Abstract
Background
Dissatisfaction following primary total knee arthroplasty (pTKA) is higher in patients with preoperative low-grade osteoarthritis (OA), but its impact on revision risk remains unclear.
Questions/purposes
This study aimed to identify predictors for revision surgery within 10 years following pTKA, focusing on radiographic severity of OA using the Kellgren-Lawrence (KL) grading system, along with other demographic and radiographic factors.
Patients and methods
This retrospective case-control study was conducted at two European tertiary referral centers. The case group included 142 patients who underwent aseptic revision total knee arthroplasty (rTKA) between 2007 and 2023, within 10 years following pTKA. A 2:1 control group of 284 patients who had pTKA between 2011 and 2014, with no revision surgery within 10 years, was selected. Collected data included age, sex, body mass index (BMI), side of surgery, and American Society of Anesthesiologists (ASA) classification. Radiographic data were collected, requiring a preoperative radiograph to assess the KL grade. Univariate analyses identified potential predictors for rTKA, and multivariable logistic regression assessed their relationship with revision risk.
Results
Patients requiring rTKA were significantly younger at the time of pTKA. For each additional year of age, the revision risk decreased by 6% (OR 0.94, 95% CI 0.92 to 0.97; p < 0.001). Men were at a fourfold higher risk for revision compared to women (OR 4.01, 95% CI 2.41 to 6.69; p < 0.001). Patients with KL grade 2 OA were ten times more likely to need revision compared to those with KL grade 4 (OR 10.12, 95% CI 5.19 to 19.74; p < 0.001).
Conclusions
Revision surgery risk following pTKA was associated with younger age, male sex, and less severe OA. These results emphasize the importance of appropriate patient selection for pTKA, enabling better identification of high-risk patients and improving preoperative counselling to optimize long-term outcomes.
Level of evidence
III.
Keywords : Total Knee Artrhoplasty , Indication , Dissatisfaction , Outcome , Revision
Introduction
Total knee arthroplasty (TKA) is widely recognized as an effective treatment for advanced osteoarthritis (OA), particularly for alleviating pain, restoring function, and improving quality of life in patients suffering from debilitating knee conditions [17, 24, 26, 34]. The favorable clinical outcomes, along with the aging population, unrelenting challenge of obesity, and expanding use to younger patients, have collectively contributed to the growing number of TKA procedures [17, 19, 30]. Recent projections suggest a 143% increase in TKA incidence in the United States by 2050 [17]. This trend raises concerns, as approximately 10% (range, 5%−16.3%) of the patients remained dissatisfied postoperatively, as recently reported by DeFrance and Scuderi [11].
In parallel, the volume of revision TKA (rTKA) is escalating even more rapidly than that of primary TKA (pTKA), with projections indicating a 149% rise by 2040 and a 520% increase by 2060 in the United States [31]. Furthermore, rTKA outcomes are generally less favorable than those of pTKA, with elevated risks of complications and mortality, as well as a higher financial burden and greater resource utilization [33]. These challenges place considerable pressure on both patients and healthcare systems [7, 36, 39]. Therefore, identifying risk factors for rTKA is instrumental, as well as to understand the underlying causes of failure. This dual approach is essential not only for developing strategies to mitigate the occurrence of revision surgeries, but also to improve patient care, inform targeted healthcare policies, and reduce the economic strain on healthcare resources [33].
Several studies have identified key factors affecting patient satisfaction following primary TKA, including sociodemographic, preoperative, intraoperative and postoperative factors [13, 37]. Notably, the severity of preoperative OA has emerged as a critical determinant of post-TKA outcomes, with patients presenting mild OA often experiencing worse clinical results than those with end-stage OA before their initial TKA [16, 28, 34]. Although end-stage OA remains a prerequisite for TKA, a subset of patients with low-grade OA continue to undergo premature surgery based on symptoms rather than radiographic severity [22]. It should be kept in mind that operating on mild OA can lead to patient dissatisfaction and remains prevalent in current orthopaedic practice, as recently reported by Tampere et al. [34].
Low-grade OA appears to be a determinant for the outcome after pTKA, as measured by Patient-Reported Outcome Measures (PROMS) [19]. However, no studies have yet explored the impact of low-grade OA on the revision risk. We hypothesize that patients with low-grade OA prior to their index TKA may undergo further investigations as they are more prone to dissatisfaction, thereby potentially increasing their likelihood of receiving revision surgery [14]. The objectives of this study were twofold: (1) Identifying the predictors for the need for revision surgery within 10 years following pTKA, with primary focus on the radiographic assessment of OA severity using the Kellgren-Lawrence (KL) grading system. (2) Analyzing whether the indications for revision differ across various degrees of OA.
Method
Study design
This was a retrospective case-control study across two European hospitals, both serving as tertiary referral centers for knee arthroplasty. Ethical approval was obtained from the institutional review board of one participating center prior to data collection (Ghent University Hospital). No ethical approval was required for the institution in the Netherlands. A schematic representation of the study is provided in Fig. 1.
Fig. 1.
Study overview. OA osteoarthritis. pTKA primary total knee arthroplasty. T time
Patient selection
The cases included patients who underwent rTKA within 10 years following pTKA, as most failures occur within this time frame. Revision TKA was defined as the revision of at least the tibial or femoral component of a pTKA. The following exclusion criteria were applied: isolated patellar revision or polyethylene exchange, revision of a patellofemoral (PF) prosthesis or unicompartmental arthroplasty (UKA), prior revision surgery, and revisions due to infection, fracture, or oncological diagnosis, as these typically involve different clinical trajectories and surgical considerations, which are not representative for mechanical failure or pain-related failure. A cutoff year of 2007 was chosen for pTKA to ensure the inclusion of recent data. Of the 519 patients who were screened at the Ghent University Hospital between January 2013 and December 2023, 65 patients were eligible for this study. Of these 65 patients, 58 (89.2%) were referred from another hospital with pTKA performed elsewhere. At Sint Maartenskliniek Nijmegen (SMK), 705 patients were screened between January 2007 and December 2023, resulting in a final selection of 77 patients, all of whom had their pTKA performed at SMK.
The control group consisted of patients who underwent pTKA with or without patellar resurfacing for OA and had no revision surgery (including isolated patellar revision) within 10 years postoperatively. Patients with UKA or PF prostheses were excluded, as well as those who died within the 10-year follow-up period. Out of the 174 patients who were operated on for primary TKA between October 2011 and December 2014 by two orthopaedic knee surgeons (JV, NA) at Ghent University Hospital, 110 patients met the inclusion criteria and were selected as controls for this study. To achieve a control group with twice as many controls as cases, 174 consecutive patients were selected based on sorted, randomly generated numbers out of a database of 1970 patients who underwent pTKA at SMK between January 2011 and December 2013. All pTKAs at SMK were performed by senior orthopaedic surgeons. Patient characteristics of the two groups are summarized in Table 1.
Table 1.
Patient characteristics of cases (revision within 10 years postoperatively) and controls (no revision within 10 years postoperatively)
| Variable | Cases (n = 142) | Controls (n = 284) | P-value | ||||||
|---|---|---|---|---|---|---|---|---|---|
| All cases | UZ Ghent | SMK | P-value | All controls | UZ Ghent | SMK | P-value | ||
| Patients | 142 | 65 | 77 | 284 | 110 | 174 | |||
| Age at time of pTKA (years) | 59 ± 9.5 | 57.6 ± 9.2 | 60.2 ± 9.6 | 0.111‡ | 64.7 ± 8.6 | 65.1 ± 9.2 | 64.5 ± 8.1 | 0.577‡ | < 0.001‡ |
| Sex | 0.775* | < 0.001* | < 0.001* | ||||||
| Women | 42 (29.6) | 20 (30.8) | 22 (28.6) | 167 (58.8) | 47 (42.7) | 120 (69) | |||
| Men | 100 (70.4) | 45 (69.2) | 55 (71.4) | 117 (41.2) | 63 (57.3) | 54 (31) | |||
| BMI (kg/m²) | 29.1 [26.4–33.9] (n = 136) | 27.6 [25.3–32.4] (n = 65) | 30.4 [27.3–35.5] (n = 71) | 0.006† | 29.4 [26.8–32.9] (n = 245) | 29.2 [26.7–32.3] (n = 102) | 29.9 [26.9–33.3] (n = 143) | 0.162† | 0.696† |
| ASA classification | (n = 134) | (n = 63) | (n = 71) | 0.798* | (n = 279) | (n = 107) | (n = 172) | 0.275* | 0.259* |
| I | 27 (20.1) | 14 (22.2) | 13 (18.3) | 77 (27.6) | 24 (22.4) | 53 (30.8) | |||
| II | 91 (67.9) | 41 (65.1) | 50 (70.4) | 178 (63.8) | 72 (67.3) | 106 (61.6) | |||
| III | 16 (11.9) | 8 (12.7) | 8 (11.3) | 23 (8.2) | 10 (9.3) | 13 (7.6) | |||
| IV | 0 (0) | 0 (0) | 0 (0) | 1 (0.4) | 1 (0.9) | 0 (0) | |||
| Side | 0.173* | 0.819* | 0.631* | ||||||
| Left | 72 (50.7) | 37 (56.9) | 35 (45.5) | 137 (48.2) | 54 (49.1) | 83 (47.7) | |||
| Right | 70 (49.3) | 28 (43.1) | 42 (54.5) | 147 (51.8) | 56 (50.9) | 91 (52.3) | |||
Data are presented as mean ± SD or median [IQR], based on normality assumptions, for continuous variables. Categorical variables are presented as absolute number (percentage). In case of missing variables, the precise number is indicated in parentheses. BMI body mass index. Kg kilogram. M, meter. ASA classification of the American Society of Anesthesiologists. ‡ Independent Student T-Test. †Mann-Whitney U-Test. *Chi-square test. The p-value at the far right of the table represents the comparison between cases and controls. The p-value listed under cases and controls individually reflects the comparison between the two centers
UZ Ghent, University Hospital Ghent. SMK, Sint Maartenskliniek Nijmegen
To be eligible for inclusion, radiographic images had to be obtained within two years prior to pTKA for both cases and controls, and the KL score had to be assessable. A flowchart for patient inclusion is presented in Fig. 2.
Fig. 2.
Flowchart of patient inclusion. Exclusion criteria are presented in hierarchical order for each center. pTKA primary total knee arthroplasty. rTKA revision total knee arthroplasty. UKA unicompartmental arthroplasty. PF patellofemoral
Data collection
A thorough review of each patient’s medical record was conducted. The following data were systematically extracted: age at time of pTKA (years), Body Mass Index (BMI, kg/m²), sex, American Society of Anesthesiologists (ASA) classification, and the side of the procedure. For revision cases, additional information was collected from the medical records, including the recorded indication for revision, age at the time of revision (years), and the time interval between the primary and revision procedure (months). The indications for revision were classified into six categories: aseptic loosening, instability, malalignment, arthrofibrosis, pain and other causes. The failure mechanism was determined based on preoperative and perioperative findings, with the definitions of aseptic loosening, instability, arthrofibrosis, and malalignment derived from those outlined in previous publications [35, 38]. Cases not fitting these definitions were categorized as ‘other causes’. In Ghent University Hospital, ‘pain’ was assigned when no other cause for the revision was found in the medical records, but pain was reported as the indication.
Radiological evaluation
When multiple images were available, the one closest to the pTKA date was selected.
Radiological evidence of tibiofemoral OA was evaluated using the KL grading system based on weight-bearing radiographs [36]. The availability of radiograph types was assessed in the following hierarchical order: fixed flexion view (FFV), anteroposterior (AP) view, and full-leg (FL) radiographs (Appendix 1). The KL grade was assigned based on the available view, in accordance with this hierarchy. When a DICOM image was available, the KL score was determined using fully-automated image processing software (Knee OsteoArthritis Labeling Assistant (KOALA), ImageBiopsy Lab, Vienna, Austria) [6]. In absence of a DICOM file, two senior orthopaedic residents (LS, HV) independently evaluated the radiographs. In the case of disagreement, a third senior orthopaedic resident (TA) made the final decision.
The skyline view was assessed for radiological evidence of patellofemoral OA using the Sperner classification [1]. Two senior orthopaedic residents (LS, HV) independently evaluated all the radiographs. In the event of disagreement, a third senior orthopaedic resident (TA) provided the final judgement.
Coronal alignment was assessed from FL radiographs, following Paley’s criteria [8]. The following parameters were calculated: the mechanical hip-knee-ankle angle (mHKA), mechanical lateral distal femoral angle (mLDFA), and the mechanical medial proximal tibial angle (mMPTA). The HKA was defined as the angle formed by the mechanical femoral axis (from the center of the femoral head to the intercondylar notch of the distal femur) and the mechanical tibial axis (from the center of the proximal tibia to the ankle center) [3]. The HKA was expressed as a deviation from 180°, with varus defined as an HKA of −3° or less, valgus as an HKA of +3° or more, and neutral as an HKA between −3° and +3°. The mLDFA was defined as the lateral angle between the mechanical femoral axis and a line tangent to both femoral condyles [5]. The mMPTA was defined as the medial angle between the mechanical tibial axis and the knee joint line of the proximal tibia [5]. Based on the previous parameters following variables were calculated: arithmetic HKA (aHKA) (mMPTA—mLDFA), joint line obliquity (mMPTA + mLDFA), and Coronal Plane Alignment of the Knee (CPAK) classification. When the DICOM file was available, these parameters were calculated by fully-automated image processing software (Leg Angle Measurement Assistant (LAMA), ImageBiopsy Lab, Vienna, Austria). In the absence of the required DICOM file, two orthopaedic residents independently measured the angles to the nearest 0.1°, and the average values were computed for subsequent analysis.
An overview of the available radiographic images and the methodology used for radiographic data collection is presented in Appendix 1.
Statistical analysis
Statistical analyses were performed using SPSS (SPSS version 29, Chicago, IL, United States) and Prism GraphPad (Prism version 10, San Diego, CA, USA). Data are reported as mean and standard deviation (SD) or median and interquartile range (IQR) for continuous variables, based on normality tests (Shapiro-Wilk combined with visualization of boxplots), and frequencies (percentage) for categorical data.
Univariate analyses were performed for each potential predictor variable to compare cases and controls, with a sufficient case-to-predictor ratio ensured. This study investigated whether the following factors were associated with the occurrence of rTKA within ten years after pTKA: center, age at pTKA, sex, BMI, ASA classification, side, KL grade, Sperner classification, and alignment values (mHKA, mMPTA, mLDFA, aHKA, joint line obliquity, CPAK). Continuous variables were compared using either the independent Student’s t-test or the Mann-Whitney U test, based on normality assumptions. Categorical variables were compared using the Chi-square test or Fisher’s exact test, as appropriate. Following the identification of the predictor variables by the univariate analyses, a multivariable logistic regression analysis was conducted. To ensure the inclusion of potentially relevant predictors, variables with a univariate significance level of ≤ 0.157 were incorporated in the model [42], provided they were expected to have clinically meaningful relevance. Concerning radiographic parameters, an initial multivariable analysis was conducted using the KL score as the sole radiographic predictor, as it was the only parameter assessable for all participants based on the inclusion criteria.
A subgroup analysis was performed in the rTKA group to assess whether the main indications for revision differed across the various KL scores. The Chi-square test was used for this analysis, with a significance level set at 0.05. A Kaplan–Meier survival analysis was performed to provide a time-to-event perspective on revision risk for the cases.
Results
Demographic variables
The mean age at time of pTKA was 59.0 years (SD, 9.5) for the case group, which was significantly younger than the 64.7 years (SD, 8.6) observed in the control group (p < 0.001). No significant age differences were found between the two surgical centers within either the case or control group. The case group comprised 42 women (29.6%) and 100 men (70.4%), while the control group included 167 women (58.8%) and 117 men (41.2%) (p < 0.001). Sex distribution in the control group was influenced by the surgical center, with a female predominance in SMK and a male predominance in UZ Ghent. For the cases, both centers exhibited a comparable sex distribution, with a male predominance. BMI was comparable between cases and controls (p = 0.696). However, within the case group, patients from SMK had a significantly higher BMI compared to those of UZ Ghent (p = 0.006). An overview of the patient characteristics is presented in Table 1. Univariate statistical tests were performed to assess the need for including each variable in the multivariable logistic regression model.
Radiographic variables
An overview of the radiographic variables is provided in Table 2. Univariate analysis of pre-operative radiographic variables indicated that the mHKA, the number of patients with overall neutral alignment, and varus alignment demonstrated p-values below the threshold of p ≤ 0.158. However, these variables were deemed clinically insignificant.
Table 2.
Radiographic characteristics of cases (revision within 10 years postoperatively) and controls (no revision within 10 years postoperatively)
| Variable | Cases (n = 142) | Controls (n = 284) | P-value | ||||||
|---|---|---|---|---|---|---|---|---|---|
| All cases | UZ Ghent (n = 65) | SMK (n = 77) | P-value | All controls | UZ Ghent (n = 110) | SMK (n = 174) | P-value | ||
| KL grading scale | < 0.001* | ||||||||
| Grade I | 11 (7.7) | 7 (10.8) | 4 (5.2) | 0.216* | 0 (0) | 0 (0) | 0 (0) | - | < 0.001** |
| Grade II | 54 (38) | 34 (52.3) | 20 (26.0) | 0.001* | 26 (9.2) | 11 (10) | 15 (8.6) | 0.695* | < 0.001* |
| Grade III | 47 (33.1) | 18 (27.7) | 29 (37.7) | 0.208* | 94 (33.1) | 31 (28.2) | 63 (36.2) | 0.162* | 1.000* |
| Grade IV | 30 (21.1) | 6 (9.2) | 24 (31.2) | 0.001* | 164 (57.7) | 68 (61.8) | 96 (55.2) | 0.269* | < 0.001* |
| Sperner classification | n = 110 | n = 36 | n = 74 | n = 266 | n = 94 | n = 172 | 0.001* | ||
| No arthrosis | 1 (0.9) | 1 (2.8) | 0 (0) | 0.327** | 1 (0.4) | 1 (1.1) | 0 (0) | 0.353** | 0.500** |
| Grade I | 22 (20) | 10 (27.8) | 12 (16.2) | 0.155* | 39 (14.7) | 12 (12.8) | 27 (15.7) | 0.518* | 0.201* |
| Grade II | 57 (51.8) | 18 (50) | 39 (52.7) | 0.790* | 93 (35) | 33 (35.1) | 60 (34.9) | 0.971* | 0.002* |
| Grade III | 17 (15.5) | 4 (11.1) | 13 (17.6) | 0.379* | 94 (35.3) | 31 (33) | 63 (36.6) | 0.552* | < 0.001* |
| Grade IV | 13 (11.8) | 3 (8.3) | 10 (13.5) | 0.540** | 39 (14.7) | 17 (18.1) | 22 (12.8) | 0.243* | 0.467* |
| Alignment | n = 88 | n = 22 | n = 66 | n = 251 | n = 99 | n = 152 | |||
| mHKA (°) | −0.85 [−5.3-3.0] | −1.9 ± 5.1 | −0.5 ± 7.5 | 0.439‡ | −3.2 [−6.8-1.6] | −3.4 [−7.2-0.8] | −3.1 [−6.8-1.9] | 0.529† | 0.032† |
| Varus | 36 (40.9) | 9 (40.9) | 27 (40.9) | 1.000* | 133 (53) | 55 (55.6) | 78 (51.3) | 0.511* | 0.051* |
| Neutral | 30 (34.1) | 10 (45.5) | 20 (30.3) | 0.194* | 64 (25.5) | 27 (27.3) | 37 (24.3) | 0.603* | 0.121* |
| Valgus | 22 (25) | 3 (13.6) | 19 (28.8) | 0.155* | 54 (21.5) | 17 (17.2) | 37 (24.3) | 0.177* | 0.500* |
| mLDFA (°) | 87.6 [85.6–89.5] | 88.8 ± 2.3 | 87.1 ± 3.2 | 0.024‡ | 87.8 [86.2–89.3] | 87.7 [86.5–89.1] | 87.9 [85.9–89.5] | 0.516† | 0.727† |
| mMPTA (°) | 88.2 [86.1–90.7] | 87.7 [86.3–88.9] | 88.4 [86–91.5] | 0.226† | 88 [85.7–89.7] | 87.5 [85.7–88.9] | 88.1 [85.8–90.5] | 0.047† | 0.294† |
| aHKA (°) | 0.6 [−3.1–4] | −1.4 [−3.1-0.7] | 1.3 [−3.1-5] | 0.043† | −0.2 [−2.6-3] | −0.4 ± 3.7 | 0.6 ± 5 | 0.052‡ | 0.606† |
| Joint line (°) | 176.5 [173.7–178.4] | 176.6 [174–177.4] | 176.4 [173.3–178.7] | 0.866† | 175.5 [173–177.7] | 174.7 [173–176.9] | 176.2 [172.9–178.1] | 0.046† | 0.330† |
| CPAK | 0.094* | 0.007* | 0.166* | ||||||
| I | 24 (27.9) | 8 (40) | 16 (24.2) | 63 (25.1) | 26 (26.3) | 37 (24.3) | |||
| II | 17 (19.8) | 5 (25) | 12 (18.2) | 56 (22.3) | 32 (32.3) | 24 (15.8) | |||
| III | 14 (16.3) | 2 (10) | 12 (18.2) | 45 (17.9) | 20 (20.2) | 25 (16.4) | |||
| IV | 3 (3.5) | 1 (5) | 2 (3) | 13 (5.2) | 4 (4) | 9 (5.9) | |||
| V | 7 (8.1) | 2 (10) | 5 (7.6) | 32 (12.7) | 11 (11.1) | 21 (13.8) | |||
| VI | 18 (20.9) | 0 (0) | 18 (27.3) | 31 (12.4) | 6 (6.1) | 25 (16.4) | |||
| VII | 2 (2.3) | 1 (5) | 1 (1.5) | 1 (0.4) | 0 (0) | 1 (0.7) | |||
| VIII | 1 (1.2) | 1 (5) | 0 (0) | 1 (0.4) | 0 (0) | 1 (0.7) | |||
| IX | 0 (0) | 0 (0) | 0 (0) | 9 (3.6) | 0 (0) | 9 (5.9) | |||
Data are presented as mean ± SD or median [IQR], based on normality assumptions, for continuous variables. Categorical variables are presented as absolute number (percentage). In case of missing variables, the precise number is indicated in parentheses. KL, Kellgren-Lawrence. mHKA, mechanical hip-knee-ankle axis. mLDFA, mechanical lateral distal femoral angle. mMPTA mechanical medial proximal tibial angle. aHKA arithmetic HKA. CPAK Coronal Plane Alignment of the Knee. ‡Independent Student T-Test. † Mann-Whitney U-Test. *Chi-square test. **Fisher’s exact test. The p-value at the far right of the table represents the comparison between cases and controls. The p-value listed under cases and controls individually reflects the comparison between the two centers
Based on the univariate statistical tests in Table 2, significant radiographic predictors for revision surgery within 10 years following pTKA included KL-grade (p < 0.001) and the grade of PF osteoarthritis according to the Sperner classification (p < 0.001). In the case group, the majority (54/142, 38%) had a KL grade 2 prior to undergoing pTKA, followed by 47 patients (33.1%) with grade 3, 30 patients (21.1%) with grade 4, and 11 patients (7.8%) with grade (1) In contrast, the distribution in the control group differed, with the majority presenting with grade 4 OA (164/284, 57.8%), followed by 94 patients (33.1%) with grade 3 and 26 patients (9.2%) with grade (2) No patients with grade 1 OA were included in the control group. An overview of the KL grade distribution across both cohorts is presented in Fig. 3A. The distribution across the two centers was largely similar (Fig. 3B and C). However, grade 2 was significantly more prevalent in the case group from UZ Ghent, where grade 4 was more common in the case group from SMK. The prevalence of PF osteoarthritis Sperner classification grade 0, 1, and 4 did not differ significantly between the case and control groups. Grade 2 was significantly more prevalent in the case group, while grade 3 was significantly more common in the control group. No statistically significant differences were observed between the two centers.
Fig. 3.
Overview of the Kellgren-Lawrence distribution for cases (revision within 10 years postoperatively) and controls (no revision within 10 years postoperatively). OA, osteoarthritis
Predictors for revision surgery
Multivariable analysis of demographic and radiographic variables demonstrated that sex, age at pTKA, KL grade, and PF grade independently predict the need for revision surgery. To ensure the inclusion of all participants, an initial model was created using the KL grade as the sole radiographic parameter, along with center, sex, and age, as the KL assessment was a mandatory inclusion criterion for all patients (Table 3; Fig. 4). This analysis revealed that men have four times higher odds (95% CI, 2.41 to 6.69) of undergoing revision surgery compared to women (p < 0.001). Additionally, for each additional year of age, the odds of undergoing revision surgery decreased by 6% (p < 0.001). When using end-stage OA (KL grade 4) as the reference, patients with grade 2 have ten times higher odds, and those with grade 3 have 2.87 times higher odds of requiring revision surgery within 10 years following pTKA (p < 0.001).
Table 3.
Multivariable analysis of predictors for revision surgery within 10 years following pTKA
| Predictors in the model (R²=0.4) | OR (95% CI) | P-value |
|---|---|---|
| Center (reference: UZ Ghent) | 1.26 (0.76 to 2.1) | 0.372 |
| Sex (reference: women) | 4.01 (2.41 to 6.69) | < 0.001 |
| Age at pTKA (years) | 0.94 (0.92 to 0.97) | < 0.001 |
| KL Classification (reference: KL Grade 4) | < 0.001 | |
| KL Grade 1 | Could not be calculated | |
| KL Grade 2 | 10.12 (5.19 to 19.74) | < 0.001 |
| KL Grade 3 | 2.87 (1.64 to 5.02) | < 0.001 |
pTKA primary total knee arthroplasty. KL Kellgren-Lawrence. R² Nagelkerke R square. OR odds ratio. CI confidence interval
Fig. 4.
Logistic regression for revision surgery within 10 years following pTKA. pTKA, primary total knee arthroplasty. KL, Kellgren-Lawrence. References: center = UZ Ghent; sex = women; age pTKA = in years; KL = grade 4
The multivariable analysis, incorporating the Sperner classification based on the significant univariate analysis, indicated that the degree of PF osteoarthritis did not have a significant impact on the model (p = 0.121) (Table 4).
Table 4.
Multivariable analysis of predictors for revision surgery within 10 years following pTKA, including the Sperner classification
| Predictors in the model (R²=0.427) | OR (95% CI) | P-value |
|---|---|---|
| Center (reference: UZ Ghent) | 2.02 (1.09 to 3.73) | 0.025 |
| Sex (reference: women) | 4.27 (2.4 to 7.59) | < 0.001 |
| Age at pTKA (years) | 0.93 (0.9 to 0.96) | < 0.001 |
| KL Classification (reference: KL Grade 4) | < 0.001 | |
| KL grade 1 | Could not be calculated | 0.999 |
| KL grade 2 | 9.2 (4.2 to 20.14) | < 0.001 |
| KL grade 3 | 2.79 (1.49 to 5.22) | 0.001 |
| PF classification (reference: Sperner Grade 4) | 0.121 | |
| No osteoarthritis | 2.07 (0.03 to 143.87) | 0.736 |
| Sperner grade 1 | 1.88 (0.77 to 4.58) | 0.166 |
| Sperner grade 2 | 2.63 (1.3 to 5.31) | 0.007 |
| Sperner grade 3 | 1.8 (0.7 to 4.64) | 0.222 |
pTKA primary total knee arthroplasty. KL Kellgren-Lawrence. R² Nagelkerke R square. OR odds ratio. CI confidence interval
Characteristics of the revision group
The median time to revision for the cases was 21.5 months (IQR, 12.8 to 39 months). Revision surgery was performed within the first 48 months postoperatively in 84.5% of patients (Fig. 5). Instability was the most common indication for revision (35.9%), followed by malalignment (33.8%) and aseptic loosening (13.4%). No significant differences in indications for revision were observed across the various KL grades (p = 0.970) (Table 5). The median time to revision was similar between low-grade OA (KL1 or KL2; 23 months; IQR, 12.5 to 41 months) and high-grade OA (KL3 or KL4; 21 months; IQR, 13 to 38 months) (p = 0.691).
Fig. 5.
Survival probability for the cases (revision within 10 years postoperatively). CI confidence interval
Table 5.
Indications for revision surgery
| Indication for revision | KL 1–2 | KL 3–4 | Total |
|---|---|---|---|
| Aseptic loosening | 9 (13.8) | 10 (13) | 19 (13.4) |
| Instability | 24 (36.9) | 27 (35.1) | 51 (35.9) |
| Arthrofibrosis | 6 (9.2) | 8 (10.4) | 14 (9.9) |
| Malalignment | 24 (36.9) | 24 (31.2) | 48 (33.8) |
| Pain | 1 (1.5) | 1 (1.3) | 2 (1.4) |
| Other | 1 (1.5) | 7 (9.1) | 8 (5.6) |
Data are presented as absolute number (percentage). KL, Kellgren-Lawrence
Discussion
This is the first study to identify preoperative radiographic OA severity as predictor for revision surgery within 10 years following pTKA. Patients with KL grade 2 had a tenfold higher revision risk compared to those with KL grade 4. Notably, the indication for revision was not influenced by the OA severity. Multivariable analysis further identified younger age and male sex as predictors for revision risk.
Age at primary TKA
The average age at pTKA was significantly lower in the revision group (59 years) compared to controls (64.7 years). Each additional year of age reduced the revision risk by 6%, consistent with previous research showing reduced implant survivorship [2, 9, 15, 23, 27, 41] and higher lifetime revision risk in younger patients [2, 4, 14, 40, 41]. This may be attributed to dissatisfaction with TKA outcomes, shaped by unmet high expectations and increased functional demands [2, 9, 15, 19, 23, 34, 40, 41]. Younger patients often experience higher activity levels, which can contribute to prosthetic loosening, primarily driven by loading forces rather than time in service [2, 15, 21]. Moreover, younger patients tend to pursue rTKA at an earlier time point following TKA, often driven by dissatisfaction [2, 4, 9, 27, 41]. Given their relatively fewer contraindications compared to an older population, coupled with their higher expectations and aspirations, younger patients are considered as more attractive surgical candidates [9, 15]. This is based on the assumption that the dissatisfaction can be mitigated through a single, successful revision procedure [40]. The growing use of TKA among younger individuals, particularly those aged 45 to 55 years, is concerning as they represent the fastest-growing demographic of recipients [2, 9, 18, 30]. While TKA can be considered for younger patients, it is crucial that they are well-informed about the technology’s limitations in comfort and performance [40].
Sex
Male sex was associated with a fourfold increase in revision risk, a finding consistent with other studies [4, 10, 12, 25, 32]. Identifying factors contributing to this male preponderance is challenging due to potentially confounding variables, such as smoking, which were not available for analysis in this study [32]. Paradoxically, although female patients report higher rates of dissatisfaction following primary TKA [37], this does not translate into a higher revision risk, a discrepancy that remains unexplained.
Body mass index
Contrary to previous reports, a higher BMI was not associated with increased revision risk in this cohort. This discrepancy may be due to the exclusion of PJI, a major failure mechanism in obese patients [19].
Radiographic osteoarthritis severity
Radiographic severity emerged as a strong and novel predictor of revision risk following TKA. Patients with KL grade 2 OA were ten times more likely to undergo revision compared to those with KL grade 4, while KL grade 3 conferred a 2.9-fold increase. These findings strongly discourage performing TKA in the early stages of OA.
Orthopaedic surgeons continuously strive to identify patients who will benefit most from TKA surgery and those at risk of poor outcomes. Previously, predictive models for patient satisfaction after TKA have been developed to offer some guidance during the decision-making process [37]. While age and sex are non-modifiable risk factors, the decision to proceed with surgery remains a modifiable variable, ultimately guided by the clinical judgment of the surgeon. Our study reinforces the critical importance of appropriate patient selection for surgery, positioning indication setting not as routine, but as a cornerstone of responsible and outcome-driven surgical practice.
Prior studies have already shown that patients with KL grade 2 have significantly higher dissatisfaction rates after TKA, with some reports exceeding 50%, and generally inferior clinical outcomes in terms of PROMs [19, 23, 26, 30, 34]. The predominant source of pain in this cohort is likely inflammatory in nature, driven by synovitis, rather than mechanical symptoms related to cartilage loss. As TKA primarily targets mechanical dysfunction, this may explain the limited symptomatic relief observed in these patients [24]. Moreover, the decision to operate is often based on subjective complaints that do not align with radiographic findings [16, 26]. This mismatch can, at least in part, be explained by individual variations in central pain processing, including cognitive and emotional modulation of nociception, which may also affect TKA outcomes beyond patient expectations [16]. So, inferior outcomes following TKA due to poor indications are well-documented, yet this study is the first to reveal the increased risk of progression to revision surgery in their prosthetic journey.
Dissatisfaction may subsequently result in worsened functionality and reduced quality of life, driving patients to a relentless quest for answers, frequently seeking second, third, or even fourth opinions in the hope of finding a solution for their persistent pain and suffering [16, 34]. Since the indications for revision are not always clearly defined, there is often a perceived justification for proceeding with revision in any patient experiencing pain. In the current study, the indication for revision did not vary according to the severity of OA. Consequently, surgeons may opt for revision surgery under the assumption that a single, successful procedure will resolve the patient’s issues [40]. The key question we must address is whether revision surgery is the appropriate strategy for dissatisfied TKA patients. Particularly, in cases without a clear indication for revision, a revision may do more harm than good. Instead, the priority must be to establish accurate and well-grounded indications for pTKA. This paper contributes to the existing predictive frameworks by highlighting low radiographic severity as a key red flag. It urges us to sharpen our surgical indications, not only to improve patient satisfaction, but to prevent avoidable revisions.
Other radiographic parameters
No association was found between preoperative alignment and revision risk, in contrast to Ritter et al., who reported increased failure in knees with alignment outside 8° varus or 11° valgus [29]. A possible explanation could be that alignment deviations in the present cohort were less pronounced.
In certain cases, extensive patellofemoral OA (Sperner grade 3 and 4) combined with low-grade tibiofemoral OA (KL grade 1 and 2) may justify TKA. However, this subgroup comprised only a small portion of both the control (5.9%) and revision (7.2%) groups, suggesting minimal confounding from patellofemoral involvement. A sensitivity analysis excluding these patients was performed and did not alter the results.
Limitations
Several potential limitations should be acknowledged. First, owing to the dual-center design of the study, some heterogeneity is inherent. For example, posttraumatic OA was excluded from the SMK control group but not from the SMK revision or UZ Ghent cohorts, although this subgroup represented fewer than 10% of cases in all groups. Radiographic data collection also varied (Table 6 in Appendix 1), but in cases where automated analysis was not feasible, two orthopaedic residents independently assessed the images, with a third adjudicator resolving discrepancies. For the 152 radiographs evaluated through this manual procedure, the interobserver agreement between the two residents demonstrated a Cohen’s kappa coefficient of 0.857, indicating a strong level of reliability [20]. The heterogeneity is further reflected in the differing timeframes applied for patient selection across centers. Second, the availability of radiographic imaging constituted a major limiting factor for patient inclusion and contributed substantially to the number of exclusions in both centers. Furthermore, imaging had to be available at least two years prior to surgery. Ideally, this timeframe should be shortened in future studies to improve representativeness of preoperative joint status. Third, OA severity was assessed using the KL scale, without MRI data to identify pathological features like chondral delamination, subchondral edema and synovitis [16]. However, the indication for TKA remains primarily based on weight-bearing radiographs [34]. Fourth, while matching cohorts on fixed variables such as age, sex, and BMI could reduce confounding, a random sampling approach was intentionally employed to enable assessment of these variables as potential risk factors, acknowledging the possibility of residual confounding. Additionally, potential confounding factors for revision risk such as implant design, surgical technique, comorbidities, and ethnicity were not included in the analysis. Another limitation is the inability to track patients who may have undergone revision elsewhere. However, given that both centers are tertiary referral hospitals for knee arthroplasty, the number of revisions performed at other institutions is likely minimal, limiting the risk of bias.
Table 6.
Overview of available images and used method for the evaluation of the degree of osteoarthritis and alignment parameters
| Variable | Cases (n = 142) | Controls (n = 284) | ||||
|---|---|---|---|---|---|---|
| All cases | Ghent (n = 65) | Nijmegen (n = 77) | All controls | Ghent (n = 110) | Nijmegen (n = 174) | |
| Radiological images available | ||||||
| Schuss | 103 (72.5) | 32 (49.2) | 71 (92.2) | 200 (70.4) | 34 (30.9) | 166 (95.4) |
| Standing AP | 91 (64.1) | 63 (96.9) | 28 (36.4) | 112 (39.4) | 102 (92.7) | 10 (5.7) |
| FL | 91 (64.1) | 22 (33.8) | 67 (87) | 251 (88.4) | 99 (90) | 152 (87.4) |
| PF view | 110 (77.5) | 36 (55.4) | 74 (96.1) | 266 (93.7) | 94 (85.5) | 172 (98.9) |
| Evaluation degree of OA | ||||||
| KOALA | 100 (70.4) | 25 (38.5) | 75 (97.4) | 174 (61.3) | 0 (0) | 174 (100) |
| Manually | 42 (29.6) | 40 (61.5) | 2 (2.6) | 110 (38.7) | 110 (100) | 0 (0) |
| Evaluation alignment | ||||||
| LAMA | 46 (52.3) | 0 (0) | 46 (69.7) | 118 (47) | 0 (0) | 118 (77.6) |
| Manually | 42 (47.7) | 22 (100) | 20 (30.3) | 133 (53) | 99 (100) | 34 (22.4) |
Data are presented as absolute number (percentage). AP, anteroposterior. FL, Full-leg. PF, patellofemoral. OA, osteoarthritis. KOALA, Knee OsteoArthritis Labeling Assistant. LAMA, Leg Angle Measurement Assistant.
Despite these limitations, this is the first well-designed study to demonstrate a significant association between radiographic OA severity and revision risk following pTKA. Future research should investigate whether the outcomes following revision surgery are worse in cases with a poor indication for primary TKA, as the underlying pain mechanism may not be biomechanical and therefore not resolvable by the revision procedure.
Conclusion
This study is the first to demonstrate a significant association between the radiographic OA severity and the risk of revision surgery following pTKA, with patients presenting with KL grade 2 OA showing a tenfold increased risk compared to those with KL grade 4. These findings support the recommendation that TKA should primarily be indicated for patients with advanced OA. Additionally, younger age and male sex were also identified as predictors of revision risk. These results emphasize the importance of appropriate patient selection and reinforce the need for comprehensive preoperative counselling and shared decision-making to optimize long-term outcomes.
Acknowledgements
The authors thank Thomas Ackerman and Catho Sadones for their valuable assistance with data collection. Furthermore, the authors thank Ewald Bronkhorst for his contribution in optimizing the statistical processing of this study.
Appendix 1
See Table 6
Author contributions
L. S.: conceptualization—data curation—formal analysis—investigation—methodology—project administration—resources—visualization—writing—original draft. H. V.: conceptualization—data curation—formal analysis—methodology—writing—review and editing. T. D.: conceptualization—data curation—formal analysis—methodology—writing—review and editing. S. V. O.: conceptualization—methodology—writing—review and editing. N. A.: conceptualization—methodology—writing—review and editing. G. v. H.: conceptualization—methodology—supervision—writing—review and editing. J. V.: conceptualization—funding acquisition—methodology—project administration—resources supervision—writing—review and editing. P. H.: conceptualization—data curation—formal analysis—investigation—methodology—project administration—software—supervision—visualization—writing—review and editing.
Funding
J. Victor reports a grant from Research Foundation Flanders (no. T001620N), related to this study.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Conflict of interest
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Akar B, Uğur F, Kaplan T, Albayrak M (2024) Relationship between patellar resurfacing and postoperative anterior knee pain in total knee arthroplasty. Anatol Curr Med J 6:267–271 [Google Scholar]
- 2.Anatone AJ, Richardson SS, Kahlenberg CA, Gausden EB, Figgie MP, Blevins JL (2022) Decreased Implant Survival is Associated With Younger Patients Undergoing Total Knee Arthroplasty. HSS J 18:290–296 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Anon. Cooke TD, Sled EA, Scudamore RA (2007) Frontal plane knee alignment: a call for standardized measurement. J Rheumatol 34(9):1796–1801 [PubMed] [Google Scholar]
- 4.Bayliss LE, Culliford D, Monk AP et al (2017) The effect of patient age at intervention on risk of implant revision after total replacement of the hip or knee: a population-based cohort study. Lancet 389:1424–1430 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bellemans J, Colyn W, Vandenneucker H, Victor J (2012) The chitranjan ranawat award. Clinical Orthopaedics and Related Research, vol 470. Springer Science and Business Media, LLC, pp 45–53 [DOI] [PMC free article] [PubMed]
- 6.Brejnebøl MW, Hansen P, Nybing JU et al (2022) External validation of an artificial intelligence tool for radiographic knee osteoarthritis severity classification. Eur J Radiol. 10.1016/j.ejrad.2022.110249 [DOI] [PubMed] [Google Scholar]
- 7.Brown LR, Clement ND, MacDonald DJ, Breusch SJ (2019) The survivorship of the link endo-rotational hinge total knee arthroplasty: 5–12-year follow-up of 100 patients. Arch Orthop Trauma Surg 139:107–112 [DOI] [PubMed] [Google Scholar]
- 8.Butcher CC, Atkins RM (2003) (II) Principles of deformity correction. Curr Orthop 17:418–435 [Google Scholar]
- 9.Charette RS, Sloan M, DeAngelis RD, Lee GC (2019) Higher Rate of Early Revision Following Primary Total Knee Arthroplasty in Patients Under Age 55: A Cautionary Tale. J Arthroplasty 34:2918–2924 [DOI] [PubMed] [Google Scholar]
- 10.Curtin B, Malkani A, Lau E, Kurtz S, Ong K (2012) Revision After Total Knee Arthroplasty and Unicompartmental Knee Arthroplasty in the Medicare Population. J Arthroplasty 27:1480–1486 [DOI] [PubMed] [Google Scholar]
- 11.DeFrance MJ, Scuderi GR (2023) Are 20% of Patients Actually Dissatisfied Following Total Knee Arthroplasty? A Systematic Review of the Literature. J Arthroplasty 38:594–599 [DOI] [PubMed] [Google Scholar]
- 12.Dy CJ, Marx RG, Bozic KJ, Pan TJ, Padgett DE, Lyman S (2014) Risk factors for revision within 10 years of total knee arthroplasty. Clin Orthop Relat Res 472:1198–1207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gunaratne R, Pratt DN, Banda J, Fick DP, Khan RJK, Robertson BW (2017) Patient dissatisfaction following total knee arthroplasty: a systematic review of the literature. J Arthroplasty 32:3854–3860 [DOI] [PubMed] [Google Scholar]
- 14.Haynes J, Sassoon A, Nam D, Schultz L, Keeney J (2017) Younger patients have less severe radiographic disease and lower reported outcome scores than older patients undergoing total knee arthroplasty. Knee 24:663–669 [DOI] [PubMed] [Google Scholar]
- 15.Julin J, Jämsen E, Puolakka T, Konttinen YT, Moilanen T (2010) Younger age increases the risk of early prosthesis failure following primary total knee replacement for osteoarthritis: a follow-up study of 32,019 total knee replacements in the Finnish Arthroplasty Register. Acta Orthop 81:413–419 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Klasan A, Rice DA, Kluger MT et al (2022) A combination of high preoperative pain and low radiological grade of arthritis is associated with a greater intensity of persistent pain 12 months after total knee arthroplasty. Bone Joint J 104–B:1202–1208 [DOI] [PubMed] [Google Scholar]
- 17.Klug A, Gramlich Y, Rudert M et al (2021) The projected volume of primary and revision total knee arthroplasty will place an immense burden on future health care systems over the next 30 years. Knee Surg Sports Traumatol Arthrosc 29:3287–3298 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kurtz SM, Lau E, Ong K, Zhao K, Kelly M, Bozic KJ (2009) Future young patient demand for primary and revision joint replacement: National projections from 2010 to 2030. Clinical Orthopaedics and Related Research, vol 467. Springer, New York, pp 2606–2612 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Leppänen S, Niemeläinen M, Huhtala H, Eskelinen A (2021) Mild knee osteoarthritis predicts dissatisfaction after total knee arthroplasty: a prospective study of 186 patients aged 65 years or less with 2-year follow-up. BMC Musculoskelet Disord. 10.1186/s12891-021-04543-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.McHugh ML (2012) Interrater reliability: the kappa statistic. Biochem Med (Zagreb) 22:276–282 [PMC free article] [PubMed] [Google Scholar]
- 21.Meehan JP, Danielsen B, Kim SH, Jamali AA, White RH (2014) Younger age is associated with a higher risk of early periprosthetic joint infection and aseptic mechanical failure after total knee arthroplasty. J Bone Joint Surg 96:529–535 [DOI] [PubMed] [Google Scholar]
- 22.Nakano N, Shoman H, Olavarria F, Matsumoto T, Kuroda R, Khanduja V (2020) Why are patients dissatisfied following a total knee replacement? A systematic review. Int Orthop 44:1971–2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Niemeläinen M, Moilanen T, Huhtala H, Eskelinen A (2019) Outcome of knee arthroplasty in patients aged 65 years or less: a prospective study of 232 patients with 2-year follow-up. Scand J Surg 108:313–320 [DOI] [PubMed] [Google Scholar]
- 24.Niinimäki TT, Murray DW, Partanen J, Pajala A, Leppilahti JI (2011) Unicompartmental knee arthroplasties implanted for osteoarthritis with partial loss of joint space have high re-operation rates. Knee 18:432–435 [DOI] [PubMed] [Google Scholar]
- 25.Ong KL, Lau E, Suggs J, Kurtz SM, Manley MT (2010) Risk of subsequent revision after primary and revision total joint arthroplasty. Clin Orthop Relat Res 468:3070–3076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Peck CN, Childs J, McLauchlan GJ (2014) Inferior outcomes of total knee replacement in early radiological stages of osteoarthritis. Knee 21:1229–1232 [DOI] [PubMed] [Google Scholar]
- 27.Perez B, Koressel J, Cohen JS, Kirchner GJ, Kerbel YE, Lee GC (2023) Why and what happens to patients younger than 60 years who need revision total knee arthroplasty? J Arthroplasty 38:2404–2409 [DOI] [PubMed] [Google Scholar]
- 28.Riis A, Rathleff MS, Jensen MB, Simonsen O, Professor A, Joint BJ (2014) Cite this article. Bone Joint J. :96–1498
- 29.Ritter MA, Davis KE, Davis P et al (2013) Preoperative malalignment increases risk of failure after total knee arthroplasty. J Bone Joint Surg 95:126–131 [DOI] [PubMed] [Google Scholar]
- 30.Scott CEH, Oliver WM, Macdonald D, Wade FA, Moran M, Breusch SJ (2016) Predicting dissatisfaction following total knee arthroplasty in patients under 55 years of age. ;98 [DOI] [PubMed]
- 31.Shichman I, Askew N, Habibi A et al (2040) Projections and epidemiology of revision hip and knee arthroplasty in the United States to 2040–2060. Arthroplast Today. 10.1016/j.artd.2023.101152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Stone B, Nugent M, Young SW, Frampton C, Hooper GJ (2022) Cite this article. Bone Joint J 104:235–241 [DOI] [PubMed] [Google Scholar]
- 33.Stroobant L, Ackerman T, Vermue H et al (2025) Revision knee arthroplasty in a tertiary center: infection remains the leading cause for primary revisions, while aseptic loosening led causes for re-revisions. J Arthroplasty. 10.1016/j.arth.2025.02.045 [DOI] [PubMed] [Google Scholar]
- 34.Tampere T, Arnout N, Victor J (2024) Total knee arthroplasty: the need for better patient selection. Knee Surg Sports Traumatol Arthrosc. 10.1002/ksa.12514 [DOI] [PubMed] [Google Scholar]
- 35.Van Kempen RWTM, Schimmel JJP, Van Hellemondt GG, Vandenneucker H, Wymenga AB (2013) Reason for revision TKA predicts clinical outcome: Prospective evaluation of 150 consecutive patients with 2-years followup knee. Clin Orthop Relat Res 471:2296–2302 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Van Laarhoven SN, Nota SPFT, Van Hellemondt GG, Schreurs BW, Wymenga AB, Heesterbeek PJC (2025) Cite this article. Bone Joint J 107:65–71 [DOI] [PubMed] [Google Scholar]
- 37.Van Onsem S, Van Der Straeten C, Arnout N, Deprez P, Van Damme G, Victor J (2016) A new prediction model for patient satisfaction after total knee arthroplasty. J Arthroplasty 31:2660-2667e.e1 [DOI] [PubMed] [Google Scholar]
- 38.van Rensch PJH, Hannink G, Heesterbeek PJC, Wymenga AB, van Hellemondt GG (2020) Long-term outcome following revision total knee arthroplasty is associated with indication for revision. J Arthroplasty 35:1671–1677 [DOI] [PubMed] [Google Scholar]
- 39.Vasso M, Beaufils P, Schiavone Panni A (2013) Constraint choice in revision knee arthroplasty. Int Orthop 37:1279–1284 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Vince KG (2014) You Can Do Arthroplasty in a Young Patient, But…. J Bone Joint Surg 96:e58 [DOI] [PubMed] [Google Scholar]
- 41.Walker-Santiago R, Tegethoff JD, Ralston WM, Keeney JA (2021) Revision total knee arthroplasty in young patients: higher early reoperation and rerevision. J Arthroplasty 36:653–656 [DOI] [PubMed] [Google Scholar]
- 42.Wood AM, White IR, Royston P (2008) How should variable selection be performed with multiply imputed data? Stat Med 27:3227–3246 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
No datasets were generated or analysed during the current study.