A novel biomechanical indicator for impaired ankle dorsiflexion function during walking in individuals with chronic stroke

Background: Ankle dorsiflexion function during swing phase contributes to foot clearance and plays an important role in walking ability post-stroke. Commonly used biomechanical measures such as foot clearance and ankle joint excursion have limited ability to accurately evaluate impaired dorsiflexor function.Research question: Can ankle angular velocity and acceleration be used as reliable measurers of dorsiflexion function in post-stroke gait?Methods: Using linear regression and Pearson's correlation we retrospectively compared peak ankle angular velocity (A omega(P)), peak ankle angular acceleration (A alpha(P)), peak dorsiflexion angle (DFA(P)) and peak foot clearance (FCLP) as direct measures for swing phase dorsiflexor function in 60 chronic stroke survivors. Intraclass correlation coefficient (ICC) analysis was used for test-retest reliability of A omega(P) and A alpha(P.)Results: Linear regression models revealed that A omega(P), A alpha(P), DFA(P), FCLP had a significant relationship (p < 0.05) with impaired dorsiflexion function. A alpha(P) and DFA(P) accounted for the most variance of dorsiflexion function. A omega(P), A alpha(P), FCLP, correlated significantly with all clinical outcome measures of walking ability. DFA(P) had a positive correlation only with FMA-LE. Post-hoc William's t-tests, used to compare the magnitude of difference between two non-independent correlations, revealed that the correlation between all clinical measures and DFA(P) were significantly weaker than with A omega(P) and A alpha(P). Correlation between FMA-LE and FCLP was weaker than with A omega(P) and A alpha(P). Excellent test-retest reliability for both A omega(P) (ICC = 0.968) and A alpha(P) (ICC = 0.947) was observed.Significance: These results suggest that DFA(P) may only be associated with dorsiflexion function during non-task specific isolated movements, but not during walking. FCLP is associated with dorsiflexion function and walking ability measures but not as strongly as A omega(P) and A alpha(P) possibly because FCLP is influenced by contribution from hip and knee joint movements. Therefore, A omega(P) and A alpha(P) are reliable measures and represent dorsiflexion function more accurately than DFA(P), and FCLP.