Prognostic Factors in Canine Acute Intervertebral Disc Disease

Knowledge of the prognosis of acute spinal cord injury is critical to provide appropriate information for clients and make the best treatment choices. Acute intervertebral disc extrusions (IVDE) are a common cause of pain and paralysis in dogs with several types of IVDE occurring. Important prognostic considerations are recovery of ambulation, return of urinary and fecal continence, resolution of pain and, on the negative side, development of progressive myelomalacia. Initial injury severity affects prognosis as does type of IVDE, particularly when considering recovery of continence. Overall, loss of deep pain perception signals a worse outcome. When considering Hansen type 1 IVDE, the prognosis is altered by the choice of surgical vs. medical therapy. Concentration of structural proteins in the plasma, as well as inflammatory mediators, creatine kinase, and myelin basic protein in the cerebrospinal fluid (CSF) can provide additional prognostic information. Finally, cross-sectional area and length of T2 hyperintensity and loss of HASTE signal on MRI have been associated with outcome. Future developments in plasma and imaging biomarkers will assist in accurate prognostication and optimization of patient management.


INTRODUCTION
Intervertebral disc disease (IVDD) is a common cause of acute spinal cord injury in dogs, due in large part to the high rates of early disc degeneration in chondrodystrophic breeds (1). Indeed, acute, severe thoracolumbar spinal cord injuries account for ∼4% of cases presenting to emergency rooms in North America, with 74% of these cases due to some form of IVDD (2) and ∼20,000 surgeries for the condition taking place annually (3). Paralysis is an extremely stressful clinical problem for pet owners, who have to process complex information on the underlying disease, the risks and costs of the procedures needed to diagnose and treat their pet, and the possibility that their pet might suffer permanent deficits or death. Thus, rapid and accurate patient assessment and a realistic portrayal of the clinical picture and prognosis is important for the owner at time of presentation to the veterinarian. Moreover, this information aids the veterinarian with appropriate patient triage and with setting realistic functional recovery goals that will allow early detection when a patient is deviating from an expected course. Indeed, the importance of prognostication to humans with traumatic spinal cord injury is such that multivariable clinical prediction models have been developed for this purpose (4). In addition, as increasing numbers of clinical trials are performed in this clinically diverse population of dogs, identifying prognostic biomarkers that quantify injury severity more accurately than clinical assessment alone can refine patient inclusion criteria or serve as covariates in trials, increasing study power and efficiency.
The prognosis, or "before (from the Greek, pro) knowledge (from the Greek, gignoskein)" of a disease is a forecast of disease course following its onset, in this case, using standard treatment. It refers to the possible outcomes of a disease and the frequency with which these outcomes can be expected to occur. A prognostic factor is a measurement that is associated with clinical outcome in the absence of therapy or with the application of a standard therapy (5). A biomarker is "a defined characteristic that is measured as an indicator of normal biological processes, pathogenic processes, or biological responses to an exposure or intervention, including therapeutic interventions. Molecular, histologic, radiographic, or physiologic characteristics are types of biomarkers" (https:// www.ncbi.nlm.nih.gov/books/NBK326791/). Moreover, a prognostic biomarker is "a biomarker used to identify likelihood of a clinical event, disease recurrence or progression in patients who have the disease or medical condition of interest." In the context of intervertebral disc extrusion (IVDE)-induced spinal cord injury, the clinical events being prognosticated include the recovery of independent walking (both speed and level of recovery) and fecal and urinary continence. Resolution of pain is also important, and usually considered for certain subsets of IVDH that cause extreme pain such as cervical IVDE. Finally, the chances of development of progressive myelomalacia, a complication that is usually fatal, is also prognosticated in Hansen type 1 IVDE specifically.
Over the last few decades, there have been numerous studies evaluating prognostic factors for recovery of ambulation and continence after IVDH-induced thoracolumbar spinal cord injury, fewer after cervical spinal cord injury. Many of these studies consider clinical parameters, but their findings can be contradictory. Quantification of lesion extent using imaging, originally using myelography and more recently, magnetic resonance imaging (MRI), has received a lot of attention. Finally, biomarkers measured within the blood and cerebrospinal fluid (CSF) have been investigated. In this article we will consider prognostic factors for dogs with acute spinal cord injury due to acute IVDH. The majority of published data reports outcomes in thoracolumbar IVDH, but data on cervical IVDH have also been included. We will focus on Hansen type 1 IVDE but have also included prognostic information on acute non-compressive nucleus pulposus extrusions (ANNPE) and FCEM (fibrocartilagenous embolic myelopathy) as well as hydrated nucleus pulposus extrusions (HNPE). Fibrocartilagenous embolism is included both because of its clinical similarity to ANNPE and because it represents a form of intervertebral disc-induced acute spinal cord injury.
When presenting data on prognosis to clients, it is extremely important to educate them on what that data represents for their individual dog and on how to use the data in the recovery period. Clients need accurate information in order to make the best decision initially and perhaps the most difficult concept for them is understanding that the data presented represents the behavior of a population; the precise prediction of outcome for an individual is not possible. Once beyond the initial decisions about treatment, providing clients with a timeline for certain thresholds to be crossed, while again explaining the variability across a population and factors that might influence their individual dog, can ensure that a dog that is not following an expected recovery curve, perhaps due to a comorbidity or a complication of the initial injury, is identified in a timely manner. This discussion can also help clients to understand that decisions can be made at many different stages of injury and recovery, and avoid a decision being made in a time of great stress that might be regretted at a later date.

Clinical Presentation
Several different parameters have been evaluated for their prognostic utility including signalment, speed of onset and duration of signs (in particular duration of paralysis), and severity of neurological deficits. The most likely underlying condition is also taken into account because it will affect prognosis. This occurs because different forms of IVDD cause differing types and distribution of spinal cord pathology. The vast majority of these studies are retrospective in nature and variation in patient populations, diagnostic and therapeutic protocols and reporting and methods of follow up make it difficult to draw strong conclusions.

Severity of Neurological Deficits
The exception to the limitation of clinical parameters as prognostic factors is severity of neurological deficits, which is the most immediately accessible, simple and reliable prognostic indicator for animals with spinal cord injury. This is true regardless of neuroanatomic location and is equally true in human spinal cord injury (6). Given the poor regenerative capabilities of the adult CNS, it is not surprising that gauging the extent of permanent tissue loss is extremely important to establish a prognosis. It is common practice to assign animals with thoracolumbar spinal cord injury to one of six categories based on the severity of their clinical signs in a scale known widely as the Modified Frankel Scale (MFS) (7, 8). The frequency of signs within each category in this scale has been estimated ( Table 1) (9). There are numerous slightly different variations on this scale in the veterinary literature, with numbers assigned in different directions and subcategories developed. This makes comparisons between studies confusing and so for the purposes of accurate reporting in this paper, clinical severity has been categorized using description of the signs ( Table 1).

Evaluation of Pain Perception
Prognosis for recovery of independent ambulation is influenced by the presence of pain perception (10)(11)(12). Not surprisingly for such an important clinical variable, there is a range of different terminology used in the literature. Historically the term "deep pain perception" (DPP) has been used-referring to the response to an extremely noxious stimulus applied over the bone of a digit. More recently, the term deep has been omitted and authors use the terms pain perception and nociception. Sometimes "deep nociception" appears. In this article we use the term DPP when discussing prognostic indicators because it is familiar to most veterinarians and conveys the importance of applying a strong noxious stimulus when determining the presence of pain perception. Because of its clinical implications, assessment of pain perception should be made extremely carefully in any animal that lacks motor function. This is performed ideally in a calm animal, using an instrument with relatively wide jaws such as needle drivers or pliers (to avoid cutting the skin as pressure is applied). While placing an animal on its side to perform this test allows a clear view of the response, if they are fighting to get up, it can be difficult to interpret their behavior. If this occurs, the animal should be placed in whatever position allows clear access to the limb being tested with the animal resting quietly. Pressure is applied over the digit being tested and a gentle squeeze is applied to produce a withdrawal reflex (if present) and then pressure is increased until the patient demonstrates perception of the stimulus such as vocalization, looking around, or moving away (Supplementary Video 1). In animals with blunted perception, the response might be as subtle as an alteration in breathing pattern or dilation of pupils. Any repeatable behavioral indication that the animal can feel the stimulus is taken as DPP being present. Both medial and lateral digits should be tested in each foot and the base of the tail should be tested (using the handles of the forceps). Presence of pain perception in any one of these locations places the animal in the prognostic category of having DPP (10,13).

Recovery in Animals With Intact DPP
The prognosis for recovery of independent ambulation and continence in animals that have intact DPP, even if apparently blunted, is good to excellent depending on the treatment pursued ( Table 2) and the type of disc herniation that occurred ( Table 3). The speed of that recovery is influenced by the severity of motor impairment at presentation, altering prognosis for walking at 2, 4-6, and 12 weeks ( Table 2). These benchmarks are extremely useful to indicate when a dog might not be recovering as expected, triggering a timely re-evaluation by the veterinarian. Recovery of fecal and urinary continence in dogs with DPP due to Hansen Type 1 IVDE matches recovery of walking. However, persistent fecal and urinary incontinence have both been reported in animals with incomplete injuries due to ANNPE and FCEM in spite of recovery of ambulation ( Table 3).
Recovery of independent ambulation and resolution of pain in dogs with cervical IVDE has also been reported with and without spinal cord decompression and in general is excellent with surgery (Table 4). However, potentially serious complications of hemorrhage, hypoventilation and bradycardia, vertebral subluxation and aspiration pneumonia have all been reported and the development of an adverse event of this manner does worsen prognosis (28,(42)(43)(44)(45). The majority of dogs with hydrated nucleus pulposus extrusions (HNPE) present with cervical extrusions. These dogs have an excellent prognosis for recovery with or without surgery, even in the presence of respiratory compromise ( Table 4).

Recovery in Animals With No DPP
The prognosis for animals that lack DPP is less certain, with recovery rates for independent walking in dogs with surgically managed thoracolumbar IVDE ranging from 30 to 75% in different studies (11)(12)(13)(14)(15)(16). Overall, ∼60% of dogs with Hansen type 1 IVDE recover DPP and ambulation by 6 months after injury ( Table 2). The timing of recovery of pain perception is important, because once it is present, the prognosis for recovery of ambulation is excellent. One study found that 62% of dogs  (10,(12)(13)(14)(15)(16)(17)(18)(19)(20)(21). that did recover DPP recovered it within 4 weeks, another 30% by 12 weeks and one dog (8%) recovered it at 36 weeks (16). The prognosis for recovery of fecal and urinary continence in these dogs is not quite the same as recovery of independent walking. In dogs with Hansen type 1 IVDE that do recover DPP and walking, ∼40% do not recover normal fecal continence and 30-53% do not recover normal urinary continence (13,16). While in the majority of cases, owners find the level of continence acceptable, it is important to note it might not be normal and accidents will be more likely to happen than prior to injury. The prognosis for recovery in dogs with ANNPE and FCEM that present with paraplegia without DPP is considered poor and the majority of these dogs are euthanized within a week of injury. As such, these cases are scarcely reported in the literature and it is extremely difficult to establish what their prognosis would be if managed long term. There is a recent report of three dogs with ANNPE that recovered walking, but none recovered fecal continence and 2/3 remained urinary incontinent (22). A review of the literature on FCEM revealed seven dogs in this category for which long term outcomes were available; three of the seven recovered ambulation but there is no information on their continence (24). The prognosis of dogs with cervical lesions that lack DPP is difficult to report because so few dogs present with this severity of injury (29). This reflects the high mortality rate due to hypoventilation and brady-arrhythmias in dogs with functionally complete cervical spinal cord injury (46).
There are special considerations in dogs that lack DPP when discussing prognosis. The first is the development of progressive myelomalacia (PMM) and the second is the prognosis for recovery of ambulation if DPP is not recovered. Progressive myelomalacia is a very important consideration due to the gravity of this condition, and the most important risk factor for development of this condition is injury severity (47, 48) ( Table 2). Indeed between 9 and 33% of DPP negative dogs with Hansen type 1 IVDE can develop this condition, with most studies reporting a rate between 9 and 17.5%, and one study reporting a rate of 33% in French bulldogs specifically (11, 14-16, 47, 48). It is extremely important to warn owners of the potential development of PMM in paraplegic DPP negative dogs and to monitor for it both pre and post-operatively (49).
The prognosis of dogs developing independent walking without recovery of DPP has also been investigated in dogs that have suffered Hansen type 1 IVDE. Two studies have reported this specifically and in total 27/88 (31%) dogs that did not regain DPP recovered the ability to walk. In both studies the median time to walking was 9 months with a range of 2-28 months (13,16). It is important to note that while these dogs did regain some ability to urinate and defecate, none had normal continence. The prognosis for recovery of walking in this population of dogs has been reported to be improved to a recovery rate of 59% with intensive physical rehabilitation (50).

Spinal Shock and Schiff Sherrington Posture
The presence of spinal shock and Schiff Sherrington posture at presentation have been described in dogs (9,51). Spinal shock occurs much more frequently with FCEM and ANNPE due to the peracute onset of signs in these conditions (52). The prognostic significance of spinal shock has been evaluated and its presence is associated with the development of fecal incontinence in ANNPE but does not appear to affect recovery of ambulation (22,52). The prognostic significance of the Schiff Sherrington posture has not been evaluated. The assumption of its importance is likely because it is easily recognized in severe thoracolumbar SCI cases, but the presence or lack thereof of DPP should be used as the indicator of prognosis in these cases, as previously discussed.  (11). However, in these examples, the prognostic factors at presentation are severity of signs and location of disc extrusion. Breed has been evaluated as a prognostic factor in several studies, but analysis is somewhat hampered by the overwhelming prevalence of Dachshunds (Table 5). Regardless, no study has found an effect of breed on prognosis. Similarly, sex does not alter prognosis (12,15,47). Several studies have evaluated the effect of age on prognosis and results have been somewhat contradictory, with one study showing increased age slows the speed of recovery in acute TL IVDE, but does not alter the final outcome, another suggesting it reduces the final recovery level, while others show no effect on final recovery (12,16). A clear conclusion on the role of age as a prognostic factor cannot be drawn. Body weight has been investigated and results are similarly conflicting. One study on TL IVDE found increased body weight slowed speed of recovery and another found that dogs that weighed >20 kg had a worse outcome. Other studies found no effect of body weight on final outcome (12,16,47,(54)(55)(56). By contrast, when evaluating a population of non-ambulatory tetraparetic dogs, small breed dogs are six times more likely to have a successful recovery than large breeds (30).

Onset and Duration of Signs
Various studies have evaluated the speed of onset and the duration of signs, in particular the duration from onset of non-ambulatory status to surgical decompression ( Table 6).     These studies are necessarily hampered by reliance on owner observations and periods during which pet dogs are not observed. In addition, the definition of the times can differ between studies with varying definition of onset (onset of ataxia, vs. pain for example), time to presentation or time to surgery as well as different populations of dog being examined. Large case cohorts are presented in Table 6. Overall, there is no consensus on an effect of speed of onset of signs or duration of signs on overall outcome, but there is some evidence that duration of signs might influence the speed of recovery. There is also some evidence that an interval of >12 h between onset of non-ambulatory status and surgical decompression increases the risk of PMM (47). Finally, there is evidence that delaying surgery until the day following presentation increases the risk of clinical deterioration from which the dog might not recover (59).

Location of Intervertebral Disc Extrusion
Several studies have compared outcome in dogs with Hansen type 1 IVDE and found no difference in dogs with T3-L3 vs. L4-S3 localization ( Table 7). One study found a worse prognosis in dogs with lower motor neuron (LMN) signs of incontinence (57). Perhaps the most important prognostic detail in this category is the increased risk of development of PMM with discs located in the caudal lumbar vertebrae in dogs that are paraplegic with no pain perception (47,48). When considering cervical IVDE, adverse events are more likely to occur with disc herniations at C7/T1 (28).

Blood and Cerebrospinal Fluid Biomarkers
While the prognosis for recovery can be established quite well in dogs with DPP prior to diagnostics and treatment, those that lack DPP pose a greater challenge. These dogs dichotomize into a group that shows a recovery comparable to those with incomplete lesions and a group in which there is no or limited recovery. These data show us that there is a profound floor effect when evaluating paraplegic DPP negative dogs and that this group includes dogs that have permanent interruption of conduction and those in which that interruption was temporary, perhaps representing conduction block due to edema, energy failure, etc. As a result, there have been many attempts to identify biomarkers that will allow differentiation of these dogs. Such attempts are plagued by the challenges of the influence of duration of injury and the ability to measure them at the time of presentation, to establish meaningful prognoses for owners. None have yet reached the point of clinical utility. A plasma or serum biomarker would be ideal because it would allow prognostication prior to embarking on expensive advanced diagnostics, but they are somewhat removed from the central nervous system (CNS) compartment and so are likely to require very sensitive measuring techniques. Biomarkers that show the most promise are CNS structural proteins, glial fibrillary acidic protein (GFAP) and phosphorylated neurofilament heavy chain (pNfH) ( Table 8) (66)(67)(68). Both can be measured in the plasma and serum using ELISA and provide some insight into the severity of CNS injury. Of the two, GFAP is the most discriminating at time of injury, predicting both recovery of ambulation and development of PMM with an accuracy of >80%. Serum pNfH concentrations are more variable at the time of injury and S100β was found to be less useful. Unfortunately, at this time there is no rapid point of care test available, but they have proven useful as a covariate of injury severity in clinical trials (69). Evaluation of CSF biomarkers takes the clinician closer to the CNS, but as a result such tests are more invasive and require general anesthesia. Markers evaluated include cytology, myelin basic protein (MBP) with or without creatine kinase (CK), tau, glutamate, matrix metalloproteinase-9 (MMP-9) and inflammatory cytokines (61)(62)(63)(64)(65). Some studies have conflicting results, but all are summarized in Table 8 and MBP and CK CSF concentrations in particular are both highly predictive of return of ambulatory function.

Myelography
The first imaging modality used to prognosticate cases of IVDE was myelography. It was based on the extension of an intramedullary pattern, which was interpreted as indirect evidence of the severity and extent of spinal cord swelling (70). Spinal cord swelling, calculated as the ratio of the length of the loss of the dorsal and ventral contrast columns to the second lumbar vertebra (spinal cord swelling: L2 ratio), was correlated with a poor prognosis when it was found to be five or more vertebral bodies. However, a subsequent study could not confirm these findings (14). In the latter study, the ratio for dogs with a successful outcome was 1.7, compared to 2.0 for those with an unsuccessful outcome. Only two dogs had intramedullary pattern longer than five bodies and both dogs recovered.
Myelographic studies of dogs with acute non-compressive nucleus pulposus extrusion (ANNPE) demonstrated an intramedullary pattern and an additional extradural pattern was seen in approximately half of the dogs. The degree of spinal cord swelling was not associated with severity of clinical signs or outcome (71).
An extensive intramedullary pattern with evidence of contrast medium infiltration into the spinal cord has been reported as an indication of progressive myelomalacia (PMM) (72). Infiltration of contrast within the spinal cord parenchyma, however, is not pathognomonic for PMM since it can also be iatrogenic or represent other intramedullary lesions such as syringohydromyelia (72,73).

Magnetic Resonance Imaging
The utility of MRI for diagnostic purposes has been very welldefined and characterized. Its ability to serve as a biomarker is not as clear, although several studies have proposed imaging markers as prognostic indicators. As MRI is routinely acquired as part of diagnostic work-up of IVDE cases, the identification of reliable imaging markers identified on MRI would be invaluable for clinicians and owners. Spinal cord hyperintensity on T2W images has been the most widely investigated parameter (Figure 1). This spinal cord (SC) hyperintensity identified on T2W images has been associated with necrosis, myelomalacia, intramedullary hemorrhage, inflammation, and edema (74)(75)(76). Without differentiating the pathologic process more specifically, T2 hyperintensity has been shown to correlate well with the severity of neurologic signs at presentation in dogs with IVDE (53,77,78). Its utility as a prognostic indicator is less clear. Even though the first report of the utility of spinal cord hyperintensity indicated that extension of the area of T2W spinal cord hyperintensity on low-field MRI was a reliable predictor of outcome, even more reliable than the absence of deep pain perception (79), these findings could not be reproduced in other studies, primarily those using high-field MRI ( Table 9). Use of high-field MRI leads to increase in signal-to noise ratio and consequently to a change in image resolution; therefore, mild intramedullary hyperintensities in sagittal T2W sequences may be more frequently evident using high field magnetic fields compared to low field ones.
Identification of MRI features suggestive of progressive myelomalacia is crucial for prognostic purposes since its identification indicates an abysmal prognosis. The length of spinal cord T2W hyperintensity and the length of intramedullary pattern reflected as loss of CSF signal on HASTE/MR myelography sequences have been used and are presented on Table 10. A recent 3T MRI study proposed that intramedullary hypointensity on T2W images was associated with PMM (86).
High field MRI changes have been associated with outcome in dogs with ANNPE (22,38) and FCEM (39). A larger lesion on transverse images, quantified as a greater percentage cross sectional area of the spinal cord, has been considered the most useful MRI variable to predict the short-and long-term outcome of dogs with ANNPE and FCEM (22,38,39). In ANNPE, dogs with a smaller lesion had a shorter interval to unassisted ambulation. In contrast, a percentage cross sectional area that equals or exceeds 90% of the spinal cord had a 92% chance of having an unsuccessful long-term outcome (38), and a lesion exceeding 40% of the transverse area has been associated with an increased likelihood of long-term urinary and fecal incontinence (22). However, a low-field MRI study of 21 dogs with ANNPE found no association between any MRI parameter with outcome (71).
Advanced MRI techniques have been proposed to increase reliability of MRI as biomarker. A recent study proposed a semiautomated assessment of SC signal changes aiming to minimize interobserver variability (86). Diffusion tensor imaging (DTI) has been used in dogs with naturally occurring spinal cord injury secondary to IVDE (81,(87)(88)(89). The spinal cord, primarily white matter, microstructural changes are captured through quantification using DTI techniques. As such DTI is able to detect abnormal SC areas that appear macroscopically normal on T2W sequences. Specific DTI parameters, including tractography, have the potential to serve as prognostic biomarkers, although no specific parameter has been identified (81,(87)(88)(89). More information regarding DTI in IVDE can be found in a companion article entitled Diagnostic imaging in intervertebral disc disease.

Electrophysiological Testing
Attempts have been made to use electrophysiological testing to quantify injury severity and predict prognosis, both at the time of injury and in chronically paralyzed dogs. The majority of this work has been completed in dogs with thoracolumbar IVDE. Two different approaches have been used. The first is to evaluate the descending pathways using magnetic stimulation of the motor cortex (90) and the second is to evaluate the ascending pathways within the spinal cord and projecting to the brain using somatosensory evoked potentials (SSEP) (91).

Motor Evoked Potentials
Transcranial magnetic motor evoked potentials (TMMEP) can be elicited reliably in dogs under sedation (92,93). However, they are extremely sensitive to spinal cord injury and are lost completely in dogs that are paraplegic (94,95). With less severe injuries, latency increases and amplitude decreases, but these values do not discriminate initial severity as well as clinical assessment and evaluation of MEPs at time of presentation does not provide prognostic information (94,95). There has been interest in the utility of repeated TMMEP evaluation in dogs that were paraplegic at presentation. Dogs that show recovery of pain perception and motor function recover TMMEPs, leading to the suggestion that this tool can be used to complement assessment of recovery (96). Two groups have evaluated the presence and latency of TMMEPs in dogs that do not regain deep pain perception, and reached different conclusions with one group finding an association between TMMEP presence and recovery or walking, and the other failing to find this association (97,98). At this time, there is no evidence that evaluation of TMMEP at time of injury can provide prognostic information in acute spinal cord injury due to IVDE.

Somatosensory Evoked Potentials
Somatosensory evoked potentials can be elicited by stimulation of a peripheral nerve in a pelvic or thoracic limb (99). Needles are introduced percutaneously to the level of the interarcuate ligament to record from different levels of the spinal cord and subcutaneously to record over the sensory cortex. Various parameters can be recorded including presence or absence of a waveform, latency, amplitude and duration of the potential at the level of the sensory cortex, the conduction velocity of ascending volleys along the spinal cord, particularly across a lesion, the presence and location of conduction block, and the presence and amplitude of an injury potential (100)(101)(102)(103). Early work suggested that lack of recordable cortical evoked potential was associated with failure to recover ambulation in dogs with TL IVDE (100). Another study evaluating conduction velocity and amplitude of spinal evoked potentials recorded at T10/11 found that a ratio of conduction velocity to amplitude was predictive of outcome (101). Later studies did not come to the same conclusions, and combined various parameters to discriminate initial injury severity to the level of clinical evaluation (102). The location of conduction block can be evaluated using the evoked injury potential and the distance between conduction block and site of compression might contribute useful information but this has not been investigated further (103). SSEPs have also been evaluated in chronically deep pain negative dogs, but results are conflicting and are not of prognostic utility (97,98). Currently, there is no clear evidence that prognosis can be established in acute IVDE using SSEPs.

CONCLUSIONS
Understanding the prognosis of spinal cord injury secondary to IVDE is important for client education, optimal patient management and clinical trial design and execution. The prognosis varies with type of IVDE and is influenced by treatment choices. Severity of initial clinical presenting signs is the most useful guide to prognosis at the time of presentation. Biomarkers within the blood, CSF and on imaging can also help to predict outcome. Understanding time to recovery and differentiating between motor and autonomic recovery (continence) can provide invaluable information for veterinarians as they manage dogs that have suffered a spinal cord injury. The manner in which these data are presented and discussed with the owner is a critical part of patient care and it is vital both that accurate information is provided and that owners understand the nature of that information as it relates to their dog. As we move into an era with increased availability of bedside tests of injury severity, and improved accuracy of imaging prognostication, it is likely that patient stratification will improve our ability to perform well-designed clinical trials and to optimize patient care on an individual basis.

AUTHOR CONTRIBUTIONS
NO, RdC, JL, VS, and CANSORT SCI contributed to conception of the study. NO, RdC, JL, and VS designed the study. NO wrote the first draft. RdC wrote a subsection. All authors reviewed, revised, and approved the submitted version.

THE CANINE SPINAL CORD INJURY CONSORTIUM (CANSORT SCI)
• Sarah A. Moore