The most frequent cause of failure in early total knee arthroplasty designs was prosthetic loosening.11,19 Cementless fixation of total knee arthroplasty was devised as an alternative to cement fixation in an attempt to diminish the rate of loosening.10 However, subsequent cemented total knee arthroplasty designs5,17 have not shown the high loosening rates reported for early designs. Ranawat et al17 reported 94.6% clinical survival at 15 years for the cemented total condylar prosthesis.
The aim of this investigation was to compare the results of cement and cementless fixation using the same implant type in a matched group of patients to determine whether there are differences in clinical results and durability of fixation at long term followup.
MATERIALS AND METHODS
A prospective study of cementless fixation of total knee arthroplasty using a finned tibial component, Press Fit Condylar (PFC), (Johnson and Johnson Orthopedics, Braintree, MA), was begun in 1985. The indications used for cementless fixation were physiologically young age, good quality bone stock, initially stable implant fixation, bone prosthesis gaps of 1 mm or less at surgery, and patient ability to cooperative with protected weightbearing for 1 month after surgery. Patients not meeting the above criteria were treated with cement fixation. All patients were operated on between 1985 and 1987. All patients with cementless fixation using the Press Fit Condylar prosthesis were entered into the study. Retrospectively, these patients were matched with a group of patients whose total knee arthroplasty of the same design was fixed with cement. Patients were matched as closely as possible by age, gender, diagnosis, and date of surgery. Sixty-one knees with cementless fixation were entered into the study. Of these, two patients (two knees) were lost to followup. Therefore, 59 knees in 50 patients with cementless fixation were compared with 59 knees in 52 patients with cement fixation.
The early results of these matched pair groups have been reported previously at a mean of 2.8 years after surgery.18 Since the previous review, five patients (eight total knee arthroplasties) in the cement group died and four patients (four total knee arthroplasties) in the cementless group died. This left 51 total knee arthroplasties in 47 patients with cement fixation and 55 total knee arthroplasties in 46 patients with cementless fixation. The age of the patients was 54 ± 8 years (mean ± standard deviation) in the cementless group and 65 ± 5 years in the cemented group (p < 0.0001). The patients' height was 170 ± 10 cm in the cementless group and 168 ± 13 cm in the cemented group (not significant). The patients' weight was 84 ± 14 kg in the cementless group (not significant) and 78 ± 13 kg in the cemented group (not significant).
In the cementless group, there were 23 males and 23 females. The right knee was involved in 30 patients and the left knee was involved in 25 patients. The diagnoses were osteoarthrosis in 42 knees, rheumatoid arthritis in nine knees, post-traumatic arthritis in three knees, and ankylosing spondylitis in one knee. Nineteen patients had bilateral total knee arthroplasties. Twenty knees had surgery before total knee arthroplasty consisting of arthroscopy in eight, medial meniscectomy in six, upper tibial osteotomy in three, debridement in three, extensor mechanism realignment in three, lateral meniscectomy in one, synovectomy in one, open reduction and internal fixation of fracture in two, patellar chondroplasty in one, and excision of osteochondritis dissecans in one. Other lower extremity joint diseases involved the ipsilateral hip in one patient, contralateral knee in 38 patients, ipsilateral ankle in two patients, and both ankles in three patients. The arthroplasties were performed by or under the supervision of eight surgeons, with 46 of the knee arthroplasties being performed by two surgeons. All knees had patellar resurfacing with 53 having a metal backed patellar component and two having a cemented polyethylene patella. Followup evaluation was performed by examination at the authors' clinic in 25 patients (32 knees), by standard telephone questionnaire in 16 patients (18 knees), and by standard letter questionnaire in five patients (five knees). The duration between surgery and followup evaluation was 10.2 ± 0.6 years (mean ± standard deviation), (range, 9.4-11.3 years).
In the cemented fixation group there were 23 males and 24 females. The right knee was involved in 24 patients and the left knee was involved in 27 patients. The diagnoses were osteoarthrosis in 42 knees, rheumatoid arthritis in six knees, ancient sepsis in two knees, and osteonecrosis in one knee. Seventeen patients had bilateral total knee arthroplasties. Fourteen knees had surgery before total knee arthroplasty consisting of arthroscopy in five, upper tibial osteotomy in five, medial meniscectomy in three, and lateral meniscectomy in one. Other lower extremity joint diseases involved the ipsilateral hip in two patients, contralateral hip in two, both hips in two, contralateral knee in 37, and both ankles in one patient. The arthroplasties were performed by or under the supervision of 13 surgeons with 29 of the arthroplasties performed by two surgeons. All knees had patellar resurfacing with 23 having a metal backed patellar component and 28 having an all polyethylene patellar component. Followup was by examination at the authors' clinic in 17 patients (28 knees), by telephone inquiry in 15 patients (18 knees), and by detailed questionnaire in five patients (five knees). The duration between surgery and latest followup was 10.0 ± 0.6 years (mean ± standard deviation) (range, 8.7-11.4 years).
Knee scores were calculated using the Knee Society scoring system.12 Implant positioning and radiolucent lines were evaluated according to the Knee Society system.8
Statistical analysis was performed using the two-sample t test or Wilcoxon's rank sum test for comparison of continuous variables. Changes in the continuous variables were assessed using the paired t test or signed ranks test. Association of pairs of continuous variables were estimated with Pearson correlation coefficient or with Spearman rank correlation coefficient. Comparisons of proportions between groups were made with the chi square test or Fisher's exact test.
The average preoperative knee score for the knee replacements that had not been revised was 32 ± 17.9 (mean ± standard deviation) for the cemented fixation group compared with 33 ± 18.9 (mean ± standard deviation) for the cementless fixation group (not significant). At latest followup the knee score was 92.4 ± 8.2 (mean ± standard deviation) for the cemented group compared with 87.8 ± 13.8 (mean ± standard deviation) for the cementless group (not significant). The preoperative Knee Society function score for knee replacements that had not been revised was 45.4 ± 22.4 (mean ± standard deviation) for the cement group and 52.3 ± 20.7 (mean ± standard deviation) for the cementless group (not significant). At latest followup the function score was 72.4 ± 25.9 (mean ± standard deviation) for the cemented group and 66.3 ± 29.1 (mean ± standard deviation) for the cementless group (not significant).
Implant Position and Alignment
The preoperative axial alignment (tibial femoral angle) was 1° ± 8° valgus (mean ± standard deviation) in the cementless fixation knee replacements compared with 0° ± 7° (mean ± standard deviation), (not significant) in the cemented fixation knee replacements. At last evaluation, the axial alignment was 5° ± 3° valgus (mean ± standard deviation) in the cementless knee replacements compared with 6° ± 3° (mean ± standard deviation), (not significant) in the cemented knee replacements.
The coronal position of the femoral component was 7° ± 2° (mean ± standard deviation) valgus in the cementless knee replacements compared with 7° ± 2° (mean ± standard deviation) valgus (not significant) in the cemented knee replacements. The coronal position of the tibial component was varus 2° ± 2° (mean ± standard deviation) in the cementless knee replacements compared with varus 2° ± 2° (mean ± standard deviation), (not significant) in the cemented knee replacements.
The sagittal position of the femoral component was rotated posteriorly 3° ± 4° (mean ± standard deviation) in the cementless knee replacements compared with 2° ± 2° (mean ± standard deviation), (not significant) in the cemented knee replacements. The sagittal position of the tibial component was sloped posteriorly 5° ± 3° (mean ± standard deviation) in the cementless knee replacements compared with 5° ± 4° (mean ± standard deviation) in the cemented knee replacements.
Knee extension improved from a flexion contracture of 6.4° ± 7° (mean ± standard deviation) in the cementless fixation and 7.3° ± 7.8° (mean ± standard deviation) in the cemented fixation group preoperatively to a flexion contracture of 0.4° ± 2.1° (mean ± standard deviation) in the cementless group and 1.3° ± 3.2° (mean ± standard deviation), (p = 0.04) in the cemented group at last evaluation. Knee flexion preoperatively was 106° ± 20° (mean ± standard deviation) in the cementless group compared with 106° ± 16° (mean ± standard deviation) in the cemented group. At last evaluation, knee flexion was 100° ± 19° (mean ± standard deviation) in the cementless group compared with 102° ± 18° (mean ± standard deviation) (not significant) in the cemented group.
Metal Backed Patellar Failure
There was a high failure rate of the metal backed patella in both groups of patients. Nineteen of 53 metal backed patellas failed in the cementless fixation group. The estimated survival free of metal backed patellar failure at 5 years was 76.5% and at 10 years was 64.8%. Nine of 23 metal backed patellas failed in the cemented fixation group. The estimated survival free of metal backed patellar failure at 5 years followup was 96.2% and at 10 years followup was 63.5%. At 5 years there was a significant difference between the survival estimates of the cementless metal backed patella compared with the cemented metal backed patella (p < 0.001); but at 10 years there was no significant difference. There were no revisions of cemented all polyethylene patellas. At the time of patellar revision 17 tibial polyethylene inserts were exchanged because of signs of wear in the cementless group and eight tibial polyethylene inserts were exchanged because of wear in the cemented group. Components were checked intraoperatively for loosening and all were deemed well fixed. Some abrasion of the trochlear surface was observed on the femoral components at the time of revision in 16 of the cementless group and in seven of the cemented group; however, none of the femoral components were thought by the operating surgeon to be damaged severely enough to warrant revision. The patients who had a patellar revision with or without tibial insert exchange were followed up and are included in the survival analysis. Since revision of the patella there have been no additional revision surgeries in this group of patients and all knee replacements were deemed well fixed at latest followup.
In the cementless fixation group there were nine revisions of the metallic tibial component or femoral components. In five of these revision surgeries the femoral component and tibial component were revised. Four of these were because of aseptic loosening of both components and one of these was because of osteolysis around both components. There were four revision surgeries involving the tibial component for loosening and the patellar component for wear. One additional patient recently was advised to have revision surgery for severe tibial osteolysis as seen on the radiographs, and that patient's knee replacement was classified as a failure. There were no revision surgeries for infection in the cementless group.
In the cemented group there have been three revision surgeries. Two patients had revision surgery for aseptic loosening of the tibial and femoral components. One cemented knee replacement had revision surgery because of infection in the knee.
Radiolucent lines were analyzed by zone8 and comparisons were made between the cementless and cemented knee replacements. On the last radiograph there were significantly more radiolucent lines for the cementless than cemented knee replacements in tibial Areas 1 (p < 0.001), 2 (p < 0.001), 3 (p < 0.001), 4 (p < 0.001), 8 (p < 0.001), 11 (p < 0.001), and 12 (p < 0.001), (tibial plateau), but not for tibial Areas 5 through 7 or 9 and 10 (stem). At last evaluation radiolucent lines were adjacent to the tibial component of 18 (41%) of the cemented compared with 31 (86%) of the cementless knees (p < 0.001). At the last evaluation there was no significant difference in radiolucent lines in Areas 1 through 7 of the femoral component between the cement and the cementless groups. However, there were radiolucent lines in at least one zone in 18 (50%) of the cementless femurs compared with radiolucent lines in at least one zone in nine (20%) of the cemented femurs (p < 0.005).
At last evaluation seven (19%) of the cementless tibias and three (8%) of the cementless femurs had evidence of bead shedding.
Radiographic loosening was defined as a complete radiolucent line greater than 1 mm in thickness in all zones or a change in implant position. Using these criteria, two (6%) femoral components and 10 (28%) tibial components in the cementless group were considered radiographically loose, whereas in the cemented group no femoral component and one (2%) tibial component was radiographically loose.
When revision of the femoral or tibial component for aseptic loosening or lysis alone was considered as a failure, the 5-year estimated survival for cementless total knee arthroplasty was 100% whereas at 10 years it was 87.6%. The 5-year estimated survival for cemented total knee arthroplasty was 98.3% at 5 years and 96.1% at 10 years. The difference at 10 years was statistically significant (p = 0.05).
When failure was defined as femoral or tibial component revision for aseptic loosening or osteolysis or radiographic loosening (mechanical failure), the 10-year survival for cementless total knee arthroplasty was 72.7% and cemented total knee arthroplasty was 94.2% (p = 0.0008).
Complications in the cementless group consisted of one deep infection that was treated with debridement and component retention, one case of fibrous ankylosis (treated with open lysis of adhesions), and one case of deep venous thrombosis (treated with anticoagulation).
In the cemented group, there were two deep venous thromboses (both treated with anticoagulation), two deep infections (one treated with component removal, debridement, intravenous antibiotics, and reimplantation and one treated with debridement and component retention and intravenous antibiotics). There was one stress fracture of the patella that was treated with the patient in a functional brace for 6 weeks.
Cementless fixation of total knee arthroplasty was developed because of concern about the long term durability of cemented component fixation. Hungerford et al10 suggested that cementless fixation would be more durable than cement. Landon et al13 stated that the future direction of knee reconstructive surgery would be toward cementless total knee arthroplasty which would be able to provide younger, more active patients a permanent and effective cure for disabling arthritis.
Although loosening of cemented tibial components was the most frequent mode of failure of early designs,18 the long term durability of the cemented condylar prosthesis of certain designs has been excellent. Ranawat et al17 reported 94.6% clinical survival at 15 years using a cemented posterior stabilized total knee prosthesis, and Colizza et al5 reported survival of 96.4% for the metal backed posterior stabilized total knee prosthesis at 11 years. Whiteside24 has reported similar good results in cementless total knee arthroplasty with survivorship of 94% at 10 years using the Ortholoc component.24 The orthopaedic surgeon must choose between the two forms of fixation, yet there are few studies directly comparing the results of the two techniques, and none that compare the two at long term followup.
The matched pair comparison method used in this study provides a means of retrospectively comparing the results of the two techniques. This method of comparison has strengths and weaknesses. All surgeries in both groups of patients were performed at the same institution by the same surgeons. Indications for cementless fixation were those that still advocated (younger age, good bone quality, ability to follow a limited weightbearing program20,21). All the prostheses were of one design, eliminating differences in implant design features. The patients were matched for weight, gender, and diagnosis thus eliminating these factors as variables. The patients in the current review could not be matched completely for age. This weakness in the study design is attributable to the selection of young patients during the period of study for cementless fixation because these were the patients thought to have the most potential for benefit from cementless fixation.
There was a high failure rate of the metal backed patellas in both groups. This complicates analysis as the failure of cementless metal backed patellas might have an effect on survival of the femoral and metallic tibial components. To minimize the influence of metal backed patellar failure on the survival analysis, the analysis was done without metal backed patellar failure as a revision. None of the knees with revised patellas in either the cement or cementless group subsequently had to have a revision of a femoral or metallic tibial component.
The early clinical results of cemented and cementless total knee arthroplasty in younger patients were similar.16,22,23 Early studies comparing cement with cementless fixation did not find any significant differences in clinical results. Dodd et al7 reported on 18 matched pairs of Porous Coated Anatomic prostheses at an average of 5 years followup. They found that the clinical and functional performance of each type of fixation were comparable and unrelated to the type of fixation method. Rosenberg et al21 compared 123 cementless Miller-Galante total knee prostheses to 116 cemented prostheses. The patients were followed up between 3 and 6 years. No cemented knee replacements failed because of loosening but three cementless knee replacements failed because of tibial loosening in two and pain of undetermined origin in one. There were no significant differences observed for pain, limp, or support scores. Recently, researchers (nonpublished data, Rorabeck CH, Neill HL: Total knee replacement: Should it be cemented, cementless, or hybrid? Presented at the annual meeting of the American Academy of Orthopaedic Surgeons, San Francisco, February 1997) reported on 555 knees using the same prosthesis implanted with either all cement, cementless, or hybrid. They described an unequivocal advantage of using cement as a preferred method of fixation in total knee replacement. The Knee Society clinical rating score at final followup for the hybrid group was 155; the cementless group, 160; and the cemented group, 173.
The group of patients studied in this report has been reported on previously.18 At an average of 2.8 years after surgery there were no significant differences in knee scores between the two groups, leading to the conclusion that cemented or cementless fixation of this prosthesis provided equivalent early clinical results. This same group of patients was reviewed to see whether early equal results would be maintained with time. This has not been the case. There was a significant difference in the 10-year survival analysis of the two groups with a higher revision rate for aseptic loosening in the cementless group and also a higher rate of mechanical failure. The high incidence of radiolucent lines seen in the cementless tibial components is also noteworthy. For patients with surviving prostheses there were slightly poorer pain and function scores in the cementless group (but not statistically significant).
The current study is consistent with other recently reported long term studies documenting the durability of cemented total knee arthroplasty. Gill et al9 reviewed 68 cemented total knee arthroplasties at 9.9 years followup. Only two total knee arthroplasties required revision for loose components. Diduch et al6 reported on 103 total knee arthroplasties in patients 55 years of age or younger with osteoarthrosis diagnosed. The average annual rate of failure of the prosthesis was 0.3% and the overall rate of survival of the prosthesis was 94% at 18 years. To date there is no long term (>10 years) report of cementless total knee arthroplasty in young patients.
The results of this paper confirm the high failure rate reported for metal backed patellas in other series1-4,14,15 and provides information quantitating the risks of this problem at 10 years after surgery. There were no patellar revisions in the all polyethylene cemented patellas. By 10 years the rate of metal backed patellar failure for this design was greater than 35%.
The present study showed better durability of femoral and tibial fixation in the cohort of patients with cemented total knee arthroplasty fixation compared with a cohort of patients with cementless total knee arthroplasty fixation of the same design. The results do not suggest any advantage of cementless fixation over cemented fixation for total knee arthroplasty of this design.
1. Andersen HN, Ernst C, Frandsen PA: Polyethylene failure of metal-backed patellar components. 111 AGC total knees followed for 7-22 months. Acta Orthop Scand 62:1-3, 1991
2. Baech J, Kofoed H: Failure of metal-backed patellar arthroplasty. 47 AGC total knees followed for at least 1 year. Acta Orthop Scand 62:166-168, 1991.
3. Bayley JC, Scott RD, Ewald FC, Holmes GB: Failure of the metal backed patellar component after total knee replacement. J Bone Joint Surg 70A:668-674, 1988.
4. Berry DJ, Rand JA: Isolated patellar component revision of total knee arthroplasty. Clin Orthop 286:110-115, 1993.
5. Colizza WA, Insall JN, Scuderi GR: The posterior stabilized total knee prosthesis. Assessment of polyethylene damage and osteolysis after a ten-year minimum follow-up. J Bone Joint Surg 77A:1713-1720, 1995.
6. Diduch DR, Insall JN, Scott WN, Scuderi GR, Font-Rodriguez D: Total knee replacement in young, active patients. Long-term follow-up and functional outcome. J Bone Joint Surg 79A:575-582, 1997.
7. Dodd CAF, Hungerford DS, Krackow KA: Total knee arthroplasty fixation. Comparison of the early results of paired cemented versus uncemented porous coated anatomic knee prostheses. Clin Orthop 260:66-70, 1990.
8. Ewald FC: The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop 248:9-12, 1989.
9. Gill GS, Chan KC, Mills DM: 5- to 18-year follow-up study of cemented total knee arthroplasty for patients 55 years old or younger. J Arthroplasty 12:49-54, 1997.
10. Hungerford DS, Krackow KA, Kenna RV: Cementless total knee replacement in patients 50 years old and under. Orthop Clin North Am 20:131-145, 1989.
11. Insall J, Scott WN, Ranawat CS: The total condylar prosthesis. A report of two hundred and twenty cases. J Bone Joint Surg 61A:173-180, 1979.
12. Insall JN, Dorr LD, Scott RD, Scott WN: Rationale of the Knee Society clinical rating system. Clin Orthop 248:13-14, 1989.
13. Landon GC, Galante JO, Casini J: Essay on total knee arthroplasty. Clin Orthop 192:69-74, 1985.
14. Levi N, Kofoed H: Early failure of metal-backed patellar arthroplasty. J Bone Joint Surg 76B:675, 1994.
15. Lombardi AV, Engh GA, Volz RG, Albrigo JL, Brainard BJ: Fracture/dissociation of the polyethylene in metal backed patellar components in total knee arthroplasty. J Bone Joint Surg 70A:675-679, 1988.
16. Rackemann S, Mintzer CM, Walker PS, Ewald FC: Uncemented press-fit total knee arthroplasty. J Arthroplasty 5:307-314, 1990.
17. Ranawat CS, Flynn WF, Saddler S, Hansraj KK, Maynard MJ: Long-term results of the total condylar knee arthroplasty. A 15-year survivorship study. Clin Orthop 286:94-102, 1993.
18. Rand JA: Cement or cementless fixation in total knee arthroplasty? Clin Orthop 273:52-62, 1991.
19. Rand JA, Peterson LFA, Bryan RS, Ilstrup DM: Revision total knee arthroplasty. Instr Course Lect 35:305-318, 1986.
20. Rorabeck CH, Bourne RB, Nott L: The cemented Kinematic II and the non-cemented porous-coated anatomic prostheses for total knee replacement. J Bone Joint Surg 70A:483-490, 1988.
21. Rosenberg AG, Borden RM, Galante JO: Cemented and ingrowth fixation of the Miller-Galante prosthesis. Clinical and roentgenographic comparison after three- to six-year follow-up studies. Clin Orthop 260:71-79, 1990.
22. Sarokhan AJ, Scott RD, Thomas WH, et al: Total knee arthroplasty in juvenile rheumatoid arthritis. J Bone Joint Surg 65A:1071-1080, 1983.
23. Stuart MJ, Rand JA: Total knee arthroplasty in young adults who have rheumatoid arthritis. J Bone Joint Surg 70A:84-87, 1988.
24. Whiteside LA: Cementless total knee replacement. Nine- to 11-year results and 10-year survivorship analysis. Clin Orthop 309:185-192, 1994.