Robot-assisted gait training yields superior functional gains in non-ambulatory subacute stroke patients

A randomized controlled trial (NCT04910217) demonstrated that the addition of robot-assisted gait training (RAGT) using the Lokomat Pro FreeD to protocol-defined conventional rehabilitation in the early subacute phase of first-ever ischemic stroke produced significant gains in functional ambulation category (FAC), Timed Up and Go (TUG), and Berg Balance Scale (BBS). Whether these benefits differ across clinically relevant subgroups remains unclear. We investigated potential effect modification by baseline characteristics.

We performed a secondary analysis of a randomized controlled trial (n = 42 randomized; 36 completed follow-up: 17 RAGT, 19 conventional). Outcomes included ≥2-point FAC improvement and changes in BBS and TUG at 3 months. Subgroups were defined by stroke localization, baseline FAC (0-1 vs. 2-3), age, and recanalization treatment. Interaction effects were assessed using logistic regression for FAC; continuous outcomes were compared within subgroups using Mann-Whitney tests. No correction happened for multiple comparisons.

The largest differential effect of RAGT was observed in patients with low baseline ambulatory capacity (FAC 0-1): 90.0% of RAGT patients versus 45.5% in the conventional group achieved ≥2-point FAC improvement at 3 months (p = 0.039, n = 15). BBS gains were higher in this subgroup (median 27 vs. 12 points; p = 0.031), although this may partly reflect lower baseline values and greater recovery potential. Greater TUG improvement was observed in patients with anterior circulation stroke (p = 0.041). No subgroup showed a reversal of treatment effect.

RAGT may confer greater benefit in patients with severe initial gait impairment. These exploratory findings support targeted evaluation in larger trials.