In Finland there is approximately 24,000 people who suffer from stroke each year. Walking disorders are the most important symptoms that limit the ability to function after cerebrovascular accident. The effectiveness of walking and endurance training has been studied in acute and subacute patients and the effectiveness of exercise is often assessed at the level of body structure and activity. The effectiveness of walking and endurance training on participation has been studied less. The purpose of this study was to find out the effectiveness of walking and endurance training on walking endurance compared to no-training/placebo, usual care or other training in patients with chronic stroke. Walking was examined at the level of activity and participation according to international classification of functioning, disability and health (ICF). This study is a part of Kela´s intensive medical rehabilitation (VAKU) -project.

A systematic literature search was conducted in VAKU-project in databases (1/2008-5/2017): Cinahl, CENTRAL, Cochrane Database of Systematic Reviews, Embase, OVID Medline, PEDro and WOS. Selection criterias were defined with PICOS-strategy: (P) diagnosed with stroke, ≥ 6 months earlier (I) walking or endurance training, (C) no-training/placebo, usual care, other training, (O) walking endurance at the level of activity and walking ability at the level of participation and (S) randomized controlled trial. Meta-analysis was performed with Review Manager 5.3 – programme using random effect model. In the main analysis effectiveness of walking and endurance training compared to combined control group was evaluated using mean difference (MD) at the level of activity and using standardized mean difference (SMD) at the level of participation. Respectively, the sub-analyses separately assessed the effectiveness of walking and endurance training compared to three different control groups. The quality assessment of the studies was performed with Cochrane Risk of Bias tool.

28 studies were included in the study and there were 1270 participants in total. In average intervention duration was 9 weeks (range 1-24) and frequency 4 times per week (SD 1.5). At the level of activity in groups of walking and endurance training walking endurance increased in the 6 minute walking test 12.58 meters (95% CI [5.55, 19.60]; p=0.0005; n=919) more compared to control group. According to results of the meta-analysis walking and endurance training are probably more effective to increase walking endurance at the level of activity than no-training/placebo, usual care or other training (level B evidence). There was no statistical difference between intervention and control group (SMD -0.04; 95% CI [-0.43, 0.35]; p=0.83; n=152) and there is a lack of reliable evidence about the effectiveness of walking and endurance training on walking at the level of participation (level D evidence). In the sub-analyses a statistical difference was only between walking training compared to other training (MD 12.81; 95% CI [3.25, 22.38]; p=0.009; n=241), but the statistical difference disappeared in the sensitivity analysis. Walking training is probably comparable training model to usual care (level B evidence) and it might be comparable to other type of training (level C evidence) to increase walking endurance. Endurance training might increase walking endurance more than low intensity training (level C evidence) and endurance training might be comparable to usual care, but there is a lack of reliable evidence of that (level D evidence). Statistical heterogeneity was low or moderate and clinical heterogeneity was observed in interventions and control groups.

According to this study walking and endurance training in the chronic stage of stroke is probably more effective to increase walking endurance compared to no-training/placebo, usual care or other training. However the result is clinically small. More research is needed on the effectiveness of walking and endurance training on participation and the effectiveness of training at different intensities. In addition, in future it would be important to explore how the training can be implemented cost-effectively.