Walking-adaptability therapy after stroke: results of a randomized controlled trial

Abstract

Background: The ability to adapt walking to environmental properties and hazards, a prerequisite for safe ambulation, is often impaired in persons after stroke.

Research question: The aim of this study was to compare the efficacy of two walking-adaptability interventions: a novel treadmill-based C-Mill therapy (using gait-dependent augmented reality) and the standard overground FALLS program (using physical context). We expected sustained improvements for both treatment groups combined but hypothesized better outcomes for C-Mill therapy than the FALLS program due to its expected greater amount of walking practice.

Methods: In this pre-registered single-centre parallel group randomized controlled trial, forty persons after stroke (≥ 3 months ago) with walking and/or balance deficits were randomly allocated to either 5 weeks of C-Mill therapy or the FALLS program. The primary outcome measure was the standard walking speed as determined with the 10-meter walking test (10MWT). Additionally, context-specific walking speed was assessed in environments enriched with either stationary physical context (10MWT context) or suddenly appearing visual images (Interactive Walkway obstacles). The walking-adaptability scores of those enriched walking tests served as secondary outcome measures. Furthermore, a cognitive task was added to all three assessments to evaluate dual-task performance in this context. Finally, the participants' experience and amount of walking practice were scored. The outcome measures were assessed at four test moments: pre-intervention (T0), post-intervention (T1), 5-week post-intervention retention (T2), and 1-year post-intervention follow-up (T3).

Results: No significant group differences were found between the interventions for the primary outcome measure standard walking speed, but we found a greater improvement in context-specific walking speed with stationary physical context of the C-Mill therapy compared to the FALLS program at the post-intervention test, which was no longer significant at retention. Both interventions were well received, but C-Mill therapy scored better on perceived increased fitness than the FALLS program. C-Mill therapy resulted in twice as many steps per session of equal duration than the FALLS program. The "change-over-time" analyses for participants of both interventions combined showed no significant improvements in the standard walking speed; however, significant improvements were found for context-specific walking speed, walking adaptability, and cognitive dual-task performance.

Significance: This study showed no between-group differences between the novel treadmill-based C-Mill therapy and the standard overground FALLS program with respect to the primary outcome measure standard walking speed. However, the greater amount of walking practice observed for the C-Mill group, an essential aspect of effective intervention programs after stroke, may underlie the reported increased perceived fitness and observed increased context-specific walking speed for the C-Mill group directly after the intervention. Although the "change-over-time" results for all participants combined showed no improvement in the standard walking speed, context-specific walking speed and walking adaptability showed sustained improvements after the interventions, underscoring the importance of including walking-adaptability training and assessment in rehabilitation post stroke.

Trial registration: The Netherlands Trial Register NTR4030 . Registered 11 June 2013.

Keywords: Gait; Rehabilitation; Stroke; Therapy; Walking adaptability; Walking speed.

Fig. 1

Snapshots of the two interventions aimed at improving walking adaptability. (Left) Walking-adaptability exercises of treadmill-based C-Mill therapy (CT); A obstacle avoidance, B goal-directed stepping, C gait acceleration and deceleration, and D a functional and interactive walking-adaptability game. (Right) Obstacle course of the overground FALLS program (FP); A obstacle avoidance, B walking over uneven terrain, C tandem walking, and D slalom walking

Fig. 2

Schematic representations of the three walking assessments and the related outcome measures; A standard 10MWT, B 10MWT with physical context, comprising three obstacles (at 2.0 m, 7.5 m, and 9.0 m, of length × width × height 9.0 × 20.0 × 4.5 cm, 4.5 × 20.0 × 9.0 cm and 33.0 × 21.0 × 11.5 cm, respectively), a 2-m tandem-walking path (with a width of 20 cm) and three stepping targets (participants’ shoe length + 4 cm by shoe width + 4 cm), C IWW obstacles, a 6.6 × 0.9-m walkway instrumented with multiple Microsoft Kinect for Windows sensors and a projector to present two suddenly appearing obstacles (projected red rectangles of 0.4 × 0.9 m) in a gait-dependent (i.e., one obstacle at a predicted foot-placement position appearing two steps ahead) and a position-dependent (i.e., one obstacle at an unpredictable but predefined position appearing when a participant’s ankle was within 2 m from that obstacle) manner. Participants performed those assessments with and without a cognitive dual task

Fig. 3

Flow chart with an overview of the procedures and group distribution