Characterizing Natural Recovery after Traumatic Spinal Cord Injury

Abstract

The predominant tool used to predict outcomes after traumatic spinal cord injury (SCI) is the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI), in association with the American Spinal Injury Association (ASIA) Impairment Scale (AIS). These measures have evolved based on analyses of large amounts of longitudinal neurological recovery data published in numerous separate studies. This article reviews and synthesizes published data on neurological recovery from multiple sources, only utilizing data in which the sacral sparing definition was applied for determination of completeness. Conversion from a complete to incomplete injury is more common in tetraplegia than paraplegia. The majority of AIS conversion and motor recovery occurs within the first 6–9 months, with the most rapid rate of motor recovery occurring in the first three months after injury. Motor score changes, as well as recovery of motor levels, are described with the initial strength of muscles as well as the levels of the motor zone of partial preservation influencing the prognosis. Total motor recovery is greater for patients with initial AIS B than AIS A, and greater after initial AIS C than with motor complete injuries. Older age has a negative impact on neurological and functional recovery after SCI; however, the specific age (whether >50 or >65 years) and underlying reasons for this impact are unclear. Penetrating injury is more likely to lead to a classification of a neurological complete injury compared with blunt trauma and reduces the likelihood of AIS conversion at one year. There are insufficient data to support gender having a major effect on neurological recovery after SCI.

Introduction

The patterns of neurological recovery after traumatic spinal cord injury (SCI) may influence acute management decisions, guide prognosis discussions, help establish functional goals, and aid in the development of individualized rehabilitation programs. Knowledge of natural recovery is essential in determining appropriate inclusion/exclusion criteria and selecting outcome measures in clinical trials to help evaluate the efficacy of treatment interventions. For these reasons, a detailed knowledge of methods to classify neurological recovery and their strengths and limitations is important.

The tools used to classify injury and predict recovery should be accessible in most medical settings. The most common tool used to predict outcomes after SCI is the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI), in association with the American Spinal Injury Association (ASIA) Impairment Scale (AIS).¹ The ISNCSCI was developed to standardize examination techniques and provide consistent terminology and definitions in SCI classification to detect changes in neurological function over time.²

The first edition, published by ASIA in 1982, used the Frankel Scale to classify injury severity.³ Significant revisions adopted in 1992 included the replacement of the Frankel Scale with the AIS, and the use of sacral sparing criteria to define injury completeness.⁴ Before 1992, a complete injury was defined by the absence of sensory and motor sparing more than three levels below the neurological level of injury (NLI). Since 1992, the determination of injury completeness has depended on the presence or absence of sensory and/or motor function at the lowest sacral levels (sensory function in the S4-S5 dermatomes, presence of deep anal pressure (DAP), or voluntary anal contraction (VAC). The sacral sparing definition was reported initially to be a more stable classification with fewer patients converting from incomplete to complete injury status,⁵ which was confirmed in a more recent analysis.⁶ The current AIS classification system is summarized in Table 1.

Table 1.

ASIA Impairment Scale¹

AIS
A = Complete	No sensory or motor function is preserved in the sacral segments S4–S5.
B = Sensory incomplete	Sensory but not motor function is preserved below the neurological level and includes the sacral segments S4–S5.
C = Motor incomplete	Motor function is preserved at the most caudal sacral segments on voluntary anal contraction (VAC) OR the patient meets the criteria for sensory incomplete status (sensory function preserved at the most caudal sacral segments [S4–S5] by LT, PP, or DAP), and has some sparing of motor function more than three levels below the ipsilateral motor level on either side of the body. More than half of the key muscles below the neurological level have a muscle grade less than 3.
D = Motor incomplete	Motor function is preserved below the neurological level, and at least half of the key muscles below the neurological level have a muscle grade of 3 or more.
E = Normal	Complete return of all motor and sensory function, but one may still have abnormal reflexes.

AIS American Spinal Injury Association Impairment Scale; LT, light touch; PP, pinprick; DAP, deep anal pressure.

¹American Spinal Injury Association (2019). International Standards for Neurological Classification of Spinal Cord Injury. ASIA: Richmond. (Used by permission.)

Several clinical studies and databases, including from the Sygen^® trial, European Multicenter Study about Spinal Cord Injury (EMSCI), National Spinal Cord Injury Model Systems (SCIMS), Rick Hansen Spinal Cord Injury Registry (RHSCIR), and North American Clinical Trials Network^® (NACTN), have been instrumental in prospectively collecting and managing neurological data from thousands of patients with traumatic SCI that have been used to characterize the patterns and extent of neurological recovery. Some previous analyses included data collected before the 1992 changes in classification^7,8; as such, awareness of the distinctive definitions of injury completeness is important when interpreting newer studies on conversion and comparing with historical data.

This article describes natural neurological recovery after traumatic SCI and focuses on AIS data published using the sacral sparing definition. Additional variables influencing outcomes, such as zones of preserved function below the NLI, patient age, injury mechanism, and gender, are also reviewed.

Using the ISNCSCI and AIS in Prognostication of Recovery

The psychometrical properties of the ISNCSCI and the AIS have been studied extensively to ensure the strength of the instrument. In general, the ISNCSCI has been shown to be a valid, reliable, and responsive tool for the neurological assessment of adults with SCI when used in the acute care, rehabilitation, and outpatient settings.⁹ Numerous studies have reported high intrarater^10–12 and interrater reliability,^11–13 with the reliability for light touch (LT) and motor examinations greater than pinprick (PP) scores, possibly because of the complexity of discriminating sharp from dull sensation.^12–14 Examination reliability varies by injury level and severity, with greater reliability for thoracolumbar relative to cervical lesions and neurologically complete relative to incomplete injuries.^12,13

Limitations of the ISNCSCI and AIS are recognized, however. Drawbacks of the comprehensive neurological examination include the dependence on a patient's ability to participate, assessor experience/training, lack of motor testing for the T2–T12 myotomes, and complexity of the sacral examination. An important challenge pertains to the heterogeneity of SCI (e.g., injury level, severity, etiology, the timing of interventions, disease-modifying complications, and patient-specific factors such as genetic polymorphisms).^15,16

Thus, patients who share the same AIS classification may have varying outcomes based on these and other not yet understood factors.^17,18 This variation may impact the ability to detect an effect on the recovery of an experimental intervention, increasing the size of standard deviations.¹⁹ To obtain adequate power in a clinical trial in the face of this heterogeneity, a large sample size may be required. Alternatively, clinical studies may seek to define responder cohorts or utilize stratification to address the heterogeneity problem.

For the AIS classification, recognized limitations include a ceiling effect for mild injuries, particularly those that are initial AIS D grade, and the potential for misclassification at the time of the initial examination or follow-up (e.g., presence of VAC or DAP) that may lead to change from neurological complete to incomplete status, and vice versa, only based on sacral sparing changes.^20–22 Despite these limitations, the ISNCSCI and AIS classifications are the most extensively studied predictors of outcome after SCI, and performance of the ISNCSCI is required for any person being considered for entry into an SCI clinical trial.¹⁸

Classification accuracy and the importance of ISNCSCI training

Successful classification of SCI demands a unique skill set, with a comprehensive understanding of ISNCSCI rules and nuances. Inherent challenges of the classification system and common areas of confusion have been described.^23–25 While the ISNCSCI has undergone several revisions over the last few decades to improve understanding and consistency, areas of confusion and high error rates in classification have been reported.^25–27 Osunronbi and Sharma²⁶ observed overall classification accuracy of 78.1%, highest for motor scores (96.8%) and complete/incomplete status (83.3%) and lowest for the motor (36.8%) and sensory (43.5%) levels; clinician experience level and accuracy were positively associated. Other studies reported that motor incomplete lesions represented greater classification challenges than motor complete injuries.^28–30

Computational algorithms have been validated to assist clinicians and researchers in ISNCSCI classification^31,32 and, when used to evaluate assessor accuracy, found a significant degree of errors in manual classification.^25,27 Armstrong and associates²⁷ reported one or more errors on 74.5% of worksheets, with most errors involving incorrect classification of the motor (30.1%) and sensory (12.4%) levels, zone of partial preservation (ZPP) (24.0%), and AIS grade (8.3%). Schuld and colleagues²⁵ similarly found the lowest agreement in the determination of motor level, motor ZPP, and AIS grade, with the highest discordance in the classification of AIS B and C grades. As such, the use of a computational algorithm is recommended to reduce classification errors.

Numerous studies demonstrated the importance of formal training in enhancing classification accuracy.^{25,28,29,33,34} For example, the overall number of correct classifications has been shown to improve from 49.6% to 91.5% on pre- and post-training assessments, respectively.³³ Therefore, sufficient training should take place before utilizing the ISNCSCI and AIS and should also be repeated regularly to maintain consistency.^18,25 Further, there should be close attention to potential ISNCSCI errors during clinical trials.

Timing of assessment and predicting outcome

It is essential to understand at what point after injury a reliable examination may be expected and the challenges to the reliability, especially in the early period after SCI. Historically, the ISNCSCI examination performed at one month post-injury served as a common baseline for predicting recovery^35–38 because this time point corresponded closely with admission to a rehabilitation facility. Over the last several decades, as acute care hospital length of stays have decreased, more recent neurological data from multiple sources (i.e., SCIMS and EMSCI databases) at earlier examination time points are now available.^20,39–41 The timing of assessment and its role in prognostication have garnered considerable research interest because of the relatively short therapeutic windows of many agents shown to be effective in pre-clinical studies.

Some studies demonstrated that the ISNCSCI examination performed ≥72 h after injury is superior to earlier assessments.^42,43 Brown and coworkers⁴² studied a single muscle group in individuals with C4 to C7 motor complete SCI; muscle recovery at the zone of injury was better predicted by the 72-h motor examination than by motor testing performed within 24 h of injury. The authors suggested that post-injury edema or hemorrhage may increase between the 24- and 72-h time points and could account for differences in the predictive capabilities of the two examinations.⁴²

More recent magnetic resonance imaging (MRI) studies have confirmed a propagation of the T2-positive edema signal in this time frame and suggested that the rate and extent are injury severity dependent.⁴⁴ Herbison and colleagues⁴³ compared motor examinations at ≤24-h and ≥72-h time points and reported that motor testing after 24 h (72 h, 1 week, and 2 weeks) demonstrated a stronger correlation with motor scores at six months; concluding that examinations completed between 72 h to one week after injury are optimal times for research and prognostication.

In contrast, other studies found ISNCSCI examinations performed in <72 h after injury to have comparable prognostic value to later examinations.^39,45–48 Using the upper extremity motor score (UEMS) to predict six-month motor scores in patients with motor complete SCI (C4–T1), Blaustein and coworkers⁴⁵ found no significant difference in predictive capability when two examination time points (≤24 h vs. 72 h to one week after injury) were compared. However, this study differed from Brown and associates⁴² because Blaustein and coworkers⁴⁵ evaluated total motor score (TMS) changes as opposed to individual muscle changes.⁴⁹

Marino and associates³⁹ dichotomized the time of the initial ISNCSCI examination to fewer than three days and three to seven days after injury and found that conversion rates from complete to incomplete status were not related to the timing of the initial examination. One important caveat to keep in mind with this study is that sacral and sensory examination data were not available for all subjects. The assumption was made that the injury was classified as motor incomplete only if the lower-extremity motor score (LEMS) was greater than 0.³⁹

Similarly, van Middendorp and colleagues⁴⁸ found no significant difference in the rate of conversion based on the timing of initial assessment (≤72 h after injury versus >72 h and within 15 days of injury). El Tecle and colleagues⁴⁶ conducted a meta-analysis of studies to characterize the natural recovery of patients with neurologically complete SCI and determined that conversion rates did not significantly differ based on the timing of the initial examination (within 72 h vs. after 72 h).

Burns and coworkers⁴⁷ evaluated the reliability of the early ISNCSCI examination (performed ≤48 h after injury) in motor complete SCI; 99% of the initial examinations were performed within one day of injury. The authors identified multiple factors, including mechanical ventilation, intoxication, concomitant closed head injury, psychological disturbance, and severe pain that could potentially impact the reliability of the early examination.⁴⁷ Compared with individuals who did not have any factors interfering with cognition or communication at the time of initial assessment, patients with initial AIS A injuries affected by one or more of these factors were more likely to convert to incomplete status on the first repeat examination (∼1 week post-injury) (9.3% vs. 2.6%; p = 0.36) as well as on the one year assessment (17.4% vs. 6.7%; p = 0.38). The authors attributed the lack of a significant difference to a small sample size. They concluded that an ISNCSCI examination performed within 48 h of injury could be used to identify reliably patients with AIS A injuries if there are no factors impacting the patient's cognition or communication.⁴⁷

Consistency of the examination within the first 24 h period of injury is of great interest to the field of SCI, and it is still unclear as to the most appropriate time for an early baseline examination. For example, the Surgical Timing in Acute Spinal Cord Injury Study (STASCIS) evaluated “early” versus “late” surgical groups based on a time frame of less than or ≥24 h. ⁵⁰ Their early intervention was performed with a mean time of 14.2 (±5.4) h after injury, suggesting a very early initial examination, and their conversion rates were significantly higher than previous studies reported in historical controls.⁵⁰

A recent study by Evaniew and associates⁵¹ evaluated the influence of the early examination (within 48 h of injury) on conversion rates in patients with AIS A tetraplegia in a single center study. They found that conversion from complete to incomplete status was significantly more likely in patients who were examined within four hours of injury relative to individuals examined after four hours, with conversion rates of 79% (11/14) and 47% (33/71), respectively.⁵¹ Similarly, 50% (7/14) of patients examined within four hours of injury experienced improvement of at least two AIS grades in comparison with only 21% (15/71) of those examined more than four hours after injury.⁵¹ In the adjusted analysis, individuals assessed within four hours of injury were significantly more likely to experience conversion than patients examined after four hours (odds ratio 5.0; p = 0.04). The presence of spinal shock may be an important factor precluding accurate classification in this early time period.⁵¹

The Spinal Emergency Evaluation of Deficits (SPEED) test is a proposed clinical assessment that could be employed by early responders (at the site of injury and emergency department) that was derived retrospectively from ambulance records to determine the severity and level of injury rapidly.⁵² Although the SPEED test may offer a potential option for early assessment, clinical use still requires validation and implementation, and this evaluation has not been studied for long-term neurologic prognostication.

Introduced in 2020, the Expedited ISNCSCI examination (E-ISNCSCI) enables clinicians to determine the NLI and AIS with a minimum number of steps and is intended for certain circumstances in which a more rapid and abbreviated examination is required (i.e., quick screening for preliminary determination of acute clinical trial eligibility).⁵³ Importantly, the E-ISNCSCI is not a replacement for the full ISNCSCI examination, which remains the standard for comprehensive characterization of SCI.⁵³ It is recommended to perform the full ISNCSCI after acute injury, on rehabilitation admission and discharge, to evaluate for possible changes in neurological status, and neurological assessment in clinical trials. If using the E-ISNCSCI to determine early trial eligibility, a full ISNCSCI examination should follow. In addition, it is recommended that the E-ISNCSCI be used only by experienced evaluators who have a comprehensive understanding of the full ISNCSCI examination.⁵³

In summary, the literature supports examinations performed within the 24 h to one week post-injury period as being the most reliable and predictive of neurological outcomes. Earlier examinations (within the 24 h period) may be reliable. There are often factors interfering with a patient's cognition and communication in this time period, however, that limit its usefulness. Results from very early examinations within four hours of injury probably should be interpreted with caution, because they may not be stable enough to determine a definitive neurologically complete injury. Further systematic study of the early examination, perhaps in concert with other diagnostic aids such as MRI⁵⁴ and intraoperative neurophysiology⁵⁵ appears warranted.

Trends in Natural Recovery after Traumatic SCI: AIS Conversion and Change

One of the most important neurological outcomes is the conversion from a neurologically complete to incomplete injury and vice versa. Numerous studies have utilized different databases and heterogeneous samples (including all levels of injury) to determine the probability of conversion after acute neurologically complete SCI.^{8,21,40,47,48,56–58} Of note, while some articles^20,40 have used the term “conversion” to describe any change in AIS grade (e.g., a change from an AIS C to D), in this review, conversion specifically denotes a change from initial neurologic complete to incomplete status (AIS A to AIS B–E) or from initial incomplete to neurological complete injury (AIS B–D to AIS A).

Approximately 30–60% of patients with SCI are classified initially as neurologically complete (AIS A).^20,21,48,59 Overall, between 20–30% of individuals with AIS A SCI on baseline examination (within 30 days of injury) will convert to incomplete status, with roughly one-half of these improving to AIS B and one-half to AIS C and D^{8,20,21,46,48,56} (Table 2). The percentage of conversion from a neurologically complete to incomplete injury in tetraplegia is greater than in paraplegia—almost double that reported in paraplegia in some studies,^40,56,57 although not corroborated in a review by El Tecle and colleagues.⁴⁶ When further stratified by spinal region, the likelihood of conversion is greatest for lumbar SCI, followed by cervical and lower thoracic levels, with the lowest probability of conversion in upper thoracic lesions.^{7,8,21,35,41,56,57,60–63}

Table 2.

Percentage of Individuals with Neurological Complete Spinal Cord Injury (AIS A) Who Convert to Incomplete Status during the First Post-Injury Year (Heterogeneous Sample Including Tetraplegia and Paraplegia)

Authors/year	Neurological level of injury	Initial exam	Follow-up exam	Sample size (N)	Overall conversion (%)	AIS A to B (%)	AIS A to C or D (%)
8Marino et al. 1999a	All levels	≤7 days	1 y	482	15.4	7.3	8.1
47Burns et al. 2003	All levels	≤48 h	1 y	53	11.3	5.7	5.7
56Fawcett et al. 2007a,b,c	Cervical and thoracic	≤30 d	1 y	NR	20	10	10
20Spiess et al. 2009b	All levels	≤15 d	1 y	139	30.2	17.3	13
	Above T10			62	24.2	12.9	11.2
48van Middendorp et al. 2009b	Above T11	≤15 d	6 m – 1 y	161	26.1	14.3	11.8
21Kirshblum et al. 2016a	All levels	≤30 d	1 y	187	27.8	10.7	17.1

AIS, American Spinal Injury Association Impairment Scale; NR, not reported;

^aSpinal Cord Injury Model Systems database.

^bEuropean Multicenter Study about Spinal Cord Injury database.

^cSygen database.

When reviewing the literature, it is important to consider the timing of initial and follow-up assessments. For determination of long-term neurological outcomes, the majority of studies have evaluated recovery from a baseline examination, typically performed within two weeks to 30 days after injury, to a follow-up assessment at approximately one year post-injury. While the most rapid recovery is observed during the first three months of injury,⁴⁰ some neurological recovery continues after discharge from rehabilitation. Kirshblum and associates⁵⁸ assessed outcomes from rehabilitation discharge (∼3 months post-injury) to one year and reported that 13.5% of individuals with AIS A injuries experienced a conversion to incomplete status (5.6% AIS B; 3.5% AIS C; 4.1% AIS D; 0.2% AIS E) during this period. The likelihood of conversion was greater for patients with tetraplegia relative to those with paraplegia.

Late neurologic recovery (after one year) for patients with complete injuries is limited. Kirshblum and coworkers⁶⁴ found that only 5.6% of patients with AIS A SCI (all levels) at one year post-injury converted to incomplete status (3.5% improved to AIS B; 1.05% each improved to AIS C and D) by five years.⁶⁴ The authors proposed that for 50% (3/6) of patients who converted to AIS D by five years, the reason for the change in classification was likely because of coding error because there was no documented improvement in motor scores or other substantiating data.⁶⁴ On the other hand, the likelihood of patient follow-up, which tends to be lower for patients with AIS D injuries relative to those with AIS A, B, and C grades, is another important consideration and may result in an underestimation of recovery to AIS D grade.⁶⁴

Trends in neurological recovery may have changed over the last three decades, with higher rates of conversion reported in more recent literature compared with historical studies,^21,51 although this was not found by Steeves and coworkers⁴⁰ in comparison with the Sygen and EMSCI databases from the 1990s to early-mid 2000s. Differences in acute traumatic SCI management, for example, recommendations for the maintenance of mean arterial pressure in the acute injury phase, earlier timing⁵⁰ and adequacy of decompression surgery,^50,65–67 and changes in the use of methylprednisolone may contribute to this trend and warrant further study. When planning clinical trials, this information is important to keep in mind when using historical controls in studies of interventions for neurological recovery in complete SCI.

For patients with initial AIS A, the length of the ZPP may help predict conversion.^8,41 For example, Marino and colleagues⁸ found that individuals with a ZPP >3 levels had an increased likelihood of conversion to motor incomplete status than those without an extended ZPP (13.3% vs. 3.6%). The role of the ZPP in predicting conversion was also supported in a study by Zariffa and associates⁴¹ that assessed neurological recovery after complete paraplegia.

Conversion in complete tetraplegia

Conversion to incomplete status for patients with initial AIS A tetraplegia is ∼30%, with roughly one-half of these improving to AIS B and one-half to AIS C and D (Table 3.) Further evidence supporting this pattern of neurological conversion was provided by El Tecle and colleagues,⁴⁶ who conducted a meta-analysis of 20 studies published after 1992 to characterize neurological recovery in patients with complete injuries and reported an overall conversion rate of 33.3%. This rate of conversion is significantly higher than an earlier and smaller study (N = 61) by Waters and coworkers,³⁶ although the number of patients who had follow-up was relatively small.

Table 3.

Percentage of Individuals from Different Databases with Complete Tetraplegia Who Converted to Neurological Incomplete Status during the First Post-Injury Year

Authors/year	Initial exam	Follow-up exam	Sample size (N)	Overall conversion (%)	AIS A to B (%)	AIS A to C (%)	AIS A to D (%)
39Marino et al. 2011a	≤7 d	1 y	336	29.8	14.6	8.0	7.1
40Steeves et al. 2011 (C4–C7 only)
Sygenb	≤72 h	1 y	228	30.2	NR	NR	NR
EMSCIc	≤7 d	50 w	77	32.9	NR	NR	NR
21Kirshblum et al. 2016a	≤30 d	1 y	72	34.7	15.3	13.9	5.5
51Evaniew et al. 2020d	≤48 h (mean 11.8 ± 9.8 h)	Mean 208 + 75.2 d	85	52	26	18	8
	≤4 h		14	78.6	28.6	42.9	7.1
	>4 h to 48 h		71	46.5	25.4	12.7	8.5

AIS, American Spinal Injury Association Impairment Scale; NR, not reported.

^aSpinal Cord Injury Model Systems database.

^bSygen database

^cEuropean Multicenter Study about Spinal Cord Injury database.

^dRick Hansen Spinal Cord Injury Registry database.

The likelihood of conversion decreases with time from injury. Using the time point of discharge from rehabilitation (mean = 97 days post-injury) as the baseline examination, Kirshblum and associates⁵⁸ reported that 14.5% of individuals with AIS A tetraplegia converted to incomplete status (7.3% improved to AIS B; 7.1% recovered to AIS C, D, E) at one year. In a study on late neurological recovery, only 9.1% of patients with AIS A tetraplegia at one year post-injury converted to incomplete status (6.7% improved to AIS B; 2.4% became motor incomplete) by 5 years.⁶⁴

Conversion in complete paraplegia

Neurological conversion is less common after complete paraplegia (on average ∼15–20%) relative to tetraplegia, but the likelihood of conversion largely depends on the level of injury, with more caudal injuries having a greater prognosis for recovery^{7,18, 59–61} (Table 4). This conversion rate is again higher than earlier work by Waters and colleagues³⁵ who reported only 4% (N = 6/148) of patients with initial (within 30 days) AIS A paraplegia converted to incomplete status during the first year post-injury.

Table 4.

Percentage of Individuals with Complete Paraplegia Who Converted to Incomplete Status during the First Post-Injury Year

Authors/year	NLI	Initial exam	Follow-up exam	Sample size (N)	Overall conversion (%)	AIS A to B (%)	AIS A to C (%)	AIS A to D (%)
35Waters et al. 1992	T2–L2	≤30 d	1 y	148	4.0	NR	NR	NR
56Fawcett et al. 2007a	Paraplegia	≤30 d	1 y	135	12.6	5.9	4.4	2.2
62Harrop et al. 2011a	T4–T9	≤24 h	1 y	31	3.2	0	0	3.2
	T10–T12			18	5.6	0	5.6	0
	L1 and below			3	66.7	33.3	33.3	0
41Zariffa et al. 2011b	T2–T12	≤30 d	48 w	209	18.2	7.7	5.7	4.8
	T2–T5			74	9.5	4.1	4.1	1.4
	T6–T10			63	15.9	3.2	7.9	4.8
	T10–T12			72	29.2	15.3	5.6	8.3
61Lee et al. 2016c	T2–T12	≤7 d	1 y	194	15.5	7.7	3.1	4.6
21Kirshblum et al. 2016c	T1–T9	≤30 d	1 y	61	16.4	4.9	9.8	1.7
	T10 and below			54	31.5	11.1	18.5	1.9
74Wilson et al. 2018b	T1–L1	≤7 d	6 m - 1 y	22	18.2	9.1	9.1	0
63Aimetti et al. 2019b,c,d	T2–T12	≤7 d	6 m – 1 y	170	21.1 (18.8b/23.4c/16.7d)	10.9b/8.5c/0d	7.8b/9.6c/16.7d	0b/5.3c/0d
	T2–T5			45	13.3	NR	NR	NR
	T6–T9			50	16.0	NR	NR	NR
	T10–T12			75	29.3	NR	NR	NR

Bolded lines represent all levels, with subdivisions from that study listed below.

NLI, neurological level of injury; AIS, American Spinal Injury Association Impairment Scale; NR, not reported.

^aSygen database.

^bEuropean Multicenter Study about Spinal Cord Injury database.

^cSpinal Cord Injury Model Systems database.

^dNorth American Clinical Trials Network registry.

Lumbar injuries have the highest proportion of individuals converting from complete to incomplete status,⁶⁸ whereas the subgroup of higher thoracic injuries has the lowest percentage of conversion.⁶² Multiple factors that may contribute to thoracic SCI having the poorest prognosis include decreased collateral vascular supply to the thoracic cord and smaller spinal canal diameter relative to cervical and lumbar regions. In addition, given the inherent stability of the thoracic spine, a high energy mechanism is often required to cause a complete thoracic SCI.^61,63 This may result in severe fracture dislocations and more severe cord damage,⁶⁹ whereas lumbar injuries are more commonly burst fractures.⁷⁰ The lumbar spinal canal is relatively large at L1, whereas the spinal cord is smaller,⁷¹ and injuries at this level may also involve the cauda equina nerve roots leading to a proportion of lower motor neuron injuries.

As with complete tetraplegia, the percentage of patients with initial neurological complete paraplegia who convert to incomplete status decreases with later baseline examinations. Kirshblum and colleagues⁵⁸ found that 12.8% of individuals with AIS A paraplegia at the time of rehabilitation discharge (mean = 97 days post-injury) converted to incomplete status (4.4% improved to AIS B; 8.4% recovered to AIS C, D, E) at one year. The likelihood of conversion is especially low after one year; 96.4% of patients with AIS A paraplegia at one year post-injury remained complete, while only 1.7% improved to AIS B and 2.0% recovered to motor incomplete status by five years.⁶⁴

Zariffa and coworkers⁴¹ further described that the length of the sensory ZPP of ≥3 segments plays a prognostic role in AIS conversion from neurological complete to incomplete paraplegia, especially for patients with an initial NLI between T6–T12. At these levels, patients with a ZPP of ≥3 segments were more likely to convert to incomplete status by 48 weeks than those with a sensory ZPP of <3 segments (40% vs. 7% [p = 0.018] for the T6–T9 subgroup and 52.9% vs. 20.7% [p = 0.027] for the T10–T12 subgroup)⁴¹ (Fig. 1). While the T2–T5 subgroup showed a similar trend, there was not a significant correlation between sensory ZPP length and conversion.⁴¹

FIG. 1.

Illustrated are four International Standards for Neurological Classification of Spinal Cord Injury patterns that have been associated with a greater degree of recovery in those with an initial neurological injury (American Spinal Injury Association Impairment Scale A). Left to right: (1) the complete to incomplete transition based on sacral sparing; (2) in paraplegia, a sensory zone of partial preservation (ZPP) of at least three segments, here indicated below the T7 sensory level including T8 with some pinprick and T9 and T10 with only light touch; (3) some proximal motor sparing of hip flexors that is three or more levels below the neurological level of injury (NLI, here T7) with a motor score of less than 3/5; and (4) in tetraplegia, a motor ZPP of at least two segments (here, an NLI of C4 and some preservation, <3/5 in C5 and C6).

Zariffa and associates⁴¹ also reported that conversion from AIS A to motor incomplete status (AIS C or D grades) by 48 weeks was greater for those with a baseline sensory ZPP of ≥3 segments (14.3% vs. 0%, 33.3% vs. 7.7%, and 41.2% vs. 3.5% in patients with T2–T5, T6–T9, and T10–T12 levels, respectively), although was statistically significant only in the T10–T12 subgroup (p = 0.002).

AIS change in incomplete injuries

Motor complete, sensory incomplete injuries (AIS B)

Approximately 13% of initial SCI cases are classified as sensory incomplete, motor complete (AIS B).^20,21,48,59 Multiple studies have reported varying percentages of AIS recovery in patients with initial AIS B grade early after injury (within 30 days)^20,21,47,48 and at later time points⁵⁸ (Table 5). Using a baseline time point of one year, Kirshblum and coworkers⁶⁴ reported that 31.6% of individuals with AIS B (N = 114) converted to AIS A, 48.2% remained AIS B, 15.8% improved to AIS C, and 4.4% transitioned to AIS D at five years. These results, both in terms of AIS grade improvement and regression, may represent small changes in motor/sensory function or inconsistencies in the sacral examination as opposed to significant neurological change. Among the 31.6% of persons who transitioned from AIS B to AIS A grade, it is unknown whether these conversions were the result of initial classification error or actual neurological deterioration.

Table 5.

Change in AIS Grade from Baseline to Follow-Up Assessment for Individuals with Initial AIS B Injuries (Heterogeneous Sample Including Tetraplegia and Paraplegia)

Authors/year	NLI	Baseline exam	Follow-up exam	Sample size (N)	AIS B to AIS A (%)	Remained AIS B (%)	AIS B to AIS C (%)	AIS B to AIS D (%)	AIS B to AIS E (%)
8Marino et al. 1999a	All levels	≤7 d	1 y	129	7.8	19.4	5.8	2.3	0
47Burns et al. 2003	All levels	≤48 h	1 y	15	0	40	40	20	0
20Spiess et al. 2009b	All levels	≤15 d	1 y	40	10.0	22.5	35	32.5	0
	Above T10			10	20	30	20	30	0
58Kirshblum et al. 2011a	All levels	Discharge (mean 97 d)	1 y	269	12.6	52.0	23.1	11.5	0.7
21Kirshblum et al. 2016a	All levels	≤ 30 d	1 y	56	10.7	35.7	32.2	21.4	0
	C1-–C8			35	11.4	40	25.7	22.9	0
	T1–T9			8	0	25.0	37.5	37.5	0
	T10 and below			13	15.4	30.8	46.1	7.7	0
39Marino et al. 2011a	Cervical levels	≤7 d	1 y	125	8.8	24.8	29.6	36.8	0
35Waters et al. 1992	T1–L3	≤30 d	Within 2 y	15		26.7%	26.7%	46.7%
48van Middendorp et al. 2009b	Above T11	≤15 d	6 m–1 y	37	5.4	21.6	35.1	35.1	2.7
62Harrop et al. 2011b	T4–T9	≤24 h	1 y	3	33.3	0	66.7	0	0
	T10–T12			2	0	100	0	0	0
	L1 and below			4	0	0	50	25	25
61Lee et al. 2016a	T2–T12	≤7 d	1 y	34	20.6	20.6	26.5	29.4	2.9
74Wilson et al. 2018c	T1–L1	≤7 d	6 m - 1 y	6	16.7	16.7	33.3	16.7	16.7

NLI, Neurological level of injury; AIS, American Spinal Injury Association Impairment Scale.

^aSpinal Cord Injury Model Systems database.

^bEuropean Multicenter Study about Spinal Cord Injury database.

^cNorth American Clinical Trials Network registry.

The modality and extent of sacral sparing may have prognostic value for patients with initial AIS B injuries, with the initial sparing of all sacral sensory components being correlated with the greatest conversion to motor incomplete status at one year.²¹ Before the sacral sparing definition of a neurological incomplete injury, Crozier and colleagues⁷² in a small study (N = 27) reported that PP sparing below the zone of injury at 72 h was a positive prognostic indicator for community ambulation. In a later study using the sacral sparing definition, Oleson and coworkers,⁷³ in the analysis of Sygen data, reported that although a higher percentage of patients with sacral PP sparing (39.4% vs. 28.3%) were ambulating at 26 weeks, the results did not reach significance.

Walking outcomes available at one year also were not significantly different in individuals with and without initial PP sparing (53.6% vs. 41.5%; p = 0.31). While ambulation was improved for patients with PP preservation in the lowest sacral segments four weeks post-injury, it was not based on S4–S5 PP preservation at the 72-h examination. In addition, lower extremity PP preservation in >50% of lower extremity dermatomes (L2–S1) at 72 h was predictive of future ambulation.⁷³

For persons with initial AIS B tetraplegia, 50–65% demonstrate recovery to motor incomplete status, while approximately 10% regress to AIS A and roughly 25–40% remain AIS B at one year post-injury.^21,39 As with complete injuries, most natural recovery in AIS B tetraplegia occurs during the first post-injury year, with 22.3% improving to motor incomplete status between one and five years.⁶⁴

For persons with initial AIS B paraplegia, the numbers are variable, but ∼60–70% improve to motor incomplete status by one year, up to ∼20% of patients with initial AIS B paraplegia regress to AIS A status, and up to 20% remain AIS B.^21,61,74 During the first year of injury, it is unknown whether the majority of AIS B to AIS A conversions was the result of initial classification error or decline in neurological function. Of those with AIS B paraplegia at one year post-injury, it was reported that 28.6% regressed to AIS A, 54.8% remained AIS B, 11.9% improved to AIS C, and 4.8% recovered to AIS D at five years post-injury.⁶⁴

Motor incomplete injuries (AIS C and D)

Approximately 15% of initial SCI cases are classified as AIS C, and roughly 20–30% are classified as AIS D.^20,21,48,59 The prognosis for AIS C recovery to AIS D is greater for patients with initial AIS C injuries than for those with initial motor complete lesions (AIS A and B). Overall, in studies that grouped all levels of injury, the majority of patients (up to 85%) with initial (≤30 days) AIS C classification improved to AIS D/E, with ∼10–20% remaining AIS C, and up to 8% regressing to AIS A/B at 6–12 month follow-up.^8,20,21,48 For those with initial AIS D, the majority (up to 100%) remain AIS D or improve to AIS E, with up to 2% regressing to AIS A/B by one year post-injury (Table 6).^8,20,21,48

Table 6.

Change in AIS Grade from Baseline to Follow-Up Assessment for Individuals with Initial Motor Incomplete Injuries (Sample Including Tetraplegia and Paraplegia)

Authors/year	NLI	Baseline exam	Follow-up exam	Initial AIS	Sample size (N)	AIS A at follow-up (%)	AIS B at follow-up (%)	AIS C at follow-up (%)	AIS D at follow-up (%)	AIS E at follow-up (%)
8Marino et al. 1999a	All levels	≤7 days	1 y	C	159	3.1	1.3	25.1	66.7	3.8
				D	72	0	0	1.4	94.4	4.2
20Spiess et al. 2009b	All levels	≤15 d	1 y	C	46	4.3	0	10.9	84.8	0
				D	49	0	1.7	0	89.8	8.5
	Above T10			C	9	0	0	22	78	0
				D	9	0	0	0	100	0
48van Middendorp et al. 2009b	Above T11	≤15 d	6 m–1 y	C	43	2.3	0	23.3	74.4	0
				D	32	0	0	0	84.4	15.6
21Kirshblum et al. 2016a	All levels (tetraplegia + paraplegia)	≤30 d	1 y	C	109	1.8	6.4	19.3	72.5	0
				D	123	0	0	0	100	0
	Tetraplegia			C	78	1.3	5.1	16.7	76.9	0
				D	94	0	0	0	100	0
	T1–T9			C	11	0	9.1	27.3	63.6	0
	T10 and below				20	5	10	25	60	0
	T1–T9			D	10	0	0	0	100	0
	T10 and below				19	0	0	0	100	0
62Harrop et al. 2011c	T4–T9	≤24 h	1 y	C	5	0	0	20	40	40
	T10–12				2	0	0	0	100	0
	L1 and below				6	0	0	16.7	66.7	16.7
	T4–T9			D	2	0	0	0	0	100
	T10–12				4	0	0	0	25	75
	L1 and below				15	0	0	0	0	100
39Marino et al. 2011a	Tetraplegia	≤7 d	1 y	C	109	0.9	2.8	13.8	80.7	1.8
				D	135	0	0.7	0.7	84.4	14.1
74Wilson et al. 2018d	T1–L1	≤7 d	6 m–1 y	C	8	0	0	0	50	50
				D	3	0	0	0	67.7	33.3
61Lee et al. 2016a	T2–T12	≤7 d	1 y	C	21	0	9.5	4.8	81.0	4.8
				D	16	0	0	0	87.5	12.5

Bolded lines represent all levels, with subdivisions of that study listed below.

NLI, neurological level of injury; AIS, American Spinal Injury Association Impairment Scale.

^aSpinal Cord Injury Model Systems database.

^bEuropean Multicenter Study about Spinal Cord Injury database.

^cSygen database.

^dNorth American Clinical Trials Network registry.

The modality and extent of sacral sparing have prognostic value for patients with initial AIS C injuries. The presence of VAC in combination with sparing of other sacral examination components was associated with the greatest percentage improvement to AIS D grade; 60–67% of patients with initial AIS C grade and preservation of VAC, DAP, and either LT or PP sensation at S4-S5 improved to AIS D by one year. This percentage increased to 87% when both LT and PP at S4-S5 were preserved.²¹ Conversely, sacral sparing of VAC alone had the poorest prognosis for recovery to AIS D.²¹

For persons with motor incomplete tetraplegia, approximately 80% of individuals with initial AIS C injuries will improve to AIS D grade from baseline assessment (within 30 days) to one year post-injury, with ∼1% converting to AIS A, 4% regressing to AIS B, and 15% remaining AIS C; almost all patients with initial AIS D tetraplegia will remain AIS D after one year^21,39 (Table 6). In terms of long-term outcomes of patients classified with AIS C tetraplegia (N = 58) at one year post-injury, 24.1% improved to AIS D, while 53.4% remained AIS C and 22.4% regressed to motor complete status at five years.⁶⁴ For those with AIS D tetraplegia (N = 94) at one year, 86.2% remained AIS D at five years, and 13.8% changed to AIS A-C grades.⁶⁴

For persons with motor incomplete paraplegia, the majority of individuals with initial AIS C grade improve to AIS D or E grades within one year of injury, and most patients with initial AIS D remain AIS D or recover to AIS E by one year^21,61,62,74 (Table 6). Regarding late neurological recovery for patients classified with AIS C paraplegia (N = 62) at one year, 19.4% improved to AIS D, 62.9% remained AIS C, and 17.7% regressed to motor complete status by five years.⁶⁴ For those with AIS D paraplegia (N = 83) at one year, 80.7% remained AIS D, 12.0% regressed to AIS C, and 7.2% became motor complete by five years.⁶⁴

Trends in Natural Recovery after Traumatic SCI: Motor Recovery

Similar to conversion from neurological complete to incomplete injury, the majority of motor recovery occurs within the first six to nine months, with the most rapid rate of recovery occurring in the first three months.^{8,36,40,56,60,75} Spontaneous motor recovery typically plateaus at 12–18 months post-injury, with the rate and extent of recovery being greater for incomplete lesions.^{35–38,56,75–77}

Generally, the degree of motor recovery varies according to the initial AIS grade in the following order: AIS C > B > D > A. Overall, patients with a neurological complete SCI (tetraplegia and paraplegia combined) have the least amount of motor score change in one year. Patients with grade D may experience less motor point recovery compared with AIS C and B injuries because of a ceiling effect (less possible improvement from baseline), but they have the highest final motor scores.^7,16,42,58

Complete tetraplegia

Motor score recovery in complete tetraplegia

In patients with initial neurological complete tetraplegia, motor change from initial assessment (usually ≤30 days) to one year after injury is ∼8–12 points,^{8,18,36,39,40,51,56,60,78} with a range from 8.2⁷⁸ to 14.0 points^18,56 (Table 7). Most motor recovery after complete cervical SCI occurs in the first 6–12 months.^40,77 For example, in evaluating the extent of natural motor recovery in patients with complete injuries at the C4–C7 levels, Steeves and associates⁴⁰ reported a significant increase in UEMS at each successive time point up to 26 weeks for both the Sygen and EMSCI database, and up to 52 weeks for the Sygen database.

Table 7.

Recovery of Upper Extremity Motor Score in Complete Tetraplegia

Authors/year	Timing of initial exam	Timing of follow-up exam	Sample size (N)	Mean motor point change
36Waters et al. 1993	30 d	1 y	39*/6	8.6 ± 4.7*/9.7 ± 7.3 (UEMS)
81Katoh and el Masry 1994	48 h	>1 y (mean 43.2 m)	40	9.2 ± 5.5 (UEMS)
8Marino et al. 1999a	≤7 d	1 y	NR	9.6 ± 12.7 (total motor score)
57Coleman et al. 2004b	≤72 h	26 w	317	9.6 (total motor score)
78Fisher et al. 2005	NR	37.8 ± 19.8 m	30	8.2 (UEMS)
56Fawcett et al. 2007
Sygenb	≤72 h	1 y	264	12.3 ± 13.7 (total motor score)
EMSCIc	≤4 w	1 y	NR	14 (total motor score)
60Curt et al. 2008c	40 days	48 w	91	11.5 (total motor score)
40Steeves et al. 2011 (C5-C7 only)
Sygenb	≤72 h	1 y	187	9.6 ± 4.3 (UEMS)
EMSCIc	≤7 d	50 w	51	11.4 ± 0.97 (UEMS)
39Marino et al. 2011a	≤7 d	1 y	315	8.8 ± 8.4 (UEMS)
51Evaniew et al. 2020d	≤48 h (mean 11.8 h)	Mean 208 ± 75.2 d	75	13.0 ± 14.1 (total motor score)

UEMS, upper extremity motor score; NR, not reported.

^aSpinal Cord Injury Model Systems database.

^bSygen database.

^cEuropean Multicenter Study about Spinal Cord Injury database.

^dRick Hansen Spinal Cord Injury Registry database.

^*Score represents UEMS change of persons who remained neurologically complete at one year versus the next score, which was the UEMS change for the six subjects who converted to incomplete status.

Fawcett and colleagues⁵⁶ documented motor recovery (from the Sygen database) from successive time points throughout the year, which is reported in Table 8. These data should be taken into account when there are large ranges of follow-up neurological data used during the first year. In a study that employed recursive partitioning instead of the AIS, Evaniew coworkers⁷⁹ found that those with initial motor scores of three or less had the least recovery.

Table 8.

Recovery of Motor Strength at One Year Assessed from Various Time Points

	From <72 h	From week 4	From week 8	From week 16	From week 26
Cervical AIS A	12.3 ± 13.7	8.0 ± 7.9	5.7 ± 6.2	2.8 ± 3.8	1.7 ± 3.5
Cervical AIS B	37.1 ± 27.8	20.2 ± 21.2	8.7 ± 12.3	6.3 ± 9.7	3.1 ± 6.6
Cervical AIS C	NR	38.0 ± 18.8	22.1 ± 16.5	9.5 ± 13.7	3.5 ± 8.4
Cervical AIS D	NR	25.5 ± 12.6	17.0 ± 11.8	8.6 ± 8.8	3.9 ± 8.6
Thoracic AIS A	4.5 ± 11.2	2.6 ± 8.1	0.9 ± 5.1	0.6 ± 3.4	-0.5 ± .8
Thoracic AIS B	22.7 ± 17.0	21.3 ± 18.6	11.1 ± 12.2	2.4 ± 6.2	0.8 ± 1.7
Thoracic AIS C	NR	31.0 ± 2 .0	17.9 ± 19.1	8.4 ± 9.8	1.0 ± 6.5
Thoracic AIS D	NR	14.1 ± 8.5	8.2 ± 6.5	3.9 ± 7.0	2.0 ± 7.6

Adapted from Fawcett et al.⁵⁶ based on the Sygen database. Please note the Benzel grades were converted to AIS grades. These numbers include all patients and not just the placebo group.

AIS, American Spinal Injury Association Impairment Scale; NR, not reported.

While the majority of motor recovery in patients with complete tetraplegia is because of gains in UEMS, some patients experience improvements in lower extremity strength. On review of Sygen data, Coleman and colleagues⁵⁷ observed that patients with complete tetraplegia (N = 317) had a mean improvement in LEMS of 2.2 points from initial assessment (within 72 h) to 26 weeks post-injury. Similarly, Marino and coworkers³⁹ reported that patients with AIS A tetraplegia experienced a mean change in LEMS of 3.3 ± 9.9 points during the first year.

Not all studies reported motor recovery in the same manner. For example, some studies reported on TMS change that incorporated the changes in UEMS as well as LEMS. In addition, the timing of the initial examination varied as well as the follow-up period. Some studies described UEMS changes at one year for all patients with an initial complete injury, regardless of whether these patients converted to neurological incomplete status or remained complete at the follow-up period. Waters and colleagues³⁶ separated the change in motor score based on those who converted (between the initial examination within 30 days and one year post-injury) versus those who did not convert, while most other studies did not make this distinction. These factors should be considered when evaluating reported changes or using motor recovery as an outcome measure in research studies.

Motor level changes in complete tetraplegia

Another important outcome measure is the change in motor level reported in studies using multiple databases. Approximately 60–70% of patients with complete tetraplegia will regain at least one motor level– (30-40% will improve one level and 20–30% will gain two or more levels³⁹), ∼5% will deteriorate one to three motor levels, and up to 35% will experience no change.^{36,39,40,75,77,80} Changes to the motor level definition, differences in timing of the initial examination, and distribution of motor function in the segment(s) below the initial motor level may account for variability in reported motor level recovery.

The likelihood of functional motor (strength of ≥3/5) recovery decreases as the number of levels below the motor level increases. For example, Fisher and colleagues⁷⁸ reported that 74% (10/27) of patients with complete tetraplegia experienced functional motor recovery at the first level below the initial motor level, whereas 16.7% (5/30) and only 3.7% (1/27) experienced functional motor recovery at the second and third levels, respectively. Waters and associates³⁶ reported that only 4% of “second zero muscles,” defined as the second most proximal key muscle with 0/5 strength, regained any measurable strength, and only 1% regained grade 3/5 strength at one year.

Motor ZPP in complete tetraplegia

The degree of strength found in the motor ZPP plays a prognostic role in recovery. Recovery to a motor grade of ≥3/5 (considered recovering functional strength) below the motor level is more likely and faster in those with baseline ≥1/5 strength, compared with myotomes with an initial 0/5 motor grade.^36,76,78 Seventy-five to 100% of muscles at the first level below the initial motor level will recover to ≥3/5 strength if they have 1–2/5 strength at initial assessment, whereas only 25–56% of muscles at this level will recover to ≥3/5 strength if their initial motor grade is 0/5.^36,76,78

For example, Fisher and colleagues⁷⁸ found that 56.3% of patients with an initial motor grade of 0/5 at the first level below the initial motor level recovered functional strength in that myotome (follow-up at 37.8 ± 19.8 months), whereas 100% of individuals with an initial motor score of 1–2/5 at that level recovered functional strength. They also reported that at the second level below the motor level, 16% (4/25) of patients with an initial motor score of 0/5 and 20% (1/5) of those with an initial score of 1–2/5 recovered functional strength at that level.⁷⁸

The length of the motor ZPP is also a positive predictor of motor recovery.³⁹ Marino and associates³⁹ reported that patients with complete tetraplegia are five times more likely to gain ≥2 motor level segments if the length of the motor ZPP is at least two levels relative to patients with a motor ZPP of 0–1 level (relative risk = 5.0; p < 0.001) (Fig. 1).

Effect of sensory preservation in tetraplegia

Several studies have found sensory preservation to be predictive of motor recovery in patients with complete tetraplegia.^36,81,82 Waters and coworkers³⁶ observed an increased likelihood of first zero muscles recovering to ≥3/5 strength at one year if there was sensory preservation in the associated dermatome on initial assessment than if there was absent sensation at that spinal level. Katoh and el Masry⁸¹ evaluated zonal recovery, defined as motor improvement from an initial grade of 0/5 to ≥2/5, and reported that more individuals with complete tetraplegia experienced zonal recovery if they had initial sensory sparing at the corresponding spinal level than individuals without initial sensory preservation (63% vs. 33.3%, respectively).

Poynton and associates⁸² reported on the positive impact of initial PP sparing (intact or impaired) in the corresponding dermatome of upper extremity key muscles with an initial grade of 0/5 with motor recovery to ≥3/5 strength in patients with complete tetraplegia (N = 18) at a mean period of 29.6 months, with recovery of 77% (20/26) of the levels relative to only 1.3% (3/235) of levels without initial PP sparing.

Effect of initial motor level of injury in tetraplegia

In patients with complete tetraplegia, the initial motor level (e.g., C4 vs. C5, or other levels through C7) has not been shown to impact UEMS improvement significantly^40,80,83 or motor level recovery at longer-term follow-up.⁴⁰ The UEMS recovery is similar for patients with injuries at C7 and those with more rostral injuries, despite the C7 level having fewer cervical segments available for motor recovery. These findings suggest that spontaneous motor recovery occurs predominantly within the first two cervical segments caudal to the initial motor level.^40,83

In a study by Dvorak and colleagues,⁸³ analyzing the magnitude of early motor recovery (at a mean of 130 days) after acute traumatic SCI using both initial NLI (C1–C4 vs. C5–T1) and severity of the injury, they found no significant differences for TMS changes, although a significant difference in change in UEMS between high and low cervical levels in AIS B (14.3 vs. 9.5), C (20.5 vs. 12.7), and D (11.4 vs. 7.5) was observed. It is unclear whether these gains changed at a later time point of follow-up.

Studies in a heterogeneous group of AIS A and B

A number of studies reported on motor recovery from a heterogeneous group of patients with cervical level motor complete injuries without differentiating between initial AIS A and B grades.^{75,77,84–87} These findings include the following:

• 1.The more caudal the cervical level, the higher percentage of recovery; those with an initial motor level of C5 recovering to C6 (75%) and C6 to C7 (85%) was greater than the recovery of C4 to C5 (70%).⁷⁵
• 2.The initial biceps (C5 motor) strength was a positive predictor of functional wrist extensor (C6 motor) recovery.^84,87 All patients with an initial biceps strength of ≥3/5 regained wrist extensor strength on that side by one year, whereas a small percentage of those with strength of 0–2/5 gained wrist extensor strength to this level.
• 3.C5 PP sensation was a predictor of C6 motor recovery, because 93% of individuals with some (impaired or normal) C5 PP sensation recovered to ≥3/5 strength at C6, while 78% of patients with absent C5 PP sensation failed to recover functional C6 strength.⁸⁷
• 4.Early motor recovery predicted recovery of functional (≥3/5) strength in first zero grade muscles with the myotome directly rostral having ≥3/5 strength in patients with motor complete tetraplegia (C4–C7).⁸⁶ While overall recovery to functional strength was 43% by one year, 86% who recovered to ≥1/5 strength by one month recovered, relative to 9% who did not achieve ≥1/5 strength by one month. Similarly, 100% of those who achieved ≥2/5 strength by three months recovered to ≥3/5 strength by one year, while no patients recovered functional strength at one year if they did not achieve ≥2/5 strength by three months.
• 5.Muscles within the ZPP that had an initial 1–2.5/5 motor grade recovered to ≥3/5 earlier, and to a greater extent, than muscles with an initial 0/5 grade.⁷⁷ In those with some initial function versus no initial function in the ZPP, 68% versus 14% and 82% versus 35% recovered to ≥3/5 strength, at three and six months post-injury, respectively. In addition, 90% versus 45% of muscles with initial 1–2.5/5 strength versus initial 0/5 motor grade recovered to a motor grade of ≥3/5 by 9–12 months.⁷⁷
• 6.Recovery for muscles with an initial 1–2.5/5 grade plateaued by ∼one year, whereas muscles with an initial 0/5 grade continued to demonstrate a significant improvement in motor scores between nine and 24 months, and motor recovery in this group did not reach a plateau by 24 months.⁷⁷
• 7.The initial muscle strength of the myotome immediately below the motor level is a significant predictor for time to recovery to grade 3 and 4/5 at six months (Mange and coworkers.⁸⁵) Those muscles with an initial (between 3–7 days post-injury) strength of 2 or 2+/5 improved to grades 3 and 4/5 six months more quickly than those muscles whose initial strength was 1 or 1+/5 (median times were 0.5 months vs. three months and three months vs. 6sixmonths, respectively).

Most of these studies had small sample sizes and comparisons of outcomes for those with motor and sensory complete (AIS or Frankel A) versus sensory incomplete/motor complete (AIS or Frankel B) injuries were not performed. As such, caution should be taken in using these data for contemporary studies with individuals classified as AIS A.

Complete paraplegia

For patients with complete paraplegia, motor recovery is overall limited, with the level of injury playing an important role^{35,41,57,61,78,83,88} (Table 9). One factor contributing to motor improvement in paraplegia is the absence of testable myotomes in the thoracic region. Lumbar injuries have the greatest prognosis for LEMS recovery, followed by mid-lower thoracic lesions, with the poorest prognosis for high thoracic injuries (T2–T6).^{7,35,41,61,63,89}

Table 9.

Average Lower-Extremity Motor Score Improvement in Persons with Complete Paraplegia within One Year

Authors/year	NLI	Timing of initial exam	Timing of follow-up exam	Sample size (N)	Mean LEMS change
35Waters et al. 1992	T2–L2	30 d	1 y	142*	T2–T8 = 0* T9–L2 = 7.8*
8Marino et al. 1999a	Paraplegia	≤7 d	1 y	NR	2.6 ± 7.3
57Coleman et al. 2004b	Thoracic	≤72 h	26 w	143	0.6
78Fisher et al. 2005	T1–L2	NR	37.8 ± 19.8 m	40	1.3
56Fawcett et al. 2007b	Thoracic	≤72 h	1 y	133	4.5 ± 11.2
60Curt et al. 2008c	Paraplegia	≤40 d	48 w	123	3.0
88Pouw et al. 2011	T2–T11	0–40 d	6 m–1 y	61	2.3
41Zariffa et al. 2011c	T2–T12	≤30 d	48 w	122	2.76 ± 7.12
	T2–T5			34	0.09 ± 0.51
	T6–T9			41	2.02 ± 6.67
	T10–T12			47	5.34 ± 9.05
61Lee et al. 2016a	T2–T9	≤7 d	1 y	194	1.5 ± 7.5
	T10–T12				4.1 ± 8.4
74Wilson et al. 2018d	T1–L1	≤7 d	1 y	22	1.5
			> 1 y	13	5.1
			Last follow-up (mean 8 m)	42	2.5

Bolded lines represent all levels, with subdivisions from that study listed below.

NLI, neurological level of injury; LEMS, lower-extremity motor score; NR, not reported.

^aSpinal Cord Injury Model Systems database.

^bSygen database.

^cEuropean Multicenter Study about Spinal Cord Injury database.

^dNorth American Clinical Trials Network registry.

^*-Score represents LEMS change of persons who remained neurologically complete at one year.

Aimetti and associates⁶³ reported that most patients with complete thoracic SCI (T2–T12) had <5 motor point improvement with only 10.6% of patients demonstrating a LEMS improvement >10 points; of those who achieved >10 point motor recovery, 77.8% were in the T10–T12 subgroup. Both Aimetti and associates⁶³ and Zariffa and coworkers⁴¹ observed that for all NLI subgroups (i.e., upper, mid, and low thoracic), the median change in LEMS was 0. Similarly, Lee and colleagues⁶¹ found that among patients with complete thoracic injuries, only those with an initial sensory level of T12 demonstrated a median change in LEMS >0 (median change of 3.5 points). For patients with complete paraplegia, recovery of motor strength at one year, starting at varied time points, is reported in Table 8.

Waters and colleagues³⁵ reported that only patients with a level of injury at and below T9 experienced LEMS recovery and described a relationship between the initial strength of lower extremity muscles and motor recovery at one year. The greater the initial strength of the hip flexors, the greater the chance of recovery of that muscle to >3/5 at one year. Approximately 7% (9/121) of hip flexors having an initial strength of 0/5 regained strength to ≥3/5 at one year relative to 68% (13/19) with an initial strength of 1–2/5.³⁵ Similarly, 3% (4/119) of muscles with an initial knee extensor strength of 0/5 regained strength to ≥3/5 at one year relative to 70% (16/23) with an initial strength of 1–2/5.³⁵

In addition, there was a relationship between initial strength of an individual muscle and recovery of adjacent muscles at one year (e.g., initial lower abdominal muscles to the recovery of hip flexors at one year and initial hip flexors to recovery of knee extensors at one year). For example, only 13% (N = 16/121) of cases with initial hip flexor strength of 0/5 regained knee extensor strength of ≥1/5 at one year, as opposed to 100% of cases (N = 21) with an initial hip flexor strength of >1/5. Further, the initial strength of the hip flexors may also predict functional motor recovery of the knee extensors at one year, as 42% (8/19) and 100% (2/2) of individuals with an initial hip flexor grade of 1–2/5 and 3/5, respectively, regained knee extensor strength to ≥3/5.³⁵

Similar to acute tetraplegia, Poynton and coworkers⁸² found a positive impact of initial PP sparing (intact or impaired) in the corresponding dermatome of the lower extremity key muscle with an initial grade of 0/5 on motor recovery to >3/5 strength in patients with complete paraplegia (N = 19), with recovery of 100% (6/6) of the levels relative to 0% (0/180) of levels without initial PP sparing.

Incomplete injuries

AIS B

Although there is a wide variation in motor recovery reported for patients with initial AIS B tetraplegia, overall recovery by one year is greater than for patients with initial AIS A injuries. Using Sygen data (control + treated groups), mean motor score change for patients with initial AIS B tetraplegia from initial examination (within three days of injury) has been reported to be 31.7 at 26 weeks post-injury⁵⁷ and 37.1 ± 27.8 at one year.⁵⁶ An earlier study by Marino and associates⁸ reported a mean motor score change of 28.2 from baseline examination (within seven days) to one year follow-up; they also found a median motor score change of 20 with an interquartile range of 7–49, which suggests individuals with AIS B tetraplegia have the most varied motor recovery relative to patients with other AIS grades.

Curt and coworkers⁶⁰ reported less overall motor recovery, with a mean change of 19.8 from initial assessment (within the first month) to the follow-up examination at 48 weeks post-injury. Marino and colleagues³⁹ distinguished between initial UE and LE motor recovery in patients with initial AIS B tetraplegia and reported an average change of 13 points in UEMS and 17.7 points in LEMS from initial assessment to one year.

In patients with initial AIS B paraplegia, the mean change in LEMS is between 12.6–22.7 from baseline assessment to one year follow-up.^8,56,60,74 Coleman and colleagues⁵⁷ (using the Sygen database) found that for individuals with thoracic AIS B injuries, the average change in LEMS was 18.9 at 26 weeks post-injury, and Fawcett and associates⁵⁶ reported a change of 22.7 ± 17.0 at one year.

Initial PP sparing (intact or impaired) in the corresponding dermatome of the UE key muscles with an initial grade of 0/5 remained a strong predictor of motor recovery to ≥3/5 strength at follow-up in patients with incomplete injuries (incomplete tetraplegia N = 17, and; incomplete paraplegia N = 5), with the recovery of 92.1% (58/63) and 68.4% (13/19) of the levels relative to 3.9% (2/51) and 7.7% (1/13) of levels without initial PP sparing, respectively.⁸²

The modality and extent of sacral sparing have prognostic value for patients with initial AIS B grade, with sparing of all sensory sacral components, including LT, PP, and DAP sensation, being associated with the greatest prognosis for motor recovery.²¹ Waters and colleagues³⁷ in a small study sample (N = 13) of patients with AIS B tetraplegia, reported that those without sacral sparing of PP sensation (N = 5) (despite having preserved bilateral S4–S5 LT sensation) did not experience any LEMS improvement. In contrast, all patients with AIS B tetraplegia and sacral sparing of PP sensation (N = 8) experienced some LEMS recovery.³⁷

AIS C and D injuries

Compared with motor complete injuries, there is greater motor recovery after initial motor incomplete SCI. For motor incomplete tetraplegia, the reported mean change in TMS is between 22–52 points from baseline assessment to follow-up evaluation six to 12 months post-injury.^{8,39,56,57,60} Marino and associates³⁹ reported a mean change of 22.6 in UEMS and 23.3 in LEMS for patients with initial AIS C tetraplegia, from baseline assessment (≤7 days after injury) to one year post-injury. During this same period, changes in motor scores were lower for patients with initial AIS D tetraplegia, because of the ceiling effect, with a mean change in UEMS of 16.1 and mean change in LEMS of 7.1. For patients with initial motor incomplete paraplegia, mean changes in LEMS range from 9.0–34.3, from initial examination to follow-up assessment six months to one year after injury.^{8,56,57,60,74} Recovery of motor strength at one year starting at varied time points for incomplete paraplegia is found in Table 8.

Trends in Natural Recovery after Traumatic SCI: Sensory Recovery

There are a limited number of articles describing sensory recovery after complete tetraplegia. Based on Sygen data, mean changes in PP and LT scores from baseline assessment (within 72 h) were 8.3 and 13.8 at 26 weeks⁵⁷ and 9.5 ± 19.7 and 15.0 ± 23.3 at one year follow- up.⁵⁶ Curt and colleagues,⁶⁰ using EMSCI data, reported a mean change of 5.4 and 13.0 for PP and LT, respectively, at 48 weeks post-injury, assessed within the first month (57% had an admission examination within 15 days, while the remainder had one between 16–40 days).

A study by Eschbach and colleagues⁹⁰ evaluated PP and LT sensory level recovery in patients with C4–C8 Frankel A tetraplegia and found that 30% of patients (N = 19) with an initial 0–1/2 sensory score in the first level below the initial sensory level improved one PP level and 29% improved one LT level from initial examination (within one week) to follow-up assessment (ranged from 3–24 months). All reported sensory level improvement occurred within three months. Interestingly, when they excluded patients with C4 level injuries (12 patients had a C4 PP level and 11 had a C4 LT level) from the analysis, 83% of C5–C8 levels recovered one PP level and one LT level.

For patients with initial complete paraplegia, the mean sensory score improvement is between 1–5 points each for PP and LT within the first year.^35,41,57,88 Zariffa and coworkers⁴¹ reported greater improvement in the T6–9 group relative to T2–T5 (mean changes in PP of 2.50 ± 8.6 versus -0.29 ± 6.21 and in LT of 4.04 ± 10.22 versus 1.34 ± 5.48). In an earlier study, however, Waters and colleagues³⁵ when reporting only on patients who remained neurologically complete at one year, did not find any difference between the T9–11 levels as opposed to the T2–8 group (mean changes in PP of 1.42 ± 2.03 and 1.57 ± 2.41 and in LT of 1.55 ± 2.27 and 2.03 ± 2.57). Waters and colleagues³⁵ also reported that among the 6/148 who converted to an incomplete injury over the first year, mean change in PP was 6.83 ± 5.67 and LT 9.67 ± 3.93.

Because there is limited change in sensory score in complete paraplegia, the sensory level (and corresponding NLI) does not change significantly after neurologically complete paraplegia between T2–L1, as most individuals (∼90%) remain within 1–2 levels of their initial level at one year.^{35,41,61–63} Between 35–70% of individuals with initial paraplegia remain at their initial level of injury, 24–32% deteriorate, and 33–40% improve their level of injury. Only a small percentage of patients (∼12%) gain or lose ≥2 levels, however.^{35,41,61–63}

Overall, while sensory recovery information in clinical trials is often reported, further analysis of the functional benefit of sensory recovery would be important.

Incomplete injuries

Individuals with initial motor complete, sensory incomplete tetraplegia (AIS B) regain significantly more sensory function than patients with initial AIS A. Using the Sygen database, for patients with tetraplegia, a mean improvement of 29.6 and 33.0 ± 26.4 points in LT sensation and 27.5 and 31.9 ± 27.4 points in PP sensation, at six months and one year, respectively, have been reported.^56,57 Curt and colleagues,⁶⁰ using the EMSCI database, observed less improvement in sensory function from initial examination (<1 month) to one year follow-up, with a mean change in LT score of 10.2 and average change in PP score of 13.4.

Using the Sygen database for patients with AIS B paraplegia, a mean improvement of 17.1 and 23.2 ± 25.7 points in LT sensation and 17.5 and 20.3 ± 22.0 points in PP sensation, at six months and one year, respectively, have been reported.^56,57 Less improvement in sensory scores was reported by Curt and colleagues,⁶⁰ with a mean improvement in LT and PP scores of 3.6 and 1.9 and from the initial examination (within the first month) to 48-week follow-up.

In patients with initial motor incomplete injuries, there is a wide range in reported sensory score changes. For patients with AIS C or D tetraplegia, the reported mean change in LT score varies between 6.9–24.7 points, while the reported mean change in PP score varies from 5.0–33.3 points from baseline to follow-up assessment (26–48 weeks post-injury).^57,60 For those with AIS C or D paraplegia, reported mean changes in LT and PP scores vary from 1.3–19.1 and 0.7–26.8 points, respectively.^57,60

In terms of sensory level changes, for those with initial thoracic AIS C injuries, 23.8% remained at their baseline sensory level, 57% gained at least one sensory level, and 19% lost at least one sensory level.⁶¹ Similarly, 19% of patients with thoracic AIS D injuries remained at their baseline sensory level, 63% gained at least one sensory level, and 19% lost one or more sensory levels.⁶¹

Age and Impact on Neurological Recovery

Most studies evaluating the impact of age on neurological and functional recovery have shown that younger age at time of injury is an important positive predictor.^{57,61,91–93} Only a few studies considered ages below 50. Pollard and coworkers⁹⁴ suggested age <18 years may be associated with increased motor recovery in patients with incomplete tetraplegia, while others evaluated age 30 and below as a positive predictor of recovery but reported conflicting data.^57,93

There is stronger literature for age >50 being associated with poorer outcomes.^{61,92,95–102} Oleson and associates⁹² (in patients with AIS B grade) and Penrod and colleagues⁹⁷ (in patients with motor incomplete injuries) found that, overall, patients age >50 years had a decreased likelihood of recovering walking ability one year post-injury, relative to patients under age 50. Burns and associates⁹⁸ reported an increased likelihood of achieving community ambulation in patients with incomplete tetraplegia and paraplegia on discharge from rehabilitation in individuals with AIS C who were <50 (91%) compared with those ≥50 years of age (42%), although for individuals presenting with an initial AIS D, age was not a factor in ambulation because all were able to ambulate 3–6 months post-injury. Lee and coworkers⁶¹ also reported less improvement for patients older than age 50 in thoracic level injuries.

The age of 60 years has also been studied and found to have a negative correlation in motor score recovery over time for patients with central cord syndrome.⁹⁹ Cifu and colleagues¹⁰⁰ reported that individuals with injury at age ≥65 had less motor score change and functional recovery relative to younger individuals at discharge from rehabilitation. Aito and associates¹⁰¹ observed a negative relationship between age >65 and neurological or functional outcome in patients with traumatic cervical SCI who had central cord syndrome. Two algorithms studying prognosis for ambulation used age 65 as the cutoff^103,104 for which recovery was dramatically impacted by this difference. In a more recent study, Engel-Haber and colleagues⁹⁶ applied these algorithms with SCIMS data and found that age 50 may be the more appropriate cutoff.

Not all studies have shown age to be a significant predictor of neurological recovery but have instead found similar ASIA motor score changes between younger and older age groups.^{61,62,105–107} Kramer and colleagues⁸⁰ did not find a significant association between age and motor level recovery in patients with complete tetraplegia. Jakob and coworkers¹⁰⁶ reported that age was not correlated with neurological recovery but was negatively associated with the change in SCIM between six and 12 months after injury, especially for those with complete paraplegia.

Harrop and coworkers⁶² evaluated factors related to the neurological improvement of thoracic, thoracolumbar, and lumbar SCIs and reported that age was not significantly related to neurologic improvement. Lee and associates⁶¹ reported that functional outcomes at one year decreased with increasing age, although no significant relationship was found between age and walking ability in thoracic SCI. Furlan and colleagues¹⁰⁷ reported that individuals ≥65 years of age did not significantly differ from younger patients in terms of neurological recovery at follow-up (6 and 12 months). While Wilson and associates⁹¹ found that older patients (≥65 years of age) demonstrated decreased functional independent measure motor scores at follow-up, this was dependent on the admission AIS grade. Age played a larger role for patients with initial AIS B and AIS C and was less prominent for AIS A and AIS D patients.⁹¹

Overall, older age seems to have a negative impact on neurological and functional recovery after SCI, most specifically in patients with a neurological complete injury, although the specific age is unclear.

Penetrating versus Blunt Trauma and Impact on Neurological Recovery

Penetrating (violence related) injury is more likely to lead to a classification of a neurological complete injury compared with blunt trauma.^7,8,61,108 While some studies have reported that penetrating injury does not impact the rate of conversion to an incomplete injury,^61,109 other reports suggest that relative to blunt trauma, patients with complete SCI because of penetrating injuries are also less likely to experience AIS improvement at one year follow-up.^7,108 Of note, those with initial incomplete injuries do not seem to have a difference in neurological recovery.^7,8,108

Gender and Impact on Predicting Outcome

Experimental studies in animal models have shown differences in neurological and functional recovery based on gender that favors females.^110,111 While the underlying mechanisms for these differences remain controversial, several studies have proposed hormonal involvement, including estrogen and/or progesterone.^112–114 The majority of clinical studies have not found gender to have an effect on neurological improvement over time.^{61,62,78,94,115–118} Sipski and colleagues¹¹⁹ reported, however, that females exhibit greater neurological recovery in ASIA motor scores from an initial examination (within 30 days of injury) to one year, in both neurologically complete and incomplete injuries, with no statistically significant difference in AIS conversion.

Overall, while the majority of published data do not report gender to have a major effect on neurological recovery after SCI, there is some evidence that this might occur, and further study is warranted.

Conclusion

This review has detailed the previous literature using the ISNCSCI (including the examination and classification using the AIS) to predict outcome after traumatic SCI (Table 10). Although the neurological recoveries, in particular from complete to incomplete status, show some variation between sources, the effects of examination timing, level, and ZPPs on recovery appear reliable. Other methods that may provide information about the severity of the initial neurological injury and inform prognostication for neurological recovery after traumatic SCI include neuroimaging, neurophysiological testing, as well as blood and cerebrospinal fluid biomarkers. Unbiased recursive partitioning may improve prediction of neurological outcome that can be used for clinical trials,^79,120 but may be sensitive to timing of early baseline evaluation.⁷⁹ As additional data become available, with potentially pooling of databases to allow for greater analysis, there will be further understanding of the prognosis of traumatic SCI.

Table 10.

Summary Findings from This Review

International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) and American Spinal Injury Association Impairment Scale (AIS)

• Formal training before utilizing the ISNCSCI and AIS, and the use of a computational algorithm, is recommended to reduce classification errors.

• Examinations performed within 24 h are reliable in the absence of factors interfering with a patient's cognition and communication. Results from very early examinations (within 4 h of injury) should be interpreted with caution.

AIS conversion and change

• ∼30% of cervical level injuries initially classified (within 30 days of injury) as AIS A convert to incomplete status, with roughly one-half improving to AIS B and the other half to motor incomplete status.

• There is a probable trend of higher rates of neurological conversion in more recent literature compared with historical studies.

• Conversion from neurological complete to incomplete injury in tetraplegia is greater than with paraplegia.

• In complete paraplegia, the likelihood of conversion largely depends on the level of injury, with lumbar injuries having a greater prognosis relative to thoracic lesions, with high thoracic (T2–T6) injuries having the poorest prognosis.

• A sensory zone of partial preservation (ZPP) of ≥3 segments plays a prognostic role in AIS conversion for persons with an initial neurological level of injury (NLI) between T6–T12.

• Persons with AIS B (tetraplegia and paraplegia) have a greater likelihood of improving to motor incomplete status at one year than those with initial AIS A.

• The modality and extent of sacral sparing has prognostic value in neurological recovery over one year. In AIS B, initial sparing of all sacral sensory components is associated with improved conversion to motor incomplete status. In AIS C, initial voluntary anal contraction (VAC) in combination with other sacral examination components is associated with the greatest percentage of improvement to AIS D grade, with sacral sparing of VAC alone having the poorest prognosis for recovery to AIS D.

Motor recovery in one year

• The majority of motor recovery occurs within the first six to nine months, with the most rapid rate of recovery occurring in the first three months after injury.

• Spontaneous motor recovery typically plateaus around 12–18 months post-injury, with the rate and extent of recovery being greater for incomplete lesions.

• Motor change in persons with initial AIS A tetraplegia from initial assessment (usually <30 days) is ∼8–12 points.

• In AIS A tetraplegia, ∼65% of persons will regain at least 1 motor level (20–30% 2 or more levels); ∼5% will deteriorate 1–3 motor levels, and up to 30% will experience no change.

• The initial strength, as well as the amount of levels of the motor ZPP, play a prognostic role in the number of motor levels recovered after cervical complete tetraplegia.

• In neurological complete paraplegia, the change in motor level is usually small with the level of injury playing an important role; lumbar lesions having greater recovery followed by low thoracic, with high thoracic levels having the least chance of lower-exremity moor score (LEMS) improvement.

• Total motor recovery in the upper and lower extremities is greater for persons with initial AIS B than AIS A tetraplegia and greater after initial motor incomplete spinal cord injury (SCI) than with motor complete injuries.

• The reported mean change in total motor score in motor incomplete (AIS C) tetraplegia is between 22–52 points from baseline to 6–12 months post-injury, usually divided between upper-extremity motor score (UEMS) and LEMS.

• Changes in motor scores are lower for persons with initial AIS D tetraplegia because of a ceiling effect.

• For persons with initial motor incomplete paraplegia, mean changes in LEMS range from 9.0–34.3 from initial examination to follow-up assessment 6–12 months post-injury.

Sensory recovery in first year

• Mean changes in pinprick (PP) and light touch (LT) scores from baseline assessment in tetraplegia are highly variable.

• For persons with initial complete paraplegia, the mean sensory score improvement is between 1–5 points each for PP and LT.

• The sensory level (and corresponding NLI) does not change significantly after neurologically complete paraplegia between T2–L1, because most individuals (∼90%) remain within 1–2 levels of their initial level. This includes ∼1/3 of patients each losing ≥1 level, staying the same, and regaining ≥1 level.

• Individuals with initial sensory incomplete tetraplegia (AIS B) regain significantly more sensory function than a person with initial AIS A.

Other factors

• Older age seems to have a negative impact on neurological and functional recovery after SCI. The specific age (whether >50 or >65 years) and the true nature of this impact on neurological recovery, however, is unclear.

• Penetrating (violence related) injury is more likely to lead to a classification of a neurological complete injury compared with blunt trauma and may reduce the likelihood of AIS conversion.

• There are insufficient data to support gender having a major effect on neurological recovery after SCI.

Authors' Contributions

All authors (SK, BS, FE, JG) were involved in the writing and editing of the manuscript.

Funding Information

This study was supported in part by a grant from the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR grant no. 90SI5026).

Author Disclosure Statement

No competing financial interests exist.