Who Falls After a Stroke? Evidence From a Prospective Stroke Cohort
Anna Kufner, Yunyou Tang, Uchralt Temuulen, Ghadir Abbas, Torsten Rackoll, Ulrike Grittner, Daniel Kroneberg, Benedikt Weigl, Andrea A. Kühn, Martin Reich, Alexander H. Nave, Matthias EndresABSTRACT
Background
Falls affect over 30% of stroke survivors within the first year, yet lesion‐related mobility and gait impairments underlying fall risk remain poorly understood. This study aimed to identify lesion‐derived functional networks associated with impaired mobility, gait, and fall risk in subacute stroke, and to determine whether disruption to these networks is associated with falls during the six‐month follow‐up.
Methods
We analyzed data from 94 patients with disabling subacute ischemic stroke from the prospective Baptize cohort, an imaging sub‐cohort of the multicenter PHYS‐STROKE trial. Principal component (PC) analysis reduced seven mobility‐related and four gait‐related baseline variables into two composites: PC1‐Mobility and PC1‐Gait, explaining 56% and 82% of variance, respectively. PC1‐Mobility indexed global disability, whereas PC1‐Gait reflected spatiotemporal walking capacity. Lesion network mapping (LNM) identified functional networks associated with each domain. Patient‐reported falls up to six months post‐enrollment were the primary endpoint.
Results
LNM revealed that the mobility‐related network predominantly involved cortical regions, whereas the gait‐related network was linked to subcortical and infratentorial connectivity. In binary multivariable logistic regression, network similarity scores were not associated with falls; only older age was significant (adjusted OR 1.08, 95% CI 1.02–1.16, p = 0.01). LNM of fall occurrence identified a cortical network with significant spatial overlap with the mobility‐related network ( p < 0.001).
Conclusion
This exploratory, hypothesis‐generating study identified distinct functional networks for post‐stroke mobility and gait impairment. Falls may be more closely linked to disruptions in cortical networks supporting voluntary motor control and whole‐body coordination than to subcortical gait‐modulating structures, potentially informing fall risk stratification and targeted prevention.
Trial Registration