Cognitive parameters can predict change of walking performance in advanced Parkinson’s disease – Chances and limits of early rehabilitation

Introduction: Links between cognition and walking performance in patients with Parkinson’s disease (PD), which both decline with disease progression, are well known. There is lack of knowledge regarding the predictive value of cognition for changes in walking performance after individualized therapy. The aim of this study is to identify relevant predictive cognitive and affective parameters, measurable in daily clinical routines, for change in quantitative walking performance after early geriatric rehabilitation.

Methods: Forty-seven acutely hospitalized patients with advanced PD were assessed at baseline (T1) and at the end (T2) of a 2-week early rehabilitative geriatric complex treatment (ERGCT). Global cognitive performance (Montreal Cognitive Assessment, MoCA), EF and divided attention (Trail Making Test B minus A, delta TMT), depressive symptoms, and fear of falling were assessed at T1. Change in walking performance was determined by the difference in quantitative walking parameters extracted from a sensor-based movement analysis over 20 m straight walking in single (ST, fast and normal pace) and dual task (DT, with secondary cognitive, respectively, motor task) conditions between T1 and T2. Bayesian regression (using Bayes Factor BF10) and multiple linear regression models were used to determine the association of non-motor characteristics for change in walking performance.

Results: Under ST, there was moderate evidence (BF10 = 7.8, respectively, BF10 = 4.4) that lower performance in the ∆TMT at baseline is associated with lower reduction of step time asymmetry after treatment (R2adj = 0.26, p ≤ 0.008, respectively, R2adj = 0.18, p ≤ 0.009). Under DT walking-cognitive, there was strong evidence (BF10 = 29.9, respectively, BF10 = 27.9) that lower performance in the ∆TMT is associated with more reduced stride time and double limb support (R2adj = 0.62, p ≤ 0.002, respectively, R2adj = 0.51, p ≤ 0.009). There was moderate evidence (BF10 = 5.1) that a higher MoCA total score was associated with increased gait speed after treatment (R2adj = 0.30, p ≤ 0.02).

Discussion: Our results indicate that the effect of ERGT on change in walking performance is limited for patients with deficits in EF and divided attention. However, these patients also seem to walk more cautiously after treatment in walking situations with additional cognitive demand. Therefore, future development of individualized treatment algorithms is required, which address individual needs of these vulnerable patients.