Implications of neuromuscular electrical stimulation on gait ability, balance and kinematic parameters after stroke: a systematic review and meta-analysis.

Certainty of the evidence was low or very low for most outcomes and high for only one, so these pooled estimates could change with further research. The review did not separate different types of electrical stimulation, different stimulation frequencies or different electrode placements. The authors name this as the study's main limitation, and several included trials did not report the type or frequency they used, so the pooled result is an average over treatments a clinician would choose between. Gait analysis speed, one of the two primary outcomes, showed evidence of publication bias on Egger's test, and 14 missing studies had to be imputed. Small trials with positive results may be over-represented. Blinding was weak in most trials: 15 used single blinding, 3 used no blinding, 7 did not say, and only 3 used double blinding. Nine trials did not describe how participants were allocated to groups, and one had a serious loss to follow-up in part of an outcome measure. Most pooled analyses had substantial unexplained heterogeneity, which the authors attribute to differences in time since stroke, sample size, overlap of confidence intervals and risk of bias. The subgroup findings about the best time after stroke, the number of treatment sessions and the number of stimulation sites rest on small numbers of trials, and in several subgroups the confidence interval crossed zero. What the stimulation was compared against was not consistent: three trials added it to treadmill training, others to conventional walking training or gave it alone, and four control groups used an ankle-foot orthosis.

The text supplied does not report a funding source or a competing-interest declaration for this review. It does state that one included trial was rated at high risk of bias for its selection of reported results because it was funded by a corporation.

Do not conclude that electrical stimulation makes stroke patients walk faster. The larger of the two primary comparisons, the ten-metre walk test in 1250 patients, found no difference between groups, and the smaller gait analysis result that did favour it showed evidence of publication bias, with 14 missing studies imputed. The authors themselves call the effect on walking speed inconclusive.