We identified 148 patients who had undergone a revision total knee replacement using a single implant system between 1990 and 2000. Of these 18 patients had died, six had developed a peri-prosthetic fracture and ten had incomplete records or radiographs. This left 114 with prospectively-collected radiographs and Bristol knee scores available for study. The height of the joint line before and after revision total knee replacement was measured and classified as either restored to within 5 mm of the pre-operative height or elevated if it was positioned more than 5 mm above the pre-operative height. The joint line was elevated in 41 knees (36%) and restored in 73 (64%).
Revision surgery significantly improved the mean Bristol knee score from 41.1 (sd 15.9) pre-operatively to 80.5 (sd 15) post-operatively (p < 0.001). At one year post-operatively both the total Bristol knee score and its functional component were significantly better in the restored group than in the elevated group (p < 0.01). Overall, revision from a unicondylar knee replacement required less use of bone graft, fewer component augments, restored the joint line more often and gave a significantly better total Bristol knee score (p < 0.02) and functional score (p < 0.01) than revision from total knee replacement.
Our findings show that restoration of the joint line at revision total knee replacement gives a significantly better result than leaving it unrestored by more than 5 mm. We recommend the greater use of distal femoral augments to help to achieve this goal.
Revision total knee replacement (TKR) accounts for between 5% and 10% of all knee replacements performed,1 and this number is increasing.2 The results after revision TKR may not match those from the primary surgery.3 One of the reasons proposed for this is the difficulty in restoring the joint line.4
Figgie et al5 have shown that elevation of the joint line by more than 8 mm at primary TKR is associated with an inferior clinical result. Partington et al1 found that elevation of the joint line occurred in 79% of revision TKRs and reduced the mean Knee Society clinical rating score from 141 points to 125 points if the elevation exceeded 8 mm. However, the difference did not reach statistical significance.
Elevation of the joint line occurs when bone lost from the distal femur is not replaced.6 This has led to the recommendation that distal femoral augments should be used more frequently and in greater thickness in revision TKR.1,7
Our aim was to review the revision TKR procedures performed in our unit, with specific reference to restoration of the joint line and the functional outcome.
Patients and Methods
Between 1990 and 2000, we carried out 148 consecutive revision TKRs using the Press Fit Condylar and its total condylar version 3 revision system (Depuy, Warsaw, Indiana). Each procedure had been carried out by a consultant orthopaedic surgeon who was a member of the Bristol Knee Group.
We excluded 34 of the 148 patients. A total of 18 had died, six had sustained a periprosthetic fracture, five had incomplete data and in five the pre- or post-operative radiographs were missing. This left 114 patients with a minimum follow-up of one year. Their details are given in Table I⇓. No re-revision cases were included in the series.
Data were collected prospectively as part of the Bristol Knee Group database. These included the Bristol knee score8–11 which was measured pre-operatively and at one, two, five, eight and ten years post-operatively. The Bristol knee score8–11 is a clinical scoring system which has been used for primary and revision TKR in our unit for many years (Fig. 1⇓). It is a 100-point score assessing joint pain, function, movement, deformity and laxity. A maximum score of 27 can be achieved for the functional component of the score. In a study by Bach et al,10 the Bristol knee score had the best interobserver reproducibility of the scoring systems studied. It has not been validated but does allow separation of knee function from overall patient function, which may be governed by comorbidities.11 The total and functional components were analysed in our study.
Operative data on bone defects, the use of bone graft, the use of prosthetic augments and stems and the thickness of the polyethylene insert used, were available from the operation records and since 1992 from specific locally-devised forms completed by the surgeon at the time of surgery.
A retrospective radiological review of the position of the tibiofemoral joint line before and after revision TKR was performed on standard anteroposterior and lateral radiographs. The position of the joint line was measured on the lateral radiograph using the method described by Figgie et al5 (Figs 2⇓ and 3⇓). Measurement was made from the top of the tibial tubercle to the superior surface of the tibial component: however, due to the radiolucency of the tibial inserts this point was taken to be the most distal part of the femoral component. Measurements were made independently by two observers (AJP, MAH) on pre- and post-revision radiographs only, as many of the radiographs performed before the primary TKR were not available, either because they had been retained by the referring hospital or had been destroyed after prolonged storage. Interobserver variability was calculated using Kappa’s measure of agreement.12
The difference between pre- and post-operative measurements was recorded. These results were then grouped into two categories based on whether the joint line had been restored either to within 5 mm or elevated by > 5 mm.
The data were analysed for the group as a whole as well as for various subgroups; these were elevated, and restored joint-line (41 patients and 73 patients, respectively) and whether it was revision from unicondylar knee replacement (UKR) or TKR (35 patients and 79 patients, respectively). The data were analysed to give means and sds with differences tested using paired and unpaired t-tests and the chi-squared test as appropriate. A p-value ≤0.05 was considered to be statistically significant.
Progression of arthritis was the most common reason for revision in the UKR group while aseptic loosening and infection were the most common causes for revision in the TKR group (Table II⇓).
The mean pre-operative total Bristol knee score was 41.1 (sd 15.9) which improved to a mean score of 80.5 (sd 15) at one year. The mean functional components of the Bristol knee score improved from 11.4 (sd 5.4) pre-operatively to 19.3 (sd 4.8) at one year. Both of these improvements were statistically significant (paired t-test, p < 0.001).
Good interobserver agreement was found between the two observers for measurements of change in the height of the joint line in millimetres with a kappa value of 0.775. When the joint line was classified as either elevated or restored there was complete agreement between the observers. The joint line was elevated in 41 knees (36%) and restored in 73 (64%). No significant difference was found between the mean pre-operative Bristol knee score or its functional component in the elevated and restored groups (unpaired t-test, p = 0.06).
A comparison of the mean scores after one year between the elevated and restored groups showed the restored group to have a significantly better Bristol knee score and function score (t-test, p < 0.01 and p < 0.005, respectively) than the elevated group (Table III⇓).
A total of 27 patients had scores available at five years post-operatively. The mean Bristol knee score and function score for each group at five years still showed the restored group to have significantly better result than the elevated group (Table IV⇓). In the elevated group there was a decrease in the mean Bristol knee score and function score between one and five years post-operatively, while in the restored group the mean Bristol knee score and function score both improved from one to five years. This change was not statistically significant.
In the knees with a five-year follow-up, 13 (48%) had elevated joint lines, while in those with follow-up for two years or less, 31 (36%) had elevated joint lines. This was not statistically significant (chi-squared test with Yates correction, p = 0.06).
Thickness of the polyethylene insert.
For all revisions the mean thickness of the polyethylene insert used was 14.8 mm (sd 4.4). In the elevated group the mean thickness was 16.9 mm (sd 3.3) while in the restored group it was 13.9 mm (sd 3.8) which was statistically significant (unpaired t-test, p < 0.001). There was a significantly higher proportion of cases with a thickness ≥ 17.5 mm in the elevated group than in the restored group (chi-squared test, p < 0.02).
Management of bone defects.
The distribution, type and size of the bone defects at revision are given in Table V⇓. These were managed in a number of ways. In 41 knees (36%) morcellised bone graft was used. In 28 of these autograft only was used, principally for small contained defects, in nine allograft alone, and in four both allograft and autograft were used.
Posterior femoral augments were used in 40 knees (35%). In four, an augment was used only on the medial side, in five only laterally and in 31 on both sides.
Distal femoral augments were used in 34 (43%) of the revisions from failed primary TKR. When a distal augment was used after a failed TKR, the joint line was elevated in only 10 knees (30%), whereas, if no distal augment was used, the joint line was elevated in 28 knees (62.5%). This difference was statistically significant (chi-squared with Yates correction, p < 0.02).
Unicondylar compared with total knee replacement.
The mean Bristol knee scores at one year for revision of UKR and TKR were compared. In the restored group there was no significant difference in the function score between UKR and TKR (unpaired t-test, p = 0.1) (Table VI⇓). Revision from UKR elevated the joint line in only 3 (9%) knees while revision from TKR elevated the joint line in 38 (48%) knees. This difference was statistically significant (chi-squared test, p < 0.001). For ten of the 35 revised UKRs (28.5%) bone graft was required, but in nine of these only autograft was used. In contrast, 31 of the 79 revised TKRs (39%) required bone graft. In 19 this was autograft alone, in nine it was allograft and in three both allograft and autograft were necessary. In 25 revisions from TKR distal femoral augments were required while none was used in the revisions from UKR.
Revision TKR is a complex procedure which requires a good understanding of the need to balance both the soft-tissue and bony components of the joint to achieve optimum knee kinematics.1,6 Elevation of the joint line occurs when the bone lost from the distal femur is not reconstructed and the defect is addressed by thickening the tibial insert. The distal femoral defect can result from excessive bone resection at primary TKR or from bone loss at revision as a result of osteolysis, infection, migration of the component or damage to the distal femur during removal of the component.
It has been suggested that a number of landmarks can be used intra-operatively to assess whether the joint line has been restored. These include the old meniscal scar if visible, or at a point 1.5 cm to 2 cm proximal to the fibular head, 2.2 cm proximal to the tibial tuberosity, 2 cm to 2.5 cm distal to the lateral femoral epicondyle or 2.5 cm to 3 cm distal to the medial femoral epicondyle.7
Potential problems caused by elevation of the joint line include patella infera, impingement of the patellar button and accelerated wear, mid-flexion laxity, weakness of quadriceps, anterior knee pain and hyperextension instability.6 Before the development of specific revision TKR implant systems the presence of femoral bone loss often required elevation of the joint line in order to obtain satisfactory bony support for the femoral component. Specialised modular revision systems were developed to address difficulties with restoring the joint line, controlling instability and dealing with bone defects.6
The Press-Fit Condylar cruciate-substituting femoral component and its associated revision total condylar version 3 component were used in our study and equipped with femoral and tibial stems, as well as various augments on the tibial component and the distal and posterior aspects of the femoral component.
In our study, we were able to show a significant difference in both the total Bristol knee score and the functional component of this score if the joint line was elevated by more than 5 mm. This improvement in scores with joint line restoration occurred at one year post-operatively and was still present in those patients followed for five years. We were also able to show a significant improvement in the ability to restore the joint line when distal femoral augments were used.
Partington et al1 found greater elevation of the joint line in cases in which distal femoral augments had been used. This apparent contradiction could be explained by the fact that distal femoral augments were probably only considered in cases in which large amounts of distal femoral bone had been lost, and the augments did not provide full correction. This is supported by their conclusion that distal femoral augments of greater thickness should be used more often in revision TKR. They also found increased thickness of the polyethylene in joints which used distal femoral augments, but this may have been due to the fact that they included second- and third-time revisions as well as revisions having significant tibial bone loss,13 whereas we only included first-time revisions which rarely involved cases of excessive tibial bone loss. In the setting of minimal tibial bone loss, it was logical that we should show a significant association between increasing thickness of the polyethylene and elevation of the joint line.
There is debate about the use of UKR as opposed to TKR in unicompartmental disease. The advocates of UKR report a similar survival rate to that of TKR and improved range of movement, function and speed of recovery as factors supporting its use.14,15 Ease of revision of UKR is also often cited. Our results provide evidence for this last assertion showing that revision TKR after a primary UKR requires less frequent use of bone graft, fewer augments, restoration of the joint line more often and improved functional results over revisions after primary TKR.
Our study has shown that elevation of the joint line by more than 5 mm in revision TKR can significantly worsen the clinical outcome and reinforces the message that thicker distal femoral augments, rather than thicker polyethylene inserts, should be used to restore the defects and allow restoration of the joint line.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received December 4, 2007.
- Accepted February 7, 2008.
- © 2008 British Editorial Society of Bone and Joint Surgery