 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1996 The Journal of Bone and Joint Surgery, Inc. Wear of Polyethylene Cups in Total Hip Arthroplasty. A Study of Specimens Retrieved Post Mortem* , ARLINGTON , CLEVELAND, ARLINGTON, VIRGINIA, CLEVELAND, OHIOInvestigation performed at The Anderson Orthopaedic Research Institute, Arlington, and The Cleveland Clinic Foundation, Cleveland
The wear of polyethylene components of total hip replacements has previously been studied radiographically and by analysis of retrieved components. The extent of wear, however, has largely been determined from components retrieved at reoperation and has therefore been based on a subset of patients in whom the prosthesis is more likely to show excessive wear. The study of cups retrieved post mortem provides a better indication of the extent of wear of well functioning components. Twenty-six polyethylene liners were retrieved at autopsy from nineteen patients, after a mean of ninety-one months (range, thirty-three to 206 months) in situ. Each component had articulated with a thirty-two-millimeter-diameter femoral head. The components were evaluated visually for evidence of polyethylene wear such as pitting, scratching, and burnishing. Additionally, a shadowgraph technique was used to examine molds of the inner surface of the liner to determine the direction and extent of wear. The mean extent of linear wear of the retrieved liners was 0.45 millimeter (range, 0.17 to 1.07 millimeters), and the mean rate of wear was 0.07 millimeter (range, 0.02 to 0.18 millimeter) per year. The mean volumetric wear was 245.3 cubic millimeters (range, 13.0 to 779.1 cubic millimeters), and the mean rate of volumetric wear was 39.8 cubic millimeters (range, 1.0 to 131.3 cubic millimeters) per year. The mean rate of wear for the twenty-six liners was 45 to 69 per cent less than the rates reported in the literature for polyethylene liners retrieved at reoperation. Examination of the articulating surface did not reveal gross evidence of surface failure such as delamination or fatigue cracks. In general, the surfaces were merely burnished and scratched. The rates of wear of press-fit, metal-backed liners were significantly higher than those of all-polyethylene cemented components (p < 0.05). Additionally, statistical analysis revealed no correlation between wear and the patient's age, weight, or gender; the duration of implantation; or the thickness of the polyethylene. These results indicate that wear of the polyethylene of well functioning hip replacements is not as excessive as reported previously.
Polyethylene wear and subsequent generation of polyethylene particles have been reported as causes of osteolysis around and loosening of total joint replacements1,5,13,16-18. It is therefore important to study the mechanisms and extent of degradation of polyethylene components of total joint replacements in situ. Although the extent of polyethylene wear can be estimated radiographically, quantification of a three-dimensional wear process on the basis of a two-dimensional radiograph is difficult. Other means of quantifying the wear process involve clinical examination of retrieved polyethylene components. These specimens, however, are typically retrieved at the time of reoperation for component failure. As a result, the wear process studied most likely does not represent the wear of a well functioning component. It is therefore important to examine well functioning components to describe patterns of wear. Specimens retrieved at autopsy provide an excellent source for such a study. The purpose of the current study was to determine the long-term patterns of wear of total hip-replacement components retrieved post mortem. Additionally, we analyzed the relationships between wear and clinical variables such as the patient's age and weight at the time of the operation, gender, the duration of the implant in situ, the type of fixation of the cup, and the thickness of the polyethylene.
Twenty-six polyethylene liners retrieved at autopsy from nineteen cadavera were included in this study. The anatomical orientation of each liner within the acetabulum was marked before removal, and care was taken to avoid damage of the polyethylene during removal. Initially, the morphological characteristics of the surface of each polyethylene liner were examined at a magnification of ten to sixty-three times with a stereoscopic zoom microscope (Nikon, Melville, New York). Semiquantitative gross examination of the articulating surface of each component was performed to identify evidence of polyethylene wear, including pitting, scratching, burnishing, abrasion, embedded particles of debris, and permanent plastic deformation. Pitting was defined as pinhole-sized circular or diamond-shaped dents in the articular surface. Scratching was identified as unidirectional lines cut into the polyethylene. Burnishing was represented by shallow multidirectional lines swirled on the articular surface, resulting in a polished appearance. Abrasion was characterized by areas where particles of material had been removed from the articular surface because of repeated rubbing against another surface. Particles of cement or metal embedded in the polyethylene were also identified. Finally, areas of the polyethylene that had deformed from the original shape were described as permanent plastic deformation. Care was taken to identify iatrogenic damage and to exclude it from the results. On the basis of its in situ orientation, each polyethylene liner was divided by superior-inferior and anterior-posterior lines into four anatomical quadrants for analysis. The quadrants were graded according to the types of wear already described and the gross estimate of the percentage of the articulating area that was affected. This semiquantitative system is similar to that developed at The Hospital for Special Surgery7,19. A grade of 0 was assigned if the type of wear was absent in a particular quadrant; a grade of 1, if less than 10 per cent of a quadrant's surface area was affected; a grade of 2, if 10 to 50 per cent of the surface was affected; and a grade of 3, if more than 50 per cent of the surface was affected. The maximum wear score for one quadrant was 18 (that is, a grade of 3 for all six possible types of wear). The wear scores for each quadrant were then combined to give an over-all score representing the relative amount of polyethylene wear of each liner. Next, linear wear was determined with a shadowgraph technique described by Hashimoto et al. With this method, a fine-grain methylmethacrylate cast (Facsimile Compound; Flexbar Machine, Islandia, New York) is made of the articular surface of each liner. Vertical and horizontal profiles of each mold are projected at a magnification of ten times onto templates of concentric circles. For each view, the center of curvature of the original surface and that of the worn surface are determined. These points represent the center of the femoral ball at the time of the original arthroplasty and at the time of death, respectively. The locations of these two points are calculated in two perpendicular planes to reconstruct a three-dimensional wear vector. This vector represents both the direction and the extent of migration of the femoral ball into the polyethylene liner. Volumetric wear was calculated from measurements of the three-dimensional linear wear vector, with use of the equation developed by Hashimoto et al. Because deformation due to polyethylene wear is difficult to distinguish from that due to polyethylene creep, all calculations of wear actually represent a combination of wear and creep. Ten polyethylene specimens were matched with unimplanted control liners of the same size and design, and volumetric wear was calculated with use of the gravimetric method6. The volumetric wear of these ten liners was measured directly with this method and compared with that calculated with use of the equation of Hashimoto et al. For gravimetric analysis, the polyethylene liners were positioned, with use of a level and a specially designed holder, so that the face of the liner was exactly horizontal. The precise weight of ethyl alcohol needed to fill each liner was then determined and compared with that needed to fill the unimplanted control liner of the same size and design. Weight was measured promptly to minimize the effects of evaporation. Seven measurements were obtained for each cup; the low and high values were discarded, and the five remaining measurements were averaged together. Weight was converted into volume by precise measurement of the weight of one milliliter of alcohol. Volumetric wear was determined as the difference between the cavity volume of the control polyethylene liner and that of the retrieved polyethylene liner. The final anteroposterior radiograph of the pelvis of each patient was analyzed for loosening of the cup (defined as migration of more than three millimeters) and signs of osteolysis or radiolucency around the cup-bone or cement-bone interface. Clinical records were also examined to determine the performance of the implant. Patients who had had a score of more than 10 points (of a possible 12 points) according to the criteria of Merle d'Aubigné and Postel and in whom the component had been radiographically stable at the last follow-up evaluation were considered to have had a well functioning total hip replacement. The angle of inclination, or the abduction angle, of the cup within the acetabulum was measured on the anteroposterior radiographs, between the horizontal teardrop line and a line drawn through the face of the cup (Fig. 1). Because the abduction angle is measured on a two-dimensional radiograph, it does not correspond with the direction of the three-dimensional wear vector derived with use of the shadowgraph technique. The three-dimensional wear vector was therefore mathematically adjusted to determine its two-dimensional component in the plane of the anteroposterior radiograph. The angle of this adjusted wear vector and the abduction angle were then combined to determine the mean direction of wear in the plane of the anteroposterior radiograph relative to horizontal (Fig. 2). This measurement, the anatomical wear angle, was used to describe an anatomical direction of wear.
Several statistical analyses were performed. First, multiple linear regression analysis was used to determine possible associations between polyethylene wear and the patient's age and weight at the time of the operation, the duration that the implant had been in situ, and the thickness of the polyethylene. Thickness was measured with a digital caliper at an area of the liner where the original machine marks were still visible, indicating an area where the liner had not been worn. Linear regression analysis was also used to determine the correlation between calculated volumetric wear and volumetric wear as measured with the gravimetric method. Second, to determine if the rate of wear had decelerated over time rather than remaining constant, as a linear relationship would suggest, we analyzed the relationship between wear and the duration that the implant had been in situ with use of a logarithmic curve fit. A logarithmic curve describes a situation in which the rate of wear is most severe initially and then decelerates over time. Finally, the Wilcoxon rank-sum test was used to determine any significant difference (p < 0.05) between the mean wear scores and the mean rates of wear for (1) the cemented all-polyethylene compared with metal-backed components, (2) the men compared with the women, (3) the cobalt-chromium compared with ceramic femoral heads, and (4) the modular compared with non-modular heads.
The twenty-six polyethylene liners were retrieved at autopsy from nine men and ten women, who had had a mean age of seventy years (range, thirty-nine to eighty-seven years) and a mean body weight of 70.4 kilograms (range, 45.4 to 107.5 kilograms) at the time of the operation. The mean duration that the implant had been in situ was ninety-one months (range, thirty-three to 206 months). Eighteen of the polyethylene liners had been used with a metal-backed, press-fit acetabular cup, and eight were non-metal-backed components that had been inserted with cement (Table I). All of the polyethylene components had articulated with a thirty-two-millimeter-diameter femoral head; five heads were ceramic and twenty-one were cobalt-chromium. All five ceramic heads and four of the cobalt-chromium heads were modular. At the last clinical follow-up evaluation, at a mean of 4.4 years after the arthroplasty, the mean Merle d'Aubigné and Postel pain and walking score for seventeen of the nineteen patients was 11.7 of a possible 12 points. The remaining two patients did not have complete clinical records with this information, but they did have radiographically stable components.
Linear Wear and Direction
From the lateral view of the liner, the mean wear vector was oriented a mean of 13.7 degrees (range, 139.4 degrees anteriorly to 168.0 degrees posteriorly) away from the most superior point of the face of the liner toward the posterior side (Fig. 3). From the anteroposterior view of the liner, this wear vector was oriented a mean of 19.3 degrees (range, 57.2 degrees laterally to 59.7 degrees medially) away from the face of the liner, toward the apex (Fig. 3).
Radiographic Evaluation The mean angle of inclination (the abduction angle) of the cup, as determined from the anteroposterior radiographs, was 47.5 degrees (range, 34 to 59 degrees) from the teardrop line. The mean direction of wear of the polyethylene liner in the plane of the anteroposterior radiograph (the anatomical wear angle) was 73.3 degrees (range, -40.8 to 119.9 degrees) from a horizontal line drawn through the center of the head (Fig. 2).
Volumetric Wear Analysis of the ten polyethylene specimens with matched, unimplanted control liners revealed a significant correlation (r2 = 0.86, p < 0.01) between the measurements of volumetric wear made with the gravimetric method and the volumetric wear as calculated from the measurements of linear wear. The slope of the regression line was 1.035, indicating that the calculated change in volume represented only a slight underestimation of the change in volume as measured with the gravimetric method.
Gross Examination of the Articulating Surfaces Burnishing was the most common type of polyethylene wear and the most severe, occurring in all of the liners that were studied. The mean score for burnishing was 10.1 of a possible 12, indicating that most of the articulating surface of the liner was burnished. Scratching and pitting also occurred in all liners, but they were less severe, with mean scores of only 7.3 and 5.1, respectively. Permanent plastic deformation, embedded particles of debris, and abrasion were seen in seventeen (65 per cent), eleven (42 per cent), and seven (27 per cent) of the twenty-six liners. In all but two liners, the embedded debris consisted of metal particles. The two liners associated with non-metal embedded debris had been cemented. Although we believe that these particles consisted of bone cement, we did not positively identify them. The wear scores for embedded particles of debris, plastic deformation, and abrasion were all less than 1.6 of a possible 12, indicating that the wear was limited to a very small area of the liner and was not severe.
Statistical Analysis Linear regression analysis revealed no correlation between either the total amount of wear or the rate of wear (linear or volumetric) and the patient's age or weight at the time of the operation, the abduction angle, or the thickness of the polyethylene (r2 < 0.16 for all variables). In addition to the lack of a linear relationship between wear and time, there was no curvilinear relationship. The extent of linear and volumetric wear was also regressed, in a multiple linear regression analysis, with the duration of implantation, the patient's weight, and the thickness of the polyethylene, to reveal possible interactive effects. Again, no correlations were found. Because the rates of wear for the cemented and metal-backed components were significantly different (p < 0.05), these two groups were separately regressed with all of the clinical variables already mentioned. As before, we found no significant correlations between wear and the clinical variables for either type of component.
Previous studies of polyethylene wear of acetabular components have focused on measurements that were obtained either radiographically or from examination of components retrieved at reoperation. Although radiographs allow a good estimation of polyethylene wear, the direction and extent of three-dimensional migration of the femoral head into a polyethylene liner are difficult to identify from a two-dimensional radiograph. Additionally, the polyethylene components typically evaluated for wear are retrieved at reoperation because of failure and therefore do not allow an accurate assessment of wear of normal, well functioning hip prostheses. It is important to assess the wear of polyethylene components of well functioning total joint replacements. We studied specimens that were retrieved post mortem, after a mean of ninety-one months in situ, to determine such wear. Like authors of similar studies, we used the shadow-graph technique to measure the direction and extent of linear wear8,11,20. Volumetric wear was calculated on the basis of linear wear with use of the equation developed by Hashimoto et al. Because this equation differed from those reported previously in the literature11,12, we validated the volumetric measurements with those made with a gravimetric method on ten liners with matched, unimplanted controls of the same size and type. For these ten liners, there was a significant correlation (r2 = 0.86, p < 0.01) between the measurements obtained with the two methods, with the calculated volume representing a slight underestimation of the measured gravimetric volume. We are therefore confident of the accuracy of our volumetric results. We found a mean rate of linear wear of 0.07 millimeter (range, 0.02 to 0.18 millimeter) per year and a calculated mean rate of volumetric wear of 39.8 cubic millimeters (range, 1.0 to 131.3 cubic millimeters) per year. Most investigators who have studied polyethylene wear with use of a mold technique have reported on the linear wear of polyethylene liners retrieved during revision operations; in those studies, the mean rates of wear ranged from 0.127 to 0.224 millimeter per year4,8,11,20. The liners that we retrieved post mortem had a 45 to 69 per cent lower mean rate of wear than did components retrieved at revision. However, components retrieved during revision procedures are expected to have high rates of wear, as they are typically removed because of failure. The rates of wear in the current study compare well with those in other reports of components retrieved at autopsy. Jasty et al. reported a mean rate of volumetric wear of 39 ± 17 cubic millimeters per year, and Berzins et al. reported mean rates of 0.11 millimeter and fifty-nine cubic millimeters per year for specimens retrieved at autopsy. We found that the rates of wear of the press-fit, metal-backed acetabular components were higher than those of the all-polyethylene, cemented components. The mean rate of linear wear of the cemented components was 0.04 millimeter per year, and the corresponding rate for the metal-backed components was 0.08 millimeter per year. However, several of the metal-backed liners in our study were subject to edge-loading (loading on the rim instead of the dome) and contained areas of unsupported polyethylene. Moreover, the metal-backed liners were significantly thinner than the cemented components (mean thickness, 4.3 compared with 5.9 millimeters; p < 0.05). Although we did not find a global relationship between the thickness of the polyethylene and wear, the combination of thinner polyethylene, metal backing, and unsupported polyethylene may have affected the wear process. The direction of wear varied widely in the liners that we examined. From an anteroposterior view, the mean wear vector was oriented 19.3 degrees medial to the face of the liner, toward the apex (Fig. 3). Some wear vectors were even directed lateral to the face of the liner. Although lateral wear was also reported by Hashimoto et al., it is relatively uncommon in the literature. Hashimoto et al. postulated that lateral wear vectors might be associated with cups positioned vertically in the pelvis, but they were unable to prove this because of a lack of clinical radiographs. In the current study, we were able to compare, on anteroposterior radiographs, the angle of inclination of the cup (the abduction angle) in hips with lateral wear vectors with the same angle in hips without lateral wear vectors. With the numbers available, we found no significant difference in the angle of inclination of the cup between these two groups. We also used the anteroposterior radiographs to examine the direction of wear with respect to the position of the cup in the body. Surprisingly, the mean anatomical angle of wear was directed 73.3 degrees from the horizontal (Fig. 2). Although this direction of wear is not intuitive on the basis of our knowledge of weight-bearing forces, it has been reported previously11. This direction of wear may be an important consideration in future cup designs and operative positioning. The liners had little gross evidence of fatigue wear mechanisms. There was no evidence of fatigue fractures, and there was little delamination or permanent plastic deformation of the articular surface. The most common types of wear observed, burnishing and scratching, were consistent with abrasive wear. Although other authors have reported associations between wear and the duration of implantation, the thickness of the polyethylene, and the weight of the patient1,10,12,19, we found no relationship between any clinical variable and linear wear, volumetric wear, or the visually evaluated wear score. This may indicate that clinical variables have little influence on polyethylene wear of well functioning hip replacements. Alternatively, it may be related to the variety of cup designs included in our study (Table I). A study of a larger number of liners retrieved post mortem and all of one design would allow a better assessment of the subtle effects of clinical variables on wear. However, we were unable to acquire such a specific group of components.
It may be argued that analyses of components retrieved post mortem do not permit an accurate assessment of wear because patients are often sedentary in the years before their death. Therefore, there is less polyethylene degradation in specimens obtained post mortem than in a corresponding group of living patients. Although the mean age of the nineteen patients in our study was seventy years, sixteen patients were able to walk outdoors independently at the time of the last clinical follow-up evaluation, at a mean of 4.4 years after the arthroplasty. Only one patient was known to be inactive, and the remaining two patients did not have complete clinical records that we could use to assess this information. Furthermore, the mean Merle d'Aubig We emphasize that the amount of wear quantified in our study does not necessarily equal the volume of polyethylene particles removed from the liner. Our calculations included the surface deformation associated with migration of the femoral head into the polyethylene socket and therefore represent the sum of wear and creep. Like previous investigators, we were unable to quantify the amount of creep, so we have referred to this combined deformation process as wear. Methods that are currently used to decrease polyethylene wear include avoiding femoral heads larger than twenty-eight millimeters in diameter, polyethylene less than six millimeters thick, and cup designs in which the metal shell and the back of the polyethylene liner are not congruent9,15. Although the liners in our study did not conform to these criteria, they had substantially less wear than polyethylene components that were previously reported on in the literature4,8,11,20. This is an important finding, indicating that even with a non-ideal liner design the polyethylene wear of well functioning total hip replacements is not as extensive as reported previously4,8,11,20. These results provide invaluable information on wear processes under circumstances not associated with failure of the implant. However, studies comprising a greater number of postmortem specimens of the same design are still needed to assess further the subtle effects of clinical factors on the wear of well functioning implants.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Technorati What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
and Postel pain and walking score for these seventeen patients at the last follow-up evaluation was 11.7 of a possible 12 points. According to this scale, 12 points indicates a well functioning hip with no pain and no limp. We therefore believe that our findings provide representative and valuable information on the wear of well functioning components. 
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |