Neurological recovery following traumatic spinal cord injury: a systematic review and meta-analysis

MirHojjat Khorasanizadeh Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by MirHojjat Khorasanizadeh in
jns
Google Scholar
PubMed
Close
 MD
,
Mahmoud Yousefifard Physiology Research Center and Department of Physiology, Faculty of Medicine, Iran University of Medical Sciences, Tehran, Iran;

Search for other papers by Mahmoud Yousefifard in
jns
Google Scholar
PubMed
Close
 PhD
,
Mahsa Eskian Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by Mahsa Eskian in
jns
Google Scholar
PubMed
Close
 MD
,
Yi Lu Department of Neurosurgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts;

Search for other papers by Yi Lu in
jns
Google Scholar
PubMed
Close
 MD, PhD
,
Maryam Chalangari Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by Maryam Chalangari in
jns
Google Scholar
PubMed
Close
 MD
,
James S. Harrop Departments of Neurological and Orthopedic Surgery, Thomas Jefferson University, Philadelphia;
Neurosurgery, Delaware Valley Regional Spinal Cord Injury Center, Thomas Jefferson University, Philadelphia, Pennsylvania; and

Search for other papers by James S. Harrop in
jns
Google Scholar
PubMed
Close
 MD
,
Seyed Behnam Jazayeri Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by Seyed Behnam Jazayeri in
jns
Google Scholar
PubMed
Close
 MD
,
Simin Seyedpour Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by Simin Seyedpour in
jns
Google Scholar
PubMed
Close
 MD
,
Behzad Khodaei Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by Behzad Khodaei in
jns
Google Scholar
PubMed
Close
 MD
,
Mostafa Hosseini Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran

Search for other papers by Mostafa Hosseini in
jns
Google Scholar
PubMed
Close
 PhD
, and
Vafa Rahimi-Movaghar Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences;

Search for other papers by Vafa Rahimi-Movaghar in
jns
Google Scholar
PubMed
Close
 MD
Full access

OBJECTIVE

Predicting neurological recovery following traumatic spinal cord injury (TSCI) is a complex task considering the heterogeneous nature of injury and the inconsistency of individual studies. This study aims to summarize the current evidence on neurological recovery following TSCI by use of a meta-analytical approach, and to identify injury, treatment, and study variables with prognostic significance.

METHODS

A literature search in MEDLINE and EMBASE was performed, and studies reporting follow-up changes in American Spinal Injury Association (ASIA) Impairment Scale (AIS) or Frankel or ASIA motor score (AMS) scales were included in the meta-analysis. The proportion of patients with at least 1 grade of AIS/Frankel improvement, and point changes in AMS were calculated using random pooled effect analysis. The potential effect of severity, level and mechanism of injury, type of treatment, time and country of study, and follow-up duration were evaluated using meta-regression analysis.

RESULTS

A total of 114 studies were included, reporting AIS/Frankel changes in 19,913 patients and AMS changes in 6920 patients. Overall, the quality of evidence was poor. The AIS/Frankel conversion rate was 19.3% (95% CI 16.2–22.6) for patients with grade A, 73.8% (95% CI 69.0–78.4) for those with grade B, 87.3% (95% CI 77.9–94.8) for those with grade C, and 46.5% (95% CI 38.2–54.9) for those with grade D. Neurological recovery was significantly different between all grades of SCI severity in the following order: C > B > D > A. Level of injury was a significant predictor of recovery; recovery rates followed this pattern: lumbar > cervical and thoracolumbar > thoracic. Thoracic SCI and penetrating SCI were significantly more likely to result in complete injury. Penetrating TSCI had a significantly lower recovery rate compared to blunt injury (OR 0.76, 95% CI 0.62–0.92; p = 0.006). Recovery rate was positively correlated with longer follow-up duration (p = 0.001). Studies with follow-up durations of approximately 6 months or less reported significantly lower recovery rates for incomplete SCI compared to studies with long-term (3–5 years) follow-ups.

CONCLUSIONS

The authors’ meta-analysis provides an overall quantitative description of neurological outcomes associated with TSCI. Moreover, they demonstrated how neurological recovery after TSCI is significantly dependent on injury factors (i.e., severity, level, and mechanism of injury), but is not associated with type of treatment or country of origin. Based on these results, a minimum follow-up of 12 months is recommended for TSCI studies that include patients with neurologically incomplete injury.

ABBREVIATIONS

AIS = ASIA Impairment Scale; AMS = ASIA motor score; ASIA = American Spinal Injury Association; CCS = central cord syndrome; LoE = level of evidence; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses; TSCI = traumatic spinal cord injury.

OBJECTIVE

Predicting neurological recovery following traumatic spinal cord injury (TSCI) is a complex task considering the heterogeneous nature of injury and the inconsistency of individual studies. This study aims to summarize the current evidence on neurological recovery following TSCI by use of a meta-analytical approach, and to identify injury, treatment, and study variables with prognostic significance.

METHODS

A literature search in MEDLINE and EMBASE was performed, and studies reporting follow-up changes in American Spinal Injury Association (ASIA) Impairment Scale (AIS) or Frankel or ASIA motor score (AMS) scales were included in the meta-analysis. The proportion of patients with at least 1 grade of AIS/Frankel improvement, and point changes in AMS were calculated using random pooled effect analysis. The potential effect of severity, level and mechanism of injury, type of treatment, time and country of study, and follow-up duration were evaluated using meta-regression analysis.

RESULTS

A total of 114 studies were included, reporting AIS/Frankel changes in 19,913 patients and AMS changes in 6920 patients. Overall, the quality of evidence was poor. The AIS/Frankel conversion rate was 19.3% (95% CI 16.2–22.6) for patients with grade A, 73.8% (95% CI 69.0–78.4) for those with grade B, 87.3% (95% CI 77.9–94.8) for those with grade C, and 46.5% (95% CI 38.2–54.9) for those with grade D. Neurological recovery was significantly different between all grades of SCI severity in the following order: C > B > D > A. Level of injury was a significant predictor of recovery; recovery rates followed this pattern: lumbar > cervical and thoracolumbar > thoracic. Thoracic SCI and penetrating SCI were significantly more likely to result in complete injury. Penetrating TSCI had a significantly lower recovery rate compared to blunt injury (OR 0.76, 95% CI 0.62–0.92; p = 0.006). Recovery rate was positively correlated with longer follow-up duration (p = 0.001). Studies with follow-up durations of approximately 6 months or less reported significantly lower recovery rates for incomplete SCI compared to studies with long-term (3–5 years) follow-ups.

CONCLUSIONS

The authors’ meta-analysis provides an overall quantitative description of neurological outcomes associated with TSCI. Moreover, they demonstrated how neurological recovery after TSCI is significantly dependent on injury factors (i.e., severity, level, and mechanism of injury), but is not associated with type of treatment or country of origin. Based on these results, a minimum follow-up of 12 months is recommended for TSCI studies that include patients with neurologically incomplete injury.

In Brief

This article quantitatively describes neurological recovery after traumatic spinal cord injury by using a holistic meta-analytical approach for the first time, and sheds light on the potential role of several injury, treatment, and study factors on patient outcomes. This unprecedented set of prognostic data is of great use for patients, clinicians, rehabilitative teams, researchers, and healthcare administrators.

Traumatic spinal cord injury (TSCI) is a life-altering, devastating condition often associated with significant morbidity, continued disability and dependency, psychological stress, and financial burden. The global prevalence of SCI has been increasing over the last 30 years, and ranges from 236 to 1298 patients per million in different countries.38 It affects more than 280,000 individuals in the US, with an annual incidence of more than 17,000.92 After any medical event and particularly after a TSCI, the affected patients and their families ask for information about the prognosis for recovery. There is a wide spectrum of outcomes that can be expected after TSCI. The heterogeneous nature of the injury characteristics and the abundance, inconsistency, and variability of the studies that have evaluated recovery after TSCI have made it a complex task for clinicians and researchers to determine what outcome to expect for patients with TSCI in general, and particularly with regard to different initial injury variables that may be present.

The value of outcome prediction is not limited to enabling clinicians to provide patients and families with accurate information and guidance. Evidence-based knowledge about the prognosis of TSCI provides physicians with the knowledge to make informed decisions that tailor treatment, follow-up, and rehabilitative plans to the needs of each individual based on his/her prognostic status and expected outcomes. This information is also crucial for clinicians, patients, and rehabilitation teams to shape realistic expectations and set achievable goals for the patients. In addition, evidence-based data on the prognosis of TSCI are necessary as a basis for future research efforts. Researchers need these data in order to direct efforts to the specific patient populations who are more likely to benefit, and to stratify patients into homogeneous groups according to their expected recovery. Moreover, unbiased interpretation of the clinical efficacy of novel therapeutic modalities is not possible unless the course of the disease and current status of recovery from TSCI are accurately established. Prognostic data could also be used by healthcare administrators to guide their expectations, to anticipate the required infrastructure and costs, and to justify decisions to care providers or other financial supporters.

Obviously, this knowledge gap cannot be filled by implementing the results of any single study, and calls for a systematic review of the large body of evidence that exists. Although a number of qualitative review studies have tried to link various factors with “better” or “worse” outcomes,5,58,138 what is not known is an overall quantitative description of the neurological outcomes associated with TSCI, and evidence-based knowledge that shows what proportion of subjects actually recover from the injury and how this proportion is different in various subgroups of patients.

Given this background, a systematic review and meta-analysis of the available evidence regarding neurological recovery after TSCI was performed. Specifically, our objective was to determine for the first time the proportion of patients with TSCI who experience neurological recovery after follow-up and to identify potential predictors of recovery through quantitative pooling of data. To achieve these goals, we sought to determine recovery rates in the general population of patients with TSCI by use of a meta-analytical approach, and also to calculate and statistically compare recovery rates in specific subgroups of these patients according to the severity, level and mechanism of injury, type of treatment, and follow-up duration. We also sought to investigate whether there are any time trends or geographic patterns in neurological recovery after SCI.

Methods

This systematic review and meta-analysis study was conducted in accordance with guidelines recommended in the Cochrane Handbook for Systematic Reviews.43 We adhered to the recommendations outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement70 during reporting of the current study’s findings.

Literature Search

A computerized search of the literature using MEDLINE (via Ovid SP) and EMBASE was performed for human studies in the English language with no date or time restrictions. Using the framework of the PICOT (population, intervention, comparison, outcome, time) format,106 keywords were selected via review of the literature and MeSH and Excerpta Medica Tree (EMTree) terms. The preliminary keyword list was then discussed by a group of authors (a neurosurgeon [V.R.M.] and 2 biostatisticians [M.Y. and M.H.]) with experience publishing numerous systematic review and meta-analysis articles. A complete list of keywords and a detailed search strategy are presented in Supplementary Table 1. The literature search was conducted on February 25, 2017, and was updated on January 26, 2018. Moreover, reference lists of 16 related systematic review articles were manually checked for potential additional results. Finally, the “similar articles” feature of the database was used for all included studies, in order to search for potentially missed relevant studies.

Study Selection

Studies were included if they met the following criteria. 1) Reported baseline and follow-up neurological status of patients with TSCI of any neurological level, severity, and mechanism, using at least 1 of the following measures: American Spinal Injury Association (ASIA) Impairment Scale (AIS), Frankel scale, or ASIA motor score (AMS). 2) Performed the baseline neurological examination within 1 month after the initial injury, to refrain from including patients with SCI who have already experienced a significant portion of their expected recovery. 3) Had a mean follow-up duration of at least 3 months, because the most rapid rate of recovery is observed during the initial 3 months,58,132 and hence the majority of recovery occurs in this period.25,76

The exclusion criteria were as follows. 1) Animal studies, case reports, case series with fewer than 10 eligible patients, editorials, and review articles. 2) Studies focusing on central cord syndrome (CCS), because CCS is known to be associated with an excellent prognosis for neurological recovery.58,85,108 3) Studies focusing on pediatric patients. 4) Studies focusing on patient groups with particular comorbidities (in all of the mentioned instances, what is meant by the term “focusing” is implementation of inclusion/exclusion criteria aimed to preferably and specifically include patients with CCS, pediatric SCI, or comorbidities). 5) Studies that included patients with SCI in whom a nontraumatic etiology was diagnosed. 6) Studies that had reported data in a way that was not suitable for analysis—including studies that had not reported changes in AIS or Frankel scale by grade, and studies that had not reported any measure of dispersion (e.g., SD) for the change in AMS. 7) Studies of clinical trial design.

We excluded clinical trials due to the inherent selection bias and unknown effect of trial medication or technique. This was a meta-analysis of prognosis, and our main aim in this study was to provide a practical description of SCI recovery by pooling studies with outcomes that can be representative of all patients with SCI. However, clinical trials usually seek to investigate the effect of novel experimental treatment modalities, and therefore cannot reflect the status of a general population of patients. Moreover, selection of patients to participate in clinical trials is influenced by ethical issues. For example, critically ill patients are usually excluded from SCI clinical trials because of issues regarding informed consent and randomization to treatment arms. Numerous studies have shown that meta-analysis of clinical trials and observational studies together raises serious statistical concerns that can hamper the reliability of results.43,112

Based on these criteria, each retrieved article was screened for eligibility by 2 independent reviewers at the title and abstract level, and then full-text level if potentially relevant. All reviewers were trained to perform the screening by senior authors after several briefing sessions and standardized pilot searches. When unavailable online, full-text reports were requested from corresponding authors via email. In case of multiple publications on the same cohort of patients, only the article containing the most comprehensive set of data was included. In case of disagreement between the 2 reviewers regarding eligibility of a study, consensus was achieved after reevaluation of the article by a senior author and then, if necessary, the corresponding author.

Data Extraction

Using a prespecified piloted data extraction form, the following information was extracted from each eligible study.

  • 1) Study characteristics: first author, year of publication, PMID or DOI, study design, mean follow-up duration, recruitment period.

  • 2) Demographics: age, sex, country.

  • 3) Neurological level of injury: choices included i) cervical (C1–T1); ii) thoracic (T2–10); iii) thoracolumbar (T11–L2); iv) lumbar or cauda equina (L3 or lower); and v) mixed/not specified.

  • 4) Type of treatment: choices included i) conservative (defined as all nonsurgical treatment approaches including isolated use of corticosteroids); ii) surgical; and iii) mixed/not specified.

  • 5) Mechanism of injury: choices included i) blunt; ii) penetrating; and iii) mixed/not specified.

  • 6) Recovery data: i) in studies that had reported follow-up changes in AIS or Frankel scale scores of the patients from baseline, the following variables were documented into the datasheet. a) Number of patients whose AIS or Frankel grade remained unchanged after follow-up; b) number of patients who experienced at least 1 grade of improvement in their AIS or Frankel scale after follow-up; c) number of patients who experienced at least 2 grades of improvement in their AIS or Frankel scale after follow-up; d) number of patients who improved to AIS or Frankel grade E (cure or full recovery) after follow-up; e) number of patients whose AIS or Frankel grade deteriorated after follow-up; and f) total number of patients. In all studies, these variables were documented both for the total patient population and for patients in each AIS or Frankel grade (A–D) separately. ii) In studies that had used AMS as their outcome measure, the mean (SD) changed in AMS from baseline to follow-up examination, along with the total number of patients who were entered into the datasheet. Moreover, whenever possible, the mean (SD) changes in AMS were extracted and documented for complete/incomplete injury groups and for each AIS or Frankel grade separately.

  • 7) Quality assessment: in order to assess the included studies for methodological quality, articles were classified according to level of evidence (LoE) by using criteria developed by the Oxford Centre for Evidence-Based Medicine17 for prognostic studies and modified by Wright et al.142 These criteria assign an LoE rating (I, II, or III) to each study according to the following qualities. i) Design of the study: prospective cohort studies, retrospective cohort studies, and case series are assigned grades I, II, and III, respectively. ii) Loss to follow-up: in case of incomplete follow-up in more than 20% of the initially included participants, LoE is downgraded by one grade. iii) Controlling for prognostic factors: based on results of the current study and those of others regarding predictors of recovery in SCI, we downgraded the LoE of any study that had not separately reported neurological recovery according to severity and level of injury. iv) Follow-up duration: again, based on results of the current study and those of others, we recognized 12 months as the minimum follow-up that is long enough for outcomes to occur in SCI, and downgraded the LoE of studies with shorter follow-ups. v) Whether patients were in a similar point in the course of their disease: because we had only included patients with recent (< 1-month-old) SCI in our study, all included data pertained to patients who were in a rather similar point in the course of their disease.

Two reviewers independently extracted data from each study in duplicate, and in case of any discrepancy a third senior investigator extracted the data and then discussed the results with reviewers in order to reach consensus.

Statistical Analysis

For the AIS/Frankel outcomes, including the proportion of patients improved by at least 1 grade (conversion rate), the proportion of patients improved by at least 2 grades, and the proportion of patients improved to grade E (full recovery), studies were weighted according to their sample size. We then calculated the pooled prevalence (proportion), obtained as a weighted average, as well as the 95% CI using the score (Wilson) procedure.94 For the AMS, the outcome measure pooled effect estimate and its 95% CI were calculated.

For each meta-analysis, the presence of publication bias was examined through Egger et al.’s precision-weighted linear regression test and symmetry of funnel plots.29 The presence of statistical heterogeneity between studies was assessed using the chi-square test (i.e., Cochrane Q statistic) and I2 statistic.44 A p value of ≤ 0.1 or an I2 statistic > 50% was considered as substantial heterogeneity. The fixed effect model was used for homogeneous meta-analyses and the random effect model using the method of DerSimonian and Laird (as described in Normand)96 was applied for heterogeneous meta-analyses. Meta-analyses were carried out only if the data were reported by at least 3 studies. All statistical analyses were performed using Stata Version 15.0 software (Stata Corp.). A p value < 0.05 was regarded as statistically significant in all analyses.

Subgroup analysis was carried out based on injury severity (AIS/Frankel grades A, B, C, and D); level of injury (cervical, thoracic, thoracolumbar, and lumbar); mechanism of injury (blunt, penetrating); type of treatment (conservative, surgical); country of study (developed, developing); and follow-up duration (very short [≤ 4 months], short [6 ± 1 months], medium [12 ± 2 months], long [24 ± 6 months], and very long [3–5 years]). Countries were classified into either developed or developing groups according to the 2017 International Monetary Fund List of Advanced Economies.46 Studies that specifically pertained to penetrating SCI were excluded from all analyses except the subgroup analysis for mechanism of injury. The AIS/Frankel conversion rate (proportion with at least 1 grade of improvement) was compared between the subgroups by using meta-regression analysis of categorical covariates, and ORs along with 95% CIs were calculated from 2 × 2 tables and reported. The potential correlation between follow-up duration or recruitment period of study and AIS/Frankel conversion rate was evaluated using linear meta-regression analysis of continuous covariates, and the β coefficient along with the 95% CI was calculated and reported.

A multivariate meta-regression analysis using the covariates that had shown significant association with recovery was performed, and the proportion of heterogeneity explained by covariates was expressed as the adjusted R2 statistic. We planned a sensitivity analysis to test the impact of the quality of included studies by categorizing studies into either high-quality (LoE I or II) or low-quality (LoE III) groups and performing meta-regression analysis to compare recovery outcomes of patients in these 2 groups.

Results

A total of 16,652 unique citations were identified, and 582 potentially relevant articles were selected for full-text review and were subsequently obtained and assessed for eligibility. Finally, a total of 114 individual articles met all eligibility criteria and were selected for data extraction and meta-analysis. Figure 1 is a flow diagram describing the stepwise study selection process performed according to the PRISMA guidelines.

FIG. 1.
FIG. 1.

Flow diagram of study identification and selection process according to PRISMA statement. Figure is available in color online only.

Study Characteristics

The selected studies included 38 prospective cohort studies, 67 retrospective cohort studies, and 9 case series.1–3,6–13,16,19–22,24–28,30,31,33–37,39–42,45,47–52,54,56,59–64,66–69,71–76,78–84,86–91,93,95,97–105,107,109–111,113–127,130–132,135–137,139,140,143,144,146–149 In total, 99 studies had reported neurological recovery using AIS or Frankel scales, and 26 had used the AMS; 11 of the studies had reported both. Overall, the included studies reported follow-up changes in AIS or Frankel scale for 19,913 patients, and changes in AMS for 6920 patients. Studies were published between 1969 and 2017 and recruited patients from 29 countries across 5 continents. The median follow-up duration of included studies was 12 months (interquartile range 6–23.4 months). The mean age of study participants ranged from 22 to 63 years, with a weighted mean of 35.4 years. The proportion of male patients ranged from 52% to 100% among included studies, with an overall proportion of 80.0%. Supplementary Table 2 presents the characteristics of all individual studies included in our analysis.

Risk of Bias Assessment

The overall methodological quality of the included studies was poor, with the majority of studies (65%) presenting level III evidence (Supplementary Table 2). Of the 114 included studies, our quality assessment identified only 7 level I (6%) and 33 level II (29%) studies. The main sources of bias included the following (Supplementary Fig. 1). 1) Poor study design: 76/114 (66%) studies were retrospective cohorts or case series. 2) Not accounting for possible prognostic factors: as mentioned earlier, studies that had not reported AIS/Frankel changes according to baseline severity of injury were not included in our study. In regard to AMS outcome measure, only 10 of 26 included studies had reported changes of AMS according to baseline completeness of injury. Furthermore, among all included studies, 27 (23%) had not clarified the level of injury or had not reported recovery of subgroups with different levels of injury separately. 3) Insufficient follow-up duration: 52 (45%) studies had followed patients for less than 12 months. 4) Loss to follow-up: 30 studies (26%) had reported loss to follow-up rates of more than 20%.

We took several measures to identify any potential effect that these sources of bias could have had on our results, and to reduce or omit their impact if possible. As will be presented in the following sections, we have calculated and reported recovery rates both in the total cohort and in different subgroups based on severity, level, and mechanism of injury, as well as treatment method and follow-up duration, aiming to mitigate allocation bias and identify sources of heterogeneity among results. Moreover, we categorized studies reporting AIS/Frankel changes into high-quality (LoE I or II; 37 studies) and low-quality (LoE III; 57 studies) groups and performed a sensitivity analysis aiming to determine whether there is any difference in the reported AIS/Frankel conversion rates between high- and low-quality studies. The results showed no significant difference between the 2 groups in the recovery of patients with either complete or incomplete SCI.

We examined the presence of small-study effects and publication bias by evaluating funnel plots for symmetry both visually and statistically. Funnel plots for both AIS/Frankel and AMS outcome measures in various subgroups are presented in Supplementary Fig. 2. In all funnel plots, studies were symmetrically distributed on both sides of the pooled outcome proportion line. Moreover, results of the Egger precision-weighted linear regression test returned no evidence of small-study effects.

Recovery and Severity of Injury

A total of 94 studies reported changes in AIS or Frankel scale in 19,460 patients with TSCI by their baseline severity of injury (Table 1). The included population comprised 9685 (49.8%) patients with grade A; 3090 (15.9%) patients with grade B; 4796 (24.6%) patients with grade C; and 1889 (9.7%) patients with grade D injury. Random pooled effect analysis showed that 19.3% (95% CI 16.2%–22.6%) of patients with AIS/Frankel A injuries experienced at least 1 grade of improvement. This rate was 73.8% (95% CI 69.0%–78.4%) for grade B, 87.3% (95% CI 77.9%–94.8%) for grade C, and 46.5% (95% CI 38.2%–54.9%) for grade D SCI (Supplementary Figs. 3–6). The results also demonstrate that full recovery was extremely unlikely for patients with grade A and B SCI, whereas 9.2% (95% CI 6.4%–12.4%) of patients with grade C and 46.5% (95% CI 38.2%–54.9%) of patients with grade D experienced full recovery. In patients with grade A and C, 9.2% gained at least 2 AIS/Frankel grades, in comparison with 40.5% (95% CI 32.5%–48.5%) of grade B subjects. Meta-regression analysis revealed that severity of injury significantly correlates with neurological recovery (p < 0.001, R2 = 68.1%). As presented in Table 2, the results of meta-regression analyses indicate that neurological recovery is significantly different between all grades of SCI severity in the following order: C > B > D > A.

TABLE 1.

Summary of meta-analysis results describing changes in AIS/Frankel scale scores in different severity and level of injury subgroups of SCI

% w/ ≥1 Grade Improvement% w/ ≥2 Grades Improvement% Gaining Full Recovery
Target Pt GroupNo. of PtsEffect Size95% CII2Effect Size95% CII2Effect Size95% CII2
Total cohort (94 studies)
 Grade A968519.316.2–22.683.5%*9.27.1–11.576.5%*0.00.0–0.024.1%*
 Grade B309073.869.0–78.472.1%*40.432.5–48.590.5%*0.30.0–1.561.0%*
 Grade C479687.377.9–94.897.0%*9.26.4–12.474.6%*9.26.4–12.474.6%*
 Grade D188946.538.2–54.989.7%*NANANA46.538.2–54.989.7%*
 Total19,46049.444.8–53.996.4%*16.814.5–19.287.2%*8.36.3–10.491.5%*
Cervical SCI (49 studies)
 Grade A173423.217.8–28.978.8%*8.35.3–11.968.0%*0.00.0–0.00.0%
 Grade B101275.267.9–82.076.0%*39.329.4–49.586.3%*1.00.0–3.455.3%*
 Grade C112087.682.9–91.866.6%*9.35.3–14.070.8%*9.35.3–14.070.8%*
 Grade D96144.434.0–55.188.2%*NANANA44.434.0–55.188.2%*
 Total482753.448.4–58.490.8%*16.913.2–20.984.3%*8.75.7–12.088.5%*
Thoracic SCI (13 studies)
 Grade A55510.05.6–15.143.6%4.82.7–7.20.0%0.00.0–0.00.0%
 Grade B4568.249.3–85.08.5%44.622.9–67.231.0%10.40.8–25.21.6%
 Grade C2187.558.9–10023.2%27.15.6–53.91.1%27.15.6–53.91.1%
 Grade D2361.321.3–95.351.8%NANANA61.321.3–95.351.8%
 Total64421.715.2–28.962.8%*10.23.8–18.364.3%*3.70.0–11.376%*
Thoracolumbar SCI (10 studies)
 Grade A28315.98.0–25.448.8%*7.02.0–13.635.5%0.00.0–0.00.0%
 Grade B6377.859.3–92.921.8%61.436.6–84.147.0%6.30.1–17.70.0%
 Grade C7894.985.7–99.90.0%14.43.1–29.525.2%14.43.1–29.525.2%
 Grade D6155.832.3–78.344.5%NANANA55.832.3–78.344.5%
 Total48545.832.9–59.181.9%*20.910.0–33.675.5%*8.91.8–18.876.2%*
Lumbar SCI (6 studies)
 Grade A2835.313.6–59.60.0%27.47.7–51.30.0%0.00.0–0.00.0%
 Grade B1993.557.6–100.058.2%45.620.4–71.80.0%1.20.0–17.90.0%
 Grade C3590.774.1–99.90.0%11.60.6–29.00.0%11.60.6–29.00.0%
 Grade D4094.868.4–10055.3%NANANA94.868.4–10055.3%
 Total12277.758.3–93.168.9%*24.312.9–37.10.0%17.10.5–43.482.4%*

NA = not applicable; pt = patient.

p < 0.1, chi-square test for heterogeneity.

TABLE 2.

Comparison of AIS/Frankel conversion rates and ASIA motor score changes between different subgroups of injury severity

AIS/Frankel Conversion RateAMS Change
Baseline AIS/Frankel GradeOR95% CIp Valueβ Coeff95% CIp Value
B vs A1.541.45–1.64<0.00118.29.9–26.4<0.001
C vs A1.761.63–1.89<0.00127.516.4–38.5<0.001
D vs A1.231.14–1.32<0.0018.01.4–14.70.021
C vs B1.121.03–1.220.0079.3−5.2 to 23.80.18
B vs D1.261.15–1.38<0.00110.40.04–20.80.049
C vs D1.421.30–1.56<0.00120.15.8–34.40.01

coeff = coefficient.

A total of 26 studies reported changes in AMS in 6920 patients with TSCI. Six studies (n = 1674) separately reported AMS changes for different baseline AIS grades of injury, and 10 studies (n = 5341) separately reported AMS changes for patients with complete and incomplete SCI. As presented in Supplementary Table 3, meta-analysis indicates that patients with neurologically complete and incomplete SCI experience a mean change of 6.6 (4.7–8.5) and 24.0 (20.3–27.8) points in their AMS, respectively. Among patients with incomplete SCI, this change is 25.6 (19.7–31.5) for grade B, 36.5 (25.1–47.8) for grade C, and 14.7 (8.5–20.9) for grade D (Supplementary Fig. 7). Results of the meta-regression analysis were similar to those of the AIS/Frankel outcome measure and are summarized in Table 2.

Recovery and Level of Injury

A total of 78 studies reported changes in AIS/Frankel scale scores for 6078 patients with TSCI by their neurological level of injury. The results of the random pooled effect analysis for each level are presented in Table 1. The proportion of patients with neurologically complete injury was significantly higher in thoracic SCI compared to all other levels of injury (86% vs 35% cervical, 58% thoracolumbar, and 22% lumbar; p < 0.01 for all). Thoracic SCI was associated with the lowest recovery rates (3.7%, 10.2%, and 21.7% for full recovery, 2 grades of improvement, and 1 grade of improvement, respectively), whereas lumbar SCI showed the highest recovery rates (17.1%, 24.3%, and 77.7%). Cervical and thoracolumbar levels of injury had similar intermediate recovery rates (8.7%, 16.9%, and 53.4% for cervical and 8.9%, 20.9%, and 45.8% for thoracolumbar SCI) (Supplementary Fig. 8). When comparing recovery rates of different levels of injury within each baseline severity grade, the previously reported pattern can still be detected in complete SCI; complete thoracic SCI is associated with the lowest (10%, 95% CI 5.6–15.1) and complete lumbar SCI is associated with the highest (35.3%, 95% CI 13.6–59.6) rate of neurological recovery. However, incomplete SCI shows more variable recovery rates among different neurological levels of injury (Table 1 and Fig. 2).

FIG. 2.
FIG. 2.

Summary of meta-analysis results depicting proportion of patients whose AIS or Frankel scoring improved by at least 1 grade, in different severity and level of injury subgroups of SCI. Figure is available in color online only.

Our meta-regression analysis revealed that level of injury is a statistically significant predictor of neurological recovery (p < 0.001, R2 = 42%, residual I2 = 48%). Lumbar SCI had a significantly superior prognosis for recovery compared to all other levels of injury, whereas the prognosis of thoracic SCI was significantly worse than that for all other levels (Table 3). Cervical and thoracolumbar SCI were not significantly different with regard to neurological recovery. When the analysis was restricted to complete SCI, meta-regression analysis showed that complete thoracic SCI has a significantly inferior recovery rate compared to cervical complete SCI (OR 0.88; p = 0.013), whereas all other between-level differences lost statistical significance (Table 3). When the analysis was restricted to incomplete SCI, meta-regression analysis indicated that there are no significant differences in recovery among different levels of injury within incomplete SCI, although there was a trend for superior recovery in the lumbar region compared to cervical SCI (OR 1.23; p = 0.08) (Table 3).

TABLE 3.

Comparison of AIS/Frankel conversion rates between different subgroups based on level and severity of injury

TotalComplete InjuryIncomplete Injury
Level of InjuryOR95% CIp ValueOR95% CIp ValueOR95% CIp Value
Thoracic vs cervical0.740.65–0.84<0.0010.880.80–0.970.0130.980.78–1.230.87
Thoracolumbar vs cervical0.920.80–1.070.320.930.81–1.060.281.060.89–1.270.44
Lumbar vs cervical1.251.01–1.560.0431.120.77–1.620.521.230.97–1.560.08
Thoracic vs thoracolumbar0.800.71–0.910.0020.940.81–1.100.460.960.74–1.240.76
Lumbar vs thoracic1.721.40–2.12<0.0011.270.86–1.880.211.220.89–1.670.19
Lumbar vs thoracolumbar1.361.07–1.740.0141.260.81–1.940.261.170.90–1.530.21

A total of 20 studies reported changes in AMS by level of injury (cervical or tetraplegia vs noncervical or paraplegia) in 2986 patients with TSCI (Supplementary Fig. 9). Ten studies (n = 2013) separately reported AMS changes for patients with complete and incomplete SCI in cervical and noncervical groups. Six studies (n = 1463) separately reported AMS changes for different baseline AIS grades of injury in cervical SCI. Only 2 studies had reported recovery of different baseline AIS grades in noncervical levels of injury; therefore a meta-analysis was not performed. As presented in Supplementary Table 3, patients with cervical and noncervical SCI experienced a mean AMS change of 20.3 (95% CI 14.9–25.8) and 7.0 (95% CI 5.3–8.8), respectively—a difference that was statistically significant (β coefficient = 13.1 [95% CI 7.7–18.5]; p < 0.001). Within the subgroup of patients with complete SCI, a cervical level of injury was again associated with significantly higher changes in AMS compared to a noncervical level (8.7 vs 1.2; p = 0.039). This was also the case when comparing incomplete cervical and noncervical SCI (26.6 vs 12.8; p = 0.039) (Supplementary Fig. 10).

Recovery and Mechanism of Injury

A total of 5 studies reporting AIS/Frankel conversion status of 453 patients with penetrating SCI were included in our analysis (Supplementary Fig. 11). The included participants comprised 321 patients with grade A, 35 with grade B, 70 with grade C, and 27 with grade D penetrating SCI. The prevalence of complete injury was significantly higher among patients with penetrating SCI compared to blunt injury (OR 1.42 [95% CI 1.09–1.84]; p = 0.008). Random pooled effect analysis revealed that 20.8% (95% CI 1.6%–51.1%) of patients with penetrating SCI experience at least 1 grade of improvement. This rate was 6.1% for patients with grade A, 73.0% for patients with grade B, 51.8% for those with grade C, and 53.4% for those with grade D. The odds of remarkable (≥ 2 grades) and full recovery were 5.0% (95% CI 0.0%–22.6%) and 5.0% (95% CI 0.0%–19.4%), respectively, for patients with penetrating SCI.

We compared recovery rates for studies of patients with penetrating SCI with those of patients with blunt SCI, and our meta-regression analysis demonstrated that penetrating SCI is associated with a significantly worse prognosis compared to blunt injury (OR 0.76 [95% CI 0.62–0.92]; p = 0.006). Unfortunately, our analysis was not powered enough to detect differences in recovery within complete SCI groups, because the significance test for effect size = 0 returned a p value of 0.18. However, we still detected a trend for inferior outcomes in complete penetrating SCI (AIS/Frankel conversion rate of 6.1% vs 19.3%; OR 0.86 [95% CI 0.73–1.02]; p = 0.08). No significant difference was found in the conversion rate of incomplete SCI between penetrating and blunt injury mechanisms (57.1% vs 68.1%; OR 0.87 [95% CI 0.71–1.07]; p = 0.20). Our multivariate meta-regression analysis confirmed that when adjusted for severity of injury, the mechanism of injury maintains its significant correlation with neurological recovery (p = 0.012).

Recovery and Method of Treatment

A total of 56 studies reported changes in the AIS/Frankel scale scores of 4372 patients with TSCI by the type of treatment they had received. Fifteen studies (n = 1472) had reported on conservatively treated patients and 47 studies (n = 2900) reported results of surgical treatment (Supplementary Fig. 12). The prevalence of complete SCI was higher in the conservatively treated group compared to the surgically treated group (53% vs 31%; p = 0.17). The recovery rates pertaining to each of the treatment groups are described in Table 4. In order to decipher the potential effect of treatment method on neurological recovery, we performed a series of meta-regression analyses comparing recovery of different treatment groups both in their total cohorts and in different subgroups based on the level and severity of injury (Supplementary Table 4). The results show that there is no significant difference in neurological recovery between conservative and surgical treatment, regardless of whether the analysis is performed within complete or incomplete injury and cervical or noncervical subgroups.

TABLE 4.

Summary of meta-analysis results describing changes in AIS/Frankel scale in different subgroups based on type of treatment and severity of injury

% w/ ≥1 Grade Improvement% w/ ≥2 Grades Improvement% Gaining Full Recovery
Target Pt GroupNo. of PtsEffect Size95% CII2Effect Size95% CII2Effect Size95% CII2
Surgical treatment (47 studies)
 Grade A92018.412.8–24.770.2%*7.24.3–10.641.5%*0.00.0–0.00.0%
 Grade B58980.771.6–88.768.3%*42.429.9–55.381.4%*0.00.0–0.838.3%*
 Grade C76289.584.2–94.159.3%*12.26.6–18.670.3%*12.26.6–18.670.3%*
 Grade D62956.546.5–66.376.1%*NANANA56.546.5–66.376.1%*
 Total290054.048.5–59.484.5%*17.413.1–22.076.9%*14.610.8–18.886.5%*
Conservative treatment (15 studies)
 Grade A78420.610.2–32.988.3%*11.64.3–21.084.8%*0.00.0–0.00.0%
 Grade B25276.268.6–83.119.3%56.246.1–66.040.5%1.90.1–5.124.7%
 Grade C25294.285.4–99.656.6%*6.01.9–11.410.7%6.01.9–11.410.7%
 Grade D18448.032.1–64.056.6%*NANANA48.032.1–64.056.6%*
 Total147248.937.6–60.292.9%*21.613.4–30.787.6%*6.93.7–10.775.3%*

p < 0.1, chi-square test for heterogeneity.

Recovery and Time and Place of Study

Based on their country of origin, all studies were categorized into 2 groups, i.e., developed countries and developing countries. A total of 18 studies had reported AIS/Frankel conversion rates of 1710 patients with TSCI from developing countries including Chile, India, Croatia, Pakistan, China, Iran, Turkey, Nigeria, and South Africa. The prevalence of complete injury was not different between developing and developed countries (38.9% vs 39.9%; p = 0.55). Random pooled effect analysis showed that in developing countries 48.5% (95% CI 38.2%–59%) of patients with SCI experienced at least 1 grade of improvement, compared to 49.5% (95% CI 44.7%–54.3%) of patients in developed nations (p = 0.83) (Supplementary Fig. 13). This rate was also not significantly different when only comparing specific severity and level of injury subgroups between developing and developed countries.

We also examined the effect of the year of study on recovery rate; the aim was to identify any potential time trends in neurological recovery of SCI. The midpoint in the recruitment period of the studies, which better reflects when the study was actually conducted than does the publication year, was used for this analysis; it ranged from 1960 to 2012. Linear meta-regression analysis revealed no significant correlation between recruitment period of studies and AIS/Frankel conversion rates (β coefficient = 0.001 [95% CI −0.002 to 0.004]; p = 0.52) (Fig. 3). Moreover, no correlation was found when the analysis was restricted to specific severity and level of injury subgroups.

FIG. 3.
FIG. 3.

Linear meta-regression analysis of the correlation between study recruitment period and AIS/Frankel conversion rate. ES = effect size. Figure is available in color online only.

Recovery and Follow-Up Duration

Four studies had a follow-up duration of more than 5 years (66, 68, 88, and 141 months); these were not included in this section of our review so as to reduce the effect of outlier data, and also because extremely long follow-up of patients is not feasible in most clinical and research settings. Meta-regression analysis revealed that longer follow-up duration is significantly associated with a higher AIS/Frankel conversion rate (β coefficient = 0.005 [95% CI 0.002–0.008]; p = 0.001) (Fig. 4A). In the subgroup with complete SCI no such association was detected (p = 0.71) (Fig. 4B); however, the significant correlation was maintained in the subgroup with incomplete SCI (β coefficient = 0.003 [95% CI 0.0003–0.007]; p = 0.03) (Fig. 4C). Within complete cervical, complete noncervical, incomplete cervical, and incomplete noncervical subgroups, only the incomplete cervical subgroup demonstrated a significant association between follow-up duration and recovery rate (p = 0.048).

FIG. 4.
FIG. 4.

Linear meta-regression analysis of the correlation between study follow-up duration in months and AIS/Frankel conversion rate in total population (A), patients with complete injury (B), and those with incomplete injury (C). ES = effect size. Figure is available in color online only.

Aiming to compare outcomes of commonly used specific follow-up durations, we categorized the studies into 5 groups based on their follow-up duration: very short (≤ 4 months; 9 studies), short (6 ± 1 months; 17 studies), medium (12 ± 2 months; 23 studies), long (24 ± 6 months; 11 studies), and very long (3–5 years; 15 studies). We performed meta-regression analysis to compare the recovery rates of different groups (Supplementary Fig. 14). In the total cohort, between any two given groups, the recovery rate was higher in the group with longer follow-up duration. Comparison of recovery rates of each group with the “very long” follow-up group showed that recovery rates were significantly lower in the “very short,” “short,” and “medium” follow-up groups (p = 0.002, 0.003, and 0.031, respectively). Consistent with our previous findings, a subgroup meta-regression analysis demonstrated that there was no difference in recovery among various follow-up groups within subgroups of patients with complete SCI or noncervical SCI. However, a trend toward inferior outcome in the “very short” and “short” follow-up groups compared to the “very long” group was still observed within incomplete SCI (p = 0.07 and 0.008, respectively) and incomplete cervical SCI (p = 0.08 and 0.06, respectively) subgroups (Supplementary Table 5).

Discussion

The current knowledge regarding the prognosis of patients with TSCI is merely based on a few individual cohort studies published from North American or European SCI registries. To our knowledge, this systematic review is the first study that describes neurological recovery after TSCI by using a meta-analytical approach. Our meta-regression analyses indicate that motor recovery after TSCI is significantly dependent on injury factors (i.e., severity, level, and mechanism of injury), but is not associated with type of treatment or country of origin, and has not significantly changed over time in the last decades.

As mentioned earlier, this is the first systematic review study to have quantitatively investigated motor recovery after SCI via pooling of data. A few systematic review articles have addressed this issue by using a qualitative approach. In their systematic review of 25 articles, Wilson et al.138 concluded that baseline severity of injury is the most important predictor of neurological outcome after TSCI. Moreover, they suggest that complete thoracic SCI is associated with worse prognosis compared to complete cervical SCI; however, the same cannot be said with regard to incomplete injury. In another systematic review, Al-Habib et al.5 concluded that neurological recovery clearly correlates with severity of injury. They also argue that complete cervical SCI is shown to have better outcomes compared to complete thoracic SCI in several studies; however, this finding “could be related to methodological and sampling issues.”

Our meta-analysis shows that 19.3% (95% CI 16.2%–22.6%) of patients with complete (grade A) SCI experience at least 1 grade of neurological improvement and convert to incomplete status. Many authors have implied that the recovery reported for patients with complete SCI can partially result from misclassification of truly incomplete SCI as complete on presentation, due to inaccurate initial neurological examination in SCI studies. Many factors, including concurrent traumatic brain injury, intoxication, chemical sedation or paralysis, mechanical ventilation, spinal shock, hemodynamic instability, severe pain, confusion, and psychological stress, can interfere with the reliability of the initial neurological examination, especially if it is done in the early hours or days after injury,15 thus leading to the reporting of high conversion rates for patients with SCI incorrectly designated as complete.14,83 However, the fact that patients who have complete SCI several months after injury still could experience neurological recovery27 proves that complete SCI is not inherently unchangeable. Nevertheless, there is no doubt that inconsistent timing of the baseline neurological assessment is one of the main sources of heterogeneity among the recovery rates of patients with complete injury reported by different studies.

Injury severity is historically considered the most important predictor of recovery in SCI. Our meta-regression analyses indicated that neurological recovery is significantly different between all grades of SCI severity in the following order: C > B > D > A. In other words, the odds of recovery, determined by either AIS/Frankel conversion rate or mean changes in AMS, increase significantly as the injury severity diminishes (Table 2), except for patients with AIS D, who experience significantly less improvement compared with those categorized as AIS B or C. The relative lack of recovery in the AIS D population is most likely due to a ceiling effect in AIS, Frankel, and AMS systems. Incomplete SCI is associated with a favorable prognosis for functionally meaningful recovery, because 40.4% and 87.3% of patients with AIS/Frankel grade B and C, respectively, will improve to grades D or E, while experiencing 25.6- and 36.5-point improvements, respectively, in their AMS. Moreover, although the chance of full recovery (conversion to grade E) is virtually nonexistent for patients with motor complete SCI (0% for patients with grade A and 0.3% for patients with grade B), patients with grade C and D attain full recovery in 9.2% and 46.5% of cases, respectively.

There are several explanations for worse prognosis in patients with complete thoracic SCI. First, the thoracic spine is well protected by the rib cage and chest wall muscles, so any trauma that can inflict a complete injury to the thoracic spinal cord is likely to be of devastatingly high energy, thus resulting in more severe destruction of the neural tissue. Second, the spinal canal is relatively narrow in the thoracic region, leaving little space for the cord in case of compressive external forces. Third, because thoracic spinal cord is located in the watershed area of spinal cord circulation, it is more susceptible to shortages of blood supply during injury because of the vascular steal phenomenon, especially considering the fact that patients with thoracic SCI have more concurrent cardiac and pulmonary injuries.5,128

Lumbar injuries were associated with significantly better prognosis compared to cervical, thoracic, and thoracolumbar injuries. This is probably due to the presence of nerve roots instead of spinal cord in the cauda equina. Lower motor neurons or peripheral nerve roots, which are known to have an enhanced ability of self-repair and nerve sprouting after injury,77 are more abundant in this region, thus providing the possibility for “root escape” phenomenon. Interestingly, our subgroup meta-regression analysis indicated that the superior prognosis of lumbar SCI is confined to patients with incomplete injury. This could be explained by the limited number of patients with complete lumbar SCI in our study (n = 28).

The mean improvement in AMS was 20.3 (complete = 8.7; incomplete = 26.6) for cervical and 7.0 (complete = 0.9; incomplete = 15.4) for thoracic/lumbar SCI (p < 0.001). However, when interpreting these results it should be noted that there is less room for improvement of AMS in paraplegia because upper extremities are already intact, hence creating a ceiling effect. Moreover, we are unable to manually test for motor recovery of T2–12 myotomes because they do not contain any key muscle groups. Any meaningful motor recovery would probably require a long-distance neural circuitry reconstruction to connect lower motor neurons to the lumbar spine.

Our meta-analysis indicates that there is no significant difference in neurological recovery between surgically and conservatively treated patients, neither in the total population nor within complete/incomplete subgroups or cervical/noncervical subgroups. In addition, the timing of the surgical decompression might be essential for the improved outcome for SCI treatment. Recent data have suggested that rapid decompression and stabilization within 24 hours or even 8 hours might provide a better chance for neurological function improvement.33,49

When interpreting the current findings, the possibility of selection bias in the reviewed studies should be kept in mind because surgical treatment was usually reserved for patients with progressive neurological deficit and persisting cord compression secondary to expanding hematomas or dislocation of bony fragments. In this subgroup of patients, surgery might be maintaining a certain extent of neurological recovery that would have been lost in case of nonsurgical management. However, in many of the other included studies, surgical treatment was performed in a broader subset of patients. On the other hand, we found that patients with surgically managed SCI are more likely to have incomplete injury, indicating an inclination to operate more expeditiously on patients with incomplete SCI. This is in line with results of an international survey of more than 900 spine surgeons, which found that the percentage of respondents who would rush a patient with incomplete SCI to the operating room for surgical decompression was much higher than the percentage who would do so for an otherwise identical patient with complete SCI (72.9% vs 46.2%).32 In summary, management principles and clinical sentiments that were already in effect at each study center must have influenced the selection of patients for conservative or surgical treatment. No definite conclusions can be drawn until randomized prospective controlled trials compare the efficacy of surgical and conservative treatment for specific indications in specific patient subgroups. To the best of our knowledge, such studies are currently lacking.

Although surgical intervention has become the standard of practice in patients with acute SCI, there is no credible evidence to support the superiority of surgical interventions compared to conservative management. Therefore, for now, there is no choice but to reevaluate the existing evidence. Now, when it comes to comparing these two treatment approaches, the relatively poor quality of the existing evidence does cause some limitations, as is mentioned above. However, at the very least the significance of these results, which provide the strongest evidence that we currently can hope for, could be to alarm us that routine use of surgical intervention might not always be beneficial in terms of neurological recovery, and to draw attention to the fact that well-designed, large-scale trials are definitely needed to address this.

To the best of our knowledge, this is the first study to have compared neurological outcome of patients with blunt and penetrating SCI by using a meta-analytical approach. We have demonstrated that penetrating injuries are associated with a significantly worse chance for recovery (20.8% vs 51.6%; p = 0.006). Due to the limited number of eligible studies of penetrating SCI, our study was not powered to detect differences between blunt and penetrating injury within complete SCI subgroups; however, the results show a trend for worse prognosis of patients with complete penetrating SCI compared to those with complete blunt SCI (6.1% vs 19.3%; p = 0.08), whereas there was no such trend in incomplete injury. Our results also suggest that penetrating injuries are more likely to result in complete SCI compared to blunt injuries (OR 1.42 [95% CI 1.09–1.84]; p = 0.008), which is in line with the findings of Marino et al.76 and Waters et al.134 All included studies of penetrating SCI involved gunshot or war missile injuries, because no stab wound study met our eligibility criteria.

This study also noted that neurological recovery of patients with SCI is not different between developing and developed nations, and that recovery has not significantly changed over time. These findings were logically expected because—as was discussed—we also found that the treatment method does not affect neurological recovery. Such results collectively indicate that neurological recovery after SCI is merely a function of injury variables and does not correlate with treatment-related factors. Other reasons might have also contributed to the acquisition of such results. Efficient emergency healthcare is less accessible in developing countries, and therefore the most severely injured patients with SCI, who would have had a poor chance for neurological recovery, are at a higher risk of death in the early hours after the incident.18

Our results also show that neurological recovery has not significantly improved in the last few decades despite all the diagnostic and therapeutic advances in medicine. It is generally recognized that there is as yet no safe and effective treatment capable of modifying the course of disease in SCI. Our results further underline this notion and raise the question as to whether steroids and surgical intervention have provided any true benefit in the management of SCI. Several novel neuroprotective agents and regenerative cell-based approaches are currently under development for treatment of SCI, some of which are showing encouraging preliminary results.4,53,55 Whether these novel interventions would enter common clinical practice and break new ground in the management of SCI remains to be seen.

Selecting an optimal follow-up duration has always been a challenge in the design of clinical trials and cohort studies of SCI. There is no consensus regarding the minimum follow-up duration that is long enough for neurological recovery to reach a plateau. Some authors perform their final neurological examination upon the patient’s discharge from the acute care hospital, which is typically 2–3 months postinjury, and many choose follow-up durations of either 6 or 12 months, on the basis of common practice or the results of a few individual studies that place the recovery plateau between 6 and 12 months postinjury.23,57,133 However, other individual studies exist that have reported remarkable conversion rates, even up to 35%, after the 1st year.103,118 Our results indicate that follow-up duration did significantly and positively correlate with the extent of reported neurological recovery in the total SCI population, in patients with incomplete SCI, and in those with cervical SCI, but it was not a significant factor in complete or noncervical SCI. In fact, complete and paraplegic patients seem to reach neurological stability sooner in the course of disease. Our subgroup meta-regression analysis revealed that studies that had followed patients with SCI for approximately 6 months or less reported significantly lower recovery rates for patients with incomplete injury compared to studies with long-term follow-up. However, approximately 12 months of follow-up did not result in the reporting of significantly different outcomes in incomplete SCI compared to long-term follow-up. In summary, we recommend a minimum follow-up of 12 months for all TSCI studies that include patients with neurologically incomplete injury.

This study is subject to various limitations and sources of bias. First, this was a meta-analysis of observational studies, and therefore the quality of evidence for the included studies was generally low, with the majority of studies presenting level II or III evidence. We took several measures to identify and minimize the impact of poor-quality data on our results, which are described in the Risk of Bias Assessment section. Another limitation is that we had to focus on motor recovery as the outcome, because studies addressing sensory recovery following SCI were limited and inconsistent in their reporting of data. Very few studies included separately reported recovery data for different age subgroups, and therefore we were unable to examine the effect of age on recovery.

Another limitation of this study was the fact that it was not equipped to investigate the effect of timing of surgery on neurological recovery. Our analyses showed that there is no significant difference in neurological recovery between conservative management and early surgery groups, or between early and late surgery groups, regardless of whether or not the analysis is performed within complete or incomplete injury and cervical or noncervical subgroups (data not shown). Several interventional studies—including randomized controlled trials—have been published that compare outcomes of early versus late surgery. There have also been several meta-analysis articles of interventional studies, which are specifically designed for this purpose and thoroughly address this issue.65,129,141,145 However, interventional studies were excluded from our study for the reasons explained earlier. Therefore, our analysis could not possibly address this issue comprehensively and accurately, which is why we refrained from commenting on this matter.

Because we found that follow-up duration can significantly affect recovery rate within the incomplete cervical SCI group, we performed multivariate meta-regression analyses adjusting for follow-up duration, in order to decipher and omit any potential confounding effect that differences in the mean follow-up duration of subgroups might have had on our results. Multivariate analysis showed that the significant difference in neurological recovery between complete and incomplete cervical SCI, as well as the lack of significant differences among various levels of injury in incomplete SCI, persists after adjusting for the effect of follow-up duration.

Another potential challenge is the heterogeneity detected in some of our reported pooled outcomes. We had set strict eligibility criteria aiming to maximize homogeneity of the included studies. Moreover, we were able to identify 4 sources for the heterogeneity among recovery outcomes through our subgroup meta-regression analyses, namely severity of injury, level of injury, mechanism of injury, and follow-up duration. A multivariate meta-regression analysis revealed that these 4 variables were able to explain 84.1% of between-study variations (R2 = 84.1%), with a residual variation due to heterogeneity of 0.0% (residual I2 = 0.0%). We hypothesize that variation in both timing and accuracy of the initial neurological examination is another potential source of heterogeneity between the results of different individual studies; however, our study was not equipped to test this hypothesis.

Conclusions

Herein, we quantitatively described neurological recovery following TSCI in different severity, level, and mechanism of injury as well as treatment subgroups by using a meta-analytical approach for the first time. Moreover, through subgroup meta-regression analyses we were able to statistically examine the potential prognostic significance of severity of injury, level of injury, mechanism of injury, treatment method, and follow-up duration—and also to investigate whether there are any time trends or geographic patterns in neurological recovery after SCI—the results of which are summarized in Table 5. It is our sincere hope that these results can help clinicians and rehabilitation teams in providing evidence-based prognostic information for affected patients and families, and facilitate the devising of realistic treatment and rehabilitation plans and goals tailored for each individual’s profile for recovery. Our results can also benefit SCI researchers in homogeneous selection and stratification of patients in clinical trials, and establish a norm for recovery, against which the efficacy of novel therapeutic interventions should be ascertained. We hope our holistic yet differentiated approach toward SCI recovery yields new insights into the course of disease and helps optimize the efforts and guide expectations of physicians, researchers, and healthcare administrators.

TABLE 5.

Summary of main findings of the current study

Conclusions From Our Systematic Review & Meta-AnalysisRelated Results
Pooled estimates of partial & full recovery rates in different subgroups of pts w/ TSCI according to severity, level, & mechanism of injury, & also type of treatment.Tables 1 & 4; Supplementary Table 3; Fig. 2; Supplementary Figs. 10 & 11
Neurological recovery is significantly different btwn all grades of SCI severity in the following order: C > B > D > A.Table 2
Level of injury is a significant predictor of recovery in TSCI; recovery rates follow this pattern: lumbar > cervical & thoracolumbar > thoracic.Tables 1 & 3; Supplementary Table 3; Fig. 2; Supplementary Fig. 10
Thoracic SCI is more likely to result in complete injury compared w/ other regions.86% vs 35% cervical, 58% thoracolumbar, & 22% lumbar; p < 0.01 for all
Complete thoracic TSCI is associated w/ significantly worse prognosis for neurological recovery compared to complete cervical TSCI. Such a difference is not observed in incomplete injury.10% vs 23.2%; OR 0.88; p = 0.01 (Table 3)
There is no significant difference in neurological recovery btwn surgically & conservatively treated pts, either in the total population or w/in complete/incomplete subgroups or cervical/noncervical subgroups.Table 4; Supplementary Table 4
Penetrating injuries are associated w/ a significantly worse chance for recovery compared to blunt TSCI.20.8% vs 51.6%; p = 0.006
There is a trend for worse prognosis of complete penetrating TSCI compared to complete blunt TSCI, whereas there is no such trend in incomplete injury. When adjusted for severity of injury, mechanism of injury maintains its significant correlation w/ neurological recovery.6.1% vs 19.9%; p = 0.08
Penetrating injuries are more likely to result in complete TSCI compared to blunt injuries.OR 1.42 (95% CI 1.09–1.84); p = 0.008
Neurological recovery of pts w/ TSCI is not different btwn developing & developed nations, & it has not significantly changed over time.Fig. 3; Supplementary Fig. 13
Follow-up duration of a study significantly & positively correlates w/ the extent of reported neurological recovery in incomplete SCI & cervical SCI, but is not a significant factor in complete or noncervical SCI.Supplementary Fig. 14
Studies w/ follow-up durations of approximately 6 mos or less report significantly lower recovery rates for pts w/ incomplete SCI compared to studies w/ long-term follow-ups. A minimum follow-up of 12 mos is recommended for TSCI studies that include pts w/ neurologically incomplete SCI.Supplementary Table 5

Acknowledgments

This study was supported by grant number 96-02-38-35036 from Tehran University of Medical Sciences, Tehran, Iran.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Rahimi-Movaghar, Khorasanizadeh, Yousefifard. Acquisition of data: Khorasanizadeh, Eskian, Chalangari, Jazayeri, Seyedpour, Khodaei. Analysis and interpretation of data: Khorasanizadeh, Yousefifard, Hosseini. Drafting the article: Khorasanizadeh, Eskian, Lu. Critically revising the article: Rahimi-Movaghar, Lu, Harrop. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Rahimi-Movaghar. Statistical analysis: Khorasanizadeh, Yousefifard, Hosseini. Administrative/technical/material support: Rahimi-Movaghar, Yousefifard. Study supervision: Rahimi-Movaghar, Harrop.

Supplemental Information

Online-Only Content

Supplemental material is available with the online version of the article.

References

  • 1

    Aarabi B, Alibaii E, Taghipur M, Kamgarpur A: Comparative study of functional recovery for surgically explored and conservatively managed spinal cord missile injuries. Neurosurgery 39:11331140, 1996

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2

    Aarabi B, Harrop JS, Tator CH, Alexander M, Dettori JR, Grossman RG, et al.: Predictors of pulmonary complications in blunt traumatic spinal cord injury. J Neurosurg Spine 17 (1 Suppl):3845, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Aarabi B, Sansur CA, Ibrahimi DM, Simard JM, Hersh DS, Le E, et al.: Intramedullary lesion length on postoperative magnetic resonance imaging is a strong predictor of ASIA impairment scale grade conversion following decompressive surgery in cervical spinal cord injury. Neurosurgery 80:610620, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Ahuja CS, Fehlings M: Concise review: Bridging the gap: novel neuroregenerative and neuroprotective strategies in spinal cord injury. Stem Cells Transl Med 5:914924, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5

    Al-Habib AF, Attabib N, Ball J, Bajammal S, Casha S, Hurlbert RJ: Clinical predictors of recovery after blunt spinal cord trauma: systematic review. J Neurotrauma 28:14311443, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Anderson DG, Voets C, Ropiak R, Betcher J, Silber JS, Daffner S, et al.: Analysis of patient variables affecting neurologic outcome after traumatic cervical facet dislocation. Spine J 4:506512, 2004

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Beaty N, Slavin J, Diaz C, Zeleznick K, Ibrahimi D, Sansur CA: Cervical spine injury from gunshot wounds. J Neurosurg Spine 21:442449, 2014

  • 8

    Biglari B, Child C, Yildirim TM, Swing T, Reitzel T, Moghaddam A: Does surgical treatment within 4 hours after trauma have an influence on neurological remission in patients with acute spinal cord injury? Ther Clin Risk Manag 12:13391346, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9

    Blauth M, Tscherne H, Haas N: Therapeutic concept and results of operative treatment in acute trauma of the thoracic and lumbar spine: the Hannover experience. J Orthop Trauma 1:240252, 1987

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    Bondurant FJ, Cotler HB, Kulkarni MV, McArdle CB, Harris JH Jr: Acute spinal cord injury. A study using physical examination and magnetic resonance imaging. Spine (Phila Pa 1976) 15:161168, 1990

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Bourassa-Moreau É , Mac-Thiong JM, Li A, Ehrmann Feldman D, Gagnon DH, Thompson C, et al.: Do patients with complete spinal cord injury benefit from early surgical decompression? Analysis of neurological improvement in a prospective cohort study. J Neurotrauma 33:301306, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Braakman R, Fontijne WP, Zeegers R, Steenbeek JR, Tanghe HL: Neurological deficit in injuries of the thoracic and lumbar spine. A consecutive series of 70 patients. Acta Neurochir (Wien) 111:1117, 1991

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13

    Bravo P, Labarta C, Alcaraz MA, Mendoza J, Verdu A: Outcome after vertebral fractures with neurological lesion treated either surgically or conservatively in Spain. Paraplegia 31:358366, 1993

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Burns AS, Lee BS, Ditunno JF Jr, Tessler A: Patient selection for clinical trials: the reliability of the early spinal cord injury examination. J Neurotrauma 20:477482, 2003

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 15

    Burns AS, Marino RJ, Flanders AE, Flett H: Clinical diagnosis and prognosis following spinal cord injury. Handb Clin Neurol 109:4762, 2012

  • 16

    Celik H, Köktekir E, Karatay M, Erdem Y, Gökçek C, Yaşitli U, et al.: A long term evaluation of result of surgically treated lower cervical spine trauma of 83 patients: a retrospective study. J Neurol Sci (Turk) 30:740747, 2013

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Centre for Evidence-Based Medicine : Levels of evidence. CEBM.net (https://www.cebm.net/2009/06/oxford-centre-evidence-based-medicine-levels-evidence-march-2009/) [Accessed November 14, 2018]

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 18

    Chiu WT, Lin HC, Lam C, Chu SF, Chiang YH, Tsai SH: Review paper: epidemiology of traumatic spinal cord injury: comparisons between developed and developing countries. Asia Pac J Public Health 22:918, 2010

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19

    Clohisy JC, Akbarnia BA, Bucholz RD, Burkus JK, Backer RJ: Neurologic recovery associated with anterior decompression of spine fractures at the thoracolumbar junction (T12-L1). Spine (Phila Pa 1976) 17 (8 Suppl):S325S330, 1992

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20

    Curt A, Dietz V: Ambulatory capacity in spinal cord injury: significance of somatosensory evoked potentials and ASIA protocol in predicting outcome. Arch Phys Med Rehabil 78:3943, 1997

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    Curt A, Keck ME, Dietz V: Functional outcome following spinal cord injury: significance of motor-evoked potentials and ASIA scores. Arch Phys Med Rehabil 79:8186, 1998

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 22

    Dakson A, Brandman D, Thibault-Halman G, Christie SD: Optimization of the mean arterial pressure and timing of surgical decompression in traumatic spinal cord injury: a retrospective study. Spinal Cord 55:10331038, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23

    Ditunno JF Jr, Stover SL, Freed MM, Ahn JH: Motor recovery of the upper extremities in traumatic quadriplegia: a multicenter study. Arch Phys Med Rehabil 73:431436, 1992

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24

    Divanoglou A, Westgren N, Bjelak S, Levi R: Medical conditions and outcomes at 1 year after acute traumatic spinal cord injury in a Greek and a Swedish region: a prospective, population-based study. Spinal Cord 48:470476, 2010

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Dobran M, Iacoangeli M, Di Somma LG, Di Rienzo A, Colasanti R, Nocchi N, et al.: Neurological outcome in a series of 58 patients operated for traumatic thoracolumbar spinal cord injuries. Surg Neurol Int 5 (Suppl 7):S329S332, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26

    Dobran M, Iacoangeli M, Nocchi N, Di Rienzo A, di Somma LG, Nasi D, et al.: Surgical treatment of cervical spine trauma: our experience and results. Asian J Neurosurg 10:207211, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 27

    Donovan WH, Cifu DX, Schotte DE: Neurological and skeletal outcomes in 113 patients with closed injuries to the cervical spinal cord. Paraplegia 30:533542, 1992

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28

    Dvorak MF, Noonan VK, Fallah N, Fisher CG, Rivers CS, Ahn H, et al.: Minimizing errors in acute traumatic spinal cord injury trials by acknowledging the heterogeneity of spinal cord anatomy and injury severity: an observational Canadian cohort analysis. J Neurotrauma 31:15401547, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 29

    Egger M, Davey Smith G, Schneider M, Minder C: Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629634, 1997

  • 30

    El Masry WS, Tsubo M, Katoh S, El Miligui YH, Khan A: Validation of the American Spinal Injury Association (ASIA) motor score and the National Acute Spinal Cord Injury Study (NASCIS) motor score. Spine (Phila Pa 1976) 21:614619, 1996

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 31

    Endo T, Suzuki S, Utsunomiya A, Uenohara H, Tominaga T: Prediction of neurological recovery using apparent diffusion coefficient in cases of incomplete spinal cord injury. Neurosurgery 68:329336, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32

    Fehlings MG, Rabin D, Sears W, Cadotte DW, Aarabi B: Current practice in the timing of surgical intervention in spinal cord injury. Spine (Phila Pa 1976) 35 (21 Suppl):S166S173, 2010

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Fehlings MG, Vaccaro A, Wilson JR, Singh A, W Cadotte D, Harrop JS, et al.: Early versus delayed decompression for traumatic cervical spinal cord injury: results of the Surgical Timing in Acute Spinal Cord Injury Study (STASCIS). PLoS One 7:e32037, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 34

    Ferbert T, Child C, Graeser V, Swing T, Akbar M, Heller R, et al.: Tracking spinal cord injury: differences in cytokine expression of IGF-1, TGF- B1, and sCD95l can be measured in blood samples and correspond to neurological remission in a 12-week follow-up. J Neurotrauma 34:607614, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 35

    Foo D, Subrahmanyan TS, Rossier AB: Post-traumatic acute anterior spinal cord syndrome. Paraplegia 19:201205, 1981

  • 36

    Frankel HL, Hancock DO, Hyslop G, Melzak J, Michaelis LS, Ungar GH, et al.: The value of postural reduction in the initial management of closed injuries of the spine with paraplegia and tetraplegia. I. Paraplegia 7:179192, 1969

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 37

    Fredø HL, Rizvi SA, Rezai M, Rønning P, Lied B, Helseth E: Complications and long-term outcomes after open surgery for traumatic subaxial cervical spine fractures: a consecutive series of 303 patients. BMC Surg 16:56, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 38

    Furlan JC, Sakakibara BM, Miller WC, Krassioukov AV: Global incidence and prevalence of traumatic spinal cord injury. Can J Neurol Sci 40:456464, 2013

  • 39

    Gertzbein SD, Court-Brown CM, Marks P, Martin C, Fazl M, Schwartz M, et al.: The neurological outcome following surgery for spinal fractures. Spine (Phila Pa 1976) 13:641644, 1988

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 40

    Ghasemi AA, Behfar B: Outcome of laminoplasty in cervical spinal cord injury with stable spine. Asian J Neurosurg 11:282286, 2016

  • 41

    Harrop JS, Naroji S, Maltenfort MG, Ratliff JK, Tjoumakaris SI, Frank B, et al.: Neurologic improvement after thoracic, thoracolumbar, and lumbar spinal cord (conus medullaris) injuries. Spine (Phila Pa 1976) 36:2125, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 42

    Heary RF, Vaccaro AR, Mesa JJ, Northrup BE, Albert TJ, Balderston RA, et al.: Steroids and gunshot wounds to the spine. Neurosurgery 41:576584, 1997

  • 43

    Higgins JPT, Green S (eds): Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1.0. London: The Cochrane Collaboration, 2011 (http://handbook.cochrane.org) [Accessed November 14, 2018]

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 44

    Higgins JPT, Thompson SG: Quantifying heterogeneity in a meta-analysis. Stat Med 21:15391558, 2002

  • 45

    Hosaka N, Kimura S, Yamazaki A, Wang X, Denda H, Ito T, et al.: Significant correlation between cerebrospinal fluid nitric oxide concentrations and neurologic prognosis in incomplete cervical cord injury. Eur Spine J 17:281286, 2008

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 46

    International Monetary Fund : World Economic Outlook. Washington, DC: International Monetary Fund, 2017

  • 47

    Iseli E, Cavigelli A, Dietz V, Curt A: Prognosis and recovery in ischaemic and traumatic spinal cord injury: clinical and electrophysiological evaluation. J Neurol Neurosurg Psychiatry 67:567571, 1999

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 48

    Jha A, Lammertse DP, Coll JR, Charlifue S, Coughlin CT, Whiteneck GG, et al.: Apolipoprotein E ε4 allele and outcomes of traumatic spinal cord injury. J Spinal Cord Med 31:171176, 2008

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 49

    Jug M, Kejžar N, Vesel M, Al Mawed S, Dobravec M, Herman S, et al.: Neurological recovery after traumatic cervical spinal cord injury is superior if surgical decompression and instrumented fusion are performed within 8 hours versus 8 to 24 hours after injury: a single center experience. J Neurotrauma 32:13851392, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 50

    Kalsi-Ryan S, Beaton D, Curt A, Popovic MR, Verrier MC, Fehlings MG: Outcome of the upper limb in cervical spinal cord injury: profiles of recovery and insights for clinical studies. J Spinal Cord Med 37:503510, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 51

    Katoh S, el Masry WS: Neurological recovery after conservative treatment of cervical cord injuries. J Bone Joint Surg Br 76:225228, 1994

  • 52

    Katoh S, el Masry WS, Jaffray D, McCall IW, Eisenstein SM, Pringle RG, et al.: Neurologic outcome in conservatively treated patients with incomplete closed traumatic cervical spinal cord injuries. Spine (Phila Pa 1976) 21:23452351, 1996

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 53

    Khorasanizadeh M, Eskian M, Vaccaro AR, Rahimi-Movaghar V: Granulocyte colony-stimulating factor (G-CSF) for the treatment of spinal cord injury. CNS Drugs 31:911937, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 54

    Kim NH, Lee HM, Chun IM: Neurologic injury and recovery in patients with burst fracture of the thoracolumbar spine. Spine (Phila Pa 1976) 24:290294, 1999

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 55

    Kim YH, Ha KY, Kim SI: Spinal cord injury and related clinical trials. Clin Orthop Surg 9:19, 2017

  • 56

    Kimura S, Hosaka N, Yuge I, Yamazaki A, Suda K, Taneichi H, et al.: Cerebrospinal fluid concentrations of nitric oxide metabolites in spinal cord injury. Spine (Phila Pa 1976) 34:E645E652, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 57

    Kirshblum S, Millis S, McKinley W, Tulsky D: Late neurologic recovery after traumatic spinal cord injury. Arch Phys Med Rehabil 85:18111817, 2004

  • 58

    Kirshblum SC, O’Connor KC: Predicting neurologic recovery in traumatic cervical spinal cord injury. Arch Phys Med Rehabil 79:14561466, 1998

  • 59

    Koo DW, Townson AF, Dvorak MF, Fisher CG: Spinal epidural abscess: a 5-year case-controlled review of neurologic outcomes after rehabilitation. Arch Phys Med Rehabil 90:512516, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 60

    Kreinest M, Ludes L, Biglari B, Küffer M, Türk A, Grützner PA, et al.: Influence of previous comorbidities and common complications on motor function after early surgical treatment of patients with traumatic spinal cord injury. J Neurotrauma 33:21752180, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 61

    Krengel WF III, Anderson PA, Henley MB: Early stabilization and decompression for incomplete paraplegia due to a thoracic-level spinal cord injury. Spine (Phila Pa 1976) 18:20802087, 1993

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 62

    Kulkarni MV, Bondurant FJ, Rose SL, Narayana PA: 1.5 tesla magnetic resonance imaging of acute spinal trauma. Radiographics 8:10591082, 1988

  • 63

    Kupcha PC, An HS, Cotler JM: Gunshot wounds to the cervical spine. Spine (Phila Pa 1976) 15:10581063, 1990

  • 64

    Kwon BK, Streijger F, Fallah N, Noonan VK, Bélanger LM, Ritchie L, et al.: Cerebrospinal fluid biomarkers to stratify injury severity and predict outcome in human traumatic spinal cord injury. J Neurotrauma 34:567580, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 65

    La Rosa G, Conti A, Cardali S, Cacciola F, Tomasello F: Does early decompression improve neurological outcome of spinal cord injured patients? Appraisal of the literature using a meta-analytical approach. Spinal Cord 42:503512, 2004

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 66

    Lauschke JL, Leong GW, Rutkowski SB, Waite PM: Changes in electrical perceptual threshold in the first 6 months following spinal cord injury. J Spinal Cord Med 34:473481, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 67

    Lee BA, Leiby BE, Marino RJ: Neurological and functional recovery after thoracic spinal cord injury. J Spinal Cord Med 39:6776, 2016

  • 68

    Lee JK, Oh CH, Kim JY, Park HC, Yoon SH: Brain activation evoked by sensory stimulation in patients with spinal cord injury: functional magnetic resonance imaging correlations with clinical features. J Korean Neurosurg Soc 58:242247, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 69

    Lenoir T, Hoffmann E, Thevenin-Lemoine C, Lavelle G, Rillardon L, Guigui P: Neurological and functional outcome after unstable cervicothoracic junction injury treated by posterior reduction and synthesis. Spine J 6:507513, 2006

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 70

    Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, et al.: The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 6:e1000100, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 71

    Liu Y, Shi CG, Wang XW, Chen HJ, Wang C, Cao P, et al.: Timing of surgical decompression for traumatic cervical spinal cord injury. Int Orthop 39:24572463, 2015

  • 72

    Louis CA, Gauthier VY, Louis RP: Posterior approach with Louis plates for fractures of the thoracolumbar and lumbar spine with and without neurologic deficits. Spine (Phila Pa 1976) 23:20302040, 1998

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 73

    Lugo LH, Salinas F, García HI: Out-patient rehabilitation programme for spinal cord injured patients: evaluation of the results on motor FIM score. Disabil Rehabil 29:873881, 2007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 74

    Machino M, Yukawa Y, Ito K, Nakashima H, Kanbara S, Morita D, et al.: Can magnetic resonance imaging reflect the prognosis in patients of cervical spinal cord injury without radiographic abnormality? Spine (Phila Pa 1976) 36:E1568E1572, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 75

    Marino RJ, Burns S, Graves DE, Leiby BE, Kirshblum S, Lammertse DP: Upper- and lower-extremity motor recovery after traumatic cervical spinal cord injury: an update from the national spinal cord injury database. Arch Phys Med Rehabil 92:369375, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 76

    Marino RJ, Ditunno JF Jr, Donovan WH, Maynard F Jr: Neurologic recovery after traumatic spinal cord injury: data from the Model Spinal Cord Injury Systems. Arch Phys Med Rehabil 80:13911396, 1999

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 77

    Marino RJ, Herbison GJ, Ditunno JF Jr: Peripheral sprouting as a mechanism for recovery in the zone of injury in acute quadriplegia: a single-fiber EMG study. Muscle Nerve 17:14661468, 1994

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 78

    Marré B, Ballesteros V, Martínez C, Zamorano JJ, Ilabaca F, Munjin M, et al.: Thoracic spine fractures: injury profile and outcomes of a surgically treated cohort. Eur Spine J 20:14271433, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 79

    Martínez-Pérez R, Cepeda S, Paredes I, Alen JF, Lagares A: MRI prognostication factors in the setting of cervical spinal cord injury secondary to trauma. World Neurosurg 101:623632, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 80

    Masuda T, Miyamoto K, Wakahara K, Matsumoto K, Hioki A, Shimokawa T, et al.: Clinical outcomes of surgical treatments for traumatic spinal injuries due to snowboarding. Asian Spine J 9:9098, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 81

    Matsushita A, Maeda T, Mori E, Yuge I, Kawano O, Ueta T, et al.: Can the acute magnetic resonance imaging features reflect neurologic prognosis in patients with cervical spinal cord injury? Spine J 17:13191324, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 82

    Matsushita A, Maeda T, Mori E, Yugue I, Kawano O, Ueta T, et al.: Subacute T1-low intensity area reflects neurological prognosis for patients with cervical spinal cord injury without major bone injury. Spinal Cord 54:2428, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 83

    Maynard FM, Reynolds GG, Fountain S, Wilmot C, Hamilton R: Neurological prognosis after traumatic quadriplegia. Three-year experience of California Regional Spinal Cord Injury Care System. J Neurosurg 50:611616, 1979

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 84

    McEvoy RD, Bradford DS: The management of burst fractures of the thoracic and lumbar spine. Experience in 53 patients. Spine (Phila Pa 1976) 10:631637, 1985

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 85

    Merriam WF, Taylor TK, Ruff SJ, McPhail MJ: A reappraisal of acute traumatic central cord syndrome. J Bone Joint Surg Br 68:708713, 1986

  • 86

    Miranda P, Gomez P, Alday R, Kaen A, Ramos A: Brown-Sequard syndrome after blunt cervical spine trauma: clinical and radiological correlations. Eur Spine J 16:11651170, 2007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 87

    Mishra A, Agrawal D, Gupta D, Sinha S, Satyarthee GD, Singh PK: Traumatic spondyloptosis: a series of 20 patients. J Neurosurg Spine 22:647652, 2015

  • 88

    Mohanty SP, Venkatram N: Does neurological recovery in thoracolumbar and lumbar burst fractures depend on the extent of canal compromise? Spinal Cord 40:295299, 2002

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 89

    Moslavac S, Dzidić I, Kejla Z: Neurological outcome in road traffic accidents with spinal cord injury. Coll Antropol 32:583586, 2008

  • 90

    Murphy KP, Opitz JL, Cabanela ME, Ebersold MJ: Cervical fractures and spinal cord injury: outcome of surgical and nonsurgical management. Mayo Clin Proc 65:949959, 1990

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 91

    Nagata K, Inokuchi K, Chikuda H, Ishii K, Kobayashi A, Kanai H, et al.: Early versus delayed reduction of cervical spine dislocation with complete motor paralysis: a multicenter study. Eur Spine J 26:12721276, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 92

    National Spinal Cord Injury Statistical Center : Spinal Cord Injury (SCI) 2016 Facts and Figures at a Glance. J Spinal Cord Med 39:493494, 2016

  • 93

    Neva MH, Roeder CP, Felder U, Kiener B, Meier W, Perler M, et al.: Neurological outcome, working capacity and prognostic factors of patients with SCIWORA. Spinal Cord 50:7880, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 94

    Newcombe RG: Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 17:857872, 1998

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 95

    Newton D, England M, Doll H, Gardner BP: The case for early treatment of dislocations of the cervical spine with cord involvement sustained playing rugby. J Bone Joint Surg Br 93:16461652, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 96

    Normand SL: Meta-analysis: formulating, evaluating, combining, and reporting. Stat Med 18:321359, 1999

  • 97

    Nwankwo OE, Katchy AU: Outcome of a 12-week programme for management of the spinal cord injured with participation of patient’s relations at Hilltop Orthopaedic Hospital, Enugu, Nigeria. Spinal Cord 41:129133, 2003

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 98

    Okada S, Maeda T, Ohkawa Y, Harimaya K, Saiwai H, Kumamaru H, et al.: Does ossification of the posterior longitudinal ligament affect the neurological outcome after traumatic cervical cord injury? Spine (Phila Pa 1976) 34:11481152, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 99

    Oleson CV, Marino RJ, Leiby BE, Ditunno JF: Influence of age alone, and age combined with pinprick, on recovery of walking function in motor complete, sensory incomplete spinal cord injury. Arch Phys Med Rehabil 97:16351641, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 100

    Park JH, Kim JH, Roh SW, Rhim SC, Jeon SR: Prognostic factor analysis after surgical decompression and stabilization for cervical spinal-cord injury. Br J Neurosurg 31:194198, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 101

    Park SE, Elliott S, Noonan VK, Thorogood NP, Fallah N, Aludino A, et al.: Impact of bladder, bowel and sexual dysfunction on health status of people with thoracolumbar spinal cord injuries living in the community. J Spinal Cord Med 40:548559, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 102

    Philippi R, Kuhn W, Zäch GA, Jacob-Chia D, Dollfus P, Molé JP: Survey of the neurological evolution of 300 spinal cord injuries seen within 24 hours after injury. Paraplegia 18:337346, 1980

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 103

    Piepmeier JM, Jenkins NR: Late neurological changes following traumatic spinal cord injury. J Neurosurg 69:399402, 1988

  • 104

    Pouw MH, Kwon BK, Verbeek MM, Vos PE, van Kampen A, Fisher CG, et al.: Structural biomarkers in the cerebrospinal fluid within 24 h after a traumatic spinal cord injury: a descriptive analysis of 16 subjects. Spinal Cord 52:428433, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 105

    Rahimi-Movaghar V: Efficacy of surgical decompression in the setting of complete thoracic spinal cord injury. J Spinal Cord Med 28:415420, 2005

  • 106

    Riva JJ, Malik KM, Burnie SJ, Endicott AR, Busse JW: What is your research question? An introduction to the PICOT format for clinicians. J Can Chiropr Assoc 56:167171, 2012

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 107

    Roquilly A, Seguin P, Mimoz O, Feuillet F, Rosenczweig E, Chevalier F, et al.: Risk factors for prolonged duration of mechanical ventilation in acute traumatic tetraplegic patients—a retrospective cohort study. J Crit Care 29:313.e7313.e13, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 108

    Roth EJ, Lawler MH, Yarkony GM: Traumatic central cord syndrome: clinical features and functional outcomes. Arch Phys Med Rehabil 71:1823, 1990

  • 109

    Sannohe A, Harata S, Ueyama K, Ito J, Sato T, Tian W, et al.: The prognosis and the treatment of patients with a C3/4 spinal cord injury. Spinal Cord 34:486487, 1996

  • 110

    Sapkas GS, Papadakis SA: Neurological outcome following early versus delayed lower cervical spine surgery. J Orthop Surg (Hong Kong) 15:183186, 2007

  • 111

    Shimada K, Tokioka T: Sequential MR studies of cervical cord injury: correlation with neurological damage and clinical outcome. Spinal Cord 37:410415, 1999

  • 112

    Shrier I, Boivin JF, Steele RJ, Platt RW, Furlan A, Kakuma R, et al.: Should meta-analyses of interventions include observational studies in addition to randomized controlled trials? A critical examination of underlying principles. Am J Epidemiol 166:12031209, 2007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 113

    Shrosbree RD: Neurological sequelae of reduction of fracture dislocations of the cervical spine. Paraplegia 17:212221, 1979

  • 114

    Silberstein M, Brown D, Tress BM, Hennessey O: Suggested MRI criteria for surgical decompression in acute spinal cord injury. Preliminary observations. Paraplegia 30:704710, 1992

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 115

    Singh R, Kumar RR, Setia N, Magu S: A prospective study of neurological outcome in relation to findings of imaging modalities in acute spinal cord injury. Asian J Neurosurg 10:181189, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 116

    Singhal B, Mohammed A, Samuel J, Mues J, Kluger P: Neurological outcome in surgically treated patients with incomplete closed traumatic cervical spinal cord injury. Spinal Cord 46:603607, 2008

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 117

    Sipski ML, Jackson AB, Gómez-Marín O, Estores I, Stein A: Effects of gender on neurologic and functional recovery after spinal cord injury. Arch Phys Med Rehabil 85:18261836, 2004

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 118

    Spiess MR, Müller RM, Rupp R, Schuld C, van Hedel HJ: Conversion in ASIA impairment scale during the first year after traumatic spinal cord injury. J Neurotrauma 26:20272036, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 119

    Spivak JM, Weiss MA, Cotler JM, Call M: Cervical spine injuries in patients 65 and older. Spine (Phila Pa 1976) 19:23022306, 1994

  • 120

    Squair JW, Bélanger LM, Tsang A, Ritchie L, Mac-Thiong JM, Parent S, et al.: Spinal cord perfusion pressure predicts neurologic recovery in acute spinal cord injury. Neurology 89:16601667, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 121

    Sumida M, Fujimoto M, Tokuhiro A, Tominaga T, Magara A, Uchida R: Early rehabilitation effect for traumatic spinal cord injury. Arch Phys Med Rehabil 82:391395, 2001

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 122

    Sun LQ, Shen Y, Li YM: Quantitative magnetic resonance imaging analysis correlates with surgical outcome of cervical spinal cord injury without radiologic evidence of trauma. Spinal Cord 52:541546, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 123

    Sun LQ, Shen Y, Liu LZ, Lu ZB: Classification and surgical decision making in cervical spinal cord injury without radiologic evidence of trauma. J Neurol Sci (Turk) 30:660673, 2013

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 124

    Tasdemiroglu E, Tibbs PA: Long-term follow-up results of thoracolumbar fractures after posterior instrumentation. Spine (Phila Pa 1976) 20:17041708, 1995

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 125

    Tominaga S: Periodical, neurological-functional assessment for cervical cord injury. Paraplegia 27:227236, 1989

  • 126

    Tsutsumi S, Ueta T, Shiba K, Yamamoto S, Takagishi K: Effects of the Second National Acute Spinal Cord Injury Study of high-dose methylprednisolone therapy on acute cervical spinal cord injury—results in spinal injuries center. Spine (Phila Pa 1976) 31:29922997, 2006

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 127

    Umerani MS, Abbas A, Sharif S: Clinical outcome in patients with early versus delayed decompression in cervical spine trauma. Asian Spine J 8:427434, 2014

  • 128

    Vale FL, Burns J, Jackson AB, Hadley MN: Combined medical and surgical treatment after acute spinal cord injury: results of a prospective pilot study to assess the merits of aggressive medical resuscitation and blood pressure management. J Neurosurg 87:239246, 1997

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 129

    van Middendorp JJ, Hosman AJ, Doi SA: The effects of the timing of spinal surgery after traumatic spinal cord injury: a systematic review and meta-analysis. J Neurotrauma 30:17811794, 2013

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 130

    Wang L, Liu C, Zhao Q, Tian J: Posterior pedicle screw fixation for complex atlantoaxial fractures with atlanto-dental interval of ≥ 5 mm or C2-C3 angulation of ≥ 11°. J Orthop Surg Res 9:104, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 131

    Waters RL, Adkins RH, Yakura JS, Sie I: Effect of surgery on motor recovery following traumatic spinal cord injury. Spinal Cord 34:188192, 1996

  • 132

    Waters RL, Adkins RH, Yakura JS, Sie I: Motor and sensory recovery following complete tetraplegia. Arch Phys Med Rehabil 74:242247, 1993

  • 133

    Waters RL, Adkins RH, Yakura JS, Sie I: Motor and sensory recovery following incomplete tetraplegia. Arch Phys Med Rehabil 75:306311, 1994

  • 134

    Waters RL, Sie I, Adkins RH, Yakura JS: Injury pattern effect on motor recovery after traumatic spinal cord injury. Arch Phys Med Rehabil 76:440443, 1995

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 135

    Weinstein DE, Ko HY, Graziani V, Ditunno JF Jr: Prognostic significance of the delayed plantar reflex following spinal cord injury. J Spinal Cord Med 20:207211, 1997

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 136

    Weyns F, Rommens PM, Van Calenbergh F, Goffin J, Broos P, Plets C: Neurological outcome after surgery for thoracolumbar fractures. A retrospective study of 93 consecutive cases, treated with dorsal instrumentation. Eur Spine J 3:276281, 1994

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 137

    Wiberg J, Hauge HN: Neurological outcome after surgery for thoracic and lumbar spine injuries. Acta Neurochir (Wien) 91:106112, 1988

  • 138

    Wilson JR, Cadotte DW, Fehlings MG: Clinical predictors of neurological outcome, functional status, and survival after traumatic spinal cord injury: a systematic review. J Neurosurg Spine 17 (1 Suppl):1126, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 139

    Wilson JR, Davis AM, Kulkarni AV, Kiss A, Frankowski RF, Grossman RG, et al.: Defining age-related differences in outcome after traumatic spinal cord injury: analysis of a combined, multicenter dataset. Spine J 14:11921198, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 140

    Wilson JR, Singh A, Craven C, Verrier MC, Drew B, Ahn H, et al.: Early versus late surgery for traumatic spinal cord injury: the results of a prospective Canadian cohort study. Spinal Cord 50:840843, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 141

    Wilson JR, Tetreault LA, Kwon BK, Arnold PM, Mroz TE, Shaffrey C, et al.: Timing of decompression in patients with acute spinal cord injury: a systematic review. Global Spine J 7 (3 Suppl):95S115S, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 142

    Wright JG, Swiontkowski MF, Heckman JD: Introducing levels of evidence to the journal. J Bone Joint Surg Am 85-A:13, 2003

  • 143

    Yisheng W, Fuying Z, Limin W, Junwei L, Guofu P, Weidong W: First aid and treatment for cervical spinal cord injury with fracture and dislocation. Indian J Orthop 41:300304, 2007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 144

    Young JS, Dexter WR: Neurological recovery distal to the zone of injury in 172 cases of closed, traumatic spinal cord injury. Paraplegia 16:3949, 1978

  • 145

    Yousefifard M, Rahimi-Movaghar V, Baikpour M, Ghelichkhani P, Hosseini M, Jafari A, et al.: Early versus late spinal decompression surgery in treatment of traumatic spinal cord injuries; a systematic review and meta-analysis. Emergency (Tehran) 5:e37, 2017

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 146

    Zariffa J, Kramer JL, Fawcett JW, Lammertse DP, Blight AR, Guest J, et al.: Characterization of neurological recovery following traumatic sensorimotor complete thoracic spinal cord injury. Spinal Cord 49:463471, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 147

    Zeilig G, Weingarden HP, Zwecker M, Rubin-Asher D, Ratner A, Ohry A: Civilian spinal cord injuries due to terror explosions. Spinal Cord 48:814818, 2010

  • 148

    Zhao LJ, Xu RM, Ma WH, Jiang WY: Evaluation of the value of pedicle screws for cervicothoracic fracture-dislocation. Orthop Surg 2:27 31, 2010

  • 149

    Zhou F, Zou J, Gan M, Zhu R, Yang H: Management of fracture-dislocation of the lower cervical spine with the cervical pedicle screw system. Ann R Coll Surg Engl 92:406410, 2010

    • Crossref
    • PubMed