Factors affecting the efficiency of walking independence in patients with subacute stroke following robot-assisted gait training with conventional rehabilitation.

Factors affecting the efficiency of walking independence in patients with subacute stroke following robot-assisted gait training (RAGT) and conventional treatment (RAGT-CT) were examined. This retrospective cohort study included 37 patients with stroke [n = 11 ischemic; n = 26 hemorrhagic; median poststroke interval, 28 days (interquartile range, IQR, 24-42)] who underwent RAGT using Welwalk for a median of 3 weeks (IQR, 2-4) followed by conventional training (median, 129 days; IQR, 114-146). The primary outcome was the change in functional independence measure (FIM)-walk item score from before to after RAGT (FIM-walk efficiency). The secondary outcome was the FIM-walk score at discharge. The independent variables included sensorimotor function [lower extremity (LE) motor and sensory scores and trunk function from the Stroke Impairment Assessment Set (SIAS) and Berg Balance Scale (BBS)] and cognitive function (FIM-cognitive, MMSE and Cognitive-related Behavioral Assessment) before RAGT-CT and RAGT dose duration per session, total steps and average treadmill speed at week 1, and number of sessions). We first determined the bivariate associations of each independent variable with the FIM-walk efficiency at the end of the RAGT period as decided by the therapists and with the FIM-walk score at discharge. Hierarchical multiple regression revealed that only the FIM-cognitive score was a significant predictor of the FIM-walk efficiency at the end of the RAGT period (β = 0.47; P < 0.01, adjusted R2 = 0.21) after accounting for age, days post-stroke, SIAS-total lower extremity (SIAS-LE) motor score, and number of RAGT sessions (all nonsignificant). Furthermore, only the SIAS-trunk score was a significant predictor of the FIM-walk score at discharge (β = 0.52; P < 0.01; adjusted R2 = 0.65) after accounting for age, days post-stroke, FIM-cognitive score, SIAS-LE motor score, and average treadmill speed at week 1 (all nonsignificant). Although patients with better cognition at the start of locomotor training achieved the pragmatic targets for terminating RAGT and proceeding with conventional therapy at a faster rate, the outcome at discharge is mainly dependent on early trunk function.