Abstract
Objective –
We aimed to 1) explore the cross-sectional relation of unilateral knee pain severity and temporal asymmetry during walking and 2) determine relations of temporal asymmetry during walking to 2-year changes in ipsilateral and contralateral knee pain in those with mild-to-moderate unilateral knee pain.
Methods –
The Multicenter Osteoarthritis Study is a prospective cohort study of adults with or at risk for knee osteoarthritis. The current study included participants with unilateral knee pain. Gait was assessed during self-selected and fast walking at baseline. Knee pain was assessed at baseline and 2-years. We calculated limb symmetry indices (LSIs; non-painful limb/painful limb×100) for stance, single limb support, and double limb support times then examined their relations to 1) unilateral knee pain severity, 2) incident contralateral knee pain, and 3) persistent ipsilateral knee pain.
Results –
Unilateral knee pain severity was not associated with temporal asymmetry during self-selected or fast walking. At 2-years, 17.1% had incident contralateral knee pain and 51.4% had persistent ipsilateral knee pain. For self-selected walking, greater LSIs (i.e., longer time on the non-painful limb) for stance and single limb support were associated with decreased odds of incident contralateral knee pain. Measures of temporal asymmetry were not associated with persistent ipsilateral knee pain, except for single limb support during fast walking.
Conclusion –
For those with unilateral knee pain, temporal asymmetry during walking is not associated with pain severity. However, select measures of stance and single limb support during self-selected and fast walking relate to longitudinal knee pain outcomes.
1. INTRODUCTION
Knee osteoarthritis (OA) is a leading cause of pain and disability that accounts for an estimated 83% of the total burden of OA (1–3). In many cases, knee OA begins as a unilateral disease (4). Yet, nearly 90% of those with unilateral knee OA progress to bilateral knee OA within 10 years (5). The progression to bilateral knee OA is concerning because it is associated with worse function, pain, and quality of life (6–8). Furthermore, bilateral knee OA may be associated with higher healthcare costs if both knees require treatment. To this end, a critical need exists for understanding the pathway from unilateral to bilateral knee OA.
If pain affects the way persons with unilateral knee OA walk, it may play a role in the progression from unilateral to bilateral OA (i.e., unilateral knee OA → unilateral knee pain → altered walking pattern that increases loading of the contralateral knee → bilateral knee OA). Research has shown that individuals with unilateral knee OA load their painful/affected knee less than their non-painful/unaffected knee while walking (9–11). However, there have been inconsistent findings when comparing biomechanical asymmetries between those with unilateral knee OA, bilateral knee OA, and healthy controls (12–14). These inconsistencies may be attributed to defining knee OA in different ways (i.e., based on structure and/or symptoms). Currently, it remains unclear if knee pain severity explains asymmetric walking patterns in those with unilateral knee OA. This is important because increased loading is associated with OA progression (15–19) and walking is commonly recommended for managing symptoms of knee OA (20–22).
Regardless of the cause, increased time on the non-painful limb relative to the painful limb may hasten the progression from unilateral to bilateral knee OA. Ideally, measures of knee joint loading would be used to quantify asymmetric walking. However, it is not feasible to directly measure or estimate knee joint loads in most clinical settings. One of the simplest ways to quantify asymmetry during walking is with temporal measures. Even though temporal measures have not been established as surrogates of knee joint loading, they are readily available to clinicians with access to wearable and portable technologies (e.g., spatiotemporal gait mats, inertial measurement units, accelerometers, pressure-sensing shoe inserts). Therefore, exploring the relation of temporal asymmetry to longitudinal knee pain outcomes may improve our understanding of the pathway from unilateral to bilateral knee OA and provide clinicians with temporal measures that can be targeted throughout rehabilitation to optimize patient outcomes.
The first objective of the current study was to explore the cross-sectional relation of unilateral knee pain severity and temporal asymmetry during walking. We hypothesized that more severe unilateral knee pain would be associated with 1) relatively longer stance and single limb support times on the non-painful limb compared to the painful limb and 2) relatively shorter double limb support time on the non-painful limb compared to the painful limb. The second objective was to determine relations of temporal asymmetry during walking to 2-year changes in ipsilateral and contralateral knee pain in those with mild-to-moderate unilateral knee pain. We hypothesized that relatively longer stance and single limb support times on the non-painful limb compared to the painful limb as well as relatively shorter double limb support times on the non-painful limb compared to the painful limb would be associated with 1) increased odds of incident knee pain in the non-painful limb (i.e., incident contralateral knee pain) and 2) decreased odds of persistent knee pain in the painful limb (i.e., persistent ipsilateral knee pain).
2. METHODS
2.1. Study Design and Sample
The Multicenter Osteoarthritis (MOST) Study is a NIH-funded prospective cohort study of 3,026 individuals, aged 50–79 years, with or at risk for knee OA. Participants were recruited for their initial clinic visit from 2003 to 2005 in two communities in the United States: Birmingham, Alabama and Iowa City, Iowa. Full details about the MOST cohort can be found elsewhere (23). Ethics approval was obtained from the institutional review boards at each of the study sites and all participants provided written consent after study procedures and risks were disclosed. The current study used data from clinic visits that occurred 60- and 84-months after the initial clinic visit. The 60-month clinic visit was the baseline visit for the current study because this was the first time point where gait data were collected in the MOST cohort.
For the current study, individuals with unilateral knee pain at the 60-month clinic visit were included. Unilateral knee pain was defined as only one knee having at least mild pain with walking as reported on the Western Ontario and McMaster Universities Arthritis Index (WOMAC) and the same knee having a score greater than zero on the Visual Analog Scale (VAS; 0–100). Individuals were excluded from all analyses for the following reasons: 1) incomplete or invalid gait data at the 60-month clinic visit, 2) had a VAS pain score greater than zero for their contralateral knee at the 60-month clinic visit, 3) reported buttock, hip, thigh, leg, ankle, or foot pain during the gait evaluation at the 60-month clinic visit, 4) underwent a hip or knee joint replacement surgery before the 60-month clinic visit, and 5) had missing covariate data. For longitudinal analyses, we additionally excluded those who: 1) withdrew or passed away before the 84-month clinic visit, 2) had missing outcome data, and 3) underwent a hip or knee joint replacement surgery before the 84-month clinic visit.
2.2. Measures
2.2.1. Temporal Asymmetry during Walking
Participants completed four self-selected and four fast walking trials over a 4.9-meter pressure-sensitive GAITRite walkway (CIR Systems, Inc., Franklin, NJ, USA) in their normal walking shoes. For the self-selected trials, participants walked at a comfortable and unhurried speed. After self-selected trials were completed, fast trials were performed. For the fast trials, participants walked as if they were late for an important meeting. All trials started and stopped 1.5 meters before and after the GAITRite walkway to avoid collecting data during acceleration and deceleration.
Mean stance, single limb support, and double limb support times were calculated for each limb by averaging the measures for each step on the GAITRite walkway. Stance time was defined as the duration between initial contact and terminal stance. Single limb support time was defined as the duration in stance where the contralateral limb was in swing (i.e., single limb on the ground). Double limb support time was defined as the duration in stance where the contralateral limb was also in stance (i.e., both limbs on the ground). Importantly, previous literature shows that the GAITRite walkway has excellent test-retest reliability (Intraclass correlation coefficients ≥ 0.89) and validity for temporal measures of gait (24, 25).
To quantify the degree of asymmetry in stance, single limb support, and double limb support, we calculated limb symmetry indices (LSI). For this, the measures from the non-painful limb were divided by the measures from the painful limb and then multiplied by 100 to transform the ratio into a percentage. This method of calculating LSI was used because it is common in clinical practice and easy to interpret. Although we kept LSI values continuous for all analyses, LSI values >100% indicated relatively longer times on the non-painful limb, LSI values <100% indicated relatively longer times on the painful limb, and LSI values of 100% indicated equal times.
Based on our hypotheses, we expected higher LSI values for stance and single limb support times to be associated with 1) more severe unilateral knee pain, 2) increased odds of knee pain in the non-painful limb two years later, and 3) decreased odds of still having knee pain in the painful limb two years later. Furthermore, we expected lower LSI values for double limb support time to be associated with these outcomes.
2.2.2. Unilateral Knee Pain Severity
To determine the cross-sectional relation of unilateral knee pain severity and temporal asymmetry during walking, VAS pain scores for the painful knee at the 60-month clinic visit were used. VAS pain scores were obtained by having participants mark an “X” on a zero to 100 scale in response to the question “How bad has the pain been in your knee, on average, in the past 30 days?” A score of zero indicated no pain, while a score of 100 indicated pain as bad as it could be. This same question was used for determining study eligibility and longitudinal outcomes.
2.2.3. Incident Contralateral Knee Pain
Incident contralateral knee pain was defined as the non-painful knee at the 60-month clinic visit having at least mild pain with walking (WOMAC) or a VAS pain score of at least 20 at the 84-month clinic visit. Because individuals in the current study sample had contralateral VAS pain scores of zero at the 60-month clinic visit, a score of at least 20 at the 84-month clinic visit was used because a 20-point change in VAS is a clinically relevant increase in pain for those with knee OA (26).
2.2.4. Persistent Ipsilateral Knee Pain
Persistent ipsilateral knee pain was defined when the painful knee at the 60-month clinic visit continued to have at least mild pain with walking (WOMAC) and a VAS pain score greater than zero at the 84-month clinic visit. In other words, the originally painful knee remained painful two years later.
2.3. Analytic Approach
Descriptive statistics were calculated for each variable. Continuous measures were reported as means and standard deviations. Categorical variables were reported as percentages and frequencies.
We determined the cross-sectional relation of unilateral knee pain severity and temporal asymmetry during walking with unadjusted and adjusted linear regression models. Adjusted models accounted for age (years), sex (female/male), body mass (kg), race (white/other), the presence of depressive symptoms (Center for Epidemiologic Studies Depression scale ≥16), gait speed (m/s), and Kellgren Lawrence (KL) grades for both knees. Associations were determined separately for self-selected and fast walking conditions. To assist with interpretation of effect estimates, beta coefficients were calculated per 20-unit increase in VAS pain scores.
To determine the longitudinal relations of temporal asymmetry during walking to 1) incident contralateral knee pain and 2) persistent ipsilateral knee pain, we used unadjusted and adjusted logistic regression models. Loess procedures were used to confirm the exposures were linearly associated with our outcomes. Similar to the cross-sectional analyses, adjusted models accounted for age, sex, body mass, race, the presence of depressive symptoms, gait speed, and KL grades. For each outcome, odds ratios (OR) and 95% confidence intervals (CI) were determined per 2.5-unit increase in exposure (i.e., 2.5% increase in LSI values). Because previous literature suggests that an absolute LSI >105% is a meaningful amount of asymmetry (27), we attempted to estimate our effects per 5-unit increase in exposure. However, in order to avoid estimating beyond the range of our data, a 2.5-unit increase was chosen. Since measures of temporal asymmetry were quantified during self-selected and fast walking trials, models were repeated for each exposure definition separately. In sensitivity analyses, we also adjusted for average steps per day, which was measured in the MOST Study with a StepWatch Activity Monitor (Orthocare Innovations, Mountlake Terrace, WA) worn for seven consecutive days.
All analyses were performed in SAS (version 9.4; SAS Institute Inc.; Cary, NC, USA). For cross-sectional analyses, significant level was set a priori to less than 0.05. For longitudinal analyses, significant associations were determined by confidence intervals not crossing 1.0.
3. RESULTS
3.1. Study Sample
Of the 3026 persons with or at risk for knee OA in the parent MOST Study, 127 were eligible for the cross-sectional analyses. Of those 127, 105 were subsequently eligible for the longitudinal analyses. Figure 1 shows the number of individuals excluded by study criterion. Participant characteristics can be found in Table 1 for those who were eligible. Additionally, a summary of the gait data collected at the 60-month clinic visit can be found in Table 2.
Figure 1.
Study Eligibility Flow Chart
Table 1.
Participant Characteristics
| Cross-sectional Analyses (n=127) | Longitudinal Analyses (n=105) | |
|---|---|---|
| Age (years); mean (SD) | 66.4 (7.2) | 65.5 (6.6) |
| Height (cm); mean (SD) | 169.8 (9.6) | 170.2 (9.0) |
| Body mass (kg); mean (SD) | 85.7 (18.1) | 86.3 (17.9) |
| Female; n (%) | 68 (53.5) | 57 (54.3) |
| White; n (%) | 115 (90.6) | 95 (90.5) |
| Depressive Symptoms; n (%) | 14 (11.0) | 10 (9.5) |
| Average Steps per day; mean (SD) | 7224 (2777)* | 7335 (2841)† |
| Kellgren Lawrence Grades for the Painful Knee | ||
| 0; n (%) | 35 (27.6) | 33 (31.4) |
| 1; n (%) | 15 (11.8) | 12 (11.4) |
| 2; n (%) | 25 (19.7) | 19 (18.1) |
| 3; n (%) | 34 (26.8) | 29 (27.6) |
| 4; n (%) | 18 (14.2) | 12 (11.4) |
| Kellgren Lawrence Grades for the Nonpainful Knee | ||
| 0; n (%) | 61 (48.0) | 56 (53.3) |
| 1; n (%) | 31 (24.4) | 23 (21.9) |
| 2; n (%) | 18 (14.2) | 15 (14.3) |
| 3; n (%) | 17 (13.4) | 11 (10.5) |
| 4; n (%) | 0 (0) | 0 (0) |
| VAS for the Painful Knee (0–100): | ||
| Baseline; mean (SD) | 22.5 (16.8) | 19.8 (13.7) |
| 2-year Follow-up; mean (SD) | N/A | 17.6 (19.1) |
| VAS for the Nonpainful Knee (0–100) | ||
| Baseline; mean (SD) | 0 (0) | 0 (0) |
| 2-year Follow-up; mean (SD) | N/A | 5.0 (11.1) |
| Pain with Walking for the Painful Knee (WOMAC): | ||
| Baseline: | ||
| None; n (%) | 0 (0) | 0 (0) |
| Mild; n (%) | 106 (83.46) | 90 (85.71) |
| Moderate; n (%) | 20 (15.75) | 15 (14.29) |
| Severe; n (%) | 1 (0.79) | 0 (0) |
| Extreme; n (%) | 0 (0) | 0 (0) |
| 2-year Follow-up: | ||
| None; n (%) | N/A | 50 (47.62) |
| Mild; n (%) | N/A | 42 (40.00) |
| Moderate; n (%) | N/A | 12 (11.43) |
| Severe; n (%) | N/A | 1 (0.95) |
| Extreme; n (%) | N/A | 0 (0) |
| Pain with Walking for the Nonpainful Knee (WOMAC): | ||
| Baseline: | ||
| None; n (%) | 127 (100) | 105 (100) |
| Mild; n (%) | 0 (0) | 0 (0) |
| Moderate; n (%) | 0 (0) | 0 (0) |
| Severe; n (%) | 0 (0) | 0 (0) |
| Extreme; n (%) | 0 (0) | 0 (0) |
| 2-year Follow-up: | ||
| None; n (%) | N/A | 90 (85.71) |
| Mild; n (%) | N/A | 14 (13.33) |
| Moderate; n (%) | N/A | 0 (0) |
| Severe; n (%) | N/A | 1 (0.95) |
| Extreme; n (%) | N/A | 0 (0) |
sample size of 100 due to missing or invalid physical activity data
sample size of 84 due to missing or invalid physical activity data
VAS – Visual Analog Scale
WOMAC – Western Ontario and McMaster Universities Arthritis Index
Table 2.
Temporal measures of gait during self-selected and fast walking in adults with unilateral knee pain
| Cross-sectional Analyses (n=127) | Longitudinal Analyses (n=105) | |
|---|---|---|
| Self-Selected Walking | ||
| Gait speed (m/s); mean (SD) | 1.20 (0.16) | 1.21 (0.16) |
| Stance time for painful limb (ms); mean (SD) | 702.7 (66.3) | 703.1 (68.8) |
| Stance time for non-painful limb (ms); mean (SD) | 706.8 (65.6) | 706.5 (68.4) |
| Stance time LSI (%); mean (SD) | 100.6 (2.9) | 100.5 (2.8) |
| Single limb support time for painful limb (ms); mean (SD) | 423.9 (39.7) | 423.8 (39.6) |
| Single limb support time for non-painful limb (ms); mean (SD) | 427.6 (38.9) | 427.0 (39.2) |
| Single limb support time LSI (%); mean (SD) | 101.0 (4.3) | 100.8 (4.1) |
| Double limb support time for painful limb (ms); mean (SD) | 278.8 (48.1) | 279.2 (49.4) |
| Double limb support time for non-painful limb (ms); mean (SD) | 279.3 (46.3) | 279.5 (47.7) |
| Double limb support time LSI (%); mean (SD) | 100.4 (4.1) | 100.3 (4.2) |
| Fast Walking | ||
| Gait speed (m/s); mean (SD) | 1.62 (0.21) | 1.64 (0.21) |
| Stance time for painful limb (ms); mean (SD) | 581.0 (61.7) | 577.4 (63.7) |
| Stance time for non-painful limb (ms); mean (SD) | 583.7 (63.1) | 580.6 (64.9) |
| Stance time LSI (%); mean (SD) | 100.5 (2.6) | 100.6 (2.4) |
| Single limb support time for painful limb (ms); mean (SD) | 373.8 (35.1) | 372.4 (35.3) |
| Single limb support time for non-painful limb (ms); mean (SD) | 378.0 (36.0) | 376.9 (36.2) |
| Single limb support time LSI (%); mean (SD) | 101.2 (3.7) | 101.2 (3.6) |
| Double limb support time for painful limb (ms); mean (SD) | 207.2 (41.2) | 205.0 (42.4) |
| Double limb support time for non-painful limb (ms); mean (SD) | 205.7 (41.1) | 203.7 (42.0) |
| Double limb support time LSI (%); mean (SD) | 99.4 (4.4) | 99.5 (4.3) |
LSI – Limb symmetry index (non-painful/painful*100)
3.2. Cross-sectional relation of unilateral knee pain severity to temporal asymmetry during self-selected and fast walking
At both self-selected and fast walking speeds, unilateral knee pain severity was not significantly associated with LSIs for stance time, single limb support time, or double limb support time (Figure 2).
Figure 2.
Cross-sectional relations of unilateral knee pain severity to measures of temporal asymmetry in stance (A&D), single limb support (B&E), and double limb support (C&F) during self-selected (A-C) and fast (D-F) walking. Beta coefficients are reported per 20-unit increase in Visual Analog Scale (VAS) pain score. Adjusted models included age, sex, body mass, race, the presence of depressive symptoms, gait speed, and KL grades as covariates.
3.3. 2-year knee pain outcomes
Out of the 105 participants with unilateral knee pain at the 60-month clinic visit who were eligible for longitudinal analyses, 18 (17.1%) had contralateral knee pain (at least mild pain with walking or a VAS pain score of at least 20) and 54 (51.4%) had persistent ipsilateral knee pain (at least mild pain with walking and a VAS pain score greater than zero) at the 84-month clinic visit. Results for each longitudinal analysis are reported in Table 3. Notably, after adjusting for age, sex, body mass, race, depressive symptoms, gait speed, and KL grades, a 2.5% increase in stance and single limb support LSIs during self-selected walking were associated with a 51% (OR=0.49; 95% CI=0.27–0.91) and 37% (OR=0.63, 95% CI=0.41–0.96) decreased odds of incident contralateral knee pain, respectively. Double limb support time LSI was not significantly associated with incident contralateral knee pain (Table 3). Furthermore, temporal measures during fast walking were not associated with incident contralateral knee pain (Table 3).
Table 3.
Relation of temporal asymmetry measures during self-selected and fast walking to 2-year knee pain outcomes
| Incident Contralateral Knee Pain | Persistent Ipsilateral Knee Pain | |||||
|---|---|---|---|---|---|---|
| Unadjusted OR (95%CI) |
Adjusted OR (95%CI)* |
Adjusted OR (95%CI)† |
Unadjusted OR (95%CI) |
Adjusted OR (95% CI)* |
Adjusted OR (95%CI)† |
|
| Self-Selected Walking | ||||||
| Stance time LSI | 0.57 (0.35–0.94) | 0.49 (0.27–0.91) | 0.46 (0.17–1.23) | 1.11 (0.79–1.56) | 1.07 (0.75–1.53) | 1.22 (0.79–1.89) |
| Single limb support time LSI | 0.78 (0.56–1.08) | 0.63 (0.41–0.96) | 0.37 (0.14–0.98) | 1.25 (0.97–1.60) | 1.22 (0.95–1.58) | 1.30 (0.93–1.82) |
| Double limb support time LSI | 0.76 (0.54–1.09) | 0.86 (0.59–1.25) | 0.92 (0.58–1.44) | 0.79 (0.61–1.02) | 0.77 (0.58–1.00) | 0.87 (0.65–1.17) |
| Fast Walking | ||||||
| Stance time LSI | 0.58 (0.33–1.00) | 0.52 (0.25–1.11) | 0.59 (0.20–1.75) | 1.24 (0.83–1.86) | 1.33 (0.86–2.08) | 1.68 (0.97–2.91) |
| Single limb support time LSI | 0.72 (0.49–1.05) | 0.67 (0.41–1.09) | 0.67 (0.41–1.09) | 1.27 (0.96–1.68) | 1.32 (0.96–1.82) | 1.59 (1.05–2.41) |
| Double limb support time LSI | 0.97 (0.73–1.30) | 1.05 (0.75–1.48) | 0.97 (0.60–1.58) | 0.93 (0.74–1.16) | 0.97 (0.76–1.25) | 1.00 (0.76–1.31) |
All odds ratios and 95% confidence intervals are reported per 2.5% increase in exposure.
Models adjusted for age, sex, body mass, race, the presence of depressive symptoms, gait speed, and Kellgren Lawrence grades;
Models additionally adjusted for average steps per day. Sample size of 84 due to missing or invalid physical activity data.
LSI – Limb symmetry index (non-painful/painful*100)
Temporal asymmetry measures were not associated with persistent ipsilateral knee pain, except for single limb support LSI during fast walking when adjusting for average steps per day (Table 3). This relation showed that a 2.5% increase in single limb support LSI is associated with a 59% (OR=1.59, 95% CI=1.05–2.41) increased odds of persistent ipsilateral knee pain.
When adding average steps per day to the models in sensitivity analyses, confidence intervals widened and point estimates changed slightly for each outcome (Table 3). The only exceptions where the point estimates appeared to change more than slightly were for relations of single limb support time and for stance time during fast walking (Table 3).
4. DISCUSSION
The objectives of this study were to determine if unilateral knee pain is associated with temporal asymmetry while walking, and subsequently if the degree of temporal asymmetry influences knee pain outcomes two years later. Our findings indicate that the severity of unilateral knee pain is not related to temporal asymmetry while walking at self-selected or fast speeds in persons with mostly mild or moderate knee pain. In other words, individuals with moderate unilateral knee pain do not spend more time on their contralateral limb during walking any more than those with mild pain. Because our sample for cross-sectional analyses included only one individual with severe knee pain with walking, further work is needed to determine if more severe knee pain has a greater effect on temporal asymmetry. From our longitudinal analyses, we found that relatively longer stance and single limb support times on the non-painful limb were protective against developing contralateral knee pain in those with unilateral knee pain. On the other hand, spending relatively more time in single limb support on the contralateral limb during fast walking was associated with persistent pain in the originally painful knee. Taken together, measures of temporal asymmetry during self-selected and fast walking may be useful for predicting knee pain outcomes.
Prior to determining if temporal asymmetry during walking influenced knee pain outcomes two years later, we were interested in understanding how unilateral knee pain severity affects temporal asymmetry. Because antalgic gait (i.e., a pain-avoidant walking pattern defined by spending more time on the non-painful limb compared to the painful limb) is a hallmark sign of unilateral lower body pain that seems more pronounced in those with worse pain, we hypothesized that more severe unilateral knee pain would be associated with relatively longer stance and single limb support times on the non-painful limb and relatively shorter double limb support times on the non-painful limb. Surprisingly, we found that unilateral knee pain severity was not associated with measures of temporal asymmetry. The lack of association may be a result of how the walking evaluation was performed and who was included in the sample. Specifically, because the walking evaluation was brief, it may not have triggered pain or gait modifications. As for the sample, relationships may have been more evident if individuals with severe or extreme unilateral knee pain with walking were included. Upon further analysis, we determined that our sample was predominantly individuals with mild or moderate unilateral knee pain because persons in the MOST cohort with severe or extreme knee pain with walking had at least some pain in their contralateral knee, which made them ineligible for the current study. Another characteristic of the sample that may have affected our results is chronicity of pain. Little is known in this area, but chronic knee pain may affect walking patterns differently than acute knee pain.
We expected the non-painful limb to have longer stance and single limb support times and shorter double limb support times relative to the painful limb and be associated with greater odds of developing knee pain in the non-painful limb. This hypothesis was grounded in prior research showing that increased loading contributes to knee OA progression (14–18). Our results indicate that relatively longer stance and single limb support times on the non-painful limb do not increase the likelihood of developing contralateral knee pain in those with mild-to-moderate unilateral knee pain. In fact, the directionality of the associations suggest that spending relatively more time on non-painful limb is protective against developing contralateral knee pain. This raises the possibility that increased stance and single limb support times could be associated with loading the knee in a way that is beneficial to the joint. For example, longer stances times may be associated with greater overall load (i.e., total impulse) per step, but reduced peak forces and loading rates, which have been historically considered harmful to the knee joint (28, 29). On the other hand, our assumption that increased loading is harmful to the joint should be challenged in future research. The idea, however, that increased loading may be beneficial should not be considered novel as physical therapy, physical activity, and walking programs increase loading and are recommended for the management of knee OA (19–21).
As for the originally painful knee, we expected the odds of persistent pain to be reduced for those who walk with relatively longer stance and single limb support times as well as shorter double limb support times on the non-painful limb. We hypothesized that pain would improve because these temporal patterns suggest that load is being shifted away from the painful knee. Contrary to our expectations, the results showed that measures of temporal asymmetry at self-selected walking speed are not associated with the originally painful knee still being painful two years later. This finding is surprising because it suggests that individuals with mild-to-moderate unilateral knee pain who spend more time on their non-painful limb while walking are just as likely to have continued pain as those who spend more time on their painful limb or neither limb (i.e., relatively symmetric). When considering this alongside the fact that 51% of our sample no longer had pain in their originally painful limb 2 years later, it seems possible that our results could have been affected by including those with persistent or transient knee pain. Future work in this area is needed to understand how different pain patterns influence walking patterns and their downstream effects. It is worth noting, however, that spending more relative time in single limb support on the contralateral limb during fast walking increased the odds of having continued ipsilateral knee pain when accounting for physical activity.
There are several limitations to consider for this study. Our primary limitation is that our sample included only individuals with mild or moderate pain with walking. We expected that more persons from the MOST cohort would have severe or extreme unilateral knee pain with walking. Likely because of this, our sample largely demonstrated only subtle temporal asymmetries. With only limited distributions, we were limited in our ability to evaluate its relation to the pain outcomes of interest, which themselves had low incidence over two years. To advance our understanding about the pathway from unilateral to bilateral knee OA, future research may need to either evaluate walking differently or include individuals with more severe pain. Another limitation is that temporal measures of gait may not be good surrogates for knee joint loading, and therefore may not capture the construct we are trying to assess. Future research should aim to elucidate relations between gait metrics that are simple to obtain in clinical and real-world settings (i.e., spatiotemporal measures) and estimates of knee joint loading. Without a better understanding of these relations, we may fail to leverage metrics provided by wearable and portable technologies for improving outcomes for individuals with knee OA. Another limitation of our study is that two years may not be enough time for gait modifications to influence outcomes. Additionally, the gait assessment may have been too short for asymmetric stance time to be identified since it may require longer duration of walking before such asymmetry may become evident. Regardless, this is the first study to leverage longitudinal, prospective data to assess cross-sectional and longitudinal relations of temporal asymmetry to pain outcomes in those with unilateral knee pain. Another limitation is that our sensitivity analyses that included physical activity may have been underpowered. Only 79% (n=84) of our sample had valid physical activity data, which lowered our sample size and limited our power by adding it as a covariate. Future biomechanics research in knee OA should strongly consider not only the quality of movement, but also the quantity.
In conclusion, for those with mild-to-moderate unilateral knee pain, spending relatively more time in stance and single limb support on the non-painful limb can be considered a biomarker associated with the contralateral knee remaining pain-free. However, it does not appear to be a biomarker for having persistent pain in the already painful knee. These findings suggest that measures of temporal asymmetry hold little value in explaining the progression from unilateral to bilateral knee OA. Further research is needed to elucidate the mechanisms that explain why nearly every individual with unilateral knee OA progresses to bilateral disease.
Significance and Innovations.
Temporal measures of gait may be surrogates of limb loading that clinicians can use to detect aberrant gait patterns in those with knee pain or osteoarthritis.
Temporal asymmetry during walking is not associated with pain severity in those with unilateral knee pain.
For those with unilateral knee pain, spending relatively more time on the non-painful limb while walking is associated with a decreased risk of developing contralateral knee pain and an increased risk of continuing to have ipsilateral knee pain.
ACKNOWLEDGEMENTS
MOST is funded by the National Institute on Aging (NIA) under the following award numbers U01AG18820 (PI: Felson), U01AG18832 (PI: Torner), U01AG18947 (PI: Lewis), and U01AG19069 (PI: Nevitt). Patrick Corrigan was supported by the Rheumatology Research Foundation and the Academy of Orthopaedic Physical Therapy. Joshua Stefanik was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) under award number K23AR070913. Tuhina Neogi was supported by NIAMS under award number K24AR070892. Other support for this study was provided by the Boston University Core Center for Clinical Research under award number P30AR072571. Funding sources had no role in the design, conduct, or reporting of this study.
Footnotes
CONFLICTS OF INTEREST
PC – None; DTF – None; CLL – None; TN – None; MPL – None; KDG – None; MCN – None; CEL – None; JCT – None; JJS – None.
REFERENCES
- 1.Guccione AA, Felson DT, Anderson JJ, Anthony JM, Zhang Y, Wilson PW, et al. The effects of specific medical conditions on the functional limitations of elders in the Framingham Study. Am J Public Health. 1994;84(3):351–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cross M, Smith E, Hoy D, Nolte S, Ackerman I, Fransen M, et al. The global burden of hip and knee osteoarthritis: estimates from the global burden of disease 2010 study. Ann Rheum Dis. 2014;73(7):1323–30. [DOI] [PubMed] [Google Scholar]
- 3.Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M, et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2163–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Metcalfe AJ, Andersson ML, Goodfellow R, Thorstensson CA. Is knee osteoarthritis a symmetrical disease? Analysis of a 12 year prospective cohort study. BMC Musculoskelet Disord. 2012;13:153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jones RK, Chapman GJ, Findlow AH, Forsythe L, Parkes MJ, Sultan J, et al. A new approach to prevention of knee osteoarthritis: reducing medial load in the contralateral knee. J Rheumatol. 2013;40(3):309–15. [DOI] [PubMed] [Google Scholar]
- 6.White DK, Zhang Y, Felson DT, Niu J, Keysor JJ, Nevitt MC, et al. The independent effect of pain in one versus two knees on the presence of low physical function in a multicenter knee osteoarthritis study. Arthritis Care Res (Hoboken). 2010;62(7):938–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.de Rooij M, van der Leeden M, Heymans MW, Holla JF, Hakkinen A, Lems WF, et al. Prognosis of Pain and Physical Functioning in Patients With Knee Osteoarthritis: A Systematic Review and Meta-Analysis. Arthritis Care Res (Hoboken). 2016;68(4):481–92. [DOI] [PubMed] [Google Scholar]
- 8.Bindawas SM, Vennu V, Al Snih S. Differences in health-related quality of life among subjects with frequent bilateral or unilateral knee pain: data from the Osteoarthritis Initiative study. J Orthop Sports Phys Ther. 2015;45(2):128–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Worsley P, Stokes M, Barrett D, Taylor M. Joint loading asymmetries in knee replacement patients observed both pre- and six months post-operation. Clin Biomech (Bristol, Avon). 2013;28(8):892–7. [DOI] [PubMed] [Google Scholar]
- 10.Briem K, Snyder-Mackler L. Proximal gait adaptations in medial knee OA. J Orthop Res. 2009;27(1):78–83. [DOI] [PubMed] [Google Scholar]
- 11.Gustafson JA, Anderton W, Sowa GA, Piva SR, Farrokhi S. Dynamic knee joint stiffness and contralateral knee joint loading during prolonged walking in patients with unilateral knee osteoarthritis. Gait Posture. 2019;68:44–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mills K, Hettinga BA, Pohl MB, Ferber R. Between-limb kinematic asymmetry during gait in unilateral and bilateral mild to moderate knee osteoarthritis. Arch Phys Med Rehabil. 2013;94(11):2241–7. [DOI] [PubMed] [Google Scholar]
- 13.Creaby MW, Bennell KL, Hunt MA. Gait differs between unilateral and bilateral knee osteoarthritis. Arch Phys Med Rehabil. 2012;93(5):822–7. [DOI] [PubMed] [Google Scholar]
- 14.Messier SP, Beavers DP, Herman C, Hunter DJ, DeVita P. Are unilateral and bilateral knee osteoarthritis patients unique subsets of knee osteoarthritis? A biomechanical perspective. Osteoarthritis Cartilage. 2016;24(5):807–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Miyazaki T, Wada M, Kawahara H, Sato M, Baba H, Shimada S. Dynamic load at baseline can predict radiographic disease progression in medial compartment knee osteoarthritis. Ann Rheum Dis. 2002;61(7):617–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sharma L, Chang AH, Jackson RD, Nevitt M, Moisio KC, Hochberg M, et al. Varus Thrust and Incident and Progressive Knee Osteoarthritis. Arthritis Rheumatol. 2017;69(11):2136–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chehab EF, Favre J, Erhart-Hledik JC, Andriacchi TP. Baseline knee adduction and flexion moments during walking are both associated with 5 year cartilage changes in patients with medial knee osteoarthritis. Osteoarthritis Cartilage. 2014;22(11):1833–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chang AH, Moisio KC, Chmiel JS, Eckstein F, Guermazi A, Prasad PV, et al. External knee adduction and flexion moments during gait and medial tibiofemoral disease progression in knee osteoarthritis. Osteoarthritis Cartilage. 2015;23(7):1099–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Brisson NM, Wiebenga EG, Stratford PW, Beattie KA, Totterman S, Tamez-Pena JG, et al. Baseline knee adduction moment interacts with body mass index to predict loss of medial tibial cartilage volume over 2.5 years in knee Osteoarthritis. J Orthop Res. 2017;35(11):2476–83. [DOI] [PubMed] [Google Scholar]
- 20.Bannuru RR, Osani MC, Vaysbrot EE, Arden NK, Bennell K, Bierma-Zeinstra SMA, et al. OARSI guidelines for the non-surgical management of knee, hip, and polyarticular osteoarthritis. Osteoarthritis Cartilage. 2019;27(11):1578–89. [DOI] [PubMed] [Google Scholar]
- 21.Loew L, Brosseau L, Wells GA, Tugwell P, Kenny GP, Reid R, et al. Ottawa panel evidence-based clinical practice guidelines for aerobic walking programs in the management of osteoarthritis. Arch Phys Med Rehabil. 2012;93(7):1269–85. [DOI] [PubMed] [Google Scholar]
- 22.Kolasinski SL, Neogi T, Hochberg MC, Oatis C, Guyatt G, Block J, et al. 2019 American College of Rheumatology/Arthritis Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee. Arthritis Care Res (Hoboken). 2020;72(2):149–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Segal NA, Nevitt MC, Gross KD, Hietpas J, Glass NA, Lewis CE, et al. The Multicenter Osteoarthritis Study: opportunities for rehabilitation research. PM R. 2013;5(8):647–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.van Uden CJ, Besser MP. Test-retest reliability of temporal and spatial gait characteristics measured with an instrumented walkway system (GAITRite). BMC Musculoskelet Disord. 2004;5:13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Webster KE, Wittwer JE, Feller JA. Validity of the GAITRite walkway system for the measurement of averaged and individual step parameters of gait. Gait Posture. 2005;22(4):317–21. [DOI] [PubMed] [Google Scholar]
- 26.Tubach F, Ravaud P, Baron G, Falissard B, Logeart I, Bellamy N, et al. Evaluation of clinically relevant changes in patient reported outcomes in knee and hip osteoarthritis: the minimal clinically important improvement. Ann Rheum Dis. 2005;64(1):29–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Patterson KK, Gage WH, Brooks D, Black SE, McIlroy WE. Evaluation of gait symmetry after stroke: a comparison of current methods and recommendations for standardization. Gait Posture. 2010;31(2):241–6. [DOI] [PubMed] [Google Scholar]
- 28.Mundermann A, Dyrby CO, Andriacchi TP. Secondary gait changes in patients with medial compartment knee osteoarthritis: increased load at the ankle, knee, and hip during walking. Arthritis Rheum. 2005;52(9):2835–44. [DOI] [PubMed] [Google Scholar]
- 29.Ewers BJ, Jayaraman VM, Banglmaier RF, Haut RC. Rate of blunt impact loading affects changes in retropatellar cartilage and underlying bone in the rabbit patella. J Biomech. 2002;35(6):747–55. [DOI] [PubMed] [Google Scholar]