Recommendations for the analysis of rodent gait data to evaluate osteoarthritis treatments

Abstract

Behavioral assays of animal pain and disability can increase the clinical relevance of a preclinical study. However, pain and symptoms are difficult to measure in preclinical models. Because animals often alter their movement patterns to reduce or avoid joint pain, gait analysis can be an important tool for quantifying OA-related symptoms in rodents. Technologies to measure rodent gait continue to advance and have been the focus of prior reviews. Regardless of the techniques used, the analysis of rodent gait data can be complex due to multiple confounding variables. The goal of this review is to discuss recent advances in the understanding of OA-related gait changes and provide recommendations on the analysis of gait data. Recent studies suggest OA-affected animals reduce vertical loading through their injured limb while walking, indicating dynamic ground reaction forces are important data to collect when possible. Moreover, gait data analysis depends on accurately measuring and accounting for the confounding effects of velocity and other covariates (such as animal size) when interpreting shifts in various gait parameters. Herein, we discuss different statistical techniques to account for covariates and interpret gait shifts. In particular, this review will discuss residualization and linear mixed effects models, including how both techniques can account for inter- and intra-animal variability and the effects of velocity. Furthermore, this review discusses future considerations for using rodent gait analysis, while highlighting the intricacies of gait analysis as a tool to measure joint function and behavioral outcomes.