Most long term studies of total knee arthroplasty (TKA) have reported on moderate sized patient cohorts treated with early generation implant designs. Studies have not been of sufficient cohort size or surveillance duration to stratify for and analyze the many factors that influence long term survivorship of TKA. Many modern TKA designs have evolved to share the common features of a condylar design, providing for posterior cruciate ligament retention or sacrifice, and a metal-backed tibial tray with a modular polyethylene insert.
Approximately 418,000 TKAs were estimated to have been performed in the United States in 2003.22 Reported mid- and long-term rates of TKA survivor- ship using revision as an endpoint have been 90% to 98%.6-11,16-19,26,27,30,31,33,35,37 Even with these excellent results, the percentage of failures would lead to a substantial number of patients with TKA failure.
We evaluated the survivorship of 1000 consecutive cemented condylar cruciate-retaining TKAs of a single, modern design with a modular tibial component and a cemented all-polyethylene patellar component. We hypothesized this TKA would provide durable long-term results. We specifically aimed to determine the long-term survivorship; patient characteristics associated with better or worse survivorship; and to examine the occurrence of and time course of complications and reoperations associated with this operative intervention.
MATERIALS AND METHODS
We retrospectively reviewed 1008 consecutive cemented cruciate-retaining TKAs using a single design (press-fit condylar design; DePuy/Johnson and Johnson, Warsaw, IN) performed on 753 patients from January 1, 1987 through August 22, 1989. The design has a modular titanium tibial tray with a keel and an all-polyethylene patellar component with three fixation pegs. All three components were cemented in each knee replacement. Per state statute, all patients must consent to the use of their medical record for research purposes. Eight patients declined such authorization, leaving 1000 knees in 745 patients available for review. Three hundred eighty-four women (509 knees) and 361 men (491 knees) were included. Two hundred fifty-five of the patients had bilateral arthroplasties. The mean age of the patients at the time of surgery was 70.2 years (range, 19-93 years) (Table 1). Ninety-six of the knees were in patients less than 60 years old, 806 were in patients 60 to 80 years old, and 98 were in patients who were older than 80 years of age. The mean body mass index (BMI) of the patients was 29.6 kg/m2 (range, 15.2-56.4 kg/m2) (Table 2). The most common tibial insert thickness used was 8.0 mm (Table 3). The diagnosis leading to TKA was osteoarthritis in 867 knees, inflammatory arthritis in 81 knees (79 of these rheumatoid arthritis), failed previous surgery/osteotomy in 28 knees (27 of these prior upper tibial osteotomy), osteonecrosis in 17 knees, and miscellaneous causes (including Paget's disease, gout, Blount's disease, and benign tumor) in seven knees. Twenty-five of the upper tibial osteotomies were lateral closing wedge (19 fixed with staples, 5 fixed with plate and screws, and 1 without fixation). In two knees the upper tibial osteotomy method and fixation were not able to be determined. Of the 1000 TKAs, 938 (93.8%) were followed for at least 14.5 years or until revision operation, component removal, or death. Sixty-two TKAs (47 patients) were lost to followup. The minimum followup was 14.5 years (mean 15.7 years, range, 14.5-17.9 years). We obtained Institutional Review Board approval before the study.
Patients were followed prospectively and were asked to return to our institution at 1 year, 2 years, 5 years, and every 5 years thereafter for an interview and clinical examination. When this was not possible, patients were asked to answer a standardized mailed questionnaire and to send radiographs. Patients who did not answer the mailed questionnaire were contacted by telephone and interviewed with use of a standardized telephone questionnaire. Patients were specifically queried about whether they had additional surgery on their knee. If they had additional surgery at another institution, a request was made to obtain operative reports and information concerning the indication for the surgery, the surgical findings, and the components that were revised. The most recent followup evaluation was conducted by interview and examination for 282 knees, mailed questionnaire for 487 knees, and telephone questionnaire for 231 knees. Dates of death were obtained by direct communication with the patient's family or from the Social Security network.
Univariate implant survival was estimated using the Kaplan Meier survival method. The association between risk factors and survival was assessed using a Cox proportional hazard survival model, adjusting for correlated data (ie, for two knees in those patients who underwent bilateral TKA).15,34 End points considered in these analyses included component removal for any reason, revision for any reason, and revision for mechanical failure (tibial insert wear, tibial-femoral instability, aseptic loosening, or component fracture). With the small number of aseptic loosening events (n = 9), statistical comparison was not possible for survivorship free of aseptic loosening. For each end point, the risk factors considered included gender, diagnosis (osteoarthritis, inflammatory arthritis, failed upper tibial osteotomy), age at the time of arthroplasty (< 60, 60-80, > 80 years), tibial insert thickness (8 mm, 10-17.5 mm), BMI (≤ 35, > 35 kg/m2), and weight (≤ 250, > 250 lbs). Ninety-five per cent confidence intervals (CI) were calculated for the survivorship estimates. A Cox proportional hazard survival model was also used to assess a multiple variable model for the end point of component removal for any reason; all other end points had too few events. This model was developed by considering any risk factors with a univariate significance of p < 0.50 and using the forward and backward selection methods to retain those factors associated with component removal for any reason. The alpha-level was set at 0.05 for statistical significance.
By 17 years after TKA, 411 patients (562 knees) had died and 43 patients (45 knees) had their knee components revised or removed. Two hundred forty four patients (331 knees) were alive, and had not had their knee components revised or removed. Seventy eight knees were known to have had an additional procedure (Table 4). Thirteen of these procedures were closed manipulation of the knee for limited motion, and the remaining 65 knees had an open reoperation. Twenty of the 65 reoperations did not involve component removal or revision (for example, open reduction and internal fixation of a periprosthetic femur fracture) (Table 5). These reoperations included five cases of irrigation and débridement for superficial (four) or deep (one) infection.
Forty-five of the 65 reoperations involved component removal (18 knees), or revision (27 knees). Of the 18 knees treated with component removal without simultaneous implant revision, 16 had component removal performed due to deep infection (Table 6). The 27 knees treated with revision were treated for aseptic loosening (nine knees), tibial insert wear (seven knees), tibialfemoral instability (three knees), treatment of periprosthetic patella fracture with associated patellar component loosening (three knees), pain of unknown etiology (three knees), and patellar-femoral instability or tibial component fracture (one knee each).
Eighteen of the 1000 knees were diagnosed with a deep infection during the followup period. Three of these deep infections occurred within the first 90 days postoperatively, and a total of five occurred within the first postoperative year. Almost ⅓ of all reoperations (21 of 65) were related to infection and just over ⅓ of the components removed or revised (16 of 45), were removed for infection. Sixteen of the 30 reoperations (53%) and 12 of 14 component removals performed in the first 5 years postoperatively were for infection. There were four knees diagnosed with deep infection among the 79 with rheumatoid arthritis, compared with 13 knees of the 867 with osteoarthritis.
Thirty-six periprosthetic fractures occurred during the study period. Fourteen of the 65 reoperations (22%) were related to periprosthetic fracture with seven femoral fractures and seven patellar fractures. Four of these operations involved component removal or revision (three for patella fracture and one for femur fracture), and 10 involved re- operations without component removal or revision (four for patella fracture and six for femur fracture). Twenty- two periprosthetic fractures (13 patellar, seven femoral, and two tibial fractures) were treated nonoperatively. Of the 36 periprosthetic fractures the majority of patellar factures (15 of 20) occurred within the first 5 years postoperatively, while the majority of femur fractures (8 of 14) occurred after 10 years (Fig 1).
Approximately ⅓ of all components removed were removed for infection (16 of 45), ⅓ for aseptic loosening and tibial polyethylene wear (16 of 45), and ⅓ for other reasons (periprosthetic fracture, instability, pain of unknown etiology, peripheral ischemia). Notably, less than ⅓ of the reoperations (20 of 65) were for mechanical failure of the prosthesis (aseptic loosening, tibial insert wear, tibial-femoral instability, and tibial component fracture).
For the 1000 knees, the 15 year rates of survivorship free of component removal for any reason, free of revision for any reason, free of revision for mechanical failure and free of revision for aseptic loosening were 93.7%, 95.9 %, 97.0%, and 98.8%, respectively (Table 7). We observed an association between age at the time of index TKA and survivorship free of component removal for any reason (p < 0.001), revision for any reason (p < 0.001), and revision for mechanical failure (p < 0.001). Patients younger than 60 years had the highest failure rates; patients 80 years or older had the lowest failure rates. There were no revisions for any reason among patients more than 80 years of age (Table 8).
The diagnosis leading to primary TKA was trended toward an association with prosthesis survival (p = 0.10). The subset of patients who had a prior upper tibial osteotomy had a lower (p = 0.02) survival, relative to those patients with a diagnosis of osteoarthrosis. Among patients with rheumatoid arthritis, there was excellent survivorship at 15 years, with a 93.1% (95% CI, 85.8-100.0%) survivorship for revision for any reason and no revisions for aseptic loosening at 15 years.
Gender had no association with respect to survivorship for any survival endpoint (Table 9).
Comparisons of the 574 knees with a tibial polyethylene insert thickness of 8 mm to the 426 knees with a thickness greater than 8 mm did not reveal any associations with any of the survival end points. However, of the 20 cases of revision for mechanical failure, 14 were in patients with an 8.0-mm insert, six failures were in patients with a 10.0-mm insert, and there were no cases of mechanical failure in patients with a 12.5 mm or greater thickness insert. Of the seven patients who had revision knee arthroplasty for tibial insert wear, five were younger than age 60 at the time of index knee arthroplasty. In all five of these cases the tibial insert thickness was 8.0 mm.
Neither BMI, dichotomized as ≤ 35 compared with > 35, nor patient weight, dichotomized as ≤ 250 lbs compared with > 250 lbs, was associated with any of the survivorship endpoints.
The result of a multiple variable analysis of risk factors associated with the end point of component removal for any reason resulted in a model containing only age. This was the only survival end point considered for multivariate analysis because the remaining end points had relatively few events.
Most studies of TKA survivorship have included early generation implant designs, different prosthetic component combinations, or reported on moderate-sized cohorts. Our study has examined 1000 patients treated consecutively with the same prosthesis and component combination and fixation at a mean 15.7 years of followup.
A limitation of the study is its lack of radiographic review and therefore the failure rates are minimum numbers representing only patients who have come to reoperation or revision. Some patients in this cohort likely have arthroplasty problems, clinical or radiographic, that have not led to revision or reoperation. However, the study included a relatively large number of consecutive arthroplasties included, few patients lost to followup, and relatively long followup. We chose to use the endpoints of revision and reoperation because they represent the most unambiguous endpoints of failure.
The survivorship of this cemented cruciate retaining implant was excellent, with survivorship free of revision for any reason 95.9% at 15 years, and survivorship free of revision for aseptic loosening 98.8% at 15 years. These data compare favorably with other reported rates of survivorship using revision as the endpoint for posterior cruciate-retaining knee prostheses at 9 to 15 years of 90% to 98%,6,8,9,12,16,19,26,27,30,31,35,37 and for cruciate- substituting knee prostheses at 10 to 16 years of 92% to 98%.7,10,17,18,33
Patient age at the time of surgery was the single most important determinant of implant survivorship. There were no failures related to mechanical failure in patients greater than 80 years of age. In contrast, in the subgroup of patients less than 60 years of age at the time of knee arthroplasty, the 15-year survivorship free of revision for mechanical failure was 88.2% (Table 8).
Implant survival for the subgroup of patients with a prior upper tibial osteotomy was poorer than for the overall group in univariate analysis, and of borderline significance in multivariate analysis. There have been conflicting reports over whether prior upper tibial osteotomy has an adverse effect on the subsequent results of TKA. Some studies have reported there are adverse clinical effects on subsequent TKA,13,21,24 and others have reported the results are similar to patients with a TKA without prior upper tibial osteotomy.1,12,20,32 Most of these studies have focused on postoperative clinical outcomes such as pain, satisfaction scores, and nonrevision complications and do not comment on survival analysis. Parvizi et al24 did report a 15-year survival rate free of revision of 89% in their series of 166 TKAs in 118 patients who had had an upper tibial osteotomy before their TKA. Thirty-four of the patients in their series had a TKA in their contralateral lower extremity without prior upper tibial osteotomy. When comparing the results of TKA both with and without prior upper tibial osteotomy in these 34 patients, they found slightly inferior results in the group having had a prior upper tibial osteotomy and were able to demonstrate with multivariate analysis part of this difference was due to these patients being younger, heavier, more active men. Our 15-year survival rate free of revision in these patients was 86.4% (95% CI, 73.1-100%). When accounting for age, gender, and patient weight in a multivariate model, prior tibial osteotomy was still a borderline risk factor for failure. This does not imply we do not believe in upper tibial osteotomy for management of osteoarthritis of the knee in appropriate candidates, particularly in light of our TKA survival data for younger patients. Our data do suggest prior upper tibial osteotomy may have an impact on the survival of a potential future TKA, or patients who are candidates for upper tibial osteotomy may also be predisposed to failure of TKA, perhaps by virtue of activity or other factors not evaluated in our multivariate analysis.
More than ⅓ of the implant failures in our series were related to deep infection. The crude rate of deep infection of 1.8% in this series is within the range of reported deep infection rates of approximately 1.0% to 3.4% in series of TKAs ranging from 146 to 6489 knees.4,11,16,18,19,25,30,35,37 We found as other causes of TKA failure have been eliminated or reduced to a greater degree, infection has become a proportionately greater cause of failure of modern TKA.
Periprosthetic fractures were the most common late complication in our series and this likely reflects the aging of this cohort. Ten of 20 reoperations without component removal were for treatment of periprosthetic fractures. Our prevalence rates of femoral (1.4%) and patellar (2.0%) periprosthetic fractures are in line with reported prevalence rates of 0.3% to 2.5% for femoral and 0.2% to 21% patellar (resurfaced) periprosthetic fractures.2,3,5,23,28,29 The rising incidence of femoral periprosthetic fractures more than 10 years after TKA in our series suggests this may be an increasing problem as numbers of patients with their TKA in situ for long time periods continues to increase.
Patient gender, BMI and weight, and tibial insert thickness did not correlate as risk factors for failure. However, for some endpoints and at all time periods, men had a poorer survivorship than women. It is possible with a larger sample size a small but important difference in survivorship between men and women could be demonstrated. Some other studies have reported gender is not a risk factor for TKA failure10,11,37; however, Rand et al,27 in a series of more than 11,000 TKAs (595 with greater than or equal to 15 years followup) found, on multivariate analysis of their series, men had poorer survivorship. A difference between their study and our series that could have led to different results is their study included multiple prostheses, including some older designs. Patient weight and BMI also are factors with conflicting reports on their influence on TKA outcome. Jared et al reported obesity of any degree (BMI > 30) had a negative effect on TKA outcome.14 Other reports have found BMI and patient weight were not risk factors for TKA failure.10,37 It is possible these findings can be reconciled by considering the effect of obesity on TKA survivorship may depend on factors such as patient activity, the presence of any TKA malalignment, presence of comorbid medical conditions (diabetes), or contribution to perioperative complications (medial collateral ligament rupture, wound healing).36
Although tibial insert thickness was not correlated as a risk factor for TKA failure in our series, there were no mechanical failures in knee arthroplasties with a 12.5 mm or greater thickness tibial insert. Five of seven failures for tibial insert wear were in arthroplasties with an 8.0 mm thick insert. Another series of 235 TKAs of similar design reported seven of nine revisions for tibial insert wear were in knee arthroplasties with tibial inserts of 8.0 mm, and there were no revisions for etiologies not related to patellar failures where the tibial insert was 12.5 mm or greater thickness.30
Notably, for a modern cemented condylar TKA only about ⅓ of reoperations and revisions were for mechanical implant failure. This suggests, in addition to ongoing efforts to minimize long term reoperation rates after TKA for mechanical failure, efforts should also concentrate on the prevention and effective treatment of prosthetic infection and periprosthetic fracture. Together these two complications made up the majority of reoperations in the first 17 years after modern TKA.
The authors thank Youlonda Loechler for assistance with data retrieval and followup questionnaire, and Karen Fasbender for her assistance in preparation of this manuscript.
1. Amendola A, Rorabeck CH, Bourne RB, Apyan PM. Total knee arthroplasty following high tibial osteotomy for osteoarthritis. J Arthroplasty
2. Ayers DC. Supracondylar fracture of the distal femur proximal to a total knee replacement. Instruct Course Lect
3. Berry DJ. Epidemiology-hip and knee. Orthop Clin North Am
4. Blom AW, Brown J, Taylor AH, Pattison G, Whitehouse S, Bannister GC. Infection after total knee arthroplasty. J Bone Joint Surg Br
5. Bourne RB. Fractures of the patella after total knee replacement. Orthop Clin North Am
6. Buehler KO, Venn-Watson E, D'Lima DD, Colwell CW. The press- fit condylar total knee system: 8 to 10 year results with a posterior cruciate-retaining design. J Arthroplasty
7. Colizza WA, Insall JN, Scuderi GR. The posterior stabilized total knee prosthesis: Assessment of polyethylene damage and osteolysis after ten year minimum follow-up. J Bone Joint Surg Am
. 1995;77: 1713-1720.
8. Dixon MC, Brown RR, Parsch DP, Scott RD. Modular fixed- bearing total knee arthroplasty with retention of the posterior cruciate ligament. J Bone Joint Surg Am
9. Ewald FC, Wright J, Poss R, Thomas WH, Mason MD, Sledge CB. Kinematic total knee arthroplasty: a 10 to 14 year prospective follow-up review. J Arthroplasty
10. Font-Rodriguez DE, Scuderi GR, Insall JN. Survivorship of cemented total knee arthroplasty. Clin Orthop Relat Res
. 1997;345: 79-86.
11. Gill GS, Joshi AB. Long term results of kinematic condylar knee replacement. J Bone Joint Surg Br
12. Haddad FS, Bentley G. Total knee arthroplasty after high tibial osteotomy: a medium term review. J Arthroplasty
. 2000;15: 597-603.
13. Jackson M, Sarangi PP, Newman JH. Revision total knee arthroplasty: comparison of outcome following primary proximal tibial osteotomy or unicompartmental arthroplasty. J Arthroplasty
. 1994; 9:539-542.
14. Jared RH, Foran J, Mont MA, Etienne G, Jones L, Hungerford DS. The outcome of total knee arthroplasty in obese patients. J Bone Joint Surg Am
15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observation. J Am Stat Assoc
16. Keating EM, Meding JB, Faris PM, Ritter MA. Long-term follow- up of non-modular total knee replacements. Clin Orthop Relat Res
17. Kelly MA, Clarke HD. Long-term results of posterior cruciate- substituting total knee arthroplasty. Clin Orthop Relat Res
. 2002; 404:51-57.
18. Li PL, Zamora J, Bentley G. The results at ten years of the Insall- Burstein II total knee replacement. Clinical, radiographic and survivorship studies. J Bone Joint Surg Br
19. Malkani AL, Rand JA, Bryan RS, Wallrichs SL. Total knee arthroplasty with the kinematic condylar prosthesis. J Bone Joint Surg Am
20. Meding JB, Keating EM, Ritter MA, Faris PM. Total knee arthroplasty after high tibial osteotomy. Clin Orthop Relat Res
. 2000;375: 175-184.
21. Mont MA, Antonaides S, Krackow KA, Hungerford DS. Total knee arthroplasty after failed high tibial osteotomy: a comparison with a matched group. Clin Orthop Relat Res
22. National Center for Health Statistics. Inpatient procedures. Available at: www.cdc.gov/nchs/fastats/insurg.htm
23. Ortiguera CJ, Berry DJ. Patellar fracture after total knee arthroplasty. J Bone Joint Surg Am
24. Parvizi J, Hanssen AD, Spangehl MJ. Total knee arthroplasty following proximal tibial osteotomy: Risk factors for failure. J Bone Joint Surg Am
25. Peersman G, Laskin R, Davis J, Peterson M. Infection in total knee replacement: A retrospective review of 6489 total knee replacements. Clin Orthop Relat Res
26. Rand JA, Ilstrup DM. Survivorship analysis of total knee arthroplasty: cumulative rates of survival of 9200 total knee arthroplasties. J Bone Joint Surg Am
27. Rand JA, Trousdale RT, Ilstrup DM, Harmsen WS. Factors affecting the durability of total knee prostheses. J Bone Joint Surg Am
28. Rolston LR, Christ DJ, Halpern A, O'Connor PL, Ryan TG, Uggen WM. Treatment of supracondylar fractures of the femur proximal to a total knee arthroplasty: a report of four cases. J Bone Joint Surg Am
29. Rorabeck CH, Angliss RD, Lewis PL. Fractures of the femur, tibia, and patella total knee arthroplasty: decision making and principles of management. Instruct Course Lect
30. Schai PA, Thornhill TS, Scott RD. Total knee arthroplasty with the PFC system: results at a minimum of ten years and survivorship analysis. J Bone Joint Surg Br
31. Sextro GS, Berry DJ, Rand JA. Total knee arthroplasty using cruciate-retaining kinematic condylar prosthesis. Clin Orthop Relat Res
32. Staeheli JW, Cass JR, Morrey BF. Condylar total knee arthroplasty after failed proximal tibial osteotomy. J Bone Joint Surg Am
. 1987; 69:28-31.
33. Thadani PJ, Vince KG, Ortaaslau SG, Blackburn DC, Cudiamat CV. Ten- to twelve-year follow-up of the Insall-Burstein I total knee prosthesis. Clin Orthop Relat Res
34. Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model (Statistics for Biology and Health)
. New York NY: Springer; 2001.
35. Weir DJ, Moran CG, Pinder IM. Kinematic condylar total knee arthroplasty. J Bone Joint Surg Br
36. Winiarsky R, Barth P, Lotke P. Total knee arthroplasty in morbidly obese patients. J Bone Joint Surg Am
37. Wright RJ, Sledge CB, Poss R, Ewald FC, Walsh ME, Lingard EA. Patient reported outcome and survivorship after kinemax total knee arthroplasty. J Bone Joint Surg Am