional motion on the wear of polyethylene, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 157–166. [22] A. Wang, C. Stark, J. Dumbleton, Mechanistic and morphological origins of ultra-high molecular weight polyethylene wear debris in total joint replacement prostheses, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 141–155. [23] A. Wang, V. Polineni, A. Essner, M. Sokol, D. Sun, C. Stark, J. Dumbleton, The significance of non-linear motion in the wear screening of orthopaedic implant materials, Test. Eval. 25 (2) (1997) 239–245.
M.E. Turell et al. / Wear 259 (2005) 984–991 [24] A. Wang, A. Essner, V. Polineni, C. Stark, J. Dumbleton, Lubrication and wear of ultra-high molecular weight polyethylene in total joint replacements, Tribol. Int. 31 (1–3) (1998) 17–33. [25] L. Caravia, D. Dowson, J. Fisher, B. Jobbins, The influence of bone and bone cement debris on counterface roughness in sliding wear tests of ultra-high molecular weight polyethylene on stainless steel, Proc. Inst. Mech. Eng. H: J. Eng. Med. 204 (1) (1990) 65–70. [26] T.D. Brown, K.J. Stewart, J.C. Nieman, D.R. Pedersen, J.J. Callaghan, Local head roughening as a factor contributing to variability of total hip wear: a finite element analysis, J. Biomech. Eng. 124 (6) (2002) 691–698.
991
[27] D. Bennett, J. Orr, R. Baker, Movement loci of selected points on the femoral head for individual total hip arthroplasty patients using three-dimensional computer simulation, J. Arthroplasty 15 (7) (2000) 909–915. [28] A. Wang, A unified theory of wear for ultra-high molecular weight polyethylene in multidirectional sliding, Wear 248 (2001) 38–47. [29] M. Turell, A. Wang, A. Bellare, Quantification of the effect of cross-path motion on the wear rate of ultra-high molecular weight polyethylene, Wear 255 (2003) 1034–1039. [30] A. Wang, V.K. Polineni, C. Stark, J.H. Dumbleton, Effect of femoral head surface roughness on the wear of ultrahigh molecular weight polyethylene acetabular cups, J. Arthroplasty 6 (1998) 615–620.