- Email: [email protected]
Revision of the Unstable Total Knee Arthroplasty
The Journal of Arthroplasty Vol. 26 No. 8 2011
Revision of the Unstable Total Knee Arthroplasty Outcome Predictors Khalid Azzam, MD, Javad Parvizi, MD, FRCS, Daniel Kaufman, BS, James J. Purtill, MD, Peter F. Sharkey, MD, and Matthew S. Austin, MD
Abstract: Revision arthroplasty is often required for confirmed cases of symptomatic tibiofemoral instability after total knee arthroplasty (TKA). We evaluated the results of revision for TKA instability in a consecutive series of 67 patients (68 knees) between 2000 and 2006. Outcome measures were surgeon-based assessment of knee stability, Knee Society Score, and Short Form Health Survey 36. At an average of 39 months of follow-up, the mean Knee Society Score and Short Form Health Survey 36 physical and mental scores were 76, 53, and 67 points, respectively. Knee instability persisted in 14 patients (22%). Data at the 95% confidence level revealed that revising both the femoral and tibial components, the use of femoral augments, and smaller joint line elevation as measured on radiographs correlated significantly with achieving a stable knee. In revision surgery for TKA instability, revision of both components and the use of femoral augments seem to offer the most predictable outcome. Keywords: revision knee arthroplasty, instability, outcome. © 2011 Elsevier Inc. All rights reserved.
Instability is a well-recognized cause of poor outcome after total knee arthroplasty (TKA). Causes of instability after TKA include inadequate soft tissue balancing, loss of ligamentous integrity, component wear, improper component sizing, and component malpositioning. It is often difficult to establish a diagnosis. Persistent or recurrent pain, especially during the first few years after surgery, accompanied by demonstration of soft tissue laxity on clinical examination can be a marker of an unstable TKA. However, the unstable TKA may present with pain without objectively demonstrable signs of instability. Therefore, exclusion of other etiologies of failure, such as infection or loosening, in all patients presenting with painful TKA is prudent. There are few studies reporting the outcome of revision arthroplasty for instability. Three studies have reported successful revision of cruciate-retaining TKA to
From the Rothman Institute of Orthopedics at Thomas Jefferson University, Philadelphia, Pennsylvania. Submitted February 19, 2010; accepted February 24, 2011. The Conflict of Interest statement associated with this article can be found at doi:10.1016/j.arth.2011.02.028. Reprint requests: Matthew S. Austin, MD, Department of Orthopaedic Surgery, Thomas Jefferson University Hospital, Rothman Institute, 925 Chestnut St, 5th Floor, Philadelphia, PA 19107. © 2011 Elsevier Inc. All rights reserved. 0883-5403/2608-0003$36.00/0 doi:10.1016/j.arth.2011.02.028
a posterior stabilized TKA [1-3]. Firestone and Eberle [4] demonstrated an acceptable overall improvement in function, as measured by the Knee Society scores (KSSs), after revision for instability. The type of revision for instability may affect surgical outcomes. For example, polyethylene exchange alone has been reported to have a high incidence of failure in treating instability [1,3,5]. This study was designed with 2 main objectives in mind. Primarily, the study was designed to evaluate the outcome of revision for TKA instability, as defined by resolution of symptoms and improvement in function, at a single institution. The second objective of the study was to identify factors predictive of successful revision TKA for instability.
Materials and Methods We studied a consecutive series of 68 knees (67 patients) that underwent revision TKA for tibiofemoral instability at our institution between 2000 and 2006. The patients were evaluated, and procedures were performed by 5 high-volume, fellowship-trained arthroplasty surgeons. There were 19 men and 48 women, with an average age of 66 years (range, 42-85 years) at time of revision surgery. Instability occurred after primary knee arthroplasty in 53 patients and after revision arthroplasty for instability-unrelated causes in 14 patients. Instability was diagnosed by clinical examination. Laxity in mediolateral and/or anteroposterior stress testing was present in
1139
1140 The Journal of Arthroplasty Vol. 26 No. 8 December 2011 all patients, as determined by subjective testing by the attending surgeon. In addition, all patients expressed additional symptoms, which may be associated with instability [2], including pain ascending and descending stairs, recurrent effusions, and tenderness over the pes anserine bursa. Thirty-seven patients (55%) presented with sensation of instability and more than 1 giving way episode. Infection was not clinically suspected in any patient. In addition, serologic testing including erythrocyte sedimentation rate and C-reactive protein was obtained to screen for the presence of infection in all patients. Aspiration, when clinically indicated, was performed; and infection was excluded before revision surgery. Radiographs did not reveal loose components in any patient, and this was confirmed intraoperatively. All patients presented within 2 years of their latest knee arthroplasty. Nonoperative treatment, consisting of kneebracing and muscle-strengthening exercises, was initially used in all patients. Revision surgery was undertaken only in those patients who failed 3 to 6 months of conservative treatment. Fifty-one knees (75%) had failure of a posterior stabilized (PS) TKA, 13 knees (19%) had failure of a cruciate-retaining (CR) design, and 4 knees (6%) had failure of a semiconstrained prosthesis. Instability was categorized as isolated coronal plane instability in 31 knees (45%), isolated flexion instability in 8 knees (12%), and combined in 29 knees (43%). Flexion instability was diagnosed in patients with anteroposterior laxity demonstrated with the knee flexed to 90° [6]. Under regional anesthesia, instability was confirmed by demonstration of soft tissue laxity upon knee examination. Prior incisions were used for skin exposure. A medial parapatellar approach was used for exposure in all cases. Scar tissue was excised, and the medial and lateral gutters were recreated; subsequently, the components were tested for stability. Only patients with well-fixed components were included in the study. Revision surgery entailed revision of both femoral and tibial components in 49 knees (72%), femoral component in 9 knees (13%), and polyethylene exchange alone in 10 knees (15%). Revision to PS prosthesis was performed in 17 knees; a semiconstrained implant in 47 knees; and a hinged prosthesis was used in 4 knees. The level of constraint in the revision implant was chosen based on the degree of soft tissue laxity and the extent of bone loss. Of 67 knees, 48 (72%) were revised to a more constrained prosthesis. It was unable to be determined what the prerevision level of constraint was for 2 of the knees. Modular augments were used on the femoral side in 28 knees (41%). Collateral ligament repair or advancement was not performed in any of the patients. The soft tissue balancing of the knee was assessed by the surgeon, and the knee was deemed stable at the conclusion of the procedure in all cases. Range-of-motion exercises were started on the
second postoperative day, and weight bearing as tolerated was allowed. Follow-up average was 39 months (range, 24 months to 8 years). One of the patients who had undergone revision to a hinged prosthesis died before 2-year followup, leaving 66 patients (67 knees) for final evaluation. The 2 main outcome measures were surgeon-based evaluation of knee stability and patient completed questionnaires (Short Form Health Survey [SF-36] and KSSs). Preoperative and postoperative radiographs (anteroposterior and lateral) were examined by 2 of the authors for component positioning, status of the joint line, evidence of prosthetic loosening, and overall alignment of the knee. Each of the surgeons who performed the procedure reviewed the radiographs during the time of the office visit. All of the radiographs were then re-reviewed by a second author who was blinded to the results of the revision at the time of the radiographic review. Femoral and tibial component positions were measured relative to the longitudinal axes of the femur and tibia, respectively. Joint line position was measured using the fibular head as the reference point [7] (Fig. 1). Prosthetic loosening was examined by detection and measurement of radiolucent lines in the Knee Society radiographic zones [8]. The change in femoral and tibial component position and the change in joint line level from preoperative x-rays were included in the statistical analysis. Other variables analyzed for statistical relevance included demographics (age, sex, and body mass index), clinical variables (type of instability and medical comorbidities), and surgical factors (revised components, extent of bone deficiency, amount of constraint in revision implant, and use of augments). Institutional review board approval was obtained before initiation of the study. Statistical Analysis All statistical analyses were performed using SAS version 9.1 software (SAS Institute Inc, Cary, NC). Analysis was done using 2 outcome measures, surgeon-based evaluation of knee stability (patients were classified into 2 groups according to final outcome: stable and unstable), and change in KSS (difference between KSS at latest follow-up and preoperative KSS). For each outcome measure, a univariate analysis with means and SDs for continuous variables and proportions for categorical variables was performed. The means of a continuous outcome were compared using t test (parametric) and Wilcoxon (nonparametric) test. Proportions of a categorical outcome were compared using χ2 (parametric) and Fisher exact (nonparametric) tests. P b .05 was considered to be significant. This unadjusted analysis was performed first to assess the differences in demographics, clinical, and surgical variables between patients who failed treatment and those who had a successful outcome. Multiple regression analysis was then performed
Revision of the Unstable TKA Azzam et al
1141
Fig. 1. (A) In the anteroposterior radiograph, femoral and tibial component positions were measured relative to the longitudinal axes of the femur (angle a) and tibia (angle b), respectively. Joint line position (*) was measured using the fibular head as the reference point. (B) In the lateral radiograph, femoral component flexion angle was calculated as the average of angles c and d. Tibial component slope was measured relative to the longitudinal axis of the tibia (angle e).
after adjusting for the potential confounders to determine the predictors of a successful revision arthroplasty.
Results There was a significant improvement in overall functional status as determined by SF-36 and KSSs. For the entire cohort, the Knee Society knee score improved significantly from an average of 43 points (range, 11-94 points) preoperatively to 76 points (range, 17-100 points) postoperatively (P b .001). The Knee Society function scores improved from an average of 47 points (range, 2080 points) preoperatively to 64 points (range, 20-100 points) postoperatively (P = .02). The SF-36 physical score improved significantly from an average of 40 points (range, 17-71 points) preoperatively to 53 points (range, 13-92 points) postoperatively (P = .0001). The SF-36 mental score also improved significantly from an average of 59 points (range, 26-93 points) preoperatively to 67 points (range, 25-95 points) postoperatively (P = .002). According to surgeon-based evaluation of knee stability, revision arthroplasty was successful in restoring stability and alleviating symptoms in 49 patients (50 knees). In this patient group, the Knee Society knee score improved significantly from 43 points (range, 1194 points) preoperatively to 77 points (range, 47-100 points) postoperatively (mean change, 32 points; P b .0001); and the Knee Society function scores improved significantly from an average of 45 points (range, 20-80 points) preoperatively to 58 points (range, 20-100
points) postoperatively (P = .01). Knee instability persisted in 14 patients (14 knees), accounting for a failure rate of 22%. In this group, the Knee Society knee scores improved from 35 points (range, 12-36 points) preoperatively to 58 points (range, 17-60 points) postoperatively (mean change, 23 points; P = .008); and the Knee Society function scores declined from an average of 43 points (range, 20-70 points) preoperatively to 38 points (range, 0-70 points) postoperatively (P = .5). All 14 patients had persistent pain at the time of their latest follow-up, and soft tissue laxity was demonstrated on clinical examination. Similar to the initial assessment, instability was deemed the most likely cause of persistent symptoms after exclusion of other causes of painful TKA. Three patients who had received hinged prostheses were excluded from surgeons' evaluation of knee stability as an outcome measure. Of the 14 patients (14 knees) who failed, 4 had undergone revision of both components, 4 had femoral component revision alone, and 6 knees had isolated polyethylene exchange. The failure rate of isolated polyethylene exchange was considerably high, as 60% (6/10) of the knees undergoing this procedure had evidence of instability at the latest follow-up. Thus, excluding knees with only polyethylene exchange, the failure rate of revision arthroplasty was 12.5% (8/64 knees). Of the 14 patients with residual instability, only 4 had symptoms that were severe enough to warrant further revision arthroplasty.
1142 The Journal of Arthroplasty Vol. 26 No. 8 December 2011 The results of univariate analysis revealed that revising both the femoral and tibial components, use of femoral augments, the use of semiconstrained revision implants, and smaller elevation of the joint line were important predictors of success of revision arthroplasty (Table 1). However, after the multivariate analysis, the factors that correlated significantly with achieving a stable knee at the time of the latest follow-up were revising both the femoral and tibial components (P b .001; 95% confidence interval [CI], 0.01-0.30) and the use of femoral augments (P = .015; 95% CI, 0.01-0.60). Elevation of the joint line measured radiographically correlated significantly with failure to achieve a stable TKA (P = .025; 95% CI, 1.03-1.63). However, when excluding polyethylene exchanges (which elevate the joint line), elevation of the joint line was not a significant predictor of persistent instability (P = .06; 95% CI, 0.98-1.70). This may have been due to the small sample size. When the change in KSS was used as an outcome measure (Table 2), revision of both the femoral and tibial components was found to be a predictor of improvement in knee function (P = .05). Polyethylene exchange alone
Table 1. Results of the Univariate Analysis Using SurgeonBased Assessment of Knee Stability at Follow-Up as the Outcome Measure Unstable, n = 14 Age (y), mean (SD) Male sex BMI (kg/m2), mean (SD) Only polyethylene exchange Both components revised Semiconstrained revision TKA Femoral augments Tibial augments Upsized polyethylene Rheumatoid arthritis Types II and III femoral bone loss Types II and III tibial bone loss Malaligned limb after revision Change in joint line, mean (SD) Lowered joint line Change in femoral component position, mean (SD) Change in femoral flexion, mean (SD) Correction of excessively flexed femoral component Change in tibial component position, mean (SD) Change in tibial slope, mean (SD) Correction of excessive tibial slope Change in KSS, mean (SD)
Stable, n = 50
Age Male sex BMI Only polyethylene exchange Both components revised Semiconstrained revision TKA Use of femoral augment Use of tibial augment Upsizing of polyethylene Rheumatoid arthritis Extent of femoral bone loss Extent of tibial bone loss Malaligned limb after revision Change in joint line Change in femoral component alignment Change in femoral flexion Correction of excessively flexed femoral component Change in tibial alignment Change in tibial slope Correction of excessive tibial slope
Coefficient (95% CI)
P
−0.11 (−0.72 to 0.51) −4.35 (−19.26 to 10.55) 0.94 (−0.19 to 2.07) −7.77 (−26.28 to 10.73) 15.24 (−0.16 to 30.63) 2.16 (−10.38 to 14.70) 9.23 (−4.31 to 22.77) −1.88 (−21.57 to 17.81) 9.52 (−14.23 to 33.27) −7.08 (−32.17 to 18.01) −9.42 (−24.18 to 5.35) −4.83 (−22.67 to 13.01) −9.91 (−62.40 to 42.58) −0.57 (−2.28 to 1.14) 1.74 (−1.64 to 5.12)
.726 .559 .100 .402 .052 .730 .177 .848 .421 .573 .205 .588 .703 .502 .302
−0.75 (−2.62 to 1.12) −5.95 (−28.08 to 16.19)
.417 .585
2.09 (−0.79 to 4.97) 0.66 (−1.41 to 2.72) −1.25 (−22.31 to 19.80)
.147 .517 .903
was associated with a higher incidence of persistent instability (P = .005).
P
63.50 (11.97) 6 (42.86%) 31.46 (5.19) 6 (42.86%) 4 (28.57%) 6 (42.86%) 1 (7.14%) 0 10 (90.91%) 3 (21.43%) 4 (28.57%)
65.56 (10.47) .531 13 (26.00%) .321 32.86 (7.42) .513 4 (8.00%) .005 41 (82.00%) b.001 41 (82.00%) .006 25 (50.00%) .005 7 (14.00%) .332 28 (82.35%) .663 3 (6.00%) .113 8 (19.05%) .470
2 (14.29%) 2 (18.18%) 7.42 (4.96) 0 −0.70 (1.77)
3 (7.14%) 0 3.69 (4.75) 5 (15.63%) −0.10 (2.75)
.590 .049 .057 .563 .586
6.20 (6.22)
2.60 (5.23)
.172
3 (60.00%)
9 (29.03%)
.307
−0.19 (4.23)
0.87 (3.56)
.498
4.35 (3.98)
2.24 (4.65)
.305
4 (66.67%)
8 (25.81%)
.073
22.75 (17.58) 32.28 (23.59)
Table 2. Results of the Univariate Analysis Using Change in KSS From Baseline as the Outcome Measure (Correlation of Each Variable With the Magnitude and Direction of Change in KSS)
.286
BMI indicates body mass index. Three patients who received hinged prosthesis were excluded from this analysis.
Discussion The present study on a relatively large cohort of patients undergoing revision arthroplasty for TKA instability highlights some important findings. This study demonstrated that patients undergoing revision of femoral and tibial components had a better outcome than those undergoing isolated polyethylene exchange. Revision of all components provides the opportunity for use of more constrained components, which may improve stability of the final construct. Forty-nine (72%) of the knees in this series underwent revision of all components, and these knees were more likely to have a successful result. Revision of all components also provides for the removal of previously malpositioned components that may have contributed to instability. In addition, revision of both femoral and tibial components may permit better restoration of the joint line and flexion-extension symmetry. This study also showed that the use of femoral augments was an important predictor of success. The use of metal augments facilitates proper positioning of the components and restoration of flexion and extension symmetry as well as restoring the joint line properly. This study has confirmed previous reports that isolated polyethylene exchange has a higher incidence of failure than revision of all components for patients with instability after TKA [1,3,5]. Using thicker polyethylene components results
Revision of the Unstable TKA Azzam et al
in tightening of both the extension and flexion gap without allowing alteration of the component positioning and overall alignment of the extremity. Elevation of the joint line was confirmed to increase the risk of failure of revision surgery to address TKA instability. Previous studies have shown that proper restoration of the joint line after revision arthroplasty correlates with better functional outcome [9,10]. Revision arthroplasty for TKA instability in this population of mostly posterior-stabilized prostheses resulted in a nearly 80% success rate as demonstrated in this study. This number may have been improved with routine use of intraarticular lidocaine injections as part of the routine workup for this diagnosis. The patients with continued pain after the revision arthroplasty also could have been aspirated more routinely in addition to the serologic markers to further exclude infection as an etiology of the failed revision. The current study demonstrates a considerably higher failure rate of revision arthroplasty in the management of an unstable TKA than previous reports [1-4]. The use of a broader definition of failure in the current study may account for this discrepancy. In a study by Firestone and Eberle [4], for example, reporting on 105 patients undergoing revision knee arthroplasty for instability, failure was deemed to occur when the patients required re-revision. In our cohort, re-revision was only necessary in 4 patients (6%). The other reasons relate to the type of cohort included in the study. Most patients (74%) in the series reported by Firestone and Eberle had failure of cruciate-retaining components [4], whereas only 13 (19%) of patients in our series were such patients. Revision knee arthroplasty for instability after use of a cruciate-retaining prosthesis has been shown to have a near perfect outcome [1-3]. Our series, on the other hand, included a variety of patients with varus-valgus, flexion, and combined instability. Not all the patients in our series had primary knee prosthesis in situ. There were 14 patients (21%) who had undergone previous revision surgery for causes unrelated to instability. Finally, this series included patients undergoing isolated polyethylene exchange, which is known to carry a higher incidence of failure. We deliberately did not exclude the latter patients from this series as we intended to show the success of each surgical procedure for unstable TKA, namely, isolated polyethylene exchange, single component revision, and complete revision arthroplasty. The current study has some shortcomings. The patients were diagnosed with instability after the surgeon excluded other mechanical causes of failure (infection, loosening, malalignment, and other). It is certainly plausible that some of the patients who failed revision for instability were simply misdiagnosed and that their pain was from another undiagnosed problem. A related shortcoming is that the surgeon performing the revision may be biased in evaluating the results of the revision
1143
surgery, which that surgeon performed. The surgeon may have been influenced by the clinical presentation of the patient and rated the knee as “stable” when the patient was satisfied and vice versa. This is an inherent shortcoming in any study that involves a diagnosis with a subjective measure such as “stability.” A particular deficiency of this study is that the degree of instability was not quantified in most cases. All attempts to add biomechanical testing and dynamic imaging for the assessment of instability remain experimental and not available in most centers. Therefore, this study represents a realistic situation in which the surgeon makes a determination that the TKA is unstable and revises the knee for that diagnosis. Therefore, the results of this study bear an important message to surgeons revising patients for the diagnosis of instability when other etiologies of the patient's pain have been eliminated. This study does not support the concept of revising the knee for “instability” if other etiologies of pain are excluded. The surgeon should review radiographs of the knee before the index procedure to determine if significant degenerative joint disease was present before the surgery. Injection of local anesthetic for diagnostic purposes may also be of use. Furthermore, the sample size might have provided insufficient power for some of the variables to reach statistical significance. Retrospective data collection could have introduced recall bias pertinent to certain variables. One of the parameters to evaluate outcome, namely, surgeon-based assessment of instability, was subjective and difficult to measure. This assessment, however, correlated well with the change in KSS from baseline and allowed us to divide patients into 2 groups so that meaningful statistical analysis could be performed. Despite the aforementioned shortfalls, the findings of this study merit special attention. The results of this study suggest that revision surgery for an unstable TKA should focus on treating the cause of instability with attention to restoration of joint line, balancing the flexion and extension gaps, and proper positioning of the components. Revision of both components seems to offer the most predictable outcome. The use of augments, resulting in better restoration of the joint line and condylar profile, was shown to provide more success in treating instability after TKA.
References 1. Pagnano MW, Hanssen AD, Lewallen DG, et al. Flexion instability after primary posterior cruciate retaining total knee arthroplasty. Clin Orthop Relat Res 1998;39. 2. Schwab JH, Haidukewych GJ, Hanssen AD, et al. Flexion instability without dislocation after posterior stabilized total knees. Clin Orthop Relat Res 2005;440:96. 3. Waslewski GL, Marson BM, Benjamin JB. Early, incapacitating instability of posterior cruciate ligament–retaining total knee arthroplasty. J Arthroplasty 1998;13:763.
1144 The Journal of Arthroplasty Vol. 26 No. 8 December 2011 4. Firestone TP, Eberle RW. Surgical management of symptomatic instability following failed primary total knee replacement. J Bone Joint Surg Am 2006;88(Suppl 4):80. 5. Babis GC, Trousdale RT, Morrey BF. The effectiveness of isolated tibial insert exchange in revision total knee arthroplasty. J Bone Joint Surg Am 2002;84-A:64. 6. Parratte S, Pagnano MW. Instability after total knee arthroplasty. Instr Course Lect 2008;57:295. 7. Kapandji IA. The physiology of the joints. Lower limb. New York: Churchill Livingston; 1978.
8. Ewald FC. The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop Relat Res 1989;9. 9. Hofmann AA, Kurtin SM, Lyons S, et al. Clinical and radiographic analysis of accurate restoration of the joint line in revision total knee arthroplasty. J Arthroplasty 2006;21:1154. 10. Figgie III HE, Goldberg VM, Heiple KG, et al. The influence of tibial-patellofemoral location on function of the knee in patients with the posterior stabilized condylar knee prosthesis. J Bone Joint Surg Am 1986;68:1035.
Revision of the Unstable Total Knee Arthroplasty Outcome Predictors Khalid Azzam, MD, Javad Parvizi, MD, FRCS, Daniel Kaufman, BS, James J. Purtill, MD, Peter F. Sharkey, MD, and Matthew S. Austin, MD
Abstract: Revision arthroplasty is often required for confirmed cases of symptomatic tibiofemoral instability after total knee arthroplasty (TKA). We evaluated the results of revision for TKA instability in a consecutive series of 67 patients (68 knees) between 2000 and 2006. Outcome measures were surgeon-based assessment of knee stability, Knee Society Score, and Short Form Health Survey 36. At an average of 39 months of follow-up, the mean Knee Society Score and Short Form Health Survey 36 physical and mental scores were 76, 53, and 67 points, respectively. Knee instability persisted in 14 patients (22%). Data at the 95% confidence level revealed that revising both the femoral and tibial components, the use of femoral augments, and smaller joint line elevation as measured on radiographs correlated significantly with achieving a stable knee. In revision surgery for TKA instability, revision of both components and the use of femoral augments seem to offer the most predictable outcome. Keywords: revision knee arthroplasty, instability, outcome. © 2011 Elsevier Inc. All rights reserved.
Instability is a well-recognized cause of poor outcome after total knee arthroplasty (TKA). Causes of instability after TKA include inadequate soft tissue balancing, loss of ligamentous integrity, component wear, improper component sizing, and component malpositioning. It is often difficult to establish a diagnosis. Persistent or recurrent pain, especially during the first few years after surgery, accompanied by demonstration of soft tissue laxity on clinical examination can be a marker of an unstable TKA. However, the unstable TKA may present with pain without objectively demonstrable signs of instability. Therefore, exclusion of other etiologies of failure, such as infection or loosening, in all patients presenting with painful TKA is prudent. There are few studies reporting the outcome of revision arthroplasty for instability. Three studies have reported successful revision of cruciate-retaining TKA to
From the Rothman Institute of Orthopedics at Thomas Jefferson University, Philadelphia, Pennsylvania. Submitted February 19, 2010; accepted February 24, 2011. The Conflict of Interest statement associated with this article can be found at doi:10.1016/j.arth.2011.02.028. Reprint requests: Matthew S. Austin, MD, Department of Orthopaedic Surgery, Thomas Jefferson University Hospital, Rothman Institute, 925 Chestnut St, 5th Floor, Philadelphia, PA 19107. © 2011 Elsevier Inc. All rights reserved. 0883-5403/2608-0003$36.00/0 doi:10.1016/j.arth.2011.02.028
a posterior stabilized TKA [1-3]. Firestone and Eberle [4] demonstrated an acceptable overall improvement in function, as measured by the Knee Society scores (KSSs), after revision for instability. The type of revision for instability may affect surgical outcomes. For example, polyethylene exchange alone has been reported to have a high incidence of failure in treating instability [1,3,5]. This study was designed with 2 main objectives in mind. Primarily, the study was designed to evaluate the outcome of revision for TKA instability, as defined by resolution of symptoms and improvement in function, at a single institution. The second objective of the study was to identify factors predictive of successful revision TKA for instability.
Materials and Methods We studied a consecutive series of 68 knees (67 patients) that underwent revision TKA for tibiofemoral instability at our institution between 2000 and 2006. The patients were evaluated, and procedures were performed by 5 high-volume, fellowship-trained arthroplasty surgeons. There were 19 men and 48 women, with an average age of 66 years (range, 42-85 years) at time of revision surgery. Instability occurred after primary knee arthroplasty in 53 patients and after revision arthroplasty for instability-unrelated causes in 14 patients. Instability was diagnosed by clinical examination. Laxity in mediolateral and/or anteroposterior stress testing was present in
1139
1140 The Journal of Arthroplasty Vol. 26 No. 8 December 2011 all patients, as determined by subjective testing by the attending surgeon. In addition, all patients expressed additional symptoms, which may be associated with instability [2], including pain ascending and descending stairs, recurrent effusions, and tenderness over the pes anserine bursa. Thirty-seven patients (55%) presented with sensation of instability and more than 1 giving way episode. Infection was not clinically suspected in any patient. In addition, serologic testing including erythrocyte sedimentation rate and C-reactive protein was obtained to screen for the presence of infection in all patients. Aspiration, when clinically indicated, was performed; and infection was excluded before revision surgery. Radiographs did not reveal loose components in any patient, and this was confirmed intraoperatively. All patients presented within 2 years of their latest knee arthroplasty. Nonoperative treatment, consisting of kneebracing and muscle-strengthening exercises, was initially used in all patients. Revision surgery was undertaken only in those patients who failed 3 to 6 months of conservative treatment. Fifty-one knees (75%) had failure of a posterior stabilized (PS) TKA, 13 knees (19%) had failure of a cruciate-retaining (CR) design, and 4 knees (6%) had failure of a semiconstrained prosthesis. Instability was categorized as isolated coronal plane instability in 31 knees (45%), isolated flexion instability in 8 knees (12%), and combined in 29 knees (43%). Flexion instability was diagnosed in patients with anteroposterior laxity demonstrated with the knee flexed to 90° [6]. Under regional anesthesia, instability was confirmed by demonstration of soft tissue laxity upon knee examination. Prior incisions were used for skin exposure. A medial parapatellar approach was used for exposure in all cases. Scar tissue was excised, and the medial and lateral gutters were recreated; subsequently, the components were tested for stability. Only patients with well-fixed components were included in the study. Revision surgery entailed revision of both femoral and tibial components in 49 knees (72%), femoral component in 9 knees (13%), and polyethylene exchange alone in 10 knees (15%). Revision to PS prosthesis was performed in 17 knees; a semiconstrained implant in 47 knees; and a hinged prosthesis was used in 4 knees. The level of constraint in the revision implant was chosen based on the degree of soft tissue laxity and the extent of bone loss. Of 67 knees, 48 (72%) were revised to a more constrained prosthesis. It was unable to be determined what the prerevision level of constraint was for 2 of the knees. Modular augments were used on the femoral side in 28 knees (41%). Collateral ligament repair or advancement was not performed in any of the patients. The soft tissue balancing of the knee was assessed by the surgeon, and the knee was deemed stable at the conclusion of the procedure in all cases. Range-of-motion exercises were started on the
second postoperative day, and weight bearing as tolerated was allowed. Follow-up average was 39 months (range, 24 months to 8 years). One of the patients who had undergone revision to a hinged prosthesis died before 2-year followup, leaving 66 patients (67 knees) for final evaluation. The 2 main outcome measures were surgeon-based evaluation of knee stability and patient completed questionnaires (Short Form Health Survey [SF-36] and KSSs). Preoperative and postoperative radiographs (anteroposterior and lateral) were examined by 2 of the authors for component positioning, status of the joint line, evidence of prosthetic loosening, and overall alignment of the knee. Each of the surgeons who performed the procedure reviewed the radiographs during the time of the office visit. All of the radiographs were then re-reviewed by a second author who was blinded to the results of the revision at the time of the radiographic review. Femoral and tibial component positions were measured relative to the longitudinal axes of the femur and tibia, respectively. Joint line position was measured using the fibular head as the reference point [7] (Fig. 1). Prosthetic loosening was examined by detection and measurement of radiolucent lines in the Knee Society radiographic zones [8]. The change in femoral and tibial component position and the change in joint line level from preoperative x-rays were included in the statistical analysis. Other variables analyzed for statistical relevance included demographics (age, sex, and body mass index), clinical variables (type of instability and medical comorbidities), and surgical factors (revised components, extent of bone deficiency, amount of constraint in revision implant, and use of augments). Institutional review board approval was obtained before initiation of the study. Statistical Analysis All statistical analyses were performed using SAS version 9.1 software (SAS Institute Inc, Cary, NC). Analysis was done using 2 outcome measures, surgeon-based evaluation of knee stability (patients were classified into 2 groups according to final outcome: stable and unstable), and change in KSS (difference between KSS at latest follow-up and preoperative KSS). For each outcome measure, a univariate analysis with means and SDs for continuous variables and proportions for categorical variables was performed. The means of a continuous outcome were compared using t test (parametric) and Wilcoxon (nonparametric) test. Proportions of a categorical outcome were compared using χ2 (parametric) and Fisher exact (nonparametric) tests. P b .05 was considered to be significant. This unadjusted analysis was performed first to assess the differences in demographics, clinical, and surgical variables between patients who failed treatment and those who had a successful outcome. Multiple regression analysis was then performed
Revision of the Unstable TKA Azzam et al
1141
Fig. 1. (A) In the anteroposterior radiograph, femoral and tibial component positions were measured relative to the longitudinal axes of the femur (angle a) and tibia (angle b), respectively. Joint line position (*) was measured using the fibular head as the reference point. (B) In the lateral radiograph, femoral component flexion angle was calculated as the average of angles c and d. Tibial component slope was measured relative to the longitudinal axis of the tibia (angle e).
after adjusting for the potential confounders to determine the predictors of a successful revision arthroplasty.
Results There was a significant improvement in overall functional status as determined by SF-36 and KSSs. For the entire cohort, the Knee Society knee score improved significantly from an average of 43 points (range, 11-94 points) preoperatively to 76 points (range, 17-100 points) postoperatively (P b .001). The Knee Society function scores improved from an average of 47 points (range, 2080 points) preoperatively to 64 points (range, 20-100 points) postoperatively (P = .02). The SF-36 physical score improved significantly from an average of 40 points (range, 17-71 points) preoperatively to 53 points (range, 13-92 points) postoperatively (P = .0001). The SF-36 mental score also improved significantly from an average of 59 points (range, 26-93 points) preoperatively to 67 points (range, 25-95 points) postoperatively (P = .002). According to surgeon-based evaluation of knee stability, revision arthroplasty was successful in restoring stability and alleviating symptoms in 49 patients (50 knees). In this patient group, the Knee Society knee score improved significantly from 43 points (range, 1194 points) preoperatively to 77 points (range, 47-100 points) postoperatively (mean change, 32 points; P b .0001); and the Knee Society function scores improved significantly from an average of 45 points (range, 20-80 points) preoperatively to 58 points (range, 20-100
points) postoperatively (P = .01). Knee instability persisted in 14 patients (14 knees), accounting for a failure rate of 22%. In this group, the Knee Society knee scores improved from 35 points (range, 12-36 points) preoperatively to 58 points (range, 17-60 points) postoperatively (mean change, 23 points; P = .008); and the Knee Society function scores declined from an average of 43 points (range, 20-70 points) preoperatively to 38 points (range, 0-70 points) postoperatively (P = .5). All 14 patients had persistent pain at the time of their latest follow-up, and soft tissue laxity was demonstrated on clinical examination. Similar to the initial assessment, instability was deemed the most likely cause of persistent symptoms after exclusion of other causes of painful TKA. Three patients who had received hinged prostheses were excluded from surgeons' evaluation of knee stability as an outcome measure. Of the 14 patients (14 knees) who failed, 4 had undergone revision of both components, 4 had femoral component revision alone, and 6 knees had isolated polyethylene exchange. The failure rate of isolated polyethylene exchange was considerably high, as 60% (6/10) of the knees undergoing this procedure had evidence of instability at the latest follow-up. Thus, excluding knees with only polyethylene exchange, the failure rate of revision arthroplasty was 12.5% (8/64 knees). Of the 14 patients with residual instability, only 4 had symptoms that were severe enough to warrant further revision arthroplasty.
1142 The Journal of Arthroplasty Vol. 26 No. 8 December 2011 The results of univariate analysis revealed that revising both the femoral and tibial components, use of femoral augments, the use of semiconstrained revision implants, and smaller elevation of the joint line were important predictors of success of revision arthroplasty (Table 1). However, after the multivariate analysis, the factors that correlated significantly with achieving a stable knee at the time of the latest follow-up were revising both the femoral and tibial components (P b .001; 95% confidence interval [CI], 0.01-0.30) and the use of femoral augments (P = .015; 95% CI, 0.01-0.60). Elevation of the joint line measured radiographically correlated significantly with failure to achieve a stable TKA (P = .025; 95% CI, 1.03-1.63). However, when excluding polyethylene exchanges (which elevate the joint line), elevation of the joint line was not a significant predictor of persistent instability (P = .06; 95% CI, 0.98-1.70). This may have been due to the small sample size. When the change in KSS was used as an outcome measure (Table 2), revision of both the femoral and tibial components was found to be a predictor of improvement in knee function (P = .05). Polyethylene exchange alone
Table 1. Results of the Univariate Analysis Using SurgeonBased Assessment of Knee Stability at Follow-Up as the Outcome Measure Unstable, n = 14 Age (y), mean (SD) Male sex BMI (kg/m2), mean (SD) Only polyethylene exchange Both components revised Semiconstrained revision TKA Femoral augments Tibial augments Upsized polyethylene Rheumatoid arthritis Types II and III femoral bone loss Types II and III tibial bone loss Malaligned limb after revision Change in joint line, mean (SD) Lowered joint line Change in femoral component position, mean (SD) Change in femoral flexion, mean (SD) Correction of excessively flexed femoral component Change in tibial component position, mean (SD) Change in tibial slope, mean (SD) Correction of excessive tibial slope Change in KSS, mean (SD)
Stable, n = 50
Age Male sex BMI Only polyethylene exchange Both components revised Semiconstrained revision TKA Use of femoral augment Use of tibial augment Upsizing of polyethylene Rheumatoid arthritis Extent of femoral bone loss Extent of tibial bone loss Malaligned limb after revision Change in joint line Change in femoral component alignment Change in femoral flexion Correction of excessively flexed femoral component Change in tibial alignment Change in tibial slope Correction of excessive tibial slope
Coefficient (95% CI)
P
−0.11 (−0.72 to 0.51) −4.35 (−19.26 to 10.55) 0.94 (−0.19 to 2.07) −7.77 (−26.28 to 10.73) 15.24 (−0.16 to 30.63) 2.16 (−10.38 to 14.70) 9.23 (−4.31 to 22.77) −1.88 (−21.57 to 17.81) 9.52 (−14.23 to 33.27) −7.08 (−32.17 to 18.01) −9.42 (−24.18 to 5.35) −4.83 (−22.67 to 13.01) −9.91 (−62.40 to 42.58) −0.57 (−2.28 to 1.14) 1.74 (−1.64 to 5.12)
.726 .559 .100 .402 .052 .730 .177 .848 .421 .573 .205 .588 .703 .502 .302
−0.75 (−2.62 to 1.12) −5.95 (−28.08 to 16.19)
.417 .585
2.09 (−0.79 to 4.97) 0.66 (−1.41 to 2.72) −1.25 (−22.31 to 19.80)
.147 .517 .903
was associated with a higher incidence of persistent instability (P = .005).
P
63.50 (11.97) 6 (42.86%) 31.46 (5.19) 6 (42.86%) 4 (28.57%) 6 (42.86%) 1 (7.14%) 0 10 (90.91%) 3 (21.43%) 4 (28.57%)
65.56 (10.47) .531 13 (26.00%) .321 32.86 (7.42) .513 4 (8.00%) .005 41 (82.00%) b.001 41 (82.00%) .006 25 (50.00%) .005 7 (14.00%) .332 28 (82.35%) .663 3 (6.00%) .113 8 (19.05%) .470
2 (14.29%) 2 (18.18%) 7.42 (4.96) 0 −0.70 (1.77)
3 (7.14%) 0 3.69 (4.75) 5 (15.63%) −0.10 (2.75)
.590 .049 .057 .563 .586
6.20 (6.22)
2.60 (5.23)
.172
3 (60.00%)
9 (29.03%)
.307
−0.19 (4.23)
0.87 (3.56)
.498
4.35 (3.98)
2.24 (4.65)
.305
4 (66.67%)
8 (25.81%)
.073
22.75 (17.58) 32.28 (23.59)
Table 2. Results of the Univariate Analysis Using Change in KSS From Baseline as the Outcome Measure (Correlation of Each Variable With the Magnitude and Direction of Change in KSS)
.286
BMI indicates body mass index. Three patients who received hinged prosthesis were excluded from this analysis.
Discussion The present study on a relatively large cohort of patients undergoing revision arthroplasty for TKA instability highlights some important findings. This study demonstrated that patients undergoing revision of femoral and tibial components had a better outcome than those undergoing isolated polyethylene exchange. Revision of all components provides the opportunity for use of more constrained components, which may improve stability of the final construct. Forty-nine (72%) of the knees in this series underwent revision of all components, and these knees were more likely to have a successful result. Revision of all components also provides for the removal of previously malpositioned components that may have contributed to instability. In addition, revision of both femoral and tibial components may permit better restoration of the joint line and flexion-extension symmetry. This study also showed that the use of femoral augments was an important predictor of success. The use of metal augments facilitates proper positioning of the components and restoration of flexion and extension symmetry as well as restoring the joint line properly. This study has confirmed previous reports that isolated polyethylene exchange has a higher incidence of failure than revision of all components for patients with instability after TKA [1,3,5]. Using thicker polyethylene components results
Revision of the Unstable TKA Azzam et al
in tightening of both the extension and flexion gap without allowing alteration of the component positioning and overall alignment of the extremity. Elevation of the joint line was confirmed to increase the risk of failure of revision surgery to address TKA instability. Previous studies have shown that proper restoration of the joint line after revision arthroplasty correlates with better functional outcome [9,10]. Revision arthroplasty for TKA instability in this population of mostly posterior-stabilized prostheses resulted in a nearly 80% success rate as demonstrated in this study. This number may have been improved with routine use of intraarticular lidocaine injections as part of the routine workup for this diagnosis. The patients with continued pain after the revision arthroplasty also could have been aspirated more routinely in addition to the serologic markers to further exclude infection as an etiology of the failed revision. The current study demonstrates a considerably higher failure rate of revision arthroplasty in the management of an unstable TKA than previous reports [1-4]. The use of a broader definition of failure in the current study may account for this discrepancy. In a study by Firestone and Eberle [4], for example, reporting on 105 patients undergoing revision knee arthroplasty for instability, failure was deemed to occur when the patients required re-revision. In our cohort, re-revision was only necessary in 4 patients (6%). The other reasons relate to the type of cohort included in the study. Most patients (74%) in the series reported by Firestone and Eberle had failure of cruciate-retaining components [4], whereas only 13 (19%) of patients in our series were such patients. Revision knee arthroplasty for instability after use of a cruciate-retaining prosthesis has been shown to have a near perfect outcome [1-3]. Our series, on the other hand, included a variety of patients with varus-valgus, flexion, and combined instability. Not all the patients in our series had primary knee prosthesis in situ. There were 14 patients (21%) who had undergone previous revision surgery for causes unrelated to instability. Finally, this series included patients undergoing isolated polyethylene exchange, which is known to carry a higher incidence of failure. We deliberately did not exclude the latter patients from this series as we intended to show the success of each surgical procedure for unstable TKA, namely, isolated polyethylene exchange, single component revision, and complete revision arthroplasty. The current study has some shortcomings. The patients were diagnosed with instability after the surgeon excluded other mechanical causes of failure (infection, loosening, malalignment, and other). It is certainly plausible that some of the patients who failed revision for instability were simply misdiagnosed and that their pain was from another undiagnosed problem. A related shortcoming is that the surgeon performing the revision may be biased in evaluating the results of the revision
1143
surgery, which that surgeon performed. The surgeon may have been influenced by the clinical presentation of the patient and rated the knee as “stable” when the patient was satisfied and vice versa. This is an inherent shortcoming in any study that involves a diagnosis with a subjective measure such as “stability.” A particular deficiency of this study is that the degree of instability was not quantified in most cases. All attempts to add biomechanical testing and dynamic imaging for the assessment of instability remain experimental and not available in most centers. Therefore, this study represents a realistic situation in which the surgeon makes a determination that the TKA is unstable and revises the knee for that diagnosis. Therefore, the results of this study bear an important message to surgeons revising patients for the diagnosis of instability when other etiologies of the patient's pain have been eliminated. This study does not support the concept of revising the knee for “instability” if other etiologies of pain are excluded. The surgeon should review radiographs of the knee before the index procedure to determine if significant degenerative joint disease was present before the surgery. Injection of local anesthetic for diagnostic purposes may also be of use. Furthermore, the sample size might have provided insufficient power for some of the variables to reach statistical significance. Retrospective data collection could have introduced recall bias pertinent to certain variables. One of the parameters to evaluate outcome, namely, surgeon-based assessment of instability, was subjective and difficult to measure. This assessment, however, correlated well with the change in KSS from baseline and allowed us to divide patients into 2 groups so that meaningful statistical analysis could be performed. Despite the aforementioned shortfalls, the findings of this study merit special attention. The results of this study suggest that revision surgery for an unstable TKA should focus on treating the cause of instability with attention to restoration of joint line, balancing the flexion and extension gaps, and proper positioning of the components. Revision of both components seems to offer the most predictable outcome. The use of augments, resulting in better restoration of the joint line and condylar profile, was shown to provide more success in treating instability after TKA.
References 1. Pagnano MW, Hanssen AD, Lewallen DG, et al. Flexion instability after primary posterior cruciate retaining total knee arthroplasty. Clin Orthop Relat Res 1998;39. 2. Schwab JH, Haidukewych GJ, Hanssen AD, et al. Flexion instability without dislocation after posterior stabilized total knees. Clin Orthop Relat Res 2005;440:96. 3. Waslewski GL, Marson BM, Benjamin JB. Early, incapacitating instability of posterior cruciate ligament–retaining total knee arthroplasty. J Arthroplasty 1998;13:763.
1144 The Journal of Arthroplasty Vol. 26 No. 8 December 2011 4. Firestone TP, Eberle RW. Surgical management of symptomatic instability following failed primary total knee replacement. J Bone Joint Surg Am 2006;88(Suppl 4):80. 5. Babis GC, Trousdale RT, Morrey BF. The effectiveness of isolated tibial insert exchange in revision total knee arthroplasty. J Bone Joint Surg Am 2002;84-A:64. 6. Parratte S, Pagnano MW. Instability after total knee arthroplasty. Instr Course Lect 2008;57:295. 7. Kapandji IA. The physiology of the joints. Lower limb. New York: Churchill Livingston; 1978.
8. Ewald FC. The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop Relat Res 1989;9. 9. Hofmann AA, Kurtin SM, Lyons S, et al. Clinical and radiographic analysis of accurate restoration of the joint line in revision total knee arthroplasty. J Arthroplasty 2006;21:1154. 10. Figgie III HE, Goldberg VM, Heiple KG, et al. The influence of tibial-patellofemoral location on function of the knee in patients with the posterior stabilized condylar knee prosthesis. J Bone Joint Surg Am 1986;68:1035.