Learn more: PMC Disclaimer | PMC Copyright Notice
Knee Biomechanics and Contralateral Knee Osteoarthritis Progression after Total Knee Arthroplasty
Abstract
Background:
Despite the success rate of Total Knee Arthroplasty (TKA), many patients undergo contralateral TKA. It is possible that altered gait mechanics after unilateral TKA play a role in the progression of contralateral OA progression.
Research question:
The purpose of this study was to identify biomechanical predictors of radiographic OA progression in the contralateral (non-surgical) knee after unilateral (primary/initial) TKA. In addition, this study quantified for patients who had contralateral OA progression.
Methods:
Biomechanical outcomes were collected 6–24 months after unilateral primary TKA and were used to predict changes in contralateral OA severity at follow-up. Participants were divided into “Progressor” and “Non-Progressor” groups based on changes in Kellgren-Lawrence (KL) OA grade and Joint Space Width (JSW) between baseline and follow-up testing sessions. Biomechanical factors during walking were peak knee adduction moment, knee flexion/extension excursions, knee angle at initial foot contact, and peak knee flexion/extension. Multiple independent t-tests were used to examine the magnitude of differences in biomechanical variables between the groups. Logistic regression was used to examine the association between the biomechanical predictors and change in KL scores and JSW.
Results:
The mean time between surgery and follow-up x-rays was 8.8 (2.4) years. Of 40 participants, 62.5–78% had contralateral radiographic knee OA progression by follow-up. There were no significant differences in the biomechanical variables between groups. For the regression analysis, none of the biomechanical variables were found to be predictors for contralateral OA progression.
Significance:
Although abnormal biomechanics are known risk factors for primary knee OA, it is possible that the mechanisms that result in OA progression of the contralateral limb are different than primary knee OA progression. Future work should evaluate other objective measures of OA progression and determine if cumulative measures of joint loading are related to OA worsening.
1. Introduction
Total knee arthroplasty (TKA) is a common surgical treatment for people with end-stage knee osteoarthritis (OA)1,2. More than 700,000 TKAs are performed annually in the United States (US)3, and that number is expected to increase 4-fold by 20404. Despite the benefits of TKA, many individuals who undergo unilateral TKA experience a progression of OA on the contralateral knee after the initial surgery5. This progression can happen quickly, and for those who require a contralateral TKA, the time to second surgery is only 3.2 years6. Even in patients who are unilaterally symptomatic at the time of the primary procedure, nearly 25% of these patients undergo contralateral TKA within several years6.
It is possible that altered gait mechanics after unilateral TKA play a role in the progression of contralateral OA progression7. Specifically, greater joint loading, as measured by higher knee adduction moments, is a known risk factor for primary OA progression8,9 and reduced knee flexion excursion is a risk for undergoing TKA in the future7,10. Individuals after unilateral TKA tend to favor the contralateral limb during functional and weight-bearing tasks like walking11 and rising from a chair12. Specifically, these individuals place greater force on the contralateral limb and have higher adduction moments in the contralateral knee11. It is possible that reliance on the contralateral knee after index TKA is an underlying cause of radiographic contralateral OA progression, but this has not been fully explored.
Previous work from our lab revealed that altered movement patterns after TKA, specifically reduced sagittal plane motion (i.e. stiff-legged gait patterns), are associated with future contralateral TKA use7. While this study explored the relationship between biomechanics and contralateral OA progression, it used time to contralateral TKA as an end-point and did not examine the association between biomechanics and radiographic changes in the contralateral knee. Undergoing TKA is usually indicative of OA progression, but the decision to undergo TKA is multifactorial and may be driven by factors other than structural and symptomatic changes in the knee joint. Identification of risk factors is an essential first step to developing interventions that preserve the long-term joint integrity of operative and non-operative.
Therefore, the primary purpose of this study was to identify biomechanical predictors of radiographic OA progression in the contralateral (non-surgical) knee after unilateral (primary/initial) TKA. We hypothesized that individuals who demonstrate OA progression on their contralateral knee will have higher knee adduction moments and lower knee flexion/extension excursions at baseline compared to those who do not demonstrate OA progression.
2. Methods
2.1. Study Design
This was a longitudinal prognostic study that utilized a participant pool from a previous observational cross-sectional study, which will be referred to as the “parent study”. As part of the parent study, biomechanical11 outcomes were collected 6–24 months after unilateral primary TKA (Table 1; Study Component 3). For the current analysis, we collected pre-operative x-rays from the medical record (Study Component 1) and had participants undergo follow-up x-rays (Study Component 4) to quantify OA severity. Biomechanical measurements collected at baseline (6–24 months after TKA) were used to predict changes in contralateral OA severity at follow-up.
Table 1.
Measurement sequence
| Additional Data for this Study | Parent Study | Additional Data for this Study | ||
|---|---|---|---|---|
|
| ||||
| Study Component | 1 | 2 | 3 | 4 |
| Baseline x-rays | Primary unilateral TKA | Biomechanical measurements | Follow up x-rays | |
| Time Relative to TKA | 1–2 months before TKA | Time of TKA | 6–24 months after TKA | 5.5–13 years after TKA |
2.2. Participants
Participants in this study were individuals who participated in the parent study and agreed to return for the follow-up radiographic assessment. In the parent study, participants were recruited through a local group of orthopedic surgeons who perform tricompartmental cemented TKA with a medial parapatellar approach using either a posterior cruciate-retaining or posterior stabilized prosthesis. In the parent study, participants were enrolled if they had end-stage unilateral knee OA and underwent primary unilateral TKA. In order to obtain a primarily unilaterally symptomatic sample, participants were excluded from the parent study if they: 1) were scheduled for or planned a contralateral TKA, 2) reported pain greater than 4 out of 10 in the contralateral knee, 3) had any other lower extremity surgeries, or 4) had limited physical function due to musculoskeletal involvement in the spine or lower extremities. Participants who underwent revision surgery were excluded from this study.
For the current analysis, all participants who participated in the parent study (N=229) were attempted to be contacted and informed about the follow-up study. Letters were sent to all participants informing them of the additional study procedures. Interested participants (n=60) contacted the research coordinator for screening and scheduling. Forty-two participants had biomechanics and x-ray data. Subjects were excluded from this analysis if they had KL Grade 4 in either the medial or lateral compartment. Prior participating in the study, participants signed a written informed consent form approved by the Human Subjects Review Board at the University of Delaware. Previous work on knee kinematics found that the minimal clinically important differences (MCID) between limbs was 3 degrees13 for knee flexion excursion. Because knee flexion excursion has been previously linked to contralateral OA progression10, we used this as the basis for our sample size. We found that a sample size of 34 individuals was sufficient to identify difference between groups, should one exist.
2.3. Biomechanical Measurements
Biomechanical Measurements were acquired during over-ground walking at a self-selected speed. Detailed information on the biomechanical setup has been previously reported10,29,30. Briefly, lower limb joint kinematics were captured using retro-reflective markers tracked via an 8-camera high-speed motion analysis system (VICON, Oxford Metrics, London, England) at 120-Hz. Kinetic (peak ground reaction force and joint moments) and kinematic (peak angles and excursions) variables were calculated with 3-dimensional instrumented motion analysis. Two force platforms (Bertec Corp., Worthington, OH, USA) (1080-Hz) synchronized with the cameras recorded the ground reaction force data to measure the lower extremity joint kinetics. Anatomical markers were bilaterally placed at acromioclavicular joints, iliac crests, greater trochanters, medial and lateral femoral epicondyles, medial and lateral malleoli, and first and fifth metatarsal heads. Non-anatomical markers (tracking markers) included 1) two markers placed on the back of the shoes, 2) one shell with fixed three markers placed on the pelvis at the sacrum, and 3) five shells with four fixed markers each placed on the posterior thoracic area, the lateral thighs, and the lateral shanks.
After placing markers bilaterally and obtaining a standing calibration, participants walked until they achieved at least 5 successful walking trials at a consistent velocity16. To control for walking velocity, the average speed of the measured successful walking trials was maintained within +/− 5% of the individual’s average walking speed that was obtained prior to the start of the data collection. A successful walking trial was described as a trial in which the participants contacted each forceplate with different leg, and without intentionally targeting the force plates. Joint angles and moments during walking using a standard inverse dynamics approach integrating kinematic and kinetic data (Visual 3D, C-Motion, Germantown, MD)16. Euler X-Y-Z sequence were used to calculate the knee joint angles (flexion/extension, abduction/adduction then rotation sequences). Force platforms data were filtered at (40-Hz) while marker trajectories were low pass filtered at (6-Hz) using a second order phase-corrected butterworth filter.
Given the previous work linking biomechanics to OA progression7,8,9,10, the biomechanical variables of interest for study (Figure 1) were overall peak external knee adduction moment (normalized to body mass and height), and knee flexion/extension excursions during stance. Since work from our lab has revealed that full knee ROM is a modifiable factor that can postpone the TKA need (population aged between 46–78)17, we also measured other biomechanical factors, such as knee angle at initial contact and peak knee flexion/extension during stance (Figure 1). Biomechanical changes on the operated and non-operated limbs are associated with contralateral TKA use; therefore, both limbs were evaluated in the current analysis.
a. Knee Angle at Initial Contact. b. Knee Flexion Excursion calculated as the difference between Initial Contact and Peak Knee Flexion. c. Knee Angle at Peak Knee Flexion. d. Knee Extension Excursion calculated as the difference between Peak Knee Flexion and Peak Knee Extension. e. Knee Angle at Peak Knee Extension.
2.4. Radiographic Reading and Digital X-Ray Assessment
Radiographic images were obtained at two time points (pre-operative and follow-up; Table 1). If the participant agreed to participate in this follow-up analysis, bilateral posterior to anterior (PA) bent-knee (20-degree) radiographs were obtained. The time between TKA and follow-up x-ray varied depending on the participant’s original testing date. In the event the individual underwent a contralateral TKA at the time they were re-contacted for this study, pre-operative x-rays of the contralateral knee from the medical records were obtained. In addition to these follow-up x-rays, bilateral weight-bearing PA bent-knee radiographs taken prior (1–2 months) to the primary TKA were retrieved from patients’ records. Osteoarthritis Research Society International (OARSI) has recommended that a semi-flexed knee view is a valid method to evaluate structural severity in knee OA clinical trials18.
Because the follow-up radiographs were digital, SigmaView software (Agfa HealthCare Corporation, Greenville, SC, USA) were used to measure Kellgren-Lawrence (KL) grades. The KL scale19 has five grades of increasing OA severity: none (0), doubtful (1), minimal (2), moderate (3), and severe (4)19. Grade-0 indicates no presence of pathological changes. Grade-1 indicates doubtful joint space narrowing and the possible formation of osteophytes, while grade-2 shows definite osteophytes and possible joint space narrowing. Grade-3 indicates definite joint space narrowing, moderate multiple osteophytes, possible bony ends deformity, and some sclerosis may present. Grade-4 indicates large osteophytes, severe sclerosis, definite deformity in bone ends, and obvious narrowing of joint space. KL grading was performed separately for the medial and lateral compartments. Previous work found that KL grade is a significant predictor for contralateral OA progression to TKA20. Although there is some subjectivity inherent in the KL scoring, it evaluates joint alignment, bony (osteophyte) formation, cartilage width changes, and articular cartilage deformation21, all of which are relevant for classifying OA severity. The evaluator (MSA) in this study underwent a standardized KL training program and graded all of the X-rays. The grader had excellent intersession reliability (ICC = 0.956) with a narrow confidence interval (95% CI = 0.921 – 0.975). This grader was blinded to participants’ biomechanical data at the time of grading.
In addition to KL scoring, the X-rays were also evaluated for joint space width (JSW) in the medial and lateral compartments. To quantify JSW (mm), the distal femoral width was measured first by drawing segment that is tangent to both medial and lateral femoral condyles. Then, drawing a 90-degree angle with a horizontal arm that is tangent to the most prominent point on femoral epicondyle for each compartment. Based on the recommendation from the Osteoarthritis Initiative (OAI)22, locations at 25% and 70% of the distal femoral width, starting from the medial compartment, will be identified. In the OAI dataset, this method was compared to Magnetic Resonance Imaging (MRI) measures of cartilage loss22. OAI suggested these locations because they found that these locations can monitor the progression based on the knee alignment. Using JSW at 25% is optimal for Varus knees, while 70% is a preferred location for Valgus knees. These location measurements made with the medial side of the joint as the reference point (0%). JSW was calculated as the distance of the vertical line between the femur and the tibia at the 25% (medial compartment) and 70% (lateral compartment) locations. An example of a measured X-ray is in Figure 2.
2.5. Defining the OA Progression
Osteoarthritis Research Society International (OARSI) suggests that OA progression should be presented in a dichotomous form; hence, to categorize individuals as “progressors or non-progressors”23. We defined OA progression using change in KL grade and joint space width (JSW). Our decision to use two different methods is because KL scores and JSW capture different anatomical structures and have a different level of subjectivity. Although KL score is less objective, it can evaluate joint alignment, bony (osteophyte) formation, cartilage width changes, and articular cartilage deformation21,24. The JSW is less subjective and evaluates the cartilage thinning and width changes in objective terms (millimeters “mm”) on a continuous scale22.
For the KL method, participants were divided into two groups based on the maximum changes in the KL score in either the medial or lateral compartments of the contralateral knee between the initial pre-operative radiographs and the follow-up period. Individuals who demonstrated worsening KL grades (at least 1-point change in either compartment) were defined as “progressor group”. Individuals without a change in KL grade in either the medial or lateral compartment were classified as “non-progressor group”. Given that individuals with KL grade 4 cannot demonstrate progression, individuals with grade 4 contralateral knee OA at baseline were excluded from the current analysis.
For the JSW method, there is a heterogeneity of reported cut-off points for JSW that define progressors in the literature. The suggested values for progressor range from (0.12 to 0.84 mm)23 and (0.50 to 0.80 mm)25 as a change in JSW from baseline to follow up25. Cooper and colleagues26 suggested using 0.5 mm as a cut-off in minimum joint space narrowing which reflects a clinical relevant progression in individuals with knee OA. Bruyere and colleagues25 found that individuals who had joint space narrowing >0.50 mm within 3 years follow-up were more likely to have total joint replacement over the next 5 years than individuals without similar radiographic progression. Given that 0.50 mm appears to be a common mid-point across studies, participants were divided into two groups based on a decrease of 0.5 mm or more in the medial or lateral compartments of the contralateral knee.
The rationale for defining OA progression using maximum KL grade and minimal JSW (as opposed to separating the medial and lateral compartment for KL, or using the 25% and 70% locations for JSW) is based on clinical observations and previous evidence22. Patients are typically concerned with symptoms (pain) and loss of function more so than the location of the impairment (medial vs. lateral compartment) and individuals undergo TKA for both medial and lateral compartment symptoms. However, it should be noted that knee adduction moment is only predictive of medial compartment progression. Therefore, we performed a sub-analysis of OA progression in the medial compartment of KL grade and 25% JSW location of the distal femoral width, which is a standard location for identifying the change in the medial compartment.
2.6. Statistical Analysis
Participants were divided into two dichotomous groups defined as the “progressor group” or “non-progressor group” based on changes in the KL score (the maximum change in either the medial or lateral compartments) in the contralateral knee between the initial pre-operative radiographs and the follow-up radiographs. Separately, we also defined “progressors” and “non-progressors” based on the change in the JSW. The statistical analyses were performed in two steps. First, multiple independent t-tests were used to examine the magnitude of differences in biomechanical variables between the two groups. Second, hierarchical logistic regressions were used to examine the association between the biomechanical predictors and the dependent variables (progressor and non-progressor). The predictors were entered in the regression model as follows: Block 1) demographics (age, BMI, sex); Block 2) biomechanical variables. This analysis was performed separately based on progression defined from KL scores and from JSW scores.
Since our primary definition of progressor was based on changes in either the medial or lateral tibiofemoral compartment, a sub-analysis was performed to identify predictors of medial compartment progression. The adduction moment is a surrogate measure of medial compartment joint loading27, so higher adduction moments are unlikely to contribute to lateral compartment disease. Therefore, an additional analysis was performed to determine if peak adduction moment was predictive of changes in the medial tibiofemoral compartment. In this sub-analysis we divided the KL group based on the medial compartment only, not the maximum KL score. The alpha level was set at 0.05 for all regression analyses. While all odds ratios are presented as a one unit change in the variable, we also calculated the odds of 1 SD change for the adduction moment, given that a 1-unit change in the adduction moment is not physiologically likely. Results for the one-unit change are shown in the table, and 1 SD change are shown in the manuscript text.
3. Results
Participants:
Sixty participants were able to be re-contacted and participated in the follow-up study. Of these participants, pre-operative x-rays were able to be obtained from 42 participants. Two of these participants were excluded because they had a KL grade of 4 at baseline x-ray in the contralateral knee in either the medial (n=1) or lateral (n=1) compartment. Forty participants were included in the KL analysis while 38 participants were included in the JSW analysis due to magnification issues of 2 baseline x-rays that precluded accurate distance measurements, but still allowed for KL scoring. No multicollinearity was found in our sample; variance inflation factor (VIF) values were less than 2 and tolerance values exceeded 0.628.
KL Score Results:
At the time of baseline testing, participants had mean age of 66.5 (6.8) years and BMI of 32.0 (4.7) kg/m2. The mean time between surgery and follow-up x-rays was 7.8 (2.5) years. There were 19 males and 21 females. No significant differences were found in sex, age, BMI, and height between progressor and non-progressor groups. At the baseline, the majority of the progressors had KL grade 1 or 2 (n=22) in the contralateral knee (Table 2).
Table 2
Participant demographics
| Based on KL scores (N = 40) | Based on JSW (N = 37) | |||||
|---|---|---|---|---|---|---|
| Progressor Group n = 25 (62.5%) | Non-Progressor Group n = 15 (37.5%) | p -value | Progressor Group n = 29 (78%) | Non-Progressor Group n = 8 (22%) | p -value | |
| Sex | ||||||
| Male n (%) | 10 (40%) | 9 (60%) | 0.220 | 14 (48%) | 4 (50%) | 0.931 |
| Female n (%) | 15 (60%) | 6 (40%) | 15 (52%) | 4 (50%) | ||
| Age (years) | 67.2 (5.9) | 65.7 (7.6) | 0.481 | 67.2 (6.6) | 63.0 (4.7) | 0.098 |
| BMI (kg/m2) | 31.5 (4.4) | 32.6 (5.0) | 0.476 | 32.2 (5.1) | 30.8 (2.7) | 0.443 |
| Height (meter) | 1.7 (0.1) | 1.7 (0.1) | 0.839 | 1.7 (0.1) | 1.7 (0.1) | 0.860 |
| Gait Speed (meter/second) | 1.33 (0.17) | 1.34 (0.15) | 0.932 | 1.33 (0.16) | 1.36 (0.18) | 0.833 |
| Time between Surgery and Follow-up X-ray (years) | 7.64 (1.96) | 8.00 (3.07) | 0.652 | 7.3 (1.8) | 10.4 (2.6) | 0.000 * |
| Baseline KL score (Grade: n) | 1: 11 | 1: 7 | ||||
| 2: 11 | 2: 4 | - | - | - | - | |
| 3: 3 | 3: 4 | |||||
Data are presented as “Means (Standard Deviations)”
BMI = Body Mass Index
Of 40 participants included in the KL analyses, 25 participants (62.5%) had contralateral radiographic knee OA progression by follow-up and 15 (37.5%), had no change. Of those 25 participants that changed KL scores, 12 changed one grade level, 10 changed two grade levels and 3 changed three grade levels.
There were no significant differences in the biomechanical variables between groups. There was also no significant difference in adduction moment in the medial compartment subset analysis (Table 3). For the regression analysis, none of the biomechanical variables were predictors for contralateral OA progression (Table 4).
Table 3
Progressor vs non-progressor (maximum KL of the medial and lateral compartments)
| Based on KL scores | Based on JSW | ||||||||||
|
| |||||||||||
| Biomechanical Variables | Side | Progressor (n=25; 62.5%) Mean (SD) | Non-progressor (n=15; 37.5%) Mean (SD) | Mean Difference (95% CI) | Cohen’s d | p-value | Progressor (n = 29; 78%) Mean (SD) | Non-progressor (n = 8; 22%) Mean (SD) | Mean Difference (95% CI) | Cohen’s d | p-value |
|
| |||||||||||
| Knee Flexion Excursion (degrees) | Operated | 12.9 (3.1) | 14.7 (3.2) | 1.78 (−0.29 – 3.86) | −0.59 | 0.089 | 13.8 (3.4) | 12.7 (3.0) | 1.17 (−1.51 – 3.84) | 0.34 | 0.382 |
| Contralateral | 15.4 (3.4) | 15.1 (4.3) | −0.27 (−2.277 – 2.22) | 0.08 | 0.825 | 15.1 (3.1) | 16.1 (6.0) | −0.95 (−4.06 – 2.17) | −0.27 | 0.542 | |
| Knee Extension Excursion (degrees) | Operated | 10.7 (4.4) | 11.9 (4.9) | 1.26 (−1.80 – 4.32) | −0.27 | 0.411 | 11.8 (4.3) | 9.4 (5.9) | 2.36 (−1.40 – 6.12) | 0.53 | 0.211 |
| Contralateral | 13.7 (5.3) | 12.6 (6.0) | −1.13 (−4.80 – 2.54) | 0.20 | 0.537 | 14.1 (4.7) | 12.7 (7.6) | 1.38 (−3.00 – 5.76) | 0.27 | 0.525 | |
| Knee Angle at Initial Contact (degree) | Operated | 5.3 (4.2) | 6.0 (6.4) | −0.69 (−4.08 – 2.70) | −0.14 | 0.684 | 5.5 (5.2) | 5.6 (5.5) | 0.14 (−4.13 – 4.42) | −0.02 | 0.946 |
| Contralateral | 3.3 (5.5) | 4.3 (7.2) | −1.04 (−5.14 – 3.05) | −0.17 | 0.609 | 3.7 (4.9) | 2.5 (6.9) | −1.23 (−5.61 – 3.15) | 0.26 | 0.573 | |
| Peak Knee Flexion (degree) | Operated | 18.2 (4.3) | 20.7 (6.6) | −2.47 (−5.95 – 1.01) | −0.49 | 0.159 | 19.3 (5.6) | 18.3 (4.5) | −1.02 (−5.43 – 3.38) | 0.19 | 0.640 |
| Contralateral | 18.7 (5.8) | 19.4 (8.4) | −0.77 (−5.34 – 3.80) | −0.10 | 0.735 | 18.9 (5.8) | 18.6 (6.2) | −0.28 (−5.08 – 4.51) | 0.05 | 0.906 | |
| Peak Knee Extension - Mid stance knee angle (degree) ** | Operated | −7.5 (4.8) | −8.8 (5.3) | −1.22 (−4.54 – 2.11) | 0.27 | 0.463 | 7.5 (5.1) | 8.8 (5.8) | 1.34 (−2.86 – 5.53) | −0.25 | 0.522 |
| Contralateral | −4.9 (6.1) | −6.8 (10.1) | −1.90 (−7.07 – 3.28) | 0.25 | 0.462 | 4.8 (4.0) | 5.9 (10.9) | 1.10 (−8.02 – 10.22) | −0.19 | 0.786 | |
| Progressor vs non-progressor (maximum KL of the medial knee compartment) | |||||||||||
| Based on KL scores | Based on JSW | ||||||||||
| Side | Progressor (n=23; 57.5%) Mean (SD) | Non-progressor (n=17; 42.5%) Mean (SD) | Mean Difference (95% CI) | Cohen’s d | p-value | Progressor (n = 26; 70%) Mean (SD) | Non-progressor (n = 11; 30%) Mean (SD) | Mean Difference (95% CI) | Cohen’s d | p-value | |
| Peak Knee Adduction Moment (Nm/kg) | Operated | 0.308 (0.106) | 0.274 (0.099) | 0.0336 (0.03 – 0.10) | 0.34 | 0.315 | 0.287 (0.101) | 0.293 (0.119) | −0.0062 (−0.07 – 0.08) | −0.06 | 0.871 |
| Contralateral | 0.386 (0.156) | 0.433 (0.127) | −0.0465 (−0.14 – 0.05) | −0.33 | 0.321 | 0.402 (0139) | 0.407 (0.170) | 0.0049 (−0.10 – 0.11) | 0.00 | 0.927 | |
Data are shown in degrees and presented as “Means (Standard Deviations)”
Table 4
Logistic regression analysis
| Based on KL scores | Based on JSW | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| Predictors | Nagelkerke R Square | Odds ratio (95% CI) | p-value | Nagelkerke R Square | Odds ratio (95% CI) | p-value |
| Contralateral Peak Knee Adduction Moment (Nm/kg) | 0.231 | 21.22 (0.14 – 3299.89) | 0.235 | 0.054 | 0.948 (0.01 – 136.66) | 0.983 |
| Contralateral Knee Flexion Excursion (degrees) | 0.002 | 1.02 (0.86 – 1.21) | 0.820 | 0.018 | 1.077 (0.855 – 1.356) | 0.531 |
| Contralateral Knee Extension Excursion (degrees) | 0.014 | 1.04 (0.92 – 1.17) | 0.527 | 0.017 | 0.953 (0.825 – 1.102) | 0.515 |
| Contralateral Knee Angle at Initial Contact “Heel Strike” (degrees) | 0.009 | 1.03 (0.93 – 1.15) | 0.599 | 0.014 | 1.045 (0.901 – 1.213) | 0.562 |
| Contralateral Peak Knee Flexion (degrees) | 0.004 | 1.02 (0.93 – 1.12) | 0.728 | 0.001 | 1.009 (0.880 – 1.156) | 0.902 |
| Contralateral Peak Knee Extension (degrees) | 0.019 | 1.03 (0.95 – 1.12) | 0.459 | 0.646 | 0.970 (0.855 – 1.101) | 0.641 |
Data are presented as “Odds ratio (95% confidence interval (CI))”
Each model was adjusted for Age, BMI, Sex
Odds ratios represent the change in risk of OA progression associated with a 1-unit increase in the metric of interest (Exp(B)).
As shown in Table 4, the odds ratio was large for the adduction moment (21.22 [95% CI: 0.14 – 3299.89]) because the 1-unit change in adduction moment is not physiologically likely. Therefore, in addition to calculating the 1-unit change in odds ratio, we also calculated the odds ratios based on 1-standard deviation (SD) change. A one-standard deviation change was 0.145 Nm/kg for the non-operated side. For every 1 SD change, the risk of KL progression was 3.005 (p=0.235).
JSW Results:
At the time of baseline testing, participants had mean age of 65.1 (5.6) years and BMI of 31.5 (3.9) kg/m2. The mean time between surgery and follow-up x-rays was 8.8 (2.2) years, p=0.000 (Table 2). There were 18 males and 19 females. No significant differences were found in sex, age, BMI, and height between progressor and non-progressor groups (Table 2).
Of 38 participants included in the JSW analyses, 29 participants (78%) had contralateral radiographic knee OA progression by follow-up. There were no significant differences in the biomechanical variables between groups (Table 3). For the regression analysis, none of the biomechanical variables were predictors for contralateral OA progression (Table 4).
As with the risk of KL progression, we also calculated the odds of JSW progression for a 1 SD change in adduction moment because a 1-unit change in the adduction moment is not physiologically likely. A one-standard deviation change was 0.147 Nm/kg for the non-operated side and the for every 1 SD change, the risk of JSW progression was −0.054 (p= 0.983).
4. Discussion
In this study, we found that the incidence of radiographic knee OA progression in the contralateral knee was 62.5% based on KL score and 78% based on joint space narrowing by an average of 8.3 years follow-up from initial TKA. This is a relatively large incidence of OA progression within an 8-year window. A previous study that evaluated primary OA progression over an average of 6.6 years, found that 22.7% of the participants demonstrated radiographic progression29. However, that was in a primary OA cohort, and the incidence of progression in the contralateral knee may differ as biomechanics are known to be altered after TKA. One previous study found that contralateral TKA use in a general sample of patients was 46% within 3 years of the index surgery5. However, these studies relied on contralateral TKA use and did not evaluate objective measures of OA progression. Clinicians should be aware that even in patients who are largely asymptomatic in the contralateral knee at baseline, OA progression is common. Despite this, we did not find any association between OA progression and movement patterns.
In this study, OA severity was assessed using both KL scores and the JSW. While they are often interpreted similarly from a clinical perspective, each outcome provides different information about anatomical structure. KL grades consider joint alignment, osteophyte (bony) formation, and articular cartilage deformation when quantifying severity21, while JSW is a more objective measure that considers solely the thickness of the cartilage on a continuous scale22. Because both approaches revealed contralateral knee OA progression, it can be assumed that the contralateral knee OA progression is due to not only cartilage thinning, but also the development of osteophytes and changes in gross joint structure. However, there was a slightly larger incidence when using the JSW method (78%) when compared to the KL scores (62.5%). This may suggest that joint space narrowing occurs even if the patient does not develop osteophytes or bony deformation. The difference in incidence may also be related to the subject’s baseline joint structure. Eleven of 25 participants (56%) in the progressor group had KL scores of 2 or 3 at baseline, suggesting they already had osteophytes and bony changes. Therefore, it is likely that subsequent changes in joint quality may be solely limited to cartilage thinning and not the development of larger or additional osteophytes. However, the difference in incidence using the two methods (KL vs JSW) was relatively small, and the majority of patients had OA progression regardless of the method used to quantify it.
Studies from tissue-level experiments have shown that excessive load placed on cartilage can cause structural deterioration30. Although previous literature has found that greater knee adduction moments at baseline predict radiographic OA progression in individuals with primary knee OA8,9, we did not find a similar relationship in the contralateral knee after. Even though patients are more likely to rely on the contralateral leg after surgery, and experience larger adduction moments on the contralateral side15, this did not appear to be a risk for contralateral OA progression. Given that the one-unit change in adduction moment represents a large increase in this metric, it is unlikely that a patient would experience a one unit change with OA progression or with an intervention. This is why the values for the odds ratio in Table 4 are exceptionally large, albeit not significant. The more clinically relevant risk values are discussed in the text for a 1 SD change in adduction moment. These risks were consistent with what was found for the 1 unit change in that an increase in adduction moment did not significantly predict an increase in the risk of OA progression.
However, the lack of associated between biomechanics and KL progression should be interpreted with caution for several reasons. First, although knee adduction moment is correlated with medial compartment loading30, it is only a partial contributor to overall joint load. Other factors, including the knee flexion moment31 and muscle activity32 need to also be considered when discussing compressive forces in the local joint environment.
Second, knee adduction moments are related to OA severity33 and individuals with higher KL grades have higher adduction moments. In this study, participants with higher KL grades at baseline had higher average adduction moments (mean [SD]): KL 1 = 0.182 (0.105), KL 2 = 0.199 (0.120), KL 3 = 0.243 (0.137); p=0.0012. The time-series curve (Figure 3) also shed insight to the progressively larger adduction moments with an increase in OA severity. It is possible that individuals with lower KL grades at baseline are more likely to progress to higher grades at follow-up. This is inherent in the way the KL grading system is designed. The change from a 1 to a 2 requires on the development of a single osteophyte. To change from a 3 to a 4, a patient must develop large osteophytes, severe sclerosis, bony deformity, and large and obvious narrowing of joint space. In this study, there were 19 participants with KL grade 1 at baseline and 11 of these participants (58%) progressed at the follow-up. There were 6 participants with KL grade 3 at baseline and only 2 (33%) progressed by follow-up. Although this is not significant (chi-square 2.44, p=0.12), it may be attributed to the low sample size.
In both the progressor and non-progressor groups, individuals walked with characteristic gait of those with knee OA and TKA. Specifically, both groups landed with knee in a flexed position at initial contact and did not come to full knee extension at midstance, which has shown to be common in this population 7,10. This supports previous work that found that gait patterns after TKA do not return to normal; often patients continue to walk with flexed knee gait patterns with attenuated sagittal plane motion on the operated side34. While this may contribute to residual functional deficits, alterations in the joint loading environment35, as well as risk of contralateral TKA, it did not appear to be a risk for radiographic changes based on KL scores (Table 3). It is worth noting that although the difference between progressors and non-progressors for knee flexion excursion was small, it did approach significant levels and had a moderate effect size, with the non-progressors having more knee flexion excursion in the operated limb (Table 3). Given the fact that previous work has shown normal joint excursions are related to functional performance and reduce the risk of contralateral TKA, clinicians should still ensure that patients move with normal knee flexion during the stance phase of gait because limited knee flexion can contribute to residual pain after TKA36. While differences in sagittal plane motion between groups may be below the minimal detectable difference (MDD) for knee flexion angles37, the MDD for knee flexion excursion has not been reported.
Biomechanical changes alone, may not be the driving cause behind OA initiation and progression. Previous work has found that greater joint loads, combined with a greater BMI, was associated with larger loss of cartilage volume in the medial compartment38. Similarly, abnormal knee alignment (excessive varus) in the presence of higher body mass, also increases the risk of OA39. We also only evaluated these subjects during gait in a controlled environment. It is possible that physical activity and cumulative loading in the real-world may also influence the rate and incidence of OA progression. Future work should determine if the risk of OA in the presence of biomechanical abnormalities increases in certain subsets of the patient population. Although we accounted for differences in BMI, abnormal biomechanics may predispose those who have systemic metabolic conditions such as diabetes, abnormal knee alignment, or greater physical activity and higher cumulative load to OA progression.
There are several limitations in this study that may also contribute to our negative findings. Our sample was limited to individuals who were enrolled in a previous study and were interested in participating in this follow-up analysis. Due to the smaller sample size, we might have been underpowered to detect statistical differences for all of the outcomes. In light of this, we have added the effect size and confidence interval for our primary outcomes (Table 3). Another limitation is that we did not record or have any data for other risk factors, such as physical activity, that may have influenced the total load experienced by the joint during the follow-up period. Gait biomechanics capture a small snapshot of an individual’s movement profile in a constrained environment. Capturing movement patterns in a variety of functional tasks, as well as a measure of cumulative loading, may provide additional insight into how movement patterns influence structural progression.
5. Conclusions
We anticipated that the knee adduction moment and knee joint excursions would be associated with contralateral knee OA progression, but our findings did not support this hypothesis. It is possible that the mechanisms that result in OA progression of the contralateral limb are different than primary knee OA progression. Future work should determine if cumulative measures of joint loading are related to OA worsening.
Acknowledgements
This work was supported by the National Institutes of Health [R56 AG048943 and P20 RR016458]; and a scholarship from Umm Al-Qura University (Makkah, Saudi Arabia) to Moiyad Aljehani through Saudi Arabian Cultural Mission (SACM).
Footnotes
Conflict of interest
The authors have no conflict of interest with the material to be published in the paper.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.