ASSESSING HOW WALKING PATTERNS CHANGE WITH KNEE OSTEOARTHRITIS PAIN IN AND OUT OF THE LABORATORY SETTING

Abstract

Introduction Knee osteoarthritis (OA) is a highly prevalent degenerative joint disease and a leading cause of musculoskeletal pain. The progression of symptomatic knee OA is associated with the knee joint loads experienced during gait. With joint loads being associated with pain, and pain being a common symptom of knee OA, understanding how gait changes with momentary fluctuations in pain may provide insight into knee OA disease progression. Few studies have assessed momentary fluctuations in pain and gait with activity and found high inter-subject variability in how gait changes with pain. Since pain is subjective, pain-related psychological factors may explain some of the high inter-subject variability and in turn why some individuals may have faster disease progression. Therefore, this project aimed to identify how pain and gait were related during activity 1) in the lab and 2) in the real-world, as well as 3) how pain-related psychological factors moderate these relationships. Methods Nineteen older healthy adults (9 female) and 18 adults with self-reported symptomatic physician-diagnosed knee OA (12 female) participated in this study. The study consisted of a campus visit for all participants, and participants with knee OA also completed a three-day real-world daily gait analysis. We assessed kinesiophobia (TSK-17), pain-catastrophizing (PCS-13), and pain self-efficacy (PSEQ) to evaluate pain-related psychological factors. For the first aim, participants were fitted with inertial measurement units (IMUs) and completed a pain inducing protocol. This protocol consisted of a 20-m walk before and after completing two flights of stairs, resting 5 minutes, then repeating the walks and stairs. Participants reported their peak pain during each of the four walks from 0-10 (0 = no pain). The average of each outcome variable (gait speed, stride length, and hip, knee, and ankle range of motion (ROM)) was calculated for each walk. Statistical analyses were run on the differences in gait between the post-stair walks and the pre-stair walks (i.e., walk 2 – walk 1 & walk 4 – walk 3). We used a mixed model 2x2 (group x time) ANOVA to determine whether changes in pain and gait in response to bouts of stairs differed between adults with and without knee OA. For the second aim, participants with knee OA wore IMUs in daily life for 3 full days. Throughout each day, participants received 5 electronic messages. Participants were instructed to complete a ~3-5 minute walk when they received each text, and to fill out the survey after the walk to report their peak pain during the instructed ~3-5 minute walk. We used a random coefficients model with 60 pairs per participant (4 outcome variables (gait speed, stride length, and knee and ankle ROM) x 5 time points x 3 days) using pain as the predictor variable to assess the relationship between pain and gait in real-world settings. For the third aim, we used the PCS-13, TSK-17 and PSEQ scores as continuous variables to test whether pain-related psychological factors moderate the relationship between gait and pain. We only included those diagnosed with knee OA and used the same pain values and gait data from in-lab (aim 1) and real-world data (aim 2). Results Aim 1 – For the in-lab protocol, there were no significant group by bout interactions for any variable (all p>0.47). There was a larger increase in knee ROM in response to the first bout of stairs compared to the second bout of stairs (bout1Δ = 0.6 ± 1.3°, bout2Δ = 0.0 ± 1.2°, p = 0.008). Those in the knee OA group who experienced a change in pain following the first stair bout also had a greater change in knee ROM (pain |bout1Δ| = 1.87 ± 1.7°, no pain |bout1Δ| = 0.86 ± 0.6°, post-hoc p = 0.02). Aim 2 – During real-world gait, greater pain was associated with less knee ROM (p = 0.03). The marginal model estimate identified a 2.7° decrease in knee ROM per unit increase in pain. The effect of pain across participants ranged from a decrease of 6.2° to an increase of 3.5° in knee ROM per one point increase in pain. Speed, stride length, and ankle ROM did not differ by pain level. Aim 3 - Pain-related psychological factors did not moderate the relationship between pain and gait mechanics during the campus visit, but kinesiophobia did moderate the relationship between pain and gait mechanics during real-world typical activity. At scores ≤35, kinesiophobia strengthened the relationship between pain and knee range of motion, with larger pain values being associated with smaller knee range of motion. At the ankle, TSK-17 scores below 29 were associated with a negative association between ankle range of motion and pain while scores above 36 were associated with a positive association between ankle range of motion and pain Conclusion We consistently found that knee ROM varied with movement-evoked pain regardless of setting (in-lab vs real-world). We also found that the negative relationship between both knee and ankle ROM and pain was more pronounced at lower levels of kinesiophobia, but only in the real-world. These findings may suggest that those with less kinesiophobia are able to vary their gait as their pain changes, while individuals with more kinesiophobia are unable to modify their gait no matter their pain level. This project highlights the value in measuring the relationship between movement-evoked pain and gait mechanics in less controlled settings and that Inertial Measurement Units can be used to measure these small differences in gait mechanics regardless of setting. While our findings are from a small sample, they also support a need for more investigation into the role of pain-related psychological factors in pain and gait knee OA research.