Retrospective study of functional outcomes and disability after non-ischaemic vascular causes of spinal cord dysfunction

Abstract

Objective: Describe demographic characteristics, functional outcomes and disability following rehabilitation for non-ischemic vascular spinal cord dysfunction (SCDys).

Design: Retrospective, open cohort, case series.

Setting: Tertiary rehabilitation unit, Victoria, Australia.

Participants: Patients with non-ischemic vascular SCDys admitted over a 21-year-period (01/01/1995–31/12/2015) were identified using International Classification of Diseases codes.

Outcome Measures: Demographic characteristics, etiology, neurologic classification, length of stay (LOS), and complications. On admission and discharge, the following were collected: functional independence measure (FIM) motor subscale, details on bowel, bladder, mobility, living arrangement, and support services.

Results: 36 patients (female 58%; mean age 69 ± 16 years) were identified. The main causes of non-ischemic vascular SCDys were epidural hematoma (39%), dural arteriovenous fistula (17%), and arteriovenous malformation (11%). 22 cases (61%) were iatrogenic. Most (86%) had incomplete paraplegia. Urinary tract infection was the most common complication (64%). Median LOS in rehabilitation was 68 days. Significant improvement in FIM motor scores was observed from admission (median 25, interquartile range [IQR] 20–38) to discharge (median 69, IQR 38–77) (P < 0.001). On discharge, 4 patients (11%) walked >100 m unaided, 6 (17%) walked >100 m with assistive device, 10 (28%) walked >10 m with assistive device, 15 (41%) were wheelchair dependent and 1 (3%) patient remained non-mobile. 20 patients (56%) were discharged home, 8 (22%) to nursing home, and 8 (22%) transferred to another hospital.

Conclusion: Most patients returned home with significantly improved functional outcomes compared to rehabilitation admission, but with the majority having ongoing major disabilities based on FIM motor scores.

Introduction

The incidence of spinal cord damage not due to trauma, referred to here as spinal cord dysfunction (SCDys), is reported to be higher than that of traumatic spinal cord injury, with an incidence of 26.3 cases per million population per year.¹ Although rehabilitation outcomes following ischemic causes of SCDys have been reported,^2,3 other vascular causes of SCDys from non-ischemic etiologies have received much less attention in the rehabilitation literature, likely due to their rarity. Non-ischemic vascular causes of SCDys have been classified into hemorrhage and vascular malformations.⁴ Non-traumatic spinal cord hemorrhage (NTSCH) can be classified by the location (epidural, subdural, subarachnoid or intramedullary), and the etiology (bleeding diathesis, medication, or other). Vascular malformations are classified as dural arteriovenous fistula (AVF), and arteriovenous malformations (AVM) with or without hemorrhage. The incidence of NTSCH and vascular malformations causing SCDys is unknown, but anecdotally these are reported to be rare. An international comparison study of SCDys reported that NTSCH accounted for 5% of cases of SCDys admitted for inpatient rehabilitation in specialist spinal rehabilitation units and vascular malformations accounted for 4%.⁵

This study aims to address the dearth of research regarding the rehabilitation outcomes for non-ischemic vascular SCDys, which includes NTSCH and vascular malformations. A range of outcomes, including the demographic characteristics, clinical, disability and functional outcomes in a cohort of these patients undergoing their initial inpatient rehabilitation after a recent onset of non-ischemic vascular SCDys was studied.

Method

Setting and study design

A retrospective cohort case series was conducted of all patients with non-ischemic vascular SCDys who were admitted between January 1, 1995 and December 31, 2015 to a tertiary spinal rehabilitation unit (Caulfield, Victoria, Australia). The spinal rehabilitation service is a 12-bed adult inpatient unit located in a public hospital and funded by the State government, but patients are referred from both private and public hospitals. The spinal rehabilitation service admits approximately 45–50 patients with spinal cord injury/damage per year.

Participants

The inclusion criteria were patients who were age ≥ 18 years at admission to rehabilitation with recent onset of SCDys due to either NTSCH or vascular malformations. Spinal cord conditions caused by trauma, other non-traumatic causes, or readmission following previous non-ischemic vascular SCDys were excluded.

Data collection

Potential participants were identified using relevant codes from International Classification of Diseases 9th edition (ICD 9) and 10th edition (ICD 10) for vascular myelopathy (P336.1 [ICD 9] and G95.1 [ICD 10]), and non-traumatic spinal cord damage (344 [ICD 9], G82 and G83.4 [ICD 10]). Medical files with these codes were subsequently reviewed by the primary author to confirm the diagnosis and obtain the relevant information.

Demographic and clinical details were extracted from the medical files notes. Demographic characteristics were age, sex, vocation and marital status. The clinical features included: the etiology of non-ischemic vascular SCDys, using the International Spinal Cord Injury Data Sets for non-traumatic spinal cord injury,⁴ classification of neurologic level (paraplegia or tetraplegia); American Spinal Injury Association Impairment Scale (AIS) grade (complete, AIS A or incomplete, AIS B to E);⁶ length of stay (LOS) in acute hospital and rehabilitation; and complications occurring during rehabilitation that were documented in the medical record as requiring management. In addition, a range of rehabilitation admission and discharge outcomes were also collected: the functional independence measure (FIM) motor subscale;⁷ mobility (walking >100 m without assistive device, walking >100 m with assistive device, walking >10 m with assistive device, wheelchair dependent, or non-mobile); bowel functional outcome (continent without aperients, continent with aperients +/− suppositories, or incontinent); bladder functional outcome (continent, intermittent self-catheterization, indwelling/suprapubic catheter, or incontinent); living arrangement at discharge (home alone, home with partner/family, hostel/retirement village, nursing home, or transfer to another hospital); formal social support services (having meals externally prepared, help with domestic duties, or paid personal care assistance); and informal supports (none required, informal support by resident family, or informal support by non-resident family/friend).

Statistical analysis

Categorical data was presented as absolute numbers and percentages. Continuous data was presented as mean and standard deviation if data was normally distributed, or median and interquartile range (IQR) for non-Gaussian data. Wilcoxon Signed-Rank test was used to compare FIM motor scores, and Pearson’s chi-square test was used to compare various categorical parameters. Statistical analysis was performed using SPSS (version 22, IBM Corp, New York, USA), and p value of <0.05 was considered to be statistically significant.

Results

There were 169 potential participants identified by the ICD codes and after reviewing the medical records 42 patients were identified with NTSCH or vascular malformations. Five patients were excluded (trauma = 2, readmissions = 3), and another participant’s full medical records could not be located, leaving 36 patients who were included in the analysis.

Majority of the patients were female (n = 21, 58%), of older age (mean age 69 ± 16 years) and were not in paid employment (aged pension n = 26, 72%; unemployed n = 4, 11%). Six patients (17%) were employed in paid workforce. Most of the patients were married or living with domestic partner (n = 21, 58%). Twelve patients (33%) were divorced or widowed, and three patients (8%) were single.

The most common etiology was epidural hematoma (39%), followed by dural AVF (17%), AVM (11%) and other hemorrhages (33%). There were twelve cases classified as other hemorrhages: 4 were intradural hematoma, 3 were subdural hematoma, 2 were subarachnoid hemorrhage, 1 was cavernoma, and 2 had bleeding at surgical site post laminectomy. Of note, 61% of cases had an iatrogenic component. These included bleeding following spinal surgeries (n = 8, 22%), warfarin therapy (n = 7, 19%), and spinal procedures (n = 7, 19%). Spinal procedures leading to spinal cord hemorrhage were epidural anesthesia (n = 2), spinal anesthesia (n = 1), computer tomography-guided nerve root injection (n = 1), arterial embolization (n = 1), epidural injection for sciatica (n = 1), and prophylactic cerebrospinal fluid drainage during thoracoabdominal aortic aneurysm repair (n = 1).

The pattern of neurologic damage was predominantly incomplete and resulting in paraplegia. The majority of patients (n = 31, 86%) had incomplete paraplegia, three patients (8%) had complete paraplegia, and two patients (6%) had incomplete tetraplegia. There was no patient who had complete tetraplegia.

The median LOS in acute hospital prior to transfer to the spinal rehabilitation service was 20 days (IQR 14–30). The median LOS in rehabilitation was 68 days (IQR 37–96). Urinary tract infection was the most common complication during the inpatient admissions (64%). The prevalence of other complications is shown in Table 1.

Table 1. Complications during inpatient rehabilitation admission.

	N = 36	Percentage (%)
Complications
Urinary tract infection	23	64
Pressure ulcer	9	25
Anxiety/depression	9	25
Spasticity	8	22
Pain – nociceptive	8	22
Pain – neuropathic	8	22
Pain – mixed	1	3
Pneumonia	5	14
Venous thromboembolism	3	8

Majority of the patients had poor functional status on admission to rehabilitation with regards to their physical disability, with poor mobility and major bowel and bladder disturbances (Table 2). By discharge from rehabilitation, however, significant improvements in functioning had occurred. On discharge, the most common functional outcomes were wheelchair dependent for mobility (41%), required indwelling/suprapubic catheterization for bladder functioning (47%), and bowel continence using aperients +/− suppositories (64%). Most patients (56%) were discharged home, but fewer were living alone. Eight patients (22%) were transferred to another hospital, which included admissions to acute hospital or transfer to another rehabilitation hospital closer to home as part of a policy to relocate those patients who live far from the spinal unit closer to home for the final discharge planning phase of their rehabilitation when they no longer needed the expertise of the unit. There were 8 patients (22%) who were discharged to nursing home for long-term care. As expected after SCDys, some patients were discharged with formal support services, such as having meals externally prepared, help with domestic duties, and paid personal care assistance, or informal support from family and friends; however, the observed changes in formal and informal supports were not statistically significant.

Table 2. Premorbid and discharge parameters.

Outcome	On admission	On discharge
FIM motor (median, IQR)	25, 20–38	69, 38–77	Wilcoxon Signed-Rank test Z = −5.1, P < 0.001
Mobility			Pearson’s χ2 = 18.0, P = 0.001
Walking >100 m, no assistive device	1 (3%)	4 (11%)
Walking >100 m, assistive device	2 (6%)	6 (17%)
Walking >10 m, assistive device	3 (8%)	10 (28%)
Wheelchair	17 (47%)	15 (41%)
Non-mobile	13 (36%)	1 (3%)
Bladder continence			Pearson’s χ2 = 10.4, P = 0.009
Continent	5 (14%)	13 (36%)
Intermittent self-catheterization	0 (0%)	4 (11%)
Indwelling/suprapubic catheter	28 (78%)	17 (47%)
Incontinent	3 (8%)	2 (6%)
Bowel continence			Pearson’s χ2 = 18.2, P < 0.0001
Continent without aperients	4 (11%)	4 (11%)
Continent with aperients ± suppositories	6 (17%)	23 (64%)
Incontinent	26 (72%)	9 (25%)
Accommodation			Pearson’s χ2 = 24.1, P = 0.0001
Home alone	11 (31%)	2 (6%)
Home with partner/family	24 (67%)	18 (50%)
Hostel/retirement village	1 (3%)	0 (0%)
Nursing home	0 (0%)	8 (22%)
Transfer to another hospital	0 (0%)	8 (22%)
Social supports			Pearson’s χ2 = 3.6, P = 0.2
Having meals externally prepared	0 (0%)	2 (17%)
Help with domestic duties	6 (17%)	6 (17%)
Paid personal care assistance	1 (3%)	6 (17%)
Informal support			Pearson’s χ2 = 1.3, P = 0.5
None	7 (19%)	11 (31%)
Resident family	16 (44%)	15 (42%)
Non-resident family/friend	13 (36%)	10 (28%)

FIM, functional independence measure; IQR, interquartile range.

Discussion

Non-ischemic vascular SCDys is quite rare. Patients tend to be aged in their 60s, have an incomplete paraplegic level of damage, and often have an iatrogenic component to their SCDys. Most of the patients were discharged home with significant improvement in their functional outcomes compared to rehabilitation admission, but nevertheless, with the majority having residual major disabilities.

This is the first study to describe the demographic characteristics, clinical features and rehabilitation outcome of non-ischemic vascular SCDys. A literature search yielded only individual case reports.^8–11 In addition, we did not identify any study reporting the rehabilitation outcomes of non-ischemic vascular SCDys. The incidence of non-ischemic vascular SCDys is unknown. Our study identified 36 participants with SCDys from NTSCH and vascular malformation over 21 years. This small number suggests that non-ischemic vascular SCDys is rare. Based on an estimate of approximately 840 patients with SCDys admitted over the study period, our findings indicate that this represents 4% of SCDys patients. An international study of SCDys reported NTSCH accounted for 5% of SCDys and vascular malformation accounted for 4%.¹²

More than half of our cases had an iatrogenic contribution to their non-ischemic vascular SCDys. We presented six different spinal procedures leading to SCDys, but further subgroup analysis on their rehabilitation outcome differences was not performed due to the small sample size. The causative anticoagulant reported in our study was warfarin, and no cases were identified as involving aspirin. Numerous case reports describe the correlation between spinal cord hemorrhage and warfarin,^13–23 and as early as in 1956.²⁴ A review of nine published cases reported minimal or no recovery following surgery for warfarin-associated spinal cord hemorrhage.²² No novel oral anticoagulant agent caused spinal cord hemorrhage in our study. These agents were first introduced in Australia in 2013, and we ended our data collection two years later. Cases of spinal cord hemorrhage related to rivaroxaban, a factor Xa inhibitor, have been reported.^25–29 Two were epidural hematoma and reported complete neurological recovery.^25,26 Another two were due to subdural hematoma, with no neurological improvement after six months,^27,28 and one subdural hematoma, with marked but incomplete recovery.²⁹

Non-ischemic vascular SCDys predominantly resulted in paraplegia. A review of 106 non-traumatic subdural hematoma reported 70% of cases occurred in lumbar or thoracolumbar spine.¹⁰ Previous studies have also reported that paraplegia is the most common level in non-traumatic SCDys, with 67% of 70 cases;³⁰ and in ischemic vascular SCDys, 80% of 104 cases were paraplegia.¹² The result of our study was similar to that of spinal cord infarction, where 93% of 44 cases had paraplegia.²

Patients with non-ischemic vascular SCDys tend to be elderly and just over half the cases were female. The age distribution of this group of patients is consistent with other causes of non-traumatic SCDys, such as degenerative causes and tumors.^12,30 Following non-traumatic SCDys, the LOS in rehabilitation is long and their rehabilitation can be challenging due to frequent interruptions by acute care. ^12,31 The median LOS in our study is comparable to that of 63 days in non-ischemic vascular group from an international comparison study between various SCDys etiologies.¹² Our median LOS is shorter than that of traumatic spinal cord injury, with median LOS 83 days.³² This is expected, given traumatic spinal cord injury is typically associated with greater severity of spinal damage and disability, requiring longer rehabilitation. The median LOS of about two months from our study may guide others in estimating the duration of inpatient rehabilitation following non-ischemic vascular SCDys, keeping in mind that factors such as social and disability supports, individual severity, comorbidities and complications may affect the LOS.

Currently, there is no literature on outcomes of bowel and bladder function following non-ischemic vascular SCDys. In this study, we compare bowel and bladder outcomes following non-ischemic vascular SCDys with those of non-traumatic SCDys.³⁰ Loss of voluntary bowel control after SCDys is socially and psychologically distressing, resulting in significant impact on quality of life.^33,34 Rehabilitation goals include bowel continence, independent defecation, and prevention of intestinal complications. In a study of non-traumatic SCDys, 76% achieved continence with a bowel program, 18% were continent without aperient, 5% were fecally incontinent, and 1 patient had colostomy for bowel management.³⁰ In our study, majority patients achieved bowel continence using a bowel program, no one received colostomy, and the number of patients with bowel incontinence was reduced by two thirds on discharge. We reported higher proportion of patients (25%) who were fecally incontinent on discharge compared to the study of non-traumatic SCDys. Most of these patients were either discharged to nursing home or transferred to another hospital. In our study, bladder function outcomes were similar to non-traumatic SCDys. In non-traumatic SCDys, 48% had bladder continence, 11% used intermittent self-catheterization, 37% used indwelling or suprapubic catheter, and 3% had reflex voiding.³⁰

Recovery of walking is a high priority for patients after spinal cord injury.³⁵ In a prospective review of 62 patients with non-traumatic SCDys, 48% regained some walking capacity after inpatient rehabilitation, but the walking performance remained severely impaired for many.³⁶ In our study, most patients were able to walk 10 m with assistive device, or better on discharge. Some patients who were non-ambulatory or walking with assistive device, may demonstrate further improvement in the post-discharge period and would benefit from having ongoing rehabilitation therapy in the community post-discharge. Significant improvement of the FIM motor subscale, as seen in our group, is comparable to those of spinal cord infarction, where the median FIM motor subscale improved from 28 to 66.² In a cohort of non-traumatic SCDys of various etiologies, the FIM motor subscale improving from 37 to 55.8.³⁷

There is no current literature available on rehabilitation outcomes following non-ischemic vascular SCDys for direct comparison of disability and functional outcomes following NTSCH and vascular malformation of SCDys. However, the overall findings of our study, with improved functional outcomes after rehabilitation, is consistent with those of non-traumatic SCDys.³⁰

Limitations

Retrospective study design is a major limitation of this project but given the rarity of NTSCH and vascular malformations causing SCDys, a prospective study would take many years to complete with adequate numbers. Despite the lengthy study period, the sample size was still relatively small, which increases the chance of bias. Due to the rarity of NTSCH and vascular malformation of SCDys, a multicenter collaboration can provide a better representation of the non-ischemic vascular SCDys population. An additional limitation was the inconsistent documentation of the AIS grade at admission and discharge, preventing us from using this as an outcome measure.

Conclusion

This retrospective study shows that non-ischemic vascular SCDys is rare. The demographic characteristics and neurologic level of classification for non-ischemic vascular SCDys are similar to other causes of non-traumatic SCDys. More than half of the cases were associated with a medical treatment or procedure. Most patients were discharged to home and despite improvement in functional outcomes compared with admission, the majority of patients sustained significant residual deficits at the time of discharge.

Disclaimer statements

Contributors C.P.T. contributed to the design, analysis, writing and editing of the manuscript. P.W.N. contributed to the design, writing, editing and provided advice for the analysis. K.C. contributed to the analysis and editing of the manuscript. The contents represent original work and have not been published elsewhere.

Funding The authors received no specific funding for this work.

Conflict of interest The authors report no conflicts of interest.

Ethics approval The project received ethics approval from the Alfred Health Human Research and Ethics Committee, Victoria, Australia (Project reference number 534/15).