The Relation of Activity Patterns with Incident Slow Gait Speed over 4 Years in Adults with Knee Osteoarthritis: Data from the Osteoarthritis Initiative

Published: 17 March 2025| Version 1 | DOI: 10.17632/h4hhcf56t3.1
Contributors:
, Sydney Liles , David Werner, Rebecca J. Cleveland , Tom Videitch Bye , Yvonne Golightly , Jason T. Jakiela, Daniel White

Description

The purpose of this study was to explore patterns of baseline physical activity using Multilevel Functional Principal Components Analysis (MFPCA) and examine the relation of these patterns with incident slow gait speed over 4 years in adults with or at high risk of knee OA utilizing data from the accelerometer sub-study of the Osteoarthritis Initiative (OAI). For the purpose of the study, Multilevel Functional Principal Components Analysis was done. As a result of the analysis, it was determined that there were 4 activity patterns (PCs), 1) high activity, 2) high evening activity, 3) high morning and evening activity, and 4) very high morning activity. Those whose daily activity patterns best matched with the pattern PC2 or PC4, were found to have 0.60 and 0.39 times the risk of developing slow gait speed, respectively; while those whose daily activity patterns least matched PC3, were found to have 1.77 times the risk. Our study results suggest that daily activity patterns may be associated with the development of slow gait speed in adults with or at risk for knee OA.

Files

Steps to reproduce

Data from the Osteoarthritis Initiative (OAI) was utilized, focusing on a longitudinal cohort of adults between the ages of 45 and 79 who either have knee osteoarthritis or are at risk for developing it. The OAI collected this data across four clinical sites located in Ohio, Pennsylvania, Maryland, and Rhode Island. The protocol and eligibility criteria can be found at ClinicalTrials.gov (NCT00080171). Participants in this study were drawn from the accelerometer substudy associated with the Osteoarthritis Initiative (OAI). The follow-up visits for these individuals occurred between August 2008 and July 2010, marking a four-year interval. Participants were directed to wear the Actigraph GT1M accelerometer throughout their waking hours for a duration of 7 days, excluding any water-related activities. Inclusion criteria required that patients had accelerometer data from a uniaxial device, the Actigraph GT1M, with at least four days of wear time exceeding ten hours per day. Slow gait speed was measured through the 20-meter walk test, which has proven to be a reliable method for assessing gait speed in individuals with knee osteoarthritis, at both the baseline of the study and after a four-year period. Participants with a prevalent slow gait speed at baseline were excluded from the study. The study utilized Multidimensional Multilevel Functional Principal Components Analysis (MFPCA), an advanced analytical technique. This method applies principal components (PC) of daily activity patterns to capture the differences in a participant's activity pattern compared to the mean of the sample. This analysis emphasized the variation in patterns at the between-subject level. Individual PC scores were segmented into tertiles, each comprising approximately one-third of the sample, and assigned participants to the appropriate tertile for each PC. Individuals in the highest tertile exhibit daily physical activity patterns that align most closely with the principal component (PC), whereas those in the lowest tertile display patterns that are contrary or opposite to this alignment. The middle tertile for each PC serves as the reference group, as it comprises participants whose activity levels throughout the day are near the sample mean, independent of the PC. In order to analyze the relationship between the four classified principal components and the development of slow gait speed during a four-year risk period, risk ratios (RR) were computed using regression methods that employed a log-link function along with robust standard errors.

Institutions

Winston-Salem State University, University of Delaware, Osmaniye Korkut Ata Universitesi, University of Nebraska Medical Center, University of North Carolina at Chapel Hill

Categories

Physical Activity

Funding

UNC Core Center for Clinical Research

Arthritis Foundation

Rheumatology Research Foundation

Licence