Predicting Complete versus Incomplete Long-Term Functional Independence after Acute AIS Grade D Spinal Cord Injury: A Prospective Cohort Study

Abstract

Background:

The proportion of patients with American Spinal Injury Association Impairment Scale (AIS) grade D traumatic spinal cord injuries (tSCI) is increasing. Although initial motor deficits can be relatively mild, some individuals fail to recover functional independence.

Objectives:

This study aims to identify factors associated with failure to reach complete functional independence after AIS grade D tSCI.

Methods:

An observational prospective cohort study was conducted at a level 1 trauma center specialized in SCI care. A prospective cohort of 121 individuals with an AIS-D tSCI was considered. The baseline characteristics, length of acute stay, need for inpatient rehabilitation, and 12-month functional status were assessed. Univariate and classification and regression tree (CART) analyses were performed to identify factors associated with reaching complete versus incomplete functional independence (defined as perfect total SCIM III score at 12-month follow-up).

Results:

There were 69.3%, 83.3%, and 61.4% individuals reaching complete independence in self-care, respiration/sphincter management, and mobility, respectively. A total of 64 individuals (52%) reached complete functional independence in all three domains. In the CART analysis, we found that patients are more likely to achieve complete functional independence when they have a baseline motor score ≥83 (65% individuals) and if they present fewer medical comorbidities (70% individuals if Charlson Comorbidity Index [CCI] ≤4).

Conclusion:

About half of individuals with AIS grade D tSCI can expect complete long-term functional independence. It is important to recognize early during acute care individuals with baseline motor score <83 or a high burden of comorbidities (CCI ≥5) to optimize their rehabilitation plan.

Introduction

It is now well established that the proportion of patients with traumatic spinal cord injury (tSCI) who are older adults (≥65 years old) is increasing over time.^1-5 Injuries in the elderly often result from low energy falls and are typically associated with functional motor-incomplete injuries (i.e., American Spinal Injury Association Impairment Scale [AIS] grade D).^6-9 With the aging of the population, this demographic shift in the tSCI landscape is only expected to accelerate in the future, such that AIS grade D injuries are becoming a clinical entity of strategic importance, both for clinicians and decision makers at the system level.

Unfortunately, patients with AIS grade D injuries represent a highly heterogeneous subgroup of patients. In particular, the spectrum of initial neurological deficits observed can vary widely. While some AIS grade D individuals display normal motor testing with only sensory deficits at the initial neurological assessment, others can present major motor deficits, such as complete paralysis below the neurological level of injury (NLI; defined as the most caudal level with preserved normal sensation and motor strength) on one side with preserved antigravitational strength contralaterally. It is also our clinical experience that the level of functional recovery after AIS grade D tSCI can vary immensely, which could be a result of this heterogeneity. Unfortunately, failing to recover complete functional independence can have profound negative impacts on long-term quality of life and social participation.^10,11 As such, it is important that these patients be identified early during the acute care in order to tailor management and potentially optimize outcomes.

The objectives of this prospective cohort study are twofold. First, we aim to characterize what the long-term expectations for individuals with AIS grade D tSCI can be to reach complete functional independence in terms of self-care, respiration and sphincter management, and mobility. Second, we aim to identify factors associated with the failure to reach complete functional independence 12 months after the injury using classification and regression tree (CART) analysis. We hypothesize that there are specific clinical variables that clinicians can identify early during the acute care period to recognize individuals who are less likely to reach complete long-term functional independence.

Methods

Study population

This prospective observational cohort study was conducted at a single level 1 trauma center specialized in SCI care in Montreal (Canada) between May 1, 2010 and November 5, 2020. The inclusion criteria were (1) age ≥18 years, (2) tSCI with an NLI from C1 to S3, (3) functional motor-incomplete AIS grade D injury upon admission to acute care, (4) follow-up with fully completed functional status questionnaire 12 months after the injury, and (5) voluntary consent to participate in the study. The exclusion criteria were (1) preexisting neurological condition (e.g., stroke, traumatic brain injury, cerebral palsy, preexisting SCI, etc.), (2) tumoral or infectious involvement of the spine, (3) requiring home services or living in a nursing home prior to the trauma, and (4) unable to comply with planned follow-up visits (e.g., in prison, living in another country, etc.). This project was approved by the local Research Ethics Board (Comité d’éthique de la recherche CIUSSS Nord-de-l’Île-de Montréal, Québec, Canada), and informed written consent for data collection and data analysis was obtained shortly after admission.

Routine SCI care

At our acute care institution, all individuals with tSCI are managed by a multidisciplinary team including physiatrists, spine surgeons, physical therapists, occupational therapists, social worker, liaison nurses, and nurses specialized in SCI care. During acute care, all patients receive similar rehabilitation resources and medical attention. Physical therapy is performed 6 days/week for the first 14 days and 4 times/week thereafter; occupational therapy is provided 3 times/week. Care from other consultants (psychology, nutrition, speech therapy, etc.) is provided as needed.

To facilitate the continuum of care, our health care system requires that specialized and coordinated SCI care be provided at the same affiliated inpatient functional rehabilitation institution following acute care at our center.¹² Patients are eligible for transfer to inpatient functional rehabilitation when (1) medically stable (i.e., stable vital signs, no impending risk of short-term deterioration, and no active treatments for acute conditions, including mean arterial pressure augmentation therapy, intravenous antibiotic treatment, and management of postoperative complications), (2) requiring daily inpatient therapies in at least two different disciplines (including physical, occupational, speech therapy, nutrition, social work, psychology), (3) showing potential and motivation to improve and reach established functional goals with inpatient functional rehabilitation, and (4) tolerating a minimum of 60 minutes/day for therapies.

Data collection

All data were collected prospectively during acute care and outpatient follow-up visits by research assistants who were neither involved in data analysis nor patient care.

Baseline neurological assessment

The baseline neurological evaluation was performed within 72 hours of admission by a trained member of the physiatry or spine surgery team using the International Standards for the Neurological Classification of Spinal Cord Injury (ISNCSCI).¹³ The AIS grade was used to determine the completeness of the lesion (AIS grade D for all patients included in this study) based on the sensory, motor, and rectal examination, including voluntary anal contraction (VAC). The NLI, as defined above, was categorized into high tetraplegia (C1-C4), low tetraplegia (C5-C8), high paraplegia (T1-T8), low paraplegia (T9-L1), and cauda equina syndrome (L2-S3). The total motor, light touch, and pinprick scores were also calculated.

Covariates: baseline variables assessed from acute care

Age, sex, body mass index (BMI), comorbidities, living arrangement, trauma severity, mechanism of injury, length of acute stay, and transfer to inpatient rehabilitation were collected. The burden of comorbidities was assessed using the Charlson Comorbidity Index (CCI).¹⁴ The CCI was dichotomized into the presence of low (0-4) and high (5 or more) burden of comorbidities, which are associated with <20% and ≥20% of risk-adjusted hazard ratio for mortality within one year in Canada, respectively.^15,16 Living arrangement was dichotomized as living alone or not. Trauma severity was quantified by the Injury Severity Score (ISS).¹⁷ Mechanism of injury was classified into falls, motor vehicle accidents (MVA), and others.

Outcome variables: reaching complete functional independence in self-care, respiration and sphincter management, and mobility

Functional status was measured 12 months after the tSCI from the third version of the Spinal Cord Independence Measure (SCIM) questionnaire. The 12-month timepoint was chosen specifically because research has shown that functional recovery largely wanes at this point after a tSCI.¹⁸ The SCIM questionnaire is a reliable and sensitive tool specific to SCI when evaluating the ability to perform daily living activities independently in three domains: self-care, respiration and sphincter management, and mobility.¹⁹ The self-care domain includes six items: (1) feeding, (2) upper body bathing, (3) lower body bathing, (4) upper body dressing, (5) lower body dressing, and (6) grooming. The respiration and sphincter domain includes four items: (1) respiration, (2) sphincter management – bladder, (3) sphincter management – bowel, and (4) use of toilet. The mobility domain includes nine items: (1) mobility in bed and action to prevent pressure sores, (2) transfers: bed-wheelchair, (3) transfers: wheelchair-toilet-tub, (4) mobility indoors, (5) mobility for moderate distances (10-100 m), (6) mobility outdoors (more than 100 m), (7) stair management, (8) transfers: wheelchair-car, and (9) transfers: ground-wheelchair. For each domain, complete functional independence is defined as top box scoring in all corresponding items. Complete long-term functional independence was defined by top box scoring for all items of all three domains (i.e., 100/100 total SCIM score).

Statistical analysis

Independent variables, covariates, and outcomes were first described using means ± SD for continuous variables and frequencies (percentages) for categorical variables. Direct comparisons between individuals reaching versus not reaching complete functional independence were then performed using chi-square and Student t tests for categorical and continuous variables, respectively. The level of significance was set at .05 for all univariate analyses. The IBM SPSS version 26 was used to perform all univariate analyses.

Finally, a CART analysis was performed to verify our main hypothesis and identify specific barriers to reaching functional independence. The following baseline characteristics were included as candidate predictors in the model: age, sex, BMI, CCI, ISS, NLI, and total motor score. The outcome variable consisted of the long-term functional independence status, that is, complete functional independence if 100/100 on total SCIM score versus incomplete functional independence otherwise. In order to build the tree, the Minitab software uses a Gini splitting rule to partition the dataset into sequential mutually exclusive subgroups until further partitioning does not add predictive value. In this case, overfitting was monitored by choosing the tree exhibiting the minimum relative cost value computed by the software. A stopping rule was used to prevent the algorithm from creating terminal nodes of fewer than five patients. During the building of the tree, a 10-fold cross-validation was used to minimize bias and variability of estimates. The relative importance of each candidate predictor was finally computed for the resulting model. Variable importance reflects the ability of a variable to split a cluster of patients, that is, how much influence a predictor has on the outcome variable according to the model. The most important variable during tree building is assigned with a score of 100, and the other variables are scaled down proportionally to their importance.

Results

Amongst the 126 eligible participants, 1 was excluded because they presented a prior history of stroke and 4 were excluded because they required home services before the accident. The final cohort available for analysis thus comprised 121 patients.

The mean age was 53.1 ± 17.2 years. There were 97 males and 24 females. The NLI was cervical (C1-C8) for 79 individuals, thoracic or thoracolumbar (T1-L1) for 24 individuals, and lumbosacral (L2 or lower) in 11 individuals. The mean total motor, light touch sensory, and pinprick sensory scores were respectively 85.4 ± 12.9 (range, 33-100), 99.7 ± 13.9 (range, 57-112), and 90.9 ± 22.2 (range, 10-112) points. The mean length of acute stay was 19.9 ± 11.8 days (range, 4-72), and 71 (57.7%) individuals were transferred to our affiliated institution for inpatient rehabilitation.

The mean subscores were 19.0 ± 2.4, 39.1 ± 2.7, and 37.7 ± 4.4 for the self-care, respiration and sphincter management, and mobility domains, respectively. There were 85 (69.1%), 102 (82.9%), and 72 (58.5%) individuals reaching complete functional independence in self-care, respiration and sphincter management, and mobility, respectively. The mean total SCIM score was 95.8 ± 6.9 (range, 69-100), and 64 (52%) individuals recovered complete functional independence simultaneously for all three domains (i.e., total SCIM score of 100).

At the bivariate level, individuals who achieved complete functional independence in all three domains were significantly younger (p = .009), had shorter acute length of stay (LOS) (p = .003), had higher AIS motor scores (p < .001), had higher AIS pinprick scores (p = .005), and were less likely to present a CCI ≥ 5 (p = .002) and to have been transferred to inpatient functional rehabilitation following the acute care (p = .005). In contrast, there were no significant differences between the two groups in terms of sex, BMI, living situation, NLI, AIS light touch sensory score, trauma severity (ISS), presence or absence of VAC, and mechanism of injury (Table 1).

Table 1.

Comparison of baseline characteristics of individuals achieving complete versus incomplete functional independence at long-term follow-up

Characteristics	Did not achieve complete functional independence	Achieved complete functional independence	p
Sex (% male)	46 (80.7)	51 (79.7)	.889
Age, mean±SD	57.4±17.0	49.3±16.7	.009
BMI, mean±SD	27.5±6.4	26.4±5.9	.374
CCI % 5 and up	6 (10.5)	0 (0)	.002*
Mechanism of injury
% Falls	28 (49.1)	26 (40.6)
% MVA	21 (36.8)	17 (26.6)	.051
% Other	8 (14)	21 (2.8)
Living situation (% alone)	17 (29.8)	20 31.3)	.865
AIS motor score, mean±SD	81.0±13.5	89.4±11.0	<.001
Neurological level of injury
% C1-C4	28 (49.1)	24 (37.5)
% C5-C8	18 (31.6)	15 (23.4)
% T1-T8	3 (5.3)	3 (4.7)	.090*
% T9-L1	6 (10.5)	12(18.8)
% L2 and below	2 (3.5)	10 (15.6)
AIS light touch score, mean±SD	97.7±15.0	101.2±13.0	.205
AIS pinprick score, mean±SD	83.3±26.5	96.8±16.1	.005
VAC (n=110) % Present	44 (86.3)	52 (88.1)	.770
ISS, mean±SD	18.7±4.5	18.7±6.0	.974
LOS acute care, mean±SD	23.4±14.0	16.8±8.4	.003
Went to IFR % Yes	41 (71.9)	30 (46.9)	.005

Note: AIS = American Spinal Injury Association Impairment Scale; BMI = body mass index; CCI = Charlson Comorbidity Index; IFR = inpatient functional rehabilitation; ISS = Injury Severity Score; LOS = length of stay; VAC = voluntary anal contraction.

^*Maximum likelihood ratio test performed because >20% of cells had expected count less than 5. Results in bold indicate statistical significance at p < .05.

The final CART model (Figure 1) showed that the most important predictor for complete long-term functional independence is the baseline total motor score. A baseline total motor score ≤82 was associated with 72.5% likelihood for having incomplete long-term functional independence, whereas only 35% of individuals with a baseline total motor score >82 had incomplete long-term functional independence. For those with a baseline total motor score >82, a CCI ≥5 was associated with 100% incomplete long-term function independence as compared to 30% for those with CCI ≤4. Of all candidate predictors included in the CART analysis, the relative importance of the total motor score was 100%, whereas the relative importance of CCI and age were respectively 58.6% and 37.0%.

Figure 1.

Final classification tree obtained from the classification and regression tree (CART) analysis of factors associated with failure to achieve compete functional independence after American Spinal Injury Association Impairment Scale (AIS) grade D traumatic spinal cord injury (tSCI). CCI = Charlson Comorbidity Index; complete = achieved complete functional independence; incomplete = failed to achieve complete functional independence.

Discussion

This study brings valuable knowledge to characterization and recovery of individuals with functional motor-incomplete AIS D injuries, which represent a growing subgroup of the tSCI population.^10,11 Although distinctive from more severe AIS A to C lesions, there is also a wide heterogeneity with regard to the neurological deficits and long-term functional outcomes. Although the majority presented with a cervical tSCI (42.3% high cervical and 26.8% low cervical), 29.8% presented a thoracic or lumbar lesion. In addition, there were large variabilities in baseline total motor scores (range, 33-100), light touch sensory scores (range, 57-112), and pinprick sensory scores (range, 10-112). Similarly, the total SCIM scores at long-term follow-up were highly variable, ranging from 69 to 100, with about half of individuals who can expect to reach complete functional independence in self-care, respiration and sphincter management, and mobility 12 months after the injury.

In the CART analysis, a baseline total motor score ≤82 was the most important predictor for not reaching complete functional independence, which supports the importance of a thorough neurological assessment using the ISNCSCI early after the injury. Another major barrier to reaching complete functional independence was the burden of comorbidities. Enquiring about comorbidities is particularly relevant in individuals with AIS grade D because this subgroup of patients is expected to involve an increasingly high number of elderly individuals in the near future who are more likely to present with severe comorbidities. There were significantly more individuals with five or more comorbidities in the subgroups not reaching complete functional independence in either self-care, respiration and sphincter management, or mobility domains. This finding is perfectly in line with the increased proportions of geriatric patients observed in these subgroups, considering that comorbidities are strongly associated with older age. This most likely explains why age was not comprised in the final CART model when also accounting for CCI in the analyses. This finding is also supported by the fact that age was the third most important predictor of the dataset, that is, third in relative importance after total motor score and CCI. It is believed that the difficulty to achieve complete independence in self-care and mobility for older patients could reflect the complex integration of the sensory, motor, and other health systems involved in these activities associated with daily living, which are likely to decline with age.²⁰

The significant proportion (48%) of individuals failing to achieve complete functional independence, particularly in self-care and mobility, raises the need to adapt functional rehabilitation care for this subgroup of patients that will only grow in the future. The early involvement of geriatricians, physiatrists, and specialized rehabilitation resources is warranted, targeting early initiation of rehabilitation therapies during acute care, the avoidance of deconditioning and contractures, and prevention of medical complications that could interfere with rehabilitation. Future studies could determine whether tailored mobility training (e.g., activity-based therapy, bed mobility and transfer techniques, etc.) and training self-care skills should be initiated early during acute care. Based on our findings regarding comorbidities and their negative impact on functional recovery, it is also possible to hypothesize that optimizing the management preexisting medical conditions during the acute care could potentially eliminate certain barriers to effective rehabilitation and improve functional recovery in the long run.

Acute LOS has been identified previously as a strong predictor of functional recovery after tSCI.²¹ However, it was not a significant predictor of reaching complete functional independence in our final classification tree. The LOS is usually a good surrogate for the course of the acute hospitalization which in turn reflects the severity of the tSCI and associated injuries and mainly comprises the time required to ensure hemodynamic control, stabilize the medical condition, treat all associated injuries, and manage complications. For individuals with AIS grade D tSCI, the course of the acute hospitalization until medical stabilization (i.e., no active treatment required) is generally less care intensive due to the lesser severity of the injury and associated conditions, and therefore it does not vary significantly between patients. In our clinical practice, longer length of acute stay for AIS grade D is mainly observed when there is a need to prepare the patient for transfer to inpatient rehabilitation because of the required counseling care, rehabilitation planning, inter-institution communications, administrative tasks, and transfer delays.

The proportion of individuals needing inpatient rehabilitation following acute care was significantly increased in those not reaching complete independence in self-care and respiration and sphincter management, and it almost reached significance for mobility. However, inpatient rehabilitation was not significantly associated with complete functional independence for either domain at the multivariable level, again suggesting that current rehabilitation strategies and protocols can still be optimized or adapted to help AIS grade D patients to reach complete functional independence.

Limitations

Our method used to define complete functional independence cannot by itself capture all the aspects involved in the functional recovery of individuals with tSCI. Nevertheless, the SCIM is a valid questionnaire specifically and commonly used to assess functional recovery after SCI. Our top box scoring method proposed for defining complete functional independence is adapted to AIS grade D individuals because SCIM scores present a bimodal distribution, with the majority of individuals reaching top box scores in one or more domains. Dichotomizing the functional outcome into complete versus noncomplete functional independence is also convenient as it will be easier in future studies to identify and/or investigate clinical subgroups of patients associated with reaching complete independence.

Considering that comorbidities and age are important predictors of reaching complete functional independence, we acknowledge that using a measure of frailty could have added to a better understanding of functional recovery after tSCI in older individuals. Frailty involves age-related processes resulting in diminished functional reserve and accumulation of pathological processes,^22,23 which can put elderly individuals with tSCI at higher risk of cascading decline-related problems such as linear functional decline and lower hospitalization efficiency.^4,6 While we have excluded individuals living in a nursing home or requiring home services, we have also taken into account other factors associated with frailty in our study, such as the living arrangement and the burden of comorbidities. In addition, regarding comorbidities, although we used a commonly used and validated index to characterize the severity and seriousness of preexisting medical conditions (Charlson Comorbidity Index), we recognize that no analyses were performed to determine which specific conditions are associated with poorer functional outcomes. However, this was outside the scope of the present study and will have to be evaluated in future work.

Finally, we recognize that the presented cohort was mostly male and comprised a majority of cervical TSCI, such that results must be carefully interpreted when applied to other populations with different compositions. However, it is important to mention that these characteristics largely reflect the current landscape of AIS grade D TSCI. ^9,24

Conclusion

About half of individuals with AIS grade D tSCI can expect complete long-term functional independence, whereas reaching complete functional independence can be challenging for the other half of individuals, particularly in terms mobility and self-care. It is important to recognize early during acute care individuals with baseline motor score <83 or a high burden of comorbidities (CCI ≥5) because they are less likely to achieve complete long-term functional independence and may require specialized resources such as inpatient rehabilitation and home services. Acknowledging these barriers can help clinicians and decision-makers to tailor the management for this growing tSCI subpopulation in the future and potentially implement new medical and rehabilitation throughout the continuum of care to optimize the functional outcomes after AIS grade D tSCI.