Long-term Training With a Brain-Machine Interface-Based Gait Protocol Induces Partial Neurological Recovery in Paraplegic Patients.

were excluded. The American Academy of Neurology level of evidence criteria (grade I, most robust, to grade IV, weakest evidence) was used to classify the strength of evidence of individual studies, and the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) system was used to evaluate the overall body of evidence. Thirty-seven studies were included in the meta-analysis, of which 36 studies were used for the analysis ofOS and 8 studies for the analysis of PFS. The meta-analysis indicated a significantly improved OS after GTR compared with STR at 1 year (relative risk [RR], 0.62; 95% confidence interval [CI], 0.56-0.69; P , .01; number needed to treat [NNT] 1⁄4 9) and after 2 years (RR, 0.84; 95% CI, 0.79-0.89; P , .01; NNT 1⁄4 17). STR was superior to biopsy only in terms of 1-year OS (RR, 0.85; 95% CI, 0.800.91; P , .01). However, no significant difference in OS was observed between STR and biopsy after 2 years (RR, 0.99; 95% CI, 0.97-1.00; P 1⁄4 .09). At 6 months, GTR was better than STR in terms of PFS, but the differences were not statistically significant (RR, 0.72; 95% CI, 0.48-1.09; P 1⁄4 .12). However, GTR was indeed significantly better than STR in terms of PFS at 1 year (RR, 0.66; 95% CI, 0.43-0.99; P , .01; NNT 1⁄4 26). The risk for progression was also significantly reduced by STR compared with biopsy alone at 6 months (RR, 0.72; 95% CI, 0.51-1.00; P 1⁄4 .05; NNT 1⁄4 321); however, these differences were not significant at 1 year (RR, 0.96; 95%CI, 0.791.17; P 1⁄4 .69). None of the individual studies included in the analysis were graded class I; 4 were class II; 15 were class III; and 18 were class IV. However, with the inclusion of only class II studies, similar trends were observed compared with the larger meta-analysis for 1-year OS (RR, 0.62; 95% CI, 0.55-0.69; P , .01; NNT 1⁄4 5) and 2-year OS (RR, 0.72; 95% CI, 0.49-1.07; P 1⁄4 .11; NNT 1⁄4 7). Final assessment by GRADE criteria for quality of evidence was moderate for OS and the metaanalysis using only class II studies and low for all other analysis. Certain limitations of this article acknowledged by the authors include the fact that GTR and STR were defined by the authors of the individual studies, introducing bias and limiting comparison between centers. GTR and STR cohorts were not matched according to prognostically important variables, and publication bias existed to exclude small studies favoring STR or biopsy over GTR. Importantly, surgical complications, adverse events, and quality of life were not compared between the GTR and STR cohorts. Despite these limitations, the article by Brown et al provides reasonable evidence, based on the sheer consistency of data across 41 117 unique patients, that GTR results in significant improvement in PFS and OS compared with STR or biopsy alone and that GTR should be preferred when clinically feasible.