The association of knee bending and physical activity with incident pain and functional limitation over 4 years in knee osteoarthritis

Abstract

Purpose: Engaging in knee bending activities, e.g. climbing stairs, kneeling, or squatting, is a risk factor for the incidence and progression of knee OA and may be deleterious for knee pain. However, engaging in some knee bending activities is, in general, beneficial for lower extremity strength, which is related to future functional outcomes. Similar benefits to physical function are observed with physical activity (PA), e.g. walking. Knee bending activities and PA can co-occur, yet little is known about the joint association of knee bending and PA with pain and functional limitation. Therefore, the purpose of this study was to investigate the joint association of knee bending and objectively-measured PA with incident pain and functional limitation in adults with or at risk for knee OA. Methods: We used data from the 48-month and 96-month visits of the Osteoarthritis Initiative (OAI), a large cohort study of adults with or at risk for knee OA. Knee bending activities and objectively-measured PA (steps/day) exposures were collected at the 48-month visit (baseline). Knee bending was measured using four Yes/No questions used to determine study eligibility: climbing up a total of 10 or more flights of stairs, kneeling 30 minutes or more, squatting 30 minutes or more, and lifting or moving objects weighing 25 pounds or more by hand. Positive responses were summed and categorized as Low Bending (0-1 Yes responses) and High Bending (2-4 Yes responses). PA was measured with an Actigraph GT1M accelerometer and was separately dichotomized into Steps Active and Steps Inactive using 6000 steps/day, based on previous studies showing this is a meaningful threshold for preventing functional limitation. To investigate the joint association of knee bending and PA, we created a four-level exposure variable: (1) Steps Inactive & Low Bending, (2) Steps Active & Low Bending, (3) Steps Inactive & High Bending, (4) Steps Active & High Bending. Incident pain and functional limitation outcomes were assessed at the 96-month visit. VAS Pain (0-10) was dichotomized using patient acceptable symptom state levels, with a score of 3/10 as the cutoff for meaningful pain. Functional limitation was defined using slow gait speed from a 20-meter walk test (<1.22 m/s) which represents the inability to safely cross a timed street intersection. To determine the association of our four-level knee bending and PA exposure with incident pain and functional limitation, we used binary logistic regression to calculate risk ratios (RR) and 95% confidence intervals (95% CI), adjusting for potential confounders at baseline. Results: Of the adults with knee bending and valid accelerometry data who were free of meaningful pain (n = 1268) and functional limitation (n = 1307) at the 48-month visit, 1115 (age: 62.9 ± 8.4 years old, sex: 50.5% female, BMI: 28.0 ± 4.5 kg/m2) had all outcomes at the 96-month visit. Of those, 393 adults reported 0 or 1 knee bending activities and 722 adults reported 2 to 4 knee bending activities. 826 adults reported engaging in at least 1 knee bending activity on at least 2-3 days of the week. Steps Active & High Bending had 23% reduced risk for incident slow gait speed compared with Inactive & Low Bending, which met statistical significance (Adjusted RR [95% CI]: 0.77 [0.68 - 0.88]). Similarly sized protective effects that met statistical significance were found for the Steps Active & Low Bending and the Steps Inactive & High Bending groups with slow gait speed (Table 1B). There was no significant association between any of the exposure groups with incident pain above a patient acceptable symptom state (Table 1A). Conclusions: Adults who were active or engaged in 2 or more knee bending activities were at reduced risk for functional limitation 4 years later, with the greatest risk reduction seen in the Steps Active & High Bending group. We did not find knee bending activities and/or PA to be associated with incident knee pain above a patient acceptable symptom state. Despite the relationship between knee bending and progression of knee OA, these findings suggest that engaging in more knee bending activities, regardless of PA level, is not adversely related to pain or physical function.