Intervention: The experimental intervention was mechanically assisted walking training, such as treadmill or gait trainer without body weight support because the participants were able to walk a priori. The control intervention was defined as no intervention or an intervention that did not involve walking

training, ie, non-walking p38 MAPK apoptosis intervention. The experimental intervention was also compared with overground training. Session duration, session frequency, and program duration were recorded in order to assess the similarity of the studies. Outcome measures: Two walking outcomes were of interest speed (typically measured using 10-m Walk Test) and distance (typically measured using 6-min Walk Test). The timing of the measurements of outcomes and the procedure used to measure walking speed and distance were recorded in order to assess the similarity of the studies. Data were extracted from the included studies by a reviewer and cross checked by another reviewer. Information about the method (ie, design, participants, intervention, outcome measures) and outcome data (ie, mean (SD) walking speed and walking distance) were extracted. Authors were contacted where there was difficulty with data. The post-intervention scores were used to obtain the pooled estimate selleck inhibitor of the effect of intervention immediately (ie, post intervention) and beyond the intervention period (ie,

after a period of no intervention). A fixed effects model was used. In the case of significant SB-3CT statistical heterogeneity (I2 > 50%), a random effects model was applied to check the robustness of the results. The analyses were performed using The MIX–Meta-Analysis Made Easy programa (Bax et al 2006, Bax et al 2009). The pooled data for each outcome were reported as the weighted mean difference (MD) (95% CI). The search returned 5305 studies. After screening the titles, abstracts and reference lists, 65 papers

were retrieved for evaluation of full text. Fifty-six papers failed to meet the inclusion criteria and therefore nine papers (Pohl et al 2002, Ada et al 2003, Eich et al 2004, Weng et al 2006, Langhammer and Stanghelle 2010, Ivey et al 2011, Kuys et al 2011, Olawale et al 2011, Ada et al 2013) were included in the review. See Appendix 2 on the eAddenda for a summary of the excluded papers. Figure 1 outlines the flow of studies through the review. Six randomised trials investigated the effect of mechanically assisted walking training on walking speed and walking distance, two on walking speed, and one on walking distance. The quality of the included studies is outlined in Table 1 and a summary of the studies is presented in Table 2. Quality: The mean PEDro score of the included studies was 6.7. Randomisation was carried out in 100% of the studies, concealed allocation in 67%, assessor blinding in 67%, and intention-to-treat analysis in 44%.