Unsupervised Multivariate Analysis of Gait Characteristics and Body Composition in Physiotherapy Students

Abstract

Aim: Measurements of locomotion performance (through quantitative gait analysis) and of body composition offer detailed information on locomotor and biomechanical functions, respectively. The correlation between gait features and body composition among physiotherapy students has yet been extensively investigated using multivariate analyses. We thus evaluated the relationships between body composition components and the parameters of the walk in physiotherapy students, through an unsupervised exploratory approach.Methods: This cross-sectional study included 48 physiotherapy students (35 women and 13 men, aged mean (SD): 21.3 (1.8) years). Measurements of body composition were done using bioelectrical impedance and consisted in body mass index (BMI), body fat percentage, muscle percentage, total body water percentage, and resting energy expenditure. Measurements of gait were performed with the OptoGait system, comprising cadence, contact time, and propulsive phase. Data was analysed using Pearson’s correlation and with both hierarchical clustering and principal component analysis (PCA) in order to find out multimodal relationships between gait and body composition variables.Results: The results of Hierarchical cluster analyses yielded two broad functional domains built on gait and body composition variables. A first domain encompassing the number of steps, proportion of propulsion, and % of muscle, and total % of water was conceptualized as a performance-based group. A second domain with % body fat, BMI and contact time was thought to represent a mass-based group linked with less desirable gait features. % total body water was seen to be a bridging element between locomotory performance and body mass variables.A positive association was seen between % muscle mass and the number of steps (r=0.62, p<0.001) while % body fat was negatively associated with this latter variable (r= -0.54, p<0.001). In principal components analyses, the first two components accounted for 73.6 % of the total variance.Conclusion: Body composition metrics were strongly related to gait patterns in apparently healthy physiotherapy students, with a greater proportion of muscle mass and total body water related to a higher frequency and a more propulsive phase of gait, whilst increased body fat proportion and BMI were correlated with a prolonged contact time and a lower frequency of gait. Further exploratory multivariate analyses would seem appropriate to expand on the biomechanical relationships of young adults.