22. European Stroke Conference 448 Small vessel stroke and white matter disease Predictors of natural trajectories in walking-speed over time in patients with age-related white matter change: Grouping patients in homogenous clusters of change using latent class growth analysis S.H. Kreisel1, C. Blahak2, H. Bäzner3, M.G. Hennerici4 Department of Neurology, Universitätsmedizin Mannheim, University of Heidelberg, Mann-heim, GERMANY1, Department of Neurology, Universitätsmedizin Mannheim, University of Heidelberg, Mannheim, GERMANY2, Department of Neurology, Klinikum Stuttgart, Bielefeld, GERMANY3, Department of Neurology, Universitätsmedizin Mannheim, University of Heidelberg, Mannheim, GERMANY4 Background: The natural course of age-related white matter change (ARWMC) is heterogenous given varying lesion volume and patterns, and diverse comorbidities. Being able to algorithmically group patients into more homogenous categories of recovery or deterioration - without prior data as-sumptions - could aid patient counseling and help in choosing interventions. Methods: The Leukoaraiosis And DIsability Study (LADIS) is a European multicenter study having ascertained 639 subjects with ARWMC with a follow-up period of three years. Numerous clinical parameters, including walking-speed were measured multiple times. While an association with the severity of ARWMC has been shown, the large variability of trajectories remains unexplained. La-tent class growth analysis (LCGA) is a statistical procedure that assumes that individuals are not necessarily drawn from a homogenous population, but are rather a heterogeneous mixture of dif-ferent subcohorts. It can be understood as a longitudinal cluster analysis. LCGA was applied to the LADIS dataset to extract groups of patients with similar temporal trajectories of change in walking speed over time – and characterizing these subcohorts in respect to predictors such as the degree of ARWMC, age and comorbidities. Results: The best solution gave three distinct trajectory clusters – most patients (64.89%) did not show change in walking-speed over time, a smaller group deteriorated (25.73%), the rest improved. Group membership is best related to walking-speed at baseline and age, next to a set of comorbidites including mood disorders, prior osteoarthritis and chronic pain. Severity of ARWMC did not differ-entiate the group staying stable from those deteriorating. Conclusion: Patients are grouped primarily according to their initial impairment in walking-speed, and other diverse factors. Though the individual’s ARWMC severity influences trajectory cluster membership, it is not its primary predictor. 524 © 2013 S. Karger AG, Basel Scientific Programme
Karger_ESC London_2013
To see the actual publication please follow the link above