Gait and heart rate: do they measure trait or state physical fatigue in people with multiple sclerosis?

• Published in: Journal of neurology 2024-07, Vol.271 (7), p.4462-4472
• Main Authors: Galperin, Irina, Buzaglo, David, Gazit, Eran, Shimoni, Nathaniel, Tamir, Raz, Regev, Keren, Karni, Arnon, Hausdorff, Jeffrey M.
• Format: Article
• Language: English
• Subjects: Adult; Autonomic nervous system; Autonomic Nervous System - physiopathology; Fatigue; Fatigue - etiology; Fatigue - physiopathology; Female; Gait; Gait - physiology; Heart rate; Heart Rate - physiology; Humans; Male; Medicine; Medicine & Public Health; Middle Aged; Multiple sclerosis; Multiple Sclerosis - complications; Multiple Sclerosis - physiopathology; Neurology; Neuroradiology; Neurosciences; Original Communication; Regression analysis; Sensorimotor integration; Walk Test
• ISSN: 0340-5354; 1432-1459; 1432-1459
• DOI: 10.1007/s00415-024-12339-8

Background Trait and state physical fatigue (trait-PF and state-PF) negatively impact many people with multiple sclerosis (pwMS) but are challenging symptoms to measure. In this observational study, we explored the role of specific gait and autonomic nervous system (ANS) measures (i.e., heart rate, HR, r–r interval, R–R, HR variability, HRV) in trait-PF and state-PF. Methods Forty-eight pwMS [42 ± 1.9 years, 65% female, EDSS 2 (IQR: 0–5.5)] completed the Timed Up and Go test (simple and with dual task, TUG-DT) and the 6-min walk test (6MWT). ANS measures were measured via a POLAR H10 strap. Gait was measured using inertial-measurement units (OPALs, APDM Inc). Trait-PF was evaluated via the Modified Fatigue Impact Scale (MFIS) motor component. State-PF was evaluated via a Visual Analog Scale (VAS) scale before and after the completion of the 6MWT. Multiple linear regression models identified trait-PF and state-PF predictors. Results Both HR and gait metrics were associated with trait-PF and state-PF. HRV at rest was associated only with state-PF. In models based on the first 3 min of the 6MWT, double support (%) and cadence explained 47% of the trait-PF variance; % change in R–R explained 43% of the state-PF variance. Models based on resting R–R and TUG-DT explained 39% of the state-PF. Discussion These findings demonstrate that specific gait measures better capture trait-PF, while ANS metrics better capture state-PF. To capture both physical fatigue aspects, the first 3 min of the 6MWT are sufficient. Alternatively, TUG-DT and ANS rest metrics can be used for state-PF prediction in pwMS when the 6MWT is not feasible.