Abstract
Background
Megaprostheses are commonly used for reconstruction after distal femoral resection in orthopaedic oncology. The polyethylene bearings in these reconstructions experience wear and wear-related complications that may result in revision surgery. Improved manufacturing and processing of polyethylene has increased the durability of components commonly used for routine arthroplasty. Alterations in the manufacture of polyethylene is expected to reduce the revision risk of oncologic megaprostheses, resulting in fewer revision procedures, but this has not been proven.
Questions/purposes
Is there a difference in the hazard of polyethylene wear or breakage leading to prosthetic revision between differences in polyethylene manufacture and processing based on a competing risk analysis?
Methods
This was a single-center, observational, retrospective comparative study of 224 patients who had distal femur megaprostheses with identical rotating hinge articulations and knee kinematics after oncologic surgery from 1993 to 2015. No differences in surgical indications, joint articular components and kinematics, age, sex, diagnosis, BMI, use of chemotherapy, or tumor stage were seen with the patient numbers available. Prosthetic survivorship free from prosthetic revision surgery because of polyethylene wear-related revisions, defined as breakage, increased excursion on varus-valgus stress, or new locking or giving way was compared between two groups of patients: group 1 polyethylene (P1) (66 patients) who had air-sterilized machined ram-extruded bar stock or group 2 polyethylene (P2) (158 patients) molded gamma-radiated argon-processed polyethylene components. The mean follow-up duration for the P1 group (89 ± 55 months) was not different from that of patients with P2 polyethylene (79 ± 63 months; p = 0.24) including 27% (18 of 66) of patients in the P1 group and 25% (40 of 158) of patients in the P2 group followed for more than 10 years. More patients in the P2 group were lost to follow-up (9.2%, 16 of 174) than in the P1 group (5.7%, 4 of 70) but this was not statistically different (chi square; p = 0.37). The hazard of revision because of polyethylene wear or breakage was calculated with a competing risk analysis using the Fine‐Gray subdistribution hazard model.
Results
The P1 implants had a higher hazard ratio for revision caused by polyethylene damage at 120 months than did the P2 polyethylene implants (P1 HR 0.24 [95% CI 0.13 to 0.36] versus HR 0.07 [95% CI 0.03 to 0.12]), which represents an estimated absolute risk reduction of 17% (95% CI 6.15 to 27.9).
Conclusion
Polyethylene damage can result in megaprosthetic revisions in patients undergoing oncologic procedures. The hazard of polyethylene failure resulting in revision surgery was lower in patients who received recent polyethylene than in patients with polyethylene produced by previous methods, enhancing the durability of distal femoral megaprosthetic reconstructions. Despite improvements in polyethylene manufacture and clinical results, revision solely because of polyethylene damage still occurs in 7% of patients by the 10-year timepoint; thus, more improvement is needed. Patients who receive these implants should be monitored for signs and symptoms of polyethylene damage.
Level of Evidence
Level III, therapeutic study.
Introduction
Megaprostheses are commonly used for reconstruction after distal femoral resection in orthopaedic oncology. These implants generally include metal surfaces that articulate with polyethylene bearings and with tibial and femoral bushings. Wear and gross damage of polyethylene components continue to compromise the long-term success of distal femur replacements [5].
The manufacturer of the rotating hinge megaprosthesis that has been used at our center almost exclusively modified its polyethylene manufacturing process in 2000 for this implant to make the material more wear-resistant by direct isostatic molding, processed in an inert environment, sterilized by gamma irradiation, and packaged, stored, and sterilized in an inert atmosphere (Table 1). These changes were designed to make the implant more durable, extend its lifespan, and reduce revision surgeries in patients with oncologic conditions [7-9, 10, 14, 15]. However, controversy exists regarding whether this manufacturing change was beneficial in conventional TKA, so it was unknown whether these changes translated into improved durability of rotating hinge megaprostheses. Reduction of the high frequency of prosthetic revision due to aseptic loosening and polyethylene wear is of utmost importance and is particularly relevant in our population of young patients with sarcomas with a long life expectancy. To our knowledge, we have not seen these two approaches to polyethylene fabrication compared clinically in patients who received megaprosthesis reconstructions of the distal femur.
Table 1.
Polyethylene manufacturing techniques
| Older manufacturing techniques | More recent manufacturing techniques | Purported benefits |
| Machined from ram-extruded bar stock material | Direct compression molding | Components uniformly consolidated |
| Processed in an air environment | Processed in an inert environment | Improved wear resistance |
| Air sterilization | Gamma irradiation sterilization | Resistance to oxidation degradation |
| Regular packaging | Argon packaging | Improved mechanical properties |
Therefore, we asked: Is there a difference in the hazard of polyethylene wear or breakage leading to prosthetic revision between differences in polyethylene manufacture and processing based on a competing risk analysis?
Patients and Methods
This single-center, observational, retrospective cohort study was performed under institutional review board protocol numbers 16-482, 16-1122, and 16-1123. We studied implants with identical knee kinematics and metal-polyethylene articulations from the same manufacturer. Osseous fixation was performed with a cemented or press-fit implant (Finn® or OSS® Biomet Inc, Warsaw, IN, USA) or a Compress® Compliant Pre-Stress Implant (Biomet Inc) according to service policy [10]. Fixation type was generally based on age; patients older than 60 years with a disease other than oligometastatic disease or who underwent local radiation therapy received cemented fixation of their implants. In other respects, they received identical articular components with identical kinematics. All operations were performed by fellowship-trained surgeons at a tertiary center from August 1993 to December 2015.
Operative reports, clinic notes, and radiographs were analyzed for each patient. The manufacture of each polyethylene component was checked by lot number. The institution is one of the 13 nationally designated Comprehensive Cancer Centers. It maintains a thorough follow-up registry and participates in the Surveillance Epidemiology and End Results (SEER) program of the National Cancer Institute. The follow-up is particularly accurate documenting patient deaths. Within the first 2 years, 3% (2 of 66) patients with P1 implants and 5% (8 of 174) of patients with P2 implants died (p = 0.59). These patients were not eligible for the long-term subdistribution competing risk analysis.
In our surgical database, we identified 3025 surgical knee procedures. In all, 224 patients who underwent limb-salvage surgery with a distal femoral rotating hinge knee arthroplasty that was performed for oncologic indications survived for more than 2 years (Fig. 1). There was no difference found in the long-term or overall follow-up between the groups. Twenty-seven percent (18 of 66) of patients with P1 polyethylene and 25% (40 of 158) of patients with P2 polyethylene were followed for more than 10 years. The follow-up duration (mean ±SD) was not different for patients with P1 polyethylene (89 ± 55 months) than it was for those with P2 polyethylene (79 ± 63 months; p = 0.24). There was no evident differential loss to follow-up between the groups: 6% (4 of 70) of P1 patients versus 12% (21 of 179) of P2 patients (chi square 0.16), leaving 66 patients in the P1 group and 158 patients in the P2 group. Overall mortality during the study period was not different between the groups; 15% (10 of 66) of P1 patients and 13% (20 of 158) of P2 patients died.
Fig. 1.
This patient selection flow chart shows the total number of patients recruited (n = 249) and the total number available for analysis (n = 224); OSS = Orthopaedic Salvage System.
Variables, Outcome Measures, Data Sources, and Bias
We compared patient demographics between the groups at presentation. The mean age was not different between patients with P1 (34 ± 20 years) and P2 (30 ± 16 years; p = 0.12) . The most common diagnoses were osteosarcoma (67% [150 of 224]) and chondrosarcoma (11% [25 of 224]). The patients’ BMIs were classified according to the World Health Organization’s criteria. With the number of patients analyzed, there were no differences in diagnoses or demographic characteristics between the two polyethylene groups (Table 2).
Table 2.
Patient demographics
| Earlier PE group, % (n = 66) | Press-fit more recent PE group, % (n = 158) | |
| Female | 34 | 79 |
| Mean age in years | 34 | 30 |
| BMI category | ||
| Underweight: < 18.5 kg/m2 | 5 (3) | 15 (24) |
| Normal weight: 18.5–25 kg/m2 | 33 (22) | 37 (58) |
| Overweight: 25–30 kg/m2 | 47 (31) | 27 (42) |
| Obese: > 30 kg/m2 | 15 (10) | 22 (34) |
| Deaths | 15 (10) | 13 (20) |
| No follow-up in the past 5 years | 18 (12) | 3 (4) |
| Most common diagnoses | Osteosarcoma (39) Chondrosarcoma (8) |
Osteosarcoma (111) Chondrosarcoma (17) |
PE = polyethylene.
The dependent variable was surgical revision for a combination of preoperative biomechanical factors reflecting polyethylene wear or insufficiency (increased pain, effusion, varus-valgus translation, clunking, locking or instability), intraoperative findings of marked polyethylene wear and/or fracture, and confirmed to be because of polyethylene if the symptoms resolved after polyethylene component revision. We excluded patients who underwent surgery for patellar pain or instability whose polyethylene was intact, although they too had polyethylene component replacement at the time of patellar resurfacing.
We assessed the time to revision arthroplasty with a competing risk analysis using the Fine‐Gray subdistribution hazard model. Our endpoint was revision for polyethylene damage. Revision for other causes (patellar pain because of degenerative arthritis, implant loosening, infection, bone fracture, implant fracture, tumor recurrence, limb length discrepancy, and patellar tendon rupture) and patient death were considered competing factors to our endpoint [3, 17]. The rate at which an event occurs (polyethylene wear or breakage resulting in revision surgery) differs over time. That rate is affected not only by the factor of interest (polyethylene characteristics), but also by other factors whose effects can vary over time. The Fine-Gray subdistribution method considers the risk of other factors (competing risks) and calculates the effect of the factor of interest based on the number of patients who are still at risk for the event to occur. The sub-distribution method controls for death and other causes of revision surgery while it calculates the cumulative incidence of polyethylene-related revision.
Statistical Analysis, Study Size
We performed all statistical analyses with SPSS Statistical Software, version 25.0 (IBM Inc, Armonk, NY, USA) except for the competing risk analysis, for which we used R version 3.5.3 [13]. To assess possible associations between different factors and the outcomes of interest, we used either the chi-square test or Fisher’s exact test for categorical variables, the Mann-Whitney test for ordinal variables, and ANOVA for continuous variables. For the survival analysis, we performed a competing risk analysis using the Fine‐Gray subdistribution hazard model, censoring patients at their last follow-up visit, at revision of the implant for polyethylene damage or other causes of revision, or at patient death [16]. We compared the absolute difference between the competing risks. It should be noted that when the coefficient is relatively small, the absolute contribution to the hazard is close to the estimate, whereas the hazard will diverge from the absolute incidence when the coefficients are larger. Thus, the absolute reduction in the hazard can be stated with certainty, but the magnitude of the relative difference is less certain when one risk is high [1]. The relative multiple of the hazards between the two polyethylene types is informative, but the magnitude of the differences can be amplified in a misleading way if there are even small differences in the effect size between the two groups. Thus, calculation of the relative hazard values (as a multiple) is avoided.
All tests were deemed significant if p was < 0.05.
Results
The P1 polyethylene implants had a higher hazard ratio for revision caused by polyethylene damage at 120 months than did P2 polyethylene implants (0.24 [95% CI 0.13 to 0.36] versus 0.07 [95% CI 0.03 to 0.12]; p = 0.04) (Fig. 2), which represents an absolute risk reduction of (24% - 7%) = 17% (95% CI 6.2 to 28). There was no difference between patients in the P1 and P2 groups regarding the other covariates contributing to revision or death, given the number of patients and events.
Fig. 2.
These Fine-Gray subdistribution competing risk curves depict the earlier polyethylene components above and the more recent polyethylene components below, with 95% CIs. The events were revision because of polyethylene damage (endpoint), other causes (nonpolyethylene-related), and death.
Discussion
Background and Rationale
Oncologic megaprostheses have high revision rates for many reasons. Improvements in implant design, bone fixation, and materials are potential methods to improve implant durability and reduce revisions. This goal is particularly important to achieve for young patients with oncologic conditions who may undergo many revision procedures. Modern polyethylene manufacturing techniques, such as direct compression molding, processing in an inert environment, gamma irradiation sterilization, and argon packaging have improved the components’ performance, leading to a decreased cumulative incidence of damage of more modern polyethylene parts in conventional joint arthroplasties [7-9, 14, 15]. However, controversy remains. For example, no improvement has been seen due to different polyethylene in posterior cruciate-sparing TKA [2], and highly crosslinked polyethylene did not have better survival in a major recent registry-based study [6]. We found that polyethylene components made by molded gamma-radiated argon-processed polyethylene (P2) were more durable than were the earlier-generation polyethylene components made by air-sterilized machined ram-extruded bar stock (P1) in patients who underwent distal femoral megaprosthesis reconstructions.
Limitations
The study has limitations. It addressed only one manufacturer and a single articulating knee design, albeit with different femoral fixation. The results may not apply to other methods of polyethylene processing or implants from other manufacturers. The femoral stem fixation methods encompassed cemented, press-fit, and compression osteointegration methods. However, the joint design and kinematics were identical in all groups, and stem fixation has not been identified as a risk factor for polyethylene wear to our knowledge. Therefore, this variable should have no influence on the outcome in question. Second, there may have been confounding indications for surgery; therefore, ascribing the surgical indication retrospectively could be inaccurate. For example, symptoms of polyethylene failure and patellar problems often overlap, and it was frequently difficult to distinguish which was the principal reason for revision surgery. To minimize this bias, we specifically excluded patients who were indicated for revision surgery because of patellar problems. Revision surgery was ascribed to polyethylene wear or damage if (1) the symptoms were consistent with this mechanism, (2) there was substantial polyethylene damage found at surgery, and (3) the patient’s symptoms resolved after polyethylene exchange. Furthermore, the potential bias introduced by this issue was minimized by verifying the absence of major polyethylene damage when operations were performed to resurface the patella.
Differential follow-up is always a potential problem in sequential studies, and we interrogated this issue in several ways. The percentage of patients in the P1 polyethylene group who were not seen in the past 5 years (28%; 16 in 56 living patients) was not different from that in the P2 polyethylene group (17%; 25 in 146 living patients; p = 0.21); although there was no statistical difference, this might still be considered a form of transfer bias. However, the patients with the P1 polyethylene who had the higher rates of polyethylene damage had a lower follow-up frequency. Assuming the worst-case scenario, polyethylene failure rates would have been even higher for the group with the P1 polyethylene implants and the difference in revision rates would have been even more substantial, thus this degree of differential follow-up strongly supported the conclusion that P2 was more durable than P1 polyethylene. There was no evident differential loss to follow-up between the groups: 6% (4 of 70) of patients in the P1 group versus 12% (21 of 179) of patients in the P2 group, leaving 66 patients with the P1 implant and 158 patients with the P2 implant. In all, 27% (18 of 66) of patients with P1 polyethylene and 25% (40 of 158) of patients with P2 polyethylene were followed for more than 10 years, so there was no difference in the number of patients with long-term follow-up conflating the results. Yet, there still may be concern about possible transfer bias influencing the results since there was a 10-month longer absolute difference in the mean duration of follow-up for patients with P1 versus P2 implants (89 ± 55 months for patients with P1 versus 79 ± 63 months for patients with P2 implants). However, since these differences were not statistically significant, and the other ways to evaluate the follow-up between the groups did not reveal a difference, the potential for meaningful transfer bias due to differences in the follow-up duration is unlikely to have contributed to the frequency of revisions due to polyethylene wear.
The method for statistical analysis is another possible limitation. Although some may prefer restricting the model to event-specific outcomes, the multiple covariates in this case alter the number of implants at risk at any given timepoint and make the subdistribution method preferable. It is the most suitable method to capture the rate of revision due to this cause over the span of the analysis.
Polyethylene Manufacture and Survivorship
No studies of which we are aware have addressed polyethylene failure and prosthetic survivorship in rotating hinge prostheses, and to our knowledge, none have addressed the factors in oncology. Patients with nononcologic conditions who have implants with the same kinematics had excellent pain relief and function, with a survival rate of 87% at 46 months [2, 12]. Berend and Lombardi’s study [2] was performed in a very different clinical context than ours. Notably, their nononcologic patients were much older than ours (mean age 76 years versus 32 years), had lower demands, and were studied for a shorter time. More importantly, specific polyethylene failure was not addressed; thus, comparison with our study is impossible.
Polyethylene implant failure has not been specifically evaluated in rotating hinge tumor endoprostheses. Among the most comprehensive studies of rotating hinge knee megaprostheses, Pala et al. [11] reported on the outcomes of 223 patients who underwent reconstruction with the GMRS® (Global Modular Replacement System®, Stryker, Rutherford , NJ, USA) prosthesis after a mean follow-up duration of 4 years, reporting on outcomes at 5 years based on Henderson criteria [4]. They reported that there was no Type 3 mechanical failure of the implant. It is impossible to reconcile these results with those of the current report, in which the mean follow-up duration was 7 years, and the outcome measure of polyethylene revision was at 10 years.
We examined a clinically relevant outcome measure: The revision arthroplasty rate in patients with distal femoral arthroplasties with identical kinematics before and after modification of the polyethylene manufacture and processing. The Fine-Gray competing risk analysis showed that the hazard of failure leading to revision arthroplasty was greater in the prior generation of polyethylene than in the recent version. A codependence of factors contributed to implant revision (for example, pain and arthritic changes that occurred in unresurfaced patellae). The variable number of implants that remained at risk of failure at any designated timepoint reduced the statistical power to analyze the result at that timepoint [16]. With the numbers of patients and events available, the differences could not be explained by other covariates. These results suggest that the recent polyethylene has made a difference by reducing the rate of polyethylene-driven revision arthroplasties.
Conclusion
Ultimately, all components are expected to wear; some of these will undergo wear-related revision surgery. The overarching goal is to prolong the implant’s durability as much as possible (Fig. 3). Further improvements in the performance of polyethylene or better knee kinematics will be needed to achieve this goal. We found that recent polyethylene components have a lower hazard of damage resulting in revision surgery than do earlier versions. The change in polyethylene processing is an important advance that has improved the durability of the implants that are currently used for oncologic reconstruction. Despite improvements in polyethylene manufacture and improved clinical results, revision solely for polyethylene damage still occurs in 7% of patients by the 10-year timepoint; therefore, more improvement is needed. In the meantime, patients who receive these implants should be monitored for signs and symptoms of polyethylene damage.
Fig. 3.
This photograph shows the new polyethylene before implantation (above) and the retrieved tibial bearing that resulted in revision surgery (below). Topside wear, delamination, and edge fractures are apparent on the tibial bearing component and in the central yoke box. Visually, polyethylene femoral and tibial bushings (not shown) had negligible wear.
Acknowledgments
We thank Julio Trecenti, PhD, of São Paulo University for statistical assistance, and Jessica Massler MSW, for editorial assistance.
Footnotes
The institution of one or more of the authors (ACB, MAE, MAY, JHH) has received funding from the National Institutes of Health/National Cancer Institute Cancer Center Support Grant (P30 CA008748) and an educational grant from the OMeGA Medical Grants Foundation (supporting ACB; grant number ujyq8083uj).
One of the authors certifies that he (JHH), or a member of his immediate family, has received or may receive payments or benefits, during the study period, an amount of USD 10,000 to USD 100,000 from Stryker (Mahwah, NJ, USA).
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research® editors and board members are on file with the publication and can be viewed on request.
Each author certifies that his or her institution approved the human protocol for this investigation and that all investigations were conducted in conformity with ethical principles of research.
References
- 1.Austin PC, Fine JP. Practical recommendations for reporting Fine-Gray model analyses for competing risk data. Stat Med. 2017;36:4391-4400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Berend KR, Lombardi AV., Jr. Distal femoral replacement in nontumor cases with severe bone loss and instability. Clin Orthop Relat Res. 2009;467:485-492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Healey JH, Morris CD, Athanasian EA, Boland PJ. Compress knee arthroplasty has 80% 10-year survivorship and novel forms of bone failure. Clin Orthop Relat Res. 2013;471:774-783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Henderson ER, Groundland JS, Pala E, Dennis JA, Wooten R, Cheong D, Windhager R, Kotz RI, Mercuri M, Funovics PT, Hornicek FJ, Temple HT, Ruggieri P, Letson GD. Failure mode classification for tumor endoprostheses: retrospective review of five institutions and a literature review. J Bone Joint Surg Am. 2011;93:418-429. [DOI] [PubMed] [Google Scholar]
- 5.Kawai A, Muschler GF, Lane JM, Otis JC, Healey JH. Prosthetic knee replacement after resection of a malignant tumor of the distal part of the femur. Medium to long-term results. J Bone Joint Surg Am. 1998;80:636-647. [DOI] [PubMed] [Google Scholar]
- 6.Kim YH, Park JW, Kim JS, Lee JH. Highly crosslinked-remelted versus less-crosslinked polyethylene in posterior cruciate-retaining TKAs in the same patients. Clin Orthop Relat Res. 2015;473:3588-3594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Krenn S, Thomsen M, Usbeck S, Scheuber LF, Boettner F, Krukemeyer MG, Huber M, Kretzer JP, Gehrke T, Krenn V. Supramacroparticulate PE in 6 different joint endoprostheses localisations: An indicator for PE damage? Pathol Res Pract. 2017;213:987-996. [DOI] [PubMed] [Google Scholar]
- 8.Kurtz SM, Walker PS. How have new designs and new types of joint replacement influenced wear behavior? J Am Acad Orthop Surg. 2008;16 Suppl 1:S107-110. [DOI] [PubMed] [Google Scholar]
- 9.Minoda Y, Hata K, Goto K, Itohara T, Nakamura H. Sequentially annealed highly cross-linked polyethylene reduced in vivo wear particle generation in total knee arthroplasty. J Orthop Surg (Hong Kong). 2017;25:2309499017718909. [DOI] [PubMed] [Google Scholar]
- 10.OSS™ Orthopedic Salvage System. Comprehensive® Segmental Revision System. 2019. Available at: www.zimmerbiomet.com/medical-professionals/limb-salvage/orthopaedic-salvage-system.html. Accessed April 16, 2020.
- 11.Pala E, Trovarelli G, Calabro T, Angelini A, Abati CN, Ruggieri P. Survival of modern knee tumor megaprostheses: failures, functional results, and a comparative statistical analysis. Clin Orthop Relat Res. 2015;473:891-899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Partridge TCJ, Baker PN, Jameson SS, Mason J, Reed MR, Deehan DJ. Conventional Versus Highly Cross-Linked Polyethylene in Primary Total Knee Replacement: A Comparison of Revision Rates Using Data from the National Joint Registry for England, Wales, and Northern Ireland. J Bone Joint Surg Am. 2020;102:119-127. [DOI] [PubMed] [Google Scholar]
- 13.R Core Team (2018). R: a language and environment for statistical computing. 2013. Accessed June 13, 2019. [Google Scholar]
- 14.Rimnac CM, Klein RW, Betts F, Wright TM. Post-irradiation aging of ultra-high molecular weight polyethylene. J Bone Joint Surg Am. 1994;76:1052-1056. [DOI] [PubMed] [Google Scholar]
- 15.Ritter MA, Worland R, Saliski J, Helphenstine JV, Edmondson KL, Keating EM, Faris PM, Meding JB. Flat-on-flat, nonconstrained, compression molded polyethylene total knee replacement. Clin Orthop Relat Res. 1995:79-85. [PubMed] [Google Scholar]
- 16.Therneau T CC, Atkinson E. Multi-state models and competing risks. 2020. Available at: https://cran.r-project.org/web/packages/survival/vignettes/compete.pdf. Accessed April 14, 2020.
- 17.Zimel MN, Farfalli GL, Zindman AM, Riedel ER, Morris CD, Boland PJ, Healey JH. Revision distal femoral arthroplasty with the Compress® prosthesis has a low rate of mechanical failure at 10 years. Clin Orthop Relat Res. 2016;474:528-536. [DOI] [PMC free article] [PubMed] [Google Scholar]