Early Trauma Indicators and Rehabilitation Outcomes in Traumatic Spinal Cord Injury

Abstract

Objectives:

To investigate the relationship between early trauma indicators and neurologic recovery after traumatic SCI using standardized outcome measures from the ISNCSCI examination and standardized functional outcome measures for rehabilitation populations.

Methods:

This is a retrospective review of merged, prospectively collected, multicenter data from the Spinal Cord Injury Model Systems (SCIMS) database and institutional trauma databases from five academic medical centers across the United States. Functional status at inpatient rehabilitation discharge and change in severity and level of injury from initial SCI to inpatient rehabilitation discharge were analyzed to assess neurologic recovery for patients with traumatic SCI. Linear and logistic regression with multiple imputation were used for the analyses.

Results:

A total of 209 patients were identified. Mean age at injury was 47.2 ± 18.9 years, 72.4% were male, 22.4% of patients had complete injuries at presentation to the emergency department (ED), and most patients were admitted with cervical SCI. Mean systolic blood pressure (SBP) was 124.1 ± 29.6 mm Hg, mean ED heart rate was 83.7 ± 19.9 bpm, mean O2 saturation was 96.8% ± 4.0%, and mean Glasgow Coma Scale (GCS) score was 13.3 ± 3.9. The average Injury Severity Score (ISS) in this population was 22.4. Linear regression analyses showed that rehabilitation discharge motor FIM was predicted by motor FIM on admission and ISS. Requiring ventilatory support on ED presentation was negatively associated with improvement of ASIA Impairment Scale (AIS) grade at rehabilitation discharge compared with AIS grade after initial injury. Emergency room physiologic measures (SBP, pulse, oxygen saturation) did not predict discharge motor FIM or improvement in AIS grade or neurological level of injury.

Conclusion:

Our study showed a positive association between discharge FIM and ISS and a negative association between ventilatory support at ED presentation and AIS improvement. The absence of any significant association between other physiologic or clinical variables at ED presentation with rehabilitation outcomes suggests important areas for future clinical research.

Introduction

Traumatic spinal cord injury (SCI) is associated with significant limitations in functional status and effects on quality of life.¹ Therefore, it is imperative that clinicians taking care of patients affected by SCI understand individual factors that predict functional outcomes. This ability allows for prognostication and planning of rehabilitation and community reintegration² and offers the potential for development of interventions to change modifiable factors. In prior research, the impact of sociodemographic factors, American Spinal Injury Association Impairment Scale (AIS) grade scores denoting severity of injury, and hospital/rehabilitation course factors (e.g., surgical management and length of hospitalization) on functional outcomes in traumatic SCI have been studied.³ On the other hand, the role of factors at or just following the time of injury have not been extensively examined.

The most common causes of traumatic SCI are motor vehicle crashes, followed by falls, violence, and sports activities.⁴ Following these events, the first point of care for most individuals with spine trauma is with emergency medical services at the scene of injury and trauma center emergency rooms. The goal during this period is overall hemodynamic stabilization and triage to the next appropriate provider. Many physiologic factors during this acute time window have been shown to predict overall mortality. For example, hypotension is well-established as a predictor of increased mortality in the trauma setting. In an analysis of the National Trauma Data Bank (NTDB), a national database of clinical trauma-related information from over 700 hospitals in the United States,⁵ mortality was 4.8% greater for every 10 mm Hg decrement in initial SBP below 110 mm Hg.⁶ Additionally, abnormal respiratory measures of hypoxia, apnea, and tachypnea in the acute trauma setting are associated with higher mortality.^7–9

We hypothesize that these acute physiologic measures also influence neurologic recovery and functional outcomes among patients with SCI. In traumatic SCI, immediate mechanical forces result in direct damage to the spinal cord and its vasculature. Decreased oxygen delivery results in further cellular injury, most severely at the injury epicenter but also at the penumbra or surrounding at-risk regions. The initial phase of injury is followed by a prolonged period of secondary injury. This phase encompasses cellular processes including inflammation, apoptosis, oxidative stress, and demyelination that lead to further cellular injury, both at the epicenter of injury and penumbra.^10,11 Given the importance of cord oxygenation during both phases, systemic physiologic parameters immediately following injury, such as blood pressure, respiratory rate, and oxygen saturation, likely impact the extent of cord injury and may influence functional outcomes.

The aim of this study was to test the hypothesis that acute physiologic parameters shortly after injury predict neurologic recovery and functional outcomes in traumatic SCI. Richard-Denis et al. recently conducted a systematic review of the literature examining early clinical predictors of SCI functional outcome measures and found a dearth of published literature examining acute physiologic data in traumatic SCI.³ Few studies reviewed included acute trauma-related factors in their analyses, and physiologic parameters at the time of injury, including heart rate, respiratory rate, and need for acute intubation, were not examined. We aimed to study the predictive value of these parameters using data from the NTDB and the National Spinal Cord Injury Model Systems (SCIMS) database, the largest available database on SCI outcomes.¹² By merging these databases, we aimed to examine the impact of these acute trauma variables on neurologic recovery and functional outcomes.

Material and Methods

Data source

The SCIMS is a national project funded by the US Department of Health and Human Services National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR). The project is the first prospective, multicenter study of recovery and outcomes in patients with SCI following participation in a coordinated system of acute hospital care and SCI rehabilitation. Institutions that are designated as SCIMS centers must have access to level 1 trauma center care, acute spine surgery care, comprehensive inpatient rehabilitation services, and long-term interdisciplinary follow-up and rehabilitation services. Currently there are 14 SCIMS centers and five additional follow-up centers across the United States. SCIMS database inclusion criteria are (1) traumatic SCI that is associated with a motor or sensory deficit, (2) admission to SCIMS center within 1 year of injury, (3) acute hospital care and inpatient rehabilitation received within a participating SCIMS center, and (4) informed consent signed by the patient and/or guardian. More information is available at https://www.nscisc.uab.edu/nscisc-database.aspx.

Data for this project were generated by merging patient data from SCIMS databases to acute hospital trauma registries from five separate trauma regions across the United States. All participating SCIMS centers in this project were also designated level 1 trauma centers by the American College of Surgeons (https://www.facs.org/search/trauma-centers). Level 1 trauma centers provide data on all traumatic injuries for inclusion into the NTDB (https://www.facs.org/quality-programs/trauma/tqp/center-programs/ntdb/about). The data collected for the trauma module reflect the data collected for the NTDB by each center’s Hospital Trauma Registry Team.

A deterministic matching method for merging each center’s trauma data (TD) with their corresponding SCIMS data was employed. Each participating SCIMS center acquired all TD for SCI patients from October 2016 through December 2018. Each participating SCIMS center organized the data into a new shared database that merged patient data from the trauma registry to patient data in the SCIMS database. At each center, the data manager assigned the patient’s corresponding SCIMS Patient ID and System ID as the identifiers for matching and then merged each patient’s SCIMS data to each patient’s trauma registry data. This new merged database was then uploaded by each SCIMS center to a new database at the SCIMS National Statistical Center. The combined database from all participating centers was then downloaded to the lead SCIMS center. The lead center made the combined trauma database available to the SCIMS participating centers. Data use agreements with the lead center were completed prior to sending out the combined trauma database.

Eligibility criteria

We analyzed demographic characteristics and early trauma variables that are known to or may influence rehabilitation and functional outcomes in SCI patients that were available in the merged database for patients with cervical, thoracic, and lumbar SCI. The neurological level of injury (NLI) was measured according to the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI), and completeness of injury was documented using the AIS grade. Measurements of NLI and AIS grade were made at the time of acute injury (within 72 hours after acute hospital admission) and discharge from inpatient rehabilitation.

Outcomes

Outcomes were measured by comparing the change in NLI and AIS grade after initial SCI and at discharge from inpatient rehabilitation and by examining total Functional Independence Measure (FIM) motor scores at discharge from inpatient rehabilitation. Changes in AIS grade from initial assessment and rehabilitation discharge were measured by recoding each AIS alpha score into a numerical score and subtracting the AIS numerical score at inpatient rehabilitation discharge from the initial AIS. Change in AIS was defined as improvement if an initial AIS of a more severe or “complete” category (e.g., AIS A) converted to a less severe or “incomplete” category (e.g., AIS B, AIS C, or AIS D) at discharge from inpatient rehabilitation.^13,14 Changes in NLI over time were measured by recoding each NLI grade into a numerical score and subtracting the numerical score at inpatient rehabilitation discharge from the initial numerical score obtained following SCI.

Statistical analyses

Multivariable linear regression analysis was performed to determine predictors of FIM motor discharge total and whether AIS or NLI improved (logistic regression) at inpatient rehabilitation discharge following SCI using multiple imputation for missing data. Analyses were performed using STATA software (16.1). Statistical significance was set at p < .05.

Results

A total of 209 SCI patients were merged across trauma registries and the SCIMS databases. The average age of the study population was 47.2 ± 18.9 years, and most patients were male. Patients with complete injuries (i.e., AIS A) at the time of initial injury accounted for 22.4% of the study population. Mean systolic blood pressure for the population on presentation to the emergency department (ED) was 124.1 ± 29.6 mm Hg, mean ED heart rate was 83.7 ± 19.9 bpm, mean oxygen saturation was 96.8% ± 4.0%, and mean Glasgow Coma Scale (GCS) score was 13.3 ± 3.9. The average Injury Severity Score (ISS) in this population was 22.4, and the greatest number of patients were admitted with cervical SCIs (Table 1).

Table 1.

Baseline characteristics of merged spinal cord injury trauma patients

Variable	Total values (missing values)	Min	Max	Average (SD)
Patient age	199 (10)	17	91	47.2 (18.9)
ED systolic blood pressure	161 (48)	0	213	124.2 (29.7)
ED heart rate	201 (8)	44	151	83.7 (19.9)
ED oxygen saturation	170 (39)	72	100	96.8 (4.0)
ED Glasgow Coma Scale (3–15)	198 (11)	3	15	13.3 (3.9)
Injury Severity Score (0–75)	172 (37)	1	75	22.5 (11.9)
Admission motor FIM score	93 (16)	13	53	21.5 (9.3)
Discharge motor FIM score	93 (16)	13	86	49.7 (21.0)
		Category	Count	Percent (%)a
Sex	199 (10)	Male	152	76.4
		Female	47	23.6
Insurance	147 (62)	Private	76	51.7
		Medicare	22	15.0
		Medicaid	30	20.4
		Worker’s comp.	6	4.1
		Other	13	8.8
Mechanical and/or external respiratory support in the EDs
	143 (66)	Yes	44	30.8
		No	99	69.2
AIS at acute hospital	133 (76)	A	47	35.3
		B	6	4.5
		C	30	22.6
		D	50	37.6
Neurologic level of injury at acute hospital	128 (81)	Cervical	85	66.4
		Thoracic	38	29.7
		Lumbar	5	3.9

Note: All continuous variables are reported as average values with standard deviations. AIS = American Spinal Injury Association Impairment Scale; ED = emergency department.

^aOf known values (missing values excluded in calculation).

Linear regression analysis showed that ISS and FIM at admission to inpatient rehabilitation were both statistically significant predictors of FIM at discharge from inpatient rehabilitation (Table 2). Age at injury, gender, ED SBP, ED heart rate, ED oxygen saturation, ED GCS, ventilatory support, and insurance did not demonstrate any significant association with FIM at discharge from inpatient rehabilitation for any specific level of injury category or AIS grade.

Table 2.

Logistic regression for discharge motor FIM

Model	B	SE	Significance	95% CI for B
				Lower	Upper
Admission FIM	2.98	0.68	<.01	1.62	4.33
ISS	−0.26	0.12	.03	−0.50	−0.02
R2 = 0.41

Note: CI = confidence interval; ISS = Injury Severity Score; SE = standard error.

Logistic regression showed a significant negative association between ventilatory support on presentation to the ED and subsequent improvement from initial AIS to AIS at discharge from inpatient rehabilitation (Table 3). No physiologic or clinical variables at ED presentation demonstrated any significant association with improvement from initial NLI to NLI at discharge from inpatient rehabilitation for any specific level of injury category or AIS grade.

Table 3.

Logistic regression for American Spinal Injury Association Impairment Scale

Model	OR	SE	Significance	95% CI for B
				Lower	Upper
Assisted ventilation	0.15	0.12	.02	0.03	0.78
C-statistic = 0.73

Note: CI = confidence interval; OR = odds ratio; SE = standard error.

Discussion

In our cohort, rehabilitation discharge FIM motor score was predicted by admission FIM motor score and ISS. It is logical that greater admission FIM score predicts greater discharge FIM scores, as those with greater functional repertoire at the beginning of rehabilitation end their rehabilitation stay at a higher functional status. The second predictor, ISS, negatively predicted discharge FIM score. The finding that greater burden of traumatic injury negatively predicts functional status is not surprising and has been reported in prior literature. Nemunaitis et al. also examined the predictive power of ISS in their cohort of 879 individuals with SCI, traumatic brain injury, and polytrauma¹⁵ and similarly found that greater ISS correlated with a lower motor FIM at rehabilitation discharge. In separate studies specifically in the SCI population, lower Spinal Cord Independence Measure scores (a functional measure that includes a measure of functional mobility) have been observed at 6 and 12 months postinjury among those with greater ISS.³

Surprisingly, we did not find that ED physiologic measures (SBP, pulse, oxygen saturation) predicted discharge motor FIM or AIS score improvement between acute hospitalization and rehabilitation discharge, as we hypothesized. A potential explanation of why these physiologic measures did not predict functional outcomes is that they represent only a temporal snapshot of the patient’s physiologic course peri-injury. One might postulate that the duration and severity of abnormalities in blood pressure, heart rate, and oxygenation, both before and after an observed time point in the emergency room, would be more clinically meaningful and more predictive of neurologic and functional outcomes. Many persons may have received partial resuscitation by the time these data were obtained. A more meaningful metric, albeit difficult to capture, may be duration of hypotension and hypoxia prior to resuscitation. Furthermore, in our cohort, only seven individuals had oxygen saturations <90% in the emergency room, likely reflecting this early resuscitation. Although our results suggest limited ability to study the impact of hypoxia using ED physiologic measures, they highlight the importance of prehospital data on physiologic parameters and resuscitation interventions to better characterize the influence of early posttraumatic hypoxia on SCI prognosis. As such, our negative findings represent an important insight into the constraints posed by existing data sources and potentially fertile areas of future investigation that would help delineate the relationship between acute physiologic measures and functional and neurological outcomes for individuals with traumatic SCI.

While emergency room vital signs were not predictive in our cohort, we did find that external ventilatory support or mechanical ventilation in the emergency room negatively predicted improvement in AIS score. A potential explanation is that need for mechanical ventilation could correlate with severity of hypoxia prior to ED presentation, regardless of level of injury. A greater degree of hypoxia may lead to both larger area of ischemia and greater severity of tissue compromise. Need for mechanical ventilation is likely more indicative of profound prehospital hypoxia than the oxygen saturation measure in the ED, which was not found to be a predictive factor. Mechanical ventilation may also act as a risk factor for decreased neurologic recovery through its association with other hospital complications that may affect trauma patients of any level of SCI, including prolonged immobilization, deconditioning, respiratory infection, and other associated medical complications such as venous thromboembolism and pulmonary embolism.

The majority of prior studies investigating the prognostic utility of acute physiologic measures in SCI have focused on in-hospital blood pressure management in the first week following injury. There is some evidence linking medically induced blood pressure elevation in the period following acute SCI with increase spinal cord perfusion pressure,¹⁶ as well as improved neurologic outcomes. Hawryluk et al. found in a retrospective cohort of 74 patients that higher MAP in the first 2 to 3 days postinjury correlated with greater improvement in AIS grade at hospital discharge.¹⁷ More recently, a prospective trial by Squair et al. of 92 individuals monitored MAP and calculated spinal cord perfusion pressure (calculated as MAP-measured cerebrospinal fluid pressure) during 7 days postinjury. The study found that both MAP and spinal cord perfusion pressure positively correlated with AIS motor improvement at 6 months postinjury.¹⁸ In contrast, in their retrospective study of 105 patients, Martin et al. found no correlation between episodes of hypotension during the first 72 hours of hospitalization and AIS motor score at hospital discharge.¹⁹ Thus, a more specific driving metric may be spinal cord perfusion, which is impacted by both MAP and intrathecal pressure.²⁰

One recent study by Gallagher et al. that examined intraoperative perfusion patterns in 22 acute SCI patients using laser speckle contrast imaging suggests a more complex relationship between MAP and spinal perfusion pressure that could shed light on these mixed findings. In 5 of the 22 patients, the use of norepinephrine to increase MAP by 20 mm Hg increased blood flow to some injured spinal cord regions but paradoxically decreased perfusion to others, which the authors termed “blood pressure-induced local steal.” This may suggest that the presumed correlations between blood pressure, spinal perfusion, and neurologic recovery may be more complicated than previously postulated.²¹

Blood pressure in the days immediately following injury has been studied, but the impact of blood pressure trajectory prior to hospitalization, at the scene of the injury, remains uncharacterized. Logistically, this is a difficult period to study, due to paucity of data typically collected in the field (for this reason, we chose to look at emergency room measures). Prehospital physiologic factors remain an important area for future research, as early intervention could save penumbral tissue. In addition, while the impact of blood pressure in the immediate postinjury days has been studied as described above, there is a paucity of data looking specifically at the influence of other vital signs, including heart rate and oxygen saturation, during the acute postinjury period on long-term neurological outcomes. Though a different measure, Biglari et al. did examine the impact of blood hemoglobin concentration on neurologic improvement in a cohort of 68 patients. They found that greater blood hemoglobin concentrations 3 days postinjury were associated with higher likelihood of AIS improvement within 3 months, implicating the importance of oxygen delivery for spinal cord tissue recovery.²²

Our study does have some limitations. First, there was a significant amount of missing data, particularly values of systolic blood pressure, respiratory status (ventilation/no ventilation), and insurance status. It is unknown whether these measures were missing at random. Also, in our cohort, unlike most previous literature, the examined blood pressure measure was SBP rather than MAP. We were unable to examine the influence of MAP because systolic but not diastolic pressures were recorded in the trauma database. Additionally, we chose to examine neurologic and functional outcomes at the time of rehabilitation discharge due significant loss of these measures at 1-year follow-up. Longer term outcome measures will likely be more impactful to examine in the future, as variable durations of inpatient rehabilitation based on insurance status and coverage type make assessing outcomes at rehabilitation discharge inherently problematic. We also did not have measures of intraspinal pressure (which can be used to calculate spinal perfusion pressure) in this group, so we could not examine the potential effect of this measure. Future studies would benefit from larger sample size, more consistent data collection, particularly at the scene of injury, and longer term follow-up. They could also incorporate other clinical factors and measures, such as comorbid medical conditions, genetic factors, and spinal perfusion pressure, which may further aid in prognostication of functional outcomes.²³ Given the relatively high motor improvement rate in participants in placebo arms of prior clinical trials, an opportunity exists to examine those with the lowest BP and oxygen levels and compare them to historical placebo group AIS score and level conversion rates.

Despite these limitations, this analysis provides a meaningful contribution to the few prior studies examining the impact of acute physiologic measures on functional outcomes in traumatic SCI. Our sample size was moderate and drew from five separate trauma regions across the United States, supporting the generalizability of our findings. Our results overall suggest caution in using many acute care variables in SCI rehabilitation prognosis but suggest that mechanical ventilation may be associated with worse neurologic recovery. This study provides a basis from which future studies on acute physiologic measures and neurologic and functional outcomes in traumatic SCI should be built.

Conclusion

Our study showed that admission FIM and ISS both predict discharge FIM for patients with traumatic SCI using merged data from the SCIMS database and institutional trauma databases from five academic medical centers. Furthermore, this retrospective analysis demonstrated that ventilatory support requirements are negatively associated with improvement in AIS grade at discharge from inpatient rehabilitation compared with AIS grade after initial injury. Notably, other physiologic and clinical variables at ED presentation such as SBP, heart rate, and oxygen saturation were not significantly associated with rehabilitation outcomes including discharge FIM, improvement in NLI, and improvement in AIS. The absence of any association between these early clinical variables with subsequent rehabilitation outcomes is counterintuitive and suggests important areas for future investigation and clinical research that may inform best practices, treatment protocols, and management paradigms for patients with traumatic SCI.