Does robot-assisted gait rehabilitation improve balance in stroke patients? A systematic review

The aim of this systematic review was to summarize the improvements in balance after robot-assisted gait training (RAGT) in stroke patients. Two databases were searched: PubMed and Web of Knowledge. The most important key words are "stroke," "RAGT," "balance," "Lokomat," and "gait trainer." Studies were included if stroke patients were involved in RAGT protocols, and balance was determined as an outcome measurement. The articles were checked for methodological quality by 2 reviewers (Cohen's κ = 0.72). Nine studies were included (7 true experimental and 2 pre-experimental studies; methodological quality score, 56%-81%). In total, 229 subacute or chronic stroke patients (70.5% male) were involved in RAGT (3 to 5 times per week, 3 to 10 weeks, 12 to 25 sessions). In 5 studies, the gait trainer was used; in 2, the Lokomat was used; in 1 study, a single-joint wearable knee orthosis was used; and in 1 study, the AutoAmbulator was used. Eight studies compared RAGT with other gait rehabilitation methods. Significant improvements (no to large effect sizes, Cohen's d = 0.01 to 3.01) in balance scores measured with the Berg Balance Scale, the Tinetti test, postural sway tests, and the Timed Up and Go test were found after RAGT. No significant differences in balance between the intervention and control groups were reported. RAGT can lead to improvements in balance in stroke patients; however, it is not clear whether the improvements are greater compared with those associated with other gait rehabilitation methods. Because a limited number of studies are available, more specific research (eg, randomized controlled trials with larger, specific populations) is necessary to draw stronger conclusions.