Revision of failed polyethylene components in total knee arthroplasty (TKA) is generally considered unsuccessful.1,2,7 Full revision has been recommended even when the implants are aligned correctly and fixed securely to bone.1,2,7 However, reports of high failure rates (17- 100%)2,5 after isolated polyethylene revision do not address four major causes of TKA failure: (1) poor polyethylene quality, (2) incompetent mechanism of locking the polyethylene component to the metal tray,7 (3) uncorrected ligament imbalance, and (4) failure of the patellar component. Since these design and engineering problems have been solved in most modern total knee systems, revision of modular polyethylene should be considered a reasonable alternative to full revision of the tibial component. Because isolated revision of the polyethylene insert decreases surgical trauma and preserves existing bone stock, it is important to question the recommendation to perform a full revision.
Because some tibial components have inadequate polyethylene insert locking mechanisms8 or are made as mono- block components without provision for polyethylene insert exchange, we devised a method of fabricating a locking mechanism using polymethylmethacrylate (PMMA) to secure the polyethylene insert to the metal tibial component. We then designed a mechanical study to evaluate the quality of fixation of this fabricated locking mechanism to measure its resistance to shear loading.
We hypothesized isolated exchange of the tibial polyethylene insert would provide long-term revision-free survival of the implant, and the fabricated locking mechanism would have shear load-bearing capacity equal to or better than currently available polyethylene locking mechanisms.
MATERIALS AND METHODS
We performed isolated revision of the polyethylene tibial component on 53 consecutive patients (56 knees) with well-aligned and well-fixed implants to correct problems related to polyethylene wear, ligament imbalance, failure of the polyethylene component locking mechanism, and osteolysis after TKA. These 57 revisions represent about 7% of all TKA revisions performed during the study period. Seven patients (7 knees) were lost to followup, leaving 46 patients (49 knees) for study. The mean patient age at revision was 74 years (range, 53-86 years), and the mean weight was 86 kg (range, 62-105 kg). Twenty patients (20 knees) were men and 26 patients (29 knees) were women. The minimum followup was 28 months (mean 48 months, range 28-81 months) for the knees revised using the original polyethylene locking mechanism, and a minimum of 23 months (mean 32 months, range 23-41 months) for knees revised using the cement fabrication method.
Twenty-nine knees (26 patients) had Ortholoc I® components (Wright Medical, Arlington, TN). The Ortholoc I® (Wright Medical) components were cobalt-chromium with a sintered- bead porous surface. All Ortholoc I® (Wright Medical) knees failed because of wear of the metal-backed patellar component and tibial polyethylene (Fig 1). The polyethylene tibial component was press-fit to the metal tray by a ridge on the metal tray and matching grooves in the polyethylene liner. Replacement polyethylene components were not available at the time of surgery, so we developed a technique to attach a new polyethylene insert to the metal tray using polymethylmethacrylate (PMMA). A carbide bit (Midas Rex MC-30, Medtronic, Fort Worth, TX) was used to cut a criss-cross network of grooves into the surfaces of the metal tray and polyethylene component (Figs 2, 3). Both components were coated with freshly mixed PMMA cement (Simplex®, Stryker, Kalamazoo, MI) (Fig 4). The components were pressed together firmly to make the cement layer uniform and as thin as possible (estimated thickness 1 mm). The components were kept compressed until the cement hardened. The Profix® (Smith & Nephew, Memphis, TN) tibial polyethylene insert matched the Ortholoc I® (Wright Medical) femoral surface so the medial and lateral femoral condyles rested in the central ⅓ of their respective condylar surfaces of the polyethylene component. This polyethylene component was used to substitute for the Ortholoc I® (Wright Medical) component, which was unavailable from the manufacturer.
Eight patients (8 knees) had Ortholoc II® components (Wright Medical). All knees with Ortholoc II® components failed because of wear in the gamma-irradiated polyethylene components. Two of these knees were revised with gamma- irradiated polyethylene inserts, and the rest were revised with gas-sterilized polyethylene components. Five patients (5 knees) had Advantim® (Wright Medical) components. All of these patients were revised with a new gas-sterilized polyethylene component. Five patients (5 knees) had AMK® components (DePuy, Warsaw, IN), and two patients (2 knees) had Miller-Gallante II® components (Zimmer, Warsaw, IN). Patients with AMK® (DePuy) and Miller-Gallante II® components (Zimmer) failed because of polyethylene wear and osteolysis. New gas-sterilized polyethylene components were used to replace the original polyethylene. When the new polyethylene components were inserted in these knees, visibly detectable motion occurred when the polyethylene components were pushed gently with an instrument. Because the locking mechanism could not be secured with the standard mechanism, the polyethylene components were secured to the metal tray with the cementing technique described for the Ortholoc I® (Wright Medical) implants. Two patients (2 knees) with major ligament imbalance not corrected at the time of the original operation had appropriate ligament release and adjustment of polyethylene component thickness to achieve ligament balance.
Patients were followed at yearly intervals. One author (LW) evaluated visible narrowing relative to the thickness seen on radiographs at the first postoperative visit and eccentric narrowing of the polyethylene insert on the weightbearing AP radio- graphs. Angular alignment of the femoral and tibial components was measured from plane radiographs of the knee taken on 14″ × 17″ film. We estimated the centers of the diaphyseal shaft and metaphysis adjacent to the implants, and drew a line connecting the two points to represent the long axis of the bone. We measured femoral component valgus angle by constructing a line perpendicular to a line through the distal-most points on the femoral component on the AP radiograph, and measuring the angle between this line and the long axis of the femur using a goniometer. The femoral component valgus angle was 5° ± 1.8° (mean ± standard deviation [SD]). The femoral component flexion angle was measured by constructing a line perpendicular to the distal inner surface of the femoral component using a goniometer, and then measuring the angle between this line and the long axis of the femur. The femoral component flexion angle was 1.2° ± 1.1° (mean ± SD). The tibial component angle was measured by constructing a line perpendicular to the undersur- face of the tibial component on the AP radiograph with a goniometer, and then measuring the angle between this line and the long axis of the tibia. Tibial component valgus angle was 0.5° varus ± 1.1° (mean ± SD). We measured the tibial component posterior slope by constructing a line perpendicular to the undersurface of the tibial component on the lateral radiograph with a goniometer, and then measuring the angle between this line and the long axis of the tibia. Tibial component posterior slope was 2.3° ± 2.2° (mean ± SD). The knees were evaluated at 1 year after surgery, then at 3-year intervals. Success was defined as intact implants without failure due to loosening or wear.
We conducted a bench test to evaluate polyethylene component fixation with a push-off test using the locking mechanism for the Profix® (Smith & Nephew) TKA as the control. The locking mechanism employed a peripheral rim extending 3 mm above the surface and had a 2 mm overhang that press-fit into a matching groove on the polyethylene component. This rim- locking device extended around the posterior ½ of the tibial component. The polyethylene component was restrained from sliding forward by an anterior rim that extended 2 mm above the surface of the metal tray.
The polyethylene components and metal trays were assembled and mounted in a machining vice. The vice was mounted in a servohydraulic tensiometer (Instron, Model 8501, Norwood, MA) with the surface of the tray parallel to the line of travel of the loading arm. The loading arm and tibial components were aligned so the applied load was centered medial to lateral on the part, and so the loading arm touched the polyethylene just past the metal tray while equally loading both posterior condyles (Fig 5). A goniometer was used to ensure the loading arm was perpendicular to the transverse axis of the tibial component. The polyethylene component was loaded from back-to-front at a rate of 5 mm/minute until failure. The force-displacement curve was recorded using the displacement and load outputs from the analog outputs serial port on the Instron (Norwood, MA) computer tower. These outputs were fed to a data acquisition device (DATAQ Instruments, Akron, OH) that used WinDaq software (DATAQ) for viewing and recording. Load-displacement curves were created using the force and displacement curves. Energy- to-failure was calculated by adding the area under the energy curve starting at a consistent starting load and continuing to the maximum load reached. Failure was defined as the initial sharp drop-off in the load-deflection curve, and always corresponded to the release of the posterior portion of the locking mechanism.
The first series of tests was performed on Profix® (Smith & Nephew) metal tibial trays and polyethylene components inserted with the standard press-fit locking mechanism. The push- off test was performed on the polyethylene. The locking lip was removed from the polyethylene component and grooves were cut in the undersurface of the component with a hand-held power burr and surgical carbide bit (Midas Rex MC30, Medtronic, Fort Worth, TX). In the samples bonded to the flat plate and Ortholoc I® (Wright Medical) tray, the upper surface of the tray also was grooved with the hand-held power burr. Simplex® cement was applied just after mixing in a low viscosity state to the mating surfaces of the metal tray and polyethylene component. The two were pressed together and held until the cement hardened. The assembled components were mounted in the tensiometer and the push-off test was repeated to complete the second series of tests.
The third series of tests was performed on flat stainless steel trays shaped to match the profile of the Profix® (Smith & Nephew) polyethylene insert. This series of tests was carried out to test the fixation of the polyethylene component in situations in which minimal locking mechanism is present, such as in revision of mobile bearing tibial polyethylene components. The trays were scored with the carbide bit on a hand-held power burr to produce a cross-hatch pattern with one groove approximately every 0.5 cm from front-to-back and side-to-side. Fresh polyethylene components were prepared, cemented, and tested as described previously.
The fourth series of tests was performed on the Ortholoc I® (Wright Medical) tibial tray. The cobalt-chromium component was flat except for a curved 1.5 mm high locking rim extending around the periphery of the upper surface. These components were prepared in a manner similar to the stainless steel trays, but additional grooves were made in the undersurface of the polyethylene insert to accommodate the locking ridge on the metal tray. The polyethylene components were cemented to the metal trays, and the push-off tests were repeated.
A two-tailed Student's t test was used to evaluate differences between the test and control groups. Confidence interval of 95% (p < 0.05) was chosen to indicate statistical significance. Power analysis was performed to test the probability of type 2 error in accepting no difference between groups with no statistically significant difference. Statistical powers greater than 0.8 are deemed adequate to accept the null hypothesis.
We observed three clinical failures in the 46 patients (49 knees). A gamma-irradiated, shelf-aged polyethylene component was inserted after failure of the polyethylene in an Ortholoc II® (Wright Medical) tibial base plate inserted 6 years earlier. However, this polyethylene component failed 23 months later because of wear. It was revised to an ethylene-oxide sterilized component, and has functioned well with no sign of wear 77 months postoperatively. The grafted osteolytic cysts healed, and the metal implant remains well-fixed to bone. One Ortholoc II® (Wright Medical) implant had its polyethylene component revised with an unconstrained polyethylene insert. Three weeks postoperatively the posterior cruciate ligament (PCL) was ruptured in a fall, and the tibia dislocated posteriorly. During attempted reduction the distal femur fractured, and revision of the knee was performed with a new femoral component and highly conforming, deep-dish polyethylene component. This patient returned to normal walking and died 38 months later of unrelated causes. One patient with an Ortholoc II® knee, which was revised 18 years after primary TKA, had massive tibial osteolysis. The cysts were curetted and grafted, and the polyethylene was exchanged. Compressive fracture of the supporting tibial bone occurred during the first 2 months after surgery, but healed in 5° varus deformity. The patient's knee was asymptomatic 25 months after revision.
Aside from the knees revised with age-damaged polyethylene components, none of the knees revised using the original polyethylene locking mechanism failed because of osteolysis, wear, or failure of fixation of the polyethylene component. None of the knees with a locking mechanism fabricated with PMMA cement failed. All knees continue to function well (Fig 6).
No knees had radiographic signs of polyethylene wear at followup, except the one revised using gamma- irradiated polyethylene.
Load-to-failure and energy-to-failure for the standard Profix® (Smith & Nephew) locking mechanism was substantial, but the cement-fabricated locking mechanism in the Profix® (Smith & Nephew) metal tray was better (p < 0.001 and p < 0.02) in both parameters (Table 1). The fabricated PMMA locking mechanism on the flat stainless steel tray had a higher (p < 0.05) load-to-failure, but lower (p < 0.02) energy-to-failure than the standard Profix® (Smith & Nephew) components. The difference was caused by substantially lower displacement to failure. The fabricated PMMA Ortholoc I® (Wright Medical) locking mechanism failed at a load similar to the standard Profix® (Smith & Nephew) components. Statistical power (1-β) was 0.822, sufficient to accept the null hypothesis for load- to-failure as compared with the Profix mechanism. However, its energy-to-failure was lower, but not significantly lower (p < 0.1) using our criteria (Fig 7). Statistical power (1-β) was 0.622, insufficient to accept the null hypothesis for energy-to-failure as compared with the Profix mechanism.
We sought to ascertain whether isolated polyethylene component revision can be successful in modular total knee tibial components and whether cementing a new polyethylene component to a tibial metal component would be effective. Our results suggest isolated exchange of the tibial polyethylene component in revision TKA can be successful in midterm followup for knees with well- fixed and correctly aligned implants. Both the clinical results and mechanical tests indicate cementing a new polyethylene component to the tibial metal tray can be a safe and effective means of revising some failed total knees. Three failures occurred in the 49 knees. One failed from wear of gamma-irradiated and shelf-aged polyethylene, one failed because of fracture of the femur during attempted reduction of a dislocated tibial component, and one of the knees had failure of the supporting tibial bone stock, but eventually healed and achieved a good result.
We note several limitations. Although the results currently seem favorable, the long-term success of isolated polyethylene component revision is not ensured. The fixation of the polyethylene component with PMMA cement appears strong but we did not conduct fatigue testing which might more directly relate to late failure. In cases such as the AMK and MGII total knee implants in which the locking mechanism is not secure even with a new polyethylene component, an alternative to removal of well-fixed metal implants would be attractive.
We addressed the problem of an inadequate polyethylene locking mechanism by using a cementing technique to bond the new polyethylene module to the metal tibial component. This method allowed salvage of osteointegrated implants, thus reducing the exposure and bone work necessary to achieve a functioning arthroplasty. It also sealed the interface between the polyethylene component and metal plate to eliminate particle generation and pressurization of fluid in the screw holes. We note in knees with smooth, round articular surfaces, for which good quality or correct size polyethylene components were not available, another closely matching polyethylene component was substituted. The preoperative plan included templating with the radiographic overlays for the original femoral component and the Profix tibial polyethylene component. This ensured close match of articular surface contours.
Review of the literature suggests revision of the polyethylene component generally is not considered effective.2,5 The failure rate was unacceptably high (100%) when isolated tibial insert exchange was combined with adhesiolysis to treat stiffness after TKA.2 However, failure was universally related to stiffness and pain, and not to recurrent failure of the polyethylene component.2,5 Isolated polyethylene exchange for knee instability had a failure rate of 29%, but failure was primarily related to persistent ligament imbalance.5 When isolated polyethylene insert revision was performed to treat a variety of failure mechanisms,2 rerevision was common because of persistent instability and pain. Five of the 24 knees revised for wear had rerevision for recurrent wear. The high failure rate suggests isolated revision of the polyethylene was ineffective, but the reasons for recurrent failure because of instability, pain, and recurrent wear are not clear. Uncorrected ligament imbalance, poor polyethylene quality, and incompetent polyethylene locking mechanism may explain the failures. The type of implants and other causes of failure were not mentioned by the authors, and valid generalizations regarding isolated polyethylene insert revision cannot be made from their study.
Deterioration of the polyethylene locking mechanism has been reported.8 The high failure rate after revision of the polyethylene components that employed an ineffective locking mechanism7 and also gamma-irradiated shelf-aged polyethylene, suggests problems with designs and materials rather than the basic concept of revision of modular implants. It follows that improvements in polyethylene quality and locking mechanism would make isolated revision of the polyethylene component acceptable in certain circumstances.
Damage to the undersurface of the polyethylene component regularly occurs during normal use.8 However, this damage seems self-limiting, and in cases of competent locking mechanisms, does not lead to catastrophic failure of the polyethylene component undersurface.6
Scratching of the femoral component also is commonly found in revision TKA, leading some authors to report isolated polyethylene component revision is seldom indicated in revision of modular total knee implants.2,3,7 The surgeon must approach femoral component revision with caution because of the magnitude of the operation and the additional destruction of bone stock. Because all femoral components scratch and roughen during use,12 revision for minor scratching seems excessive. Also, the low activity level of most older patients would militate for retention of an existing well-fixed femoral component if a wear- resistant polyethylene module were available.
The method of testing fixation of the tibial polyethylene component differs from the usual push-off test required for implant clearance by the Food and Drug Administration, but was designed to evaluate performance of the components in the mode in which failure occurs clinically, ie in shear.8,11
The testing results in this study suggested isolated revision of a polyethylene component in a well-designed locking mechanism is a reasonable clinical choice. The resistance to shear load exhibited by this type of locking mechanism is substantial, and the peripheral capture and tight initial press-fit inherent in the Ortholoc II® (Wright Medical), Advantim® (Wright Medical), Genesis® (Smith and Nephew), NexGen® (Zimmer) and Profix® (Smith & Nephew) TKA systems have excellent long-term success.11,16 There are no reports of clinical failure caused by loosening of the polyethylene component or excessive backside wear with these implants, so it is likely direct exchange of the polyethylene module would be acceptable practice if the original cause of failure could be corrected.
The dilemma created by well-fixed and well-aligned implants with poor polyethylene locking mechanisms or inadequate constraint in the available polyethylene surface may be addressed by fashioning a locking mechanism with PMMA cement. The data in the current study indicate the shear resistance of this bond is considerably higher than a robust, clinically tested, press-fit peripheral capture system. The fabricated PMMA locking mechanism in a tibial component with robust peripheral rim resulted in load-to- failure and energy-to-failure values considerably better than those for standard press-fit polyethylene components in the same tibial tray. The fabricated PMMA locking mechanism applied to a flat plate gave load-to-failure values comparable to or better than the standard peripheral capture metal and polyethylene combination. However, energy to failure was somewhat compromised as compared with the standard capture mechanism. In this configuration, the fabricated PMMA locking mechanism had a more brittle failure mechanism with lower displacement before failure. Although the polyethylene cemented to a CoCr Ortholoc tray had lower load to failure and energy to failure, the study did not have adequate power to achieve statistical significance. Larger numbers in these groups likely would demonstrate this method is inferior to the robust peripheral locking mechanism of the Profix components. Our findings suggest this method of fixating the polyethylene module to a relatively flat metal tray would be effective if the interface were not subjected to impact loading in shear. Moderate constraint in the articular surface is unlikely to subject the interface between the polyethylene module and metal tray to impact shear loads, but posterior-stabilized implants may generate substantial impact loads in shear at this interface. This may present a higher risk of failure in patients with low-profile metal locking mechanisms.
Isolated revision of the polyethylene component in total hip arthroplasty (THA) has been a successful and effective alternative to revising the metal acetabular shell.4,10,11,13 Incompetent polyethylene locking mechanisms are common in the hip,9,14,15 but cement fixation has become a satisfactory alternative. Testing the cement-to-metal bond in THA showed impressive mechanical strength.11 Revising the polyethylene component in TKA offers similar advantages. In patients with well-aligned, well-fixed implants, revising the polyethylene component can be an effective alternative to a full revision of the metal components. The surgeon should be aware locking mechanisms with central capture, even when reinforced with a metal clip, likely are inadequate. Pushing the polyethylene component firmly with an instrument while looking for motion relative to the tray is a reasonably good test of the locking mechanism. If there is a question regarding the quality of the locking mechanism, the new polyethylene component should be attached with a fabricated acrylic cement locking mechanism. Revising the tibial polyethylene component by using a new component from a different knee system is a solution that should be used only as a last resort, because articular surface contact area may not be adequately large. However, in cases with inadequate or even unavailable polyethylene components, this method can be a simple solution for lightweight, inactive patients.
The authors thank Diane Morton, MS, for editorial assistance during manuscript preparation.
1. Babis GC, Trousdale RT, Morrey BF. The effectiveness of isolated tibial insert exchange in revision total knee arthroplasty. J Bone Joint Surg Am
2. Babis GC, Trousdale RT, Pagnano MW, Morrey BF. Poor outcomes of isolated tibial insert exchange and arthrolysis for the management of stiffness following total knee arthroplasty. J Bone Joint Surg Am
3. Bert JM, Reuben J, Kelly F, Gross M, Elting J. The incidence of modular tibial polyethylene insert exchange in total knee arthroplasty when polyethylene failure occurs. J Arthroplasty
. 1998;13: 609-614.
4. Bonner KF, Delanois RE, Harbach G, Bushelow M, Mont MA. Cementation of a polyethylene liner into a metal shell: factors related to mechanical stability. J Bone Joint Surg Am
. 2002;84: 1587-1593.
5. Brooks DH, Fehring TK, Griffin WL, Mason JB, McCoy TH. Polyethylene exchange only for prosthetic knee instability. Clin Orthop Relat Res
6. Conditt MA, Stein JA, Noble PC. Factors affecting the severity of backside wear of modular tibial inserts. J Bone Joint Surg Am
7. Engh GA, Koralewicz LM, Pereles TR. Clinical results of modular polyethylene insert exchange with retention of total knee arthroplasty components. J Bone Joint Surg Am
8. Engh GA, Lounici S, Rao AR, Collier MB. In vivo deterioration of tibial baseplate locking mechanisms in contemporary modular total knee components. J Bone Joint Surg Am
9. Gonzalez della Valle A, Ruzo PS, Li S, Pellicci P, Sculco TP, Salvati EA. Dislodgment of polyethylene liners in first and second- generation Harris-Galante acetabular components: a report of 18 cases. J Bone Joint Surg Am
10. Haft GF, Heiner AD, Dorr LD, Brown TD, Callaghan JJ. A biomechanical analysis of polyethylene liner cementation into a fixed metal acetabular shell. J Bone Joint Surg Am
11. Jiranek WA. Acetabular liner fixation by cement. Clin Orthop Relat Res
12. Lakdawala A, Todo S, Scott G. The significance of surface changes on retrieved femoral components after total knee replacement. J Bone Joint Surg Br
13. LaPorte DM, Mont MA, Pierre-Jacques H, Peyton RS, Hungerford DS. Technique for acetabular liner revision in a nonmodular metal- backed component. J Arthroplasty
14. Tradonsky S, Postak PD, Froimson AI, Greenwald AS. A comparison of the dissociation strength of modular acetabular components. Clin Orthop Relat Res
15. Williams VG 2nd, Whiteside LA, White SE, McCarthy DS. Fixation of ultrahigh-molecular-weight polyethylene liners to metal- backed acetabular cups. J Arthroplasty
16. Whiteside LA. Effect of porous-coating configuration on tibial osteolysis after total knee arthroplasty. Clin Orthop Relat Res
. 1995; 321:92-97.