- Email: [email protected]
Unicondylar Knee Retrieval Analysis
The Journal of Arthroplasty Vol. 25 No. 6 Suppl. 1 2010
Unicondylar Knee Retrieval Analysis Theodore T. Manson, MD,* Natalie H. Kelly, BS,y Joseph D. Lipman, MS,y Timothy M. Wright, PhD,y and Geoffrey H. Westrich, MDy
Abstract: Unicondylar knee arthroplasty (UKA) is considered an alternative to total knee arthroplasty for patients who have arthritis limited to one compartment of the knee. This study examined surface damage of 3 contemporary UKA designs that were retrieved at revision surgery. Two of the UKA designs were fixed bearing and one was mobile bearing. Demographic information was collected, as well as information about the implants used at revision surgery. Articular surface damage was greater in the fixed-bearing designs as compared to the mobile bearing, although the mobile-bearing implants had significantly shorter length of implantation. Backside damage was also graded for the mobile bearing and when combined with articular wear resulted in overall damage scores higher than both fixed-bearing designs. The fixed-bearing designs showed delamination and surface deformation, whereas the mobile bearing had no evidence of these damage modes. However, mobile-bearing components showed other types of wear, and significant wear damage was present on the bearing surfaces of the mobile-bearing implants despite a short time of implantation. At the time of conversion to a total knee arthroplasty, more than 50% of cases required the use of stems, augments, or constrained inserts for the tibial reconstruction. In conclusion, wear modes differed among UKA prosthesis designs. Revision of a UKA to a total knee arthroplasty remains complex with the tibial preparation more complicated than in the primary setting. Keywords: articular wear, mobile-bearing implants, delamination, unicondylar knee, retrieval analysis. © 2010 Published by Elsevier Inc.
Unicondylar knee arthroplasty (UKA) is an alternative to high tibial osteotomy or total knee arthroplasty (TKA) in patients with limited arthrosis. Although results of early designs were disappointing [1-4], success with more recent designs has increased interest in these devices [5-8]. Two studies have determined polyethylene wear rates [9,10] in retrieval of UKA specimens. No formal analysis of wear patterns has been conducted in modern UKA. The objective of this study was to quantify the differences in wear patterns among 3 modern UKA designs: the Oxford mobile-bearing UKA (Biomet, War-
From the *Department of Orthopaedics, R Adams Cowley Shock Trauma Center, University of Maryland, Baltimore, Maryland; and yHospital for Special Surgery, Department of Adult Reconstruction and Joint Replacement, Weill Cornell Medical College, New York, New York. Submitted July 20, 2009; accepted May 5, 2010. Benefits or funds were received in partial or total support of the research material described in this article. These benefits and/or support were received from the Clark and Kirby Foundations. Investigation Performed at Hospital for Special Surgery, New York, New York. Reprint requests: Geoffrey Westrich, MD, Hospital for Special Surgery, Department of Adult Reconstruction and Joint Replacement, 535 East 70th St, New York, NY 10021. © 2010 Published by Elsevier Inc. 0883-5403/2506-0022$36.00/0 doi:10.1016/j.arth.2010.05.004
saw, Ind), the Repicci fixed-bearing all-polyethylene tibia UKA (Biomet), and the Miller-Galante fixed-bearing metal-backed tibia UKA (Zimmer, Warsaw, Ind). The mobile-bearing design is intended to reduce the stresses associated with fatigue wear by providing a second articulation between the polyethylene tibial insert and the underlying metal-backing. Therefore, we hypothesized that the bearing surfaces of the retrieved components from the mobile-bearing UKA design would show pitting and scratching, but little delamination or surface deformation, whereas those of the 2 fixed-bearing designs would show delamination and surface deformation in addition to pitting and scratching.
Materials and Methods The study design was a retrospective cohort analysis. Forty-three UKA tibial components were retrieved at revision as part of an ongoing institutional review board– approved retrieval analysis program. This collection represents all the medial Oxford mobile bearing, Repicci, and Miller-Galante components retrieved from 1999 to 2008 at our center. We evaluated the retrieved components using visual and stereomicroscopic examination. We quantified wear damage based on a subjective scale from 0 to 3 for each of 7 wear modes (scratching, pitting, burnishing,
108
Unicondylar Knee Retrieval Analysis Manson et al
109
Table 1. Description of Augments, Stems, and Constrained Liners Used at the Time of Revision Surgery to a TKA
Oxford (n = 12) Repicci (n = 17) Miller-Galante (n = 14)
Required Femoral Stem
Required Femoral Augments
Required Tibial Stem
Required Tibial Augments
Required Constrained Articular Insert
Noncomplex Revision
2 2 1
0 0 1
5 5 7
4 2 1
3 9 4
5 8 7
Noncomplex revision represents those patients who did not require the use of a stem, augment, or constrained liner at the time of revision.
embedded debris, abrasion, delamination, and surface deformation) [11]. Scoring of the wear damage was repeated for each of 4 quadrants of the bearing surfaces of the tibial polyethylene components (medial, anterior, lateral, and posterior). A cumulative score was assigned to each component, in addition to calculating the mean score for each of the 7 wear modes. The highest possible cumulative score was 84 for a component (7 modes × maximum score of 3 × 4 quadrants). We also examined the backside of the mobile-bearing components, using the same scoring system as for the tibiofemoral bearing surface. We obtained demographic data (age, sex, height, weight, length of implantation, reason for revision, and the need for revision stems, augments, and constrained liners at the time of revision) from a review of the patients' charts from whom the implants had been retrieved. We determined tibial component varus/valgus and flexion/extension alignment and weight-bearing tibiofemoral alignment from the radiographs that had been obtained immediately pre-revision. Tibial component alignment was reported relative to the diaphysis of the tibia. Most of the primary surgeries were performed at outside institutions, so we had limited access to radiographs taken in the immediate postoperative period after the primary surgery. Statistical comparisons were made using Student t test, analysis of variance (ANOVA), and Tukey HSD post hoc analysis, as well as Kruskal-Wallis with Dunn's method where appropriate.
Repicci implants (range, 6-80 months), and 20 months for the Miller-Galante implants (range, 3-217 months). The Repicci implants were implanted significantly longer than the Oxford implants (P b .02, Kruskal-Wallis). More than 50% of the revision surgeries required the use of stem extensions, augments, or constrained liners (Oxford, 58%; Repicci, 53%; and Miller-Galante, 50%). Most of the stems and augments were used on the tibial side (Table 1). Radiographs were available for all but 2 of the study patients. No significant differences were found in radiographic tibiofemoral alignment among the groups (Oxford, 3° varus; Repicci, 2° varus; and Miller-Galante, 1° valgus). Likewise, no significant differences were found in tibial component flexion (posterior slope) among the groups (Oxford, 8°; Repicci, 7°; and Miller-Galante, 8°). With respect to the diaphysis of the tibia, Repicci
Results Clinical Data The reasons for revision were loosening in 56% of the cases, unexplained pain in 30%, progression of arthritis in 7%, infection in 2%, and bearing dislocation in 5%. No differences existed among the 3 design groups regarding the reason for revision except for bearing dislocation, which only occurred in the Oxford group. The mean age of the patients at the time of implant removal was 61.7 years with no significant differences among the 3 groups (ANOVA). Likewise, no differences existed in patient weight (mean weight, 84 kg) or body mass index (mean body mass index, 29 kg/m2). The median time of implantation was 14 months for the Oxford implants (range, 2-30 months), 47 months for the
Fig. 1. Scratching and pitting present on the front side (A) and backside (B) of an Oxford mobile-bearing polyethylene insert.
110 The Journal of Arthroplasty Vol. 25 No. 6 Suppl. 1 September 2010
Fig. 2. Individual mean wear damage separated according to wear mode for each of the 3 different prostheses.
components were in significantly more varus (14°) than either the Oxford (4°) or Miller-Galante (3°) prostheses. Retrieval Analysis Cumulative wear scores were higher for the Repicci and Miller-Galante prostheses (33.6 and 33.7, respectively) than for the Oxford prostheses (22.6, P b .01, ANOVA with Tukey post hoc HSD). The backside of the Oxford prostheses also showed wear damage with an average score of 16.3 (Fig. 1). Individual mean wear damage scores separated according to wear mode are shown in Fig. 2. Repicci and Miller-Galante prostheses had mild delamination scores (0.9 and 0.8, respectively), whereas none of Oxford Prostheses showed delamination. Likewise, Repicci and Miller-Galante prostheses had mild surface deformation scores (1.8 and 1.6, respectively), whereas the Oxford prostheses showed no surface deformation. All 3 groups showed pitting with no significant differences among the pitting scores for the 3 groups (ANOVA). Scratching was significant in the Oxford and Repicci groups (2.6 in each group) and significantly less prominent in the Miller-Galante Group (1.9, P b .01, ANOVA with Tukey post hoc HSD). The Oxford and Miller-Galante designs showed more abrasion (0.7 in each group) than the Repicci design (0.2, P b .01, Kruskal-Wallis with Dunn method). Embedded debris was similar among the groups: Repicci, 0.3; Oxford, 0.2; and Miller-Galante, 0.2. No correlation was found in this group of 43 implants between pre-revision radiographic alignment and the location of wear damage or the wear mode observed on the retrieval specimens.
Discussion We reviewed a large number of UKA retrieval specimens in an effort to determine if differences in wear damage existed between fixed-bearing and mobilebearing implants. By design, this group excludes patients with well-functioning prostheses. Most retrieved implants were revised for loosening (56%) or unex-
plained pain (30%) with 2 of the Oxford prostheses revised for bearing dislocation (5%). We had a relatively small number of patients who were revised for radiographic progression of arthritis (7%). However, it is possible that some of the patients with unexplained pain had progression of arthritis not visible on the prerevision radiographs. On the articular side, we noted that the cumulative wear scores were higher for the fixed-bearing unicompartmental designs (Repicci [33.6] and Miller-Galante [33.7]) than for the mobile-bearing Oxford prostheses (22.6). The Oxford mobile-bearing prostheses showed less influence of shear-induced wear patterns (surface deformation and delamination) than the Repicci and Miller-Galante prostheses. We also noted that the wear damage modes were different for the fixed-bearing prostheses compared to the mobile-bearing design with delamination and surface deformation noted only in the fixed-bearing designs. All 3 groups had a similar amount of pitting and embedded debris. The increased articular-side wear in the fixed-bearing unicompartmental designs is most likely related to the greater contact stresses that are noted in fixed-bearing designs. It appears from our retrieval analysis that mobile-bearing designs are able to achieve less articular-side contact stresses that result in less wear damage on the articular side of the polyethylene. However, we also noted that the backside of the Oxford prostheses also showed wear damage with an average score of 16.3. When the cumulative wear scores of the Oxford prosthesis on the articular side (22.6) and the backside (16.3) were added, the score (38.9) was greater than either of the fixed-bearing designs. Therefore, the overall wear (articular and backside) that was noted with the Oxford mobile-bearing design may implicate potentially a greater amount of total wear damage shared between both bearing surfaces of the polyethylene insert. This is particularly concerning given the relatively short time of implantation of the Oxford implants. We did not assess volumetric or gravimetric wear with these retrieved implants, but further clinical and laboratory studies would better elucidate these concerns. Oxford prostheses demonstrated wear patterns similar to total hip arthroplasty acetabular polyethylene inserts (scratching, abrasion) and therefore more consistent with abrasive and adhesive wear mechanisms. Repicci and Miller-Galante prostheses showed wear patterns more akin to TKA polyethylene inserts (delamination, surface deformation), consistent with fatigue and shear stress– related mechanisms. All implant designs showed similar amounts of pitting. One limitation of the study is that the mean time of implantation was shorter for the Oxford prostheses (14 months) compared to the Miller-Galante implants (20 months) and the Repicci implants (47 months). In theory, the cumulative wear may be
Unicondylar Knee Retrieval Analysis Manson et al
greater with the Oxford over a longer time of implantation, but the wear modes may still be similar to those we observed in our retrieval analysis. In addition, our conclusion that the wear scores were lower with the Oxford prostheses should be tempered by the fact that the length of implantation was much shorter than the fixed-bearing prostheses. Also, our study is a description of wear damage, not of absolute wear. Combined with the relatively short time of implantation of the Oxford components in our study, we can make no endorsement of decreased wear for a mobile-bearing construct. With respect to the radiographic alignment, the Repicci tibial components were in significantly more varus than the other 2 designs at the time of revision. This was so, even allowing for the fact that the Repicci design is implanted to match the natural 3° varus alignment of the tibial plateau. The lack of immediate postoperative alignment radiographs precludes us from determining if the components were implanted in this orientation or whether settling of the implant with subchondral collapse had occurred by the time they came to revision. Revision of a UKA to a TKA was complex, requiring the use of stems, augments, or constrained liners in more than 50% of cases. However, only rarely was a femoral stem or augment required. In general, the claims that revision of a unicondylar prosthesis to a TKA is equivalent to performing a primary TKA seem unreasonable, at least with regard to the tibial component in our experience. Likewise, there were no differences among prosthesis designs with respect to revision difficulty. In conclusion, wear modes differed among UKA prosthesis designs with no obvious evidence of more idealized wear performance between fixed-bearing and mobile-bearing implants. Revision of a UKA to a TKA
111
remains complex with the tibial preparation more complicated than in the primary setting.
References 1. Insall J, Walker P. Unicondylar knee replacement. Clin Orthop Relat Res 1976;83. 2. Insall J, Aglietti P. A five to seven-year follow-up of unicondylar arthroplasty. J Bone Joint Surg Am 1980;62: 1329. 3. Laskin RS. Unicompartmental tibiofemoral resurfacing arthroplasty. J Bone Joint Surg Am 1978;60:182. 4. Padgett DE, Stern SH, Insall JN. Revision total knee arthroplasty for failed unicompartmental replacement. J Bone Joint Surg Am 1991;73:186. 5. Berger RA, Meneghini RM, Jacobs JJ, et al. Results of unicompartmental knee arthroplasty at a minimum of ten years of follow-up. J Bone Joint Surg Am 2005;87:999. 6. Naudie D, Guerin J, Parker DA, et al. Medial unicompartmental knee arthroplasty with the Miller-Galante prosthesis. J Bone Joint Surg Am 2004;86-A:1931. 7. Romanowski MR, Repicci JA. Minimally invasive unicondylar arthroplasty: eight-year follow-up. J Knee Surg 2002; 15:17. 8. Price AJ, Waite JC, Svard U. Long-term clinical results of the medial Oxford unicompartmental knee arthroplasty. Clin Orthop Relat Res 2005;171. 9. Collier MB, Engh Jr CA, McAuley JP, et al. Factors associated with the loss of thickness of polyethylene tibial bearings after knee arthroplasty. J Bone Joint Surg Am 2007;89:1306. 10. Psychoyios V, Crawford RW, O'Connor JJ, et al. Wear of congruent meniscal bearings in unicompartmental knee arthroplasty: a retrieval study of 16 specimens. J Bone Joint Surg Br 1998;80:976. 11. Hood RW, Wright TM, Burstein AH. Retrieval analysis of total knee prostheses: a method and its application to 48 total condylar prostheses. J Biomed Mater Res 1983; 17:829.
Unicondylar Knee Retrieval Analysis Theodore T. Manson, MD,* Natalie H. Kelly, BS,y Joseph D. Lipman, MS,y Timothy M. Wright, PhD,y and Geoffrey H. Westrich, MDy
Abstract: Unicondylar knee arthroplasty (UKA) is considered an alternative to total knee arthroplasty for patients who have arthritis limited to one compartment of the knee. This study examined surface damage of 3 contemporary UKA designs that were retrieved at revision surgery. Two of the UKA designs were fixed bearing and one was mobile bearing. Demographic information was collected, as well as information about the implants used at revision surgery. Articular surface damage was greater in the fixed-bearing designs as compared to the mobile bearing, although the mobile-bearing implants had significantly shorter length of implantation. Backside damage was also graded for the mobile bearing and when combined with articular wear resulted in overall damage scores higher than both fixed-bearing designs. The fixed-bearing designs showed delamination and surface deformation, whereas the mobile bearing had no evidence of these damage modes. However, mobile-bearing components showed other types of wear, and significant wear damage was present on the bearing surfaces of the mobile-bearing implants despite a short time of implantation. At the time of conversion to a total knee arthroplasty, more than 50% of cases required the use of stems, augments, or constrained inserts for the tibial reconstruction. In conclusion, wear modes differed among UKA prosthesis designs. Revision of a UKA to a total knee arthroplasty remains complex with the tibial preparation more complicated than in the primary setting. Keywords: articular wear, mobile-bearing implants, delamination, unicondylar knee, retrieval analysis. © 2010 Published by Elsevier Inc.
Unicondylar knee arthroplasty (UKA) is an alternative to high tibial osteotomy or total knee arthroplasty (TKA) in patients with limited arthrosis. Although results of early designs were disappointing [1-4], success with more recent designs has increased interest in these devices [5-8]. Two studies have determined polyethylene wear rates [9,10] in retrieval of UKA specimens. No formal analysis of wear patterns has been conducted in modern UKA. The objective of this study was to quantify the differences in wear patterns among 3 modern UKA designs: the Oxford mobile-bearing UKA (Biomet, War-
From the *Department of Orthopaedics, R Adams Cowley Shock Trauma Center, University of Maryland, Baltimore, Maryland; and yHospital for Special Surgery, Department of Adult Reconstruction and Joint Replacement, Weill Cornell Medical College, New York, New York. Submitted July 20, 2009; accepted May 5, 2010. Benefits or funds were received in partial or total support of the research material described in this article. These benefits and/or support were received from the Clark and Kirby Foundations. Investigation Performed at Hospital for Special Surgery, New York, New York. Reprint requests: Geoffrey Westrich, MD, Hospital for Special Surgery, Department of Adult Reconstruction and Joint Replacement, 535 East 70th St, New York, NY 10021. © 2010 Published by Elsevier Inc. 0883-5403/2506-0022$36.00/0 doi:10.1016/j.arth.2010.05.004
saw, Ind), the Repicci fixed-bearing all-polyethylene tibia UKA (Biomet), and the Miller-Galante fixed-bearing metal-backed tibia UKA (Zimmer, Warsaw, Ind). The mobile-bearing design is intended to reduce the stresses associated with fatigue wear by providing a second articulation between the polyethylene tibial insert and the underlying metal-backing. Therefore, we hypothesized that the bearing surfaces of the retrieved components from the mobile-bearing UKA design would show pitting and scratching, but little delamination or surface deformation, whereas those of the 2 fixed-bearing designs would show delamination and surface deformation in addition to pitting and scratching.
Materials and Methods The study design was a retrospective cohort analysis. Forty-three UKA tibial components were retrieved at revision as part of an ongoing institutional review board– approved retrieval analysis program. This collection represents all the medial Oxford mobile bearing, Repicci, and Miller-Galante components retrieved from 1999 to 2008 at our center. We evaluated the retrieved components using visual and stereomicroscopic examination. We quantified wear damage based on a subjective scale from 0 to 3 for each of 7 wear modes (scratching, pitting, burnishing,
108
Unicondylar Knee Retrieval Analysis Manson et al
109
Table 1. Description of Augments, Stems, and Constrained Liners Used at the Time of Revision Surgery to a TKA
Oxford (n = 12) Repicci (n = 17) Miller-Galante (n = 14)
Required Femoral Stem
Required Femoral Augments
Required Tibial Stem
Required Tibial Augments
Required Constrained Articular Insert
Noncomplex Revision
2 2 1
0 0 1
5 5 7
4 2 1
3 9 4
5 8 7
Noncomplex revision represents those patients who did not require the use of a stem, augment, or constrained liner at the time of revision.
embedded debris, abrasion, delamination, and surface deformation) [11]. Scoring of the wear damage was repeated for each of 4 quadrants of the bearing surfaces of the tibial polyethylene components (medial, anterior, lateral, and posterior). A cumulative score was assigned to each component, in addition to calculating the mean score for each of the 7 wear modes. The highest possible cumulative score was 84 for a component (7 modes × maximum score of 3 × 4 quadrants). We also examined the backside of the mobile-bearing components, using the same scoring system as for the tibiofemoral bearing surface. We obtained demographic data (age, sex, height, weight, length of implantation, reason for revision, and the need for revision stems, augments, and constrained liners at the time of revision) from a review of the patients' charts from whom the implants had been retrieved. We determined tibial component varus/valgus and flexion/extension alignment and weight-bearing tibiofemoral alignment from the radiographs that had been obtained immediately pre-revision. Tibial component alignment was reported relative to the diaphysis of the tibia. Most of the primary surgeries were performed at outside institutions, so we had limited access to radiographs taken in the immediate postoperative period after the primary surgery. Statistical comparisons were made using Student t test, analysis of variance (ANOVA), and Tukey HSD post hoc analysis, as well as Kruskal-Wallis with Dunn's method where appropriate.
Repicci implants (range, 6-80 months), and 20 months for the Miller-Galante implants (range, 3-217 months). The Repicci implants were implanted significantly longer than the Oxford implants (P b .02, Kruskal-Wallis). More than 50% of the revision surgeries required the use of stem extensions, augments, or constrained liners (Oxford, 58%; Repicci, 53%; and Miller-Galante, 50%). Most of the stems and augments were used on the tibial side (Table 1). Radiographs were available for all but 2 of the study patients. No significant differences were found in radiographic tibiofemoral alignment among the groups (Oxford, 3° varus; Repicci, 2° varus; and Miller-Galante, 1° valgus). Likewise, no significant differences were found in tibial component flexion (posterior slope) among the groups (Oxford, 8°; Repicci, 7°; and Miller-Galante, 8°). With respect to the diaphysis of the tibia, Repicci
Results Clinical Data The reasons for revision were loosening in 56% of the cases, unexplained pain in 30%, progression of arthritis in 7%, infection in 2%, and bearing dislocation in 5%. No differences existed among the 3 design groups regarding the reason for revision except for bearing dislocation, which only occurred in the Oxford group. The mean age of the patients at the time of implant removal was 61.7 years with no significant differences among the 3 groups (ANOVA). Likewise, no differences existed in patient weight (mean weight, 84 kg) or body mass index (mean body mass index, 29 kg/m2). The median time of implantation was 14 months for the Oxford implants (range, 2-30 months), 47 months for the
Fig. 1. Scratching and pitting present on the front side (A) and backside (B) of an Oxford mobile-bearing polyethylene insert.
110 The Journal of Arthroplasty Vol. 25 No. 6 Suppl. 1 September 2010
Fig. 2. Individual mean wear damage separated according to wear mode for each of the 3 different prostheses.
components were in significantly more varus (14°) than either the Oxford (4°) or Miller-Galante (3°) prostheses. Retrieval Analysis Cumulative wear scores were higher for the Repicci and Miller-Galante prostheses (33.6 and 33.7, respectively) than for the Oxford prostheses (22.6, P b .01, ANOVA with Tukey post hoc HSD). The backside of the Oxford prostheses also showed wear damage with an average score of 16.3 (Fig. 1). Individual mean wear damage scores separated according to wear mode are shown in Fig. 2. Repicci and Miller-Galante prostheses had mild delamination scores (0.9 and 0.8, respectively), whereas none of Oxford Prostheses showed delamination. Likewise, Repicci and Miller-Galante prostheses had mild surface deformation scores (1.8 and 1.6, respectively), whereas the Oxford prostheses showed no surface deformation. All 3 groups showed pitting with no significant differences among the pitting scores for the 3 groups (ANOVA). Scratching was significant in the Oxford and Repicci groups (2.6 in each group) and significantly less prominent in the Miller-Galante Group (1.9, P b .01, ANOVA with Tukey post hoc HSD). The Oxford and Miller-Galante designs showed more abrasion (0.7 in each group) than the Repicci design (0.2, P b .01, Kruskal-Wallis with Dunn method). Embedded debris was similar among the groups: Repicci, 0.3; Oxford, 0.2; and Miller-Galante, 0.2. No correlation was found in this group of 43 implants between pre-revision radiographic alignment and the location of wear damage or the wear mode observed on the retrieval specimens.
Discussion We reviewed a large number of UKA retrieval specimens in an effort to determine if differences in wear damage existed between fixed-bearing and mobilebearing implants. By design, this group excludes patients with well-functioning prostheses. Most retrieved implants were revised for loosening (56%) or unex-
plained pain (30%) with 2 of the Oxford prostheses revised for bearing dislocation (5%). We had a relatively small number of patients who were revised for radiographic progression of arthritis (7%). However, it is possible that some of the patients with unexplained pain had progression of arthritis not visible on the prerevision radiographs. On the articular side, we noted that the cumulative wear scores were higher for the fixed-bearing unicompartmental designs (Repicci [33.6] and Miller-Galante [33.7]) than for the mobile-bearing Oxford prostheses (22.6). The Oxford mobile-bearing prostheses showed less influence of shear-induced wear patterns (surface deformation and delamination) than the Repicci and Miller-Galante prostheses. We also noted that the wear damage modes were different for the fixed-bearing prostheses compared to the mobile-bearing design with delamination and surface deformation noted only in the fixed-bearing designs. All 3 groups had a similar amount of pitting and embedded debris. The increased articular-side wear in the fixed-bearing unicompartmental designs is most likely related to the greater contact stresses that are noted in fixed-bearing designs. It appears from our retrieval analysis that mobile-bearing designs are able to achieve less articular-side contact stresses that result in less wear damage on the articular side of the polyethylene. However, we also noted that the backside of the Oxford prostheses also showed wear damage with an average score of 16.3. When the cumulative wear scores of the Oxford prosthesis on the articular side (22.6) and the backside (16.3) were added, the score (38.9) was greater than either of the fixed-bearing designs. Therefore, the overall wear (articular and backside) that was noted with the Oxford mobile-bearing design may implicate potentially a greater amount of total wear damage shared between both bearing surfaces of the polyethylene insert. This is particularly concerning given the relatively short time of implantation of the Oxford implants. We did not assess volumetric or gravimetric wear with these retrieved implants, but further clinical and laboratory studies would better elucidate these concerns. Oxford prostheses demonstrated wear patterns similar to total hip arthroplasty acetabular polyethylene inserts (scratching, abrasion) and therefore more consistent with abrasive and adhesive wear mechanisms. Repicci and Miller-Galante prostheses showed wear patterns more akin to TKA polyethylene inserts (delamination, surface deformation), consistent with fatigue and shear stress– related mechanisms. All implant designs showed similar amounts of pitting. One limitation of the study is that the mean time of implantation was shorter for the Oxford prostheses (14 months) compared to the Miller-Galante implants (20 months) and the Repicci implants (47 months). In theory, the cumulative wear may be
Unicondylar Knee Retrieval Analysis Manson et al
greater with the Oxford over a longer time of implantation, but the wear modes may still be similar to those we observed in our retrieval analysis. In addition, our conclusion that the wear scores were lower with the Oxford prostheses should be tempered by the fact that the length of implantation was much shorter than the fixed-bearing prostheses. Also, our study is a description of wear damage, not of absolute wear. Combined with the relatively short time of implantation of the Oxford components in our study, we can make no endorsement of decreased wear for a mobile-bearing construct. With respect to the radiographic alignment, the Repicci tibial components were in significantly more varus than the other 2 designs at the time of revision. This was so, even allowing for the fact that the Repicci design is implanted to match the natural 3° varus alignment of the tibial plateau. The lack of immediate postoperative alignment radiographs precludes us from determining if the components were implanted in this orientation or whether settling of the implant with subchondral collapse had occurred by the time they came to revision. Revision of a UKA to a TKA was complex, requiring the use of stems, augments, or constrained liners in more than 50% of cases. However, only rarely was a femoral stem or augment required. In general, the claims that revision of a unicondylar prosthesis to a TKA is equivalent to performing a primary TKA seem unreasonable, at least with regard to the tibial component in our experience. Likewise, there were no differences among prosthesis designs with respect to revision difficulty. In conclusion, wear modes differed among UKA prosthesis designs with no obvious evidence of more idealized wear performance between fixed-bearing and mobile-bearing implants. Revision of a UKA to a TKA
111
remains complex with the tibial preparation more complicated than in the primary setting.
References 1. Insall J, Walker P. Unicondylar knee replacement. Clin Orthop Relat Res 1976;83. 2. Insall J, Aglietti P. A five to seven-year follow-up of unicondylar arthroplasty. J Bone Joint Surg Am 1980;62: 1329. 3. Laskin RS. Unicompartmental tibiofemoral resurfacing arthroplasty. J Bone Joint Surg Am 1978;60:182. 4. Padgett DE, Stern SH, Insall JN. Revision total knee arthroplasty for failed unicompartmental replacement. J Bone Joint Surg Am 1991;73:186. 5. Berger RA, Meneghini RM, Jacobs JJ, et al. Results of unicompartmental knee arthroplasty at a minimum of ten years of follow-up. J Bone Joint Surg Am 2005;87:999. 6. Naudie D, Guerin J, Parker DA, et al. Medial unicompartmental knee arthroplasty with the Miller-Galante prosthesis. J Bone Joint Surg Am 2004;86-A:1931. 7. Romanowski MR, Repicci JA. Minimally invasive unicondylar arthroplasty: eight-year follow-up. J Knee Surg 2002; 15:17. 8. Price AJ, Waite JC, Svard U. Long-term clinical results of the medial Oxford unicompartmental knee arthroplasty. Clin Orthop Relat Res 2005;171. 9. Collier MB, Engh Jr CA, McAuley JP, et al. Factors associated with the loss of thickness of polyethylene tibial bearings after knee arthroplasty. J Bone Joint Surg Am 2007;89:1306. 10. Psychoyios V, Crawford RW, O'Connor JJ, et al. Wear of congruent meniscal bearings in unicompartmental knee arthroplasty: a retrieval study of 16 specimens. J Bone Joint Surg Br 1998;80:976. 11. Hood RW, Wright TM, Burstein AH. Retrieval analysis of total knee prostheses: a method and its application to 48 total condylar prostheses. J Biomed Mater Res 1983; 17:829.