In the present study, the mean (SD) change of the outcome measures were calculated at four and

12 months for the experimental and control groups of the two subgroups (walking speed ≤ 0.4 m/s and > 0.4 m/s). To determine whether treadmill training to improve walking has more effect PLX3397 on community-dwelling people after stroke who can walk faster (ie, baseline 10-m walk test of > 0.4 m/s), the mean difference (95% CI) between the experimental and control groups between subgroups (walking speed ≤ 0.4 m/s and > 0.4 m/s) for outcomes in the short-term (four months) and the long-term (12 months) were calculated.11 Sixty-eight community-dwelling people with stroke participated in this subgroup analysis. Adriamycin At baseline, all participants completed the six-minute walk test, a 10-m walk test at comfortable and fast speed, and the EuroQol 5Q-3L. However, five control participants did not complete the 10-m walk test at four months, and four control and one experimental participant did not complete it at 12 months. At baseline, 23 participants (34%) had a walking speed of ≤ 0.4 m/s and 45 participants (66%) had a walking speed of > 0.4 m/s.

Table 1 shows the baseline characteristics of the participants. Table 2 presents the six-minute walk test distance, the 10-m walk test at comfortable and fast speeds, and EuroQol EQ-5D-3L health status in the short term (four months) and in the long term (12 months) for both the experimental and control groups of the two subgroups. In the short term, there were statistically significant differences between the experimental and control groups between subgroups for the six-minute walk test distance and for the 10-m walk test comfortable speed. At four months, treadmill and overground walking training produced an extra distance of 72 m (95% CI 23 to 121) and an extra comfortable speed

of 0.16 m/s (95% CI 0.00 to 0.32) in the subgroup of participants with a baseline walking speed of > 0.4 m/s, compared with the subgroup with a baseline speed of ≤ 0.4 m/s. There was also a trend towards an extra fast speed of 0.17 m/s (95% CI –0.04 to 0.36). There was no extra effect of treadmill training in the faster walkers in terms of EuroQol 5Q-5D-3L. There were no statistically significant differences between the experimental and control groups between Ergoloid subgroups in the long term for any outcome. This study has shown that patients who walk slowly do worse on some outcomes at four months and 12 months than those with a moderate-to-fast walking speed. Whilst acknowledging the general limitations of post hoc secondary analyses, the chance of spurious findings was limited by dividing participants into subgroups based on previous evidence7 prior to analysis.12 At four months, treadmill and overground walking training for faster walkers (> 0.4 m/s) had a significant additional benefit in terms of walking distance and speed compared with slower walkers (≤ 0.4 m/s).