During the first 4-8 hours after spinal cord injury, partial tissue necrosis begins to appear in the impacted area, and within 24-48 hours, tissue necrosis becomes apparent and progresses extensively to the adjacent spinal cord segments that have been impacted. This change reaches its peak at 24 h, following which there are minimal changes in the next 24 h [3, 8]. Spinal cord injury produces two types of damages: primary and secondary. Features of primary damage include the ejection light of electrolytes because of mechanical damage, bleeding, and cell damage, while features of secondary damage include edema, inflammation, ischemia, growth defects, the release of cytokines, reduction in blood flow, calcium accumulation, peroxide glass, etc [9-11].

Since most patients with acute spinal injuries die from breathing difficulties and acute circulatory disorders, it is important to align the treatment of damaged vertebrae with neurological examination during the acute stage to prevent further damage [6]. It is very important to supply oxygen to patients during the acute period as it can reduce secondary ischemic damage to the spinal cord. Patients with spinal cord injuries suffer from interruptions to the cardiac accelerator nerves, affecting the sympathetic neuroanatomy, which result in a decrease in heart rate along with a decrease in cardiac output [12, 13]. Surgery for acute spinal cord injury involves removing structures pressing against the spinal cord and aligning the spine using instruments or bone transplants, which is done when the deformation is not conquered in a non-surgical manner, or if the spine is contaminated with progressive neurological or open wounds [4, 10]. Furthermore, it has been reported that administering large amounts of steroids to patients within 8 hours of damage is effective in recovery [14]. The mechanism of action of steroids is not exactly clear, but it is known to stabilize the cell membrane, neutralize the peroxide, reduce the accumulation of calcium in the cell, decrease the excitant amino acid, decrease the edema of tissue, and increase blood flow to the spinal cord. The guidelines for steroid medication instruct to first inject methylprednisolone (30 mg per kilogram of weight) into the vein for 15 min, followed by 45 minutes of rest, and then injecting 5.4 mg per kilogram of weight per hour over 23 hours for 3-5 days [15-17].

The effects of steroid therapy are unclear for various reasons. First, the positive effect of high-dose steroid injection was reported only in a small number of groups [2]; second, improvements in spinal cord segments did not lead to improvements in survival rate or quality of life [7]; third, a lack of reproducibility of results in other studies; and fourth, its use results in a combination of side effects such as high blood sugar, infection, delayed wound treatment, and gastrointestinal bleeding [18]. There are no reports of getting better results from using steroids for longer than 24 hours in spinal cord damage. Long-term steroid use only increases the risk of complications such as delayed ulcer healing and infection. Thus, attention should be paid to the various side effects of steroid therapy.

Treatment outcomes for acute spinal cord injury vary depending on the location and scope of the damage. Moreover, it is not clear whether a treatment is really effective or what factors affect the treatment because of insufficient prognosis for acute spinal cord injury and its associated factors. Furthermore, the primary and secondary damage mechanisms can proceed more broadly and rapidly in the acute period, therefore various factors of impact need to be identified.

Despite the advantages and disadvantages, clinical trials apply steroid therapy as a primary post-damage treatment for acute spinal cord injury. Therefore, this study aims to analyze the effects of high-dose steroid injection therapy applied to patients with acute spinal cord injuries using meta-analysis and identify the factors related to its effectiveness.

The specific objectives of this study are as follows:

1. To quantitatively understand or estimate the effectiveness of steroid pulse therapy in treating patients with acute spinal cord injuries.
2. To investigate the effects of factors influencing the effectiveness of steroid pulse therapy in patients with acute spinal cord injuries.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols [PRISMA-P] 19 will be set as a guidebook for the protocol, and the review methods will be designed based on the meta-analysis of observational studies in epidemiology: a proposal for reporting PRISMA and the Cochrane Collaboration Handbook [18].

Two independent reviewers (SH and one meta-analysis specialist) selected and reviewed studies. The reviewers independently screened the titles and abstracts of the selected studies to determine whether each citation met the inclusion criteria and assessed eligibility based on a full-text review. The reviewers compared their lists and resolved any differences in opinion through discussion; potential conflicts were resolved with the help of one external specialist who is a nursing science professor. We attempted to plan, perform, and report this meta-analysis in compliance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses [PRISMA] [18-25]. Related articles published in English were identified and selected by searching databases of Pubmed, Medline, the Cochrane Central Register of Controlled Trials, Embase, and CINAHL (until July 31, 2019) using the following mesh search terms: “spinal cord/traumatic/acute spinal cord injury,” “steroid/corticosteroid/methylprednisolone/dexamethasone/hydrocortisone/naloxone hydrochloride,” “mega-dose/high-dose/steroid pulse therapy,” “randomized controlled trial/RCT,” and “observational study.” We combined these terms in accordance with the instructions of the database. In addition, the reference lists of retrieved studies and previous reviews and meta-analyses were reviewed and manually searched. No attempt to identify unpublished reports were made. Our search strategies will be peer-reviewed by a second information specialist using the peer-review of the electronic search strategy method

To select study titles, we first screened identified titles or abstracts and then reviewed the full-text of the articles. Studies were considered eligible if the following criteria were met [1]: had randomized controlled trial design or observational studies involving adult patients [2]; study subjects had undergone high-dose steroid therapy at the time of admission after injuries [3]; had comparison groups of high-dose steroid injection therapy and non-steroid therapy [4]; had reported the factors affecting the effectiveness of steroid pulse therapy, and [5] had mentioned relative outcomes such as related complications.

The following data and information were extracted: first author, year of publication, study design, injury levels of spinal cords, time of treatment application, steroid pulse therapy and non-steroid pulse therapy treatment, main outcomes, side effects, factors relating to main outcomes, and study results. The study selection and data extraction were conducted by two authors independently, and discussions were held to resolve disagreements. The following main outcomes were extracted: sensory change and impairment score, which states whether, after applying steroid pulse therapy, the patients suffer from complications such as urinary tract infection, pneumonia, and sepsis. The extracted data and information from studies are presented in Tables 1,2, and 3.

The Cochrane Collaboration’s “risk of bias [RoB]” and “Risk of Bias Assessment tool for Non-randomized Studies [RoBANS]” were used to evaluate the methodological quality and risk of bias of included randomized controlled trials and non-randomized controlled trials; seven or eight specific domains were examined and measured using this tool: RoB contains domains of sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and “other” issues. RoBANS contains 6 domains, including the selection of participants, confounding variables, measurement of intervention (exposure), blinding of outcome assessment, incomplete outcome data, and selective outcome reporting. Every domain can be classified as “low risk of bias,” “high risk of bias,” or “unclear risk of bias” in accordance with the judgment criteria (Cochrane Handbook for Systematic Reviews of Intervention. Part 2: 8.5) [26-28].

The extracted data were fed into the freeware program Review Manager [RevMan] Version 5.3. Binary outcomes were presented as Mantel–Haenszel style Odds Ratios (ORs) with 95% Confidence Intervals (CIs), and continuous outcomes were reported as inverse variance mean differences [MDs]. A fixed-effect model was adopted in cases of homogeneity [p value of χ² test > 0.10 and I²< 50%], while a random-effects model was used in cases of obvious heterogeneity (p value of χ² test < 0.10 and I² ≥ 50%). Publication bias was evaluated by the demonstration of funnel plots and assessed using the Rosenthal and Rosenberg fail-safe numbers; the latter was weighted by study variance. Fail-safe numbers of less than 5n+10 (where n is the number of studies in the meta-analysis) were considered indicators of publication bias [29].

Computerized and manual searches resulted in 54 citations, 34 of which were excluded because of duplication. From the remaining 20, 7were excluded after reviewing their titles and abstracts. In total, 13 potentially eligible articles were retrieved for full-text review, out of which 5 were excluded. Finally, we included 3 randomized controlled trials and 5 observation studies for synthesized analysis. Detailed review process is presented in Fig. (1).

The basic characteristics of the 8 included studies are presented in Table 1. A total of 2418 patients were involved, including 934 for non-steroid pulse therapy, 1484 for steroid pulse therapy (NASCIS-II or III regimen), and the study took place in the US, France, Italy, Japan, and Canada. The sample size ranged from 19 to 1624.

Fig. (1). Flowchart of the study search and inclusion criteria (Literature search and results).

Figs. (2 and 3) illustrate the methodological quality of included randomized controlled trials. Random sequence generation was detailed in all studies (100%), “allocation concealment” was described in all studies (100%), and blindness to “blinding of participants and personnel” in two studies (66.7%). The antecedents of outcome assessments was detailed in one study (33.3%) and “incomplete outcome data” was described in two studies (66.7%). “Selective reporting” was described in all studies (100%). “Other bias” assessed the professionalism of the arbitrator, existence of the intervention manual, and a number of research samples according to the criteria of the previous studies, and was described in three studies (100%), determining that the bias risk was low. In a non-randomized controlled trial qualitative assessment, “comparability of parts” to evaluate selection bias and “selection of parts” were found to be unbiased in all the five studies (100%), while “confounding variables” was less bias risk in all studies. The “Measurement of exposure” of the performance bias, “Blinding of Outcome Assessment” and “Selective reporting” of the detection bias, and “Outcome Evaluation” of the reporting bias were all low in all five studies (100.0%). The “incomplete outcome data” to evaluate attrition bias was found to be highly perverse in only one study. The results of the bias risk assessment for the three randomized controlled trial studies were judged to be low, since more than 50% of these studies were assessed to be low bias in the assessment items except for the cover-up of the “allocation concealment” and “blinding of participants and personnel.” In addition, the five non-randomized controlled trial studies were found to be mostly low in quality assessment other than the “Blinding of Outcome Assessment” items, and were suitable for presenting the findings collectively.

Fig. (2). Summarized results of the quality evaluation for RCT literatures using RoB.

Fig. (3). Summarized results of the quality evaluation for NRCT literatures using RoBANS.