We analysed the long-term clinical and radiological results of 63 uncemented Low Contact Stress total knee replacements in 47 patients with rheumatoid arthritis. At a mean follow-up of 12.9 years (10 to 16), 36 patients (49 knees) were still alive; three patients (five knees) were lost to follow-up. Revision was necessary in three knees (4.8%) and the rate of infection was 3.2%. The mean clinical and functional Knee Society scores were 90 (30 to 98) and 59 (25 to 90), respectively, at final follow-up and the mean active range of movement was 104° (55° to 120°). The survival rate was 94% at 16 years but 85.5% of patients lost to follow-up were considered as failures. Radiological evidence of impending failure was noted in one knee.
Total knee replacement (TKR) is the treatment of choice for relieving pain and improving function in patients with rheumatoid arthritis. The Low Contact Stress (LCS) mobile bearing knee replacement (DePuy Orthopaedics Inc., Warsaw, Indiana) has been in extensive use since 1977. Most of the intermediate and long-term results have been reported1–7 in predominantly osteoarthritic patients.1–7 While the long-term results for TKR have been published for rheumatoid arthritis by various authors,8–19 there have been no studies into the performance of the uncemented LCS knee replacement in rheumatoid arthritis over an intermediate to long-term period. The purpose of this retrospective study was to report our ten- to 16-year clinical and radiological results in these patients.
Patients and Methods
Between October 1987 and May 1993, the senior author (SWM) performed 63 uncemented primary LCS mobile bearing total knee replacements in 47 patients with rheumatoid arthritis. Meniscal-bearing tibial components were used in all cases. The mean age of the patients at the time of their knee replacement was 69 years (53 to 81). The final follow-up date was designated as May 2003 so that all patients had a minimum follow-up of ten years (mean 12.9; 10 to 16). A previous operation had been performed in ten patients (ten knees), eight having a synovectomy and two an arthroscopic washout.
A midline skin incision with a medial parapatellar splitting of the quadriceps was used in 36 patients (50 knees) and a lateral parapatellar approach in 11 patients (13 knees). After achieving ligamentous balance, the proximal tibia was resected in order to achieve a surface that was perpendicular to the shaft of the tibia in the coronal plane but with a minimal posterior slope in the sagittal plane. The distal femur was then excised to achieve a tibiofemoral alignment of 5° to 7° in the coronal plane. A quantity of bone equal to the size of the component to be implanted was removed from the patella. Special attention was paid to achieving equal gaps in flexion and extension and to the correction of any pre-operative flexion contracture. The patella was resurfaced in 41 patients (53 knees).
After operation, patients undertook active and passive movements of their knees and began walking with crutches or walking sticks on the day after the procedure. Continuous passive motion machines were not used. Patients were then assessed at regular intervals of six weeks, three months and then annually thereafter.
At the most recent assessment, pain and function were assessed using the Knee Society scores.20 A score ≥ 85 was considered as excellent and ≤ 60 as poor. The range of movement, alignment, and contractures were assessed with a goniometer. Early post-operative and final follow-up standing anteroposterior and lateral radiographs were evaluated, according to the method of the Knee Society,21 for radiolucency at the bone-implant interfaces, the lateral and medial joint spaces, any change in the position of the components and osteolysis. Because of a lack of long pre-operative and post-operative standing radiographs, we cannot comment on any change in the mechanical axis.
Radiolucency at the bone-implant interface was rated in seven zones in the anteroposterior radiograph and three zones in the lateral radiograph of the tibial component, and in seven zones in the lateral film of the femoral prosthesis. The total of the widths of the radiolucencies were added for each of the three components in order to generate a score. A score of ≤ 4 was considered insignificant, 5 to 9 as indicating that a patient should be followed closely for progression and ≥ 10 as possible or impending failure of the prosthesis.21 Patients were assessed by three of the authors (SS, FN, MGH), independently of the surgeon performing the procedure.
We used a Kaplan-Meier analysis in order to calculate survivorship22 with the end-point being removal or revision of a component for any reason.
Between ten and 16 years after the procedure, 36 patients (49 knees) were alive and 11 (14 knees) had died at a mean of 7.1 years after surgery, although one had died intra-operatively and another in the first post-operative year. One patient had undergone a revision because of collapse and shift of the tibial component four months after surgery. This was thought to be because of soft-tissue imbalance. This patient also had cancellous bone grafting of a defect under the lateral tibial condyle and a lateral release. Three of the surviving patients (five knees) were lost to follow-up while two (three knees) could not attend the clinic for follow-up because of medical infirmity. Neither had undergone removal or revision of their primary knee replacement. Thirty-one of the surviving patients (41 knees) were followed clinically and radiographically and form the basis of this report.
Fifteen patients (48%) had surgery to other joints in addition to their TKR with a total hip arthroplasty in 11, ankle fusion in one and forefoot surgery in two. One patient sustained a supracondylar fracture above the femoral component 11 years after her knee replacement. This was treated with a supracondylar nail, the fracture healing seven months later. The mean clinical and functional Knee Society ratings were 90 (30 to 98) and 59 (25 to 90), respectively, at the final follow-up. The mean active range of movement was 104° (55° to 120°). At this time 37 knees (90.2%) were painfree, two (4.9%) mildly painful, and two (4.9%) gave moderate pain. Five patients could walk an unlimited distance, one could walk for 30 to 60 minutes, 14 for 10 to 30 minutes, and 11 for less than ten minutes. Seven patients required no walking aid, although five used them outdoors, and 19 patients needed them permanently.
We revised another two knees during the ten- to 16-year period of follow-up. One patient had a revision of the tibial component because of its collapse and tilt on the lateral side ten years after the primary procedure. The other underwent a revision to replace a worn meniscal bearing ten years after the primary procedure. There were no revisions of the patellar component.
We found that 28 patients (36 knees) were satisfied with the results of their operation. The remaining three patients (five knees) were dissatisfied because of a poor range of movement in one and poor function in two.
Radiographs of 28 knees showed minor radiolucency at the bone-prosthesis interface (Table I⇓). Twelve knees scored between 5 and 9 while a high risk of component failure was noticed in the tibial component of one patient at follow-up of 12 years. This patient declined a revision. No radiolucency was noted at the bone-prosthesis interface of the 35 patellar components which we reviewed.
Table II⇓ shows survivorship for all 63 knees. With removal or revision of a component for any reason as an endpoint, there was a survival rate of 94% at 16 years (95% confidence interval (CI) 90 to 98). When all patients lost to follow-up were assumed to have undergone revision surgery the survival rate was 85.5%. Post-operative infection occurred in two knees (3.2%) within six weeks of surgery and resolved with antibiotics.
Rheumatoid arthritis affects about 1% of the population. Most patients with long-standing rheumatoid arthritis have at least one, and often both knees affected.12 Total knee replacement has become the principal procedure in the treatment of such knees.
Although there have been previous accounts of the LCS knee replacement,1–7 these have largely concerned patients with osteoarthritis. However, in rheumatoid arthritis there may be differences in terms of the activity levels of patients, osteoporosis around the knee, disorder of other joints and the age of the patient at surgery. The information concerning follow-up for the two conditions are therefore not comparable. Our results match other studies on rheumatoid populations receiving different prostheses, which have shown good to excellent results in between 77% and 93% of patients.8–19 The functional status of rheumatoid patients after knee replacement is less than for those with osteoarthritis because of the polyarticular involvement and the declining functional status.23 In our series, 15 patients (48%) had surgery to joints other than the knee. The pain scores and the range of movement are not normally influenced by other disorders, which probably explains why in our patients the pain scores were better than their function scores, with 90% painfree, which compares with other studies.11,14,19
The prevalence of radiolucent lines is reported to be between 20% and 60%, the higher rates being seen in studies with a more than ten-year follow-up.18,24–26 In our series the prevalence of significant radiolucent lines was 32% at 16 years after surgery. One of the reasons for this lower figure could be that our population had rheumatoid rather than osteoarthritis. Reduced levels of activity because of multiple joint involvement may be a factor.
Apart from loosening and infection a significant complication after a TKR for rheumatoid arthritis is a peri-prosthetic fracture.18,27,28 In our patients, one knee sustained a supracondylar fracture in a fall ten years after surgery. Although a reduced range of movement predisposes to an increased risk of fracture,18 this patient had 90° of flexion in the knee before sustaining the fracture. The fracture was managed effectively with a supracondylar nail.
In long-term studies, the number of patients often decreases because of death from natural causes. Higher mortality rates have been reported in patients with rheumatoid arthritis than for those with osteoarthritis.29 The mortality rate of 23.4% in our patients compares with that in other studies.8–19 Our rate of infection of 3.2% was also similar to that previously reported by Gill and Joshi.8
Survivorship including patients lost to follow-up is presented on the assumption that all patients lost to follow-up failed. Our worst-case survivorship (85.5%) compares with Ito et al18 who reported survival of 76.7% at 15 years and Weir, Morgan and Pinder30 who reported a figure of 80.6% at ten years.
In conclusion, our study has uniformity, as a single surgeon performed all the operations and all patients received the same level of post-operative care. With this regime we found good ten to 16-year clinical and radiological results for the uncemented meniscal-bearing LCS knee replacement in patients with rheumatoid arthritis.
No benefits in any form have been received or will be received from any commercial party related directly or indirectly to the subject of this article.
- Received November 8, 2004.
- Accepted February 14, 2005.
- © 2005 British Editorial Society of Bone and Joint Surgery