Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 16.
Published in final edited form as: Clin Biomech (Bristol). 2021 Jul 17;88:105427. doi: 10.1016/j.clinbiomech.2021.105427

Walking Energetics and Fatigue are Associated with Physical Activity in People with Knee Osteoarthritis

Kharma C Foucher 1, Burcu Aydemir 1, Chun-Hao Huang 1
PMCID: PMC9841508  NIHMSID: NIHMS1859270  PMID: 34303950

Abstract

Background:

Aberrant biomechanics may influence osteoarthritis-associated physical activity limitations. Our purpose was to evaluate the association of walking energetics, fatigue, and fatigability on physical activity in people with knee osteoarthritis. We hypothesized that using increased energy for walking, experiencing more fatigue, or being more fatigable are associated with less activity, and that fatigue and fatigability mediate the relationships between walking energetics and physical activity.

Methods:

We tested our hypothesis in 30 people with knee osteoarthritis (age 58 ± 9 years, 10 Male/20 Female). Physical activity was assessed using the University of California Los Angeles score. We used a six-minute walk test to predict VO2max. Next we used a portable oxygen exchange system to measure relative energy used (100 * VO2rate/VO2max) and VO2cost during walking at preferred speeds. We used the Knee injury and Osteoarthritis Outcome Score subscale to quantify pain, and the Patient Reported Outcome Measurement Instrument System Fatigue survey and a treadmill-based fatigability test to assess fatigue and fatigability. Spearman correlations, regression, and mediation analysis were used to test our hypotheses.

Findings:

Greater energy used during walking, fatigue, and fatigability were all associated with lower physical activity (rho=−0.585 to −0.379, P = 0.001 to 0.043). These associations persisted when incorporating pain into the models. Fatigue and fatigability mediated the associations between walking energetics and physical activity.

Interpretation:

Walking energetics could be a useful target to promote physical activity in people with osteoarthritis. Further, the effect of walking energetics on physical activity may work through its impact on fatigability.

Keywords: osteoarthritis, physical activity, fatigue, walking energetics

1. Introduction

Osteoarthritis (OA) and aging are independently associated with activity limitations and reduced levels of physical activity (Centers for Disease Control and Prevention, 2011; Murphy et al., 2017). Physical activity, however, has great benefits for those with OA. Physical activity is associated with reduced pain and better function in OA (Fransen et al., 2015; Skou and Roos, 2019), and is important for healthy aging more broadly. There are many potential reasons that OA is associated with reduced physical activity. A better understanding of these factors is needed so that we can design new interventions to effectively promote physical activity.

Energy capacity has been proposed as a factor that limits activity in older adults (Schrack et al., 2010; Schrager et al., 2014). It has been proposed that age-related reductions in total energy capacity (Fleg et al., 2005) combined with inefficient walking patterns, i.e. using excessive energy for walking, can result in having less “energy reserve” that can be used for lifestyle physical activity. Moreover, people with less energy reserve may be inclined to conserve energy to stave off fatigue. This model, dubbed the energetic model of activity limitation, has been used to understand physical activity restriction in older adults with mobility disability (Schrager et al., 2014; Wert et al., 2013), and even to develop interventions (VanSwearingen et al., 2011). More recently, this model has been modified to include the role of pain. The Pain Energy model posits that pain itself can result in increased energy expenditure during walking, via alterations in motor strategies (Coyle et al., 2018). Thus, pain may exacerbate the role of energetics in limiting physical activity. These models have not been specifically evaluated in people with knee OA.

In addition to pain, fatigue is a prominent symptom of OA (Power et al., 2008; Snijders et al., 2011). Fatigue refers to an overall subjective feeling of tiredness or lack of energy (Eldadah, 2010). Fatigability is a separate, but related concept that can be defined as movement-related fatigue (Kim et al., 2018). Fatigability can be measured in a repeatable way by evaluating change in tiredness or performance over a standardized physical task (Eldadah, 2010; Kim et al., 2018). Two individuals can report the same amount of fatigue but have different responses to an activity challenge. Thus, fatigue and fatigability may measure different constructs. Both fatigue and fatigability can limit physical activity in people with OA (Egerton et al., 2016b, 2016a; Murphy et al., 2016). Based on the energetic model and the pain energy model, we hypothesize that the reduced energy capacity associated with aging (Fleg et al., 2005), and gait inefficiency can exacerbate fatigue or fatigability and thereby limit physical activity. Pain may further exacerbate fatigue in this population.

The purpose of this study was to evaluate the associations of physical activity with walking energetics, fatigue, and fatigability in people with knee OA in the context of pain. We tested three hypotheses. First, we hypothesized that walking energetics, specifically Relative of energy used for walking relative to total energy capacity, are associated with physical activity. Second, we hypothesized that fatigue and fatigability are associated with physical activity. Third, we hypothesized that walking energetics are associated with fatigue and fatigability. Finally, we determined the extent to which fatigue and fatigability mediate the association between walking energetics and physical activity.

2. Methods

2.1. Subjects

This cross-sectional study was approved by the University of Illinois at Chicago (Chicago, Illinois, USA) Institutional Review Board (IRB). We recruited participants from an IRB-approved contact list obtained from medical records. The inclusion criteria were diagnosed knee OA (based on ICD-10 codes in the medical record) and the self-reported ability to walk for 15 minutes without stopping. Exclusion criteria included other actively symptomatic joints, history of total joint replacement within 2 years, inability to walk without assistive devices, and any medical condition that interfered with gait or the ability to safely complete the protocol. Data from 30 people who satisfied enrollment criteria and provided written informed consent to participate were used in this study. Subjects were clinically characterized using the Knee injury and Osteoarthritis Outcome Score (KOOS) (Roos and Lohmander, 2003).

2.2. Walking energetics

Relative energy use during walking was our primary measure of walking energetics. First total aerobic capacity was assessed by predicting the VO2max from performance and heart rate during a treadmill-based six-minute walk test using a published regression equation (Laskin et al., 2007). The regression equation estimates VO2max based on heart rate, body mass, and distance walked: VO2max (L/min) = −1.732 + (weight [kg] × 0.049) + (distance [m] × 0.005) + (HR [beats/min] × [−0.015]) (Laskin et al., 2007). For this test, subjects were asked to walk “as far as possible”. Subjects were encouraged throughout the test to maximize distance but were allowed to alter the treadmill speed up or down or take breaks. Heart rate was measured during the test using a sensor strapped to the subjects’ chests (Garmin 010-10997-00, Olathe, KS, USA). Predicted VO2max was calculated, normalized to body mass, and reported as ml/min*kg. Next, during a separate evaluation, VO2 in ml/min*kg used during treadmill walking at preferred speeds was assessed using a portable gas exchange system (COSMED K5, Concord, CA, USA) after a 3 minute period of familiarization and 2.5 minutes of steady state walking. Relative energy use during walking was calculated as 100* VO2 during walking at preferred speeds/total aerobic capacity. We also measured VO2 cost per unit distance (ml/kg*m) during the steady-state walking period as a second measure of walking energetics. VO2 cost indicates the amount of energy needed to walk a unit of distance.

2.3. Fatigue and fatigability

Self-reported fatigue was assessed using the Patient-Reported Outcomes Measurement Information System (PROMIS) Fatigue version 1.0 Computerized Adaptive Test (CAT) (Broderick et al., 2013; Cella et al., 2016). Raw scores are converted to T-Scores, which are based on a representative sample of the US population and range from 0 to 100. A T-score of 50 represents the average of the representative population with a standard deviation of 10. For PROMIS instruments, higher scores represent having more of the given construct. Thus, a T-score of 60 represents a fatigue level that is one standard deviation more severe than average.

Fatigability was measured using a modified assessment previously developed and validated in older adults (Schnelle et al., 2012). Subjects walked on a treadmill for 10 minutes with a 3 minute acclimation period. They began walking at their preferred speed for 2.5 minutes. Then, after every 2.5 minute interval they were given the opportunity to reduce or increase their speed to their comfort. They indicated their preference by hand signals. The treadmill speed was increased or decreased in 0.2 m/s increments until the participant indicated that the new comfortable speed had been reached. Before beginning the test, and then again after completion, the test subjects were asked to rate their tiredness on a scale of 1 to 7, where 1 represented extremely energetic and 7 represented extremely tired. The fatigability score was calculated as the difference in tiredness divided by the distance walked. Higher scores reflect greater levels of fatigability.

2.4. Physical activity

Physical activity was characterized using the University of California, Los Angeles (UCLA) activity score. This score is widely used in orthopedic populations, has been validated against pedometers (Zahiri et al., 1998), and shows good construct validity (Terwee et al., 2011). The UCLA activity score assesses self-reported activity level ranging from a score of 1 – “Wholly inactive; dependent on others; cannot leave residence” to 10 “Regularly participate in impact sports such as jogging, tennis, skiing, acrobatics, ballet, heavy labor, or backpacking.”

2.5. Statistical Analysis

In the absence of prior studies in this area, we used G*Power (Faul et al., 2009, 2007) to determine an appropriate sample size to detect a medium effect size (R=0.5). We determined that 29 participants would be needed to detect a correlation coefficient of this magnitude with 80% power at the alpha < 0.05 level. For a linear regression, with 4 total predictors, and each term contributing a f2 of 0.35 (large effect size), we determined that 25 participants would be needed. Thus, we were adequately powered for the analyses. We note, however, that the sample size calculation was not prespecified. Subsequent statistical analyses were conducted using R version 4.1.0. Before assessing hypotheses, first, all variables were investigated for normality using the Shapiro-Wilk test. After finding that several variables were not normally distributed, we decided to use Spearman correlations to investigate our hypothesized associations. Next, we assessed the association of the UCLA score with age and BMI to determine whether these factors needed to be accounted for statistically when testing our hypotheses using Pearson correlations. Then, we assessed the association of KOOS pain scores with our variables of interest. Note that for additional context, we also assessed the association between UCLA scores and the other domains of the KOOS.

To evaluate hypothesis 1, that walking energetics are associated with physical activity, we used Spearman correlations to test the association between relative energy used for walking or VO2cost and UCLA scores. We then used regression analysis, using a log transformation to account for the non-normality, to confirm that these relationships were the same when including pain as a covariate. We used the same procedure to evaluate hypothesis 2, that fatigue and fatigability are associated with physical activity, with Spearman correlations to test the association between these variables and UCLA scores, and regression analysis to confirm the relationship including pain as a covariate. To test hypothesis 3, that fatigue and fatigability mediate the association between walking energetics and physical activity we used mediation analysis. Because we hypothesized that fatigue and fatigability operate through different pathways, we used parallel multiple mediation analysis using the PROCESS macro in R to simultaneously test the indirect effects of both factors (Hayes, 2017). In these analyses, pain was also included as a covariate.

3. Results

Demographic and physical characteristics of the participants (Table 1) were not associated with physical activity. Specifically, there were no associations between age or BMI and UCLA activity scores (respectively, rho=−0.159, P=0.409 and rho=−0.251, P=0.190) and there were no differences in activity levels between men and women (P=0.192). There was, however, a significant association between all KOOS scores and UCLA scores with the exception of Sport/Recreation Function (Table 2). Further pain scores were associated with walking energetics (relative energy for walking rho = −0.379, P = 0.043; VO2 cost rho = −0.407, P = 0.029), and fatigue (rho = −0.684, P < 0.001), but not fatigability (rho = −0.203, P = 0.294). Because of the number of relevant significant relationships (Fig. 1), and in keeping with the Pain-Energy model, pain was included in subsequent regression models.

Table 1.

Physical characteristics, clinical characteristics, and activity levels of the 20 female and 10 male participants.

Mean ± standard deviation Median (IQR) Range
Age (years) 58.3 ± 9.3 60 (11) 35 - 75
BMI (kg/m2) 33.5 ± 5.8 33.5 (7.8) 23.3 - 45.0
KOOS Pain Score 53.9 ± 22.4 55.6 (33.3) 13.9 - 100
KOOS Symptoms Score 54.3 ± 23.3 50.0 (32.0) 3.6 - 100
KOOS Function (ADL) Score 58.3 ± 23.5 60.9 (28.1) 14.1 - 100
KOOS Function (Sport/Rec) Score 34.5 ± 30.9 25.0 (40.0) 0 - 100
KOOS Quality of Life Score 34.9 ± 22.8 31.3 (31.3) 0.0 - 93.8
Fatigue T-score 53.9 ± 10.9 55.4 (16.0 31.9 - 74.3
Fatigability score (x10−3) 5.5 ± 7.5 3.0 (10.3) −4.6 - 25
Walking speed (m/s) 0.61 ± 0.22 0.55 (0.38) 0.40 - 1.08
Relative energy used during walking 36.3 ± 17.3 30.5 (19.5) 17.2 - 79.3
VO2 cost (ml/kg*m) 0.28 ± 0.12 0.26 (0.16) 0.15 - 0.54
Six-minute walking test distance (m) 290 ± 156 274 (249) 64 - 563
UCLA Activity Score 5.1 ± 2.1 5 (3) 2 - 10
Open in a new tab

Table 2.

Associations between KOOS subscores and UCLA scores.

Pain Symptoms ADL Function Sport/Recreation Function Quality of Life
UCLA Scores rho=0.570 P=0.001 rho=0.598 P < 0.001 rho=0.498 P=0.006 rho=0.305 P=0.176 rho=0.533 P=0.003
Open in a new tab

Fig. 1.

Fig. 1.

Open in a new tab

Reporting more pain was associated with (a) lower activity levels, (b) using more energy for walking relative to total energy capacity, (c) a higher energy cost of walking, (d) more fatigue).

Walking energetics were significantly associated with physical activity in accordance with our first hypothesis. People who used more energy for walking relative to their total energy capacity had lower UCLA scores (Fig. 2a, rho=−0.434, P=0.019). The R2 value for a model with ln(UCLA scores) as the dependent variable and ln(relative energy used for walking) as the predictor variable, including KOOS pain scores, was 0.366 (P=0.003). People with a higher VO2 cost also had lower UCLA scores (Fig. 2b, rho=−0.554, P=0.002). The R2 value for a model with ln(UCLA scores) as the dependent variable and ln(VO2 cost) as the predictor variable, including KOOS pain scores, was 0.436 (P < 0.001).

Fig. 2.

Fig. 2.

Open in a new tab

(a) People who used more energy for walking relative to their total energy capacity had lower UCLA scores (rho = −0.485, P = 0.007). (b) People with a higher VO2 cost also had lower UCLA scores (rho = −0.571, P = 0.001). R2 values shown include pain as a covariate.

Fatigue and fatigability were associated with physical activity in accordance with our second hypothesis (Fig. 3). People who reported more overall fatigue were less physically active (rho=−0.606, P < 0.001). The R2 value for a model with ln(UCLA scores) as the dependent variable and PROMIS fatigue scores as the predictor variable, controlling for KOOS pain scores, was 0.374 (P=0.002). People who were more fatigable were also less physically active (rho=−0.634, P < 0.001). The R2 value for a model with ln(UCLA score) as the dependent variable and fatigability scores as the predictor variable, controlling for KOOS pain scores, was 0.590 (P < 0.001).

Fig. 3.

Fig. 3.

Open in a new tab

(a) People who reported more overall fatigue were less physically active (rho = −0.588, P = 0.001). (b) People who were more fatigable were also less physically active (rho = −0.665, P < 0.001). R2 values shown include pain as a covariate.

Relative energy used for walking was not significantly correlated with fatigue at the 0.05 level (rho=0.345, P=0.067) but was significantly correlated with fatigability scores (rho=0.374, P=0.045). The same trends were seen for VO2 cost. VO2 cost was not associated with fatigue (rho=0.259, P=0.175) but was associated with fatigability was 0.570 (P=0.001). Fatigue and fatigability were not correlated with each other (rho=0.317, P=0.094).

In the parallel mediation analysis model, fatigability mediates the association between relative energy used for walking and UCLA scores. The 95% bootstrapped confidence interval for the effect of fatigability score did not include zero (Fig. 4). By contrast, the confidence interval for the effect of fatigue did include zero, which indicates a lack of mediation. Notably, the 95% confidence interval for the direct effect of relative energy used for walking on UCLA scores included zero. This indicated that the influence of this variable on physical activity is exerted entirely through fatigability in our model. Including pain as a covariate, the four variables explained 57.7% of the variance in UCLA scores (Table 3).

Fig. 4.

Fig. 4.

Open in a new tab

(a) Direct and indirect effects of relative energy use during walking on UCLA scores through fatigue and fatigability adjusting for pain (not shown). (b) Direct and indirect effects of VO2 cost on UCLA scores through fatigue and fatigability adjusting for pain (not shown). Unstandardized regression coefficients (percentile bootstrap 95% confidence interval) shown.

Table 3.

Walking energetics, fatigue, fatigability, and pain predict self-reported physical activity measured as the UCLA activity score.

Dependent variable R2 P Independent variable B (95% CI) Standardized β P
UCLA scores 0.577 <0.001 Relative energy used during walking −0.001 (−0.039, 0.038) −0.006 0.971
Fatigue Score −0.045 (−0.120, 0.030) −0.231 0.228
Fatigability Score −121.6 (−207.6, −35.6) −0.435 0.008
Pain Score 0.031 (−0.006, 0.067) 0.328 0.093
UCLA scores 0.584 <0.001 VO2 cost −2.0 (−8.6, 4.6) −0.112 0.527
Fatigue Score −0.05 (−0.12, 0.03) −0.253 0.191
Fatigability Score −105.9 (−201.2, −10.7) −0.379 0.031
Pain Score 0.026 (−0.01, 0.06) 0.277 0.177
Open in a new tab

Fatigability similarly mediated the association between VO2 cost and UCLA scores. The 95% bootstrapped confidence interval for the effect of fatigability score did not include zero (Fig. 4). The confidence interval for the effect of fatigue also included zero, which similarly indicates a lack of mediation. The 95% confidence interval for the direct effect of VO2 cost on UCLA scores again included zero (Fig. 4). However, in the full model, VO2 cost, fatigue, fatigability, and pain together explained 58% of the variance in UCLA scores (Table 3). Fatigability was the best predictor of UCLA scores based on the magnitude of the standardized coefficients.

4. Discussion

The purpose of this study was to investigate the associations of physical activity with walking energetics, fatigue, and fatigability in people with knee OA. The rationale was that there could be physical barriers to physical activity specific to people with knee OA and that the knowledge of these barriers could lead to more effective physical activity promoting interventions. We found that using more energy for walking, whether measured as relative total energy capacity or oxygen consumption per unit distance, was associated with reduced physical activity. Higher fatigue and fatigability were also associated with reduced physical activity. Fatigability mediated the associations between walking energetics and physical activity.

The theoretical underpinnings of this study were the energetic model of activity limitation, which was described as a way to understand activity limitation in older adults, (Schrack et al., 2010; Schrager et al., 2014; Wert et al., 2013) and the pain energy model, which was described as a way to understand the contributing role of pain in limiting activity (Coyle et al., 2018). As people age, their total energy capacity (aerobic fitness) tends to decrease (Fleg et al., 2005). Thus, the amount of energy used for walking relative to total energy capacity is increased, and the leftover energy or “energy reserve” is decreased, even in the absence of mobility impairment. It has been hypothesized that with lower energy reserve, the ability to be physically active will decrease (Schrack et al., 2010). Chronic pain has further been shown to increase energy cost of walking (Coyle et al., 2018) in people with musculoskeletal pain, including hip impairment (Gussoni et al., 1990). In accordance with this hypothesis, we found a negative relationship between measures of walking energetics and UCLA activity scores.

We postulate that the gait adaptations associated with painful knee OA are energetically costly and exacerbate the normal reduction of energy reserve that accompanies aging. Further, we speculate that increasing pain is accompanied by a variety of gait-based mitigation strategies, that may have different effects on energy use. This variability may have led to increased variability in the association between pain and energy use at higher pain scores. It is also possible that habitual gait speed, which is known to be reduced in people with knee OA, may slow below the speed that minimizes energy consumption, given the characteristic U-shaped relationship between gait speed and energy consumption. While we do not know the specific gait alterations adopted by the participants in this study, the link between walking energetics and physical activity suggests that energy consumption should be considered when advising or investigating therapeutic gait modifications.

While fatigue and fatigability have been previously linked to physical activity, the novel finding of this study is the association between fatigability and walking energetics in this population. While this association has not previously been documented, others have speculated that this connection exists. For example, abnormal motion of the center of mass has been observed during gait in people with OA and linked to mechanical energy exchange, which is another way to characterize walking energetics (Queen et al., 2016; Sparling et al., 2014). In these studies, it was speculated that fatigue, which has been well document in people with OA, could be linked to these energy and gait-related abnormalities. Our findings support the notion that walking energetics are linked to the construct of fatigue, particularly fatigability – movement associated fatigue.

The mediation analysis further suggests that that walking energetics, as measured via relative energy used during walking, may exert its effect on physical activity through fatigability as indicated by a statistically significant indirect effect of fatigability. An interpretation of this finding is that people who use more energy for walking may be less active in part because they tire more easily with activity. Interestingly, self-reported fatigue, while independently associated with physical activity, did not mediate the association between walking energetics and physical activity. This suggests that fatigue and fatigability exert their effects through different pathways – behavioral and physical. This hypothesis is in agreeance with findings by Murphy and Smith (Murphy and Smith, 2010) that fatigue and fatigability are different constructs in OA. Physical interventions that see to address physical activity should consider targeting fatigability. Based on our recent study in people with hip OA (Foucher et al., 2020), and work in other populations (Keyser et al., 2015), targeting aerobic fitness may be one viable strategy.

This study has implications for the development of interventions to improve symptoms and physical activity in people with OA. For example, increased trunk lean is sometimes adopted by people with knee OA and sometimes recommended to reduce loading on the medial compartment of the knee (Hunt et al., 2008; Mundermann et al., 2008; Simic et al., 2012). This recommendation may raise the energy cost of walking (Caldwell et al., 2013; Marques et al., 2013; Takacs et al., 2014). Linking this finding to the energetic/pain energy models suggests that such a protective gait modification could have unintended effects on physical activity. Other interventions such as medial thrust gait could similarly negatively affect walking energetics, but this has not yet been specifically addressed. Positive effects on loading and symptoms could be offset by negative impacts on physical activity. In contrast, in one study of older adults with mobility impairments, a motor sequence learning program focused on “walking skill” was effective at improving gait efficiency and physical activity. A change in −0.04 ml/kg*m was statistically significant and proved to be clinically meaningful. Such a gait retraining approach may be fruitful in people with OA. We note that it is conceivable that any gait retraining approach could result in a transient increase in pain, however the literature shows that the long-term effects of physical activity in knee OA include reduced pain and improved function (Fransen et al., 2015; Skou and Roos, 2019).

There are several limitations to this study that should be considered. First, there are many ways to assess fatigability. While there is general agreement on the definition of the construct, there is no standard fatigability test (Van Geel et al., 2020). Other studies have used different assessments, but overall, our results with respect to the association between fatigability and physical activity are in line with others. A potential drawback of the test we selected, however, was its length. Hence, the second limitation was potential selection bias. Participants needed to be able to walk on a treadmill for 10 minutes without assistive devices. This could have biased our sample toward higher functioning individuals. However, we do not believe that such bias was a factor as there was a range of functional levels represented as indicated by KOOS ADL scores ranging from 14 to 100. Next, we did not explore the walking impairments that could have led to increased energy consumption during gait. Future work in that area is needed to do so that less energetically costly gait interventions can be developed. Next, we cannot comment on the radiographic severity since we did not obtain radiographs to reduce participant burden. However, we are not aware of any published associations between structural disease severity and physical activity. Further, pain and symptoms, which are often found to be associated with physical activity, are typically poorly corelated with radiographic disease severity. Next, while the six-minute walk test is widely used as a measure of function in people with OA, its use for predicting VO2max has not been specifically validated in this population. Next, the UCLA scale is not without limitations. In the 1998 validation study, the authors noted that moderate-highly active persons may have underestimated their habitual activity using the UCLA scale (Zahiri et al., 1998). If that were the case in this study, it would not change our findings, however it may have changed the shapes or linearity of the curves. Finally, direction of correlation and causality cannot be inferred from this cross-sectional analysis. It is possible that people who are less active become less physically fit and therefore use more energy for walking, rather than the opposite association hypothesized. This relationship is likely bidirectional.

5. Conclusions

In conclusion, we found that walking energetics were significantly associated with physical activity in people with knee OA and that the effect on physical activity was mediated by fatigability. These findings highlight the importance of considering walking energetics and fatigability, in addition to pain, when developing biomechanically based interventions for knee OA since gait alterations may be energetically costly. These findings also suggest that targeting walking energetics could be a viable strategy to improve physical activity in people with knee OA.

Acknowledgements:

This work was funded by a University of Illinois at Chicago Center for Clinical and Translational Science pilot grant. The funder had on involvement in study design, collection, analysis or interpretation of data, writing of the report, or the decision to submit the article for publication.

Footnotes

Declarations of interest: none

References

  1. Broderick JE, DeWitt EM, Rothrock N, Crane PK, Forrest CB, 2013. Advances in Patient-Reported Outcomes: The NIH PROMIS((R)) Measures. EGEMS (Washington, DC) 1, 1015-9214.1015. eCollection 2013. 10.13063/2327-9214.1015 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Caldwell LK, Laubach LL, Barrios JA, 2013. Effect of specific gait modifications on medial knee loading, metabolic cost and perception of task difficulty. Clinical biomechanics (Bristol, Avon) 28, 649–654. 10.1016/j.clinbiomech.2013.05.012; 10.1016/j.clinbiomech.2013.05.012 [DOI] [PubMed] [Google Scholar]
  3. Cella D, Lai JS, Jensen SE, Christodoulou C, Junghaenel DU, Reeve BB, Stone AA, 2016. PROMIS Fatigue Item Bank had Clinical Validity across Diverse Chronic Conditions. Journal of clinical epidemiology 73, 128–134. 10.1016/j.jclinepi.2015.08.037 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Control, C. for D., Prevention, (CDC), 2011. Arthritis as a potential barrier to physical activity among adults with obesity–United States, 2007 and 2009. MMWR.Morbidity and mortality weekly report 60, 614–618. https://doi.org/mm6019a4 [pii] [PubMed] [Google Scholar]
  5. Coyle PC, Schrack JA, Hicks GE, 2018. Pain Energy Model of Mobility Limitation in the Older Adult. Pain Med 19, 1559–1569. 10.1093/pm/pnx089 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Egerton T, Chastin SF, Stensvold D, Helbostad JL, 2016a. Fatigue May Contribute to Reduced Physical Activity Among Older People: An Observational Study. The journals of gerontology.Series A, Biological sciences and medical sciences 71, 670–676. 10.1093/gerona/glv150 [doi] [DOI] [PubMed] [Google Scholar]
  7. Egerton T, Helbostad JL, Stensvold D, Chastin SF, 2016b. Fatigue Alters the Pattern of Physical Activity Behavior in Older Adults: Observational Analysis of Data from the Generation 100 Study. Journal of Aging and Physical Activity 24, 633–641. https://doi.org/2015-0237 [pii] [DOI] [PubMed] [Google Scholar]
  8. Eldadah BA, 2010. Fatigue and fatigability in older adults. PM & R : the journal of injury, function, and rehabilitation 2, 406–413. 10.1016/j.pmrj.2010.03.022 [doi] [DOI] [PubMed] [Google Scholar]
  9. Faul F, Erdfelder E, Buchner A, Lang AG, 2009. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behavior research methods 41, 1149–1160. 10.3758/BRM.41.4.1149 [DOI] [PubMed] [Google Scholar]
  10. Faul F, Erdfelder E, Lang AG, Buchner A, 2007. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior research methods 39, 175–191. [DOI] [PubMed] [Google Scholar]
  11. Fleg Jerome L, Morrell Christopher H, Bos Angelo G, Brant Larry J, Talbot Laura A, Wright Jeanette G, Lakatta Edward G, 2005. Accelerated Longitudinal Decline of Aerobic Capacity in Healthy Older Adults. Circulation 112, 674–682. 10.1161/CIRCULATIONAHA.105.545459 [DOI] [PubMed] [Google Scholar]
  12. Foucher KC, Aydemir B, Huang C-H, Horras M, Chmell SJ, 2020. Aerobic capacity and fatigability are associated with activity levels in women with hip osteoarthritis. J Orthop Res. 10.1002/jor.24856 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fransen M, McConnell S, Harmer AR, Esch M.V. der, Simic M, Bennell KL, 2015. Exercise for osteoarthritis of the knee: a Cochrane systematic review. British journal of sports medicine 49, 1554–1557. 10.1136/bjsports-2015-095424 [doi] [DOI] [PubMed] [Google Scholar]
  14. Gussoni M, Margonato V, Ventura R, Veicsteinas A, 1990. Energy Cost of Walking with Hip Joint Impairment. Physical Therapy 70, 295–301. 10.1093/ptj/70.5.295 [DOI] [PubMed] [Google Scholar]
  15. Hayes AF, 2017. Introduction to Mediation, Moderation, and Conditional Process Analysis, 2nd ed. Guilford Publications. [Google Scholar]
  16. Hunt MA, Birmingham TB, Bryant D, Jones I, Giffin JR, Jenkyn TR, Vandervoort AA, 2008. Lateral trunk lean explains variation in dynamic knee joint load in patients with medial compartment knee osteoarthritis. Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society 16, 591–599. 10.1016/j.joca.2007.10.017 [DOI] [PubMed] [Google Scholar]
  17. Keyser RE, Christensen EJ, Chin LMK, Woolstenhulme JG, Drinkard B, Quinn A, Connors G, Weir NA, Nathan SD, Chan LE, 2015. Changes in fatigability following intense aerobic exercise training in patients with interstitial lung disease. Respiratory Medicine 109. 10.1016/j.rmed.2015.01.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kim I, Hacker E, Ferrans CE, Horswill C, Park C, Kapella M, 2018. Evaluation of fatigability measurement: Integrative review. Geriatric nursing (New York, N.Y.) 39, 39–47. https://doi.org/S0197-4572(17)30130-1 [pii] [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Laskin JJ, Bundy S, Marron H, Moore H, Swanson M, Blair M, Humphrey R, 2007. Using a treadmill for the 6-minute walk test: reliability and validity. Journal of cardiopulmonary rehabilitation and prevention 27, 407–410. 10.1097/01.HCR.0000300270.45881.d0 [doi] [DOI] [PubMed] [Google Scholar]
  20. Marques NR, LaRoche DP, Hallal CZ, Crozara LF, Morcelli MH, Karuka AH, Navega MT, Goncalves M, 2013. Association between energy cost of walking, muscle activation, and biomechanical parameters in older female fallers and non-fallers. Clinical biomechanics (Bristol, Avon) 28, 330–336. 10.1016/j.clinbiomech.2013.01.004; 10.1016/j.clinbiomech.2013.01.004 [DOI] [PubMed] [Google Scholar]
  21. Mundermann A, Asay JL, Mundermann L, Andriacchi TP, 2008. Implications of increased medio-lateral trunk sway for ambulatory mechanics. Journal of Biomechanics 41, 165–170. 10.1016/j.jbiomech.2007.07.001 [DOI] [PubMed] [Google Scholar]
  22. Murphy LB, Hootman JM, Boring MA, Carlson SA, Qin J, Barbour KE, Brady TJ, Helmick CG, 2017. Leisure Time Physical Activity Among U.S. Adults With Arthritis, 2008–2015. American Journal of Preventive Medicine 53, 345–354. https://doi.org/S0749-3797(17)30207-6 [pii] [DOI] [PubMed] [Google Scholar]
  23. Murphy SL, Niemiec SS, Lyden AK, Kratz AL, 2016. Pain, Fatigue, and Physical Activity in Osteoarthritis: The Moderating Effects of Pain- and Fatigue-Related Activity Interference. Archives of Physical Medicine and Rehabilitation 97, 201. 10.1016/j.apmr.2015.05.025 [doi] [DOI] [PubMed] [Google Scholar]
  24. Murphy SL, Smith DM, 2010. Ecological Measurement of Fatigue and Fatigability in Older Adults With Osteoarthritis. J Gerontol A Biol Sci Med Sci 65A, 184–189. 10.1093/gerona/glp137 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Power JD, Badley EM, French MR, Wall AJ, Hawker GA, 2008. Fatigue in osteoarthritis: a qualitative study. BMC musculoskeletal disorders 9, 63–63. 10.1186/1471-2474-9-63 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Queen RM, Sparling TL, Schmitt D, 2016. Hip, Knee, and Ankle Osteoarthritis Negatively Affects Mechanical Energy Exchange. Clinical orthopaedics and related research 474, 2055–2063. 10.1007/s11999-016-4921-1 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Roos EM, Lohmander LS, 2003. The Knee injury and Osteoarthritis Outcome Score (KOOS): from joint injury to osteoarthritis. Health and quality of life outcomes 1, 64–64. 10.1186/1477-7525-1-64 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Schnelle JF, Buchowski MS, Ikizler TA, Durkin DW, Beuscher L, Simmons SF, 2012. Evaluation of Two Fatigability Severity Measures in Elderly Adults. Journal of the American Geriatrics Society 60, 1527–1533. 10.1111/j.1532-5415.2012.04062.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Schrack JA, Simonsick EM, Ferrucci L, 2010. The energetic pathway to mobility loss: an emerging new framework for longitudinal studies on aging. Journal of the American Geriatrics Society 58 Suppl 2, 329. 10.1111/j.1532-5415.2010.02913.x; 10.1111/j.1532-5415.2010.02913.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Schrager MA, Schrack JA, Simonsick EM, Ferrucci L, 2014. Association between energy availability and physical activity in older adults. American Journal of Physical Medicine & Rehabilitation / Association of Academic Physiatrists 93, 876–883. 10.1097/PHM.0000000000000108 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Simic M, Hunt MA, Bennell KL, Hinman RS, Wrigley TV, 2012. Trunk lean gait modification and knee joint load in people with medial knee osteoarthritis: the effect of varying trunk lean angles. Arthritis care & research 64, 1545–1553. 10.1002/acr.21724 [doi] [DOI] [PubMed] [Google Scholar]
  32. Skou ST, Roos EM, 2019. Physical therapy for patients with knee and hip osteoarthritis: supervised, active treatment is current best practice. Clinical and experimental rheumatology 37 Suppl 120, 112–117. https://doi.org/14740 [pii] [PubMed] [Google Scholar]
  33. Snijders GF, van den Ende CH, Fransen J, Riel P.L. van, Stukstette MJ, Defoort KC, Arts-Sanders MA, van den Hoogen FH, Broeder A.A. den, Group, N.O.C.S., 2011. Fatigue in knee and hip osteoarthritis: the role of pain and physical function. Rheumatology (Oxford, England) 50, 1894–1900. 10.1093/rheumatology/ker201 [doi] [DOI] [PubMed] [Google Scholar]
  34. Sparling TL, Schmitt D, Miller CE, Guilak F, Somers TJ, Keefe FJ, Queen RM, 2014. Energy recovery in individuals with knee osteoarthritis. Osteoarthritis and Cartilage 22, 747–755. 10.1016/j.joca.2014.04.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Takacs J, Kirkham AA, Perry F, Brown J, Marriott E, Monkman D, Havey J, Hung S, Campbell KL, Hunt MA, 2014. Lateral trunk lean gait modification increases the energy cost of treadmill walking in those with knee osteoarthritis. Osteoarthritis and Cartilage 22, 203–209. [DOI] [PubMed] [Google Scholar]
  36. Terwee CB, Bouwmeester W, Elsland S.L. van, Vet H.C. de, Dekker J, 2011. Instruments to assess physical activity in patients with osteoarthritis of the hip or knee: a systematic review of measurement properties. Osteoarthritis and cartilage 19, 620–633. 10.1016/j.joca.2011.01.002 [doi] [DOI] [PubMed] [Google Scholar]
  37. Van Geel F, Moumdjian L, Lamers I, Bielen H, Feys P, 2020. Measuring walking-related performance fatigability in clinical practice: a systematic review. Eur J Phys Rehabil Med 56, 88–103. 10.23736/S1973-9087.19.05878-7 [DOI] [PubMed] [Google Scholar]
  38. VanSwearingen JM, Perera S, Brach JS, Wert D, Studenski SA, 2011. Impact of exercise to improve gait efficiency on activity and participation in older adults with mobility limitations: a randomized controlled trial. Physical Therapy 91, 1740–1751. 10.2522/ptj.20100391; 10.2522/ptj.20100391 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wert DM, Brach JS, Perera S, VanSwearingen J, 2013. The association between energy cost of walking and physical function in older adults. Archives of Gerontology and Geriatrics 57, 198–203. 10.1016/j.archger.2013.04.007; 10.1016/j.archger.2013.04.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zahiri CA, Schmalzried TP, Szuszczewicz ES, Amstutz HC, 1998. Assessing activity in joint replacement patients. The Journal of arthroplasty 13, 890–895. [DOI] [PubMed] [Google Scholar]

RESOURCES