J. R. Thompson, W. M. Harper, P. J. Gregg. From Glenfield Hospital, Leicester and the University of Newcastle upon Tyne, England. Early implants for total knee ...
Randomised, prospective study comparing cemented and cementless total knee replacement RESULTS OF PRESS-FIT CONDYLAR TOTAL KNEE REPLACEMENT AT FIVE YEARS A. W. McCaskie, D. J. Deehan, T. P. Green, K. R. Lock, J. R. Thompson, W. M. Harper, P. J. Gregg From Glenfield Hospital, Leicester and the University of Newcastle upon Tyne, England
arly implants for total knee replacement were fixed to bone with cement. No firm scientific reason has been given for the introduction of cementless knee replacement and the long-term survivorship of such implants has not shown any advantage over cemented forms. In a randomised, prospective study we have compared cemented and uncemented total knee replacement and report the results of 139 prostheses at five years. Outcome was assessed both clinically by independent examination using the Nottingham knee score and radiologically using the Knee Society scoring system. Independent statistical analysis of the data showed no significant difference between cemented and cementless fixation for pain, mobility or movement. There was no difference in the radiological alignment at five years, but there was a notable disparity in the radiolucent line score. With cemented fixation there was a significantly greater number of radiolucent lines on anteroposterior radiographs of the tibia and lateral radiographs of the femur. At five years, our clinical results would not support the use of the more expensive cementless fixation
E
A. W. McCaskie, MD, FRCS Orth, Lecturer in Orthopaedic Surgery D. J. Deehan, MD, FRCS Orth, Specialist Registrar in Orthopaedic Surgery P. J. Gregg, MD, FRCS, Professor of Orthopaedic Surgery Department of Trauma and Orthopaedic Surgery, School of Surgery, The Medical School, University of Newcastle, Newcastle upon Tyne NE2 4HH, UK. T. P. Green, FRCS Orth, Consultant Orthopaedic Surgeon Leicester General Hospital NHS Trust, Gwendolen Road, Leicester LE5 4PW, UK. K. R. Lock, MCSP, SRP, Research Physiotherapist (during study) W. M. Harper, MD, FRCS, Professor of Orthopaedic and Trauma Surgery Department of Orthopaedic Surgery, University of Leicester, The Glenfield Hospital NHS Trust, Groby Road, Leicester LE3 9QP, UK J. R. Thompson, PhD, Senior Lecturer in Epidemiology and Public Health University of Leicester, Robert Kilpatrick Clinical Sciences Building, Leicester Royal Infirmary, Leicester LE1 5WW, UK. Correspondence should be sent to Mr A. W. McCaskie ©1998 British Editorial Society of Bone and Joint Surgery 0301-620X/98/68558 $2.00 VOL. 80-B, NO. 6, NOVEMBER 1998
whereas the radiological results are of unknown significance. Longer follow-up will determine any changes in the results and conclusions. J Bone Joint Surg [Br] 1998;80-B:971-5. Received 18 November 1997; Accepted after revision 6 May 1998
Early implants for total knee replacement were fixed to bone with cement. Examination of cemented components from these prostheses removed at revision showed extrusion of cement on to articulating surfaces, unequal distribution over the surface of the implant and deficient interposition into cancellous bone. At the same time concern was being expressed about ‘cement disease’ in hip 1 replacement. Observed radiolucent lines were believed to 1 result from a local host response to cement or particulate 2 debris. Migration of the femoral component and subsidence of the prosthesis were thought to be the end-result of cement fracture or cement-induced bone resorption. Similar results with migration of components were 3 observed in total knee arthroplasty. These problems resulted in the development of methods of cementless fixation. Porous-coated prostheses allowed direct bone apposition 4 and subsequent bone ingrowth without cement. It has not been confirmed that cement is the initiator of osteolytic problems and indeed osteolysis has been reported 5 with uncemented prostheses. Uncemented hip replacements do not appear to have significantly greater survivor6 ship and cemented hip replacement has proved to be a 7 very reliable and successful procedure. No firm scientific reason has been given for the introduction of cementless knee replacement and the long-term survivorship of such implants has not shown an advantage over cemented 8,9 forms. We have compared the results of cemented and cementless knee replacement to test the null hypothesis of no difference in outcome between them.
Patients and Methods In 1987 we began a randomised, prospective study using the press-fit Condylar Knee Replacement System (PFC; Johnson & Johnson, Raynham, Massachusetts). This uses a cobalt-chrome femoral component and a titanium tibial 971
972
A. W. MCCASKIE,
D. J. DEEHAN,
Table I. Patients who were not entered or randomised during the surgical procedure Reason for exclusion
Number of knees (patients)
Bone-related Bone not suitable for cementless prosthesis Bone graft required
8 (7) 8 (6)
Other Randomisation error Cementless prosthesis unavailable
3 (3) 2 (2)
Total
21 (18)
component with a fixed polyethylene insert. Each size is available in cemented and cementless forms, with identical articulating geometry. For the cemented version there is a small recess for cement on the reverse of the component, which is otherwise smooth; the cementless form has a porous coating in the corresponding area. This implant is therefore ideal for study of the single variable of fixation. There was a total of 501 prostheses (393 patients) in the study. Between June 1987 and December 1990, 197 prostheses were implanted in 160 patients. Of these, 21 implants (18 patients) were not randomised or included in the study for the reasons given in Table I. Of the 176 prostheses which were randomised, 37 were not formally assessed at five years; in 27 (23 patients) the patient had died and in ten (8 patients) they were unable to attend the five-year follow-up because they were unwell or unwilling. Although formal assessment was not possible their basic data were collected and, in this group of 37 implants, there were no revisions and satisfactory formal assessment had been maintained until the lapse in follow-up. The remaining 139 prostheses (113 patients) were assessed at five years. There were 81 knees (67 patients) in the cemented and 58 knees (47 patients) in the uncemented group. Patients having the second knee replaced received the same fixation as on the other side. There was no significant difference in the gender ratio, diagnosis, age, and weight between the two groups (Table II). All analyses of response variables were adjusted for these characteristics but in no instance did the decision to adjust or not make any difference to the result. The operation was performed by or under the direct supervision of one surgeon (PJG) at the Glenfield Hospital NHS Trust in a theatre without a laminar flow enclosure but with antibiotic prophylaxis. No chemical thromboprophylaxis was used. The operating technique was standardised using a longitudinal incision and a medial parapatellar approach. The bone was cut using the manufacturer’s jigs
T. P. GREEN,
ET AL
and the implant positioned using the recommended surgical technique. The patella was not routinely resurfaced. Randomisation was performed during surgery. After making the bone cuts the quality of the bone stock and cuts was assessed and most patients were then randomised to either cemented or cementless fixation (pseudorandomisation according to year of birth). After operation the knee was then rested in a Robert Jones bandage for two days and a standard rehabilitation programme followed after wound review. Independent clinical review was carried out by a research physiotherapist. All patients were assessed before operation and afterwards at six months, one year and annually until five years. The detailed clinical assessment followed the 10,11 which assesses Nottingham data collection system pain, mobility, alignment and other physical characteristics. For pain a five-point scale graded the frequency, total severity and amount at night, and for mobility a similar evaluation was made of standing, walking, step-climbing, sitting and rising from a chair, with grade 5 being the highest, corresponding to the best level of performance, and grade 1 the lowest. Stability was subjectively assessed and the alignment of the leg as measured on long-leg films was recorded. Other factors measured included flexion contracture, further flexion, total flexion, extensor lag, and hip abduction. Standard anteroposterior and lateral radiographs were obtained to determine the orientation of the prostheses and the presence of lucent lines using the Knee Society 12 scoring system. Statistical analysis. We needed to allow for the potential correlation between the results obtained from both knees of the same patient. This was achieved with continuous variables, such as the angles measured from radiographs, by using a mixed model which incorporated random terms for subject and knee within subject and fixed effects for the treatment group and any covariates. With ordered categorical variables such as the pain scores, generalised estimating equations (GEE) were used incorporating a common odds ratio model for the association. The association estimates were updated according to the method suggested by 13 Lumley. Categorical baseline covariates were compared by a chi-squared test on numbers of subjects for gender and on the number of knees for diagnostic group and side. Continuous baseline covariates were compared using a mixed model. Baseline pain scores were compared using GEE. Five-year responses were adjusted for age, gender, diagnostic group, side, weight at follow-up and interval between operation and follow-up.
Table II. Details of the patients in the two groups. See text for tests used Male:female Osteoarthritis (%) Mean age at operation in years (SD; range) Mean weight at operation (kg) Mean weight gain at five years (kg)
Cemented
Cementless
p value
32:49 (male 40%) 68 (84) 68.8 (8.2; 41 to 87) 72.9 2.2
26:32 (male 46%) 50 (86) 70.2 (7.2; 59 to 86) 72.3 2.6
0.92 0.71 0.35 0.75 0.73
THE JOURNAL OF BONE AND JOINT SURGERY
RANDOMISED, PROSPECTIVE STUDY COMPARING CEMENTED AND CEMENTLESS TOTAL KNEE REPLACEMENT
Table III. Grades of pain (frequency, severity and night) observed before and after surgery in both groups Grade* 1
2
Cemented Cementless
8 8
64 40
8 9
1 1
0 0
Severity
Cemented Cementless
0 1
55 35
23 22
3 0
0 0
Night
Cemented Cementless
5 7
14 15
29 16
25 13
7 7
Cemented Cementless
1 1
3 0
8 4
18 21
51 32
Severity
Cemented Cementless
1 1
3 0
5 4
10 10
62 43
Night
Cemented Cementless
1 1
2 0
2 0
6 6
70 51
Preoperative Frequency
Five years Frequency
3
4
5
* see text
Table IV. Results for flexion contracture and range of movement before surgery and at five years for both groups Mean Preoperative Flexion contracture Range of movement Five years Flexion contracture Range of movement
Median
Range
Cemented Cementless Cemented Cementless
11.1 10.0 88.1 97.0
9.0 8.0 92.5 100.0
0.0 0.0 30.0 40.0
to to to to
38.0 40.0 127.0 130.0
Cemented Cementless Cemented Cementless
1.8 1.8 103.0 103.9
0.0 0.0 107.0 104.5
-5.0 to 15.0 -5.0 to 26.0 45.0 to 141.0 74.0 to 140.0
Results Power. If we consider the main outcome to be pain severity, a study of this size would have 85% power to detect a difference between a 75% success rate in one group and a 50% success rate in the other, with success defined as a pain severity of grade 5. Complications. None of the 139 prostheses (113 patients) required revision but there was one case of deep infection in a woman with a cementless implant. The infection was thought to be secondary to sepsis in the foot and occurred one year after operation. The same organism was cultured from both the foot and knee. Treatment with arthrotomy and antibiotics was successful. Three wounds were explored and resutured for minor necrosis or stitch abscess, all in the cemented group. Two patients with uncemented prostheses developed cellulitis which was successfully treated with antibiotics. Patellar resurfacing was subsequently required in two patients with uncemented knees and in one with a cemented prosthesis. In the whole group there were 33 positive venograms (27%) of which three (2%) were proximal. In the cemented group, there were 20 positive venograms, of which 19 were VOL. 80-B, NO. 6, NOVEMBER 1998
973
below the knee, with one above and below. In the cementless group, there were 13 positive venograms, with 11 below the knee, one above and below, and one above the knee. Clinical outcome. Table III shows the pain scores before operation and at five years using the Nottingham knee score. Most patients had frequent pain before their knee replacement (grade 2) but never or hardly ever five years after surgery (grade 5). There were similar results for severity of pain and pain at night (Table III). The three pain scales correlated highly within knee and less so within subject. Analysing the three pain scores as a set, the GEE model applied at baseline with adjustment found no difference between cemented and cementless fixation (p = 0.49). Neither was there any difference at follow-up (p = 0.83). None of the potential covariates was significantly associated with pain either at baseline or at follow-up. The only observable differences between the three scales were that the numbers responding with grade 5 were less for severity of pain than for pain at night and less for frequency of pain than for severity. Using the same method for mobility, similar patterns of improvement were noted for both techniques. There were no statistical differences at the 5% level between cemented and cementless fixation either before operation or at five years. Table IV shows the degrees of flexion contracture and the total range of movement before and at five years after operation. Before surgery there was no significant difference between cemented and cementless fixation in flexion contracture but the difference for range of movement was significant at the 5% level. At five years, however, no significant differences for both flexion contracture and range of movement were noted. For the whole group of 139 prostheses, the preoperative range of movement was significantly greater in men than in women explaining in part the differences between cemented and cementless knees. Radiological outcome. Prosthetic alignment in relation to the femur and tibia was measured using anteroposterior and lateral radiographs. The data analysed included the baseline and five-year measurements, along with the difference between them, i.e., the change in the measurement (Table V). Comparing the alpha, beta, gamma, omega and sigma angles at baseline and at five years showed no significant differences, although the measurement of the gamma angle was close to significance at baseline (p = 0.04), with the mean for the cemented group being 2.2° (SD 3.0) and for the cementless group 3.2° (SD 2.8). The difference between the alpha angle measured at baseline and follow-up (i.e., the change in alpha) was significant (p = 0.01), with the cemented group changing –2.2° (SD 3.3) and the cementless group –0.7° (SD 2.2). The number and thickness of radiolucent lines in each zone for each projection were added to produce a total. Scores for the lateral femoral, anteroposterior tibial and
974
A. W. MCCASKIE,
D. J. DEEHAN,
T. P. GREEN,
ET AL
Table V. Results of radiological measurement of prosthesis alignment and lucent line formation for cemented (72) and cementless (49) knees Mean ± Early postoperative Alignment angles
Alpha Beta Gamma Omega Sigma
Radiolucent lines
Lateral femur Anteroposterior tibia Lateral tibia
At five years Alignment angles
Alpha Beta Gamma Omega Sigma
Radiolucent lines
Lateral femur Anteroposterior tibia Lateral tibia
lateral tibial components treated as continuous variables, did not differ significantly at baseline. At five years, however, there was a difference in anteroposterior tibial scores (p = 0.02), with cemented at 2.19 (SD 1.83) and cementless at 1.41 (SD 1.67). In the five-year period, there was also a significant difference for both anteroposterior tibial (p = 0.03) and lateral femoral scores (p = 0.03). For the tibia the mean change for cemented was 1.58 (SD 1.73) and for cementless 0.96 (SD 1.38) and for the femur the mean change for cemented was 0.71 (SD 1.18) and for cementless 0.21 (SD 1.03).
Discussion We found no difference in the clinical outcome of the cemented and cementless knees. Both gave improvement in pain, function and joint movement and were equally effective. We have observed, however, a significantly greater number of radiolucent lines with cemented fixation at five years with both the tibial and femoral component. There was no difference in angular alignment of the prostheses after operation and at five years. There has been no associated clinical deterioration and it is unclear whether these changes will progress and present either the
SD
Cemented Cementless Cemented Cementless Cemented Cementless Cemented Cementless Cemented Cementless
97.6 96.7 85.8 86.4 2.2 3.2 6.9 6.7 86.1 85.7
± ± ± ± ± ± ± ± ± ±
3.1 2.7 2.6 2.3 3.0 2.8 3.7 3.5 3.3 2.9
Cemented Cementless Cemented Cementless Cemented Cementless
0.21 0.40 0.61 0.45 0.48 0.33
± ± ± ± ± ±
0.58 0.89 1.00 1.02 0.84 0.60
Cemented Cementless Cemented Cementless Cemented Cementless Cemented Cementless Cemented Cementless
95.4 96.0 97.4 97.5 1.2 1.8 5.2 4.6 86.4 86.2
± ± ± ± ± ± ± ± ± ±
2.9 2.3 2.5 2.4 2.8 3.0 2.6 2.5 3.2 2.8
Cemented Cementless Cemented Cementless Cemented Cementless
0.92 0.59 2.19 1.41 1.32 0.98
± ± ± ± ± ±
1.24 0.98 1.83 1.67 1.12 1.03
(range)
(92 to 108) (92 to 103) (78 to 90) (80 to 90) (-8 to ±10) (0 to 10) (-2 to ±16) (0 to 14) (80 to 94) (78 to 90) (0 (0 (0 (0 (0 (0
to to to to to to
3) 5) 6) 5) 4) 2)
(89 to 102) (90 to 101) (82 to 97) (93 to 92) (-10 to +12) (-5 to ±10) (0 to 12) (0 to 11) (78 to 93) (79 to 90) (0 (0 (0 (0 (0 (0
to 5) to 4) to 8) to 6) to 4) t 3)
symptoms or radiological signs of aseptic loosening. There are other differences between the two designs. Cemented knee replacement is said to be technically easier 14 to perform than cementless. The bone cuts need not fit the prosthesis precisely, with the layer of cement filling the defects. It has been said that surgeons who develop the necessary skills for cementless implants will “reap the 15 advantages”. One such advantage is a more effective biological seal from wear particles which ultimately leads to less osteolysis and loosening. Polyethylene wear particles appear to be important factors in loosening and may be more sensitive to changes in design than in fixation of the prosthesis. The final difference is in the cost. When there is no clinically proven superior performance there is no reason for choosing an expensive implant. Generally, the cemented form of fixation is cheaper than the cementless. At five years, our clinical results do not support the use of the more expensive cementless fixation whereas the radiological results are of unknown significance. Longer follow-up will determine any changes in the results and conclusions. We are currently awaiting the results of the ten-year assessments, after which we will present a more extensive survival analysis. THE JOURNAL OF BONE AND JOINT SURGERY
RANDOMISED, PROSPECTIVE STUDY COMPARING CEMENTED AND CEMENTLESS TOTAL KNEE REPLACEMENT
Although none of the authors have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received but are directed solely to a research fund, educational institution, or other non-profit institution with which one or more of the authors is associated.
References 1. Jones LC, Hungerford DS. Cement disease. Clin Orthop 1987;225: 192-206. 2. Howie DW, Haynes DR, Rogers SD, McGee MA, Pearcy MJ. The response to particulate debris. Orthop Clin North Am 1993;24: 571-81. 3. Nilsson KG, Karrholm J, Linder L. Femoral component migration in total knee arthroplasty: randomized study comparing cemented and uncemented fixation of the Miller-Galante I design. J Orthop Res 1995;13:347-56. 4. Engh CA, Bobyn JD, Glassman AH. Porous-coated hip replacement: the factors governing bone ingrowth, stress shielding and clinical results. J Bone Joint Surg [Br] 1987;69-B:45-55. 5. Maloney WJ, Woolson ST. Increasing incidence of femoral osteolysis in association with uncemented Harris-Galante total hip arthroplasty: a follow-up report. J Arthroplasty 1996;11:130-4. 6. Maric Z, Karpman RR. Early failure of noncemented porous coated anatomic total hip arthroplasty. Clin Orthop 1992;278:116-20.
VOL. 80-B, NO. 6, NOVEMBER 1998
975
7. Malchau H, Herberts P, Ahnfelt L. Prognosis of total hip replacement in Sweden. Acta Orthop Scand 1993;64:497-506. 8. Havelin LI, Espehaug B, Vollset SE, Engesaeter LB. Early aseptic loosening of uncemented femoral components in primary total hip replacement: a review based on the Norwegian arthroplasty register. J Bone Joint Surg [Br] 1995;77-B:11-7. 9. Albrektsson BEJ, Carlsson LV, Freeman MAR, Herberts P, Ryd L. Proximally cemented versus uncemented Freeman Samuelson knee arthroplasty: a prospective, randomised study. J Bone Joint Surg [Br] 1992;74-B:233-8. 10. Tew M, Waugh W. Guide to recording information about knee replacements: a manual for use in outpatient clinics and hospitals. Department of Orthopaedic Surgery, University of Nottingham, 1980. 11. Tew M. Evaluation. In: Jackson JP, Waugh W, eds. Surgery of the knee joint. London: Chapman & Hall, 1984:392-410. 12. Ewald FC. The knee society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop 1989;248:9-12. 13. Lumley T. Generalized estimating equations for ordinal data: a note on working correlation structures. Biometrics 1996;52:354-61. 14. Booth RE Jr. Cementless TKA: an endangered species? In the affirmative. 12th Annual Current Concepts in Joint Replacement (abs), 1996:123. 15. Whiteside LA. Cementless TKA: an endangered species? In opposition. 12th Annual Current Concepts in Joint Replacement (abs), 1996:125.
