The effect of the training on health status did not differ between the subgroups at any assessment point. Therefore, although treadmill and overground walking training is recommended for people with stroke to improve walking capacity

and speed, the present study’s findings showed that the effect of intervention was different depending on initial walking speed. In the present trial, a walking speed of 0.4 m/s was used to separate participants into two subgroups. Those with speeds ≤ 0.4 m/s were considered to be severely impaired slow walkers and those with speeds above 0.4m/s were considered to be moderate-to-fast walkers. A cut off of 0.4 m/s meant Abiraterone molecular weight that the subgroup of slow walkers included the lowest four categories (physiological walker, limited household walker, unlimited household walker and most-limited community walker) and the moderate-to-faster walkers included the highest

two categories (least-limited community walker and community walker).7 This same cut off was used to define the slow walkers in the recent LEAPS trial.13 The additional benefit of treadmill and overground walking training related to baseline walking speed declined over time. Immediately after four months of intervention, the faster walkers had an additional benefit of 72 m over PD0332991 six minutes compared with the slower walkers. By 12 months, the additional benefit had disappeared. The additional benefit in comfortable and fast-walking speeds for the moderate-to-fast walkers mirrored the changes in six-minute walking distance. The size of the additional benefit at 0.16 m/s and 0.175 m/s for comfortable and fast, respectively, indicate that these benefits are clinically meaningful.14 and 15 The finding that there is a differential effect of treadmill and overground walking training based on baseline comfortable walking speed is consistent with other intervention

trials after stroke, with slower walkers performing worse compared Oxygenase to faster walkers. In a community stroke trial of exercise classes and a home program, larger improvements in walking speed and six-minute walking distance were found for faster walkers compared with slower walkers.5 The major clinical implication of this study and others, which find significant subgroup intervention effects, is the need to target intervention. Given the heterogeneity of stroke, the ‘one size fits all’ approach of clinical trials runs the risk of discounting worthwhile intervention. The present study’s findings suggest that the treadmill and overground walking intervention should be implemented for those with initial walking speeds of greater than 0.4 m/s, whereas poor walkers may need additional and/or different interventions to enhance their community participation.