Secondary Logo

Journal Logo

The Failure of All-Polyethylene Patellar Components in Total Knee Replacement

Berend, Michael E., MD; Ritter, Merrill A., MD; Keating, E. Michael, MD; Faris, Philip M., MD; Crites, Brian M., MD

Clinical Orthopaedics and Related Research®: July 2001 - Volume 388 - Issue - p 105-111
SECTION I SYMPOSIUM: Long-Term Followup of Total Knee Arthroplasty

The patellar component of total knee replacement is the most frequent source of nonseptic complications after total knee arthroplasty. The purpose of the current study was to review the radiographic factors associated with loosening of all-polyethylene patellar components. Four thousand five hundred eighty-three cases of Anatomic Graduated Components total knee replacements were performed at the authors’ institution during the past 15 years. Radiographs were reviewed and loosening was defined as global radiolucency or component migration. Four thousand two hundred eighty-seven all-polyethylene patellar components were implanted. There were 180 (4.2%) loose all-polyethylene patellar components. The mean time to loosening was 2.6 years (± 1.75 years). Fifteen (0.3%) patellar components required revision. Five radiographic features were associated with failure. The incidence and mean time of appearance were recorded: (number; incidence; time) (1) Bone-cement radiolucency, n = 174, 96.7%, 1.4 years; (2) increased density, n = 118, 65.6%, 1.8 years; (3) trabecular collapse of the bone, n = 160, 88.9%, 2.3 years; (4) patella fracture and fragmentation, n = 133, 73.9%, 2.5 years; and (5) lateral subluxation of the residual patella bone, n = 146, 81.1%, 2.9 years. Lateral retinacular release was associated with an increased rate of patellar loosening. Loosening of the all-polyethylene patella component is an avascular process strongly associated with lateral retinacular release and for which the patient infrequently requires revision surgery.

From the Center for Hip and Knee Surgery, St Francis Hospital–Mooresville, Mooresville, IN.

Reprint requests to Michael E. Berend, MD, The Center for Hip and Knee Surgery, 1199 Hadley Road, Mooresville, TN 46158.

The patella component in total knee arthroplasty has been reported to be the most frequent source of nonseptic complications, including component loosening, in several series of long-term followup of total knee arthroplasty. 4,7,9,26 Complications associated with the patellofemoral articulation and extensor mechanism 22 include extensor mechanism disruption, avulsion of infrapatellar tendon from the tibial tubercle, lateral patellar subluxation and tilt, fracture of the patella, and failure associated with use of metal-backed patella components. 1,10 The radiographic natural history of the failure process of loosening and associated factors have not yet been described.

The purpose of the current study was to document the radiographic features associated with loosening of an all-polyethylene patella component after total knee arthroplasty with the AGC (Anatomic Graduated Components, Biomet, Warsaw, IN) knee prosthesis. The incidence and order of occurrence of these features is documented.

Back to Top | Article Outline


There were 4913 AGC total knee replacements done between September 1983 and December 1996 at the authors’ institution. Exclusion criteria for this study included infection and revision surgery. Sixty-two knees were excluded because of infection for a rate of 1.2%. There were 58 revision procedures. Two hundred ten knees were lost to followup (4.3%), leaving 4583 total knee replacements performed on 3054 patients. There were 407 (8.9%) deaths during the followup period. The diagnosis was osteoarthritis in 87% of the patients. The average age of patients at the time of surgery was 70.4 years (± 8.9 years; range, 18–93 years). Sixty percent of the patients were women. All patients had femoral, tibial, and patellar components inserted. There were 296 metal-backed patellar components, and 4287 cemented single central peg design all-polyethylene patellar components. All tibial and femoral components were cemented except for 348 femoral components in an early portion of this series. Functional outcome was assessed with postoperative pain scores and Knee Society scores. 11,20

Anteroposterior (AP) radiographs obtained with the patient standing and lateral and patella view radiographs were obtained at 1-to 2-year intervals and at the most recent clinic visit and reviewed for signs of aseptic loosening. Radiographic features associated with failure and the timing of appearance were documented. A patella prosthesis was determined to be loose if there was global radiolucency or it had migrated. Statistical analysis was done using chi square analysis and two sample t tests. A p value of 0.05 was selected to determine statistical significance.

Back to Top | Article Outline


Of the 4287 all polyethylene patellar components, there were 180 cases of patellar component loosening for a loosening rate of 4.2%. The mean time to loosening was 2.6 years (range, 0.3–11 years). The mean followup for the group of patients in whom the patellar component failed was 5.5 years (range, 0.5–12 years). Five distinct radiographic features were associated with process of loosening of the patellar component. These features included: (1) bone-cement radiolucency; (2) increased radiographic density; (3) trabecular bone collapse; (4) fragmentation and fracture of the patella; and (5) lateral subluxation of the patella bone. The incidence of these findings and the timing of their respective appearance are shown in Table 1. Radiolucency appeared significantly sooner than when the components were loose and lateral subluxation was the last feature of the failure process to appear (p < 0.05).



Seventy-three percent (132 of 180) of the patients who had loosening of the component had a lateral retinacular release compared with 59% (2320 of 4104) of patients with a stable implant. Lateral release was associated with an increased likelihood of failure (p = 0.0004). In the patients with loose patellar components, the average knee score was 85 points (range, 44–99 points) and the average pain score was 45 points (range, 10–50 points). Fifteen all-polyethylene patella components were revised or excised for a patellar revision rate of 8.3% for the failed components and a patellar revision rate of 0.3% for the total number of knee components implanted.

The radiographic features and diagrammatic representation of the failure process are shown in Figures 1 to 4. Loosening of the patellar component is not an immediate radiographic finding; rather it is a process that takes approximately 1.5 years to complete.

Fig 1A–B.

Fig 1A–B.

Fig 2A–B.

Fig 2A–B.

Fig 3A–B.

Fig 3A–B.

Fig 4A–B.

Fig 4A–B.

Back to Top | Article Outline


The patella component of a total knee arthroplasty is the most frequent component to show signs of radiographic loosening. The incidence of loosening of the patella implant after total knee arthroplasty has been reported to be approximately 1%. 4,9 The overall incidence of radiographic loosening in the current series was 4.2% (180 of 4287) and the longitudinal process leading to loosening was characterized.

From this investigation, the authors propose a combined biologic and biomechanical mechanism leading to loosening of the all-polyethylene patella component. The majority of the patients in the current series had femoral components implanted with posterior referencing instrumentation. Therefore, the components were placed into relative internal rotation compared with the epicondylar axis. This increased the need for intraoperative lateral retinacular release to balance the soft tissues of the patellofemoral articulation, which increased the risk of subsequent avascular necrosis of the patella and eventual component loosening. The femoral component rotation also would biomechanically predispose the patella component to lateral subluxation and increased lateral patella biomechanical forces that may contribute to loosening of the patella component.

The radiographic features that were documented in the current series follow the biologic pathway of osteonecrosis such as those recognized in the talus, scaphoid, and the femoral head. These features include increased density of the bone, subchondral and trabecular bone resorption, progressive radiolucency, and fracture and fragmentation of the bone. Once loose, there was lateral subluxation of the residual patella bone in 146 of 180 (81%) of the knees caused by the biomechanical forces resulting from relative internal rotation of the femoral component. The process of loosening of the patellar component is not entirely biologic and not fully attributable to abnormal biomechanical forces but has features of both mechanisms contributing to the failure process. The current study may serve as a worst case scenario for future cohorts of patients to compare patellar loosening when the femoral component rotation is determined with the epicondylar axis and no lateral retinacular release is performed.

The design of patella components including the role of three-peg versus one-peg has been suggested to play a role in implant failure. In the current series, patellar component breakage always was associated with other radiographic features of loosening and avascular necrosis. Francke and Lachiewiez, 12 reported peg shearing of three-peg patellar components at the interface of the peg with the component. Using single-peg patellar components, Huang et al 19 reported four cases of failure of the peg junction with the patella component as a cause of failure. Berry and Rand 1 reported an additional seven cases of fracture of the fixation peg in a large series of isolated patellar component revisions. Fatigue fractures at the peg-plate junction were reported in 12 knees by Rosenberg et al 30 in ingrowth-type implants. It remains unclear which is a more durable implant design.

Thickness of the patella has not been shown to have an effect on patella fracture or loosening or both. 29 Biomechanical studies by Hsu et al 17 have shown that reapproximation of the original patella thickness may lead to more normal contact areas and help restore postoperative range of motion and decreased the incidence of patella subluxation. Patella thickness and patella tilt were not measured in the current series because they have been documented previously. 29

Instability and subluxation of the patella have been reported to be related to errors in surgical technique, quadriceps imbalance, and trauma. 5,14,24 Proposed treatment of instability includes correction of proximal and distal malalignment with or without component revision, which has been associated with a serious complication rate including patella tendon rupture. 14 Bindelglass et al 2 reported that subluxation was not associated with decreased pain scores. Briard and Hungerford 5 suggested that treatment of patellar instability should be focused on correcting the underlying cause of a patellofemoral instability and not solely patella component revision. The authors found that lateral instability or subluxation was the final radiographic feature to appear in the process of loosening of the patellar component and not a primary source of patellar complications.

Patella tilt has not been correlated with poor outcome or increased incidence of radiolucency or loosening of the patella component. The incidence of patella tilting after knee replacement surgery varies and has been reported to be as high as 31%. 2,25 Chan and Gill 8 reported that the greater the preoperative patellar tilt, the more likely the patella would tilt laterally postoperatively. Tilt has not been found to be associated with the incidence of lateral release or decrease in pain scores 2 although it may predispose to increase loading at the lateral periphery of the component and wear of metal-backed components. 2 In an earlier series of total knee replacements from the authors’ institution, lateral retinacular release was shown to decrease lateral tilting of the patella. 27 From the current study, one may conclude that proper balancing of the patellofemoral articulation through proper component rotation with efforts to alleviate the need for lateral release may be the best way to prevent tilt, subluxation, and decrease the overall likelihood of patellar loosening.

Fracture of the residual patella beneath the patella component may be a catastrophic complication after total knee arthroplasty if the extensor mechanism is disrupted. The incidence of patella fracture after total knee replacement surgery ranges from 3% to 21%. 31,33 These results compare with the results of the current series in which a fracture was seen in 133 patellas for a rate of 3.1%. The revision of such fractures and reconstruction of the extensor mechanism has been found to have a high complication rate. 1 Fracture type and function of the extensor mechanism has been shown to affect outcome of operative and nonoperative treatment of patients with patella fractures. 13 In general, it is recommended that patients with nondisplaced fractures of the patella can be treated conservatively with immobilization. 15,16 Even patients with displaced fractures without extensor lag can be treated nonoperatively. Keating et al reported on fractures with and without patellar loosening and found that operative treatment with excision had a high complication rate with four of nine patients having deep infection and two patients treated with internal fixation had nonunion. (Keating, EM, Berend, ME, Crites, BM, et al: Patella Fracture Post Total Knee Replacements. Presented at the Knee Society Meeting, Boston, September 2000.)

Much attention has been given to the role of lateral release and its effects on patella complications and the disruption in the patellar blood supply postoperatively. 3,6,21,27,28,32 The relative role of the biomechanical imbalance that leads to the need for lateral release predisposing to patellar loosening in comparison with the vascular disruption that results after lateral retinacular release is unclear. Lateral release may alter the patellar tracking and joint contact characteristics 18 but may lead to other more detrimental effects on the viability of the patellar bone. Lateral release has been associated with decreased patella viability and osteonecrosis. 23,32 Sacrificing or saving the lateral geniculate artery during lateral release does not seem to influence patellofemoral complications after total knee arthroplasty. 28 In an early series, Ritter and Campbell 28 found no increased evidence of osteonecrosis, bone-cement radiolucency, or patella fracture after lateral release. Subsequent studies 28,29 with longer-term followup reported that lateral retinacular release was associated with increased incidence of initial patella fractures within 1 year. Both of these series had a high incidence of lateral retinacular release ranging from 15% to 35%. In the current series, lateral retinacular release was associated with a significantly increased rate of loosening of the patellar component. This correlates with the radiographic features of an avascular process leading to patellar component loosening.

One limitation of the current study as it describes the timing of the failure process is the timing of the postoperative radiographs. The authors standard followup protocol is for radiographs to be obtained at 2 months, 6 months, 1 year, 3 years, 5 years, 7 years, and 10 years. The majority of the radiographic features were found between the radiographs obtained at 6 months and 3 years. Perhaps with more frequent intervals of followup the exact timing could be identified more precisely. The progressive nature of the failure process including progression through the phases of bonecement radiolucency, increased density, collapse, fracture and fragmentation, and finally lateral subluxation was reproducible and predictable with the interval of followup used in the current study.

Pain scores and Knee Society scores 11,20 were not affected adversely by the presence of a loose patella averaging 45 points and 85 points, respectively. Revision or excision of the patellar component was done in only 15 patients (0.3%) in the series when the component completely dislodged from the patellar bone or when there was a complete disruption of the extensor mechanism with significant extensor lag.

Patellar component loosening occurred in 4.2% of AGC total knee replacements in the current series. A loose patellar component did not adversely affect the functional outcome or increase knee pain and infrequently requires revision. Loosening was strongly associated with lateral retinacular release and was present in 73% of the cases. The mechanism of failure was found to have characteristics of an avascular process with bone-cement radiolucency, increased density, collapse, fracture, and eventual lateral subluxation. The radiographic features and timing of this process were described. It is concluded that patellar loosening can be reduced by establishing proper patellofemoral balancing through appropriate femoral component rotation and alleviating the need for lateral release and maintaining patellar viability.

Back to Top | Article Outline


1. Berry DJ, Rand JA: Isolated patellar component revision of total knee arthroplasty. Clin Orthop 286: 110–115, 1993.
2. Bindelglass DF, Cohen JL, Dorr LD: Patellar tilt and subluxation in total knee arthroplasty: Relationship to pain, fixation, and design. Clin Orthop 286: 103–109, 1993.
3. Bonutti PM, Miller BG, Cremens MJ: Intraosseous patellar blood supply after medial parapatellar arthroplasty. Clin Orthop 352: 202–214, 1998.
4. Boyd AD, Ewald FC, Thomas WH, et al: Long-Term complications after total knee arthroplasty with or without resurfacing of the patella. J Bone Joint Surg 75A: 674–681, 1993.
5. Briard J-L, Hungerford DS: Patellofemoral instability in total knee arthroplasty. J Arthroplasty 4 (Suppl): 87–97, 1989.
6. Brick GW, Scott RD: Blood supply to the patella: Significance in total knee arthroplasty, J Arthroplasty 4 (Suppl):75–79, 1989.
7. Brick GW, Scott RD: The patella component of total knee arthroplasty. Clin Orthop 231: 163–178, 1988.
8. Chan KC, Gill GS: Postoperative patellar tilt in total knee arthroplasty. J Arthroplasty 14: 300–303, 1999.
9. Clayton ML, Thirupathi R: Patellar complications after total condylar arthroplasty. Clin Orthop 170: 152–155, 1982.
10. Emerson RH, Higgins LL, Head WC: The AGC total knee prosthesis at average 11 years. J Arthroplasty 15: 418–423, 2000.
11. Ewald FC: The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop 248: 9–12, 1989.
12. Francke EL, Lachiewicz PF: Case report: Failure of a cemented all-polyethylene patellar component of a press-fit condylar total knee arthroplasty. J Arthroplasty 15: 234–237, 2000.
13. Goldberg VM, Figgie HE, Inglis AE, et al: Patellar fracture type and prognosis in condylar total knee arthroplasty. Clin Orthop 236: 115–122, 1988.
14. Grace JN, Rand JA: Patellar instability after total knee arthroplasty. Clin Orthop 237: 184–189, 1988.
15. Grace JN, Sim FH: Fracture of the patella after total knee arthroplasty. Clin Orthop 230: 168–175, 1988.
16. Hozack WJ, Goll SR, Lotke PA, et al: The treatment of patellar fractures after total knee arthroplasty. Clin Orthop 236: 123–127, 1988.
17. Hsu HC, Luo ZP, Rand JA, et al: Influence of patellar thickness on patellar tracking and patellofemoral contact characteristics after total knee arthroplasty. J Arthroplasty 11: 69–80, 1996.
18. Hsu HC, Luo ZP, Rand JA, et al: Influence of lateral release on patellar tracking and patellofemoral contact characteristics after total knee arthroplasty. J Arthroplasty 12: 74–83, 1997.
19. Huang CH, Lee YM, Lai JH, et al: Failure of the all-polyethylene patellar component after total knee arthroplasty. J Arthroplasty 14: 940–944, 1999.
20. Insall JN, Dorr LD, Scott RD, et al: Rationale of the Knee Society clinical rating system. Clin Orthop 248: 13–14, 1989.
21. Kayler DE, Lyttle D: Surgical interruption of patellar blood supply by total knee arthroplasty. Clin Orthop 229: 221–227, 1988.
22. Lynch AF, Rorabeck CH, Bourne RB: Extensor mechanism complications following total knee arthroplasty. J Arthroplasty 2: 135–140, 1987.
23. McMahon MS, Scuderi GR, Glashow JL, et al: Scintigraphic determination of patellar viability after excision of infrapatellar fat pad and/or later retinacular release in total knee arthroplasty. Clin Orthop 260: 10–16, 1990.
24. Merkow RL, Soudry M, Insall JN: Patellar dislocation following total knee replacement. J Bone Joint Surg 67A: 1321–1327, 1985.
25. Ranawat CS: The patellofemoral joint in total condylar knee arthroplasty: Pros and cons based on five to ten-year follow-up observations. Clin Orthop 205: 93–99, 1986.
26. Rand JA: Current concepts review: The patellofemoral joint in total knee arthroplasty. J Bone Joint Surg 76A: 612–620, 1994.
27. Ritter MA, Campbell ED: Postoperative patellar complications with or without lateral release during total knee arthroplasty. Clin Orthop 219: 163–168, 1987.
28. Ritter MA, Herbst BA, Keating EM, et al: Patellofemoral complications following total knee arthroplasty. J Arthroplasty 11: 368–372, 1996.
29. Ritter MA, Pierce MJ, Zhou H, et al: Patellar complications (total knee arthroplasty). Clin Orthop 367: 149–157, 1999.
30. Rosenberg AG, Andriacchi TP, Barden RN, et al: Patellar component failure in cementless total knee arthroplasty. Clin Orthop 236: 106–114, 1988.
31. Scott RD, Turoff N, Ewald FC: Stress fracture of the patella following duopatellar total knee arthroplasty with patellar resurfacing. Clin Orthop 170: 147–151, 1982.
32. Scuderi G, Scharf SC, Meltzer LP, et al: The relationship of lateral releases patella viability in total knee arthroplasty. J Arthroplasty 2: 209–214, 1987.
33. Windsor RE, Scuderi GR, Insall JN: Patellar fractures in total knee arthroplasty. J Arthroplasty (Suppl):4:63–67, 1989.

Section Description

Merrill A. Ritter, MD; and John B. Meding, MD, Guest Editors

© 2001 Lippincott Williams & Wilkins, Inc.