P167. Dual tasking deficits in motor tasks predict falls in Parkinson’s disease

Abstract

Objective To evaluate the predictive value of dual task performance for future falls in patients with Parkinson’s disease (PD). Background Falls are a severe health and injury risk in PD patients. While several factors have been linked to future falls, evidence supporting their predictive value remains inconsistent, and not yet defined factors may also contribute to this risk. One promising factor is dual task costs (DTC), in particular during the challenging, simultaneous performance of two motor tasks including walking. Increasing evidence suggests that performance of two motor tasks may lead to a “bottleneck” situation of competing processes in central motor circuits, which may contribute to falls in PD. Methods DTC (reduction of walking/checking boxes speeds or walking/subtracting serial 7 s speeds compared to respective single task speeds; performed at maximum speed without any prioritization) were biannually assessed in PD patients under OFF conditions over 3.5 years in the “Modeling epidemiological data to study PD progression” (MODEP) study. Falls were reported retrospectively for a period of 6 months before each assessment. DTC of patients 6–12 months prior to the first reported fall incident ( n = 14; fallers) were compared to DTC of 14 patients who reported no falls at all (non-fallers; matched regarding age, sex and UPDRS III). The predictive value of DTC for future falls was investigated using logistic regression and ROC analysis. Results DTC for walking while checking boxes were higher ( p = .035) and DTC for checking boxes while walking showed a trend ( p = .090) towards higher DTC in fallers compared to non-fallers. The sum of standardized z -values of these two motor DTCs (total motor DTC) was higher in fallers compared to non-fallers ( p = .012), independent of PD duration and UPDRS III. The total motor DTC predicted future falls with an odds ratio of 2.64 (95% confidence interval: 1.05–6.63), and differentiated PD fallers from non-fallers with an accuracy of 0.81, reaching 71% sensitivity and 86% specificity. DTC for walking/subtracting tasks were comparable between PD fallers and non-fallers. Conclusion The total motor DTC of two simultaneously performed motor tasks, but not of a dual motor/cognitive task, differentiate PD future fallers from non-fallers with high accuracy, independent from disease severity. If confirmed in other studies, this parameter may have the potential to serve as one of the best predictors of falls in PD currently available. Moreover, our results may have implications for fall prevention strategies in PD ( Fig. 1 , Fig. 2 ).