Clinical Benefit of Rehabilitation Training in Spinal Cord Injury

A Systematic Review and Meta-Analysis

Ruimeng Duan, MS; Mingjia Qu, MS; Yashuai Yuan, MS; Miaoman Lin, MS; Tao Liu, MS; Wei Huang, MS; Junxiao Gao, MS; Meng Zhang, MS; Xiaobing Yu, MD


Spine. 2021;46(6):E398-E410. 

In This Article

Materials and Methods

Ethics Statement

This study is a systematic review and meta-analysis using comprehensive databases. Neither animals nor human specimens were used in this study, and therefore the ethics committee approval is not applicable. This study was carried out according to the Cochrane Handbook for systematic reviews of interventions.[20]

Study Sources and Search Strategy

This study was carried out in February 2020. The meta-analyst attempts to ensure a comprehensive literature search (including grey literatures) to fully reflect the existing evidential base and assess the impact of methodological variations through meta-analysis.[21] RCTs and CTs were searched from PubMed (1966–Nov, 2018), Embase (1966–Nov, 2018), Cochrane (~Nov 2018), the database of the U.S. National Institutes of Health (, and World Health Organization International Clinical Trials Registry Platform independently by two authors. Eligible English publications (including conference papers, academic papers, and research projects) were searched using the searching terms and strategies as follows: "spinal cord injury," "transcranial magnetic stimulation," "functional electrical stimulation," "exercise," "training," and "Locomotor and robot." Publication searching were implemented using free-text, title, and abstract terms with controlled vocabulary (MeSH). The searching strategy is "transcranial direct current stimulation[Title/Abstract]" OR "tDCS[Title/Abstract]" OR "repetitive transcranial magnetic stimulation[Title/Abstract]" OR "rTMS[Title/Abstract]" OR "deep brain stimulation[Title/Abstract]" OR "Epidural[Title/Abstract]" OR "Surface FES for UE" OR "Surface FES for UE and LE" OR "Implanted FES for LE" OR "Implanted FES for trunk" OR "functional electrical stimulation" OR "upper extremity" OR "lower extremity" AND "spinal cord injury[MeSH Terms]" AND "training." RCTs/CTs were filtered using "trial." Additional publications of RCT or CTs in the reference lists were identified.

Inclusion and Exclusion Criteria for Study Selection

Titles, abstracts, and the full-texts of all articles were scanned and the publications of (1) non-RCTs or non-CTs; (2) cross-sectional studies; (3) reviews; (4) case reports; (5) and articles without available data were excluded. Reports were included if they met the inclusion criteria: (1) RCTs or CTs; (2) patients (>15 years of age) with acute or subacute SCI for less than 1 year or chronic SCI for more than 1 year; (3) patients were treated with physical intervention for improving nerve recovery including rTMS, FES, ABT, Robotic-assisted locomotor training, instead of surgery, pharmacology, or stem cells.

Data Extraction and Outcomes

This step was independently implemented by two reviewers. Available data were extracted from each study. Demographic (patient number, age, and illness duration), clinical, and treatment characteristics (intervention strategy) of individuals with SCI were extracted from included studies. Available data of therapeutic efficacy in walking capacity and motor function that were evaluated using different methods like American Spinal Injury Association (ASIA) motor score and class, ASIA upper/lower extremities motor score (ASIA UEMS/LEMS), upper extremity functional index (UEFI), modified Ashworth scale (MAS) score, 10 m walking test (10MWT, step length and cadence), and 6-minute walking test (6MWT) were extracted. Safety data were the incidence of the adverse events during intervention. Necessary consultations on the authors or journals were used if the data were incomplete or unavailable for current publications via email. Otherwise the publications with unavailable data were removed.

Quality Assessment

This step was independently implemented by two reviewers. The available data extracted from the studies were assessed using from the Physiotherapy Evidence Database (PEDro; scale scores. The method of using PEDro to assess non-RCT trials is based on previous studies.[22,23] PEDro scale is consisted of 11 items for assessing the quality, including random allocation, concealed allocation, baseline similarity, blinding of subjects, blinding of therapists, blinding of all assessors, measures of key outcomes from more than 85% of subjects, intention-to-treat analysis, between-group statistical comparison, point measures of variability, and total of them.

Trials with score less than six were regarded as low quality, otherwise high quality. The third-reviewer adjudication was used for the disagreements. Funnel plot was performed to assess the publication bias for the included studies.

Statistical Analysis

Meta-analysis was performed using Revman 5.0 software (Cochrane Collaboration, Oxford, United Kingdom). 95% confidence interval (CI) and individual mean difference (MD) or odd ratio (OR) values were used. Publication heterogeneity was assessed using Chi-test (χ2) and I-square (I2 ). Publication heterogeneity (I 2 ≥ 50%) was explored using meta-regression models and assessed. Sensitivity analysis was conducted for the heterogeneity sources of publications with low quality. Meta-analysis of publications with homogeneity (I 2 < 50%) and heterogeneity (I 2 ≥ 50%) was performed using fixed- and random-effects model, respectively.

Weighted MD values (95% CI) were calculated using the inverse variance (I-V) and DerSimonian–Laird (D–L) method for the included studies in fixed- and random-effects model, respectively. For adverse events analysis, OR was used. Statistical differences by meta-analysis were identified as P < 0.05.