Eight studies (161 participants) were included. Two of the studies were randomised controlled trials (RCTs; 56 participants), one was a controlled trial (19 participants), one a one-group crossover study (16 participants), one a two-group crossover study (20 participants) and three were before-and-after studies (50 participants).
The mean score for methodological quality was 15 out of 32. The mean score was 16 for the controlled trials and 14 for the before-and-after studies. For each study, the two independent raters were in initial agreement on at least 70% of the 27 items assessed.
The meta-analysis showed a statistically significant improvement in gait speed with FES compared with control (WMD 0.18 m/s, 95% CI: 0.08, 0.28, p<0.01), based on an RCT, a controlled trial and a two-group crossover study including 71 participants. Although statistical heterogeneity was not significant at the 0.05 level (p=0.09), the RCT showed no difference between the treatment and control groups while the controlled trial and two-group crossover study showed significant benefit in favour of FES.
The mean effect size in the two-group crossover study of FES in the acute or sub-acute stage of stroke recovery was larger (d=1.43) than the mean effect size of the four studies of FES in the chronic stage of recovery (d=0.40). The mean effect size of the two studies of multi-channel FES (d=1.38) was larger than the mean from the three studies that used single-channel FES (d=0.09). The mean effect size from the five studies of FES (d=0.61) was larger than the mean effect size from the three studies of TENS (d=0.30).