Quantifying Poststroke Gait Deviations: A Meta-analysis of Observational and Cross-sectional Experimental Trials

Abstract

The hemiparetic walking pattern is a common feature of an upper motor neuron impairment caused by stroke. Previous studies have recorded deviations in hemiparetic walking patterns. We aimed to quantify gait alterations poststroke relative to healthy gait across a range of walking speeds by a systematic review and meta-analysis. We searched databases including MEDLINE via PubMed, Web of Science, SPORTDiscus, Cumulative Literature Index to Nursing and Allied Health Literature, and Cochrane Central Register of Controlled Trials from 1927 to the end of July 2015 for all instrumented measured gait variables in (a) observational studies with stroke patients only and (b) interventional studies in healthy, age-matched (elderly) participants, and those with stroke published in English. We analyzed data using a patient-weighted random-effects method. The heterogeneity of methods and measures employed across studies limited the intended analyses. The search pooled 2,223 articles, of which we included 49 observational and 124 interventional studies involving 960 healthy elders and 4,691 participants with stroke. As expected, those with stroke have gait-speed-related statistical differences from those without stroke. Specifically, this meta-analysis determined that (a) only 2 spatial parameters consistently changed according to gait speed – step length and stride length – and (b) irrespective of speed, 9 spatiotemporal measures, 5 sagittal kinematic, and 3 kinetic measures from the analysis were statistically different from the healthy counterparts ( P < .05). These findings suggest the importance of using instrumented measurement techniques for evaluating stroke gait rehabilitation.