Complications of polyethylene wear debris-induced osteolysis led to the development of ceramic-on-ceramic total hip replacements and other alternative bearing surfaces such as highly cross-linked polyethylene and metal-on-metal bearings1-4. However, concerns have been raised about the increased stiffness associated with all-ceramic and metal-backed ceramic acetabular components, which potentially leads to increased rates of migration and loosening2,3. The design goal of so-called “sandwich” cup designs, with polyethylene interposed between the ceramic bearing surface and the outer metal shell, is to reduce the rigidity of the ceramic-on-ceramic coupling and prevent impingement between the rim of the ceramic liner and the neck of the femoral stem5. One such design (Hedrocel ceramic bearing cup; Implex, Allendale, New Jersey) consists of a porous tantalum shell, compression-molded polyethylene, and a ceramic bearing insert (Fig. 1).
In 2003, patient enrollment into a prospective United States Food and Drug Administration (FDA)-approved clinical study of the Hedrocel ceramic bearing cup was halted following reports of three failed ceramic acetabular liners. As of the time of writing, eleven additional failed ceramic liners have been reported, for a total of fourteen. The resulting investigation, failure analysis, and review of the literature led us to hypothesize that a combination of high patient body weight and deep flexion activities caused high frictional interaction and torque, which in turn led to dislodgment and subsequent fracture of the ceramic acetabular liner.
Materials and Methods
Beginning in October 1999, a prospective, randomized, controlled FDA-regulated investigational device exemption clinical trial was initiated in the United States to assess the clinical performance of the Hedrocel ceramic bearing cup. The study protocol was approved by the investigational review boards at twenty-two institutions, and all subjects gave their informed consent to participate in the study. The nonmodular cup comprised a porous tantalum shell, direct-compression-molded ultra-high molecular weight polyethylene, and an alumina ceramic bearing insert (Biolox Forte; CeramTec, Stuttgart, Germany). The ceramic bearing was mechanically secured to the polyethylene by an interference fit (Fig. 1). The study was designed to compare this cup with a nonmodular design consisting of the same porous tantalum shell and an articulating surface of direct-compression-molded polyethylene. All cups had a 28-mm internal diameter, and the range of outer diameters was 48 to 64 mm in 2-mm increments. All subjects received a 28-mm alumina ceramic head, which had 0, +3, and +6-mm neck-length options (Biolox Forte). The modular femoral stems used in the study included press-fit hydroxyapatite-coated titanium-alloy and porous-coated cobalt-chromium-alloy implants (both ProxiLock designs; Implex) and cemented cobalt-chromium-alloy implants (Cobrex; Implex).
The ratio of the random group assignment was two ceramic-on-ceramic arthroplasties for every one ceramic-on-polyethylene arthroplasty. There were fewer control patients because we had thought that another population of control subjects would be available from an identically structured and randomized FDA-regulated study of metal-on-metal implants that had used the same control implants and our intention had been to pool the controls. However, with the termination of this study, the two studies were divorced so we did not have access to the second group of controls for the present analysis. No attempt was made to blind the thirteen principal investigator surgeons or the eleven coinvestigator surgeons with regard to the bearing surface assignment.
Clinical outcome measures included radiographic analysis, Harris hip scores6, and assessments with the Short Form-12 (SF-12) Health Survey7. Radiographic analysis was performed by a radiologist who had not been involved in the patient's care and included anteroposterior pelvic, anteroposterior hip, and frog-leg lateral views. Harris hip scores and SF-12 data were collected preoperatively and at twelve and twenty-four months postoperatively. Hip range-of-motion values were derived from the motion subscale of the Harris hip score. Radiographs were made at the preoperative visit; at six weeks and three, six, twelve, and twenty-four months following the surgery; and annually thereafter. The primary criteria for including a patient in the study were an age between eighteen and seventy-five years, a body-mass index of <40, a preoperative Harris hip score of <70 points, and clinical indications for total hip replacement. Subjects with bilateral hip disease were randomized once and received the same cup in both hips.
Biomechanical Data and Failure Analysis
Failed acetabular components retrieved at the time of revision were analyzed both macroscopically and microscopically. Ceramic liners that were retrieved intact (not fractured) and the corresponding femoral heads were analyzed with profilometry and a coordinate measuring system (Mitutoyo FJ805; Mitutoyo America, Aurora, Illinois), and the wear areas were documented with scanning electron microscopy.
The mechanical integrity of the interference fit between the ceramic liner and the polyethylene was measured in unimplanted cups with use of a lever-out test, with torque calculated from the measured force and apparatus dimensions. The lever-out test was performed by tangentially loading the ceramic liner within a water bath, with testing performed at both room and body temperatures (20° and 37°C). The load was increased, and maximum load was recorded when the ceramic liner escaped capture. The maximum torque at failure was calculated from the load applied and the distance from the center of rotation of the ceramic liner (14 mm).
In an attempt to experimentally replicate the in vivo failures, unimplanted cups were tested with use of hip-wear simulators under a variety of adverse conditions, including creation of neck impingement, introduction of third-body debris, and testing of the effects of deliberately roughened heads, partially dislodged ceramic liners, and microseparation conditions. The hip simulators included an MTS 3 Station 4-DOF (four-degrees-of-freedom) Hip Wear Simulator (MTS Systems, Eden Prairie, Minnesota) with the components in the anatomical position at room temperature and an AMTI-Boston 12 Station Hip Simulator (AMTI, Watertown, Massachusetts) with the components in the inverted position at 37°C. All testing was performed with a modified Paul curve in bovine serum.
Finite element analysis of the ceramic bearing acetabular cup was performed for extreme loading conditions (a near maximum range of motion of 60° with a 5-kN [1124-lb] load) by EndoLab Mechanical Engineering (Rosenheim, Germany). The purpose of the finite element analysis was to determine the location and magnitude of the maximum stress within the materials that make up the prosthesis (porous tantalum, polyethylene, and the alumina ceramic liner). The femoral neck had a 120° maximum arc of motion before it impinged on the acetabular component. Thus, the end point of motion used in the finite element model was 60° as measured between the axis of the femoral neck and the polar axis of the acetabular component. As a result of the nonaxisymmetric load condition, a three-dimensional model was constructed with use of the program MSC.MARC 2001 (MSC. Software, Ann Arbor, Michigan). Cups with 48 and 54-mm outer diameters were modeled. The finite element analysis model assumed that the acetabular cup was loaded with a 28-mm alumina ceramic femoral head.
Summary clinical data and group comparisons were made for the ceramic-on-ceramic and ceramic-on-polyethylene arms of the study as well as within the ceramic-on-ceramic group to compare data between the failed and non-failed cases. For the latter data analysis, patient weight, body-mass index, gender, age, size of the acetabular cup, abduction of the acetabular component, type of femoral stem, range of motion at one year, and Harris hip score at one year were compared by using Mann-Whitney rank sum or Fisher exact testing followed by logistic regression in univariate analysis, with failure as the outcome of interest. Variables with p values of ≤0.25 in the univariate analysis were entered into a multivariate logistic model. Alpha was set at 0.05 for all assessments. Statistical calculations were performed with MINITAB 14.11 (Minitab, State College, Pennsylvania).
Three hundred and fifteen ceramic-on-ceramic total hip prostheses were implanted in 282 patients, and 157 ceramic-on-polyethylene total hip prostheses were implanted in 147 patients. Demographics were comparable between the ceramic-on-ceramic and ceramic-on-polyethylene groups. In both groups, 77% of the hips had a diagnosis of osteoarthritis, 17% had aseptic necrosis, 3% had rheumatoid arthritis, 2% had a fracture, and 0.3% had another diagnosis. Fifty-five percent of the patients were male, and 45% were female. The mean age (and standard deviation) was 54 ± 12 years (range, twenty-three to seventy-six years), and the mean body-mass index was 29 ± 4.6 kg/m2 (range, 18.4 to 39.9 kg/m2). No differences in the above factors were seen between the ceramic-on-ceramic and ceramic-on-polyethylene groups.
At the time of analysis, 98.8% of patients had been followed for two years or more. The mean preoperative and one-year postoperative Harris hip scores were 45 points (range, 14 to 91 points) and 92 points (range, 36 to 100 points), respectively, in the ceramic-on-ceramic group and 43 points (range, 10 to 78 points) and 93 points (range, 51 to 100 points) in the control group; these scores did not differ significantly between the groups. The SF-12 Mental and Physical Component Summary scores were comparable between the ceramic-on-ceramic and ceramic-on-polyethylene groups at all time-points, and the radiographic results were also comparable. No progressive radiolucency, cup migration, or evidence of osteolysis was seen in either patient group at any follow-up time. Excluding hip revision necessitated by failure of the ceramic liner, only two ceramic acetabular components were revised: one because of recurrent dislocation and the other because of groin pain of uncertain etiology. In the control group, one cup was revised at nineteen months because of recurrent dislocation, one was revised because of persistent groin pain, and one was revised during surgery for a periprosthetic femoral fracture.
As of April 15, 2005, fourteen of the 315 implanted ceramic bearings were reported to have failed. The time to failure averaged twenty-five months and ranged from eight to forty-two months. The median weight and body-mass index of the patients with a failed ceramic-on-ceramic prosthesis were 102.5 kg and 30.7 kg/m2, respectively (Table I), with all but three failures occurring in patients weighing >91 kg. In contrast, the median weight and body-mass index of the patients with no report of failure were 83.4 kg (p = 0.004) and 27.9 kg/m2 (p = 0.09), respectively. Twelve of the fourteen failures occurred in men. The results of univariate logistic regression analysis of subject characteristics for the fourteen failed and 301 non-failed ceramic-on-ceramic implants are shown in Table II. No significant association was found between failure and age, range of motion, Harris hip score at one year, acetabular cup size, stem size, stem type, or cup abduction angle. Male gender (p = 0.03), a weight of >91 kg (p = 0.006), and an increased body-mass index (p = 0.045) increased the odds of the ceramic liner failing.
On the basis of the univariate regression analysis, gender, a weight of >91 kg, and acetabular cup size were assessed in a multivariate logistic regression model. Body-mass index was excluded from the model as it is calculated from patient weight and height and is therefore not an independent variable. The results of the multivariate logistic regression analysis are shown in Table III. After we controlled for gender and acetabular cup size, we found that patients with a body weight of >91 kg had a 4.76 times greater odds (95% confidence interval, 1.14 to 19.92) of having a ceramic liner failure than those who weighed ≤91 kg (p = 0.03). The difference based on gender seen in the univariate analysis was explained when weight data were stratified by gender for the entire ceramic-on-ceramic group. The mean weight for men was 94 ± 15 kg, whereas it was 74 ± 15 kg for women (p < 0.0005). The increased odds of failure seen in men were due to the confounding of the increased mean weight of men compared with that of women in the study group.
Anecdotal clinical information reported by the surgeons who performed the revision procedures in five of the fourteen hips indicated that these patients had experienced a noise or sensation during a deep-hip-flexion activity at the approximate time of ceramic failure. One patient involuntarily experienced deep flexion during a work-related accident. Another patient had frequent unusual sensations about the replaced hip when he tied his shoe. One patient with a failed bearing stated that she gardened often and typically in a squatting position. The fourth patient reported the hip “going” while in a deep-squat position, and the fifth such failure was in a body builder who often engaged in deep-hip-flexion activities. The “noise” or “sensation” was described as a “squeak” emanating from the hip. This finding was not analyzed statistically.
At the time of revision, twelve of the fourteen acetabular bearing surfaces were found to have fractured. In the remaining two hips, the acetabular bearing surface had disengaged from the polyethylene portion of the acetabular component but remained intact. In both of these unfractured implants, the ceramic liner was apparently entrapped by the femoral neck, as illustrated in Figures 2 and 3. Visual analysis of the acetabular polyethylene at the time of revision indicated that little or no impingement had occurred between the femoral neck and the edge of the polyethylene layer, the first point of contact in impingement, prior to displacement of the ceramic liner. Metal transfer from the neck of the femoral stem was observed on the rims of the two ceramic liners that were retrieved intact (Fig. 3) and on the rims of all of the fractured ceramic liners (Fig. 4). The ceramic heads corresponding to the two liners that were retrieved intact showed evidence of stripe wear. This could not be definitively identified in the remaining heads because they were damaged by the fractured acetabular liners. Scanning electron microscopy identified abrasion of the ceramic surface with grain pull-out within the striped wear area of the ceramic heads (Fig. 5). Metrological analysis of the two retrieved intact liners and their corresponding heads indicated that the components had deformation of the surfaces outside of the original manufacturing tolerances due to in vivo wear.
The strength of the ceramic liner-and-polyethylene assembly of five unimplanted cups of each size from the manufacturer's inventory (48 to 64 mm) was measured with lever-out testing at both body and room temperatures. The strength of the assembly (defined here as the resistance to torsional dislodgment of the ceramic liner) at body temperature averaged 33.4 ± 3.8 Nm (295 ± 34 in-lb), with a range of 24.9 to 41.2 Nm. The lever-out torque at body temperature was approximately 40% less than that at room temperature, which is consistent with the decrease in elastic modulus and yield strength of polyethylene for this change in temperature. The results of the lever-out testing also showed no apparent loss of integrity of the capture mechanism caused by shelf aging within the package. (The shelf ages of the tested cups ranged from two to four years.)
In the hip wear/failure simulation testing, femoral neck impingement did not result in displacement of the ceramic liner. With current hip-simulator technology, we could not reproduce microseparation of the articulation in combination with load and motion of the hip joint. Complete displacement of the ceramic liner occurred only after partial displacement (experimentally caused prior to testing). Finally, fracture of the ceramic liner occurred subsequent to complete displacement of the ceramic liner from the remainder of the acetabular component. Hip wear simulation testing, with use of standard and modified gait loads as well as abrasives added to the lubricant for acetabular cups that had not been altered prior to testing, could not replicate the failure mechanism observed in vivo.
Finite element calculation of maximum principal tensile stress within the ceramic liner indicated a value of 100 MPa (14,500 psi) for a 5-kN (1124-load) load at a nearly maximum range of motion (60°). In contrast, the manufacturer of the alumina ceramic liner (Biolox Forte) has reported the four-point strength to be 580 MPa (84,000 psi), far more than the maximal load predicted by our finite element analysis8.
With the exception of the failed ceramic liners, the excellent short-term clinical results associated with the ceramic-on-ceramic and ceramic-on-polyethylene articulations in our study are consistent with the results in other clinical studies of ceramic articulations in primary total hip replacement9-11. The demographic and clinical similarity between the subjects with a ceramic-on-ceramic articulation and those with a ceramic-on-polyethylene articulation in our study suggests that selection bias did not play a role in the observed failure rate. Analysis of the clinical data indicated that the odds of ceramic liner failure were 4.76 times greater in patients who weighed >91 kg than in those who weighed ≤91 kg after we controlled for gender and the size of the acetabular cup. The Harris hip scores indicated that the patients with and without failure of the liner were quite active with an essentially unrestricted range of hip motion. Anecdotal information about the activities of five of the patients with a failed bearing indicated a relationship between symptoms and deep hip flexion.
Retrieval analysis of the two intact liner-and-head pairs demonstrated striped head wear and rim wear. The wear damage observed macroscopically and microscopically showed evidence of abrasion and grain pull-out, indicating both abrasive and adhesive wear mechanisms12. This observation is in agreement with the findings of other reports. Alumina-alumina ceramics have been reported to wear in vivo as a result of rim/edge loading at the extremes of motion and/or by means of microseparation4,12-14. Walter et al. reported a 52% prevalence of rim and/or striped head wear of ceramic-ceramic hip bearings12. Retrieved ceramic bearings in other studies have demonstrated stripe and rim wear, and scanning electron microscopy has shown that wear was caused by grain pull-out (adhesive wear) and abrasion (by the pulled-out grains)12,15. By carefully documenting the locations of the stripe and rim wear relative to the osseous anatomy at the time of retrieval, Walter et al. concluded that this mechanism was due to head-rim engagement with subluxation in a deep hipflexion position.
The relevance of this finding is that adhesive wear occurs when two opposing solid surfaces come into physical contact (e.g., when there is complete breakdown of fluid boundary lubrication), thereby leading to high friction/traction forces15. This causes high, localized friction at the contact interface. Prosthetic hip bearings are known to demonstrate mixed-film boundary lubrication under biological conditions (i.e., an absence of lubrication during stop-start, direction reversals, and low-relative-velocity conditions)16-18. Such mixed-film lubrication conditions increase the probability of adhesive wear, which becomes more likely for bearing surfaces damaged by wear because of disruption of the fluid boundary layer. The coefficient of friction for nonlubricated alumina-alumina contact has been reported to range from 0.5 to 1.0, whereas the value for lubricated contact ranges from 0.002 to 0.05, depending on the lubricant19-21.
Our laboratory simulations conclusively showed that neck impingement alone cannot cause dislodgment of this ceramic liner. In addition, hip simulation of normal gait cycles under a variety of adverse conditions did not generate cyclic or singular torsional forces sufficient to dislodge the ceramic liner. However, with the liner partially displaced, normal gait cycles resulted in complete dislodgment and eventually fracture.
We suggest that high torsion across the articulation interface that leads to ceramic dislodgment must originate with, and depends on, one or more of the following: friction between the head and liner, subluxation and edge loading of the head and liner, and the patient's weight. The relevance of patient weight to the failure mechanism is that frictional torque at the articulation interface is directly proportional to patient weight; the higher the patient's weight, the larger the frictional force and associated torque transmitted about the center of rotation of the hip. Given the force multiplier of five times body weight in deep flexion, most of the patients in whom the ceramic liner failed could exceed a load of 4460 N at the hip articulation8. Coupled with a high coefficient of friction of 0.5 to 1.0, there is a resultant 2230 to 4460-N force on the ceramic liner acting tangentially 14 mm from the center of the 28-mm-diameter head. This results in a torque of 31.2 to 62.4 Nm, which exceeds the strength of the assembly of the ceramic liner within the polyethylene portion of the ceramic cup (33.4 ± 3.8 Nm at body temperature).
Finite element analysis of the cup at the near end point of the range of motion and with high load (5 kN) indicated a maximum stress within the ceramic (100 MPa) that was substantially less than that required to cause fracture of the ceramic (flexural strength = 580 MPa)8. Analysis of the retrieved fractured and non-fractured inserts indicated that overload or fatigue fracture of the ceramic liner alone did not cause failure but rather that displacement of the liner from the polyethylene socket was a prerequisite to failure.
In their case report, Akagi et al. noted head and liner wear of the ABS nonmodular ceramic cup1. They speculated that torque generated by damage to the articulation caused by microseparation during normal gait led to the gradual propagation of a defect in the polyethylene layer, causing a loss of support of the ceramic liner and eventual failure. Hasegawa et al. reported two ceramic liner fractures and one liner dissociation in a study of thirty-five nonmodular ceramic-polyethylenemetal acetabular components22. They suggested that the fractures were likely due to edge loading with the liner in situ and the dissociation was due to the mechanism described by Akagi et al. In contrast, we agree with Walter et al.12 and believe that the displacement of the ceramic liner occurs during subluxation and reengagement of the head and liner during deep flexion.
In summary, we believe that the failures of the ceramic liner of the nonmodular, so-called sandwich-design ceramic-on-ceramic cup were caused by high torque transmitted from the femoral head to the ceramic liner, causing dislodgment of the ceramic liner from the polyethylene socket. We postulated that the origin of high torque is frictional interference between a wear-damaged ceramic head and the rim of the ceramic liner. Striped wear damage of the head occurs as a result of edge loading and rim wear, due to subluxation and microseparation. The high frictional torque transmitted across the articulation interface probably occurs during deep flexion or other high load/extreme range-of-motion activities. Subsequent to dislodgment of the ceramic liner and with continued articulation of the hip joint, the ceramic liner completely displaces and eventually fractures in most cases.
A potential solution to the ceramic liner displacement is the addition of a geometric irregularity such as a central peg or a change in the geometry of the outer surface of the ceramic liner from a hemisphere to an angled surface between the ceramic-polyethylene interface. This would increase the torque required to displace the liner and reduce this mode of failure.
This investigation revealed a previously unknown phenomenon that is inherent in ceramic-ceramic hip articulations, which is high friction and torque associated with the tribological process that causes striped head wear and rim liner wear. ▪
Disclosure: The authors did not receive any outside funding or grants in support of their research for or preparation of this work. One or more of the authors, or a member of his or her immediate family, received, in any one year, payments or other benefits in excess of $10,000 or a commitment or agreement to provide such benefits from a commercial entity (Implex Corporation and Zimmer Incorporated [employee]). No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, division, center, clinical practice, or other charitable or nonprofit organization with which the authors, or a member of their immediate families, are affiliated or associated.
Investigation performed at Zimmer Trabecular Metal Technology, Inc., Parsippany, New Jersey, and Malcolm and Dorothy Coutts Institute for Joint Reconstruction and Research, San Diego, California
1. Akagi M, Nonaka T, Nishisaka F, Mori S, Fukuda K, Hamanishi C. Late dissociation of an alumina-on-alumina bearing modular acetabular component. J Arthroplasty. 2004;19: 647-51.
2. Bohler M, Schachinger W, Wolfl G, Krismer M, Mayr G, Salzer M. Comparison of migration in modular sockets with ceramic and polyethylene inlays. Orthopedics. 2000;23: 1261-6.
3. Boutin P, Christel P, Dorlot JM, Meunier A, de Roquancourt A, Blanquaert D, Herman S, Sedel L, Witvoet J. The use of dense alumina-alumina ceramic combination in total hip replacement. J Biomed Mater Res. 1988;22: 1203-32.
4. Butterfield M, Stewart T, Williams S, Ingham E, Stone M, Fisher J. Wear of metal-metal and ceramic-ceramic hip prostheses with swing phase microseparation. Read at the Annual Meeting of the Orthopaedic Research Society; 2004 ; Dallas, TX.
5. Hasegawa M, Sudo A, Hirata H, Uchida A. Ceramic acetabular liner fracture in total hip arthroplasty with a ceramic sandwich cup. J Arthroplasty. 2003;18: 658-61.
6. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end-result study using a new method of result evaluation. J Bone Joint Surg Am. 1969;51: 737-55.
7. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34: 220-33.
8. Heros RJ, Willman G. Ceramics in total hip arthroplasty: history, mechanical properties, clinical results, and current manufacturing state of the art. Semin Arthroplasty. 1998;9: 114-22.
9. D'Antonio JA, Capello WN, Manley MT, Naughton M, Sutton K. A titaniumencased alumina ceramic bearing for total hip arthroplasty: 3- to 5-year results. Clin Orthop Relat Res. 2005;441: 151-8.
10. Kim YH. Comparison of polyethylene wear associated with cobalt-chromium and zirconia heads after total hip replacement. A prospective, randomized study. J Bone Joint Surg Am. 2005;87: 1769-76.
11. Yoo JJ, Kim YM, Yoon KS, Koo KH, Song WS, Kim HJ. Alumina-on-alumina total hip arthroplasty. A five-year minimum follow-up study. J Bone Joint Surg Am. 2005;87: 530-5.
12. Walter WL, Insley GM, Walter WK, Tuke MA. Edge loading in third generation alumina ceramic-on-ceramic bearings: stripe wear. J Arthroplasty. 2004;19: 402-13.
13. Nevelos J, Ingham E, Doyle C, Streicher R, Nevelos A, Walter W, Fisher J. Microseparation of the centers of alumina-alumina artificial hip joints during simulator testing produces clinically relevant wear rates and patterns. J Arthroplasty. 2000;15: 793-5.
14. Tipper JL, Hatton A, Nevelos JE, Ingham E, Doyle C, Streicher R, Nevelos AB, Fisher J. Alumina-alumina artificial hip joints. Part II: characterisation of the wear debris from in vitro hip joint simulations. Biomaterials. 2002;23: 3441-8.
15. Nevelos JE, Ingham E, Doyle C, Fisher J, Nevelos AB. Analysis of retrieved alumina ceramic components from Mittelmeier total hip prostheses. Biomaterials. 1999;20: 1833-40.
16. Rieker C, Konrad R, Schon R. In vitro comparison of the two hard-hard articulations for total hip replacements. Proc Inst Mech Eng [H]. 2001;215: 153-60.
17. Scholes SC, Unsworth A. Comparison of friction and lubrication of different hip prostheses. Proc Inst Mech Eng [H]. 2000;214: 49-57.
18. Scholes SC, Unsworth A, Goldsmith AA. A frictional study of total hip joint replacements. Phys Med Biol. 2000;45: 3721-35.
19. Saikko V, Pfaff HG. Low wear and friction in alumina/alumina total hip joints: a hip simulator study. Acta Orthop Scand. 1998;69: 443-8.
20. Santavirta SS, Lappalainen R, Pekko P, Anttila A, Konttinen YT. The counterface, surface smoothness, tolerances, and coatings in total joint prostheses. Clin Orthop Relat Res. 1999;369: 92-102.
21. Schuh C, Kinast EJ, Mezzomo E, Kapczinski MP. Effect of glazed and polished surface finishes on the friction coefficient of two low-fusing ceramics. J Prosthet Dent. 2005;93: 245-52.
Copyright 2007 by The Journal of Bone and Joint Surgery, Incorporated
22. Hasegawa M, Sudo A, Uchida A. Alumina ceramic-on-ceramic total hip replacement with a layered acetabular component. J Bone Joint Surg Br. 2006; 88: 877-82.