Biomechanical differences and variability during sustained motorized treadmill running versus outdoor overground running using wearable sensors.

Abstract

This study aimed to compare running biomechanics and biomechanical variability across 3 run segments and between conditions for 5-km outdoor overground and indoor treadmill running. Seventy-one recreationally-active adults (31F, 40 M; age: 37 ± 11 years; body mass index: 22.9 ± 2.5 kg/m²) completed aerobic fitness assessments at baseline (VO₂max), outdoor overground 5 km runs on a standardized route, and indoor treadmill 5 km runs on a motorized system (12.6 ± 4.9 days apart). Wearable sensors recorded step-by-step spatiotemporal, kinetic, and kinematic biomechanics. Repeated measures analyses of covariance were used to compare mean and coefficient of variation (CV) of sensor-derived metrics across run segments, conditions, and limbs (covariates: pace, VO₂max). Tukey's post-hoc tests with mean differences and Cohen's d effect sizes were used to determine the difference magnitudes across comparisons. Most biomechanical measures significantly differed between running conditions (p < 0.001); contact time (mean difference and standard error: 8 ± 3 ms; d = 0.20), stride length (0.20 ± 0.12 m; d: 0.31), kinetics (shock, impact, braking; 0.17-1.30 g; d-range: 0.36-0.57), and pronation velocity (138 ± 16°/s; d: 0.61) were all higher during indoor treadmill running. Indoor treadmill running biomechanics CV were significantly higher for most measures compared to outdoor overground running (p < 0.001; d-range: 0.18-0.52). Only spatiotemporal measures and CV significantly differed across run segments (d-range: 0.16-0.68). Clinicians should expect that indoor treadmill biomechanics, particularly kinetic and pronation, will be significantly higher than patients' outdoor overground running biomechanics and tailor subsequent recommendations accordingly. Furthermore, clinicians should expect that indoor treadmill running analyses may result in more variable biomechanics, potentially attributed to consistent speed and surface, and tailor assessments to preferred run environments.