Search strategy and data extraction

We used the methodological recommendations of the Cochrane Collaboration and the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. Electronic databases, including PubMed, Science Citation Index, Cochrane Library, China National Knowledge Infrastructure, Chinese Journals Full-text database, China Biology Medicine disc, and Wanfang Database, were searched to retrieve related studies published before September 2017. The MeSH terms (the Medical Subject Headings) including “myelotomy”, “spinal cord injuries”, “rats”, “randomized controlled trial” and all related entry terms were searched. No restrictions were established on language, publication data, or publication status.

Two independent authors screened citations and publications identified by the initial search, to select potentially relative titles, review their abstracts, and determine whether they were eligible. The reference lists in included studies were also screened for any relevant publications that were not identified by the primary search. If data were not available, the authors were contacted by email. Disagreements were solved by a debate and consensus between both reviewers.

Two authors independently abstracted data from the selected articles, recording the following information: First author's name, publication year, model used to induce SCI (contusion or compression), injury level of spinal cord, modeling parameters, operation time after injury, SCI degree (severe or moderate), the number of rats in each group, rat characteristics, myelotomy procedures, Basso, Beattie and Bresnahan (BBB) score, and other experimental results. For each comparison, data were collected for the mean outcome, standard deviation, and the number of animals per group. If any data were only shown in graphs, GetData Graph Digitizer software was used to estimate data.