Utility and Repeatability of Quantitative Outcome Measures to Assess Recovery After Canine Spinal Cord Injury

Utility and repeatability of quantitative outcome measures to assess recovery after canine spinal cord injury

Thesis

Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in the Graduate School of The Ohio State University

Rachel Baek Song, VMD

Graduate Program in Comparative Veterinary Medicine

The Ohio State University

2015

Master’s Examination Committee:

Professor Sarah A. Moore, Advisor

Professor Ronaldo C. da Costa

Professor D. Michele Basso

Copyrighted by

Rachel Baek Song

2015

Abstract

Intervertebral disc extrusion (IVDE) is the most common cause of acute spinal cord injury (SCI) in dogs, and chondrodystrophic breeds such as the dachshund are commonly represented in the literature. The high incidence of spontaneous SCI in dogs makes them an important large animal model for human SCI. Clinical evaluation of dogs with SCI has historically focused on locomotor scoring and only a crude assessment of the ‘presence’ or ‘absence’ of nociception. The development of repeatable, sensitive and objective outcomes measures are of immense importance in establishing dogs with spontaneous SCI as a successful model for human SCI. Clearly defined outcome measures would allow for the identification of potential interventional therapy that may progress into clinical trials in humans. The goals of this study were to (1) identify differences in sensory threshold (ST) values between normal dogs and dogs with spontaneous IVDE and (2) to identify differences in footprint parameters between normal dogs and dogs with spontaneous IVDE to be used as outcome measures from SCI.

Twenty normal dogs and 29 dogs with 30 discrete incidences of spontaneous SCI due to acute IVDE were prospectively enrolled for both studies. Sensory threshold (ST) was measured using an electric von Frey anesthesiometer (VFA) in all limbs of normal dogs at three separate time points and SCI-affected dogs at day 3, 10 and 30 following decompressive surgery. ST values were compared between groups and correlated with locomotor recovery in SCI-affected dogs. ST values were significantly higher (consistent ii with hypoalgesia) in the pelvic limbs of SCI-affected dogs at day 3, day 10 and day 30 when compared to normal dogs while no significant difference in thoracic limb ST values was observed between groups. A progressive decrease in pelvic limb ST values occurred in SCI-affected dogs over time, consistent with improvement toward normal sensation or development of allodynia. This finding correlated inversely with locomotor score. A significant overall decline in ST values across testing sessions was observed for all limbs of normal and SCI-affected dogs. This finding may be related to patient acclimation, operator training effect, or effect of analgesic medications.

Footprint parameters of stride length (SL), base-of-support (BS) and coefficients of variance (COV) were made using a simple ‘finger painting’ technique in all limbs of the same 20 normal dogs and 29 dogs with 30 discrete incidences of IVDE.

Measurements were made at three separate time points in normal dogs and on day 3, 10 and 30 following decompressive surgery in dogs with SCI. Values for SL and BS were compared between groups at each time point. Mean SL in all limbs was significantly lower in SCI-affected dogs at day 3, 10 and 30 compared to normal dogs. The COV of

SL was significantly higher in both thoracic limbs and one pelvic limb in SCI-affected dogs only at day 3 following surgery compared to normal dogs. Additionally, BS in the thoracic limbs was found to be significantly higher in SCI-affected dogs at day 3 and day

30 following surgery compared to normal dogs. BS-PL was not significantly different between SCI-affected dogs and normal dogs.

iii

To my grandfathers Lee Eun Chul and Baek Myung Ik. Thank you for your

teachings on life, love, and perseverance.

Acknowledgments

I would like to thank my advisor, Dr. Sarah A. Moore for her guidance and support throughout this process. I would also like to thank my examination committee for their encouragement and advice. I would especially like to thank Lesley Fisher, who was instrumental in data analysis and manuscript preparation.

I would like to thank Amanda Disher and Heather Myers in their assistance with data collection. Additionally, I would like to thank the entire Neurology and Surgery service for their help in case recruitment.

I would additionally like to thank the OSU Neurology service including Dr. Laurie Cook,

Dr. Michele Provencher, Dr. Ronaldo da Costa, Dr. Sarah Moore, Amanda Disher and

Heather Myers for their constant encouragement through tough times.

Thank you to all my past and present mentors, who have touched my life and shaped me in immeasurable ways. Thank you to my family for making me to be the person I am.

Thank you to my husband Moses Song, for being my emotional support and rock throughout the years and allowing me to pursue my dreams.

Vita

June 2003 ...... Ridgewood High School

2007 ...... B.A. Biological Sciences, Cornell

University

2011...... V.M.D. University of Pennsylvania

2011-2012 ...... Small Animal Rotating Intern, Red Bank

Veterinary Hospital

2012 to present ...... Graduate Teaching Associate, Department

of Veterinary Clinical Sciences, The Ohio

State University

2012 to present ...... Residency in Neurology and Neurosurgery,

Veterinary Medical Center, The Ohio State

University

Publications

1. Song RB, Vitullo C, da Costa RC, Daniels JB. Long term survival in a dog with meningocencephalitis and epidural abscessation due to Actinomyces sp. Journal of

Veterinary Diagnostic Investigation. 2015;27. Accepted April 17, 2015. In press.

2. Song RB, Kent M, Glass EN, Castro F, de Lahunta A. Dyke-Davidoff-Masson like syndrome in a cat. Australian Veterinary Journal. Accepted Apr 13, 2015.

3. Himmel LE, Song RB, da Costa RC, Oglesbee MJ. Pathology in Practice: Primitive neuroectodermal tumor. Journal of the American Veterinary Medical Association.

Accepted Dec 2, 2014.

4. Song RB, Glass EN, Kent M, Sanchez MD, Smith DM, de Lahunta A. Surgical correction of a sacral meningomyelocele in a dog. Journal of the American Animal

Hospital Assocation. 2014;50:436-443.

5. Song RB, Cross JR, Bradley CW, Vite C. Retrospective postmortem evaluation of 435 cases of canine intracranial neoplasia: relationship of neoplasm to breed, age, and body weight (1986-2010). Journal of Veterinary Internal Medicine. 2013;27:1143-1152.

6. Song RB, Cross JR, Golder FS, Callan MB. Chronic urinary and bowel dysfunction in a cat following epidural morphine analgesia. Journal of Feline Medicine and Surgery.

2011;13:602-605.

vii

7. Song RB, Basso DM, da Costa RC, Fisher LC, Mo X, Moore SA. von Frey anesthesiometry to assess sensory impairment after acute spinal cord injury caused by thoracolumbar intervertebral disc extrusion in dogs. The Veterinary Journal. Submitted

Jan 12, 2015.

8. Song RB, Oldach M, Basso DM, da Costa RC, Fisher LC, Mo X, Moore SA. A simplified method of walking track analysis to assess locomotor recovery after acute spinal cord injury caused by thoracolumbar intervertebral disc extrusion in dogs. The

Veterinary Journal. Submitted April 8, 2015.

Fields of Study

Major Field: Comparative Veterinary Medicine

viii

Table of Contents

Abstract ...... ii

Acknowledgments...... v

Vita ...... vi

Publications ...... vii

Fields of Study ...... viii

Table of Contents ...... ix

List of Tables ...... xii

List of Figures ...... xiv

List of Abbreviations ...... xvi

Chapter 1 Introduction ...... 1

Chapter 2 Literature Review ...... 4

2.1 Canine spinal cord injury ...... 4

2.1.1 Causes ...... 4

2.1.2 Pathophysiology of injury ...... 6

2.2 Clinical trials in canine models of spinal cord injury...... 13

2.3 Behavioral tests as outcome measures ...... 18

2.3.1 Locomotor gait scales ...... 19 ix

2.3.2 Electronic von Frey anesthesiometry ...... 26

2.3.3 Footprint analysis ...... 29

Chapter 3 Von Frey Anesthesiometry to Assess Sensory Impairment After Acute Spinal

Cord Injury Caused by Thoracolumbar Intervertebral Disc Extrusion in Dogs ...... 33

3.1 Abstract ...... 33

3.2 Introduction ...... 34

3.3 Materials and Methods ...... 35

3.4 Results ...... 38

3.5 Discussion ...... 42

3.6 Conclusion ...... 46

3.7 Conflict of Interest Statement ...... 46

3.8 Acknowledgements ...... 46

3.9 References ...... 47

Chapter 4 Utility and Repeatability of Footprint Analysis to Assess Gait Recovery After

Thoracolumbar Intervertebral Disc Extrusion in Dogs ...... 58

4.1 Abstract ...... 58

4.2 Introduction ...... 59

4.3 Materials and Methods ...... 61

4.4 Results ...... 64

4.5 Discussion ...... 70

4.6 Conclusions ...... 74

4.7 Conflict of Interest Statement ...... 74

4.8 Acknowledgements ...... 74

4.9 References ...... 74

Chapter 5 Conclusions, Limitations, and Future Directions ...... 87

5.1 Conclusions ...... 87

5.2 Limitations ...... 90

5.3 Future directions ...... 95

References ...... 98

Appendix: Summary of Data ...... 113

List of Tables

Table 1 von Frey anesthesiometry sensory threshold values of thoracic and pelvic limbs across three testing sessions in normal dogs (n=20) (S1= testing session 1, S2 = testing session 2, S3 = testing session 3) ...... 50

Table 2 von Frey anesthesiometry sensory threshold values of thoracic and pelvic limbs in dogs with acute spinal cord inury and T3-L3 myelopathy caused by intervertebral disc extrusion (n=30). ST values are inversely correlated with Olby Spinal Cord Injury Scale

(OSCIS) locomotor scores at days, 10, and 30 after decompressive surgery...... 51

Table 3 Ratio of pelvic limb to thoracic limb sensory threshold values (ST ratio) in normal and spinal cord injury-affected dogs at three time points. ST ratios were significantly higher in SCI-affected dogs compared to controls at 3 days after decompressive surgery (p<0.0001). For SCI-affected dogs, ST ratios were also inversely correlated with locomotor scores at 3, 10 and 30 days after surgery...... 53

Table 4 Mean and coefficient of variance (COV) of stride lengths (cm) for each limb across 3 testing sessions in normal dogs (n=20) ...... 78

Table 5 Mean and coeffcient of variance of stride lengths (cm) for each limb in dogs wih thoracolumbar spinal cord injury due to acute intervertebral disc extrusion at day 3, 10 and 30 following decompressive surgery (n=30). Missing values represent dogs who were notable to consistently walk without support in the pelvic limbs at the time of testing...... 79

xii

Table 6 Mean and coefficient of variance of base of support (cm) in the thoracic and pelvic limbs across 3 testing sessions in normal dogs (n=20) ...... 81

Table 7 Mean and coefficient of variance of stride lengths (cm) for thoracic and pelvic limbs in dogs with thoracolumbar spinal cord injury due to acute intervertebral disc extrusion at day 3, 10 and 30 following decompressive surgery (n=30). Missing values represent dogs that were not able to consistently walk without support in the pelvic limbs at time of testing...... 82

Table 8 Comparison in mean stride lengths of all limbs between the first testing session of normal dogs and all testing sessions in SCI-affected dogs at day 3, 10 and 30 following decompressive surgery Mean SL is significantly lower (p<0.05) in in SCI-affected dogs compared to normal dogs in all limbs at all SCI-affected testing sessions...... 83

Table 9 Comparison in mean base of support of the thoracic and pelvic limbs between the first testing session of normal dogs and all testing session in SCI-affected dogs at day 3,

10 and 30 following decompressive surgery. Statistically significant values (p<0.05) are italicized...... 84

Table 10 Summary data for 20 normal small breed dogs ...... 114

Table 11 Summary data for 30 dogs with acute, spontaneous spinal cord injury due to intervertebral disc extrusion ...... 115

Table 12 von Frey anesthesiometry data for normal dogs (n=20) ...... 118

Table 13 von Frey anesthesiometry data for spinal cord injury affected dogs (n=30) ... 122

Table 14 Locomotor gait scores in dogs with spinal cord injury ...... 128

xiii

List of Figures

Figure 1 Olby spinal cord injury scale (OSCIS) (Olby 2001) ...... 22

Figure 2 Texas spinal cord injury scale (TSCIS) (Levine 2009) ...... 23

Figure 3 Basso-Beattie-Bresnahan (BBB) scale for rat models of spinal cord injury

(Basso 1995) ...... 25

Line represents the median ST value; small box represents the mean ST value. Whiskers represent the 95% confidence interval. Asterisk indicates statistically significant differences from normal dogs (p<0.05)...... 55

Figure 5 Relationship between Olby Spinal Cord Injury Scale (OSCIS) locomotor score and pelvic limb sensory threshold (ST) value in spinal cord injury affected dogs over time following decompressive surgery. A significant inverse correlation between locomotor score and pelvic limb ST value is observed across the 30 day recovery period. ST value displayed is the mean ± SEM value for the pelvic limb with the highest ST value...... 56

Figure 6 Pelvic limb/thoracic limb sensory threshold (ST) ratios in normal dogs (session

1) and spinal cord injury dogs at 3, 10, and 30 days following decompressive surgery. ST ratios were significantly higher in SCI-affected dogs compared to normal dogs at day 3, xiv but not at days 10 or 30. Line represents the median ST value; small box represents the mean ST value. Whiskers represent the 95% confidence interval. Asterisk indicates statistically significant differences from normal dogs (p<0.05)...... 56

Figure 7 Example of footprints acquired and used for analysis. Blue represents the left thoracic limb, purple the right thoracic limb, pink the right pelvic limb, and yellow the left pelvic limb. Lines perpendicular to the walking track drawn through each footprint’s interdigital space (IDS) are shown. Lines that were used to measure stride length (SL), base of support of the thoracic limbs (BS-TL) and pelvic limbs (BS-PL) are depicted. .. 84

Figure 8 Mean stride length (SL) of each limbs in normal dogs (session 1) compared to spinal cord injury (SCI) - affected dogs at day 3, 10 and 30 following decompressive surgery. Whiskers represent ± standard error of the mean. Asterisk denotes statistically significant differences from normal dogs (p<0.05). Mean SL was significantly lower in all limbs in SCI-affected dogs compared to normal dogs. A gradual increasing trend in mean SL is seen in all limbs of SCI-affected dogs with recovery...... 85

Asterisk denotes statistically significant differences from normal dogs (p<0.05). Mean

BS-TL was higher in SCI-affected dogs at all time points when compared to normal dogs.

Mean BS-PL in SCI-affected dogs was not significantly different from normal dogs at any time point...... 86

List of Abbreviations

4-AP: 4-aminopyridine OSCIS: Olby spinal cord injury scale

BBB: Basso-Beattie-Bresnahan PEG: polyethylene glycol locomotor scale QST: quantitative sensory testing

BS: base of support ROS: reactive oxygen species

BS-TL: base of support-thoracic limbs RTL: right thoracic limb

BS-PL: base of support-pelvic limbs RPL: right pelvic limb

COV: coefficient of variance SCI: spinal cord injury

IVDE: intervertebral disc extrusion SEM: standard error of the mean

LTL: left thoracic limb SL: stride length

LPL: left pelvic limb ST: sensory threshold

MFS: modified Frankel score TSCIS: Texas spinal cord injury scale

MRI: magnetic resonance imaging vFA: von Frey anesthesiometry

NSAID: non-steroidal anti-inflammatory drug

xvi

Chapter 1

Introduction

Acute spinal cord injury (SCI) is a common, debilitating and life threatening condition in people. An estimated 276,000 people in the United States live with disabilities from SCI and approximately 12,500 new cases of SCI occur each year

(NSCISC 2014). Depending on the cause and severity of the injury, patients can experience permanent residual motor and sensory deficits that can lead to wheelchair- dependence, development of neuropathic pain, and other life limiting co-morbidities such as urinary tract infections, pneumonia, pressure sores and neuropsychiatric problems

(Kloos 2005, NSCISC 2014). SCI results in tremendous financial burden, with an estimated cost of $1.1 to 4.7 million per patient in lifetime medical expense (NSCISC

2014).

Domestic animals also have a high incidence of SCI due to a variety of different causes including intervertebral disc extrusion (IVDE), fibrocartilaginous embolism and trauma. Intervertebral disc extrusion is by far the most common cause of SCI in dogs. It is estimated that dogs less than 12 years of age have 3.5% prevalence of intervertebral disc disease (Bergknut 2012), and 2% of all cases admitted to a veterinary hospital are due to intervertebral disc disease (Brisson 2010, Webb 2010, Levine 2011).

Intervertebral disc degeneration-related disease incidence for all dogs was estimated to be

27.8/10,000 dog years at risk, with associated mortality estimated to be 9.4/10,000 dog

year at risk, and a case fatality rate of 34% based on an epidemiological study of insurance claims (Bergknut 2012). IVDE can affect any portion of the vertebral column, but approximately 85% of intervertebral disc problems in dogs occur in the thoracolumbar vertebral column (Kinzel 2005). The high incidence of IVDE in dogs offers a valuable spontaneous large animal model for human SCI because of similarities in pathophysiology, and predictable recovery patterns based on clinical severity of injury.

Spontaneous recovery from SCI may occur over time without intervention due to reversible lesions (ie restoration of ionic concentrations, reestablishment of axonal conduction, etc) and plasticity (ie reorganization of white matter pathways) (Jeffery

1999i). However, interventional therapy is typically indicated to remove the primary injury, prevent ongoing damage, and provide the best chance for a functional recovery.

Interventional therapies for SCI have been under intense investigation in both humans and in canine clinical models. However, one of the challenges investigators using canine models face is determining objective measurable outcomes from the intervention under investigation. Small improvements attributable to the interventional therapy in sensory or motor function may represent dramatic improvements in quality of life in people with

SCI such as the ability to operate a wheel chair, sense human touch, and perceive pain for early detection of potentially life threatening infections (McDonald 2002, Kloos 2005).

Although self-reportable in humans, such small changes in sensory or motor function can be difficult to determine in dogs using currently available outcome measures. Therefore, establishing more sensitive and objective outcome measures that can be used to assess recovery in dogs with SCI is of immense importance. Ultimately, this would facilitate 2

identification of modest but important treatment effects in canine clinical studies and would allow further investigation that might progress to clinical trial in humans with SCI.

Chapter 2

Literature Review

2.1 Canine spinal cord injury

2.1.1 Causes

There are many causes of SCI in dogs including intervertebral disc extrusion

(IVDE), fibrocartilaginous embolism, and trauma (Olby 2010). IVDE is the most common cause of acute SCI in dogs, and chondrodystrophic breeds such as the dachshund, cocker spaniel, basset hound, beagle, Pekingese, shih tzu, miniature poodle and bichon frise are commonly represented in the literature (Olby 2003, Ito 2005, Levine

2011, Aikawa 2012ii, Bergknut 2012, Packer 2013). Dachshunds represent the vast majority of reported cases of IVDE and are estimated to have a lifetime incidence of nearly 20% (Brisson 2010, Levine2011, Bergknut 2012) and a relative risk 10-12 times higher than other breeds (Mogensen 2012). Dogs with shorter distal limbs, long back lengths relative to height of the withers, miniaturization, obesity and advanced age were shown to have an increased risk of thoracolumbar IVDE (Levine 2006i, Bergknut 2012,

Packer 2013).

Clinical presentations of thoracolumbar IVDE range from mild back pain to paraplegia with absent nociception of the affected limbs. Prognosis for recovery varies based on the clinical presentation and the treatment pursued. Studies have shown recovery rates of 86%-100% in dogs with thoracolumbar IVDE with intact nociception

following decompressive surgery (Davis 2002, Ferreira 2002, Olby 2004, Ito 2005,

Ruddle 2006, Aikawa 2012i). Dogs presenting with absent nociception have a less favorable outcome with reported recovery rates of 0%-78% following decompressive surgery (Scott 1999, Olby 2003, Ito 2005, Laitinen 2005, Loughin 2005, Ruddle 2006).

Additionally, histopathological examination of the spinal cord in dogs with acute thoracolumbar IVDE found that absent nociception did not always correlate with severe structural spinal cord damage, suggesting a functional rather than structural impairment

(Henke 2013). Due to the variations in outcome in dogs with clinically severe lesions

(absent nociception), it is clear that more objective measurements for severity of SCI are needed to determine an accurate prognosis (Srugo 2011). In addition to absence of nociception, other factors that may indicate unfavorable outcome after decompressive surgery have been studied. These include myelographic evidence of spinal cord swelling

(Duval 1996), hyperintensity within the spinal cord seen on T2-weighted images on magnetic resonance imaging (MRI) (Ito 2005), longer duration of clinical signs (Ferreira

2002), peracute progression of signs (Scott 1999), cranial movement of the cutaneous trunci muscle reflex postoperatively (Muguet-Canoit 2012), increased macrophage to monocyte ratio in cerebrospinal fluid (CSF) (Srugo 2011), and elevated tau protein in

CSF (Roerig 2013). Other factors that have been examined but have not demonstrated prognostic value include motor evoked potentials (Sylvestre 1993), presence or absence of the cutaneous trunci reflex (Gutierrez 2012), and concentrations of glutamate, matrix metalloproteinase-9 and myelin basic protein in the cerebrospinal fluid (CSF) (Olby

1999ii, Levine 2006ii, Levine 2010).

2.1.2 Pathophysiology of injury

Regardless of the cause and species, pathophysiology of acute SCI is similar.

Damage caused by acute SCI has been described in two phases – primary and secondary.

The primary phase of injury refers to the initial mechanical forces responsible for the traumatic event. Primary injuries vary with underlying etiology, but generally include compression, shear, laceration, distraction and contusion to the spinal cord (Jeffery 1999i,

Olby 2010, Webb 2010, Park EH 2012). Depending on the cause, primary injury may lead to loss of axonal impulse conduction, disruption of blood flow, and impaired venous drainage of the spinal cord (Jeffery 1999i). Initial treatment following acute injury is focused on removing the cause of the primary injury (ie disc extrusion, vertebral fracture, foreign body, hematoma, etc), preventing continued damage to the spinal cord and improving spinal cord tissue perfusion (Blaser 2012). Frequently following SCI, a cavitation in the central gray matter at the epicenter of the primary injury site occurs with variable sparing of the peripheral tissues due to cell death during the secondary phase of injury (Lu 2000). Primary injury causes changes on the cellular, molecular and biochemical level that lead to subsequent secondary tissue destruction. It is this secondary phase of injury that is the target of many new pharmacological interventions.

Many portions of the complex secondary injury cascade pathways have been under investigation in attempts to discover novel and targeted interventional therapies that may improve outcome from SCI.

At the start of the secondary phase of SCI, injured neurons release excitatory neurotransmitters such as glutamate and aspartate that cause direct toxicity and ongoing 6

ischemic damage to the spinal cord (Lu 2000, Olby 2010, Park EH 2012). These excitatory neurotransmitters can lead to neuronal cell depolarization and direct opening of voltage gated Ca2+ channels or through activation of ligand gated channels, N-methyl-

D aspartate (NMDA), α-amino-3hydroxy-5-methyl-4-isoxazoleproprionic acid (AMPA) and kainite receptors (Lu 2000). Experimental models targeting ongoing excitations with inhibitors of glutamate membrane transporters, calcium channel antagonists, sodium channel antagonists, NMDA and AMPA receptor antagonists have shown some promising preliminary results (Jeffery 1999ii, Lu 2000, Olby 2010). For example, magnesium, a physiologic antagonist of NMDA receptors can block massive calcium influx and calcium-induced cellular damage, leading to improved functional recovery in rodent models (Nagoshi 2015). This has led to the completion of Phase I clinical trials, and the initiation of Phase II clinical trials in humans (Nagoshi 2015). Hypothermia has also been suggested to have neuroprotective effects by decreasing cell metabolic demand, excitotoxicity, vascular permeability, edema and inflammation (Gensel 2011).

Vascular changes from hemorrhage and ischemia, ionic disturbance and excitotoxicity from cell membrane damage, free radical release, and inflammation can lead to neuron, glial cell and endothelial cell death via apoptosis (Olby 2010). The effects of hemorrhage and ischemia are most pronounced immediately in the gray matter, and later in the white matter (Jeffery 1999i). Secondary hemorrhage from failure of the structural integrity of capillaries causes petechial hemorrhage that eventually becomes a coalescing, hemorrhagic lesion larger than the initial primary site of injury (Gerzanich

2009). Recent studies in rat models of spinal cord injury have shown the role of Trpm4, a 7

nonselective cation channel in perpetuating secondary hemorrhage (Gerzanich 2009).

Additionally, adverse effects on the sympathetic nervous system due to injury involving the neurons of the paravertebral ganglia can also contribute to loss of vascular tone, bradyarrythmias, loss of autoregulation of spinal perfusion, and cause subsequent systemic and local hypotension (Atkinson 1996, Olby 1999i, Smith 2005, Olby 2010,

Park EH 2012). Autonomic dysreflexia characterized by an increase in arterial blood pressure and other signs of autonomic overactivation such as sweating, piloerection, facial flushing, and headaches has been recognized as possible life threatening complications in people with severe thoracic SCI (Krassioukov 2003, Sedý 2008).

Ischemic damage further leads to depletion of cellular ATP stores and decreased

Na+/K+ exchange pump activity. As a result, excessive intracellular Na+ accumulates, causing cytotoxic edema, activation of the Na+/Ca2+ exchange pump, and ultimately increased intracellular Ca2+ concentrations (Jeffery 1999i, Park EH 2012). This has devastating local effects such as inhibition of mitochondrial function, formation of reactive oxygen species (ROS) and induction of apoptosis. Additionally, ROS can go on to cause cell membrane damage through lipid peroxidation, causing dysfunction of proteins and mitochondria. Attempts at preventing damage from increased intracellular calcium through calcium channel blockers, Na+/Ca2+ exchange pumps, and Na+ channel blockers have been studied experimentally with some promising initial results (Jeffery

1999ii). Riluzole, a Na+ channel blocker and anti-glutamatergic, has shown evidence of improved functional recovery, tissue preservation with reduced tissue cavitation, reduced inflammation and apoptosis, and improved somatosensory-evoked potential and 8

locomotor studies in multiple rodent models of SCI (Wu 2013, Grossman 2014). Such findings in rodent models of SCI have led to phase I and phase II clinical trials in human

SCI (Gensel 2011, Grossman 2014). Antioxidant treatment with superoxide dismutase and α-tocopherol has also shown beneficial effects experimentally following SCI (Lu

2000), although another antioxidant N-acetylcysteine failed to show improvement in neurological outcome in a canine clinical trial of SCI (Baltzer 2008). Corticosteroids such as MPSS therapy had once been recommended to inhibit lipid peroxidation and counteract ROS (Bracken 1997), but more recent clinical trials have failed to prove a clear positive outcome for patients and some have reported more adverse effects than benefits (Toombs 1980, Moore 1982, Short 2000, Boag 2001, Levine 2008, Olby 2010,

Shimamura 2010, Hurlbert 2013). Tirilazad mesylate, a derivative of MPSS with similar

ROS scavenging effects but without the adverse effects of glucocorticoids has also shown similar effects as MPSS in human clinical trials (Bracken 1997). Needless to say, corticosteroid use with SCI remains controversial in experimental and clinical settings

(Hurlbert 2013). The role of other free radicals such as nitric oxide has also been investigated in SCI, with some studies suggesting the improved outcomes with inhibition of nitric oxide production in acute SCI (Lu 2000, Zakeri 2014). Dimethyl superoxide

(DMSO) has also been suggested to improve clinical outcomes from acute SCI canine models due to its neuroprotective effects as a free radical scavenger (Levine 2014).

Inflammation occurs shortly after SCI, initially consisting of neutrophils by 12 to

24 hours post injury, then macrophage infiltration by 24 hours post injury with monocyte numbers peaking at 4 to 7 days post injury (Olby 2010, Lee 2011, Srugo 2011). Most of 9

the mechanisms for secondary SCI are initiated by the cellular and humoral mediators of the immune system, with the activation of macrophages as a pathologic hallmark of SCI

(Gensel 2011). Inflammation, especially macrophage function, is initially beneficial, to help remove cellular debris and stimulate axonal regeneration and sprouting through release of protective cytokines (Assina 2008, Shimizu 2010, Srugo 2011). However, overzealous inflammation can also ultimately cause a detrimental effect by releasing toxic chemicals, increasing the permeability of the blood-spinal cord barrier, production of autoantibodies, depression of systemic lymphocyte function, and ultimately causing cell death (Olby 2010, Gensel 2011). Investigators have evaluated the effects of implantation of activated macrophages into rodent and canine models of SCI with mixed results (Assina 2008). Inflammatory cells responsible for phagocytosis can also cause local thrombosis, which further exacerbates ischemic injury (Park EH 2012).

Minocycline, a second-generation tetracycline derivative has been shown to significantly reduce microglia and macrophage activation and decrease the activation of apoptosis.

Several rodent models of SCI have shown improved oligodendrocyte survival, reduction in lesion size, axonal sparing, decrease neuropathic pain, and improved neurological function with minocycline (Lee 2003, Gensel 2011, Sonmez 2013, Nagoshi 2015). Based upon these promising effects in rodent models of SCI, Phase I and II trials in human SCI have been completed, and Phase III trials are currently ongoing (Casha 2012, Nagoshi

2015). Additionally, inhibition of the intraspinal accumulation of monocyte-derived macrophages using agents such as clodronate liposomes, anti-CD11d, anti-Ly6G antibodies found on macrophages and neutrophils, appears to be neuroprotective and improves recovery (Gensel 2011, Lee 2011). 10

Matrix metalloproteinases (MMP), especially gelatinases (MMP-2, MMP-9) have also been shown to play a critical role in inflammation following acute SCI with unique temporal profiles (Zhang 2011). Studies have suggested that early inhibition of MMPs can confer neuroprotection by stabilizing the blood-brain barrier and reducing apoptotic cell death (Levine 2014). The distinctive temporal profiles expressed by different MMPs allows for the opportunity to study the effects of specific gelatinase inhibitors while maintaining the beneficial effects of the MMPs such as axonal regeneration and require apoptosis (Zhang 2011).

Following acute SCI, death of neuronal and oligodendrocytes can occur through necrosis and apoptosis (Emery 1998, Olby 2010). Necrosis is a passive process in which mitrochondrial damage leads to energy loss and disruption of internal homeostasis, whereas apoptosis is an active process of destruction characterized by cell shrinkage, chromatin aggregation, genomic fragmentation and nuclear pyknosis (Emery 1998, Lu

2000). Such programmed cell death has been associated with continuing tissue damage,

Wallerian-like degeneration, and protracted demyelination of ascending and descending spinal tracts in the white matter of the spinal cord. Apoptosis can be initiated by various intracellular and extracellular signals such as activation of the Fas receptor or Bcl-2- related proteins, leading to the triggering of the caspase cascade. Inhibition of different portions of the complex cascade can prevent apoptotic cell death, and have potentially protective effects again SCI (Emery 1998, Lu 2000). Apoptotic cell death occurs acutely after SCI but also continues to occur chronically following the primary injury. This 11

protracted temporal profile of apoptotic injury presents a unique window of potentially neuroprotective therapy that could be applied months after SCI (Lu 2000).

Erythropoeitin, a hematopoietic growth factor is another agent that is being studied for its ability to reduce apoptosis, reduce inflammation, modulate excitability and mobilize proliferation of neuronal stem cells (Gensel 2011, Gorio 2002, Matis 2009).

Acute SCI can trigger the production of inhibitory cues and prevent positive cues for axonal growth, causing overall failure of axonal regeneration in the injured adult spinal cord (Ferraro 2004, Nagoshi 2015). Many regenerative therapies aimed at targeting mechanisms that inhibit axonal growth by neutralization of the inhibitory myelin proteins, or through pharmacological manipulations of inhibitory signaling pathways are under investigation (Ferraro 2004, Gensel 2011, Nagoshi 2015).

Exogenous application of complex acidic glycolipids called gangliosides

(monosialotetrahexosylganglioside sodium salt, or GM-1) was shown to facilitate nerve regeneration (Fehlings 2005). Such promising experimental studies prompted clinical trials in human patients, which failed to demonstrate statistically significant differences from placebo treated patients in larger trials (Hurlbert 2013). Rho protein antagonists such as C3 transferase, an enzyme from Clostridum botulinum, have shown axonal regeneration through areas of SCI with improved motor function, with reduction of apoptotic cells in rodent models of SCI (Nagoshi 2015). Phase I and II clinical trials in human SCI have suggested therapeutic efficacy, leading to the establishment of Phase III clinical trials (Nagoshi 2015). Axonal regenerative and cell replacement therapy using embryonic, mesenchymal and neural stem cells have represented another possible therapy 12

in both experimental and clinical settings (Lim 2007, Ryu 2009, Garbossa 2012, Park SS

2012, Ryu 2012). Recently, the unique properties of specialized glial cells, the olfactory ensheathing cells that reside in the olfactory nerve and outer layer of the olfactory bulb, has been under investigation for its ability to extend primary axons and its possible application for transplant-mediated spinal cord repair (Barnett 2004, Jeffery 2005,

Granger 2012). Clinical trials in dogs with SCI have shown that transplantation of olfactory ensheathing cells in safe and can improve thoracic and pelvic limb coordination

(Jeffery 2005, Granger 2012).

2.2 Clinical trials in canine models of spinal cord injury

In recent years, experimental rodent models of SCI have identified many promising interventional therapeutic. However, translation from rodent studies to human clinical trials has been difficult. Successful “proof of principle” studies performed in laboratory animals with highly uniform injuries have not translated to the larger, heterogeneous human patient population (Jeffery 2005). Potential reasons for failure include species differences, inhomogeneity in lesion type in human clinical trials, and difference in patient demographics in clinical settings. (Jeffery 2006, Jeffery 2011,

Granger 2012). Reproducibility of promising studies at the basic science level have also been challenging, despite most scientists recognizing the importance of replication prior to testing in clinical trials (Steward 2012). Many important factors determine the success of a clinical trial including designing a prospective randomized control parallel group study, ability to confirm positive findings, choosing appropriate outcome measures with high inter-rater reliability that are statistically and clinically relevant, and having clearly 13

defined primary outcomes prior to trial initiation (Lammertse 2013). Several clinical trials in people with SCI have been initiated based on promising data in animal models

(Gensel 2011, Lammertse 2013, Nagoshi 2015). However, despite three decades of clinical trials for SCI in humans, an intervention that has achieved consensus standard of care status has yet to be discovered (Lammertse 2013).

There have been several groups around the world that have used experimental canine models of SCI for various therapeutic interventions (Lim 2007, Assina 2008, Ryu

2009, Park SS 2012, Ryu 2012). However, the high incidence of spontaneous SCI in dogs makes them an important large animal model for spontaneous SCI in humans (Rice

2009). Dogs have comparable mechanisms of injury and resultant pathology to that seen in humans. They also offer a genetically similar, but environmentally heterogenous study population that can bridge the gap between experimental rodent models and the human

SCI population. (Olby 2004, Jeffery 2006). Clinical trials in dogs with spontaneous SCI may lead to the development of interventional therapies that can help both dogs and humans.

Since the 1990’s, there have been several completed clinical trials with spontaneous SCI in dogs (Blight 1991, Borgens 1999, Laverty 2004, Baltzer 2008,

Granger 2012, Levine 2014, Lim 2014). In almost all cases, human studies were initiated based on results of the canine trials. The first canine clinical trial evaluated the effects 4- aminopyridine (4-AP), a voltage gated potassium channel blocker that restores conduction in demyelinated nerve fibers. The study evaluated the effects of oral and 14

intravenous administration of 4-AP and demonstrated transient improvement of proprioceptive placement of the pelvic limbs, pain perception, and cutaneous trunci reflex in a heterogenous population of dogs with SCI of various causes (Blight 1991). Recently, another placebo controlled, blinded clinical trial in dogs with chronic, functionally complete SCI was completed evaluating the effects of 4-AP and t-butyl, a more potent derivative of 4-AP (Lim 2014). This recent study showed improvements in previously validated outcomes measures of treadmill based stepping scores and open field gait scores (Lim 2014). Another clinical trial evaluated the effects of oscillating electrical fields to encourage neuronal sprouting and regeneration in dogs with complete SCI due to

IVDE or trauma (Borgens 1999). A neurologic score combining the findings of ambulation, pain perception and proprioceptive placement, as well as somatosensory evoked potentials, were used as outcome measures. The study showed improved combined neurologic scores in treated dogs compared to control animals (Borgens 1999).

Another clinical trial examining the effects of polyethylene glycol (PEG), an intravenously administered surfactant, and Poloxamer 188, a related co-polymer, was conducted in dogs with complete SCI due to acute IVDE. Similar outcome measures of a combined neurologic score and somatosensory evoked potentials as the previous trial of oscillating electrical fields were used (Laverty 2004). This study also showed encouraging improvements in sensory, proprioception and locomotion in dogs, although a concurrent control population was not evaluated (Laverty 2004). The study suggested the ability of PEG to seal and repair small membrane breaches in damaged axons to improve functional recovery from SCI. A subsequent clinical trial in dogs with SCI comparing the effects of PEG, methylprednisolone sodium succinate (MPSS) and placebo detected no 15

significant difference in locomotor recovery (Natasha Olby, personal communication).

The results of this trial are not yet published (NC State). Another clinical trial in dogs investigated the role of N-acetylcysteine, a precursor to glutathione and anti-oxidant in dogs with acute IVDE (Baltzer 2008). A modified Frankel score ranging from 0 to 2 based upon the presence or absence of motor and pain perception was used as an outcome measure. No significant effect was seen with N-acetylcysteine on neurological outcome in this study, although the outcome measure used may not been sensitive enough to detect any changes. Transplantation therapy of olfactory ensheathing cells have also been under investigation for evaluation of these unique neural cells to regenerate, induce axonal growth, bridge tissue, reduce cavity formation, and improve angiogenesis. Clinical trials in dogs with chronic, complete SCI were recently completed (Jeffery 2005, Granger

2012). Treadmill assisted gait scores were used to determine thoracic-pelvic limb coordination, lateral stability of the pelvic limb steps and electrodiagnostics including somatosensory-evoked potentials, transcranial magnetic motor-evoked potentials and urodynamics were used as outcome measures (Granger 2012). The study revealed that transplantation was safe, and transplanted dogs had improved thoracic-pelvic limb coordination, suggesting improved connectivity across the damaged region of the spinal cord (Granger 2012). Canine clinical trials in dogs with IVDE investigating the effects of intraspinal Schwann cell implantation harvested from nerve roots are also currently ongoing (Andrea Tipold, personal communication). There have been several other studies in experimental canine models of SCI evaluating the effects of transplantation of various stem cells, which show promising results (Lim 2007, Ryu 2009, Park SS 2012,

Ryu 2012). The most recently published canine clinical trial was a double-blinded, 16

randomized, placebo-controlled trial evaluating the effects of DMSO, DMSO with a broad-spectrum MMP inhibitor (GM6001), and saline in dogs with acute SCI due to

IVDE (Levine 2014). Outcome measures used in the study was the Texas Spinal Cord

Injury Score, an ordinal locomotor score. Results of the study revealed improvement in locomotor score in subjects with severe SCI receiving DMSO only and DMSO with

GM6001, suggesting DMSO’s role in recovery.

Over the years, the field of SCI has seen an improvement in canine clinical trial quality and design, although there are still major improvements that need to be made.

The canine population with spontaneous acute SCI due to IVDE can offer a unique large animal model, bridging the gap between rodent models of SCI and the clinical human population. Use of appropriately powered, randomized, placebo controlled, double- blinded studies with strict inclusion and exclusion criteria, and the use of sensitive, objective, reliable, repeatable and valid outcome measures are of upmost importance to establish successful canine clinical trials. Standardized grading of clinical trial quality can also be used in order to detect truly effective therapeutic interventions in spinal cord injury (Kwon 2011). Such critical evaluation of animal models of SCI can lead to the establishment of equally well-designed clinical trials in human SCI for promising potential therapies, without wasting resources and lives in clinical trials on falsely detected potential therapies from poorly designed studies.

2.3 Behavioral tests as outcome measures

Objective determination of SCI and recovery by examining locomotor, sensory, reflex and autonomic function before and after treatment is of the upmost importance in

SCI research (Goldberger 1990, Goldberger 1991, Muir 2000, Basso 2004, Webb 2004).

Improvements in clearly defined, objectively measured outcome assessments can provide opportunities for further studies of a particular therapeutic intervention, initiate clinical trials in animal and human patients, and eventually establish new treatment protocols for

SCI (Lammertse 2013). To this end, there are a variety of different behavioral methods that can be considered in research and clinical settings for functional assessment

(Goldberger 1990). The goal of testing is to associate functional deficits with lesion severity, to document the extent of recovery following interventions, and identify the integrity of motor and sensory systems that may be the substrates of recovery from SCI

(Basso 2004). It is important that each behavioral test chosen as outcome measures for

SCI be sensitive, reliable, standardized and generalizable across different testers, injury severity and species (Basso 2004). Each behavioral test will have associated advantages and disadvantages (Sedý 2008). As long as the examiner is aware of the differences between the behavioral tests and minimizes the limitations, each test may provide valuable information on recovery assessment. It is important to keep other variables that may affect animal behavior as constant as possible such as testing time of day, order of testing, testing environment, and the examiner (Basso 2004). It has further been suggested that prospective clinical trials be controlled for the severity of the lesion and its location by grouping animals with similar lesions (ie similar quantitative gait scores), include both sensory and motor tasks in behavioral analysis, control for the morphology 18

of the animal, and ensure identical environmental enrichment and activity levels (Webb

2004).

2.3.1 Locomotor gait scales

Quantitative gait analysis is recognized to be an important clinical tool as well as an important outcome measure in clinical trials to determine the efficacy or non-efficacy of a particular intervention in people and animals with SCI (Ditunno 2009). Various locomotor scales have been developed to assess recovery from SCI in people and domestic animals. In people, different scales measuring capacity or disability

(performance) are used to assess recovery from SCI, determine need for device assistance, and measure an individual’s independence from human assistance (Ditunno

2013). An ideal rating scale needs to be sensitive (able to detect differences in performance with varying injury severity), reliable (produces consistent scoring between different investigators) and valid (able to capture different stages of locomotor recovery)

(Basso 2006). Normal locomotion in quadrupeds is typified by weight supported plantar steps, coordination between the forelimbs and hindlimbs, consistent positioning of the paw during stepping, adequate toe clearance and maintenance of a stable trunk (Basso

2006). In rats, weight distribution is equal over the thoracic and pelvic limbs on a level surface, and the thoracic limbs are used for breaking, whereas the pelvic limbs are used for propulsion (Webb 2002). Differently in dogs, the thoracic limbs carry approximately

60% of body weight, whereas the pelvic limbs carry 40% of body weight. Additionally, the thoracic limbs spend 50% of the stance phase in braking and the other 50% in propulsion, whereas the pelvic limbs spend 35% in braking and 65% in propulsion (Foss 19

2013). Such differences in normal locomotion should be considered when using gait scales as outcome measures from SCI in the particular species of interest. Due to gait differences among species, gait scales established for a specific species should not be used for others. Additionally, different locomotive patterns are believed be influenced by particular spinal pathways. For example, overground locomotion is believed to be initiated by descending supraspinal pathways, quadrupedal locomotion evaluated on a treadmill is dependent upon propriospinal pathways between the thoracic limbs and pelvic limbs, while bipedal locomotion of the pelvic limbs is dependent primarily upon segmental spinal pathways (Goldberger 1991). Prior studies in rodent and feline models of SCI have revealed the importance of lesion location in locomotor function. Animals with preservation of the ventral spinal cord segments and associated reticulospinal tracts recover overground locomotor function much more readily than animals with lesions within the dorsal spinal cord segments and associated corticospinal and rubrospinal pathways (Schucht 2002, Collazos-Castro 2006). The modified Frankel score (MFS),

Olby spinal cord injury scale (OSCIS; Olby 2001) and the Texas spinal cord injury score

(TSCIS; Levine 2009) are examples of quantitative scales of overground locomotion that have been utilized in dogs with SCI.

The 5-point MFS is a practical, widely used, easily learned scale that requires minimal observer training. It has proven to be a useful scale when assigning clinical prognosis to patients. There is also good agreement between prospectively and retrospectively determined MFS scores on the same dogs (Van Wie 2013). However, it is not sensitive enough to monitor recovery patterns for experimental models of spinal cord 20

injury as there is a wide spectrum of presentation for each single point on the scale. The

MFS also has not been formally examined for inter-observer variability or external validity despite its widespread use (Levine 2009). Additionally, the MFS scale itself is inconsistent within the literature with multiple variations and sub-scoring.

The OSCIS is a 14-point canine gait scale that was loosely based upon the Basso-

Beattie-Bresnahan (BBB) scale developed for rats (Olby 2001, Figure 1). Advantages of the OSCIS include a more detailed scoring system that allows better quantification of recovery from SCI, relative ease of learning the scale, and is a validated gait scale to be used in dogs with thoracolumbar spinal cord injuries. Some disadvantages of the OSCIS gait scale include the need for video-based training, the need to provide weight support in non-ambulatory dogs, which may distort clinical findings and pelvic limb movements

(Handa 1986, Naito 1990i, Naito 1990ii), and the lack of operational definitions for behavioral categories and outcomes can make scoring difficult for one unfamiliar to the scale. Additionally, large changes in function may correspond to only a small number change along the scale.

Figure 1 Olby spinal cord injury scale (OSCIS) (Olby 2001)

The TSCIS is a 40-point scale that has been developed for assessment of SCI- affected dogs (Levine 2009, Figure 2). The primary advantage of the scale is that all four limbs may be scored separately enabling its use for both thoracolumbar and cervical SCI.

Additionally, video-based training is not necessary for the observer, and other aspects of the neurological examination including postural reactions and nociceptive testing are included within the gait score. Disadvantages of the TSCIS include use of weight

support for gait determination, which may distort the clinical findings, use of the controversial “deep” and “superficial” nociception evaluation (de Lahunta 2009), and large variations in gait in dogs with an identical score due to the additive nature of the scoring. Additionally, incorporation of clinical findings such as quality of postural reactions and nociception introduces subjectivity. The TSCIS has also been shown to be prone to inter-abstractor variability and bias when prospectively and retrospectively assigned score for the same dogs was compared (Van Wie 2013).

Figure 2 Texas spinal cord injury scale (TSCIS) (Levine 2009)

The BBB scale is a locomotor rating scale originally developed for rat models of thoracolumbar SCI (Basso 1995, Figure 3). The BBB scale was developed through observations of rats with established scoring categories for early, intermediate and late phases of recovery from SCI (Basso 1995). The BBB scale is based on a 21-point scoring system that spans all lesion severity types. It is currently the most widely accepted, descriptive and sensitive scale in discriminating phases of recovery available to date in rats (Olby 2001, Koopmans 2005). The scale minimizes user variability, one of the inherent limitations of subjective and semi quantitative scales, by eliminating vague terminology and using strict operational definitions (Basso 2004). Sub-scoring can also be used when certain treatments affect some, but not all aspects of locomotion (Basso

2004). Lastly, the BBB can be employed across all injury severities, a characteristic that is uncommon amongst other behavioral assessments in SCI. The BBB scale has been successfully adapted for other species including mice (Basso 2006) and opossums (Wang

1998). It has been used in canine SCI studies (Jeffery 2005, Fukuda 2005, Park 2012) although no validation studies have been performed to date in dogs. Some limitations of the BBB scale include its utility only in rats with mid to lower thoracic spinal contusion injuries, non-linearity of the scale, the prerequisite for thoracic limb and pelvic limb coordination for higher scores, and the lack of validation studies for dogs with spinal cord injury (Basso 1996, Hamers 2001, Basso 2004, Koopmans 2005, Kloos 2005, Hamers

2006, Gordon-Evans 2009).

Figure 3 Basso-Beattie-Bresnahan (BBB) scale for rat models of spinal cord injury (Basso 1995)

2.3.2 Electronic von Frey anesthesiometry

Quantitative sensory testing (QST) enables objective evaluation of sensory function and dysfunction including those of pain such as allodynia and hyperalgesia in response an array of stimuli (Shy 2003, Felix 2009, Walk 2009). QST is an important part of outcome assessment in animal models of acute and chronic pain where therapeutic interventions are intended for eventual translation to human patients (Vierck 2008, Felix

2009, Mogil 2009, Rice 2009, Lascelles 2013). QST modalities that have been evaluated in animals and humans include mechanical stimuli such as light touch, punctate, pinprick, pressure and vibration and thermal stimuli including heat and cold. Other types of quantitative measures used in SCI include somatosensory evoked potentials, contact heat- evoked potentials, electric, thermal or vibrational perceptual threshold, and autonomic measures such as the sympathetic skin response, postural challenge, quantitative sudomotor axon reflex, Valsalva maneuver, and cardiovagal heart rate (Boakye 2012).

Von Frey anesthesiometry (VFA) is one such sensory evaluation technique first developed by physiologist Maximilian von Frey at the end of the 19th century. Since then, there have been several adaptations of his technique that have been applied primarily in rodent models and human patients. VFA has been applied in both conscious and anesthetized animals where measured end points include a behavioral response, nerve traffic and neuronal discharge (Lambert 2009). Initially, von Frey monofilaments

(“hairs”) were used for sensory threshold (ST) testing but recently, an electronic von Frey anesthesiometer was developed for simplification. The electronic von Frey anesthesiometer delivers a punctate stimulus at a given pressure, which activates low 26

threshold mechanoreceptors such as Merkel discs and Ruffini endings in the skin, which stimulates the A-delta and C-fibers at a given pressure (Walk 2009, Tena 2012). ST is measured by determining at which pressure point the patient senses the rigid probe that is being applied to the skin. Patients with hypoalgesia have increased sensory thresholds, whereas patients with hyperalgesia and allodynia are expected to have decreased sensory threshold when compared to those with normal sensory function (Hoschouer 2010). The value of sensitive QST is increasingly recognized for evaluation of interventions that treat acute or chronic pain. In dogs, VFA has been used to evaluate hyperalgesia in orthopedic disease (Brydges 2012), antinociceptive effects of analgesics (KuKanich 2005i, Kukanich

2005ii, Kukanich 2011) and ST in a small group of dogs with acute SCI (Moore 2013).

Sensory testing in dogs with SCI has been historically limited to testing for the presence or absence of a behavioral response to pinching the toes, which is interpreted as the presence or absence of nociception (Lascelles 2013). Although the presence or absence of nociception remains an important prognostic indicator for recovery, it is a subjective and crude assessment of sensory function. At best, response to pinching of the toes can be evaluated as normal, diminished or absent. Recently, histopathological examination of the spinal cord of dogs with absent nociception showed varying degrees of damage from mild to severe, which suggested that severe clinical signs do not always correlate with severe structural damage (Henke 2013). In rodent models of SCI, it has been found that only small amounts of white matter sparing are required for normal sensory function. This suggests that the small diameter, peripheral, ascending sensory

axons (ie spinothalamic fibers) may be less susceptible to injury or display greater neuronal plasticity in comparison to descending motor axons (Kloos 2005).

A limitation of VFA is the varied behavioral response evoked by the test in individual animals. Some animals may be more stoic than others and thus may tolerate a higher pressure prior to a behavioral signal such as withdrawal, is given. Conversely, an anxious animal may display premature behavioral responses in response to stimulus.

Additionally, the determinants of normal or abnormal behavioral responses remain controversial, with inconsistencies reported in the literature (Kloos 2005).

Despite these limitations, VFA shows promising potential for objective determination of sensory dysfunction seen with acute SCI, and documenting changes in sensory function in dogs with therapy following SCI. Objective evaluation of sensory recovery would allow researchers to measure small improvements that otherwise may have remained undetectable. Detection of such small, measurable improvements could support the progression of a treatment intervention from a pilot study to clinical trials.

Such information may also provide important prognostic information for clinicians on quality of life issues by predicting improvements in proprioception, incidence of pressure sores and self-mutilation, and associated infections and its medical care costs (McDonald

2002, Moore 2013). The electronic VFA is relatively inexpensive, simple to use, portable, well tolerated by animals, does not cause tissue damage, and produces reliable results (Tena 2012, Moore 2013). Additionally, VFA has been used to evaluate animals for the development of allodynia, or painful responses to normally innocuous stimuli 28

following experimental nerve injury or SCI (Christensen 1997, Kloos 2005, Detloff 2010,

Hoschouer 2010). It has been previously shown in rat and mouse models of SCI that mechanical allodynia can occur in those with more severe injury when compared to those with mild injuries (Kloos 2005, Hoschouer 2010). VFA thus may be applied in companion animals to investigate causes for continued pain and discomfort following treatment for SCI. Longitudinal changes in sensory function during recovery from SCI will require further investigation.

2.3.3 Footprint analysis

Quantitative kinematics, kinetic gait analysis and footprint analysis are some of the most reliable methods of functional behavioral testing in spinal cord research (Basso

2004, Gordon-Evans 2009). Footprint analysis is most often used to assess return of pelvic limb function following sciatic nerve injury (de Medinaceli 1982, Varejao 2004) and SCI (Cheng 1997, Klapdor 1997, Hamers 2001, Hamers 2006, Gordon-Evans 2009,

Rangasamy 2013). Measurements such as base-of-support (BS), stride length (SL) and paw rotation can provide valuable information regarding an animal’s pattern of locomotion (Kunkel-Bagden 1993, Klapdor 1997, Hamers 2001, Hamers 2006,

Rangasamy 2013). Gait differences not detected through visual assessment may be detected through walking track analysis via footprint assessment (Ballermann 2006).

Additionally, SCI-affected animals may adapt locomotor gait patterns that differ from normal animals and only detectable with quantitative measurements, despite perceived full locomotor recovery on gait score (Webb 2002, Ballermann 2006). In such cases, footprint analysis may allow for more reliable detection of objective signs of recovery in 29

higher functioning animals, such as those in the latter stages of recovery or those with mild SCI. Gait analysis in humans and animals with neurologic impairments has shown altered spatial and temporal aspects of gait. Dogs with spinal cord disease exhibit an uncoordinated gait that is quantifiably different from dogs with lameness due to orthopedic disease (Gordon-Evans 2009). Dogs with orthopedic disease develop a predictable lameness and gait pattern whereas dogs with spinal cord disease have increased variation in their gait that is unpredictable (Gordon-Evans 2009). Coefficients of variance (COV) for stride length and swing time have previously been shown to be significantly larger in dogs with neurologic disease when compared to normal dogs

(Gordon-Evans 2009). Therefore, kinetic gait analysis can be helpful in differentiating neurologically abnormal dogs from dogs with orthopedic disease, which can be difficult with just visual gait examination.

The CatWalk method of walking track gait analysis was originally developed to better assess thoracic limb and pelvic limb coordination (Hamers 2001), an important biological marker for recovery from SCI (Basso 2004). The CatWalk consists of an elevated glass walkway on which an animal’s gait parameters such as pressure estimates, inter-limb coordination, regularity of step patterns, stride length (SL), base-of-support

(BS), swing duration, paw print size, and relative paw positions are recorded (Hamers

2001, Hamers 2006). Thus, the CatWalk allows for quantitation of a large number of locomotion parameters over a set distance (Koopmans 2005). Previous studies in animal models of thoracolumbar spinal cord injuries have revealed decreased SL in the thoracic limbs (Hamers 2001, Plemel 2008, Gordon-Evans 2009) and pelvic limbs (Kunkel- 30

Bagden 1990, Stokes 1992, Kunkel-Bagden 1993, Keirstead 2005, Hamers 2006, Plemel

2008, Gordon-Evans 2009), decreased BS in the pelvic limbs (Cheng 1997, Hamers

2006), increased BS in the pelvic limbs (Kunkel-Bagden 1990, Behrmann 1992, Stokes

1992, Kunkel-Bagden 1993, Metz 2000, Hamers 2001), increased print area (Hamers

2001, Hamers 2006), and changes in step sequence distribution (Hamers 2001). Based on the multitude of measurable gait parameters, the CatWalk has the potential to differentiate between injury type, and identify specific spinal cord tracts involved with the lesion such as the corticospinal and rubrospinal tract (Hamers 2001). For example, studies in rat dorsal spinal cord contusion models using the CatWalk showed the importance of ventral spinal cord segments in preservation of coordination, similar to previous cat models of dorsal spinal cord hemisection (Hamers 2001). Future application of the CatWalk would allow further study of specific spinal cord tracts and their roles in locomotion in different species. The CatWalk has also been utilized to detect conditions of mechanical allodynia in rats, and may be a powerful additional tool for assessing sensory function in addition to VFA (Vrinten 2003).

The biggest disadvantage of footprint and automated walking track kinetic gait analysis (such as the CatWalk and Tekscan®) is that the necessary equipment is expensive and not widely available. The other disadvantage is that animals need to be in the more advanced stages of neurologic recovery (i.e. consistently stepping) for accurate assessment. Animals also need to be trained to properly walk across the walkway prior to injury, which is not possible in a veterinary clinical trial.

Overall, footprint analysis allows the investigator to evaluate specific spatial gait parameters that are not attainable through other behavioral tests. With CatWalk, additional temporal gait parameters such as swing phase and stance time can also be determined to further assess recovery. Lastly, components of inter-limb coordination including pattern, consistency and step ratio, can be evaluated that may otherwise be missed in open field visual observation (Kloos 2005).

Chapter 3

Von Frey Anesthesiometry to Assess Sensory Impairment After Acute Spinal Cord Injury Caused by Thoracolumbar Intervertebral Disc Extrusion in Dogs

3.1 Abstract

Sensory threshold (ST) was measured using an electric von Frey anesthesiometer

(VFA) in all limbs of 20 normal dogs and 29 dogs with acute thoracolumbar spinal cord injury (SCI) caused by spontaneous intervertebral disc extrusion. ST values were measured at three separate time points in normal dogs and on day 3, 10 and 30 following decompressive surgery in dogs with SCI. ST values were compared between groups and correlated with locomotor recovery in SCI-affected dogs.

ST values were significantly higher (consistent with hypoalgesia) in the pelvic limbs of SCI-affected dogs at day 3, day 10 and day 30 when compared to normal dogs while no significant difference in thoracic limb ST values was observed between groups

(p < 0.05). A progressive decrease in pelvic limb ST values occurred in SCI-affected dogs over time, consistent with improvement toward normal sensation or development of allodynia. This finding correlated inversely with locomotor score (p <0.05, r range -0.44 to -0.63 across 30 days). A significant overall decline in ST values across testing sessions was observed for all limbs of normal and SCI-affected dogs. This finding may be related to patient acclimation, operator training effect, or effect of analgesic medications. This study supports the use of VFA to assess differences in ST between normal and SCI-affected dogs. However, future studies must focus on techniques to 33

minimize or compensate for clinical, environmental and behavioral factors that cannot be controlled during clinical trials but may impact ST values in the clinical setting.

3.2 Introduction

Acute spinal cord injury (SCI) is a common neurologic problem in dogs (Olby

2003). Despite the prognostic significance of diminished nociception in canine SCI, clinical evaluation of dogs with SCI has historically focused on locomotor scoring and only a crude assessment of the ‘presence’ or ‘absence’ of nociception (Olby 2001, Levine

2009, Lascelles 2013). Abnormalities in sensory processing such as allodynia, hyperesthesia, and neuropathic pain have also yet to be explored in dogs with SCI, despite being well documented in rodent models and humans with SCI (Lindsey 2000,

Hayes 2002, Carlton 2009, Felix 2009, Densmore 2010, Hoschouer 2010).

Recent studies suggest using an electronic von Frey anesthesiometer (VFA) is feasible in dogs as an objective mechanical quantitative sensory test (QST) (Moore 2013,

Briley 2014). This technique has been used previously to assess hyperalgesia in dogs with orthopedic disease (Brydges 2012), anti-nociceptive effects of analgesics (KuKanich

2005i, KuKanich 2005ii, KuKanich 2011), and evaluate sensory threshold (ST) in a small number of dogs with acute SCI (Moore 2013). When assessing patients with SCI, increases in ST above baseline are generally interpreted to represent hypoalgesia while decreases in ST below baseline are representative of allodynia or hyperesthesia (Detloff

2010, Moore 2013).

Our aim was to compare ST values obtained using VFA between normal dogs and a cohort of dogs with acute thoracolumbar SCI caused by intervertebral disc extrusion

(IVDE), and to document how ST values changed in SCI-affected dogs over a 30-day period of neurologic recovery. We hypothesized that pelvic limb ST values would differ between normal dogs and those with thoracolumbar SCI, while thoracic limb ST values would not. We also hypothesized that pelvic limb ST values in SCI-affected dogs would have an inverse correlation with improving locomotor scores, consistent with recovery of sensory function and/or development of allodynia in the weeks following SCI.

3.3 Materials and Methods

The study was approved by The Ohio State University (OSU) Clinical Research

Advisory Committee and the Institutional Animal Care and Use Committee

(2012A00000149). Written owner consent was obtained prior to study enrollment.

Normal dogs

Twenty apparently healthy adult dogs were recruited from The OSU Veterinary

Medical Center. Dogs had no prior history of neurologic or orthopedic disease and were of a small breed (≤ 20 kg). All dogs were assessed to be neurologically and orthopedically normal based on examination by two of the investigators (RBS and SAM), with the exception that valgus and varus conformational limb abnormalities typical for chondrodystrophic breeds were considered acceptable for enrollment to facilitate generalization of our results across a realistic clinical population.

An electronic VFA device (IITC, Woodland Hills, CA) was used for ST measurement in all four limbs. ST testing was performed in a quiet room with minimal traffic as previously described (Moore 2013). Testing order of the limbs was decided by a coin toss and was recorded. For pelvic limbs, the electronic VFA probe was applied perpendicular to the dorsal surface of the metatarsus, halfway between the tarsometatarsal and metatarsophalangeal joints between digits IV and V (Moore 2013). This region lies within the cutaneous autonomous zone of the fibular branch of the sciatic nerve. For thoracic limbs, the electronic VFA probe was applied perpendicular to the dorsal surface of metacarpus, halfway between the carpalmetacarpal and metacarpophalangeal joints between digits IV and V. This region lies within the cutaneous autonomous zone of the radial nerve. Dogs were prevented from visualizing the device during application to ensure limb withdrawal or other behavioral responses were due to tactile stimulation

(Detloff 2010). Gentle, constant and progressively increased pressure was applied until the dog displayed a behavioral response to the stimulus, regarded as withdrawal of the limb, in conjunction with a conscious response such as vocalization, or lip licking.

Immediate withdrawal of the limb upon application of the probe without any pressure was determined to be a reflexive movement or a product of proprioceptive input rather than a conscious response to tactile stimulus and was discarded and the stimulus repeated in one minute (Kloos 2005, KuKanich 2005i, Detloff 2010). The evaluator (RBS) was blinded to the pressure reading during testing. The minimum pressure required to elicit a behavioral response was recorded. The test was repeated five times in each limb, with each test separated by one minute to avoid windup, ST decay, and hypersensitization

(KuKanich 2005ii, Detloff 2010, Detloff 2012). The highest and lowest ST values were 36

excluded and the three middle values averaged to assign a single ST value to each limb

(Moore 2013). ST testing was repeated three times at least 48 hours apart in all normal dogs.

Affected dogs

Twenty-nine dogs adult dogs with 30 discrete episodes of acute T3-L3 myelopathy caused by IVDE were consecutively and prospectively enrolled from the general patient population at OSU Veterinary Medical Center. Dogs were eligible for enrollment if diagnostic testing (CT, CT and myelogram, or MRI) confirmed IVDE, and they weighed ≤ 20 kg. A modified Frankel Score (MFS) on admission of ≥ 1, indicating intact nociception, as assessed by both the attending clinician and the investigators, was required for enrollment. All dogs underwent surgical decompression for their IVDE. ST testing of all four limbs using the technique described above was performed at three time points: 3, 10 and 30 days after surgery. Each affected dog was also assigned a locomotor score by the investigators using the Olby Spinal Cord Injury Scale (OSCIS) (Olby et al.,

2001) at each time point. Analgesic and/or anti-inflammatory medications were prescribed for all patients during the perioperative period with dosing at the discretion of the attending clinician. All medications that the subjects were receiving at the time of testing were recorded.

Statistics

Summary statistics including mean and standard deviation or median and range where appropriate are reported for clinical data on all dogs, and for ST values for all testing sessions. Data for ST values and pelvic limb to thoracic limb ST ratios were compared across three testing sessions in normal dogs using a mixed effect model, incorporating repeated measures for each subject (Verbeke 2000). Comparison of ST values across three testing sessions for SCI-affected dogs was performed using a trend analysis. With this model, a p-value of < 0.05 was considered significant for all analyses.

Analyses were conducted using SAS software (SAS, Inc; Cary, NC).

3.4 Results

Normal dogs

Normal dogs ranged in age from 8 months to 6.5 years (median 3 years) and weighed between 3.7 kg to 17.2 kg (median 9.4 kg). There were 8 spayed females and 12 castrated males. Breeds were as follows: mixed breed dogs (6), dachshunds (4), miniature schnauzers (2), Sealyham terriers (2), beagle (1), bichon frise (1), cocker spaniel (1), Pembroke Welsh corgi (1), miniature pinscher (1), and shih tzu (1). Time period between testing sessions for each dog ranged from 2 to 27 days (median 6 days).

Affected dogs

A total of 29 dogs with 30 discrete episodes of acute SCI caused by IVDE were enrolled. One dog with an acute onset of IVDE and surgical decompression recovered fully, then experienced a second episode of IVDE six months later. This dog was enrolled 38

twice in the study, with each episode analyzed as a separate event. Dogs ranged in age from 2 to 11 years (median 5 years) and weighed between 3.9 kg to 17.0 kg (median 8.2 kg). There were 14 spayed females, 13 castrated males, and 2 intact males. Breeds were as follows: dachshunds (12), mixed breed dog (6), French bulldog (4), beagle (2),

Pembroke Welsh corgi (2), shih tzu (2), and cocker spaniel (1).

All dogs underwent decompressive hemilaminectomy or pediculectomy at one or multiple sites between T10-11 and L3-4 intervertebral disc spaces, with or without one or more lateral disc fenestrations dependent on imaging results and discretion of the surgeon. Postoperative analgesic dosage and type was dependent upon the surgeon’s preference but included a fentanyl constant rate infusion for12-24 hours postoperatively, a fentanyl patch placed immediately postoperatively, and combinations of tramadol, gabapentin, methocarbamol, or diazepam. Postoperative anti-inflammatory therapy generally included tapering anti-inflammatory doses of prednisone, or a non-steroidal anti-inflammatory drug (NSAID). The medication doses, frequency of administration, and number of total medications were recorded for each dog at each session.

Von Frey Anesthesiometry Sensory Threshold (ST) Values of Normal Dogs

Mean ST values for normal dogs across three testing sessions are summarized in

Table 1. There was no significant difference between ST values obtained from the left thoracic limb (LTL), right thoracic limb (RTL), left pelvic limb (LPL), or right pelvic limb (RPL) of normal dogs between sessions 1 and 2 (p = 0.18, 0.43, 0.25, 0.39 respectively) or between sessions 2 and 3 (p = 0.31, 0.68, 0.39, 0.41). When comparing 39

sessions 1 and 3, a significant decrease in session 3 was noted in ST values for the LTL

(p = 0.02) and LPL (p = 0.04) and values for the RPL approached significance (p = 0.09).

ST values differ in pelvic but not in thoracic limbs between normal and SCI-affected dogs

ST values obtained from SCI-affected dogs at three time points after injury are summarized in Table 2. An overall trend towards decreasing ST values was seen in all limbs across testing sessions in SCI-affected dogs. ST values from the limbs of normal dogs at session 1 were compared to ST values from SCI-affected dogs at days 3, 10, and

30 (Figure 4). ST values derived from session 1 was used in order to minimize the effects of acclimation and/or investigator training that may contributed to lower ST values at subsequent testing sessions 2 and 3. Since ST values of the affected limbs in

SCI-affected dogs was hypothesized to be higher compared to normal dogs, the use of ST values from session 1 thereby represented the most stringent test to find a difference between normal and SCI-affected dogs.

Statistical analysis showed no significant differences between the mean ST values in the thoracic limbs of normal dogs when compared to the mean ST values in the thoracic limbs of SCI-affected dogs at any time point after injury (Figure 4A). Significant differences were observed in ST values in the pelvic limbs between normal and SCI- affected dogs on day 3 (p < 0.0001 RPL, p < 0.0001 LPL), day 10 (p = 0.0078 RPL, p =

0.048 LPL) and day 30 (p = 0.0086 RPL, p = 0.0061 LPL) (Figure 4B). A trend analysis incorporating repeated measures indicated a statistically significant decrease in pelvic limb ST values over time in SCI-affected dogs (p = 0.032). 40

ST values correlate inversely with locomotor recovery in SCI-affected Dogs

Locomotor scores for SCI-affected dogs at days 3, 10, and 30 are summarized in

Table 2. At 3 days postoperatively, the median locomotor score was 6.5 (mean 6.8, range

1-11), which increased to a median score of 10 (mean 11.1, range 4-13) by day 10 and a median score of 11 (mean 11.1, range 7-14) by day 30. A significant inverse correlation was observed between locomotor score and pelvic limb ST value across the 30 day period

(Figure 5). For the LPL, this was observed at day 3 (r = -0.63, p = 0.0002) and day 10 (r

= -0.49, p = 0.006). For the RPL, this was also observed at day 3 (r = -0.50, p = 0.005), day 10 (r = -0.51, p = 0.004) and day 30 (r = -0.44, p = 0.01).

Thoracic limb ST values change with repeated measures in SCI-affected Dogs

ST values for the thoracic limbs of SCI-affected dogs significantly decreased between days 3 and 10 (p = 0.04 LTL, p = 0.01 RTL) and from day 10 to day 30 (p =

0.02 LTL only). There was not a correlation between thoracic limb ST values and locomotor score in SCI-affected dogs at any time point.

ST ratios in normal and SCI-affected dogs

Thoracic limb ST values in SCI-affected dogs represent an intra-animal control since the thoracic limbs do not experience sensorimotor impairment after thoracolumbar

IVDE. Changes in thoracic limb ST values would likely represent systemic and environmental factors such as patient stress or analgesic medications. A simple ratio of pelvic limb ST values to thoracic limb ST values was calculated to account for such confounding factors for each SCI-affected dog at every testing session, and this ratio was 41

compared to the same value in normal dogs (Table 3). This value was obtained by taking the average of both pelvic limb ST values and dividing this number by the average of both thoracic limb values for a given dog during each session (PL/TL).

The mean PL/TL ST ratio in normal dogs for sessions 1, 2, and 3 was 0.9 for all sessions, with ranges of 0.6-1.8, 0.6-1.6 and 0.6-1.5 respectively. There was no significant difference between ST ratios obtained in normal dogs from sessions 1 and 2, sessions 2 and 3, or sessions 1 and 3 (p = 0.55, 0.45, 0.88 respectively).

For SCI-affected dogs, the mean PL/TL ST ratios at day 3, day 10, and day 30 were 1.9 (range 1.0-4.6), 1.8 (range 0.0-7.7) and 1.4 (range 0.5-3.6) respectively. ST ratios were compared between normal and SCI-affected dogs (Figure 6). ST ratios obtained from SCI-affected dogs at day 3 were significantly higher than ST ratios from normal dogs at session 1 (p < 0.0001). Statistical analysis showed no significant difference in ST ratios between days 3 and 10 (p = 0.57) and between days 10 and 30 (p

= 0.11) in SCI-affected dogs. However, a significant decrease in pelvic limb to thoracic limb ST ratio was observed between days 3 and 30 (p = 0.03) in SCI-affected dogs. ST ratios were inversely correlated with locomotor score at day 3 (r = -0.61, p < 0.001), day

10 (r = -0.32, p = 0.089), and day 30 (r = -0.36, p = 0.048).

3.5 Discussion

Our study provides the first objective evaluation of ST in a large cohort of dogs with acute SCI. ST values obtained from the pelvic limbs of dogs with thoracolumbar 42

SCI were significantly higher than ST values obtained from the pelvic limbs of normal dogs in our study, while no differences were observed between ST values from the thoracic limbs of the same two groups.

Pelvic limb ST values significantly decreased in the 30-day postoperative period in dogs with acute thoracolumbar SCI. This change correlated inversely with locomotor scores, indicating that as motor function improves, sensory thresholds decrease in neurologically affected limbs. Given the significant difference in pelvic limb ST values observed between normal dogs and SCI-affected dogs at all three time points during neurologic recovery, coupled with the lack of statistically significant difference when comparing thoracic limb ST values between the same groups, it is likely that this change represents a true decline in ST. This finding may be explained by improvement of sensory function towards pre-injury status, or may represent trends towards development of central sensitization or mechanical allodynia (Kloos 2005, Walk 2009).

Interpretation of the observed changes in pelvic limb ST values in SCI-affected dogs during neurological recovery are complicated by a concurrent smaller but statistically significant decrease in ST values measured from the thoracic limbs of the same patients over the same time period. Because patients with thoracolumbar SCI should have neurologically normal thoracic limbs, improvement of sensory function to pre-injury status cannot explain this observation. It is possible that central sensitization or the development of allodynia could explain this finding (Carlton 2009, Densmore 2010).

It is equally possible that the changes in thoracic ST values represent a decline in 43

analgesic administration, acclimation of subjects to the testing environment, investigator training effect, or a combination of all of these factors. These factors also likely contribute in part to changes noted in pelvic limb ST values.

All of the SCI-affected dogs were administered analgesics including fentanyl, gabapentin, NSAIDs and tramadol. Such medications have been shown to influence the results of QST in various species (Lascelles 1998, Matthews 2002, Wegner 2008,

KuKanich 2011, Kögel 2014). Although the specific effect of some of these medications on ST values in dogs is unknown, the expected effect is an increase in ST values for the duration of administration. This effect could be expected to manifest in thoracic and pelvic limbs equally, and would also be expected to cease with discontinuation of medication administration.

A small but significant decline in ST values was also observed in normal dogs between sessions one and three in thoracic and pelvic limbs. This finding may be explained by ST decay, acclimation, or investigator training effect. Tactile sensory threshold decays (lowered sensory thresholds) can occur within a testing session if too many stimuli are given, or repeated stimuli are given too closely together (Detloff 2010).

We adhered to a one-minute delay between stimuli in order to minimize this concern

(KuKanich 2005ii, Detloff 2010, Detloff 2012, Moore 2013). Feeding during the testing session is suggested to minimize the effect of sensory threshold decay in rodent studies

(Detloff 2010, Detloff 2012), but proved too distracting in dogs during pilot studies

(Moore, unpublished data). With repeated testing sessions dogs may acclimate to the 44

testing environment, which can also decrease ST values (Detloff 2010). Testing sessions in all dogs were separated by no less than 48 hours, but a longer period may be needed to minimize this phenomenon.

Our data highlights several hurdles to the use of VFA in future clinical trials. For obvious reasons, it would be unethical to withhold or restrict analgesics for veterinary patients in a clinical trial. Prolonged acclimation of subjects to the testing environment is also not feasible for studies using client-owned animals with spontaneous SCI. For these reasons we considered the utility of a ratio of pelvic limb to thoracic limb ST values. For dogs with thoracolumbar injuries, this approach could provide a method to account for environmental and behavioral factors that may affect ST values for a patient on any given day by allowing the thoracic limbs to serve as an intra-animal control. Theoretically, analgesics, behavioral, and environmental factors should influence thoracic and pelvic ST values equally during a given testing session. The PL/TL ST ratio would remove those factors from the ST measurement, but could also minimize true differences in ST values.

The use of ST ratios provided more consistent values across testing sessions in normal dogs when compared to raw ST values. Significant differences in ST ratio were still appreciated between normal and SCI-affected dogs, and SCI-affected dogs showed a significant decrease in ST ratio as locomotor scores improved. However, during a pair- wise comparison of sessions, the use of ST ratios may have minimized important differences in measured ST values for SCI-affected dogs. Given our results, the utility of this measurement for assessment of ST in dogs with SCI is promising, but requires further investigation. 45

3.6 Conclusion

Our results support the utility of VFA to objectively measure changes in sensory function for dogs with acute SCI. A significant decline in pelvic limb ST values correlated inversely with locomotor recovery in SCI-affected dogs, supporting the utility of this technique as a quantitative assessment of neurologic recovery for future veterinary clinical trials in SCI. We also observed smaller but significant declines in ST values in normal dogs with repeated testing and in thoracic limbs of SCI-affected dogs over time.

These changes may be explained by analgesic medications in SCI-affected dogs, or by environmental and behavioral confounders in both groups. Future studies must focus on techniques to minimize or compensate for clinical, environmental, and behavioral factors that cannot be controlled for during clinical trials but may impact ST values in the clinical setting.

3.7 Conflict of Interest Statement

The authors report no conflict of interest.

3.8 Acknowledgements

This study was funded by the Morris Animal Foundation D13CA-024. The authors also gratefully acknowledge Mrs. Amanda Disher and Ms. Heather Myers for their assistance with data collection and Mr. Tim Vojt for his assistance with figure preparation.

3.9 References

Briley JD, Williams MD, Freire M, Griffith EH, Lascelles BD. Feasability and repeatability of cold and mechanical quantitative sensory testing in normal dog. Vet J 2014;199:245-250.

Brydges NM, Argyle DJ, Mosley JR, Duncan JC, Fleetwood-Walker S, Clements DN. Clinical assessments of increased sensory sensitivity in dogs with cranial cruciate ligament rupture. Vet J 2012;193:545-550.

Carlton SM, Junhui D, Tan HY, Nesic O, Hargett GL, Bopp AC, Yamani A, Lin Q, Willis WD, Hulsebosch CE. Peripheral and central sensitization in remote spinal cord regions contribute to central neuropathic pain after spinal cord injury. Pain 2009;147:265- 276.

Densmore VS, Kalous A, Keast JR, Osborne PB. Above-level mechanical hyperalgesia in rats develops after incomplete spinal cord injury but not after cord transection, and is reversed by amitriptyline, morphine and gabapentin. Pain 2010;151:184-193.

Detloff MR, Clark LM, Hutchinson KJ, Kllos AD, Fisher LC, Basso DM. Validity of acute and chronic tactile sensory testing after spinal cord injury in rats. Exp Neurol 2010;225:366-376.

Detloff MR, Fisher LC, Deibert RJ, Basso DM. Acute and chronic tactile sensory testing after spinal cord injury in rats. J Vis Exp 2012;62:e3247.

Felix ER, Widerstrom-Noga EG. Reliability and validity of quantitative sensory testing in persons with spinal cord injury and neuropathic pain. J Rehabil Res Dev 2009;46:69- 84.

Hayes KC, Wolfe DL, Hsieh JT, Potter PJ, Drassioukov A, Durham CE. Clinical and electrophysiologic correlates of quantitative sensory testing in patients with incomplete spinal cord injury. Arch Phys Med Rehabil 2002;83:1612-1619.

Hoschouer EL, Basso DM, Jakeman LB. Aberrant sensory responses are dependent on lesion severity after spinal cord contusion injury in mice. Pain 2010;148:328-342.

Kloos AD, Fisher LC, Detloff MR, Hassenzahi DL, Basso DM. Stepwise motor and all- or-none sensory recovery is associated with nonlinear sparing after incremental spinal cord injury in rats. Exp Neurol 2005;191:251-265.

Kögel B, Terlinden R, Schneider J. Characterisation of tramadol, morphine and tapentadol in an acute pain model in Beagle dogs. Vet Anaesth Analg 2014;41:297-304.

KuKanich B, Lascelles BD, Papich MG. Assessment of a von Frey device for evaluation of the antinociceptive effects of morphine and its application in pharmacodynamics modeling of morphine in dogs. Am J Vet Res 2005i;66:1616-1622.

KuKanich B, Lascelles BD, Papich MG. Use of a von Frey device for evaluation of pharmacokinetics and pharmacodynamics of morphine after intravenous administration as an infusion or multiple doses in dogs Am J Vet Res 2005ii;66:1968-1974.

KuKanich B, Papich MG. Pharmacokinetics and antinociceptive effects of oral tramadol hydrochloride administration in Greyhounds. Am J Vet Res 2011;72:256-262.

Lascelles BD, Cripps PJ, Jones A, Waterman-Pearson AE. Efficacy and kinetics of carprofen, administered preoperatively or postoperatively, for the prevention of pain in dogs undergoing ovariohysterectomy. Vet Surg 1998;27:568-82.

Lascelles BD. Getting a sense of sensations. Vet J 2013;197:115-117.

Levine GJ, Levine JM, Budke CM, Kerwin SC, Au J, Vinayak A, Hettlich BF, Slater MR. Description and repeatability of a newly developed spinal cord injury scale for dog. Prev Vet Med 2009;89:121-127.

Lindsey AE, LoVerso RL, Tovar A, Hill CE, Beattie MS, Bresnahan JC. An analysis of changes in sensory threholds to mild tactile and cold stimuli after experimental spinal cord injury in the rat. Neurorehabil Neural Repair 2000;14:287-300.

Matthews EA, Dickenson AH. A combination of gabapentin and morphine mediates enhanced inhibitory effects on dorsal horn neuronal responses in a rat model of neuropathy. Anesthesiology 2002;96:633-640.

Moore SA, Hettlich BF, Waln A. The use of an electronic von Frey device for evaluation of ST in neurologically normal dogs and those with acute spinal cord injury. Vet J 2013;197:216-219.

Olby NJ, De Risio L, Muñana K, Wosar MA, Skeen TM, Sharp NJ, Keene BW. Development of a functional scoring system in dogs with acute spinal cord injuries. Am J Vet Res 2001;62:1624-1628.

Olby N, Levine J, Harris T, Muñana K, Skeen T, Sharp N. Long-term functional outcome of dogs with severe injuries of the thoracolumbar spinal cord: 87 cases (1996-2001). J Am Vet Med Assoc 2003;222:762-769.

Verbeke G, Molenberghs G. Linear Mixed Models for Longitudinal Data, First Ed. Springer, Verbag, NY, 2000.

Walk D, Sehgal N, Moeller-Bertram T, Edwards RR, Wasan A, Wallace M, Irving G, Argoff C, Backonja MM. Quantitative sensory testing and mapping – A review of nonautomated quantitative methods for examination of the patient with neuropathic pain. Clin J Pain 2009;25:632-640.

Wegner K, Horais KA, Tozier NA, Rathbun ML, Shtaerman Y, Yaksh TL. Development of a canine nociceptive thermal escape model. J Neurosci Methods 2008;168:88-97.

Tables

Von Frey (g) LTL von Frey (g) LPL von Frey (g) RTL von Frey (g) RPL

S1 S2 S3 S1 S2 S3 S1 S2 S3 S1 S2 S3 Normal Dogs Ave Ave Ave Ave Ave Ave Ave Ave Ave Ave Ave Ave

Beagle 177.3 134.4 86.3 205.7 184.9 111.8 100.8 157.1 90.9 264.6 206.2 147.4 Bichon Frise 111.4 95.4 60.4 161.6 59.7 61.4 80.6 71.6 84.8 92.4 58.5 41.9 Cocker 134.1 169.0 114.4 117.6 137.4 133.4 154.2 138.6 112.0 139.9 146.0 132.4 Spaniel Corgi 188.9 131.7 94.5 178.5 68.8 77.8 227.4 145.9 99.4 158.2 129.8 94.6 Dachshund 174.8 120.7 136.4 159.7 117.2 139.5 162.6 161.5 138.3 108.6 150.2 117.4 Dachshund 133.0 85.6 66.0 143.4 90.1 89.0 108.1 132.6 39.1 74.8 101.1 65.2 Dachshund 95.0 81.4 57.7 77.8 90.8 56.9 229.6 83.6 107.9 202.0 128.6 72.2 Dachshund 170.7 128.7 160.1 181.2 111.5 82.2 169.9 165.6 169.2 75.9 126.5 155.5 Miniature 143.8 141.2 234.4 126.2 169.0 166.8 132.2 161.1 164.5 156.8 180.7 203.0 pinscher Mixed Breed 269.4 133.3 118.4 204.3 102.7 65.1 123.3 120.2 112.2 197.3 127.1 80.7 Mixed breed 186.5 68.8 55.8 158.1 100.2 55.6 180.6 69.5 104.2 178.1 123.8 89.5 Mixed breed 244.8 225.9 138.8 192.5 233.4 178.2 151.1 197.3 201.8 237.6 218.3 159.3 Mixed breed 126.5 123.3 93.7 135.1 149.1 113.3 144.8 144.3 93.9 185.9 128.2 89.0 Mixed breed 311.6 274.4 121.0 229.7 225.0 101.1 204.5 168.9 161.0 147.0 96.3 87.5 Mixed breed 135.1 106.9 96.7 109.1 71.3 79.2 156.8 116.0 142.3 106.4 90.5 115.1 Schnauzer 93.6 75.6 70.1 92.3 41.3 80.7 86.0 101.4 122.7 61.5 59.7 72.8 Schnauzer 105.5 62.3 66.8 106.9 80.9 70.4 92.5 90.6 84.8 109.7 66.9 43.0 Sealyham 198.4 146.6 117.5 113.6 113.7 95.4 236.6 131.7 123.6 154.2 92.3 76.8 terrier Sealyham 209.1 165.1 123.3 144.7 119.5 76.0 147.2 144.0 133.9 131.9 150.3 102.7 terrier Shih Tzu 27.7 97.5 42.1 75.3 62.7 61.2 52.9 39.7 46.6 73.3 46.9 66.9 MEAN 161.9 128.4 102.7 145.7 116.5 94.8 147.1 127.1 116.7 142.8 121.4 100.6 STD ERROR OF THE 14.8 11.7 10.0 9.8 11.9 8.0 11.6 9.0 9.0 12.6 10.5 9.3 MEAN

Table 2 von Frey anesthesiometry sensory threshold values of thoracic and pelvic limbs in dogs with acute spinal cord inury and T3- L3 myelopathy caused by intervertebral disc extrusion (n=30). ST values are inversely correlated with Olby Spinal Cord Injury Scale (OSCIS) locomotor scores at days, 10, and 30 after decompressive surgery.

Von Frey (g) LTL von Frey (g) LPL von Frey (g) RTL von Frey (g) RPL OSCIS

Day 3 Day 10 Day 30 Day 3 Day 10 Day 30 Day 3 Day 10 Day 30 Day 3 Day 10 Day 30 Day Day Day SCI-affected dogs Ave Ave Ave Ave Ave Ave Ave Ave Ave Ave Ave Ave 3 10 30

Beagle 160.8 97.9 105.7 246.9 199.8 286.4 104.4 148.5 168.0 229.0 167.5 243.5 4 10 11 Beagle 258.7 137.3 178.5 487.4 380.7 235.6 170.5 155.4 170.5 383.6 251.1 240.1 7 10 13 Beagle 180.1 199.8 110.9 284.8 258.5 165.2 218.9 224.7 162.3 269.0 300.3 138.9 11 13 11 Cocker Spaniel 176.8 249.1 172.3 327.3 266.4 231.2 219.9 256.5 172.3 248.5 319.3 281.3 9 10 13 Corgi 251.5 152.5 107.1 407.4 256.9 276.4 279.5 157.1 260.6 381.6 282.2 325.8 4 5 7 Corgi 155.2 154.6 126.4 343.1 269.6 293.7 219.2 100.9 195.7 589.8 218.5 231.6 6 9 9

51 Dachshund 98.1 82.4 77.6 168.3 131.0 85.6 164.6 123.2 81.7 174.0 116.8 114.6 11 11 14

Dachshund 74.7 77.9 90.5 239.0 71.7 44.0 128.3 98.8 105.5 302.9 142.8 109.2 9 11 11 Dachshund 293.9 194.0 232.4 573.8 405.3 187.8 190.8 168.9 197.8 713.6 99.7 288.5 1 8 11 Dachshund 88.3 50.3 129.0 489.7 210.2 141.8 350.3 148.1 130.4 120.1 89.8 115.8 10 11 14 Dachshund 139.2 57.4 103.9 550.6 356.0 126.9 84.2 34.4 110.9 481.5 352.4 149.9 1 7 13 Dachshund 150.3 174.5 200.3 295.8 280.6 224.1 160.7 180.8 197.2 237.2 220.1 154.9 7 9 11 Dachshund 178.3 132.7 174.5 503.4 444.2 326.7 211.4 145.2 92.2 394.1 358.4 275.6 4 9 11 Dachshund 94.7 173.6 228.1 440.5 237.4 305.0 84.0 167.9 253.2 222.1 141.2 248.0 4 6 13 Dachshund 271.3 124.4 97.2 511.7 275.3 132.4 209.0 117.0 80.1 636.1 333.6 171.3 3 8 10 Dachshund 109.3 83.7 94.6 168.5 109.6 167.3 160.6 109.9 98.1 251.5 135.4 103.8 11 11 11 Dachshund 184.4 118.9 120.8 319.1 213.3 173.8 142.2 98.7 120.6 340.0 210.7 245.7 4 11 11 Dachshund 166.8 113.3 237.0 194.8 166.0 226.1 236.4 85.8 184.0 227.4 146.2 228.2 11 14 11 Dachshund 114.7 113.7 115.0 187.6 217.1 158.4 153.7 165.6 142.6 220.1 189.6 176.9 10 11 11 51 Continued

Table 2 Continued

French Bulldog 171.4 92.8 125.6 253.3 134.2 120.4 238.0 69.6 128.6 188.8 95.9 103.0 9 13 13 French Bulldog 212.0 350.8 202.8 318.7 277.9 284.8 285.0 331.7 227.4 313.6 368.6 308.2 11 13 13 French Bulldog 368.1 268.9 177.8 640.7 482.5 339.8 645.7 248.6 197.9 783.8 580.7 323.9 3 4 9 French Bulldog 407.4 164.4 173.0 291.3 139.4 145.2 396.7 175.2 248.1 615.2 187.3 158.3 10 11 11 Mixed breed 235.9 146.4 145.8 334.5 226.7 289.0 163.5 128.3 238.9 361.2 186.4 220.4 11 13 13 Mixed breed 147.6 122.3 54.2 161.0 99.5 41.8 200.1 63.5 80.9 173.2 60.4 84.4 11 10 11 Mixed breed 82.1 116.4 127.3 171.4 175.8 119.0 106.1 103.2 109.0 145.9 149.3 192.1 6 9 10 Mixed breed 252.1 235.4 204.1 299.1 344.8 203.6 187.8 226.7 220.3 348.0 170.6 182.0 6 11 11 Mixed breed 263.5 219.6 124.3 701.2 621.8 374.2 164.8 177.7 87.7 555.7 568.9 390.3 4 5 8 Shih Tzu 222.7 221.6 281.6 300.4 318.9 379.1 104.2 173.0 265.6 289.0 247.1 424.2 1 6 8 Shih Tzu 349.2 157.7 109.9 416.2 25.0 19.7 77.0 167.6 140.4 273.2 231.1 106.6 4 7 9

52 MEAN 195.3 152.8 147.6 354.3 253.2 203.5 201.9 151.8 162.3 349.0 230.7 211.2

STD ERROR OF 15.8 12.3 10.0 27.0 23.6 17.9 20.5 11.5 10.9 32.0 23.1 16.3 THE MEAN

ST ratio ST ratio ST ratio Normal Dogs Session 1 Session 2 Session 3 Beagle 1.7 1.3 1.5 Bichon Frise 1.3 0.7 0.7 Cocker Spaniel 0.9 0.9 1.2 Corgi 0.8 0.7 0.9 Dachshund 0.8 0.9 0.9 Dachshund 0.9 0.9 1.5 Dachshund 0.9 1.3 0.8 Dachshund 0.8 0.8 0.7 Miniature pinscher 1.0 1.2 0.9 Mixed Breed 1.0 0.9 0.6 Mixed breed (Puggle/Corgi) 0.9 1.6 0.9 Mixed breed (Dachshund) 1.1 1.1 1.0 Mixed breed (Beagle/hound) 1.2 1.0 1.1 Mixed (Dachshund/Scottie) 0.7 0.7 0.7 Mixed breed (Pomeranian) 0.7 0.7 0.8 Schnauzer 0.9 0.6 0.8 Schnauzer 1.1 1.0 0.7 Sealyham terrier 0.6 0.7 0.7 Sealyham terrier 0.8 0.9 0.7 Shih Tzu 1.8 0.8 1.4 MEAN 0.9 0.9 0.9 ST ratio ST ratio ST ratio SCI-Affected dogs Day 3 Day 10 Day 30 Beagle 1.8 1.5 1.9 Beagle 2.0 2.2 1.4 Beagle 1.4 1.3 1.1 Cocker Spaniel 1.5 1.2 1.5 Corgi 1.5 1.7 1.6 Corgi 2.5 1.9 1.6 Dachshund 1.3 1.2 1.3 Dachshund 2.7 1.2 0.8 Dachshund 2.7 1.4 1.1 Continued

Table 3 Continued

Dachshund 1.4 1.5 1.0 Dachshund 4.6 7.7 1.3 Dachshund 1.7 1.4 1.0 Dachshund 2.3 2.9 2.3 Dachshund 3.7 1.1 1.1 Dachshund 2.4 2.5 1.7 Dachshund 1.6 1.3 1.4 Dachshund 2.0 1.9 1.7 Dachshund 1.0 1.6 1.1 Dachshund 1.5 1.5 1.3 French Bulldog 1.1 1.4 0.9 French Bulldog 1.3 0.9 1.4 French Bulldog 1.4 2.1 1.8 French Bulldog 1.1 1.0 0.7 Mixed breed (pitbull/basset) 1.7 1.5 1.3 Mixed breed (yorkie) 1.0 0.9 0.9 Mixed breed (Dachshund) 1.7 1.5 1.3 Mixed breed (Cocker spaniel/poodle) 1.5 1.1 0.9 Mixed breed (Dachshund/yorkie) 2.9 3.0 3.6 Shih Tzu 1.8 1.4 1.5 Shih Tzu 1.6 0.8 0.5 MEAN 1.8 1.6 1.3

Figures

A 700 B 700 Left Left Right Right * 600 600 *

500 500 *

e 400 e 400

a a *

* *

S 300 S 300

200 200

100 100

0 0 Normal Day 3 Day 10 Day 30 Normal Day 3 Day 10 Day 30

Figure 4 Comparison of left and right thoracic limb (A) and pelvic limb (B) sensory threshold (ST) values between normal dogs at session one and spinal cord injury (SCI)- affected dogs at days 3, 10 and 30 following decompressive surgery. ST values were significant higher in the pelvic limbs of SCI-affected dogs at all three time points evaluated, while no differences were noted in thoracic limb ST values between groups. Line represents the median ST value; small box represents the mean ST value. Whiskers represent the 95% confidence interval. Asterisk indicates statistically significant differences from normal dogs (p<0.05).

12 500 450 10 400

350 e

8 u

I 300 a

O 6 250

200 4 150 100 2 50

0 0 Day 3 Day 10 Day 30

3.0 *

2.5

2.0

r 1.5

1.0

0.5

0.0 Normal Day 3 Day 10 Day 30

Figure 6 Pelvic limb/thoracic limb sensory threshold (ST) ratios in normal dogs (session

1) and spinal cord injury dogs at 3, 10, and 30 days following decompressive surgery. ST ratios were significantly higher in SCI-affected dogs compared to normal dogs at day 3,

but not at days 10 or 30. Line represents the median ST value; small box represents the mean ST value. Whiskers represent the 95% confidence interval. Asterisk indicates statistically significant differences from normal dogs (p<0.05).

Chapter 4

Utility and Repeatability of Footprint Analysis to Assess Gait Recovery After Thoracolumbar Intervertebral Disc Extrusion in Dogs

4.1 Abstract

Measurements of stride length (SL), base-of-support (BS) and coefficients of variance (COV) were made using a ‘finger painting’ technique for footprint analysis in all limbs of 20 normal dogs and 29 dogs with 30 discrete episodes of acute thoracolumbar spinal cord injury (SCI) caused by spontaneous intervertebral disc extrusion.

Measurements were determined at three separate time points in normal dogs and on day

3, 10 and 30 following decompressive surgery in dogs with SCI. Values for SL and BS were compared between groups at each time point.

Mean SL in all limbs was significantly (p<0.05) lower in SCI-affected dogs at day

3, 10 and 30 compared to normal dogs, and gradually increased with time in SCI-affected dogs. The COV of SL was significantly higher in both thoracic limbs (TL) and one pelvic limb (PL) in SCI-affected dogs only at day 3 compared to normal dogs. Significant changes in mean SL of one PL of SCI-affected dogs was detected between day 3 and 30.

Additionally, BS-TL was found to be significantly higher in SCI-affected dogs at day 3 and day 30 following surgery compared to normal dogs. BS-PL was not significantly different between SCI-affected dogs and normal dogs. The findings of this study support the use of footprint parameters to compare locomotor differences between normal and

SCI-affected dogs, and to assess recovery from SCI. Additionally, the results of the study underscore the importance of evaluating the TL in thoracolumbar SCI-affected dogs.

4.2 Introduction

Intervertebral disc extrusion (IVDE) is the most common cause of acute spinal cord injury (SCI) in dogs, and chondrodystrophic breeds such as the dachshund, cocker spaniel, basset hound, beagle, Pekingese, shih tzu, miniature poodle and bichon frise are commonly represented in the literature (Olby 2003, Ito 2005, Levine 2011, Aikawa 2012,

Bergknut 2012, Packer 2013). The high incidence of spontaneous SCI in dogs makes them an important large animal model for human SCI (Rice 2009). Dogs offer a genetically similar, but environmentally heterogenous study population, with comparable mechanisms of injury and resultant pathology to that in humans, which can bridge the gap between experimental rodent models and the human SCI population (Borgens 1999,

Laverty 2004, Olby 2004, Jeffery 2006). Successful clinical trials in dogs with spontaneous SCI may lead to the development of interventional therapies that can help both dogs and humans.

Footprint analysis has been used previously to assess return of pelvic limb function following animal models of nerve injury (de Medinaceli 1982, Varejao 2004) and SCI (Kunkel-Bagden 1993, Cheng 1997, Klapdor 1997, Hamers 2001, Hamers 2006,

Gordon-Evans 2009, Rangasamy 2013). Measurements such as base-of-support, stride length, inter-limb coordination, regularity of step patterns and paw position can provide 59

valuable information regarding the animal’s pattern of locomotion which may reflect the injury type, severity of injury, and specific spinal tract involved in the lesion (Kunkel-

Bagden 1993, Klapdor 1997, Hamers 2001, Hamers 2006, Rangasamy 2013). Dogs with spinal cord disease exhibit an uncoordinated gait that is quantifiably different from dogs with lameness due to orthopedic disease (Gordon-Evans 2009). Coefficients of variance for stride length, swing time and lateral paw positioning have previously been shown to be significantly larger in dogs with neurologic disease when compared to normal dogs

(Hamilton 2008, Gordon-Evans 2009). Additionally, footprint analysis may reveal gait deficits that can be objectively measured, and not readily detected through visual assessment only (McEwen 2006).

Recently, footprint analysis is more often performed using specialized equipment such as the Tekscan® or Catwalk (Hamers 2001, Hamers 2006, Gordon-Evans 2009).

However, such specialized equipment is costly and not readily available. Lack of availability makes use of footprint analysis impractical for some researchers and precludes its use in multi-center SCI veterinary clinical trials.

The primary purpose of this study was to use a simple ‘finger painting’ method as a means of comparing footprint analysis between a large cohort of normal dogs and dogs with acute thoracolumbar SCI caused by IVDE. We aimed to document the change in easily measurable footprint parameters in SCI-affected dogs over a 30-day recovery period, specifically stride length (SL) and base of support (BS). We reasoned that footprint analysis using this method would produce many of the same measurable 60

parameters obtained via commercially available equipment. Additionally, we hypothesize that variation in SL and BS is different between normal dogs and SCI-affected dogs, and footprint parameters will change as dogs with SCI progress toward neurologic recovery.

4.3 Materials and Methods

The study was conducted in accordance with the guidelines and approval of The

Ohio State University Clinical Research Advisory Committee and the Institutional

Animal Care and Use Committee (2012A00000149). Written owner consent was obtained prior to study enrollment for all dogs.

Normal and SCI-affected dogs

Normal control dogs that were skeletally mature (≥ 8 months of age) and of a small-breed (≤ 17 kg) were recruited from the pet population of The Ohio State

University Veterinary Medical Center. All dogs underwent a full neurologic and orthopedic examination by two of the investigators (RBS and SAM). Dogs determined to be neurologically and orthopedically normal without any history of neurologic or orthopedic disease were enrolled in the control population of the study. Valgus and varus conformational limb variations typical for chondrodystrophic breeds were considered acceptable for enrollment to facilitate generalization of results across a realistic clinical population.

SCI-affected dogs from the general patient population at The Ohio State

University Veterinary Medical Center were screened for the following inclusion criteria: 61

(1) diagnosis of a T3-L3 myelopathy caused by acute IVDE determined by computed tomography with or without myelography, or magnetic resonance imaging (2) intact nociception of both pelvic limbs and tail (minimum Olby Spinal Cord Injury Scale

(OSCIS) = 1) (3) small-breed dog ≤ 17 kg (4) behaviorally amenable. While not a specific criterion for enrollment, all dogs enrolled underwent surgical decompression for their IVDE. Dogs that fit the inclusion criteria were consecutively and prospectively enrolled.

Footprint acquisition

Different colored non-toxic, washable paint was applied to each paw. The colors applied to each paw were kept consistent between dogs and recorded (Figure 7). Dogs were then walked with a leash at a natural, consistent pace by the same investigator

(RBS) down 3 meters of butcher paper. Five walking trials were collected during each testing session. Dogs that were reluctant to walk were encouraged and enticed with treats and verbal cues by a second investigator at the opposite end of the butcher paper.

Notations on the butcher paper were made by the investigator during the testing to indicate if dogs stopped or deviated from the butcher paper path.

Control dogs were tested on three days, separated by at least 48 hours. SCI- affected dogs were tested and assigned an OSCIS score at days 3, 10 and 30 following decompressive surgery. Only dogs that were ambulatory without external support were used at each time point for data acquisition (OSCIS ≥ 9).

Footprint analysis

The first and last steps for each paw within each trial were excluded from analysis to account for the animal’s adjustment to a unique walking surface. A single investigator

(MO) performed all measurements. A reference point for each paw print was located at the intersection of the intermetacarpophalangeal space and the P3-P4 interdigital space

(IDS). This space was chosen to be consistent with previous research using footprint analysis (Kunkel-Bagden 1993). In the case that the print was incomplete, the investigator measured this location on a complete print from the same dog from the same testing session and determined the theoretical point of interest by extrapolating from the front of the third or fourth digit. To correct for rotational variation of the paw, lines were drawn perpendicular to the edge of the walking track through each print at the IDS as previously described (Kunkel-Bagden 1993). The distance between these lines was measured for each step for each paw, and was designated as stride length (SL). The distance between the IDS on the right and left thoracic limbs was measured for each step cycle and designated as thoracic limb base-of-support (BS-TL). Similarly, the distance between the IDS on the right and left pelvic limbs was measured for each step cycle and designated as the pelvic limb base-of-support (BS-PL) (Figure 7).

Prints were excluded from analysis in the following cases: (1) the print was incomplete such that the IDS could not be identified, either by direct identification or measured extrapolation (2) the dog was noted to have stopped walking in the middle of a trial (3) multiple prints for the same paw appeared within 2 cm of one another. When a print was excluded, the strides including the excluded print were excluded from analysis. 63

When a dog was noted to have stopped mid-trial, the stride following the stop was excluded from analysis.

Mean SL and coefficient of variance (COV) of SL was calculated for each limb at each testing session for control and SCI-affected dogs. Mean BS and COV of BS for the thoracic limbs (TL) and pelvic limbs (PL) were calculated at each testing session for control and SCI-affected dogs.

Statistics

Summary statistics including mean and COV for clinical data on all dogs are reported. A mixed effect model incorporating repeated measures, was used to compare the changes in SL and BS for each limb between sessions. With this model, a p-value of

< 0.05 was considered significant for all analyses. Analyses were conducted using SAS software (SAS, Inc; Cary, NC).

4.4 Results

Normal dogs

Twenty normal dogs were recruited for the study. Ages ranged from 8 months to

6.5 years (median 3 years) and weight ranged from 3.7 kg to 17.2 kg (median 9.4 kg).

There were 8 spayed females and 12 castrated males. Breeds represented included the following: mixed breed dogs (6), dachshunds (4), miniature schnauzers (2), Sealyham terriers (2), beagle (1), bichon frise (1), cocker spaniel (1), Pembroke Welsh corgi (1),

miniature pinscher (1), and shih tzu (1). Time period between each testing session for each dog ranged from 2 to 27 days (median 6 days).

Affected dogs

A total of 29 dogs with 30 discrete episodes of acute SCI caused by IVDE were enrolled. The age of the dogs ranged from 2 to 11 years (median 5 years) and weight ranged from 3.9 kg to 17.0 kg (median 8.2 kg). There were 14 spayed females, 13 castrated males, and 2 intact males. Breeds represented included the following: dachshunds (12), mixed breed dog (6), French bulldog (4), beagle (2), Pembroke Welsh corgi (2), shih tzu (2), and cocker spaniel (1).

All dogs underwent decompressive hemilaminectomy or pediculectomy at one or multiple sites between T10-11 and L3-4 intervertebral disc spaces, with or without one or more lateral disc fenestrations dependent on imaging results and discretion of the surgeon. Post-operative analgesic treatment was dependent upon the surgeons’ preference but typically included a fentanyl constant rate infusion (CRI) for 12-24 hours immediately postoperatively, a fentanyl patch placed immediately post-operatively, and combinations of tramadol, gabapentin, methocarbamol, or diazepam at various doses.

Postoperative anti-inflammatory therapy generally included tapering anti-inflammatory doses of prednisone, or a non-steroidal anti-inflammatory drug (Meloxicam, Carprofen,

Firocoxib, or Deracoxib). The medication doses, frequency of administration, and number of total medications were recorded for each dog at each session. An OSCIS score was assigned on days 3, 10 and 30 post-operatively to determine eligibility for 65

footprint acquisition (OSCIS ≥ 9). Based upon examination, 12 dogs were available on day 3, 21 dogs were available on day 10, and 28 dogs were available on day 30 for footprint acquisition. Two dogs did not gain consistent ambulation without support in the pelvic limbs during the study time period.

Mean and COV of SL of normal dogs

The mean SL (cm) of the left TL (LTL) across the three testing sessions was

43.69, 45.30 and 45.82 with a COV of 0.13, 0.10 and 0.11 respectively (Table 4). The mean SL of the right TL (RTL) across the three testing sessions was 43.68, 45.22 and

45.74 with a COV of 0.13, 0.11 and 0.12. The mean SL of the left PL (LPL) across the three testing sessions was 43.43, 44.09, 45.40 with a COV of 0.13, 0.12 and 0.11.

Similarly, the SL of the right PL (RPL) across the three testing sessions was 43.54, 44.16 and 45.19 with a COV of 0.13, 0.12 and 0.12. The mean COV of any limb at any testing session in normal dogs was ≤ 0.13.

Mean and COV of SL of SCI-affected dogs

The mean SL of the LTL across the three testing sessions was 30.37, 33.91 and

34.78 with a COV of 0.22, 0.16 and 0.15 respectively (Table 5). The mean SL of the RTL across testing sessions was 30.69, 34.47 and 35.09 with a COV of 0.20, 0.15 and 0.15.

The mean SL of the LPL across testing sessions was 32.17, 36.05, 37.52 with COV of

0.18, 0.17 and 0.15. The mean SL of the RPL across testing sessions was 31.46, 36.25 and 37.32 with a COV of 0.18, 0.17 and 0.16.

Mean and COV of BS of normal dogs

The mean BS-TL (cm) across the three testing sessions of normal dogs was 6.28,

6.05 and 6.18 with a COV of 0.48, 0.46 and 0.48 respectively (Table 6). The mean BS-

PL across testing sessions was 8.18, 7.85 and 7.89 with a COV of 0.41, 0.37 and 0.39.

Mean and COV BS of affected dogs

The mean BS-TL across the three testing sessions was 8.41, 7.61 and 7.88 with a

COV of 0.37, 0.41 and 0.37 respectively (Table 7). The mean BS-PL across testing sessions was 7.90, 7.75 and 8.04 with a COV of 0.49, 0.46 and 0.41.

Mean and COV of SL and BS does not differ between sessions in normal dogs

Changes in mean SL values of all limbs across testing sessions 1 and 2, and session 1 and 3 were compared to detect significant differences. No significant differences were detected across any testing session comparison in any limb of all normal dogs (LTL p=0.57, 0.45; RTL p=0.58, 0.46; LPL p=0.81, 0.48; RPL p=0.82, 0.55).

Additionally, comparisons between testing sessions 1 and 2, 1 and 3, and 2 and 3 were performed for the COV of SL in all limbs in normal dogs, which revealed no significant differences (LTL p=0.22, 0.45, 0.63; RTL p=0.35, 0.46, 0.83; LPL p=0.69, 0.56, 0.85;

RPL p=0.58, 0.55, 0.96). Similarly, changes in mean BS-TL and BS-PL across testing sessions 1 and 2, and 1 and 3 did not reveal significant differences (BS-TL p=0.78, 0.90;

BS-PL p=0.69, 0.72). Comparisons between testing sessions 1 and 2, 1 and 3 and 2 and 3 performed for the COV of BS-TL and BS-PL revealed no significant differences (BS-TL p=0.73, 0.94, 0.78; BS-PL p=0.61,0.73, 0.87). 67

Mean SL of all limbs differs significantly between normal dogs and SCI-affected dogs

The mean SL for each limb from the first testing session of normal dogs was compared to the SL for the corresponding limb in SCI-affected dogs at day 3, 10 and 30 postoperatively (Table 8). The SLs were significantly (p<0.05) lower in all limbs at all testing sessions in SCI-affected dogs when compared to normal dogs (Figure 8). The difference in the SLs between normal and SCI-affected dogs decreased at each testing session of SCI-affected dogs in all limbs. The difference decreased from 13.32, 9.78, 8.91

(p<0.0001, p=0.0004, p=0.0006 respectively) in the LTL, from 11.26, 7.37, 5.91

(p=0.0004, p=0.0079, p=0.023) in the LPL, from 12.99, 9.21 8.59 (p<0.0001, p=0.0009, p=0.001) in the RTL, from 12.08, 7.29, 6.22 (p=0.0001, p=0.0086, p=0.017) in the RPL.

Mean SL of the RPL changes significantly across testing sessions in SCI-affected dogs

Changes in mean SL values of all limbs across testing sessions at day 3 and 10, and days 3 and 30 post-operatively were compared to detect significant differences in

SCI-affected dogs. Comparisons between days 3 and 30 revealed a significant difference in the RPL while the significance was approached in the LPL (LTL p=0.26, 0.24; RTL p=0.23, 0.14; LPL p=0.21, 0.072; RPL p=0.13, 0.049).

COV of SL in PL and TL differs significantly between normal dogs and SCI-affected dogs

Comparisons of the COV SL of each limb were made between the first testing session of normal dogs to all three testing sessions of SCI-affected dogs. COV SL of the 68

LTL (p=0.0024) and RTL (p=0.019) and LPL (p=0.034) at day 3 of SCI-affected dogs were significantly higher when compared normal dogs. COV SL of the RPL between normal dogs and day 3 of SCI-affected dogs (p=0.085) and COV SL of the LPL between normal dogs and day 10 of SCI-affected dogs (p=0.073) approached significance.

BS-TL differs significantly between normal dogs and SCI-affected dogs

The mean BS-TL and BS-PL were compared between the first testing session of normal dogs to SCI-affected dogs at day 3, 10 and 30 following decompressive surgery

(Figure 9). The BS-TL was higher in SCI-affected dogs compared to normal dogs at all three testing sessions; day 3 (p=0.023), day 10 (list p=0.11), and day 30 (p=0.038) (Table

9).

COV of BS does not differ significantly between normal dogs and SCI-affected dogs

Comparisons of the COV of BS-TL and BS-PL were made between the first testing session of normal dogs to all three testing sessions of SCI-affected dogs. No significant differences were observed in COV of BS between normal and SCI-affected dogs at any time point. However, COV of BS-TL between normal dogs and day 3

(p=0.076) and day 30 (p=0.053) of SCI-affected dogs approached significance.

There is no significant correlation between mean and COV SL or mean and COV BS-PL with locomotor scores in SCI-affected dogs

Correlations were made between mean and COV SL and mean and COV BS-PL between the OSCIS at all three testing sessions in SCI-affected dogs. No significant correlations were found.

4.5 Discussion

To our knowledge, this is the first study that has evaluated footprint analysis parameters of overground locomotion using a ‘finger paint’ technique in a large cohort of normal dogs and dogs with spontaneous acute SCI. This study demonstrated that dogs with thoracolumbar SCI have significantly lower SL in all limbs at least thirty days following injury when compared to normal dogs. Mean SL increased significantly between day 3 and 30 of the RPL in SCI-affected dogs. COV of SL was significantly higher in both TLs and one PL in SCI-affected dogs compared to normal dogs in the first three days following injury. Additionally, BS-TL was demonstrated to be significantly higher in SCI-affected dogs when compared to normal dogs.

The decrease in PL SL may be due to loss of supraspinal excitatory input to the motor neurons innervating the extensor muscles, causing paresis and associated decreased ability of the limb to support body weight, reduced propulsion of the affected

PLs, and decreased stance or swing duration of the PLs (Hamers 2001, McEwen 2006,

Collazos-Castro 2006, Rangasamy 2013). A few studies have additionally demonstrated a decrease in SL of the TLs following thoracolumbar SCI (Hamers 2001, McEwen 2006,

Plemel 2008, Gordon-Evans 2009). The decrease in TL SL in thoracolumbar SCI- affected dogs demonstrated in this study may reflect compensatory changes in response 70

to the affected PLs. Dogs with thoracolumbar SCI likely have increased contribution by the TLs to their overall gait, and likely use the TLs at a higher speed, and thus a decreased SL and an increased vertical force to compensate for the decreased PL function

(Cheng 1997, Hamers 2001, McEwen 2006, Gordon-Evans 2009). Additionally, the increase in COV of both TLs and the PL in the acute phase of SCI found in this study may reflect increased instability of the trunk cranial and caudal to the site of injury demonstrating involvement of the propriospinal pathways.

Several studies of rodent models of SCI reveal an increase in BS-PL following

SCI (Kunkel-Bagden 1990, Behrmann 1992, Stokes 1992, Kunkel-Bagden 1993, Metz

2000, Hamers 2001), with a gradual decrease in BS-PL with recovery. BS-PL was considered to be lower in animals receiving effective therapy interventions for acute SCI in comparison to non-treated animals (Stokes 1992, Keirstead 2005). However, other studies have found the BS-PL to vary depending on the severity, where animals with more severe lesions had decreased BS-PL and those with milder lesions had increased

BS-PL (Cheng 1997). Additionally, asymmetry of the lesion and recovery may also play a role in BS.

In dogs with SCI, measurements similar to BS to assess recovery of lateral stability have been made (Hamilton 2008, Jeffery 2011, Granger 2012, Lim 2014).

Lateral stability of the PL is largely determined by descending supraspinal pathways from the brainstem rather than propriospinal pathways or central pattern generators (Hamilton

2008). Previous studies have shown that dogs with thoracolumbar SCI of varying degree 71

had increased variability of lateral pelvic limb placement suggesting instability in comparison to normal dogs on a treadmill (Hamilton 2008). However, the use of a treadmill, as well as the use of an abdominal support band in dogs (Hamilton 2008) may add artificial stability and reflect differences from overground locomotion as was used in our study. An increase in BS-PL following SCI was not demonstrated in this study, which may reflect the variable injury severity and lesion asymmetry seen in this population.

However, our study did find significantly higher BS-TL in SCI-affected dogs at day 3 and 30 compared to normal dogs. Similar to our findings, a previous study reported significantly increased BS-TL in rats with thoracic SCI compared to rats without SCI

(McEwen 2006). This finding may reflect attempts to stabilize the trunk cranial to the lesion with use of a wider center of gravity, to compensate for instability and paresis of pelvic limbs (McEwen 2006).

Prior studies in rodent models of thoracic SCI have revealed functional reorganization of the sensory and motor cortex for increased representation of the trunk and thoracic limbs, and structural reorganization of damaged motor pathways in the form of increased collateral sprouting of the corticospinal pathway to increase connections in the cervical spinal cord immediately and weeks after injury (Fouad 2001, Bazley 2014,

Oza 2014, Yagüe 2014). Furthermore, changes in thoracic limb and trunk activity in rodent models of thoracic SCI have been shown, such as increased thoracic limb and back extensor muscle activity, increased stepping frequency of the thoracic limbs, increased weight bearing of the thoracic limbs, and increased peak vertical forces of the 72

thoracic limbs (Webb 2002, Ballermann 2006). The changes in SL and BS detected in the

TL of the SCI-affected dogs in our study underscore the importance of the adaptations of the trunk and TL in quadrupedal locomotor recovery from SCI. Future studies in footprint analysis in dogs with thoracolumbar SCI should include investigations of the TL, rather than limited PL parameters with locomotor recovery.

In rats, strain related differences in the amount of weight-bearing of the thoracic and pelvic limbs have been recognized (McEwen 2006). Similar breed and conformation

(ie. chondrodystrophic vs. non-chondrodystrophic) related differences in weight-bearing, footprint parameters and adaptive TL gait parameters may exist in dogs as well. Further studies will be needed in order to determine the presence and significance of such breed related differences as it relates to SCI and locomotor recovery.

There are some limitations to the use of footprint analysis in the method described herein as a sole outcome measure of SCI in dogs. The largest limitation was that dogs needed to be consistently ambulatory in order for the footprints to be of measurable value. Therefore, footprint analysis in dogs with more SCI, or in the earlier stages of recovery could not be performed. Additionally, although significantly cheaper in cost than available gait analytic equipment such as the Tekscan® or CatWalk, this method proved time and labor intensive.

4.6 Conclusions

This study has demonstrated that footprint analysis can be performed in dogs with

SCI using a simple ‘finger painting’ method with affordable supplies. Significant differences in footprint parameters in both the PLs and TLs were found between normal and SCI-affected dogs. This study supports the use of footprint analysis using this technique as an objective outcome measure of SCI in dogs.

4.7 Conflict of Interest Statement

The authors report no conflict of interest.

4.8 Acknowledgements

4.9 References

Aikawa T, Fujita H, Shibata M, Takahashi T. Recurrent thoracolumbar intervertebral disc extrusion after hemilaminectomy and concomitant prophylactic fenestration in 662 chondrodystrophic dogs. Vet Surg 2012;41:381-390.

Ballermann M, Tse ADY, Misiaszek JE, et al. Adaptations in the walking pattern of spinal cord injured rats. J Neurotrauma 2006;23:897-907.

Bazley FA, Maybhate A, Tan CS, et al. Enhancement of bilateral cortical somatosensory evoked potentials to intact forelimb stimulation following thoracic contusion spinal cord injury in rats. IEEE Trans Neural Syst Rehabil Eng 2014;22:953-964

Bergknut N, Egenvall A, Hagman R, Gustås P, Hazewinkel HA, Meij BP, Lagerstedt AS. Incidence of intervertebral disk degeneration-related diseases and associated mortality rates in dogs. J Am Vet Med Assoc 2012;240:1300-1309.

Borgens RB, Toombs JP, Breur G, Widmer WR, Waters D, Harbath AM, March P, Adams LG. An imposed oscillating electrical field improves the recovery of function in neurologically complete paraplegic dogs. J Neurotrauma 1999;16:639-657.

Bregman BS, Kunkel-Bagden E, Schnell L, Dai HN, Gao D, Schwab ME. Recovery from spinal cord injury mediated by antiboides to neurite growth inhibitors. Nature 1995;378:498-501.

Cheng H, Almström S, Giménex-Llort L, Chang R, Ove Ogren S, Hoffer B, Olson L. Gait Analysis of adult paraplegic rats after spinal cord repair. Exp Neurology 1997;148:544-557.

Collazos-Castro JE, López-Dolado E, Nieto-Sampedro M. Locomotor deficits and adaptive mechanisms after thoracic spinal cord contusion in the adult rat. J Neurotrauma 2006;23:1-17. de Medinaceli L, Freed WJ, Wyatt JR. An Index of the functional condition of rat sciatic nerve based on measurements made from walking tracks. Exp Neurol 1982;77:634-643.

Fouad K, Pedersen V, Schwab ME, et al. Cervical sprouting of corticospinal fibers after thoracic spinal cord injury accompanies shifts in evoked motor responses. Curr Biol 2001;11:1766-1770.

Gordon-Evans WJ, Evans RB, Conzemius MG. Accuracy of spatiotemporal variables in gait analysis of neurologic dogs. J Neurotrauma 2009;26:1055-1060.

Granger N, Blamires H, Franklin RJ, Jeffery ND. Autologous olfactory mucosal cell transplants in clinical spinal cord injury: a randomized double-blinded trial in a canine translational model. Brain 2012;135:3227-3237.

Hamers FP, Lankhorst AJ, Van Laar TJ, Veldhuis WB, Gispen WH. Automated quantitative gait analysis during overground locomotion in the rat: its application to spinal cord contusion and transection injuries. J Neurotrauma 2001;18:187-201.

Hamers FP, Kiipmans GC, Joosten EAJ. Catwalk-assisted gait analysis in the assessment of spinal cord injury. J Neurotrauma 2006;23:537-548.

Hamilton L, Franklin RJ, Jeffery ND. Quantification of deficits in lateral paw positioning after spinal cord injury in dogs. BMC Vet Res 2008;4:47.

Ito D, Matsunaga S, Jeffery ND, Sasaki N, Nishimura R, Mochizuki M, Kasahara M, Fujiwara R, Ogawa H. Prognostic value of magnetic resonance imaging in dogs with paraplegia caused by thoracolumbar intervertebral disk extrusion: 77 cases (2000-2003). J Am Vet Med Assoc 2005;227:1454-1460.

Jeffery ND, Smith PM, Lakatos A, Ibanez C, Ito D, Franklin RJ. Clinical canine spinal cord injury provides an opportunity to examine the issues in translating laboratory techniques into practical therapy. Spinal Cord 2006;44:584-593.

Jeffery ND, Hamilton L, Granger N. Designing clinical trials in canine spinal cord injury as a model to translate successful laboratory interventions into clinical practice. Vet Rec 2011;168:102-107.

Keirstead HS, Nitor G, Bernal, Totoiu M, Cloutier F, Sharp K, Steward O. Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remyelinate and restore locomotion after spinal cord injury. J Neurosci 2005;25:4694-4705.

Klapdor K, Dulfer BG, Hammann A, Van der Staay FJ. A low-cost method to analyse footprint patterns. J Neurosci Methods 1997;75:49-54.

Kunkel-Bagden E, Bregman BS. Spinal cord transplants enhance the revoery of locomotor function after spinal cord injury at birth. Exp Brain Res 1990;81:25-34.

Kunkel-bagden E, Dai HN, Bregman BS. Methods to assess the development and recovery of locomotor function after spinal cord injury in rats. Exp Neurol 1993;119:153-164.

Laverty PH, Leskovar A, Breur GJ, Coates JR, Bergman RL, Widmer WR, Toombs JP, Shapiro S, Borgens RB. A preliminary study of intravenous surfactants in paraplegic dogs: polymer therapy in canine clinical SCI. J Neurotrauma 2004;21:1767-1777.

Levine JM, Levine GJ, Porter BF, Topp K, Noble-Haeusslein LJ. Naturally occurring disk herniation in dogs: an opportunity for pre-clinical spinal cord injury research. J Neurotrauma 2011;28:675-688.

Lim JH, Muguet-Chanoit AC, Smith DT, Laber E, Olby NJ. Potassium channel antagonists 4-aminopyridine and the T-butyl carbamate derivative of 4-aminopyridine improve hind limb function in chronically non-ambulatory dogs; a blinded, placebo- controlled trial. PLoS One 2014;9:e116139.

Metz GA, Merkler D, Dietz V, Schwab ME, Fouad K. Efficient testing of motor function in spinal cord injured rats. Brain Res 2000;883:165-177.

McEwen ML, Springer JE. Quantification of locomotor recovery following spinal cord contusion in adult rats. J Neurotrauma 2006;23:1632-1653. 76

Olby N, Harris T, Burr J, Muñana K, Sharp N, Keene B. Recovery of pelvic limb function in dogs following acute intervertebral disc herniations. J Neurotrauma 2004;21:49-59.

Oza CS, Giszter SF. Plasticity and alterations of trunk motor cortex following spinal cord injury and non-stepping robot and treadmill training. Exp Neurology 2014;256:57-69.

Packer RMA, Hendricks A, Volk HA, Shihab NK, Burn CC. How long and low can you go? Effect of conformation on the risk of thoracolumbar intervertebral disc extrusion in domestic dogs. PLoS One 2013;8: 1-11.

Plemel JR, Duncan G, Chen KW, Shannon C, Park S, Sparling JS, Tetzlaff W. A graded forceps crush spinal cord injury model in mice. J Neurotrauma 2008;25:350-370.

Rangasamy SB. Locomotor recovery after spinal cord hemisection/contusion injuries in Bonnet Monkeys: footprint testing-a minireview. Synapse;2013:427-453.

Rice AS, Cimino-Brown D, Eisenach JC, Kontinen VK, Lacroix-Fralish ML, Machin I; Preclinical Pain Consortium, Mogil JS, Stöhr T. Animal models and the prediction of efficacy in clinical trials of analgesic drugs: A critical appraisal and call for uniform reporting standards. Pain 2009;139:243-247.

Stokes BT, Reier PJ. Fetal grafts alter chronic behavioral outcome after contusion damage to the adult rat spinal cord. Exp Neuro 1992;116:1-12.

Varejao AS, Cabrita AM, Meek MF, Bulas-Cruz J, Melo-Pinto P, Raimondo S, Geuna S, Giacobini-Robecchi MG. Functional and morphological assessment of a standardized rat sciatic nerve crush injury with a non-serrated clamp. J Neurotrauma 2004;21:1652-1670.

Webb AA, Muir GD. Compensatory locomotor adjustments of rats with cervical or thoracic spinal cord hemisections. J Neurotrauma 2002;19:239-256.

Yagüe JG, Humanes-Valera D, Aguilar J, et al. Functional reorganization of the forepaw cortical representation immediately after thoracic spinal cord hemisection in rats. Exp Neurol 2014;257:19-24.

Table 4 Mean and coefficient of variance (COV) of stride lengths (cm) for each limb across 3 testing sessions in normal dogs (n=20)

Dog Stride Length LTL Stride Length RTL Stride Length LPL Stride Length RPL Session 1 Session 2 Session 3 Session 1 Session 2 Session 3 Session 1 Session 2 Session 3 Session 1 Session 2 Session 3

Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV

1 42.75 0.13 42.84 0.07 41.51 0.12 40.72 0.09 42.16 0.08 40.47 0.12 41.28 0.13 41.96 0.07 41.19 0.12 40.80 0.12 41.60 0.08 41.38 0.12

2 46.03 0.16 54.74 0.05 51.43 0.17 44.99 0.14 54.48 0.05 47.64 0.19 45.89 0.16 54.20 0.05 48.97 0.18 47.00 0.16 53.92 0.05 47.62 0.18

3 50.69 0.20 45.19 0.07 52.68 0.09 50.58 0.19 45.01 0.08 52.78 0.08 52.02 0.16 44.17 0.09 53.60 0.06 51.97 0.17 43.39 0.09 53.79 0.06

4 46.66 0.08 49.70 0.06 49.13 0.06 47.38 0.07 49.86 0.05 49.71 0.06 45.10 0.12 49.20 0.07 48.96 0.06 46.66 0.09 49.10 0.07 48.50 0.09

5 36.32 0.22 43.43 0.11 36.80 0.09 36.25 0.21 44.07 0.11 36.45 0.11 40.23 0.11 43.22 0.12 36.83 0.10 38.87 0.13 43.47 0.12 36.24 0.12

6 34.42 0.07 27.67 0.16 24.24 0.10 34.08 0.08 28.95 0.17 24.10 0.09 34.34 0.07 27.05 0.16 24.21 0.10 34.03 0.08 27.57 0.16 24.21 0.10

7 33.07 0.06 32.50 0.10 37.48 0.10 33.08 0.07 32.40 0.09 36.64 0.11 32.33 0.08 32.13 0.10 37.12 0.10 32.77 0.06 32.34 0.11 36.83 0.11

8 35.64 0.10 40.51 0.11 41.40 0.13 34.77 0.14 40.63 0.12 41.63 0.11 35.27 0.10 40.40 0.11 42.63 0.09 34.99 0.14 40.01 0.12 42.46 0.10

78 9 44.73 0.17 57.85 0.12 50.30 0.15 47.08 0.20 55.92 0.16 49.56 0.16 43.22 0.21 51.56 0.16 47.95 0.17 44.05 0.22 53.19 0.16 47.70 0.19

10 40.23 0.18 37.57 0.13 42.07 0.21 41.24 0.17 37.98 0.12 46.26 0.16 39.12 0.19 38.40 0.13 43.70 0.25 39.60 0.18 37.09 0.13 42.48 0.21

11 44.23 0.05 40.85 0.09 42.18 0.05 44.62 0.04 40.67 0.12 42.09 0.06 44.37 0.04 39.67 0.10 40.09 0.08 43.68 0.05 40.26 0.11 40.31 0.06

12 46.06 0.17 60.61 0.05 56.64 0.11 46.45 0.16 61.24 0.05 57.49 0.10 47.63 0.21 60.30 0.06 54.96 0.12 49.25 0.20 60.78 0.07 56.03 0.10

13 38.50 0.14 37.71 0.08 38.95 0.10 38.71 0.11 38.37 0.07 38.47 0.11 39.00 0.12 37.86 0.09 38.88 0.11 38.50 0.12 37.58 0.11 38.54 0.10

14 44.60 0.10 45.95 0.16 46.21 0.06 46.22 0.09 44.43 0.18 46.32 0.07 45.02 0.10 45.09 0.18 45.91 0.06 45.40 0.10 45.82 0.18 46.10 0.07

15 47.99 0.05 48.39 0.07 47.66 0.07 47.58 0.05 48.54 0.07 46.87 0.09 47.74 0.06 47.09 0.09 47.48 0.07 47.41 0.06 47.93 0.08 46.67 0.12

16 40.83 0.19 44.55 0.15 51.74 0.09 40.97 0.23 46.22 0.12 50.96 0.10 41.58 0.17 44.24 0.14 50.98 0.06 41.94 0.18 44.83 0.14 51.40 0.06

17 50.00 0.09 44.76 0.09 54.68 0.09 49.53 0.09 44.42 0.09 55.14 0.06 49.25 0.08 43.73 0.08 54.23 0.09 48.78 0.08 43.45 0.09 55.47 0.07

18 28.43 0.25 30.60 0.21 33.33 0.29 28.33 0.27 28.38 0.30 33.91 0.33 30.73 0.13 30.50 0.25 33.90 0.21 29.51 0.21 31.20 0.20 34.30 0.25

19 55.82 0.09 57.94 0.08 56.25 0.12 55.74 0.11 57.94 0.07 57.38 0.13 51.64 0.15 53.87 0.13 56.88 0.11 52.58 0.13 54.18 0.15 56.28 0.14

20 66.81 0.10 62.59 0.11 61.73 0.08 65.23 0.11 62.79 0.11 60.99 0.06 62.75 0.11 57.22 0.16 59.56 0.12 63.02 0.15 55.52 0.16 57.58 0.11

MEAN 43.69 0.13 45.30 0.10 45.82 0.11 43.68 0.13 45.22 0.11 45.74 0.12 43.43 0.13 44.09 0.12 45.40 0.11 43.54 0.13 44.16 0.12 45.19 0.12

Dog Stride Length LTL Stride Length RTL Stride Length LPL Stride Length RPL

Day 3 Day 10 Day 30 Day 3 Day 10 Day 30 Day 3 Day 10 Day 30 Day 3 Day 10 Day 30 CO CO CO CO CO CO CO CO CO CO CO Mean Mean Mean Mean COV Mean Mean Mean Mean Mean Mean Mean Mean V V V V V V V V V V V 1 36.09 0.11 36.23 0.16 36.17 0.13 35.94 0.11 37.31 0.16 37.93 0.14 35.43 0.14 38.28 0.12 40.67 0.09 36.76 0.15 38.38 0.13 39.51 0.16

2 34.80 0.20 45.72 0.08 42.50 0.08 34.48 0.19 45.20 0.09 42.22 0.07 34.40 0.20 45.03 0.10 41.17 0.09 34.09 0.22 44.08 0.13 41.00 0.10

3 27.05 0.18 32.43 0.18 30.87 0.15 26.82 0.17 32.43 0.18 31.10 0.14 26.38 0.16 32.98 0.19 30.74 0.15 26.84 0.15 32.68 0.19 30.59 0.16

79 5 17.79 0.23 34.92 0.09 33.40 0.06 18.43 0.22 35.22 0.09 33.32 0.08 27.69 0.39 35.47 0.07 32.86 0.11 18.58 0.19 35.12 0.07 33.13 0.08

6 34.56 0.20 44.57 0.05 38.58 0.07 35.29 0.17 45.04 0.05 38.50 0.07 34.91 0.18 44.93 0.05 38.39 0.05 35.75 0.15 44.47 0.07 37.92 0.06

7 49.41 0.10 51.22 0.07 50.24 0.08 50.44 0.10 59.14 0.15 51.72 0.06 58.41 0.12 52.23 0.07

8 21.86 0.17 26.21 0.17 22.14 0.20 27.14 0.14 24.54 0.28 34.14 0.21 27.71 0.39 31.91 0.16

9 43.94 0.19 52.42 0.08 45.50 0.15 50.51 0.15 42.91 0.22 51.26 0.12 42.90 0.20 53.43 0.12

11 34.02 0.15 36.52 0.18 34.20 0.13 37.11 0.16 35.41 0.27 35.23 0.17 33.95 0.12 36.11 0.16

12 46.90 0.07 41.80 0.14 47.36 0.14 45.86 0.06 42.36 0.15 47.58 0.14 47.27 0.06 41.16 0.14 48.18 0.14 45.27 0.04 41.83 0.18 46.45 0.14

13 41.98 0.16 45.42 0.12 51.02 0.09 40.91 0.19 44.49 0.13 51.48 0.08 44.59 0.15 45.44 0.13 51.02 0.06 42.41 0.20 45.11 0.15 51.55 0.04

14 38.91 0.13 38.65 0.16 40.44 0.11 40.51 0.07

15 22.19 0.21 23.49 0.19 28.71 0.18 30.39 0.18

16 41.70 0.19 39.02 0.12 38.64 0.15 40.44 0.15 39.57 0.13 39.53 0.15 41.08 0.08 45.43 0.08 48.20 0.10 42.00 0.07 46.49 0.08 47.47 0.11 Continued 79

Table 5 Continued

17 14.58 0.37 22.40 0.17 24.96 0.29 15.94 0.27 22.38 0.16 25.23 0.27 19.42 0.15 23.06 0.15 27.11 0.17 17.34 0.25 22.38 0.17 25.12 0.27

18 29.13 0.12 29.58 0.13 30.59 0.11 30.99 0.13

19 21.51 0.30 22.03 0.34 20.91 0.29 21.13 0.35 27.76 0.20 24.51 0.25 24.89 0.17 23.77 0.24

20 22.80 0.17 22.95 0.17 33.39 0.11 31.62 0.19

21 25.07 0.21 25.07 0.20 27.10 0.21 28.52 0.16

22 34.09 0.21 33.56 0.25 44.41 0.36 47.75 0.35

23 24.44 0.40 23.64 0.47 32.92 0.31 29.49 0.24 30.69 0.25 37.30 0.25 28.16 0.24 29.00 0.30 35.94 0.23 29.61 0.23 31.49 0.23 37.29 0.24

24 19.54 0.25 27.02 0.13 19.92 0.23 27.70 0.11 24.91 0.39 30.04 0.24 26.31 0.23 28.57 0.23

25 23.49 0.21 24.22 0.14 37.85 0.24 32.65 0.24

26 22.36 0.27 35.42 0.07 36.60 0.09 22.43 0.25 35.83 0.05 37.09 0.07 23.10 0.22 35.57 0.08 35.68 0.17 23.35 0.25 35.61 0.09 34.17 0.24

27 25.77 0.14 33.13 0.08 24.95 0.17 34.00 0.08 25.97 0.13 33.62 0.10 30.47 0.40 35.21 0.12

28 24.65 0.28 25.77 0.20 35.33 0.15 23.85 0.32 26.16 0.21 35.58 0.15 25.81 0.22 26.11 0.19 36.36 0.13 26.69 0.21 26.32 0.16 36.29 0.12

80 29 39.35 0.09 43.63 0.12 39.72 0.07 43.35 0.13 44.33 0.12 45.21 0.07 42.84 0.12 44.99 0.07

30 27.88 0.22 29.41 0.13 37.58 0.16 29.08 0.18 29.54 0.11 36.74 0.16 29.91 0.16 29.65 0.12 36.01 0.13 30.28 0.16 29.81 0.12 35.80 0.16

MEAN 30.37 0.22 33.91 0.16 34.78 0.15 30.69 0.20 34.47 0.15 35.09 0.15 32.17 0.18 36.05 0.17 37.52 0.15 31.46 0.18 36.25 0.17 37.32 0.16

Table 6 Mean and coefficient of variance of base of support (cm) in the thoracic and pelvic limbs across 3 testing sessions in normal dogs (n=20)

Dog Base of Support TL Base of Support PL

Session 1 Session 2 Session 3 Session 1 Session 2 Session 3 Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV 1 6.66 0.48 6.44 0.41 6.13 0.44 7.30 0.34 6.80 0.34 7.07 0.41 2 7.77 0.54 7.41 0.47 9.73 0.54 8.99 0.53 6.19 0.64 8.17 0.59 3 10.30 0.34 8.22 0.29 9.43 0.31 11.18 0.37 12.88 0.28 13.04 0.24 4 8.64 0.31 7.54 0.34 8.27 0.40 9.25 0.22 7.77 0.19 8.24 0.36 5 4.94 0.44 5.02 0.48 4.99 0.46 4.26 0.60 4.26 0.60 5.48 0.37 6 3.77 0.55 3.64 0.59 5.32 0.40 6.47 0.35 6.85 0.32 6.20 0.25 7 6.39 0.20 5.70 0.34 6.49 0.34 7.07 0.23 7.28 0.22 6.33 0.32 8 4.82 0.49 4.73 0.32 4.24 0.47 8.96 0.31 8.85 0.29 8.66 0.21 9 6.06 0.54 4.74 0.85 5.63 0.80 10.33 0.36 9.04 0.37 9.05 0.47 10 4.91 0.68 5.84 0.43 3.83 0.73 8.20 0.53 8.29 0.37 8.00 0.41 11 6.82 0.41 7.73 0.36 7.26 0.40 5.32 0.47 5.05 0.67 5.02 0.50 12 7.53 0.49 7.74 0.43 7.60 0.27 15.49 0.34 13.00 0.24 17.23 0.17 13 4.12 0.57 3.27 0.61 3.38 0.52 7.07 0.33 7.80 0.27 6.96 0.30 14 5.85 0.45 5.92 0.36 5.09 0.50 6.67 0.39 7.77 0.24 7.44 0.28 15 5.80 0.27 5.80 0.40 6.65 0.35 9.61 0.17 8.72 0.24 9.47 0.28 16 7.23 0.52 7.18 0.50 6.42 0.45 12.90 0.31 11.14 0.28 10.82 0.19 17 6.30 0.40 5.68 0.46 4.83 0.44 4.02 0.66 5.96 0.43 2.93 0.66 18 7.93 0.64 7.58 0.35 7.80 0.55 10.03 0.35 9.97 0.30 10.98 0.28 19 3.81 0.79 5.88 0.61 4.21 0.65 3.37 0.75 3.38 0.64 2.56 0.73 20 6.00 0.48 4.92 0.66 6.30 0.54 7.15 0.60 5.96 0.55 4.11 0.70 MEAN 6.28 0.48 6.05 0.46 6.18 0.48 8.18 0.41 7.85 0.37 7.89 0.39

81 Table 7 Mean and coefficient of variance of stride lengths (cm) for thoracic and pelvic limbs in dogs with thoracolumbar spinal cord injury due to acute intervertebral disc extrusion at day 3, 10 and 30 following decompressive surgery (n=30). Missing values represent dogs that were not able to consistently walk without support in the pelvic limbs at time of testing.

Base of Support FL Base of Support HL Dog

Day3 Day 10 Day 30 Day 3 Day 10 Day 30 Mean COV Mean COV Mean COV Mean COV Mean COV Mean COV 1 3.95 0.57 4.48 0.73 4.02 0.73 9.12 0.44 10.74 0.40 11.27 0.31 2 6.60 0.40 5.67 0.43 5.74 0.31 7.10 0.40 6.38 0.35 8.06 0.36 3 7.06 0.29 6.59 0.30 5.91 0.28 4.18 0.40 5.70 0.31 6.46 0.31 4 5 7.34 0.32 7.43 0.26 8.71 0.31 7.23 0.89 5.81 0.28 6.53 0.31 6 10.30 0.19 10.30 0.17 10.12 0.15 10.36 0.24 10.12 0.17 10.58 0.23 7 9.33 0.43 11.26 0.30 6.86 0.84 9.74 0.31 8 5.54 0.47 6.33 0.33 4.65 0.66 2.44 0.65 9 8.93 0.54 8.81 0.52 8.31 0.72 8.21 0.44 10 11 6.18 0.42 4.80 0.54 4.52 0.50 5.87 0.27 12 12.05 0.15 9.99 0.48 10.02 0.33 11.75 0.32 16.74 0.35 15.95 0.13 13 10.07 0.43 8.01 0.40 6.56 0.48 4.89 0.77 4.98 0.58 5.79 0.51 14 6.86 0.32 2.05 1.05 15 6.49 0.33 6.88 0.45 16 9.58 0.38 8.37 0.31 10.15 0.33 10.64 0.27 10.95 0.24 8.76 0.36 17 9.00 0.35 6.99 0.26 7.17 0.34 6.71 0.49 6.20 0.31 7.52 0.30 18 9.45 0.23 9.86 0.22 19 5.02 0.43 6.28 0.45 8.46 0.44 8.05 0.45 20 5.09 0.46 6.42 0.49 21 12.49 0.28 12.31 0.24 22 12.24 0.32 10.10 0.71 23 11.24 0.39 14.46 0.36 11.15 0.36 12.31 0.46 14.23 0.28 13.31 0.28 24 7.53 0.73 5.60 0.53 7.62 0.77 7.22 0.38 25 10.66 0.30 12.29 0.53 26 9.26 0.43 6.20 0.55 6.67 0.34 6.17 0.65 5.21 0.45 5.19 0.71 27 8.08 0.27 7.18 0.28 7.60 0.35 7.23 0.29 28 6.73 0.58 4.52 0.61 5.91 0.46 8.51 0.41 7.64 0.36 3.79 0.49 29 10.06 0.28 8.85 0.35 3.85 0.84 6.64 0.32 30 6.16 0.36 6.14 0.27 6.09 0.37 3.68 0.63 6.15 0.48 6.69 0.33 Mean 8.41 0.37 7.61 0.41 7.88 0.37 7.90 0.49 7.75 0.46 8.04 0.41

82 Table 8 Comparison in mean stride lengths of all limbs between the first testing session of normal dogs and all testing sessions in SCI-affected dogs at day 3, 10 and 30 following decompressive surgery Mean SL is significantly lower (p<0.05) in in SCI-affected dogs compared to normal dogs in all limbs at all SCI-affected testing sessions.

Difference between Testing Session of Limb Normal and SCI- p-value 95% CI SCI-affected dogs affected dogs Left Day 3 13.32 <.0001 (7.1, 19.5) Thoracic Day 10 9.78 0.0004 (4.4, 15.2) Limb Day 30 8.91 0.0006 (3.8, 14) Right Day 3 12.99 <.0001 (6.8, 19.2) Thoracic Day 10 9.21 0.0009 (3.8, 14.6) Limb Day 30 8.59 0.001 (3.5, 13.7) Left Day 3 11.26 0.0004 (5.1, 17.4) Pelvic Day 10 7.37 0.0079 (1.9, 12.8) Limb Day 30 5.91 0.0229 (0.8, 11) Right Day 3 12.08 0.0001 (5.9, 18.3) Pelvic Day 10 7.29 0.0086 (1.9, 12.7) Limb Day 30 6.22 0.0166 (1.1, 11.3)

83 Table 9 Comparison in mean base of support of the thoracic and pelvic limbs between the first testing session of normal dogs and all testing session in SCI-affected dogs at day 3, 10 and 30 following decompressive surgery. Statistically significant values (p<0.05) are italicized.

Testing Session Difference between Base of p- of SCI-affected Normal and SCI- 95% CI Support value dogs affected dogs Day 3 -2.13 0.0234 (-4, -0.3) Thoracic Day 10 -1.33 0.1055 (-2.9, 0.3) Limbs Day 30 -1.6 0.0382 (-3.1, -0.1) Day 3 0.29 0.7591 (-1.6, 2.1) Pelvic Day 10 0.43 0.5961 (-1.2, 2) Limbs Day 30 0.14 0.8562 (-1.4, 1.6

Figures

Figure 9 Mean base of support (BS) of the thoracic limbs (TL) and the pelvic limbs (PL) in normal dogs compared to spinal cord injury (SCI) - affected dogs at day 3, 10 and 30 following decompressive surgery. Whiskers represent ± standard error of the mean. Asterisk denotes statistically significant differences from normal dogs (p<0.05). Mean BS-TL was higher in SCI-affected dogs at all time points when compared to normal dogs. Mean BS-PL in SCI-affected dogs was not significantly different from normal dogs at any time point.

Chapter 5

Conclusions, Limitations, and Future Directions

5.1 Conclusions

The primary goal of this study was to determine the utility and repeatability of novel quantitative outcome measures to assess recovery in dogs with SCI. Specifically this study evaluated the use of electronic VFA to objectively measure ST in normal dogs and dogs with SCI throughout their recovery. Additionally, this study evaluated the use of a simple ‘finger painting’ technique in order to evaluate footprint parameters of SL and

BS in normal dogs and dogs with SCI throughout their recovery.

The results of this study supported the utility of VFA to objectively measure changes in sensory function for dogs with acute SCI. A significant decline in pelvic limb

ST values correlated inversely with locomotor recovery in SCI-affected dogs. However, smaller but significant declines in ST values in normal dogs with repeated testing and in thoracic limbs of SCI-affected dogs over time were also detected. Overall, these findings support the use of VFA as a quantitative assessment of neurologic recovery for future veterinary clinical trials in SCI. However, future investigators should be aware of the decline in ST values in normal dogs and unaffected limbs over testing sessions. Future studies should focus on techniques to minimize or compensate for clinical, environmental, and behavioral factors that may impact ST values across testing sessions.

Additionally, this study has demonstrated that footprint analysis is easily performed in a clinical setting using readily available materials. Parameters SL and BS were determined in all limbs of normal dogs and SCI-affected dogs at various time-points during their recovery. No significant changes were detected in SL and BS across testing sessions of normal dogs. Therefore, it appears that footprint analysis is not subject to confounding environmental or behavioral factors that the VFA study results suggested.

SL in both pelvic and thoracic limbs was significantly lower in dogs with thoracolumbar

SCI. A gradual increase in SL in all limbs was documented with locomotor recovery from SCI. In normal dogs, the extensor muscles are crucial for the majority of the mechanical work involved in the stance phase of locomotion (Collazos-Castro 2006). The initial decrease in SL of the pelvic limbs may be due to weakness associated loss of supraspinal excitatory input to the pelvic limb motor neurons and subsequent loss of extensor muscle, leading to a shortened stride (Collazos-Castro 2006). The decrease in

SL of the thoracic limbs may represent compensatory mechanisms by increasing body weight onto the thoracic limbs, resulting in increased step frequency and a shortened stride (Cheng 1997, Hamers 2001, McEwen 2006, Gordon-Evans 2009). BS-TL was also significantly increased (wider) in SCI-affected dogs compared to normal dogs at day 3 and 30 following spinal cord injury whereas no significant differences in BS-PL were detected. The reason for the lack of significance at day 10 is unknown and my be due to a variety of factors such as a statistical type II error, lack of statistical power, etc. Prior studies in rodent models of thoracic SCI have revealed functional reorganization of the

88 sensory and motor cortex for increased representation of the trunk and thoracic limbs, and structural reorganization of damaged motor pathways in the form of increased collateral sprouting of the corticospinal pathway to increase connections in the cervical spinal cord immediately and weeks after injury (Fouad 2001, Bazley 2014, Oza 2014, Yagüe 2014).

Furthermore, changes in thoracic limb and trunk activity in rodent models of thoracic SCI have been shown, such as increased thoracic limb and back extensor muscle activity, increased stepping frequency of the thoracic limbs, increased weight bearing of the thoracic limbs, and increased peak vertical forces of the thoracic limbs (Webb 2002,

Ballermann 2006). Therefore, it is likely that the thoracic limb and trunk stability is further enhanced with increased motor and sensory cortical representation in dogs with

SCI. Furthermore changes seen in the thoracic limbs of SCI-affected dogs may be explained by compensatory mechanisms, or may be due to disruption of the long ascending propriospinal neurons important for inter-limb coordination (Webb 2002,

Collazos-Castro 2006, Kanagal 2008). Additionally, it is difficult to determine whether gait changes following SCI reflect represent true recovery and return to normalcy, or if they represent adaptive mechanisms (Webb 2002). Regardless, these findings underscore the importance of evaluating all limbs, rather than just the affected pelvic limbs in quadruped animals with thoracolumbar SCI.

Most outcome measures of SCI-affected dogs today are focused on locomotor recovery. This study has established the utility and repeatability of two additional outcome measures that can be used to document recovery from SCI. The identification

89 of such objective, sensitive and valid outcome measures will prove to be invaluable in order for canine spontaneous SCI to be established as a successful model for human SCI.

The easy application and use of VFA and footprint parameters will additionally allow for large, multi-center studies needed to establish efficacy of a potential therapeutic intervention for SCI, towards the common goal for finding a cure for this devastating condition.

5.2 Limitations

The use of behavioral outcomes measures in animals has inherent limitations due to the animal’s inability to self-report and the subjective nature of the human’s interpretation of the animal’s behavior. The use of electronic VFA and footprint parameters attempts to minimize the subjectivity of behavioral outcome measures.

Despite our attempts to make ST values objective by use of a standard, electronic

VFA device, there is still some subjective interpretation of the endpoint behavioral response among dogs. For example, dogs of a stoic nature with a calm demeanor and minimal emotional expression, may have had a falsely elevated ST values due to the lack of supraspinal cues to VFA probe stimulation. Conversely, anxious dogs with higher emotional expression may have led to falsely lowered ST values due to rapid supraspinal cues such as vocalization and attempts to bite VFA probe stimulation. Several techniques were employed in order to minimize the potential for anxiety and environment related behaviors such as allowing a minimum of 15 minutes of acclimation to the room prior to

90 each testing session, use of the same quiet and dim room with minimal traffic, use of the least amount of restraint possible, minimizing conversation between investigators during testing session, use of a single trained investigator with the same technique to minimize intra and inter-rater variability. Future investigators should similarly identify and minimize controllable environmental factors that may affect VFA ST values.

Another limitation in this study was the inability to evaluate the individual effects of systemically administered anti-inflammatory medications and analgesic medications on VFA ST values in SCI-affected dogs. In this study, subjects were recruited from multiple clinicians, who used their medical judgment to determine the best anti- inflammatory and analgesic medication protocol for each patient. For obvious reasons, withholding of such medications following decompressive surgery was deemed unethical in this clinical trial. The majority of the SCI-affected dogs were administered steroids or

NSAIDs, a fentanyl patch (Duragesic ®), tramadol, gabapentin, diazepam or methocarbamol. Fentanyl patches have previously been shown to provide analgesia up to

72 hours in dogs following application (Bellei 2011). It has been shown that fentanyl can significantly increase the withdrawal latency in response to thermal stimulus in a dog nociceptive model (Wegner 2008). Similarly, gabapentin has been shown to increase

VFA ST values in rat models of nerve injury (Matthews 2002) although another study did not show significant differences in pain between dogs that received or did not receive gabapentin (Aghighi 2012). Similarly, mixed evidence on the anti-nociceptive effects of tramadol have been reported in dogs. Studies in dogs and rodent models have shown that

91 tramadol significantly increases VFA sensory thresholds (KuKanich 2011, Kögel 2014) but other studies shown no effect in different dog models of acute pain such as thermal heat induced tail flick test (Kögel 2014). NSAIDs have also previously been shown to increase mechanical nociceptive threshold measurements (Lascelles 1998). Based on these studies, it is likely that the administration of medications increased VFA ST values in SCI-affected dogs compared to normal dogs that were not on medications. Such effects of medications may have artificially increased VFA ST values in SCI-affected dogs. However, systemic medication would be expected to affect all limbs, and cease to have an effect with discontinuation of medications. Future studies should evaluate the individual effects of each medication, as well as the different combinations of medications on electronic VFA ST in normal dogs and SCI-affected dogs. Ideally, future clinical trials should use a standardized medication protocol in all SCI-affected dogs in order to minimize medication related effects.

The largest limitation of the footprint analysis technique is that dogs need to be consistently ambulatory in order for the footprints to be of measurable value. Therefore, footprint analysis of the pelvic limbs in dogs with more severe SCI, or in the earlier stages of recovery cannot be performed. Therefore, footprint analysis should not be used as the sole outcome measure in dogs with variable severity of spinal cord injury.

However, it can be advantageous to evaluate objective footprint parameters in higher functioning animals. Prior studies in rat models of SCI have shown that assignment of overground locomotor scores was most reliable in the mid-range of the score scale

92 compared to the lower the higher ends of the score scale (Basso 1996). Our experience in this study revealed similar challenges in using the currently most commonly employed overground locomotor scale in dogs with SCI, the OSCIS. The OSCIS proved difficult to interpret in the higher functioning scores due to vague operational definitions, leading to inter-rater variability. Objective measurements of footprint parameters can allow improved evaluation of gait characteristics and compensate for the limitations of the human eye in assigning locomotor gait scores.

Another limitation of the footprint analysis method used in this study is the variable velocity at which the subjects were walked during footprint acquisition.

Different gait velocities may change footprint parameters and step sequences (Metz 2000,

Sedý 2008). However, other studies in dogs with SCI have revealed no significant changes in thoracic or pelvic limb lateral paw positioning with differing velocities in treadmill assisted gait assessment (Hamilton 2008). In this study, the dogs were walked at a steady, controlled walk on a leash. Ideally, future studies should either standardize gait velocity, or further establish whether differences in gait velocity have a significant impact on footprint parameters in normal dogs or dogs with SCI during overground locomotion.

Additionally, although significantly cheaper in cost than available automated gait analytic equipment such as the Tekscan® or CatWalk, this method proved time and labor intensive. Required equipment included a non-carpeted room with easily cleanable floors, a room large enough to accommodate five strips of butcher paper at least 10 feet

93 in length, a generous supply of non-toxic paint in four different colors, at least two investigators during each testing session to administer the paint and walk the dog across each butcher paper, and sufficient time and space to allow the paint to dry prior to manual measurements of each footprint parameter and storage of the prints.

In general, behavioral response tests such as the VFA and footprint analysis may be affected by other non-neurologic factors such as systemic health, environmental factors, body part stimulated, instrument used, time of day, and different investigator technique (Lascelles 2013). Prior studies have shown that diversions of attention in animals can alter perceptions of noxious stimulus (Lascelles 1997). For this study, controllable variables such as room, acclimation, technique, stimulating probe, and investigator were kept consistent but other factors such as time of day, patient temperament, outside distractions (ie construction noise, smells, distant voices, dog barking) were more difficult to control and may have attributed to the variation in data.

There also may be environmental and psychological factors experienced by hospitalized and injured patients such as stress related endogenous opioid release, sleep disturbances and decreased peripheral limb temperature that may affect an animal’s cognitive concentration and thus, testing results (Lascelles 1995, Lascelles 1997, Tena 2012).

Future investigators using the various behavioral response tests as outcome measures of

SCI should be aware of such factors and limit controllable variables as best as possible.

94 An ideal behavioral response test in dogs is yet to be determined. Every behavioral response test will have its strengths and limitations (Basso 2004, Sedý 2008).

However, one test’s weakness can be mitigated by use of another test’s strengths. As dogs with spontaneous SCI become a more popular large animal model for human SCI, the establishment and identification of different behavioral outcome measures with unique strengths and weakness is important. Future investigators should be aware of the strengths and weaknesses of behavior tests established in the dog, and use multiple behavioral outcomes measures within a single study that is most appropriate for the clinical trial (Muir 2000, Webb 2002, Basso 2004, Sedý 2008).

5.3 Future directions

There are many more applications of the VFA and footprint analysis that remains largely unexplored in evaluating normal dogs and SCI-affected dogs. Future studies will likely provide investigators with much more information and knowledge towards the common goal of improving the quality of life of animals and humans affected with SCI.

One application of the VFA that was not evaluated in this study is the evaluation of SCI-affected dogs for the development of mechanical allodynia and/or chronic neuropathic pain following therapy. Nearly 65% of human patients with SCI can experience spontaneous or evoked chronic neuropathic pain (Hoschouer 2010). Previous studies in rat and mouse models of SCI have shown that mechanical allodynia can occur in those with more severe injury when compared to those with mild injuries (Kloos 2005,

95 Hoschouer 2010). VFA can be further used to determine whether dogs with SCI also experience mechanical allodynia and/or neuropathic pain. However, such conditions typically occur minimally several weeks following SCI. Therefore, future studies evaluating dogs with SCI for the development of chronic painful conditions will likely need longer follow-up times than the 30 days employed for the purposes of this study.

For the purposes of this study, only SL and BS footprint parameters were determined in normal dogs and dogs with SCI. However, other footprint parameters using the same technique should be studied such as foot angle, gait patterns, thoracic limb and pelvic limb coordination, print area, etc. Such information may further reveal valuable outcome measures that may be used in SCI-affected dogs, as well as reveal breed and conformation differences and their effects on normal and abnormal gait. Additionally, although footprint analysis in thoracolumbar SCI has focused primarily on the pelvic limbs, this study has shown the important role of the thoracic limbs during recovery.

Therefore, the measurements of footprint parameters even in the lower functioning dogs

(i.e. paraplegic) may be useful to study the compensation of the trunk and thoracic limbs seen thoracolumbar SCI.

Studies in rat models of SCI have shown strain related differences between breed, age and weight in locomotor and sensory recovery from SCI (Sedý 2008). It is reasonable to assume that such differences may exist between breeds in dogs with SCI. It is also likely that pelvic limb and thoracic limb footprint parameters vary depending on

96 lesion severity (Cheng 1997, Collazos-Castro 2006). In this study, dogs with varying severity of SCI were compared to normal dogs. Further segregation of dogs with similar degrees of SCI to evaluated VFA ST and footprint parameters for each degree of SCI (i.e. mild, moderate, severe) may reveal further differences from normal dogs and changes with functional recovery. Ideally, future clinical trials should employ breed, age and weight matched controls in order to minimize such differences.

A power analysis was performed prior to the initiation of the study in order to determine the adequate number of dogs to enroll into each group. The analysis revealed that 20 dogs in the normal group would provide 80% power to detect an intraclass correlation of 0.5 (minimal) using an F-test when using 3 observations (3 separate time points of measurement) per dog. The analysis further revealed that with 30 SCI-affected dogs, a power of 80% can be obtained to detect a correlation of 0.55 between the OSCIS,

VFA or footprint analysis at each of the three time points. Therefore, 20 normal dogs and

30 SCI-affected dogs were enrolled. However, larger, multi-center studies may be able to reduce any possible type I or type II errors that may have been present in this study, as well as improve on correlations between each behavioral response test.

References

Aghighi SA, Tipold A, Piechotta M, Lewczuk P, Kästner SB. Assessment of the effects of adjunctive gabapentin on postoperative pain after intervertebral disc surgery in dogs. Vet Anaesth Analg 2012;39:636-646.

Aikawa T, Fujita H, Kanazono S, et al. Long-term neurologic outcome of hemilaminectomy and disk fenestration for treatment of dogs with thoracolumbar intervertebral disk herniation: 831 cases (2000-2007). J Am Vet Med Assoc 2012i; 241:1617-1625.

Aikawa T, Fujita H, Shibata M, et al. Recurrent thoracolumbar intervertebral disc extrusion after hemilaminectomy and concomitant prophylactic fenestration in 662 chondrodystrophic dogs. Vet Surg 2012ii;41:381-390.

Assina R, Sankar T, Theodore N, et al. Activated autologous macrophage implantation in a large-animal model of spinal cord injury. Neurosurg Focus 2008;25:E3.

Atkinson PP, Atkinson JLD. Spinal shock. Mayo Clin Proc 1996;71:384-389.

Ballermann M, Tse ADY, Misiaszek JE, et al. Adaptations in the walking pattern of spinal cord injured rats. J Neurotrauma 2006;23:897-907.

Baltzer WI, McMichael MA, Hosgood GL, et al. Randomized, blinded, placebo- controlled clinical trial of N-acetylcysteine in dogs with spinal cord trauma from acute intervertebral disc disease. Spine 2008;33:1397-1402.

Barnett 2004. Olfactory ensheathing cells: unique glial cell types. J Neurotrauma 2004;21:375-382.

Basso DM, Beattie MS, Bresnahan JC. A sensitive and reliable locomotor rating scale for open field testing in rats. J Neurotrauma 1995;12:1-21.

Basso DM, Beattie MS, Bresnahan JC, et al. MASCIS Evaluation of open field locomotor scores: effects of experience and teamwork on reliability. J Neurotrauma 1996;13:343-359.

Basso DM. Behavioral testing after spinal cord injury-congruities, complexities, and controversies. J Neurotrauma 2004;21:395-404.

98 Basso DM, Fisher LC, Anderson AJ, et al. Basso mouse scale for locomotion detects differences in recovery after spinal cord injury in five common mouse strains. J Neurotrauma 2006;23:635-659.

Behrmann DL, Bresnahan JC, Beattie MS, et al. Spinal cord injury produced by consistent mechanical displacement of the cord in rats: behavioral and histologic analysis. J Neurotrauma 1992;9:197-217.

Bellei E, Roncada P, Pisoni L, Joechler M, Zaghini A. The use of fentanyl-patch in dogs undergoing spinal surgery: plasma concentration and analgesic efficacy. J Vet Pharmacol Therap 2011;34:437-411.

Bergknut N, Egenvall A, Hagman R, et al. Incidence of intervertebral disk degeneration- related diseases and associated mortality rates in dogs. J Am Vet Med Assoc 2012;240:1300-1309.

Blaser A, Lang J, Henke D, et al. Influence of durotomy on laser-doppler measurement of spinal cord blood flow in chondrodystrophic dogs with thoracolumbar disk extrusion. Vet Surg 2012;41:221-227.

Blight AR, Toombs JP, Bauer MS, et al. The effects of 4-aminopyridine on neurological deficits in chronic cases of traumatic spinal cord injury in dogs: a phase I clinical trial. J Neurotrauma 1991;8:103-119.

Boag AK, Otto CM, Drobatz KJ. Complications of methylprednisolone sodium succinate therapy in dachshunds with surgically treated intervertebral disc disease. J Vet Emerg Crit Care 2001;11:105-110.

Boakye M, Harkema S, Ellaway PH, et al. Quantitative testing in spinal cord injury: overview of reliability and predictive validity. J Neurosurg Spine (Suppl) 2012;17:141- 150.

Borgens RB, Toombs JP, Breur G, et al. An imposed oscillating electrical field improves the recovery of function in neurologically complete paraplegic dogs. J Neurotrauma 1999;16:639-657.

Bracken MB, Shepard MJ, Holford TR, et al. Administration of methylprednisolone for 24 or 48 hours or tirilazad mesylate for 48 hours in the treatment of acute spinal cord injury. Results of the third national acute spinal cord injury randomized controlled trial. National acute spinal cord injury study. JAMA 1997;277:1597-1604.

Bregman BS, Kunkel-Bagden E, Schnell L, et al. Recovery from spinal cord injury mediated by antiboides to neurite growth inhibitors. Nature 1995;378:498-501.

Briley JD, Williams MD, Freire M, et al. Feasability and repeatability of cold and mechanical quantitative sensory testing in normal dog. Vet J 2014;199:245-250.

Brisson BA. Intervertebral disc disease in dogs. Vet Clin Small Anim 2010;40:829-858.

Brydges NM, Argyle DJ, Mosley JR, et al. Clinical assessments of increased sensory sensitivity in dogs with cranial cruciate ligament rupture. Vet J 2012;193:545-550.

Carlton SM, Junhui D, Tan HY, et al. Peripheral and central sensitization in remote spinal cord regions contribute to central neuropathic pain after spinal cord injury. Pain 2009;147:265-276.

Casha S, Zygun D, McGowan MD, et al. Results of a phase II placebo-controlled randomized trial of minocycline in acute spinal cord injury. Brain 2012;135:1224-1236.

Cheng H, Almström S, Giménex-Llort L, et al. Gait Analysis of adult paraplegic rags after spinal cord repair. Exp Neurology 1997;148:544-557.

Christensen MD, Hulsebosch CE. Chronic central pain after spinal cord injury. J Neurotrauma 1997;14:517-537.

Collazos-Castro JE, López-Dolado E, Nieto-Sampedro M. Locomotor deficits and adaptive mechanisms after thoracic spinal cord contusion in the adult rat. J Neurotrauma 2006;23:1-17.

Davis GJ, Brown DC. Prognostic indicators for time to ambulation after surgical decompression in nonambulatory dogs with acute thoracolumbar disk extrusions: 112 cases. Vet Surg 2002;31:513-518.

de Lahunta A, Glass E. General sensory systems: General proprioception and general somatic afferent. In: Veterinary neuroanatomy and clinical neurology. 3rd edn. Saunders, 2009:221-242. de Medinaceli L, Freed WJ, Wyatt JR. An Index of the functional condition of rat sciatic nerve based on measurements made from walking tracks. Exp Neurol 1982;77:634-643.

Densmore VS, Kalous A, Keast JR, et al. Above-level mechanical hyperalgesia in rats develops after incomplete spinal cord injury but not after cord transection, and is reversed by amitriptyline, morphine and gabapentin. Pain 2010;151:184-193.

100 Detloff MR, Clark LM, Hutchinson KJ, et al. Validity of acute and chronic tactile sensory testing after spinal cord injury in rats. Exp Neurol 2010;225:366-376.

Detloff MR, Fisher LC, Deibert RJ, et al. Acute and chronic tactile sensory testing after spinal cord injury in rats. J Vis Exp 2012;62:e3247.

Ditunno J, Scivoletto G. Clinical relevance of gait research applied to clinical trials in spinal cord injury. Brain Res Bull 2009;78:35-42.

Ditunno JF Jr, Ditunno PL, Scivoletto G, et al. The walking index for spinal cord injury (WISCI/WISCI II): nature, metric properties, use and misuse. Spinal Cord 2013;51:346- 355.

Duval J, Dewey C, Roberts R, et al. Spinal cord swelling as a myelographic indicator of prognosis: a retrospective study in dogs with intervertebral disc disease and loss of deep pain sensation. Vet Surg 1996;25:6-12.

Emery E, Aldana P, Bunge MB, et al. Apoptosis after traumatic human spinal cord injury. J Neurosurg 1998;89:911-920.

Fehlings MG, Baptiste DC. Current status of clinical trials for acute spinal cord injury. Injury 2005; 36: SB113-SB122.

Felix ER, Widerstrom-Noga EG. Reliability and validity of quantitative sensory testing in persons with spinal cord injury and neuropathic pain. J Rehabil Res Dev 2009;46:69- 84.

Ferraro GB, Alabed YZ, Fournier AE. Molecular targets to promote central nervous system regeneration. Curr Neurovasc Res 2004;1:61-75.

Ferreira AJA, Correia JHD, Jaggy A. Thoracolumbar disc disease in 71 paraplegic dogs: influence of rate of onset and duration of clinical signs on treatment results. J Small Anim Pract 2002;43:158-163.

Foss K, da Costa RC, Rajal-Shultz PJ, et al. Force plate gait analysis in Doberman pinschers with and without cervical spondylomyelopathy. J Vet Intern Med 2013;27:106- 111.

Fouad K, Pedersen V, Schwab ME, et al. Cervical sprouting of corticospinal fibers after thoracic spinal cord injury accompanies shifts in evoked motor responses. Curr Biol 2001;11:1766-1770.

Fukuda S, Nakamura T, Kishigami Y, et al. New canine spinal cord injury model free from laminectomy. Brain Res Brain Res Protoc 2005; 14:171-180.

101

Garbossa D, Boido M, Fontanella M, et al. Recent therapeutic strategies for spinal cord injury treatment: possible role of stem cells. Neursurg Rev 2012;35:293-311.

Gensel JC, Donnelly DJ, Popovich PG. Spinal cord injury therapies in humans: an overview of current clinical trials and their potential effects on intrinsic CNS macrophages. Expert Opin Ther Targets 2011;15:505-518.

Gerzanich V, Woo SK, Vennekens R, et al. De novo expression of Trpm4 initiates secondary hemorrhage in spinal cord injury. Nat Med 2009;15:185-191.

Goldberger ME, Bregman BS, Vierck CJ, et al. Criteria for assessing recovery of function after spinal cord injury: behavioral methods. Exp Neurol 1990;107:113-117.

Goldberger ME. The use of behavioral methods to predict spinal cord plasticity. Restor Neurol Neurosci 1991;2:339-350.

Gordon-Evans WJ, Evans RB, Conzemius MG. Accuracy of spatiotemporal variables in gait analysis of neurologic dogs. J Neurotrauma 2009;26:1055-1060.

Gorio A, Gokmen N, Erbayraktar S, et al. Recombinant human erythropoietin counteracts secondary injury and markedly enhances neurological recovery from experimental spinal cord trauma. Proc Natl Acad Sci 2002;99:9450-9455.

Granger N, Blamires H, Franklin RJM, et al. Autologous olfactory mucosal cell transplants in clinical spinal cord injury: a randomized double-blinded trial in a canine translational model. Brain 2012;135:3227-3237.

Grossman RG, Fehlings MG, Frankowski RF, et al. A prospective, multicenter, phase I matched-comparison group trial of safety, pharmacokinetics, and preliminary efficacy of riulzole in patients with traumatic spinal cord injury. J Neurotrauma 2014;31:239-355.

Gutierrez-Quintana R, Edgar J, Wessmann A, et al. The cutaneous trunci reflex for localizing and grading thoracolumbar spinal cord injures in dogs. J Small Anim Pract 2012;53:470-475.

Hamers FPT, Lankhorst AJ, Van Laar TJ, et al. Automated quantitative gait analysis during overground locomotion in the rat: its application to spinal cord contusion and transection injuries. J Neurotrauma 2001;18:187-201.

Hamers FPT, Kiipmans GC, Joosten EAJ. Catwalk-assisted gait analysis in the assessment of spinal cord injury. J Neurotrauma 2006;23:537-548.

102 Hamilton L, Franklin RJ, Jeffery ND. Quantification of deficits in lateral paw positioning after spinal cord injury in dogs. BMC Vet Res 2008;4:47.

Handa Y, Naito A, Watanabe S, et al. Functional recovery of locomotive behavior in the adult spinal dog. Tohoku J Exp Med 1986;148:373-384.

Hayes KC, Wolfe DL, Hsieh JT, et al. Clinical and electrophysiologic correlates of quantitative sensory testing in patients with incomplete spinal cord injury. Arch Phys Med Rehabil 2002;83:1612-1619.

Henke D, Vandevelde M, Doherr MG, et al. Correlations between severity of clinical signs and histopathological changes in 60 dogs with spinal cord injury associated with acute thoracolumbar intervertebral disc disease. Vet J 2013;198:70-75.

Hoschouer EL, Basso DM, Jakeman LB. Aberrant sensory responses are dependent on lesion severity after spinal cord contusion injury in mice. Pain 2010;148:328-342.

Hurlbert RJ, Hadley MN, Walters BC, et al. Pharmacological therapy for acute spinal cord injury. Neurosurgery 2013;72:93-S2105.

Ito D, Matsunaga S, Jeffery ND, et al. Prognostic value of magnetic resonance imaging in dogs with paraplegia caused by thoracolumbar intervertebral disk extresio: 77 cases (2000-2003). J Am Vet Med Assoc 2005;227:1454-1460.

Jeffery ND, Blakemore WF. Spinal cord injury in small animals: 1. Mechanisms of spontaneous recovery. Vet Rec 1999i;144:407-413.

Jeffery ND, Blakemore WF. Spinal cord injury in small animals: 2. Current and future options for therapy. Vet Rec 1999ii;145:183-190.

Jeffery ND, Lakatos A, Franklin RJM. Autologous olfactory glial cell transplantation is reliable and safe in naturally occurring canine spinal cord injury. J Neurotrauma 2005;22:1282-1293.

Jeffery ND, Smith PM, Lakatos A, et al. Clinical canine spinal cord injury provides an opportunity to examine the issues in translating laboratory techniques into practical therapy. Spinal Cord 2006;44:584-593.

Jeffery ND, Hamilton L, Granger N. Designing clinical trials in canine spinal cord injury as a model to translate successful laboratory interventions into clinical practice. Vet Rec 2011;168:102-107.

Kanagal SG, Muir GD. The differential effects of cervical and thoracic dorsal funiculus lesions in rats. Behav Brain Res 2008;187:379-386.

103

Keirstead HS, Nitor G, Bernal, et al. Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remyelinate and restore locomotion after spinal cord injury. J Neurosci 2005;25:4694-4705.

Kinzel S, Wolff M, Buecker A, et al. Partial percutaneous discectomy for treatment of thoracolumbar disc protrusion: retrospective study of 331 dogs. J Sm Anim Pract 2005;46:479-484.

Klapdor K, Dulfer BG, Hammann A, et al. A low-cost method to analyse footprint patterns. J Neurosci Methods 1997;75:49-54.

Kloos AD, Fisher LC, Detloff MR, et al. Stepwise motor and all-or-none sensory recovery is associated with nonlinear sparing after incremental spinal cord injury in rats. Exp Neurol 2005;191:251-265.

Kögel B, Terlinden R, Schneider J. Characterisation of tramadol, morphine and tapentadol in an acute pain model in Beagle dogs. Vet Anaesth Analg 2014;41:297-304.

Koopmans GC, Deumens R, Honig WMM, et al. The assessment of locomotor function in spinal cord injured rats: the importance of objective analysis coordination. J Neurotrauma 2005:22;214-225.

KuKanich B, Papich MG. Pharmacokinetics and antinociceptive effects of oral tramadol hydrochloride administration in Greyhounds. Am J Vet Res 2011;72:256-262.

Kunkel-Bagden E, Bregman BS. Spinal cord transplants enhance the revoery of locomotor function after spinal cord injury at birth. Exp Brain Res 1990;81:25-34.

Kunkel-bagden E, Dai HN, Bregman BS. Methods to assess the development and recovery of locomotor function after spinal cord injury in rats. Exp Neurol 1993;119:153-164.

Krassioukov AV, Furlan JC, Fehlings MG. Autonomic dysreflexia in acute spinal cord injury: an under-recognized clinical entity. J Neurotrauma 2003;20:707-716.

104 Kwon BK, Okon EB, Tsai E, et al. A grading system to evaluate objectively the strength of pre-clinical data of acute neuroprotective therapies for clinical translation in spinal cord injury. J Neurotrauma 2011;28:1525-15543.

Laitinen OM, Puerto DA. Surgical decompression in dogs with thoracolumbar intervertebral disc disease and loss of deep pain perception: a retrospective study of 46 cases. Acta Vet Scan 2005;46:79-85.

Lambert GA, Mallos G, Zagami AS. Von Frey’s hairs – a review of their technology and use – a novel automated von Frey device for improved testing for hyperalgesia. 2009;177:420-426,

Lammertse DP. Clinical trials in spinal cord injury: lessons learned on the path to translation. The 2011 International Spinal Cord Society Sir Ludwig Guttmann lecture. Spinal Cord 2013;51:2-9.

Lascelles BD, Waterman AE, Cripps PJ, Livingston A, Henderson G. Central sensitization as a result of surgical pain: investigation of the pre-emptive value of pethidine for ovariohysterectomy in the rat. Pain 1995;62:201-212.

Lascelles BD, Cripps PJ, Jones A, Waterman AE. Post-operative central hypersensitivity and pain: the pre-emptive value of pethidine for ovariohysterectomy. Pain 1997;73:461- 471.

Lascelles BD, Cripps PJ, Jones A, et al. Efficacy and kinetics of carprofen, administered preoperatively or postoperatively, for the prevention of pain in dogs undergoing ovariohysterectomy. Vet Surg 1998;27:568-82.

Lascelles BD. Getting a sense of sensations. Vet J 2013;197:115-117.

Laverty PH, Leskovar A, Breur GJ, et al. A preliminary study of intravenous surfactants in paraplegic dogs: polymer therapy in canine clinical SCI. J Neurotrauma 2004;21:1767-1777.

Lee SM, Yune TY, Kim SJ, et al. Minocycline reduces cell death and improves functional recovery after traumatic spinal cord injury in the rat. J Neurotrauma 2003;10:1017-27.

Lee SM, Rosen S, Weinstein P, et al. Prevention of both neutrophil and monocyte recruitment promotes recovery after spinal cord injury. J Neurotrauma 2011;28:1893- 1907.

Levine GJ, Levine JM, Budke CM, et al. Description and repeatability of a newly developed spinal cord injury scale for dog. Prev Vet Med 2009;89:121-127.

105

Levine GJ, Levine JM Witsberger TH, et al. Cerebrospinal fluid myelin basic protein as a prognostic biomarker in dogs with thoracolumbar intervertebral disk herniation. J Vet Intern Med 2010;24:890-896.

Levine JM, Levine GJ, Kerwin SC, et al. Association between various physical factors and acute thoracolumbar intervertebral disk extrusion or protrusion in Dachshunds. J Am Vet Med Assoc 2006i;229:370-375.

Levine JM, Ruaux CG, Bergman RL, et al. Matrix metalloproteinase-9 activity in the cerebrospinal fluid and serum of dogs with acute spinal cord trauma from intervertebral disk disease. Am J Vet Res 2006ii;67:283-287.

Levine JM, Levine GJ, Boozer L, et al. Adverse effects and outcome associated with dexamethasone administration in dogs with acute thoracolumbar intervertebral disk herniation: 161 cases (2000-2006). J Am Vet Med Assoc 2008;232:411-417.

Levine JM, Levine GJ, Porter BF, et al. Naturally occurring disk herniation in dogs: an opportunity for pre-clinical spinal cord injury research. J Neurotrauma 2011;28:675-688.

Levine JM, Cohen ND, Heller M, et al. Efficacy of a metalloproteinase inhibitor in spinal cord injured dogs. PLoS One 2014;9:e96408.

Lim JH, Byeon YE, Ryu HH, et al. Transplantation of canine umbilical cord blood- derived mesenchymal stem cells in experimentally induced spinal cord injured dogs. J Vet Sci 2007;8:275-282.

Lim JH, Muguet-Chanoit AC, Smith DT, et al. Potassium channel antagonists 4- aminopyridine and the T-butyl carbamate derivative of 4-aminopyridine improve hind limb function in chronically non-ambulatory dogs; a blinded, placebo-controlled trial. PLoS One 2014;9:e116139.

Lindsey AE, LoVerso RL, Tovar A, et al. An analysis of changes in sensory thresholds to mild tactile and cold stimuli after experimental spinal cord injury in the rat. Neurorehabil Neural Repair 2000;14:287-300.

Loughin CA, Dewey CW, Ringwood PB, et al. Effect of durotomy on functional outcome of dogs with type I thoracolumbar disc extrusion and absent deep pain perception. Vet Comp Orthop Traumatol 2005;18:141-146.

Lu J, Ashwell KWS, Waite P. Advances in secondary spinal cord injury. Spine 2000;25:1859-1866.

106 Matis GK, Birbilis TA. Erythropoietin in spinal cord injury. Eur Spine J 2009;18:314- 323.

McDonald JW, Becker D, Sadowsky CL, et al. Late recovery following spinal cord injury. J Neurosurg (Spine 2) 2002;97:252-265.

McEwen ML, Springer JE. Quantification of locomotor recovery following spinal cord contusion in adult rats. J Neurotrauma 2006;23:1632-1653.

Metz GA, Merkler D, Dietz V, et al. Efficient testing of motor function in spinal cord injured rats. Brain Res 2000;883:165-177.

Mogensen MS, Scheibye-Alsing K, Karlskov-Mortensen P, et al. Validation of genome- wide intervertebral disk calcification associations in Dachshund and further investigation of the chromosome 12 susceptibility locus. Front Genet 2012;3:doi: 10.3389/fgene.2012.00225

Mogil JS. Animal models of pain: progress and challenges. Nat Rev Neurosci 2009;10:283-294.

Moore SA, Hettlich BF, Waln A. The use of an electronic von Frey device for evaluation of sensory threshold in neurologically normal dogs and those with acute spinal cord injury. Vet J 2013;197:216-219.

Moore RW, Withrow SJ. Gastrointestinal hemorrhage and pancreatitis associated with intervertebral disk disease in the dog. J Am Vet Med Assoc 1982;180:1443-1447.

Muguet-Chanoit AC, Olby NJ, Lim J, et al. The cutaneous trunci muscle reflex: a predictor of recovery in dogs with acute thoracolumbar myelopathies caused by intervertebral disc extrusions. Vet Surg 2012;41:200-206.

Muir GD, Webb AA. Mini-review: assessment of behavioural recovery following spinal cord injury in rats. Eur J Neurosci 2000;12:3079-3086.

Nagoshi N, Fehlings, MG. Investigational drugs for the treatment of spinal cord injury: review of preclinical studies and evaluation of clinical trials from Phase I to II. Expert Opin Investig Drugs. 2015;1-14.

Naito A, Shimizu Y, Handa Y. Analyses of airstepping movement in adult spinal dogs. Tohoku J Exp Med 1990i;162:41-48.

107

Naito A, Shimizu Y, Handa Y. Analyses of treadmill locomotion in adult spinal dogs. Neurosci Res 1990ii;4:281-290.

National spinal cord injury statistical center (NSCISC). Spinal cord injury facts and figures at a glance February 2014. (cited February 21, 2015). Available from: https://www.nscisc.uab.edu/PublicDocuments/fact_figures_docs/Facts%202014.pdf

North Carolina State University, College of Veterinary Medicine. Canine spinal cord injury program. (cited February 21, 2015). Available from: http://www.cvm.ncsu.edu/docs/scip/ct.html

Olby N. Current concepts in the management of acute spinal cord injury. J Vet Intern Med 1999i;13:399-407.

Olby NJ, Sharp NJH, Munana KR, et al. Chronic and acute compressive spinal cord lesions in dogs due to intervertebral disc herniation are associated with elevation in lumbar cerebrospinal fluid glutamate concentration. J Neurotrauma 1999ii;16:1215- 1224.

Olby NJ, De Risio L, Munana KR, et al. Development of a functional scoring system in dogs with acute spinal cord injuries. Am J Vet Res 2001;62:1624-1628.

Olby N, Levine J, Harris T, et al. Long-term functional outcome of dogs with severe injuries of the thoracolumbar spinal cord: 87 cases (1996-2001). J Am Vet Med Assoc 2003;222:762-769.

Olby N, Harris T, Burr J, et al. Recovery of pelvic limb function in dogs following acute intervertebral disc herniations. J Neurotrauma 2004;21:49-59.

Olby N. The pathogenesis and treatment of acute spinal cord injuries. Vet Clin Small Anim 2010;40:791-807.

Oza CS, Giszter SF. Plasticity and alterations of trunk motor cortex following spinal cord injury and non-stepping robot and treadmill training. Exp Neurology 2014;256:57-69.

Packer RMA, Hendricks A, Volk HA, et al. How long and low can you go? Effect of conformation on the risk of thoracolumbar intervertebral disc extrusion in domestic dogs. PLoS One 2013;8: 1-11.

Park EH, White GA, Tieber LM. Mechanisms of injury and emergency care of acute spinal cord injury in dogs and cats. J Vet Emerg Crit Care 2012;22:160-178.

108 Park SS, Le YJ, Lee SH, et al. Functional recovery after spinal cord injury in dogs treated with combination of Matrigel and neural-induced adipose-derived mesenchymal stem cells. Cytotherapy 2012;14:584-597.

Plemel JR, Duncan G, Chen KW, et al. A graded forceps crush spinal cord injury model in mice. J Neurotrauma 2008;25:350-370.

Rangasamy SB. Locomotor recovery after spinal cord hemisection/contusion injuries in Bonnet Monkeys: footprint testing-a minireview. Synapse;2013:427-453.

Rice ASC, Cimino-Brown D, Eisenach JC, et al. Animal models and the prediction of efficacy in clinical trials of analgesic drugs: A critical appraisal and call for uniform reporting standards. Pain 2009;139:243-247.

Roerig A, Carlson R, Tipold A, et al. Cerebrospinal fluid tau protein as a biomarker for severity of spinal cord injury in dogs with intervertebral disc herniation. Vet J 2013;197:253-258.

Ruddle TL, Allen DA, Schertel ER, et al. Outcome and prognostic factors in non- ambulatory Hansen Type I intervertebral disc extrusions: 308 cases. Vet Comp Orthop Traumatol 2006;19:29-34.

Ryu HH, Lim JH, Byeon YE, et al. Functional recovery and neural differentiation after transplantation of allogenic adipose-derived stem cells in a canine model of acute spinal cord injury. J Vet Sci 2009;10:273-284.

Ryu HH, Kang BJ, Park SS, et al. Comparison of mesenchymal stem cells derived from fat, bone marrow, Wharton’s jelly, and umbilical cord blood for treating spinal cord injuries in dogs. J Vet Med Sci 2012;74:1617-1630.

Scott HW, McKee WM. Laminectomy for 34 dogs with thoracolumbar intervertebral disc disease and loss of deep pain perception J Small Anim Pract 1999;40:417-422.

Schucht P, Raineteau O, Schwab ME, et al. Anatomical correlates of locomotor recovery following dorsal and ventral lesion of the rat spinal cord. Exp Neuro 2002;176:143-153.

Sedý J, Urdzíková L, Jendelová P, et al. Methods for behavioral testing of spinal cord injured rats. Neurosci Biobehav Rev 2008;32:550-580.

Shimizu J, Mochida K, Kobayashi Y, et al. Inflammatory reaction in the herniated degenerative disc materials in miniature dachshunds. J Vet Med Sci 2010;72:81-84.

Shimamura S, Kanayama K, Himada T, et al. Evaluation of the function of polymorphonuclear neutrophilic leukocytes in healthy dogs given a high dose of

109 methylprednisolone sodium succinate. Am J Vet Res 2010;71:541-546.

Short DJ, El Masry WS, Jones PW. High dose methylprednisolone in the management of acute spinal cord injury – a systematic review from a clinical perspective. Spinal Cord 2000;38:273-286.

Shy ME, Frohman EM, Arezzo JC, et al. Quantitative sensory testing: Report of the therapeutics and technology assessment subcommittee of the American academy of neurology. Neurology 2003;60:898-904.

Smith PM, Jeffery ND. Spinal shock – Comparative aspects and clinical relevance. J Vet Intern Med 2005;19:788-793.

Sonmez E, Kabatas S, Ozen O, et al. Minocycline treatment inhibits lipid peroxidation, preserves spinal cord ultrastructure, and improves functional outcome after traumatic spinal cord injury in the rat. Spine (Phila Pa 1976) 2013;38:1253-1259.

Stokes BT, Reier PJ. Fetal grafts alter chronic behavioral outcome after contusion damage to the adult rat spinal cord. Exp Neuro 1992;116:1-12.

Srugo I, Aroch I, Christopher MM, et al. Association of cerebrospinal fluid analysis findings with clinical signs and outcome in acute nonambulatory thoracolumbar disc disease in dogs. J Vet Intern Med 2011;25:846-855.

Steward O, Popovich PG, Dietrich WD, et al. Replication and reproducibility in spinal cord injury research. Exp Neurol 2012;233:597-605.

Sylvestre AM, Cockshutt JR, Parent JM, et al. Magnetic motor evoked potentials for assessing spinal cord integrity in dogs with intervertebral disc disease. Vet Surg 1993;22:5-10.

Tena B, Escobar B, Arguis MJ, et al. Reproducibilily of electronic von Frey and von Frey monofilaments testing. Clin J Pain 2012;28:318-323.

Toombs JP, Caywood DD, Lipowitz AJ, et al. Colonic perforation following neurosurgical procedures and corticosteroid therapy in four dogs. J Am Vet Med Assoc 1980;222:68-72.

Van Wie EY, Fosgate GT, Mankin JM, et al. Prospectively recorded versus medical record-derived spinal cord injury scores in dogs with intervertebral disk herniations. J Vet Intern Med 2013;27:1273-1277.

110 Varejao ASP, Cabrita AM, Meek MF, et al. Functional and morphological assessment of a standardized rat sciatic nerve crush injury with a non-serrated clamp. J Neurotrauma 2004;21:1652-1670.

Verbeke G, Molenberghs G. Linear Mixed Models for Longitudinal Data, First Ed. Springer, Verbag, NY, 2000.

Vierck CJ, Hansson PT, Yezierski RP. Clinical and pre-clinical pain assessment: Are we measuring the same thing? Pain 2008;135:7-10

Vrinten DH, Hamers FFT. ‘CatWalk’ automated quantitative gait analysis as a novel method to assess mechanical allodynia in the rat; a comparison with von Frey testing. Pain 2003;102:203-209.

Walk D, Sehgal N, Moeller-Bertram T, et al. Quantitative sensory testing and mapping – A review of nonautomated quantitative methods for examination of the patient with neuropathic pain. Clin J Pain 2009;25:632-640.

Wang XM, Basso DM, Terman JR, et al. Adult opossums (Didelphis virginiana) demonstrate near normal locomotion after spinal cord transection as neonates. Exp Neurol 1998;151:50-69.

Webb AA, Muir GD. Compensatory locomotor adjustments of rats with cervical or thoracic spinal cord hemisections. J Neurotrauma 2002;19:239-256.

Webb AA, Jeffery ND, Olby NJ, et al. Behavioural analysis of the efficacy of treatments for injuries to the spinal cord in animals. Vet Rec 2004;155:225-230.

Webb AA, Ngan S, Fowler. Spinal cord injury 1: a synopsis of the basic science. Can Vet J 2010;51:485-492.

Wegner K, Horais KA, Tozier NA, et al. Development of a canine nociceptive thermal escape model. J Neurosci Methods 2008;168:88-97.

Wu Y, Satkunendrarajah K, Teng Y, et al. Delayed post-injury administration of Riluzole is neuroprotective in a preclinical rodent model of cervical spinal cord injury. J Neurotrauma 2013;30:441-452.

Yagüe JG, Humanes-Valera D, Aguilar J, et al. Functional reorganization of the forepaw cortical representation immediately after thoracic spinal cord hemisection in rats. Exp Neurol 2014;257:19-24.

Zakeri M, Afshari K, Gharedaghi MH, et al. Lithium protects against spinal cord injury in rats: role of nitric oxide. J Neurol Surg A Cent Eur Neurosurg 2014;75:427-433.

111

Zhang H, Chang M, Hansen CN, et al. Role of matrix metalloproteinases and therapeutic benefits of their inhibition in spinal cord injury. Neurotherapeutics 2011;8:206-220.

112

Appendix: Summary of Data

113 Table 10 Summary data for 20 normal small breed dogs

Breed Gender Age (years) Weight (kg) Beagle FS 4.5 10.2 Mixed Breed FS 2.5 12.6 Bichon FS 2.5 4.92 Corgi MC 6 16 Dachshund FS 2 3.72 Dachshund FS 6.5 4.93 Dachshund MC 1.5 6.1 Dachshund MC 0.7 6 Mixed Breed MC 1.5 12.6 Miniature Pinscher FS 4.5 6.7 Mixed Breed FS 2 7.7 Mixed Breed MC 2 14.7 Mixed Breed MC 5 17.2 Schnauzer MC 5 8.5 Schnauzer FS 4 7 Sealyham Terrier MC 3 11.9 Sealyham Terrier MC 3 12.1 Shih Tzu MC 2.5 3.75 Mixed Breed MC 4.5 13 Cocker Spaniel MC 4.5 12.8 Median 3 9.35 FS, female spayed; MC, male castrated

114 Table 11 Summary data for 30 dogs with acute, spontaneous spinal cord injury due to intervertebral disc extrusion

Age Weight Gender Breed Surgery site (years) (kg)

FS Mixed Breed 8 17 R T12-13 (h/f)

FS Mixed Breed 5 3.9 R T12-13 (h)

L L3-L4 (h), L L3-L4/L1-2 FS Dachshund 5 5.6 (f)

Pembroke welsh MC 5.5 17 R T12-13 (h/f) corgi

MC Standard Dachshund 7 9.8 L L1-2 (h)

FS French Bulldog 3 8.8 L T13-L1 (h), T12-L3 (f)

FS Beagle 11 16 R T13-L1, L1-L2 (h/f)

M Mixed Breed 3 4 R L1-L2 (h)

FS Mixed Breed 6 14.6 R T13-L1 (h)

FS Shih Tzu 5 5.6 L T12-13 (h)

Continued

115 Table 11 Continued

MC French Bulldog 5 14.8 L L3-4 (h)

MC Cocker Spaniel 5 12.7 R L2-3 (h)

MC Cockapoo 3 7.6 L L2-3 (h)

MC Shih Tzu 4.5 6.3 R T10-13 (h)

MC Dachshund 5 9.8 L T13-L2 (h)

FS French Bulldog 5 10.8 L L3-4 (h)

L T12-13 (h), T11-12/ T12- FS Dachshund 6 6 13/T13-L1 (f)

L T12-13 (h), T12-13/ T13- MC Dachshund 4 7.5 L1 (f)

MC Dachshund 6 6.5 L T11-12 (h)

L L2-3 (h), T13-L1/L1-2/L2- M Dachshund 3 5.4 3 (f)

L T13-L1 (h), T12-13/ T13- FS Dachshund 5 13.7 L1/L1-2 (f)

Continued

116 Table 11 Continued

MC French Bulldog 4 10.8 R T13-L3 (h)

FS Beagle 2 10.2 R T11-L2 (h)

FS Dachshund 3.5 4.5 R T11-12 (h)

FS Mixed Breed 5.5 11 R T13-L1 (h), L T12-13 (p)

FS Dachshund 12 5.2 R T13-L1 (h)

MC Dachshund 8 5.6 R L3-4 (h), L2-3 (f)

MC Dachshund 7 6.4 R T13-L1 (h)

Pembroke welsh MC 6.5 12.8 L T11-12 (h), T11-12 (f) corgi

FS Dachshund 3 5.8 R T13-L1(h), T13-L1 (f)

Median 5 8.2

FS, female spayed; MC, male castrated; M, male intact; h, hemilaminectomy; f, fenestration; p, pediculectomy

117 Table 12 von Frey anesthesiometry data for normal dogs (n=20)

Session Session Session Dog Limb Session 1 Session 2 Session 3 1 Mean 2 Mean 3 Mean

RHL 88.3 96.8 92.0 92.4 36.7 62.7 76.2 58.5 49.0 31.2 45.5 41.9 LHL 200.3 155.5 128.9 161.6 57.9 45.6 75.7 59.7 64.6 55.9 63.6 61.4 1 RFL 108.7 46.5 86.6 80.6 83.6 50.9 80.4 71.6 82.7 110.8 60.8 84.8 LFL 85.2 144.2 104.8 111.4 141.4 29.6 115.3 95.4 57.6 67.9 55.6 60.4 RHL 218.3 233.1 140.6 197.3 136.4 125.3 119.5 127.1 70.7 76.7 94.6 80.7 LHL 201.2 242.1 169.6 204.3 105.2 114.7 88.2 102.7 72.9 55.1 67.2 65.1 2 RFL 121.0 144.8 104.0 123.3 106.9 138.8 115.0 120.2 115.6 112.8 108.2 112.2

118 LFL 316.3 276.2 215.8 269.4 153.7 147.9 98.2 133.3 123.3 121.2 110.6 118.4

3 5

RHL 258.2 265.3 270.3 264.6 240.7 176.6 201.4 206.2 146.6 146.2 149.4 147.4

LHL 223.9 176.7 216.4 205.7 155.5 220.6 178.7 184.9 83.5 130.2 121.7 111.8

RFL 74.2 148.2 79.9 100.8 154.4 122.2 194.7 157.1 112.8 83.1 76.7 90.9 3 RHL 192.6 186.8 152.5 177.3 137.1 139.0 127.1 134.4 63.6 107.7 87.6 86.3 RHL 190.5 160.2 183.7 178.1 110.4 138.1 123.0 123.8 99.5 95.1 73.8 89.5 LHL 149.5 144.2 180.6 158.1 112.0 89.7 98.8 100.2 66.5 54.0 46.3 55.6 4 RFL 188.6 179.5 173.6 180.6 100.9 56.6 50.9 69.5 79.5 94.8 138.2 104.2 LFL 226.2 216.8 116.6 186.5 77.5 74.8 54.0 68.8 55.0 50.3 62.0 55.8 RHL 90.0 44.5 85.4 73.3 41.3 55.4 44.0 46.9 74.7 50.4 75.6 66.9 LHL 86.7 82.4 56.9 75.3 66.5 69.5 52.0 62.7 52.4 70.2 61.1 61.2 5 RFL 47.7 74.7 36.4 52.9 50.6 38.3 30.3 39.7 48.7 52.9 38.3 46.6 LFL 26.6 25.7 30.7 27.7 106.4 100.3 85.7 97.5 52.5 24.4 49.3 42.1 Continued

118

Table 12 Continued

RHL 103.9 82.7 37.8 74.8 99.3 93.6 110.4 101.1 55.8 50.0 89.9 65.2 LHL 156.3 128.9 144.9 143.4 109.2 91.4 69.7 90.1 90.7 77.2 99.0 89.0

6 RFL 104.6 115.3 104.5 108.1 115.8 154.0 128.1 132.6 31.9 34.3 51.0 39.1 LFL 189.2 113.7 96.2 133.0 85.5 75.2 96.1 85.6 61.7 54.3 81.9 66.0 RHL 233.3 234.1 138.6 202.0 134.8 129.4 121.7 128.6 75.9 70.9 69.7 72.2 LHL 78.3 51.8 103.2 77.8 64.4 108.1 100.0 90.8 61.5 51.3 58.0 56.9 7 RFL 297.7 180.7 210.5 229.6 73.5 69.2 108.0 83.6 123.5 98.7 101.6 107.9 LFL 123.8 80.4 80.7 95.0 69.5 116.6 58.2 81.4 67.4 64.6 41.2 57.7 RHL 42.2 67.0 75.2 61.5 66.6 53.5 59.1 59.7 83.9 65.2 69.2 72.8 LHL 95.4 85.7 95.7 92.3 66.7 28.7 28.5 41.3 85.2 69.6 87.3 80.7

119

3 5 8 RFL 102.8 68.7 86.6 86.0 117.7 100.5 86.0 101.4 110.0 137.4 120.6 122.7

LFL 92.4 95.0 93.5 93.6 61.6 90.0 75.3 75.6 81.4 69.6 59.4 70.1 RHL 165.4 150.8 154.3 156.8 155.0 194.4 192.6 180.7 175.8 223.1 210.2 203.0 LHL 130.8 104.1 143.7 126.2 197.3 138.8 170.8 169.0 184.3 174.7 141.4 166.8 9 RFL 156.2 135.2 105.2 132.2 134.8 188.8 159.8 161.1 130.1 211.8 151.7 164.5 LFL 185.4 67.2 178.8 143.8 120.0 183.2 120.3 141.2 209.6 295.2 198.5 234.4 RHL 234.1 238.9 239.7 237.6 249.5 190.3 215.0 218.3 168.1 149.6 160.1 159.3 LHL 172.1 217.1 188.3 192.5 262.8 218.5 219.0 233.4 187.0 214.3 133.4 178.2 10 RFL 152.5 166.1 134.8 151.1 184.6 221.7 185.7 197.3 215.2 195.8 194.4 201.8 LFL 214.4 237.6 282.4 244.8 241.0 214.2 222.4 225.9 121.5 155.3 139.7 138.8 Continued

119

Table 12 Continued

RHL 118.2 200.8 155.5 158.2 79.8 149.6 160.1 129.8 94.3 95.4 94.1 94.6 LHL 132.1 200.6 202.7 178.5 72.4 81.4 52.6 68.8 86.8 73.8 72.9 77.8 11 RFL 269.0 272.8 140.4 227.4 130.8 160.2 146.6 145.9 103.1 111.3 83.9 99.4 LFL 215.8 207.9 142.9 188.9 147.9 126.6 120.5 131.7 97.3 93.2 92.9 94.5 RHL 213.3 190.3 154.2 185.9 114.8 166.0 103.9 128.2 79.5 86.3 101.2 89.0 LHL 146.0 155.0 104.4 135.1 176.7 122.2 148.4 149.1 135.9 98.7 105.3 113.3 12 RFL 129.3 151.9 153.3 144.8 169.7 140.0 123.3 144.3 135.8 62.8 83.1 93.9 LFL 148.8 159.0 71.7 126.5 111.1 135.3 123.6 123.3 99.6 84.6 96.9 93.7 RHL 108.4 11.5 107.9 75.9 127.3 124.3 127.9 126.5 159.5 156.0 151.0 155.5 LHL 172.4 172.9 198.2 181.2 115.1 110.8 108.6 111.5 51.1 93.8 101.7 82.2

120

3 5 13 RFL 159.5 169.7 180.5 169.9 178.7 164.4 153.7 165.6 213.0 129.1 165.6 169.2

LFL 192.6 209.9 109.5 170.7 131.3 114.8 140.1 128.7 146.8 171.6 162.0 160.1 RHL 103.6 164.6 194.4 154.2 88.4 84.1 104.4 92.3 86.3 78.3 65.8 76.8 LHL 89.3 120.9 130.6 113.6 127.6 89.8 123.6 113.7 82.6 90.3 113.3 95.4 14 RFL 211.8 238.4 259.7 236.6 144.4 122.5 128.1 131.7 103.4 151.5 115.8 123.6 LFL 170.5 194.2 230.5 198.4 153.9 163.0 122.8 146.6 146.8 95.3 110.4 117.5 RHL 122.9 150.1 122.6 131.9 138.2 169.6 143.1 150.3 105.4 103.9 98.7 102.7 LHL 131.6 164.8 137.7 144.7 113.7 123.7 121.0 119.5 79.7 72.7 75.6 76.0 15 RFL 152.6 128.6 160.5 147.2 156.5 161.7 113.9 144.0 129.4 133.7 138.6 133.9 LFL 252.4 135.0 240.0 209.1 225.1 153.5 116.6 165.1 115.7 135.5 118.7 123.3 Continued

120 Table 12 Continued

RHL 125.5 155.1 160.3 147.0 117.9 87.6 83.4 96.3 86.9 98.4 77.1 87.5 LHL 281.2 240.5 167.5 229.7 243.7 222.7 208.5 225.0 107.3 95.3 100.6 101.1 16 RFL 203.2 192.8 217.5 204.5 180.3 192.3 134.2 168.9 149.4 200.0 133.6 161.0 LFL 186.7 432.4 315.7 311.6 281.4 256.1 285.8 274.4 140.4 130.1 92.6 121.0 RHL 87.8 125.2 116.0 109.7 75.0 59.3 66.5 66.9 30.6 48.7 49.8 43.0 LHL 98.8 101.7 120.2 106.9 52.2 90.8 99.6 80.9 62.5 77.8 71.0 70.4 17 RFL 81.1 92.9 103.6 92.5 90.0 81.5 100.3 90.6 93.7 86.2 74.6 84.8 LFL 119.3 108.9 88.4 105.5 58.3 49.0 79.6 62.3 76.7 37.6 86.0 66.8 RHL 94.0 110.6 121.1 108.6 142.6 177.3 130.8 150.2 96.3 142.5 113.5 117.4 LHL 175.0 134.1 169.9 159.7 124.4 115.2 111.9 117.2 147.6 128.3 142.7 139.5

121

3 5 18 RFL 202.7 120.8 164.3 162.6 140.9 166.2 177.4 161.5 140.6 133.7 140.6 138.3

LFL 118.7 164.5 241.2 174.8 127.8 110.5 123.8 120.7 168.1 125.8 115.3 136.4 RHL 101.7 102.6 115.0 106.4 77.9 74.3 119.2 90.5 134.2 101.0 110.0 115.1 LHL 111.3 113.2 102.8 109.1 72.2 79.9 61.7 71.3 92.9 65.3 79.3 79.2 19 RFL 135.6 166.9 167.8 156.8 127.8 100.9 119.4 116.0 128.8 163.6 134.5 142.3 LFL 121.3 137.6 146.5 135.1 119.6 93.1 107.9 106.9 97.9 91.0 101.3 96.7 RHL 133.1 157.9 128.6 139.9 193.3 139.9 104.8 146.0 130.8 143.8 122.6 132.4 LHL 94.4 98.2 160.1 117.6 134.6 108.4 169.1 137.4 131.9 111.0 157.4 133.4 20 RFL 156.4 145.7 160.5 154.2 144.6 123.5 147.6 138.6 124.0 146.8 65.2 112.0 LFL 116.3 157.4 128.7 134.1 179. 1 183.5 144.4 169.0 133.7 112.5 97.0 114.4

121

Table 13 von Frey anesthesiometry data for spinal cord injury affected dogs (n=30)

Dog / Day 3 Day 10 Day 30 Surgery Limb Day 3 Day 10 Day 30 Mean Mean Mean side RHL 348.7 385.9 348.9 361.2 184.7 178.9 195.7 186.4 224.5 276.7 160.1 220.4 1 LHL 289.7 347.7 366.0 334.5 201.4 252.0 226.6 226.7 259.4 330.6 277.0 289.0 Right RFL 146.5 167.8 176.3 163.5 109.2 144.6 131.1 128.3 176.1 257.0 283.7 238.9 LFL 210.5 279.7 217.6 235.9 143.8 208.9 86.6 146.4 171.7 144.9 120.9 145.8 RHL 226.4 165.1 128.2 173.2 48.2 72.1 61.0 60.4 80.1 88.7 84.4 84.4 2 LHL 174.7 151.6 156.8 161.0 108.9 106.7 82.8 99.5 30.9 36.6 57.9 41.8 Right

122 RFL 212.0 188.9 199.3 200.1 65.2 71.4 54.0 63.5 61.0 75.6 106.0 80.9

3 5

LFL 112.4 160.4 170.0 147.6 121.9 123.8 121.1 122.3 65.0 55.0 42.7 54.2

RHL 189.5 183.3 149.2 174.0 99.6 125.6 125.2 116.8 104.9 119.3 119.6 114.6 3 LHL 173.6 146.8 184.4 168.3 124.4 133.2 135.3 131.0 66.1 90.7 99.9 85.6 Left RFL 157.1 144.8 192.0 164.6 133.6 106.6 129.5 123.2 84.9 60.7 99.6 81.7 LFL 117.2 66.0 111.2 98.1 109.5 78.8 59.0 82.4 81.4 82.3 69.0 77.6 Continued

122

Table 13 Continued

RHL 396.4 347.1 401.2 381.6 330.9 225.6 290.2 282.2 325.4 369.1 282.9 325.8 4 LHL 324.2 413.3 484.6 407.4 274.1 244.0 252.5 256.9 257.4 309.1 262.8 276.4 Right RFL 293.6 267.7 277.3 279.5 125.8 164.3 181.3 157.1 300.2 177.1 304.5 260.6 LFL 236.5 272.0 245.9 251.5 176.6 172.5 108.3 152.5 143.6 79.8 98.0 107.1 RHL 297.7 272.5 338.5 302.9 132.9 143.8 151.7 142.8 110.0 110.7 106.8 109.2 5 LHL 233.9 255.5 227.5 239.0 76.1 59.1 79.8 71.7 42.6 46.3 43.0 44.0 Left RFL 143.9 132.1 108.8 128.3 102.0 103.6 90.8 98.8 96.0 97.5 122.9 105.5 LFL 74.3 93.6 56.2 74.7 51.9 89.2 92.7 77.9 90.5 63.2 117.7 90.5 RHL 171.5 128.4 266.5 188.8 112.1 85.5 90.2 95.9 94.1 96.3 118.5 103.0 6

123 LHL 236.8 313.5 209.7 253.3 95.9 166.8 139.8 134.2 118.3 125.4 117.5 120.4 Left RFL 185.9 243.5 284.5 238.0 69.2 72.6 67.1 69.6 122.7 127.5 135.7 128.6

LFL 189.4 169.9 154.9 171.4 82.8 91.8 103.9 92.8 124.2 129.5 123.1 125.6 RHL 220.2 233.5 233.3 229.0 179.8 157.8 164.9 167.5 251.8 232.5 246.1 243.5

7 LHL 246.4 252.3 241.9 246.9 166.0 239.7 193.7 199.8 344.7 288.7 225.9 286.4 Right RFL 120.7 106.5 86.0 104.4 160.1 158.5 127.0 148.5 162.6 134.9 206.4 168.0 LFL 155.7 164.1 162.6 160.8 79.7 124.7 89.2 97.9 141.7 84.8 90.7 105.7 RHL 154.9 155.3 127.4 145.9 124.9 162.9 160.1 149.3 204.0 172.0 200.4 192.1 8 LHL 160.9 183.8 169.5 171.4 210.8 165.9 150.6 175.8 113.1 109.3 134.6 119.0 Right RFL 100.8 109.0 108.6 106.1 115.3 88.4 105.9 103.2 105.0 118.3 103.8 109.0 LFL 71.6 81.7 93.1 82.1 105.2 143.0 100.9 116.4 126.0 130.5 125.4 127.3 Continued

123 Table 13 Continued

RHL 390.3 418.1 342.3 383.6 281.3 251.5 220.6 251.1 216.4 275.0 229.0 240.1 9 LHL 449.8 563.3 449.2 487.4 255.7 472.9 413.6 380.7 285.0 279.3 142.4 235.6 Right RFL 207.6 207.9 95.9 170.5 151.8 142.3 172.1 155.4 150.8 173.7 187.1 170.5 LFL 261.7 232.6 281.9 258.7 128.5 172.1 111.2 137.3 191.6 190.5 153.5 178.5 RHL 290.4 288.8 287.9 289.0 312.2 205.2 224.0 247.1 424.9 393.6 454.2 424.2 10 LHL 314.6 314.5 272.2 300.4 349.3 279.7 327.6 318.9 385.4 292.2 459.7 379.1 Left RFL 42.0 165.9 104.8 104.2 189.1 159.3 170.7 173.0 248.5 272.3 276.1 265.6 LFL 209.5 228.6 229.9 222.7 263.2 200.5 201.1 221.6 334.3 240.5 269.9 281.6 RHL 395.9 252.0 396.2 348.0 189.2 172.1 150.5 170.6 182.9 135.4 227.8 182.0 11

124 LHL 335.9 258.0 303.4 299.1 353.2 373.9 307.2 344.8 216.5 189.1 205.2 203.6 Left RFL 173.9 219.1 170.3 187.8 233.8 238.2 208.2 226.7 259.3 166.0 235.5 220.3

LFL 273.4 238.0 244.8 252.1 212.1 238.4 255.8 235.4 209.0 218.7 184.6 204.1 RHL 322.4 320.1 298.2 313.6 365.2 346.1 394.5 368.6 326.6 276.6 321.4 308.2

12 LHL 392.5 251.7 311.9 318.7 264.4 248.8 320.6 277.9 239.8 328.2 286.5 284.8 Left RFL 310.1 339.4 205.6 285.0 374.3 332.2 288.6 331.7 236.2 286.4 159.6 227.4 LFL 170.7 251.8 213.5 212.0 378.2 324.8 349.4 350.8 232.5 233.6 142.4 202.8 RHL 250.4 259.7 235.5 248.5 285.7 337.2 335.1 319.3 260.5 300.4 282.9 281.3 13 LHL 295.2 393.4 293.4 327.3 235.5 289.0 274.6 266.4 290.6 238.9 164.1 231.2 Right RFL 231.5 227.8 200.4 219.9 259.9 241.6 268.0 256.5 214.4 148.6 153.8 172.3 LFL 177.7 201.3 151.3 176.8 296.5 250.9 199.8 249.1 208.3 179.5 129.2 172.3 Continued

124 Table 13 Continued

RHL 233.0 221.1 365.5 273.2 214.8 211.0 267.6 231.1 101.8 77.9 140.0 106.6 14 LHL 396.5 444.6 407.4 416.2 15.1 44.1 15.9 25.0 12.7 16.2 30.1 19.7 Right RFL 35.1 75.9 120.1 77.0 162.5 154.7 185.7 167.6 136.6 150.8 133.8 140.4 LFL 362.9 370.6 314.0 349.2 162.0 141.9 169.3 157.7 127.0 102.0 100.7 109.9 RHL 681.2 702.3 757.4 713.6 66.8 119.0 113.4 99.7 343.7 283.7 238.1 288.5 15 LHL 572.5 611.3 537.7 573.8 416.3 378.3 421.3 405.3 226.5 178.3 158.5 187.8 Left RFL 165.5 191.2 215.7 190.8 206.3 146.7 153.7 168.9 215.3 234.8 143.4 197.8 LFL 253.3 317.0 311.4 293.9 216.9 208.9 156.1 194.0 188.3 289.8 219.0 232.4 RHL 882.0 584.5 379.1 615.2 252.2 138.8 171.0 187.3 177.4 144.3 153.3 158.3 16

125 LHL 297.9 254.7 321.3 291.3 151.8 130.5 136.0 139.4 172.6 154.3 108.7 145.2 Left RFL 480.9 393.7 315.4 396.7 114.7 154.7 256.1 175.2 239.0 218.4 286.9 248.1

LFL 379.9 438.7 403.6 407.4 149.1 150.7 193.4 164.4 185.9 164.4 168.8 173.0 RHL 78.2 158.9 123.3 120.1 107.8 66.8 94.7 89.8 128.4 113.7 105.2 115.8

17 LHL 520.0 526.9 422.2 489.7 221.2 205.9 203.6 210.2 127.1 175.2 123.1 141.8 Left RFL 370.6 332.0 348.4 350.3 153.5 148.6 142.3 148.1 165.1 112.7 113.4 130.4 LFL 89.5 69.7 105.7 88.3 34.4 52.8 63.6 50.3 189.5 133.1 64.4 129.0 RHL 473.7 471.5 499.2 481.5 367.9 382.1 307.1 352.4 174.0 152.3 123.5 149.9 18 LHL 555.4 534.7 561.7 550.6 401.2 347.9 319.0 356.0 165.0 112.9 102.8 126.9 Left RFL 104.8 76.2 71.5 84.2 43.6 20.7 39.0 34.4 125.7 112.7 94.3 110.9 LFL 151.0 83.7 183.0 139.2 45.0 61.0 66.1 57.4 95.3 127.7 88.7 103.9 Continued

125

Table 13 Continued

RHL 218.7 251.7 241.2 237.2 238.7 164.9 256.6 220.1 179.8 167.3 117.6 154.9 19 LHL 284.0 322.9 280.5 295.8 266.7 282.4 292.8 280.6 239.4 234.8 198.0 224.1 Left RFL 158.7 146.8 176.6 160.7 173.7 215.5 153.1 180.8 197.5 191.1 203.0 197.2 LFL 170.7 171.5 108.6 150.3 201.1 165.0 157.4 174.5 172.1 231.4 197.4 200.3 RHL 468.2 330.7 383.4 394.1 373.6 331.8 369.8 358.4 233.1 313.5 280.2 275.6 20 LHL 529.6 488.6 492.0 503.4 363.1 438.7 530.7 444.2 359.6 309.2 311.4 326.7 Left RFL 205.4 208.8 220.0 211.4 142.5 142.4 150.6 145.2 110.5 64.7 101.5 92.2 LFL 204.9 152.9 177.2 178.3 131.6 140.5 126.0 132.7 220.3 145.5 157.7 174.5 RHL 217.4 207.0 242.0 222.1 116.6 159.3 147.8 141.2 279.2 286.4 178.5 248.0 21

126 LHL 461.8 502.7 357.0 440.5 236.5 268.3 207.3 237.4 290.7 325.4 298.9 305.0 Left RFL 63.9 56.4 131.7 84.0 130.8 167.5 205.4 167.9 257.2 267.3 235.1 253.2

LFL 70.7 133.2 80.1 94.7 117.2 172.1 231.4 173.6 183.2 270.6 230.5 228.1 RHL 942.1 605.6 803.8 783.8 664.6 481.6 596.0 580.7 322.4 321.8 327.6 323.9

22 LHL 576.8 666.6 678.7 640.7 458.2 493.2 496.2 482.5 433.1 301.2 285.2 339.8 Right RFL 626.3 706.0 604.8 645.7 250.2 233.6 262.1 248.6 196.6 202.4 194.6 197.9 LFL 339.7 319.5 445.1 368.1 271.5 264.9 270.4 268.9 204.6 169.5 159.2 177.8 RHL 306.4 221.6 279.0 269.0 282.6 333.0 285.3 300.3 124.1 126.8 165.9 138.9 23 LHL 302.1 247.8 304.5 284.8 224.7 199.5 351.3 258.5 209.9 134.7 150.9 165.2 Right RFL 239.5 192.5 224.6 218.9 238.1 230.9 205.1 224.7 117.2 201.6 168.2 162.3 LFL 172.4 182.9 185.1 180.1 226.6 174.0 198.7 199.8 95.9 112.2 124.7 110.9 Continued

126

Table 13 Continued

RHL 693.7 620.4 594.2 636.1 354.2 295.7 350.8 333.6 181.2 165.9 166.7 171.3 24 LHL 532.4 440.8 562.0 511.7 277.7 279.6 268.5 275.3 138.8 126.1 132.2 132.4 Right RFL 210.3 258.8 157.9 209.0 91.1 115.8 144.1 117.0 75.1 69.5 95.6 80.1 LFL 308.3 366.0 139.6 271.3 78.1 157.2 138.0 124.4 83.2 108.8 99.6 97.2 RHL 511.2 563.4 592.6 555.7 650.3 535.0 521.4 568.9 377.0 402.5 391.3 390.3 25 LHL 575.2 760.3 768.1 701.2 558.9 558.4 748.2 621.8 335.3 422.2 365.0 374.2 Right / Left RFL 210.4 132.6 151.4 164.8 189.5 188.1 155.4 177.7 78.4 103.1 81.7 87.7 LFL 273.8 243.3 273.3 263.5 179.0 215.2 264.5 219.6 122.1 131.7 119.0 124.3 RHL 268.8 252.8 233.0 251.5 131.0 134.6 140.6 135.4 99.6 108.8 103.1 103.8 26

127 LHL 180.2 151.3 174.1 168.5 122.2 136.6 70.0 109.6 159.4 178.0 164.6 167.3 Right RFL 146.9 192.5 142.5 160.6 115.1 114.6 100.1 109.9 93.0 102.2 99.1 98.1

LFL 128.7 97.7 101.5 109.3 78.1 91.5 81.6 83.7 74.7 112.1 97.1 94.6 RHL 324.3 377.7 318.1 340.0 197.6 222.5 211.9 210.7 249.3 232.3 255.5 245.7

27 LHL 319.7 327.9 309.7 319.1 202.5 219.4 217.9 213.3 204.8 181.5 135.1 173.8 Right RFL 139.8 148.4 138.5 142.2 92.1 104.4 99.6 98.7 129.2 109.7 122.9 120.6 LFL 221.4 157.2 174.5 184.4 83.8 162.4 110.5 118.9 143.5 104.8 114.1 120.8 RHL 223.5 275.3 183.3 227.4 149.9 181.5 107.2 146.2 231.2 250.1 203.2 228.2 28 LHL 189.9 178.3 216.3 194.8 162.6 181.6 153.9 166.0 237.9 256.0 184.5 226.1 Right RFL 193.3 202.7 313.3 236.4 72.0 82.5 103.0 85.8 175.6 172.9 203.4 184.0 LFL 177.9 146.5 175.9 166.8 101.6 107.3 131.1 113.3 270.6 241.7 198.7 237.0 Continued

127

Table 13 Continued

RHL 656.2 619.5 493.7 589.8 208.5 232.6 214.3 218.5 221.9 238.9 233.9 231.6 29 LHL 282.6 397.1 349.5 343.1 356.5 254.2 198.2 269.6 286.1 293.4 301.6 293.7 Right RFL 255.5 124.9 277.1 219.2 85.5 134.5 82.8 100.9 164.1 188.1 234.9 195.7 LFL 189.9 178.5 97.1 155.2 152.6 155.6 155.5 154.6 78.3 140.9 159.9 126.4 RHL 224.0 231.5 204.9 220.1 159.9 209.3 199.7 189.6 194.7 174.9 161.1 176.9 30 LHL 206.5 175.4 180.8 187.6 199.3 216.0 236.1 217.1 184.3 142.8 148.1 158.4 Right RFL 150.9 151.3 159.0 153.7 170.2 165.6 161.1 165.6 143.8 133.0 151.0 142.6 LFL 112.3 120.4 111.3 114.7 118.5 106.8 115.9 113.7 123.1 106.9 115.0 115.0

128

128 Table 14 Locomotor gait scores in dogs with spinal cord injury

3 Days Postop 10 Days Postop 30 Days Postop Left BBB Right BBB OSCIS MFS Left BBB Right BBB OSCIS MFS Left BBB Right BBB OSCIS MFS 1 17 17 11 4 18 18 13 4 19 19 13 4 2 19 16 11 4 19 17 10 4 19 19 11 4 3 18 18 11 4 19 17 11 4 19 19 14 5 4 1 1 4 3 7 8 5 3 10 10 7 3 5 10 11 9 4 17 18 11 4 17 18 11 4 6 13 13 9 4 18 18 13 4 18 18 13 4 7 5 0 4 3 12 12 10 4 18 17 11 4 8 10 9 6 3 11 11 9 4 13 13 10 4 9 10 9 7 3 13 13 10 4 19 17 13 4

129 10 0 0 1 2 4 9 6 3 11 11 8 4

11 1 10 6 3 18 17 11 4 17 17 11 4 12 17 17 11 4 18 18 13 4 18 18 13 4 13 13 13 9 4 17 17 10 4 19 19 13 4 14 5 1 4 3 10 10 7 3 12 12 9 4 15 0 0 1 2 11 11 8 4 19 17 11 4 16 17 17 10 4 17 17 11 4 17 17 11 4 17 14 18 10 4 17 19 11 4 19 19 14 5 Continued

129

Table 14 Continued

18 0 0 1 2 10 11 7 3 18 18 13 4 19 9 10 7 3 12 12 9 3 17 19 11 4 20 1 1 4 3 11 11 9 4 18 17 11 4 21 2 8 4 3 10 10 6 3 18 17 13 4 22 1 0 3 3 2 1 4 3 11 11 9 4 23 17 17 11 4 18 18 13 4 18 18 11 4 24 1 0 3 3 11 11 8 4 13 13 10 4 25 1 2 4 3 8 9 5 3 11 11 8 4 26 1 2 11 4 18 18 11 4 18 18 11 4 27 13 13 4 3 13 13 11 4 17 17 11 4 28 19 17 11 4 18 18 14 5 19 17 11 4

130 29 10 10 6 3 13 13 9 4 15 15 9 4

30 13 13 10 4 18 17 11 4 17 17 11 4 Median 10 10 6.5 3 13 13 10 4 18 17 11 4 Range 0-19 0-18 1-11 2-4 2-19 1-19 4-13 3-5 10-19 10-19 7-14 3-5 BBB, Basso-Beattie-Bresnahan scale adapted for dogs (Range 0-19); OSCIS, Olby Spinal Cord Injury Scale; MFS, Modified Franklin Score

130