Factors influencing, and associated with, physical activity patterns in dogs with osteoarthritis-associated pain.

Background: Accelerometry can be used to measure physical activity and is a validated objective measure for evaluating the impact of osteoarthritis (OA) pain in companion animals. However, several factors other than OA pain can affect physical activity in dogs, and relatively little is understood about their influence. Functional linear modeling (FLM) is an approach for analyzing and visualizing high-frequency longitudinal data such as physical activity and can be used to assess the influence of factors on activity patterns. This study aimed to use FLM to investigate the effect of various factors on physical activity patterns in a cohort of dogs with OA pain. Methods: Ninety-nine client-owned dogs with radiographic and clinical evidence of OA were fitted with a collar-based activity monitor (Actigraph GT3X). Average vector magnitudes were recorded once per minute over 7 days and averaged to create 24-h, per-minute activity profiles for each dog. Demographic information, owner completed OA Clinical Metrology Instruments (Liverpool Osteoarthritis in Dogs and Canine Brief Pain Inventory), and veterinary examination findings (joint pain, muscle atrophy) were collected. Data were analyzed using FLM and a custom R package to evaluate the effect of each factor on 24-h patterns of physical activity. Results: At times of peak activity within a 24-h period, dogs with hindlimb OA pain, higher age, higher Clinical Metrology Instrument scores, higher joint pain, greater Body Condition Score and greater muscle atrophy all had decreased activity profiles. However, only age, hindlimb joint pain, and hindlimb muscle atrophy had statistically significant effects on physical activity. Conclusions and clinical relevance: Several factors influence activity patterns in dogs with OA pain. Understanding what and how factors influence patterns in dogs with OA pain will help refine the usage of physical activity as an objective outcome measure in clinical pain studies.