From January 1996 to December 1998, forty-seven patients who had
bilateral arthritis of the knees with similar deformity and preoperative range
of motion were offered simultaneous bilateral total knee replacement and were
invited to have one knee replaced with an LCS rotating-platform
(mobile-bearing) prosthesis and the other with an Insall Burstein-II
(fixed-bearing) prosthesis. Patients were requested to agree to a random
selection as to which knee would receive the fixed bearing and which would
receive the mobile bearing. Thirty-four patients agreed to participate in the
investigation. The study was approved by the institutional review board, and
all patients provided informed consent to participate in it. The surgeon was
informed on the night before the surgery as to which prosthesis was to be
implanted in which knee. As a convention, the first operation was always
performed on the right knee. Two patients (four knees) were subsequently lost
to follow-up. The last follow-up in both of these patients was three months
after surgery. Since the follow-up period was so short, these two patients
were excluded from the study. The demographic data and mean duration of
follow-up for the remaining thirty-two patients are given in
Table I. All patients had the
bilateral total knee arthroplasty performed under a single anesthetic session.
Regional anesthesia (spinal or epidural) was used in twenty-four patients,
whereas general anesthesia was used in eight patients.
All of the replacements were performed by the senior author (S.B.).
Antibiotic prophylaxis with intravenous cefazolin (2 g given one-half hour
before inflation of the tourniquet followed by 1 g every eight hours for three
days) and anti-thrombotic prophylaxis with subcutaneous enoxaparin (40 mg on
the night before surgery and 40 mg daily continued through the tenth
postoperative day) were used in all patients.
Technique
The surgeries were performed under tourniquet control with use of a medial
parapatellar approach in all knees. The posterior cruciate ligament was
released in all knees. The patella was not replaced in any patient; however, a
patelloplasty was done in
all10. In all of
the patients with rheumatoid arthritis and in twelve of the patients with
osteoarthritis with thick, deformed patellae, 2 to 4 mm of the articular
surface was
removed11 and the
medial and lateral facets were reformed.
The Insall Burstein-II and the LCS rotating-platform prostheses were used
in the thirty-two patients (sixteen with osteoarthritis and sixteen with
rheumatoid arthritis). Appropriate soft-tissue procedures to realign the knee
were performed in both groups. Eight knees in each group required extensive
posterior capsulotomy, and four required an additional 5-mm distal femoral cut
to correct a severe preoperative flexion deformity (40°). The so-called
no-thumb technique was used to assess the adequacy of patellar tracking and
the need for lateral retinacular
release12. A
lateral retinacular release was required in four knees in each group.
Both the tibial and the femoral components were cemented in sixty-two
knees, while hybrid fixation (a cementless femoral component with a cemented
tibial component) was used in two knees managed with a mobile-bearing
prosthesis. Pulsatile lavage of the osseous surface was used before cementing
the implants. The postoperative regimen included use of a knee immobilizer for
two days, gravity-assisted motion to regain flexion from the third day, and
walking with support from the fourth postoperative day.
Follow-up evaluation included examination in the outpatient department at
one, three, and six months postoperatively and then at yearly intervals. The
average duration of follow-up was six years (range, 4.5 to 7.5 years).
Clinical Evaluation
Preoperative and postoperative clinical evaluations were done according to
the Knee Society
recommendation13.
This includes a knee score and a function score. Both were recorded, but only
the Knee Society knee score was used for comparison. The function score was
not used for comparison of the two implants, since it is difficult to make a
meaningful comparison of function when the patient is his or her own
control.
Patients were also categorized into three groups according to the status of
other joints and medical
comorbidities13.
The postoperative status of the patellofemoral joint was graded according to a
system developed by Stern and
Insall14.
The overall patient satisfaction after surgery was evaluated with use of
the British Orthopaedic Association patient-satisfaction
score15. The
response was rated as enthusiastic, satisfied, noncommittal, or disappointed
with the result of the knee arthroplasty.
Clinical evaluation was done on an outpatient basis by two residents who
did not know which prosthesis was used in each knee.
Radiographic evaluation was done by the senior author (S.B.) on the
preoperative and the postoperative radiographs, which included an
anteroposterior radiograph made with the patient standing, lateral radiographs
made with the patient supine, and a skyline patellar radiograph. The overall
alignment of the limb, the position of each of the implants, and the location
of radiolucent lines at the bone-cement interface were analyzed according to
the guidelines of the Knee
Society16. A loose
prosthesis was diagnosed when there was a progressive lucency of >2 mm
surrounding the entire border of the prosthesis, subsidence of the component,
or a change in alignment of the prosthesis compared with its immediate
postoperative status. Patellar tilt, subluxation, or dislocation, if any, were
assessed with the skyline patellar radiograph.
Statistical Analysis
The results in the two groups were compared, with use of the paired t test,
on an overall basis and by diagnosis with respect to the preoperative and
postoperative knee scores, improvement in knee scores, and range of motion
achieved postoperatively. Within the two groups, the distribution of
preoperative and postoperative knee scores and range of motion among the
patients with rheumatoid arthritis and those with osteoarthritis were analyzed
with use of the independent samples t test. The independent t test was used to
evaluate the overall complication rates and the prevalence of radiolucent
lines in the two groups. The Kaplan-Meier survivorship analysis was performed
to assess the survival of the fixed-bearing and mobile-bearing total knee
replacements at the time of final follow-up. A post hoc power calculation was
done on the difference in the survival rates of the implants, the knee scores,
and the postoperative range of motion in the two groups. The study would have
90% power to detect a difference of 30% in the survival rates of the two types
of implants, a 10-point difference in the improvement in the knee scores, and
a 12° difference in the improvement in the range of motion with a sample
size of thirty-two subjects.
Statistical analysis was performed with use of SPSS software (SPSS,
Chicago, Illinois).
The mean follow-up period was six years (range, 4.5 to 7.5 years).
At the time of the index arthroplasty, twenty-five patients (78%) were
categorized in group B and seven patients (22%) were categorized in group C,
according to the Knee Society clinical rating
system13, on the
basis of the status of their other joints and medical comorbidities.
Clinical Results
The mean preoperative Knee Society knee score (and standard deviation) was
35 ± 13.9 points for the LCS rotating-platform (mobile-bearing) group
and 34.2 ± 15 points for the Insall Burstein-II (fixed-bearing) group
(p = 0.27). The mean postoperative knee score was 90 ± 4.86 points in
the mobile-bearing group and 89.4 ± 5.64 points in the fixed-bearing
group (p = 0.084). The mean improvement in the Knee Society knee score was 55
± 11.7 points (95% confidence interval, 50.8 to 59.2 points) for the
fixed-bearing group and 55.2 ± 13 points (95% confidence interval, 50.1
to 59.5 points) for the mobile-bearing knees. The difference was not found to
be significant, with the numbers available (p = 0.94).
The mean preoperative arc of motion was 77.5° ± 15° (range,
55° to 105°) in the mobile-bearing group and 74.5° ±
15.6° (range, 50° to 110°) in the fixed-bearing group. Motion
improved to 106.9° ± 7.8° (range, 90° to 125°) and to
105.6° ± 7.7° (range, 90° to 120°), respectively; the
difference was not significant (p = 0.161). The mean improvement in the
postoperative arc of flexion was 31.1° ± 14.2° (95% confidence
interval, 26.7° to 37°) for the fixed-bearing knees and 29.4°
± 13.6° (95% confidence interval, 24.9° to 34.8°) for the
mobile-bearing knees (p = 0.54).
The mean preoperative flexion deformity was 15° (range, 5° to
40°) in both groups. Postoperatively, a flexion deformity (mean, 5°;
range, 2° to 10°) persisted in six knees (three with an LCS prosthesis
and three with an Insall Burstein-II prosthesis) at the latest follow-up
evaluation. Preoperatively, the flexion deformity had been 40° in four of
those knees and 35° in the other two. The deformity could not be corrected
completely at the time of surgery in four of the knees (the mean flexion
deformity following surgery was 9°, which improved to 5° at the time
of the latest follow-up). Two knees had recurrence of this deformity
postoperatively.
Excellent or good results based on the Knee Society knee score were
obtained in 88% (twenty-eight) of the thirty-two knees in both groups. All but
three patients could walk independently without support for an unlimited
distance. Two patients attributed their inability to walk to the knee with the
mobile-bearing prosthesis, and one attributed it to the knee with the
fixed-bearing prosthesis. On the basis of the British Orthopaedic Association
patient-satisfaction index, 82% (twenty-six) of the thirty-two patients were
enthusiastic about the results of their knee
replacements14.
The distribution of preoperative and postoperative knee scores and range of
motion, according to diagnosis, was similar for both groups
(Table II).
The mean preoperative function score was 32.5 ± 6 (range, 0 to 50).
At the time of the latest follow-up, the mean function score was 73.75
± 5.9 for the sixteen patients with osteoarthritis and 58.1 ±
4.4 for the sixteen patients with rheumatoid arthritis (p < 0.001).
Radiographic Results
The preoperative femorotibial alignment (anatomic axis) averaged 15° of
varus (range, 10° to 40° of varus) in fifty-eight knees (twenty-six
with rheumatoid arthritis and thirty-two with osteoarthritis). In six knees
with rheumatoid arthritis, the preoperative femorotibial alignment averaged
12° of valgus (range, 5° to 20° of valgus). At the time of the
final follow-up, the femorotibial alignment was 6° of valgus (range,
0° to 9° of valgus) in the fixed-bearing subgroup and 5° of valgus
(range, 1° to 9° of valgus) in the mobile-bearing subgroup. The
component alignment was evaluated in both groups with use of the Knee Society
radiographic index (Table III).
Four (13%) of the thirty-two knees with a fixed-bearing and ten (31%) of the
thirty-two knees with a mobile-bearing had lucencies around the tibial
prosthesis (p = 0.1). In thirteen of them, the radiolucencies were <2 mm
and were nonprogressive. One mobile-bearing knee developed progressive
radiolucency involving all four zones with loosening and migration of the
tibial prosthesis.
Complications
Patellar tilt was seen in four knees (two in the fixed-bearing group and
two in the mobile-bearing group) and was associated with mild stiffness and
anterior knee discomfort during stair-climbing or rising from a sitting
position. In the fixed-bearing group, inadvertent malposition of the tibial
component in internal rotation led to patellar tilt in one knee.
In the mobile-bearing group, so-called overstuffing of the anterior
compartment was believed to be the cause of patellar tilt in two knees. In
both knees, the smallest implant size available was used. Neither of these
patients required subsequent surgery during the follow-up period.
Deep infection with septic loosening of the tibial component necessitated
implant removal and arthrodesis in one patient with a mobile-bearing knee
replacement one year after the index operation.
One knee in the mobile-bearing group had recurrent dislocation of the
rotating bearing six months after the index operation
(Fig. 1). This patient had a
severe preoperative deformity (26° of varus and 20° of flexion), and
the flexion gap intraoperatively was 3.0 mm more than the extension gap, which
led to the recurrent bearing dislocation. This knee was subsequently revised
to a fixed-bearing prosthesis.
Survivorship Analysis
With use of repeat surgery for any reason as the end point, Kaplan Meier
survivorship analysis revealed that the survival rate at six years of
follow-up was 100% in the fixed-bearing subgroup and 94% in the mobile-bearing
subgroup. With the numbers available, the difference was not found to be
significant.
Fixed-bearing prostheses have provided long-term fixation with
prosthetic survival rates of 95% to 97% at ten to fifteen years of
follow-up17.
Various independent studies for both mobile-bearing and fixed-bearing
prostheses involving sixty-two to 473 total knee arthroplasties have
documented results that are comparable in terms of performance and survival,
with overall revision rates of approximately 1% per year for both types of
implants2,17-28.
No previous controlled comparison that we are aware of has been able to show
any advantage for a mobile-bearing over a fixed-bearing total knee prosthesis
either in terms of clinical function or
longevity29.
Favorable mid-term and long-term results for both the Insall Burstein-II and
the LCS prosthesis have been shown by several
authors8,15,17,30-33.
The purpose of this study was to analyze the individual performance of
fixed-bearing and rotating-bearing knee replacements in an identical clinical
setting by eliminating variables such as age, weight, and activity level. The
particular strength of this trial design is that the statistical comparisons
are within the patient (paired) and not between the patients as is more
typical. All surgeries were performed by the senior author (S.B.). The
clinical evaluation was done by observers blinded to the type of implant in a
particular knee, and the patients were also blinded as to which prosthesis had
been implanted in which knee. Observer and patient-related bias were thus
minimized.
The clinical results of both arthroplasties were similar. No benefit of the
mobile-bearing knee over the fixed-bearing knee could be seen with respect to
the overall knee score, postoperative range of motion, and survival rate on
the basis of the size of the series. Excellent or good results were obtained
in 88% (twenty-eight) of the thirty-two patients in both groups. Most et al.,
in an experimental study, showed that both mobile-bearing and fixed-bearing
implants had similar kinematic patterns with regard to posterior femoral
translation and tibiofemoral rotation despite the fact that their designs are
different34. They
suggested that the mobile tibial insert stops moving at <90° of flexion
and, after this point, the prosthesis performs essentially as a fixed-bearing
implant. Similar findings were reported by D'Lima et
al.35. The clinical
results of the present study are consistent with the findings of these
experimental studies. Both the fixed-bearing and the mobile-bearing group had
similar postoperative range of motion, which suggests that the in vivo
kinematics of these implants may, in fact, be similar.
Patients with rheumatoid arthritis in both groups had significantly lower
preoperative and postoperative knee scores (p < 0.001). Rehabilitation in
these patients is likely to be sub-optimal because of the soft bone quality,
scarred soft tissues, flare-up of disease, use of prolonged corticosteroid
therapy, and extensive
synovitis36. We
believe that this led to the lower postoperative range of motion in these
patients. The standard methods of assessing general functional improvement and
patient satisfaction after total joint replacement depend upon bipedal
activities (distance the patient is able to walk, stair-climbing, and use of
support) and cannot be used to compare one leg with the other. Hence, function
score and patient satisfaction scores were not used for comparison in this
study.
Patellar resurfacing was not done in any of the knees as we believe that
the presence of a small patella, typical of patients in our part of the world,
and the higher rate of lateral retinacular release (thus potentially
compromising the vascularity of the patella) would increase the risk of
patellar fracture if resurfacing was
attempted10.
Lateral retinacular release was required in 13% (four) of the thirty-two knees
in both groups. We believe that this high rate of lateral release in our
patients was related to the advanced deformities seen preoperatively in our
patients.
Dislocation is a potential complication with any mobile-bearing knee
replacement, and the LCS prosthesis is no
exception37,38.
In our series, one of the thirty-two knees with an LCS prosthesis had a
bearing dislocation that required revision to a fixed-bearing knee design six
months after the index operation. We believe that the severe preoperative
deformity in this rheumatoid knee, which required extensive soft-tissue
release, led to this complication. Both reoperations in our study were in the
knees with an LCS implant.
No significant difference was detected in the rates of survival between the
two prostheses, with the numbers available. The lack of statistical power is a
potential drawback of this study, owing to the small number of patients
evaluated. Our study lacked adequate power to compare the survival rates and
the improvement in knee scores and range of motion between the two groups.
Our series showed satisfactory results with both the Insall Burstein-II and
LCS prostheses, on the basis of the clinical and radiographic results,
patellofemoral complications, and rates of survival. No benefit of the
mobile-bearing design over the fixed-bearing design could be demonstrated,
with the numbers available. The risk of bearing subluxation and dislocation
associated with the mobile-bearing knee replacement is a cause for concern and
may necessitate early revision. ?