 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1998 The Journal of Bone and Joint Surgery, Inc.
The Multifactorial Nature of Polyethylene Wear in Vivo*In the 1980s, the focus of the joint-replacement industry was on fixation of the implants, and progress has been made both with devices that are inserted with cement and with those that are inserted without cement. As has always been the case in the history of joint replacement, each advance has unmasked another limitation. More durable fixation allows for more frequent, more intense, and more variable use of the joint, and the indications for the procedure have gradually been expanded to include younger and more active patients. With greater anticipated longevity for both patients and devices, there are justified concerns about long-term skeletal remodeling, but this has not yet been demonstrated to be a widespread clinical problem. The current problem is osteolysis, which most commonly occurs in association with polyethylene wear particles. The clinical assessment of the performance of a bearing couple has traditionally been based on radiographic studies. With this approach, which presumably measures a change in the thickness of the polyethylene, there is a tendency to attribute the change in thickness to variables inherent to the polyethylene bearing. This is especially true when issues related to the manufacturing and sterilization of polyethylene have been put in the spotlight by both science and industry. We must be cautious about this tendency. As polyethylene wear is a focal issue in joint replacement, it is essential that we appreciate the complexities of studying such wear in vivo. The October 1997 issue of The Journal contained an article by Livingston et al. entitled, "Complications of Total Hip Arthroplasty Associated with the Use of an Acetabular Component with a Hylamer Liner."18 While we share the authors' concern about the in vivo performance of Hylamer, we wish to point out that there were differences in their comparison groups, in addition to the difference in the type of polyethylene, that also affect wear. Livingston et al.18, referring to a previous report, compared the average rate of wear of six of 143 Hylamer liners, which had been specifically identified because of a high rate of wear, with the average rate of wear in several series rather than with the average rate of wear for the high-wear components in those series. There are components with high rates of wear in all series (Table I). For example, in one of the series referred to by Livingston et al., there were rates of wear of as high as 1.41 millimeters per year13. This rate is substantially higher than the average for that series and higher than the rates for the Hylamer liners reported on by Livingston et al. If one looks not only at average rates of linear wear but also at the range of rates of wear, one realizes that all studies that included a range demonstrated substantial variability (Table I). Regardless of the duration of follow-up, there are hip components that have no or only slight radiographically measurable wear and there are components that demonstrate wear that is several times greater than the average for that study. Such large case-to-case variations in rates of wear have not previously been explained simply by differences in the wear resistance of the polyethylene28.
Polyethylene wear in vivo is multifactorial with a complex interaction of variables; it is therefore not surprising that rates of polyethylene wear are highly variable. There are patient-related variables, such as age and gender, that are associated with the activity of the patient and the use of the prosthesis. There are variables related to the hip prostheses, which include all aspects of the acetabular and femoral implants (not just the polyethylene). There are also variables related to the operative procedure for implantation, which include operative techniques and the initial, as well as the long-term, fixation of the implants. Loosening of the implant can adversely affect wear, and vice versa. These variables are important as they can affect the loads on and the motions of the bearing and the degree of three-body-wear mechanisms. There is also variability due to differences in the method with which the wear is assessed and to limitations of the techniques for measurement. In vivo rates of wear are higher in the short term and decrease with time for several reasons. Penetration of the femoral head into the acetabular polyethylene is due to a combination of creep (plastic deformation) and wear. Creep decreases exponentially with time, such that most of the penetration that occurs after the first twelve to eighteen months is due to wear. There is also an initial period of higher wear from so-called running-in of the bearing: with use, the contacting surface of the polyethylene wears into a higher degree of conformity with the specific femoral head with which it articulates, and this results in a larger contact surface, lower contact stresses, and a lower rate of wear26,27,33. Furthermore, with modular components, a change in the position of the polyethylene liner relative to the metal shell can also cause a change in the relative position of the femoral head27. These issues should be considered when rates of wear are compared among hips that have been followed for different amounts of time. A fundamental limitation of all radiographic studies of wear is that clinical rates of wear have traditionally been expressed with the use of time as the denominator. This has been done for reasons of convenience, not accuracy. More appropriately, the number of cycles should be used as the denominator in in vitro laboratory studies of simulated wear. Similar to the wear of a set of automobile tires, the wear of a prosthetic hip is a function of use or the number of cycles; it is not simply a function of time. The assumption made in clinical studies is that the activity of patients who have a joint replacement—that is, the actual use of the joint or the number of cycles to which the bearing is subjected—is similar from patient to patient or, if it is not, any differences will tend to "average out" with a large sample size. In view of the broad range of patients who now have total joint implants, the limitations of this assumption must be recognized. We reported on the walking activity, as measured with an electronic digital pedometer, of 111 patients who had a total joint replacement24. These patients averaged just a little more than 0.9 million cycles per year. The most important result, however, was that there was a forty-fivefold difference in the number of gait cycles between the least active and the most active patient. The most active patient averaged 3.2 million cycles per year, about 3.6 times higher than the average. Age was associated with daily walking activity (p = 0.048) but with a high degree of variability (standard deviation, 3040 steps per day). Patients who were less than sixty years old walked about 30 per cent more on the average than those who were sixty years old or more (p = 0.023). Men walked about 30 per cent more on the average than women (p = 0.037). Men who were less than sixty years old walked about 40 per cent more on the average than the other patients (p = 0.011). These data demonstrate that the activity of the patient can contribute substantially to the variability in rates of wear seen in in vivo studies. Many clinical studies of wear have involved retrospective comparisons of wear on the basis of a specific variable, such as the type of polyethylene, the type of femoral head, the presence of a metal backing, or the type of acetabular or femoral fixation. The strength of the conclusions of these studies is limited because of the tremendous number of potentially confounding variables. Furthermore, caution should be exercised when the results are extrapolated to other reconstructions with the same generic variable. An example is the issue of metal backing. One study4 indicated an increase in the rate of polyethylene wear with a specific type of metal-backed acetabular component designed to be inserted with cement compared with that associated with an all-polyethylene component designed to be inserted with cement. In contrast, in other reports, a different type of metal-backed acetabular component was associated with lower rates of polyethylene wear3,14. The issue cannot be as simple as the presence or absence of metal backing. The apparently contradictory results in these series could be due to specific differences in the design and manufacture of the different types of metal-backed acetabular components or to the fact that the femoral heads were composed of titanium alloy in one study and of cobalt-chromium alloy in the others, or a combination of these or other factors. Livingston et al. stated: "As with the initial six failed arthroplasties, there were no identifiable risk factors associated with the latter five."18 Youth has generally been considered a risk factor for wear. The average age of the eleven patients who had a revision because of a high rate of wear in the study by Livingston et al. was only forty-four years. Wear is a function of use, not time. With the same amount of time in situ, a specific bearing combination in a younger patient will, in general, show a greater linear penetration than the bearing combination in an older patient because of the greater amount of use, in general, by the younger patient24,27. Even though Livingston et al. acknowledged differences in the ages of their patients in Groups 1 (DePuy stem, cobalt-chromium femoral head, and Hylamer liner), 1a (DePuy stem, alumina femoral head, and Hylamer liner), 2 (Osteonics stem, cobalt-chromium femoral head, and Hylamer liner), and 3 (Osteonics stem, cobalt-chromium femoral head, and conventional polyethylene liner), they stated that age did not correlate with the rate of wear. However, Table I from their article indicates that a younger age is associated with a higher rate of wear, as does our plot of the linear wear rates versus the patient ages given in the study by Livingston et al. (Fig. 1). Group 2, the 138 hips in which the components of the implant had been made by different manufacturers, is an outlier. When the rate of wear is plotted against age for the groups in which both bearing components had been made by the same manufacturer, the correlation coefficient (r2) is greater than 0.99 (Figs. 2 and 3).
If there were an inherent difference between the wear resistances of the polymers, then one would expect a difference in the rates of wear between groups that were matched for other covariables, such as age. This is not the case in the study by Livingston et al.18. In the groups with similar ages and femoral components inserted with cement, Hylamer liners that articulated with a DePuy femoral head had virtually the same low rate of wear (average, 0.13 millimeter per year) as conventional polyethylene liners that articulated with an Osteonics femoral head (average, 0.12 millimeter per year). The patients in whom a Hylamer liner articulated with a DePuy femoral head were actually slightly younger than the patients who had a conventional polyethylene liner and an Osteonics femoral head. The low rate of wear of Hylamer in these patients suggests that there were factors other than the inherent wear resistance of the polymer that contributed to higher rates of wear in the other groups of patients. The countersurface (the femoral head) is a factor that, in previous studies, has been demonstrated to affect polyethylene wear22,23,33. Despite having a slightly higher average age (62.8 compared with 58.3 years; average difference, 4.5 years), Group 2 (patients with an Osteonics cobalt-chromium femoral head) had a higher average rate of wear (0.29 compared with 0.20 millimeter per year) than Group 1 (patients with a DePuy cobalt-chromium femoral head)18. In this short-term study, the higher rate of linear penetration in the Group-2 hips may, at least in part, have been the result of a mismatch of manufacturing tolerances resulting in a higher degree of creep and running-in of the bearing combination in that group. Another factor that can increase the rate of wear is suggested by Figures 3-A through 3-D in the article by Livingston et al.18. Although the published images are somewhat blurred, the radiograph made two years postoperatively suggests that there is radiolucency at the metal-cement interface, indicative of so-called debonding of the femoral component. This radiolucency was followed by progressive femoral endosteal osteolysis, but there was no apparent pelvic osteolysis. These features raise the possibility that loosening with local fragmentation of the cement mantle caused the femoral osteolysis. The position of Livingston et al. was that a high rate of wear of the bearing surface caused the femoral osteolysis. If the osteolysis were due to Hylamer wear particles, this would imply that the particles migrated to the femoral endosteum through the space between the stem and the cement mantle. Another explanation for the outcome in this patient, suggested by these radiographs, is that loosening of the femoral component at the stem-cement interface generated metal and cement particles, which could account for both the localized osteolysis and an increased rate of wear from three-body mechanisms. Livingston et al.18 did not indicate any criteria for the selection of the different implants used by the five surgeons in their study, and they did not discuss any differences in outcome among the surgeons. The data indicate that ceramic femoral heads were used in the youngest patients; implants inserted without cement were preferentially used in younger patients; and, regardless of the femoral implant, Hylamer was used in younger patients. In retrospective studies such as the one by Livingston et al., there are limitations related to demographic variables such as age and gender. With surgeon selection bias, a particular implant or particular implants may be selected on the basis of the anticipated activity level of a particular patient, regardless of age, gender, or other demographic data. For example, a forty-year-old man who had symptomatic coronary artery disease would be expected to be less active than a sixty-five-year-old female aerobics instructor. It is probable that the surgeons in the study by Livingston et al. selected bearing implants that they believed to be more durable, such as Hylamer liners and ceramic femoral heads, for use in more active patients, regardless of age and gender. This would explain the apparently paradoxical result of higher wear associated with alumina ceramic heads. Livingston et al.18 apparently presumed that failure to find a significant difference between two groups at the 95 per cent confidence level means that there is no difference between them11. This error affects the presentation of the data, as in Table I where the authors reported "NS," meaning that "no significant difference could be detected," for certain subgroups rather than reporting the average rates of wear for those subgroups. This suggests that the actual values were, in fact, not equal. Given the small size of some of the subgroups, the values could have differed substantially without having been significantly different. To evaluate the relative wear resistance of different bearing materials, in vitro studies are conducted under carefully controlled and monitored conditions. The hip-simulator method of McKellop et al., which has been demonstrated to produce wear similar to that found in vivo20, demonstrated a 9 per cent reduction in the rate of wear of Hylamer compared with that of conventional GUR 415 polyethylene19. This difference had low significance (p = 0.5). One set of controlled conditions is used in hip-simulator studies. As we discussed, the conditions and the rates of wear in vivo are highly variable. From a materials-testing perspective, the report by Livingston et al.18 does not diminish the value of hip-simulator testing. It has been suggested that, because of greater stiffness, Hylamer may demonstrate higher rates of wear in a more abrasive environment, such as in the presence of hard third bodies. Furthermore, all of the Hylamer liners in the study by Livingston et al. were sterilized with gamma irradiation in air. Shelf-life is another clinical variable. For given oxidation levels, the values for ultimate tensile strength, elongation to breaking, and toughness of Hylamer have been shown to be lower than those for conventional polyethylene8. Such time-dependent effects of oxidation may result in higher in vivo rates of wear of these components. This time-dependent effect would not have been detected by the wear-simulator studies that have been performed on unaged Hylamer. Over the next decade, we will see a number of studies that compare the clinical performance of one bearing with that of another. The issues that we outlined involve general principles and illustrate the complexities in the clinical evaluation of wear.
Thomas P. Schmalzried, M.D. Frederick J. Dorey, Ph.D. Harry McKellop, Ph.D. Joint Replacement Institute at Orthopaedic Hospital 2400 South Flower Street Los Angeles, California 90007
Footnotes
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this commentary. In addition, benefits have been or will be directed to a research fund, foundation, educational institution, or other non-profit organization with which one or more of the authors is associated. Funds were received in total or partial support of the research or clinical study presented in this commentary. The funding source was DePuy-DuPont Orthopaedics.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |