Biomarkers of canine

Kathleen L. Graham

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Clinical Ophthalmology and Health Sydney Medical School The University of Sydney

2019

P a g e | 2

DECLARATION

This thesis is submitted to the University of Sydney in fulfilment of the requirement for the degree of Doctor of Philosophy. This is to certify that the content of this thesis is my own work. This thesis has not previously been submitted for any degree or other purposes. I certify that the intellectual content of this thesis is the product of my own work and that all the assistance received in preparing this thesis and sources has been acknowledged.

Kathleen Graham 27/09/2019 P a g e | 3

DISCLOSURE AND AUTHOR CONTRIBUTIONS

The following chapters have been re-formatted from manuscripts published or under review in peer-reviewed journals. For each of the manuscripts I was responsible for study design; obtaining and complying with animal ethics approval and requirements; recruitment; conducting clinical examinations and imaging; collating and interpreting data; statistical analysis; manuscript composition and submission. Co-authors Associate Professor Andrew White, Dr Christina McCowan, Dr Eve Diefenbach, Dr Pip Johnson, Professor Jacky Reid, Professor Mitchell Lawlor, Erica Barry, Dr Cameron Whittaker, Dr Evelyn Hall, Dr Kelly Caruso, Sarah-Elizabeth Byosiere, Lynna Feng, Dr Matthew Saunders, Professor Pauleen Bennett, Usha Pattamatta, Marisa Perez Orrico, Domenic Soligo, Dr Mark Billson, Dr David Donaldson, Dr Charles Caraguel, and Professor Frank Billson assisted with conduct of the studies, analyses and revision of the manuscripts. The manuscripts produced for this thesis are as follows: CHAPTER 1 Graham KL, McCowan C, White A. Genetic and biochemical biomarkers in primary canine glaucoma. Veterinary Pathology 2016; 54 (2): 194-203 doi: 10.1177/030098581666661. CHAPTER 2 Graham KL, Reid J, Whittaker CJG, Hall EJS, Caruso K, McCowan CI, White A. Development of a vision impairment score for the assessment of functional vision in dogs: initial evidence of validity, reliability and responsiveness. Veterinary Ophthalmology. 30th January 2019, DOI: 10.1111/vop.12656. CHAPTER 3 Graham KL, Byosiere S-E, Feng LC, Sanders M, Bennett PC, Caruso K, McCowan CI, White A. A forced-choice preferential looking task for the assessment of vision in dogs: a pilot study. Journal of Small Animal Practice 2019; 60(6): 340-347 doi: 10.1111/jsap.12965 . CHAPTER 4 Graham KL, McCowan CI, White AJR. A modified protocol for assessment of the pupillary light reflex in dogs predisposed to glaucoma. Journal of Small Animal Practice. Under review CHAPTER 5 Graham KL, Whittaker CJG, Caruso K, McCowan CI, White A. Anterior segment ultrasound biomicroscopy of the unaffected eye in dogs with primary angle closure glaucoma. Veterinary Ophthalmology. Under review CHAPTER 6 Graham KL, McCowan CI, Caruso K, Billson FM, Whittaker CJG, White A. Optical coherence tomography of the , nerve fiber layer and head in dogs with glaucoma. Veterinary Ophthalmology. 16th June 2019 DOI: 10.1111/vop.12694. CHAPTER 7 Graham KL, Johnson PJ, Barry E, Perez Orrico M, Soligo D, Lawlor M, White AJR. Diffusion tensor imaging of the visual pathway in dogs with primary angle closure glaucoma. Veterinary Ophthalmology. Under review P a g e | 4

CHAPTER 8 Graham KL, Diefenbach E, McCowan CI, White AJR. Shotgun proteomic analysis of the precorneal tear film in dogs with primary glaucoma. Veterinary Ophthalmology. Under review CHAPTER 9 Graham KL, Donaldson D, Billson FA, Billson FM. Use of a 350mm2 Baerveldt glaucoma drainage device to maintain vision and control in dogs with glaucoma: a retrospective study (2013-2016). Veterinary Ophthalmology. 2017; 20(5): 427-434. DOI: 10.1111/vop.12443. CHAPTER 10 Graham KL, Hall EJS, Caraguel C, White A, Billson FA, Billson FM. Comparison of diode laser trans-scleral cyclophotocoagulation versus implantation of a 350mm2 Baerveldt glaucoma drainage device for the treatment of glaucoma in dogs (a retrospective study: 2010-2016). Veterinary Ophthalmology. 2018; 21(5): 487-497 DOI: 10.1111/vop.12536.

FORMAL PRESENTATIONS

Diffusion tensor imaging of the visual pathway in dogs with angle closure glaucoma. American College of Veterinary Radiology annual conference; Baltimore Maryland, USA. November 2019 Posterior segment optical coherence tomography in dogs predisposed to primary glaucoma. American College of Veterinary Ophthalmology annual conference; Minneapolis Minnesota, USA. September 2018 Biomarkers in canine glaucoma – structure and function. Australian and New Zealand College of Veterinary Scientists Annual Conference; Ophthalmology chapter; Gold Coast Queensland. July 2018 Optics of vision. Royal Veterinary College seminar; Royal Veterinary College, University of London; Hertfordshire, United Kingdom. October 2017 Vision assessment in dogs. UNSW School of Psychology Vision Research presentation; University of New South Wales, Randwick NSW. May 2017 Biomarkers of canine glaucoma. Save Sight Institute Annual Research Symposium; Sydney Eye Hospital Sydney NSW. January 2017

P a g e | 5

DECLARATION BY SUPERVISOR

This is to certify that all co-authors have provided their consent for inclusion of the manuscripts presented in this thesis. The co-authors accept the Kathleen’s contribution to each manuscript and the description of the co-authors’ contribution.

Clinical Associate Professor Andrew White 30/09/2019

P a g e | 6

ACKNOWLEDGEMENTS

There are many people who have been instrumental in the development and conduct of this project. Most importantly, I want to acknowledge the dogs that were included and their owners who saw the value of participating in this clinical research. To my supervisor, Clin. A/Professor Andrew White, I am grateful for your voice of reason, practical approach to clinical research, and your eternal optimism. To my auxiliary supervisor Dr Christina McCowan, I am grateful for your common sense, grammatical corrections, and for your interest in the details and all the possibilities. To all the veterinarians and nurses who allowed me to spend time in their hospitals, meet their clients, and examine their patients, especially Dr Anne Fawcett, Dr Belinda Parsons and Dr Kelly Caruso. I very much appreciate being allowed into your hospitals to work with your clients and your patients. And to those who collaborated on the studies and made them a reality, I remain grateful. Especially Dr Philippa Johnson from Cornell University College of Veterinary Medicine for sharing her talents and kindness; Dr Eve Diefenbach from the Westmead Institute for Medical Research proteomics department for her patience, time, knowledge, skill and for always seeing the funny side; and Professor Jacky Reid for her time and expertise. P a g e | 7

TABLE OF CONTENTS

DECLARATION ...... 2 DISCLOSURE AND AUTHOR CONTRIBUTIONS ...... 3 FORMAL PRESENTATIONS ...... 4 DECLARATION BY SUPERVISOR ...... 5 ACKNOWLEDGEMENTS ...... 6 TABLE OF CONTENTS ...... 7 ABSTRACT ...... 13 LIST OF FIGURES ...... 15 LIST OF TABLES ...... 23 LIST OF ABBREVIATIONS ...... 26

CHAPTER ONE AIM OF THE THESIS AND LITERATURE REVIEW Literature review ...... 32 What is glaucoma? ...... 32 Classification ...... 33 Aetiology ...... 33 Iridocorneal angle morphology ...... 38 Stage of disease ...... 39 dynamics ...... 39 Genetic and biochemical biomarkers in primary canine glaucoma ...... 43 Abstract ...... 43 Introduction ...... 43 Genetic markers ...... 45 Biochemical markers ...... 49 Future work ...... 52 Conclusion ...... 53 Diagnosis ...... 54 Clinical diagnosis ...... 54 Treatment ...... 60 Medical treatment ...... 60 Surgical treatment ...... 66 P a g e | 8

CHAPTER TWO DEVELOPMENT OF A VISION IMPAIRMENT SCORE FOR THE ASSESSMENT OF FUNCTIONAL VISION IN DOGS: INITIAL EVIDENCE OF VALIDITY, RELIABILITY AND RESPONSIVENESS Abstract ...... 70 Introduction ...... 71 Materials and Methods ...... 72 Questionnaire development ...... 72 Testing for validity, reliability and responsiveness ...... 73 Statistical analysis of sample population ...... 75 Results ...... 77 Content validation ...... 78 Assessment of validity, reliability and responsiveness of the VIS ...... 78 Discussion ...... 84 Footnotes ...... 88 Acknowledgements ...... 88

CHAPTER THREE A FORCED-CHOICE PREFERENTIAL LOOKING TASK FOR THE ASSESSMENT OF VISION IN DOGS: PILOT STUDY ABSTRACT...... 89 Introduction ...... 90 Materials and methods ...... 90 Observer (human) reliability assessment ...... 91 Animals ...... 91 Canine vision (estimated visual acuity) assessment ...... 93 Results ...... 95 Observer (human) reliability outcomes ...... 95 Canine vision (estimated visual acuity) outcomes ...... 96 Discussion ...... 99

CHAPTER FOUR A MODIFIED PROTOCOL FOR ASSESSMENT OF THE PUPILLARY LIGHT REFLEX IN DOGS PREDISPOSED TO GLAUCOMA Abstract ...... 103 P a g e | 9

Introduction ...... 104 Materials and Methods ...... 104 Animals ...... 104 Procedures ...... 105 Statistical analyses ...... 106 Results ...... 107 Baseline pupil size ...... 110 Red light stimulation ...... 110 Blue light stimulation ...... 110 White light stimulation ...... 110 Test-retest reliability ...... 110 Discussion ...... 114 Acknowledgements ...... 116

CHAPTER FIVE ANTERIOR SEGMENT ULTRASOUND BIOMICROSCOPY OF THE UNAFFECTED EYE IN DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA Abstract ...... 117 Introduction ...... 118 Materials and Methods ...... 118 Case selection ...... 119 Ultrasound biomicroscopy ...... 119 Data analyses ...... 119 Statistical analysis ...... 123 Results ...... 123 Group effect ...... 127 Quadrant effect ...... 128 Interaction of group and quadrant ...... 128 Measurements corrected for body weight ...... 129 Reproducibility of images...... 132 Intra-rater reliability of measurements ...... 132 Discussion ...... 132 Acknowledgements ...... 134

P a g e | 10

CHAPTER SIX OPTICAL COHERENCE TOMOGRAPHY OF THE RETINA, NERVE FIBRE LAYER AND OPTIC NERVE HEAD IN DOGS WITH GLAUCOMA Abstract ...... 135 Introduction ...... 136 Materials and Methods ...... 139 Animals ...... 139 Procedures ...... 139 Image acquisition ...... 140 Scan adjustments and segmentation ...... 141 Repeatability of scanning technique ...... 146 Reliability of observer measurements ...... 146 Statistical Analyses ...... 146 Results ...... 146 Retinal thickness ...... 147 Nerve fibre layer thickness...... 150 Optic nerve head analyses ...... 152 Ganglion cell complex ...... 152 Repeatability of scanning technique ...... 152 Reliability of observer measurements ...... 153 Discussion ...... 155 Acknowledgements ...... 158

CHAPTER SEVEN DIFFUSION TENSOR IMAGING OF THE VISUAL PATHWAY IN DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA ABSTRACT...... 160 Introduction ...... 161 Material and Methods ...... 162 Subjects ...... 162 Ophthalmic examination ...... 162 Imaging protocol ...... 163 Data processing ...... 163 Statistical analysis ...... 165 Results ...... 166 P a g e | 11

Subjects ...... 166 T1-weighted magnetic resonance imaging ...... 167 Intra-observer reliability ...... 167 Diffusion tensor imaging parameters ...... 167 Discussion ...... 172 Acknowledgements ...... 174

CHAPTER EIGHT SHOTGUN PROTEOMIC ANALYSIS OF THE PRECORNEAL TEAR FILM IN DOGS WITH PRIMARY GLAUCOMA Abstract ...... 175 Introduction ...... 176 Materials and methods ...... 176 Subjects ...... 177 Tear collection and storage ...... 177 Tear analyses ...... 177 Mass spectrometry ...... 178 Data processing and statistical analysis ...... 179 Results ...... 183 Protein recovery with and without dye in the tear strip, and diluent in the storage tube ...... 183 Proteins unique to disease states ...... 183 Comparative proteomics analysis by label-free mass spectrometry ...... 185 ...... 188 Molecular function and pathway analysis of differential proteins ...... 191 Discussion ...... 191 Acknowledgements ...... 194

CHAPTER NINE USE OF A BAERVELDT-350MM2 GLAUCOMA DRAINAGE DEVICE TO MAINTAIN VISION AND CONTROL INTRAOCULAR PRESSURE IN DOGS WITH GLAUCOMA: A RETROSPECTIVE STUDY (2013-2016) ABSTRACT...... 195 INTRODUCTION ...... 196 MATERIALS AND METHODS ...... 197 Selection criteria ...... 197 P a g e | 12

Surgical procedure ...... 198 Follow up ...... 200 Outcomes evaluated ...... 200 RESULTS ...... 201 Clinical findings ...... 201 Implant and technique modifications ...... 201 Surgical outcome ...... 202 Complications ...... 203 DISCUSSION ...... 203 CONCLUSION...... 206

CHAPTER TEN COMPARISON OF DIODE LASER TRANSSCLERAL CYCLOPHOTOCOAGULATION VERSUS IMPLANTATION OF A 350MM2 BAERVELDT GLAUCOMA DRAINAGE DEVICE FOR THE TREATMENT OF GLAUCOMA IN DOGS (A RETROSPECTIVE STUDY: 2010-2016) Abstract ...... 207 Introduction ...... 208 Materials and Methods ...... 209 Patient selection...... 209 Surgical Procedures ...... 209 Outcomes evaluated ...... 210 Postoperative management...... 211 Data analysis ...... 212 Results ...... 212 Outcomes following TSCP ...... 214 Surgical outcome ...... 214 Additional surgery ...... 216 Complications ...... 217 Discussion ...... 218 Conclusion ...... 223 Acknowledgements ...... 223

CHAPTER ELEVEN ...... 224 CONCLUSION ...... 224 REFERENCES ...... 226 P a g e | 13

ABSTRACT

Glaucoma is a leading cause of blindness worldwide1 and successful management of the disease is complicated by a limited understanding of pathophysiological mechanisms, a slow and insidious onset and progression, and difficulties directly observing endpoints of disease (and treatment). Progress in our understanding of the disease is therefore limited by the need for prolonged and costly clinical trials. These difficulties mean monitoring the effect of, and response to novel therapeutic interventions in clinical trials is limited. Therefore, the use of animal models that show the characteristic progressive loss of (RGC) function becomes necessary. Primary glaucoma occurs in dogs at a similar rate to that in humans, and the dog has been suggested as a suitable model of the disease due to the occurrence of natural disease in heterogenous populations, the size of the eye, and the visual system which functions at day and at night. Although presented as a model of glaucoma, the canine eye has never been validated as such. To be an effective model of disease, similarities and differences between these species must be known, quantifiable methods of identifying and monitoring disease progression established, and objective methods of determining response to treatment are required. With establishment of these entities, the validity of a potential biomarker, or more likely a group of biomarkers, may be investigated for relevant outcomes of interest in both human and veterinary patients. Chapter 1 outlines the current knowledge regarding canine glaucoma. Currently used diagnostic and therapeutic strategies are summarised and gaps in current understanding of the disease as it relates to the dog, are identified. In Chapter 2 and 3, preliminary studies describing tests to obtain quantifiable measures of functional vision in dogs for use in a clinical setting are presented. Initial development and validity of the canine Vision Impairment Score (VIS) is described (Chapter 2). Using a proxy (owner)-based questionnaire about daily performance and abilities, the VIS provides a measure that differed between blind and sighted , between dogs based on the number of sighted eyes, and between sighted dogs with some vision impairment compared to those without impairment. In Chapter 3, a forced-choice preferential looking task that measures estimated visual acuity in dogs is described. This paradigm, modelled on testing used in human infants, demonstrated measures of visual acuity in emmetropic dogs that showed no evidence of vision impairment and had no prior training. In Chapter 4, a modified pupillometry assessment using equipment that is both available and cost effective for interested veterinarians to have, enabling evaluation of function that is not dependent on vision. Having investigated methods of assessing visual function, Chapters 5-7 describe cross- sectional studies using imaging techniques to investigate potential structural biomarkers of glaucoma. In these studies, structural measures in eyes with and without glaucoma, and P a g e | 14 those predisposed to the development of glaucoma but before the onset of clinically recognisable disease, were compared. In Chapter 5 structures affecting aqueous humour outflow in the anterior segment are evaluated with the use of ultrasound biomicroscopy. We identified measures of the ciliary cleft and iridocorneal angle that differed between groups. In Chapter 6 optical coherence tomography (OCT) is used to compare the peripapillary retina, retinal nerve fibre layer, and the optic nerve head between normal eyes and those predisposed to glaucoma. Structural differences are identified that discriminate predisposed eyes from healthy eyes. The potential that these changes are suggestive of early disease, and that serial monitoring may facilitate monitoring of disease progression at earlier stages than are currently possible, is proposed In Chapter 7, diffusion tensor imaging (DTI), a magnetic resonance technique that measures the properties of water diffusion, is used to evaluate the integrity and architecture of white matter tracts of the visual pathway beyond the retina. Using this non- invasive technique where indices correlate with tissue microstructure and pathology, we demonstrate measurable differences in diffusivity within the visual pathway of dogs with and without PACG. In Chapter 8 mass spectrometry is used to investigate the potential use of the canine tear film as a source of biomarkers of disease in glaucoma. Tear samples were collected using Schirmer tear test strips and protein expression as well as pathway analyses were performed. Tear proteins unique to the tears from eyes with both primary open and closed angle glaucoma are identified, and proteins uniquely expressed in the tears of eyes predisposed to PACG are also identified. Chapter 9 outlines the technique and outcomes in a series of dogs with naturally occurring glaucoma that is refractory to medical treatment. Placement of a Baerveldt-350mm2 implant with modifications commonly used in human glaucoma patients is described. The outcomes of this surgical technique, which is not previously described in dogs, are described. In Chapter 10, a review of clinical outcomes seen in dogs undergoing surgical treatment of glaucoma at a veterinary referral hospital is performed. Comparisons between dogs treated with the technique described in Chapter 7 and dogs treated with trans-scleral destruction using a diode laser are made. The merits of techniques to shunt aqueous humour from the eye compared to cyclodestructive techniques are discussed with the suggestion these techniques treat the disease in a more physiologically appropriate manner. Chapter 11 forms the conclusion of this thesis. A discussion of how the objectives of this research to establish a paradigm for the assessment of glaucoma in dogs is presented. The significance of the findings to clinical veterinary practice as well as vision science researchers is outlined. An outline of initial proposals for subsequent work to validate and build on this research is provided with a description of how this body of knowledge may contribute to our understanding and management of glaucoma. P a g e | 15

LIST OF FIGURES

FIGURE 1. 1. SCHEMATIC DIAGRAM OF THE CANINE DRAINAGE ANGLE DEPICTING AQUEOUS HUMOUR OUTFLOW PATHWAYS IN THE DOG. AFTER PASSING BETWEEN PECTINATE LIGAMENT STRANDS, AQUEOUS HUMOUR PASSES THROUGH THE TRABECULAR MESHWORK TM), TO THE ANGULAR AQUEOUS PLEXUS (AAP), AND THEN THE SCLERAL VENOUS PLEXUS (SVP). THE AQUEOUS HUMOUR DRAINS (1) TO THE EPISCLERAL AND CONJUNCTIVAL VEINS, (2) INTO THE SCLERAL VENOUS PLEXUS AND VORTEX VENOUS SYSTEMS. THE ALTERNATIVE OUTFLOW PATHWAY (UVEOSCLERAL OUTFLOW) HAS AQUEOUS HUMOUR FLOW THROUGH THE CILIARY MUSCLE INTERSTITIUM TO THE SUPRACILIARY OR SUPRACHOROIDAL SPACES BETWEEN THE CILIARY BODY (CB), THROUGH THE SCLERA (S) AND INTO THE ORBIT. PS: POSTERIOR SEGMENT; L: LENS; PC: POSTERIOR CHAMBER; I: IRIS. (MODIFIED FROM MARTIN CL: GLAUCOMA, IN SLATTER D, EDITOR: TEXTBOOK OF SMALL ANIMAL SURGERY, ED 2, PHILADELPHIA, 1993, SAUNDERS.) ...... 41

FIGURE 1. 2. PHOTOGRAPHS OF PECTINATE LIGAMENT ABNORMALITIES (A) NORMAL PECTINATE LIGAMENT (PL) VISIBLE WHEN RAREFACTION OF THE INITIAL FIBRILLAR SHEET WAS ALMOST COMPLETE, BY 2 TO 4 WEEKS AFTER BIRTH, LEAVING STRANDS OF INTERTWINING COLLAGEN, PROGRESSIVELY ENCASED BY ATTENUATE TRABECULAR CELLS, CONFLUENT WITH THE ANTERIOR SURFACE OF THE IRIS (STAGE 0). (B- D) THE ENTIRE ICA AND OPENING OF THE CC WERE SYSTEMATICALLY EXAMINED FOR THE PRESENCE OF A PLA, DEFINED AS ABNORMALLY BROAD AND THICKENED PECTINATE LIGAMENT FIBRES (STAGE 1) (FROM BOILLOT T, ET AL. DETERMINATION OF MORPHOLOGICAL, BIOMETRIC AND BIOCHEMICAL SUSCEPTIBILITIES IN HEALTHY EURASIER DOGS WITH SUSPECTED INHERITED GLAUCOMA. PLOS ONE 2014; 9(11): PE111873...... 56

FIGURE 1. 3. ASSESSING THE IRIDOCORNEAL ANGLE USING . TOP SCHEMATIC DIAGRAM OF GRADING SYSTEM PROPOSED IN SAMOYEDS. AN ESTIMATE IS MADE OF THE RATIO OF THE CILIARY CLEFT OPENING WIDTH (A) TO THE DISTANCE FROM THE PECTINATE LIGAMENT TO THE ANTERIOR CORNEAL SURFACE (B) (FROM EKESTEN B, NARFSTROM K. CORRELATION OF MORPHOLOGIC FEATURES OF THE IRIDOCORNEAL ANGLE TO INTRAOCULAR PRESSURE IN SAMOYEDS. AM J VET RES 1991; 52:1875)...... 57

FIGURE 2. 1. BOXPLOT DEMONSTRATING LOWER VISION IMPAIRMENT SCORES (VIS) ASSOCIATED WITH INCREASING NUMBER OF SIGHTED EYES IN DOGS. THE HORIZONTAL LINE WITHIN THE BOX REPRESENTS MEAN SCORE OUT OF A MAXIMUM POSSIBLE SCORE OF 67...... 79

FIGURE 2. 2. VISION IMPAIRMENT AND SUBCATEGORY SCORES IN DOGS ACCORDING TO NUMBER OF SIGHTED EYES. STATISTICALLY SIGNIFICANT DIFFERENCE BETWEEN DOGS WITH 0, 1 AND 2 SIGHTED EYES IN ALL CATEGORIES EXCEPT FOR OCULAR PAIN (P=0.246). FOR ALL CATEGORIES, INCLUDING OVERALL VIS, EXCEPT FOR OCULAR PAIN HIGHER SCORES (INDICATING INCREASED IMPAIRMENT) WERE ASSOCIATED WITH DECREASING NUMBER OF SIGHTED EYES. THE GENERAL HEALTH QUESTION IS ILLUSTRATED, THOUGH NOT USED IN CALCULATION OF THE VIS ...... 79

P a g e | 16

FIGURE 2. 4. VISION IMPAIRMENT SCORE IN GROUPS WITH KNOWN CHARACTERISTICS (‘KNOWN GROUPS’) ...... 82

FIGURE 2. 5. PLOTS OF SCORES FOR FOUR DOMAINS OF HRQL (ENERGETIC/ENTHUSIASTIC, HAPPY/CONTENT, ACTIVE/COMFORTABLE, CALM/RELAXED) GENERATED BY OWNERS OF 25 BLIND DOGS, 28 DOGS WITH SIGHT IN 1 EYE AND 43 DOGS WITH SIGHT IN BOTH EYES, USING THE WEB-BASED VETMETRICA GENERIC HRQL QUESTIONNAIRE INSTRUMENT FOR DOGS. EACH BLUE BOX REPRESENTS THE SCORES OBTAINED FOR BETWEEN 25% (BOTTOM LINE) AND 75% (TOP LINE) OF THE GROUP WITH THE LINE IN THE MIDDLE REPRESENTING THE MEDIAN SCORE...... 83

FIGURE 2. 6. THE RELATIONSHIP BETWEEN THE HRQL PROFILE AND VISION IMPAIRMENT SCORE FOR 96 DOGS WITH VISION IMPAIRMENT, BUT NO CO-MORBIDITIES. PEARSON’S CORRELATION COEFFICIENTS FOR ENERGETIC/ENTHUSIASTIC, HAPPY/CONTENT, ACTIVE/COMFORTABLE AND CALM/RELAXED ARE - 0.61, -0.58. -0.45 AND -0.30 RESPECTIVELY ...... 84

FIGURE 3. 1. RELATIVE DIFFERENCES IN THE SIZE OF OPTOTYPES AND THEIR CORRESPONDING VISUAL ACUITY VALUES, AS USED IN THE ASSESSMENT OF HUMAN OBSERVER RELIABILITY...... 91

FIGURE 3. 2. A) SCHEMATIC DIAGRAM REPRESENTING STIMULUS PANEL USED TO ASSESS VISUAL ACUITY. THE POSITION OF THE GRATING STIMULUS (TOP LEFT OF SCREEN) AND HOMOGENOUS GREY STIMULUS (BOTTOM RIGHT OF SCREEN) WERE RANDOMLY GENERATED BY SOFTWARE. CONTRAST AND LUMINANCE ARE NOT MEASURED IN THESE REPRESENTATIVE DIAGRAMS; B) SAMPLE ANIMATED DIAGRAM DISPLAYED ON SCREEN (WITH ACCOMPANYING SOUND) TO RE-FOCUS DOG’S ATTENTION AT THE DISCRETION OF THE INVESTIGATOR (ISTOCK.COM/BANDIAN1122); C-D) ALTERED POSITIONS OF STIMULI ON PANEL DISPLAYING GRATINGS OF THE SAME SPATIAL FREQUENCY. THIS SEQUENCE WAS SELECTED BY THE INVESTIGATOR (UP ARROW KEY) WHEN THE INVESTIGATOR COULD NOT MAKE A ‘CHOICE’ ON WHERE/WHETHER THE DOG LOOKED AT THE STIMULUS, AND COULD BE REPEATED UNTIL THE INVESTIGATOR MADE A CHOICE...... 94

FIGURE 3. 3. BOX AND WHISKER PLOT SHOWING MEDIAN (RANGE) OF ESTIMATED VISUAL ACUITY IN DOGS. MEAN VA REPRESENTED BY CROSS (+)...... 96

FIGURE 3. 4. CORRELATION BETWEEN ESTIMATES OF VISUAL ACUITY AS MEASURED FROM 1-METRE (X-AXIS) AND 3-METRES (Y-AXIS) ...... 99

FIGURE 3. 5. SCHEMATIC OF FRONTAL VIEW DISPLAYING STIMULUS (AS PRESENTED TO DOG) SHOWING DISPLAY MONITOR IN FRONT OF A SCREEN CONCEALING THE INVESTIGATOR. THE INVESTIGATOR VIEWS THE DOG THROUGH A SMALL OPENING IN THE SCREEN (BLACK RECTANGLE)...... 102

FIGURE 3. 6. SCHEMATIC REPRESENTATION OF TESTING ROOM IN WHICH VISION ASSESSMENTS WERE CONDUCTING DEMONSTRATING LAYOUT OF APPARATUS AND PERSONNEL (INVESTIGATOR IN CHAIR BEHIND DESK USING LAPTOP WHICH CONTROLS THE STIMULUS PRESENTATION; HANDLER/OWNER RESTRAINING DOG ON STOOL AT 1M OR 3M TESTING DISTANCE) ...... 102 P a g e | 17

FIGURE 4. 1. DOG’S HEAD POSITIONED RELATIVE TO THE CAMERA SO THAT THE IRIS PLANE IS PARALLEL TO THE CAMERA...... 106

FIGURE 4. 2. MEASUREMENT OF RELATIVE (A-B) RESTING AND (C-D) STIMULATED PUPIL SIZE: THE RELATIVE SIZE OF THE PUPIL (OUTLINED IN RED) TO THE CORNEA (INDICATED BY THE PERIPHERAL IRIS MARGINS; OUTLINED IN GREEN) WAS DETERMINED BY CALCULATING THE PUPIL DIAMETER (YELLOW LINE) AS A PROPORTION OF THE CORNEAL DIAMETER (ORANGE LINE); (C-D) TO DETERMINE THE DEGREE OF PUPIL CONSTRICTION IN RESPONSE TO A LIGHT, THE SAME MEASUREMENTS WERE MADE, AND THE RESULTING RELATIVE PUPIL SIZE MEASURED AS A PROPORTION OF THE RESTING PUPIL SIZE...... 107

FIGURE 4. 3. UNSUITABLE IMAGES FOR PUPILLOMETRY DUE TO A) EYELID/DAZZLE REFLEX OBSCURING >180° OF CORNEAL CIRCUMFERENCE, AND B) HEAD AND GLOBE OFF-CENTRE SO THAT IRIS PLANE IS NOT PARALLEL TO LIGHT SOURCE AND CAMERA ...... 109

FIGURE 4. 4. RELATIVE BASELINE PUPIL SIZES IN NORMAL, PREDISPOSED AND SARDS EYES WHEN STIMULATED WITH LOW AND HIGH INTENSITY RED LIGHT. THE PUPIL SIZE AFTER STIMULATION IS CALCULATED RELATIVE TO THE BASELINE PUPIL SIZE FOR EACH INDIVIDUAL. THE BOX REPRESENTS 95% CONFIDENCE INTERVAL WITH THE MEAN REPRESENTED BY A HORIZONTAL LINE. THE WHISKERS INDICATE THE RANGE OF RELATIVE PUPIL SIZES. THE GROUPS BETWEEN WHICH THERE WAS A STATISTICALLY SIGNIFICANT DIFFERENCE ARE INDICATED WITH THEIR RESPECTIVE P VALUE...... 111

FIGURE 4. 5. RELATIVE BASELINE PUPIL SIZES IN NORMAL, PREDISPOSED AND SARDS EYES WHEN STIMULATED WITH LOW AND HIGH INTENSITY BLUE LIGHT. THE PUPIL SIZE AFTER STIMULATION IS CALCULATED RELATIVE TO THE BASELINE PUPIL SIZE FOR EACH INDIVIDUAL. THE BOX REPRESENTS 95% CONFIDENCE INTERVAL WITH THE MEAN REPRESENTED BY A HORIZONTAL LINE. THE WHISKERS INDICATE THE RANGE OF RELATIVE PUPIL SIZES. THE GROUPS BETWEEN WHICH THERE WAS A STATISTICALLY SIGNIFICANT DIFFERENCE ARE INDICATED WITH THEIR RESPECTIVE P VALUE...... 112

FIGURE 4. 6. PUPIL CONSTRICTION AFTER STIMULATION WITH HIGH INTENSITY BLUE LIGHT IN (A) NORMAL, AND (B) PREDISPOSED EYES ...... 113

FIGURE 4. 7. RELATIVE BASELINE PUPIL SIZES IN NORMAL, PREDISPOSED AND SARDS EYES WHEN STIMULATED WITH WHITE LIGHT. THE PUPIL SIZE AFTER STIMULATION IS CALCULATED RELATIVE TO THE BASELINE PUPIL SIZE FOR EACH INDIVIDUAL. THE BOX REPRESENTS 95% CONFIDENCE INTERVAL WITH THE MEAN REPRESENTED BY A HORIZONTAL LINE. THE WHISKERS INDICATE THE RANGE OF RELATIVE PUPIL SIZES. THE GROUPS BETWEEN WHICH THERE WAS A STATISTICALLY SIGNIFICANT DIFFERENCE ARE INDICATED WITH THEIR RESPECTIVE P VALUE...... 114

FIGURE 5. 1. VARIABLES MEASURED (DEFINITIONS OUTLINED IN TABLE 1) OF A) THE IRIDOCORNEAL ANGLE INCLUDING THE [A] GEOMETRIC IRIDOCORNEAL ANGLE (ICA) (ANGLE FORMED BY RED LINES); [B-C] ANGLE OPENING DISTANCES (AOD1, PALE BLUE LINE AND AOD2, YELLOW LINE); [D] ANGLE RECESS AREA (ARA, ENCLOSED BY GREEN DOTTED LINES); B) THE CILIARY CLEFT (CC) INCLUDING [A] CC ENTRY (RED LINE); [B] CC LENGTH (YELLOW LINE); [C] MID-CC WIDTH (GREEN LINE); [D] CC WIDTH (PINK P a g e | 18

LINE); [E] CC AREA (ENCLOSED BY PALE BLUE DASHED LINE); C) MEASUREMENT OF THE AXIAL ANTERIOR CHAMBER DEPTH (ACD, RED LINE); AND D) GENERAL MEASUREMENTS WITHIN THE ANTERIOR CHAMBER INCLUDING [B] THE DISTANCE FROM THE END OF DESCEMET’S MEMBRANE (SCHWALBE’S LINE) TO THE ANTERIOR LENS CAPSULE (SLD, PURPLE LINE); [C] THE DISTANCE FROM THE LIMBUS TO THE FIRST CILIARY PROCESS (DLCP, GREEN LINE); [D] THE DISTANCE FROM THE END OF DESCEMET’S MEMBRANE TO THE IRIS (SID, RED SOLID LINE); AND [E] THE IRIDOCILIARY PROCESS DISTANCE (ICP, RED DOTTED LINE)...... 122

FIGURE 5. 2. SAMPLE IMAGES OF ULTRASOUND BIOMICROSCOPIC EXAMINATION OF THE DRAINAGE ANGLE IN THE SUPERIOR QUADRANT OF A A) NORMAL/HEALTHY CONTROL (CASE 5); AND AN EYE B) PREDISPOSED TO DEVELOPING GLAUCOMA WITH GONIODYSGENESIS (CASE 25) ...... 124

FIGURE 5. 3. MEASUREMENTS OF THE CILIARY CLEFT. THE P VALUES ARE OBTAINED FROM TWO-WAY ANOVA INVESTIGATING THE EFFECT OF GROUP (NORMAL/CONTROL VERSUS PREDISPOSED) ON EACH VARIABLE. ALL VARIABLES MEASURED IN MILLIMETRES (MM) ON THE LEFT Y-AXIS, EXCEPT FOR CILIARY CLEFT AREA, WHICH IS MEASURED ON THE RIGHT Y-AXIS...... 127

FIGURE 5. 4. MEASUREMENT OF VARIABLES REPRESENTING THE DRAINAGE ANGLE IN NORMAL (CONTROL) AND PREDISPOSED EYES. THE P VALUES ARE OBTAINED FROM TWO-WAY ANOVA INVESTIGATING THE EFFECT OF GROUP (NORMAL/CONTROL VERSUS PREDISPOSED) ON EACH VARIABLE. ICA = GEOMETRIC IRIDOCORNEAL ANGLE; ARA = ANGLE RECESS AREA; AOD = ANGLE OPENING DISTANCE...... 128

FIGURE 5. 5. BOX PLOT DEMONSTRATING MEDIAN (LINE WITHIN BOX) AND RANGE (WHISKERS) MEASUREMENTS OF ANGLE RECESS AREA (ARA), ANGLE OPENING DISTANCE 2 (AOD2), AND MID-IRIS THICKNESS (MID-IRIS) MEASUREMENTS IN NORMAL (RED BOXES WITH CROSSHATCHES) AND PREDISPOSED (BLUE BOX WITH DIAGONAL LINES) EYES USING RAW AND CORRECTED DATA. P-VALUES SHOWING STATISTICALLY SIGNIFICANT FINDINGS (P<0.004) ARE INDICATED IN RED...... 129

FIGURE 6. 1. (A) OPTIC NERVE HEAD LOCATED AT THE CENTRE OF THE SCAN (LEFT EYE) TO ALLOW COMPARISON OF MEASUREMENTS BETWEEN SUBJECTS; B) SCHEMATIC OVERLAY DEMONSTRATING THE SECTORS OF THE PERIPAPILLARY RETINA THAT WERE ANALYSED FOR ASSESSMENT OF RETINAL THICKNESS. RETINAL THICKNESS WAS ASSESSED IN QUADRANTS (SUPERIOR, NASAL, INFERIOR, TEMPORAL); EIGHT SECTORS OF THE RNFL WERE ASSESSED (SN SUPERIOR, NASAL; ST SUPERIOR, TEMPORAL; NU NASAL, UPPER; NL NASAL, LOWER; IN INFERIOR, NASAL; IT INFERIOR, TEMPORAL; TU TEMPORAL, UPPER; TL TEMPORAL, LOWER; SUP SUPERIOR; NAS NASAL; INF INFERIOR; TEMP TEMPORAL)...... 140

FIGURE 6. 2. SELECTION FOR APPROPRIATE SCAN QUALITY AND POSITION. A) RETINAL SCAN WITH A POOR SCAN QUALITY INDEX SHOWING DISTORTION OF THE EN FACE IMAGE TO THE LEFT (DARK REGIONS AT THE TOP AND BOTTOM EDGES OF THE IMAGE), INADEQUATE DEFINITION OF RETINAL LAYERS FOR MANUAL SEGMENTATION, AND THE RETINA IS NOT CONFINED TO THE OCT WINDOW. THICKNESS MAP OF CASE 1 SHOWING APPROPRIATE POSITIONING OF THE OPTIC NERVE HEAD (B) TAKEN AFTER DISCARDING THE INITIAL SCAN (C) DUE TO EXCESSIVE ROTATION OF THE GLOBE SHOWN WITH THE DORSAL RETINAL VEIN ORIENTED TO THE TEMPORAL QUADRANT. RETINAL CROSSLINE SCAN (SCAN P a g e | 19

ORIENTATION HORIZONTAL OR VERTICAL AS INDICATED BY THE ARROW IN THE TOP RIGHT CORNER) SHOWING (D) INAPPROPRIATE SCAN ANGLE, AND (E) REPEAT SCAN WITH APPROPRIATE POSITIONING...... 142

FIGURE 6. 3. SCHEMATIC REPRESENTATION OF WHERE MEASUREMENTS FOR DETERMINING RETINAL THICKNESS WERE OBTAINED (HIGHLIGHTED YELLOW REGIONS) WHEN THE CIRCLES ARE CENTRED OVER THE OPTIC NERVE HEAD. THE REGION OF EACH QUADRANT THAT WAS WITHIN THE INNER CIRCLE WAS EXCLUDED TO MINIMISE THE IMPACT OF THE OPTIC NERVE AND INTRAOCULAR MYELIN ON MEASURES OF RETINAL THICKNESS...... 142

FIGURE 6. 4. RETINAL LAYER SEGMENTATION ON B-SCAN IMAGE OF THE RETINA FROM CASE 1 (NORMAL EYE); A) RETINAL LAYERS IDENTIFIED FOR MANUAL SEGMENTATION OF SCANS TO MEASURE RETINAL THICKNESS INCLUDING INNER, OUTER AND TOTAL RETINA, NERVE FIBRE LAYER, GANGLION CELL COMPLEX. NFL = NERVE FIBRE LAYER; GCL = GANGLION CELL COMPLEX LAYER; IPL = INNER PLEXIFORM LAYER; INL = INNER NUCLEAR LAYER; OPL = OUTER PLEXIFORM LAYER; ONL = OUTER NUCLEAR LAYER; ELM = EXTERNAL LIMITING MEMBRANE; PR = PHOTORECEPTORS; RPE = RETINAL PIGMENTED EPITHELIUM; B) DEFINITION OF LAYERS FOR MEASUREMENT OF RETINAL THICKNESS: INNER RETINA (FROM INNER LIMITING MEMBRANE TO INNER PLEXIFORM LAYER), OUTER RETINA (FROM INNER PLEXIFORM LAYER TO RETINAL PIGMENTED EPITHELIUM)...... 143

FIGURE 6. 5. THREE OF THE 20 B-SCAN IMAGES FROM CASE 2 SHOWING MANUAL ADJUSTMENT OF THE LINES DEMARCATING THE BOUNDARIES OF THE INNER AND OUTER RETINA (RED ARROWS IN [A]; THE LEVEL AT WHICH RETINAL THICKNESS IS MEASURED IN EACH IMAGE IS DEPICTED IN THE EN FACE IMAGE TO THE RIGHT). WHERE THE OPTIC NERVE HEAD AND MYELIN WERE PRESENT, THE LINES WERE ALIGNED TO NEGATE ANY MEASURE OF THICKNESS IN THESE REGIONS (YELLOW ARROWS IN [B-C]). D) RESULTING THICKNESS MAP SHOWING THE REGION OF THE ONH AND INTRAOCULAR MYELIN IN BLACK, REFLECTING THE MANUAL ADJUSTMENTS. S= SUPERIOR; T= TEMPORAL; I= INFERIOR; N= NASAL. THE POORER IMAGE QUALITY EVIDENT IN FIG 5A-C IS PRESENTED HERE AS THE ONLY MANNER TO SHOW IMAGES OF THE MODIFIED INDIVIDUAL B-SCANS WAS TO TAKE PHOTOGRAPHS OF THE COMPUTER SCREEN AS THE ANALYSES WERE BEING CONDUCTED. POSITION OF TEMPORAL AND NASAL QUADRANTS VARIES DEPENDING ON WHETHER THE LEFT OR RIGHT EYE IS IMAGED...... 144

FIGURE 6. 6. OPTIC NERVE HEAD ANALYSIS. FOR MANUAL IDENTIFICATION OF THE OUTLINE OF THE ONH, 3D SCANS WERE ASSESSED. MARKERS (ON THE RIGHT AND BELOW EN FACE IMAGE) WERE MANOEUVRED ALONG THE 6MM MARGINS TO SHOW CROSS-SECTIONAL IMAGES OF THE RETINA IN THE HORIZONTAL AND VERTICAL PLANES AT THAT LEVEL. IDENTIFICATION OF THE TERMINATION OF THE RETINAL PIGMENTED EPITHELIUM AND INCREASED THICKNESS ASSOCIATED WITH MYELINATION COULD THEN BE MADE PRIOR TO MARKING THE ONH BOUNDARIES ON THE EN FACE IMAGE (CASE 22).145

FIGURE 6. 7. B-SCAN IMAGE (LEFT) AND CORRESPONDING THICKNESS MAP (RIGHT) DEMONSTRATING A THINNER TOTAL RETINAL THICKNESS IN A PREDISPOSED LEFT EYE (CASE 17, BOTTOM IMAGES) COMPARED TO A NORMAL RIGHT EYE (CASE 10, TOP IMAGES)...... 149

P a g e | 20

FIGURE 6. 8. SCATTERPLOT COMPARING MEASURES OF INNER RETINAL THICKNESS OF THE PERIPAPILLARY RETINA IN EACH QUADRANT SHOWING A STATISTICALLY SIGNIFICANTLY THINNER INNER RETINA IN PREDISPOSED (DIAMONDS) COMPARED TO NORMAL EYES (CIRCLES)...... 149

FIGURE 6. 9. IMAGES OF THE OPTIC NERVE HEAD AND RETINAL NERVE FIBRE LAYER OBTAINED FROM NORMAL (A. CASE 3, B. CASE 8) AND PREDISPOSED EYES (C. CASE 15, D. CASE 19). THE SOFTWARE WHICH AUTOMATICALLY DETECTS THE OPTIC CUP, FAILED TO DO SO IN SOME CASES (E.G. A, C). THE OPTIC CUP IS REPRESENTED IN LIGHT GREY IN B AND D...... 151

FIGURE 6. 10. SCATTERPLOT COMPARING MEASURES OF PERIPAPILLARY RETINAL NERVE FIBRE LAYER (RNFL) THICKNESS BETWEEN NORMAL (CIRCLES) AND PREDISPOSED (DIAMONDS) EYES IN EACH QUADRANT...... 151

FIGURE 6. 11. SEPARATE SCANS OF THE OPTIC NERVE HEAD/RETINAL NERVE FIBRE LAYER OBTAINED TO DETERMINE REPEATABILITY (CONSISTENCY) OF MEASUREMENTS (CASE 22, PREDISPOSED EYE). GOOD CONSISTENCY WAS SEEN WHEN COMPARING THE OPTIC DISC OUTLINE (DARK GREY) WHICH REQUIRED MANUAL ADJUSTMENTS, BUT THE CONSISTENCY OF AUTOMATED MEASURES SUCH AS THE OPTIC CUP (LIGHT GREY) WHICH COULD NOT BE MANUALLY ADJUSTED WAS OF LIMITED APPLICABILITY. ... 153

FIGURE 6. 12. ASSESSMENT OF INTRA-RATER RELIABILITY DETERMINED BY TAKING SEPARATE MEASUREMENTS (INCLUDING MANUAL ADJUSTMENTS) ON THE SAME OPTIC NERVE HEAD/RETINAL NERVE FIBRE LAYER SCANS WITH AN INTERVAL OF 6 MONTHS BETWEEN ANALYSES (CASE 6; NORMAL EYE). EXCELLENT RELIABILITY WAS IDENTIFIED FOR MEASURING OPTIC DISC AREA (DARK GREY), WHILE RELIABILITY OF AUTOMATED MEASURES OF OPTIC CUP (LIGHT GREY) WAS ADEQUATE...... 155

FIGURE 7. 1. MAGNETIC RESONANCE IMAGES DEMONSTRATING REGIONS OF INTEREST (ROI) PLACEMENT. A) OPTIC NERVE ROIS (PINK) IN TRANSVERSE PLANE, B) OPTIC CHIASM ROI (YELLOW) IN TRANSVERSE PLANE, C) OPTIC TRACT ROIS (BLUE) IN TRANSVERSE PLANE, D) LATERAL GENICULATE NUCLEI ROIS (GREEN) IN TRANSVERSE PLANE, E) OPTEIC NERVE (PINK), OPTIC CHIASM (YELLOW) AND OPTIC TRACT (BLUE) ROIS IN DORSAL PLANE AND F) LATERAL GENICULATE NUCLEI (GREEN) ROIS IN DORSAL PLANE. THE IMAGES ARE COMPOSED OF 3-DIMENSIONAL T1-WEIGHTED IMAGES OVERLAID WITH FRACTIONAL ANISOTROPY (FA) HEAT MAPS (RED = LOW FA, WHITE = HIGH FA)...... 164

FIGURE 7. 2. DIFFUSIVITY MAPS IN DORSAL (TOP) AND TRANSVERSE (BOTTOM) PLANE, GENERATED FOR (FROM LEFT TO RIGHT) FRACTIONAL ANISOTROPY (FA), MEAN DIFFUSIVITY (MD), RADIAL DIFFUSIVITY (RD), AND AXIAL DIFFUSIVITY (AD)...... 165

FIGURE 7. 3. INTRAOCULAR PRESSURE (MMHG) AT THE TIME OF IMAGING IN EYES DIAGNOSED WITH PRIMARY ANGLE CLOSURE GLAUCOMA COMPARED TO HEALTHY CONTROL EYES. THERE WAS NO STATISTICALLY SIGNIFICANT DIFFERENCE IN THE MEDIAN IOP OF DOGS WITH GLAUCOMA (MEDIAN 15MMHG, RANGE 3-32MMHG) COMPARED TO CONTROL DOGS (MEDIAN 15MMHG, RANGE 11- 16MMHG) (P=0.469)...... 170

FIGURE 7. 4. FRACTIONAL ANISOTROPY (FA) OF THE (A) OPTIC NERVE, (B) OPTIC TRACT, AND (C) LGN . 171 P a g e | 21

FIGURE 7. 5. AXIAL DIFFUSIVITY (AD) OF THE (A) OPTIC NERVE, (B) OPTIC TRACT, AND (C) LGN ...... 171

FIGURE 7. 6. RADIAL DIFFUSIVITY (RD) OF THE (A) OPTIC NERVE, (B) OPTIC TRACT, AND (C) LGN ...... 171

FIGURE 8. 1. REPRESENTATION OF PROTEINS UNIQUE TO THE TEARS OF DOGS WITH PRIMARY GLAUCOMA. ALL DOGS WITH PRIMARY OPEN ANGLE GLAUCOMA (POAG) WERE BEING TREATED WITH TOPICAL OCULAR HYPOTENSIVE MEDICATIONS AT THE TIME OF TEAR SAMPLING. THE EYES OF DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA (PACG) EITHER HAD PACG (AFFECTED) THAT WAS ALREADY BEING TREATED (TREATED) OR WAS NEWLY DIAGNOSED (UNTREATED), WHILE THE CONTRALATERAL (PREDISPOSED) EYE WAS ON TOPICAL OCULAR HYPOTENSIVE MEDICATION (TREATED), OR YET TO BE STARTED ON PROPHYLACTIC MEDICATION (UNTREATED)...... 182

FIGURE 8. 2. ILLUSTRATION OF THE NUMBER OF PROTEINS UNIQUE TO, AND COMMON BETWEEN DISEASE STATES WHEN COMPARING (A) DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA (PACG) AND HEALTHY CONTROLS; AND (B) DOGS WITH PRIMARY GLAUCOMA AND HEALTHY CONTROLS. NORM = NORMAL/HEALTHY CONTROLS; PD = PREDISPOSED TO PACG; PACG = PRIMARY ANGLE CLOSURE GLAUCOMA; POAG = PRIMARY OPEN ANGLE GLAUCOMA ...... 185

FIGURE 8. 3. PIE CHARTS SHOWING THE CELLULAR LOCATION OF PROTEINS IDENTIFIED IN (A) NORMAL EYES; (B) EYES PREDISPOSED TO PACG; (C) EYES WITH PACG; AND (D) EYES WITH POAG WERE SIMILAR BETWEEN DISEASE STATES. THE PERCENTAGE LISTED REPRESENTS THE NUMBER OF PROTEINS IN EACH CELLULAR LOCATION AS A PROPORTION OF ALL PROTEINS IDENTIFIED IN THAT GROUP...... 186

FIGURE 8. 4. RELATIVE EXPRESSION OF UBIQUITIN-40S RIBOSOMAL PROTEIN S27A (WHITE CIRCLES) AND COAGULATION FACTOR V (BLACK CIRCLES) IN THE TEARS OF EYES PREDISPOSED TO PRIMARY ANGLE CLOSURE GLAUCOMA. THE CIRCLES REPRESENT MEAN RELATIVE EXPRESSION AND THE LINES INDICATE 95% CONFIDENCE INTERVAL. P VALUES ARE LISTED FOR ALL COMPARISONS WHERE THERE WAS A STATISTICALLY SIGNIFICANT DIFFERENCE IN THE AMOUNT OF PROTEIN BETWEEN GROUPS...... 186

FIGURE 8. 5. NETWORK PATHWAY GENERATED BY INGENUITY PATHWAY ANALYSIS FOR THE PROTEINS IDENTIFIED IN THE TEARS OF DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA. THE TWO MAJOR NETWORKS IDENTIFIED BY INGENUITY PATHWAY ANALYSIS IN THIS DISEASE GROUP INCLUDED INFECTIOUS DISEASES (GREY LINES) AND INFLAMMATORY DISEASE (PINK LINES). THE NETWORK PATHWAY DEMONSTRATES DIRECT (SOLID LINE) AND INDIRECT RELATIONSHIPS (BROKEN LINE) BETWEEN MOLECULES. AN ARROW BETWEEN THE PROTEINS INDICATES THE RELATIONSHIP BETWEEN MOLECULES AS: ACTIVATION; CAUSATION; EXPRESSION; LOCALISATION; MEMBERSHIP; MODIFICATION; MOLECULAR CLEAVAGE; PHOSPHORYLATION; PROTEIN-DNA INTERACTIONS; PROTEIN-RNA INTERACTIONS; REGULATION OF BINDING; OR TRANSCRIPTION. A LINE WITH NO ARROWHEAD INDICATES THE RELATIONSHIP BETWEEN MOLECULES AS: CHEMICAL-CHEMICAL INTERACTIONS; CHEMICAL-PROTEIN INTERACTION; CORRELATION; PROTEIN-PROTEIN INTERACTION; OR RNA-RNA INTERACTION. INCREASING BRIGHTNESS OF COLOUR INDICATES INCREASED RELATIVE EXPRESSION OF THE MARKER. 188

FIGURE 9. 1. BAERVELDT-350MM2 DRAINAGE DEVICE ...... 198

P a g e | 22

FIGURE 9. 2. THE DORSAL (A) AND LATERAL (B) RECTUS MUSCLES ARE IDENTIFIED AND ISOLATED; (C) A MMC-SOAKED SPONGE IS USED TO SOAK THE SCLERAL BED AT THE SURGEON’S DISCRETION; (D) THE IMPLANT IS PREPARED WITH AN INTRALUMINAL SUTURE AND EXTRALUMINAL LIGATURE BEFORE BEING PLACED BENEATH THE RECTUS MUSCLES; (E) THE IMPLANT IS SECURED TO THE SCLERA...... 199

FIGURE 9. 3. STAB INCISIONS ARE MADE THROUGH THE TUBING ANTERIOR TO THE EXTRALUMINAL LIGATURE...... 200

FIGURE 9. 4. ENTRY INTO THE ANTERIOR CHAMBER USING A 23-GAUGE NEEDLE UNDER A SCLERAL FLAP. 200

FIGURE 9. 5. A) LEFT: MATTRESS SUTURES MAY BE USED TO ANCHOR THE TUBING TO THE SCLERA WITH CAUTION NOT TO DIRECT THE TIP OF THE TUBING TOWARD THE CORNEAL ENDOTHELIUM; B) RIGHT: TUBING POSITIONED WITHIN THE ANTERIOR CHAMBER WITHOUT CONTACTING CORNEAL ENDOTHELIUM OR IRIS...... 200

FIGURE 10. 1. COMPARISON OF POSTOPERATIVE OUTCOMES BETWEEN TREATMENT GROUPS EXPRESSED AS A PROPORTION OF THE DOGS IN THAT TREATMENT GROUP. *OUTCOMES WITH A STATISTICALLY SIGNIFICANT DIFFERENCE BETWEEN TREATMENT GROUPS; **NOT INCLUDING DE NOVO GLAUCOMA SURGERY...... 215

FIGURE 10. 2. INTRAOCULAR PRESSURES (MEDIAN, MEAN, RANGE) AT FOLLOW UP POINTS DURING THE POSTOPERATIVE PERIOD...... 216

P a g e | 23

LIST OF TABLES

TABLE 1. 1. SYSTEMS USED TO CLASSIFY CANINE GLAUCOMA ...... 33 TABLE 1. 2. HISTOLOGIC FEATURES OF THE FILTRATION ANGLE SUGGESTIVE OF PLD ...... 38 TABLE 1. 3 FACTORS THAT CAUSE DECREASED AQUEOUS HUMOUR SECRETION ...... 40 TABLE 1. 4. USE OF MOLECULAR AND BIOCHEMICAL BIOMARKERS IN CANINE PRIMARY GLAUCOMA ... 45 TABLE 1. 5. GENETIC MUTATIONS ASSOCIATED WITH CANINE PRIMARY GLAUCOMA ...... 46 TABLE 1. 6. TONOMETERS RECOMMENDED FOR USE IN VETERINARY PRACTICE...... 55 TABLE 1. 7. CARBONIC ANHYDRASE INHIBITORS USED IN THE TREATMENT OF CANINE GLAUCOMA ..... 61 TABLE 1. 8. PROSTAGLANDIN ANALOGUES IN DOGS ...... 62 TABLE 1. 9. PARASYMPATHOMIMETICS USED IN CANINE GLAUCOMA ...... 63 TABLE 1. 10. BETA-BLOCKERS IN CANINE GLAUCOMA ...... 64 TABLE 1. 11. HYPEROSMOTIC AGENTS IN CANINE GLAUCOMA ...... 65 TABLE 1. 12. PROPHYLACTIC GLAUCOMA MEDICATIONS IN EYES PREDISPOSED TO PRIMARY GLAUCOMA ...... 66 TABLE 1. 13. GLAUCOMA DRAINAGE DEVICES USED IN DOGS ...... 68

TABLE 2. 1. SIGNALMENT AND BASELINE CHARACTERISTICS OF DOGS GROUPED ACCORDING TO THE NUMBER OF VISUAL EYES ...... 76 TABLE 2. 2. CHARACTERISTICS OF DOGS CLASSIFIED WITH ‘LOW VISION’ ...... 77 TABLE 2. 3. QUESTIONNAIRE ITEMS ...... 80 TABLE 2. 4. FACTOR ANALYSIS: SORTED UNROTATED FACTOR LOADINGS AND COMMUNALITIES ...... 81 TABLE 2. 5. RESPONSIVENESS OF IMPAIRMENT SCORE TO SURGICAL INTERVENTION FOR VISION IMPROVEMENT OR RESTORATION ...... 83

TABLE 3. 1. CRITERIA FOR INVESTIGATOR DECISION MAKING IN PROOF OF CONCEPT TRIALS (DETERMINING INTER- AND INTRA-RATER RELIABILITY IN A FORCED-CHOICE LOOKING TASK) ...... 93 TABLE 3. 2. CHARACTERISTICS AND BASELINE DATA OF SUBJECTS ...... 97 TABLE 3. 3. HIGHEST ESTIMATED VISUAL ACUITY RECORDED FOR EACH DOG EXPRESSED IN COMMONLY REPORTED NOTATIONS...... 98

TABLE 4. 1. COMPARISON OF SUBJECT CHARACTERISTICS IN DOGS WITH NORMAL EYES, EYES PREDISPOSED TO PRIMARY ANGLE CLOSURE GLAUCOMA AND THE EYES OF DOGS DIAGNOSED WITH SUDDEN ACQUIRED RETINAL DEGENERATION SYNDROME (SARDS) ...... 108 TABLE 4. 2. RELATIVE PUPIL DIAMETER AFTER STIMULATION ...... 109

TABLE 5. 1. DEFINITION OF VARIABLES MEASURED ...... 121 TABLE 5. 2. INDIVIDUAL SUBJECT CHARACTERISTICS ...... 125 TABLE 5. 3. MEASUREMENTS OF THE CILIARY CLEFT IN EYES PREDISPOSED TO GLAUCOMA AND HEALTHY CONTROLS ...... 126 P a g e | 24

TABLE 5. 4. ANGULAR MEASUREMENTS ASSOCIATED WITH THE IRIDOCORNEAL ANGLE IN EYES PREDISPOSED TO GLAUCOMA AND HEALTHY CONTROLS ...... 130 TABLE 5. 5. GENERAL ANTERIOR SEGMENT MEASURES IN EYES PREDISPOSED TO GLAUCOMA AND HEALTHY CONTROLS ...... 131

TABLE 6. 1. INDIVIDUAL SUBJECT CHARACTERISTICS ...... 137 TABLE 6. 2. CHARACTERISTICS OF DOG POPULATION STUDIED ...... 147 TABLE 6. 3. MEASURES OF RETINAL THICKNESS IN NORMAL AND PREDISPOSED EYES ...... 148 TABLE 6. 4. RETINAL NERVE FIBRE LAYER MEASUREMENTS...... 150 TABLE 6. 5. OPTIC NERVE HEAD ANALYSES ...... 152 TABLE 6. 6. GANGLION CELL COMPLEX ANALYSES ...... 152 TABLE 6. 7. INTRA-CLASS CORRELATION COEFFICIENTS FOR DETERMINATION OF REPEATABILITY (CONSISTENCY) AND INTRA-OBSERVER RELIABILITY...... 154 TABLE 6. 8. DESCRIPTION OF QUALITY AND PATTERN OF SCANS ...... 159

TABLE 7. 1. INDIVIDUAL SUBJECT CHARACTERISTICS ...... ERROR! BOOKMARK NOT DEFINED. TABLE 7. 2. AUTOMATED MEASURES OF CENTRAL CORNEAL THICKNESS...... ERROR! BOOKMARK NOT DEFINED. TABLE 7. 3. MEAN PERIPAPILLARY RETINAL THICKNESS (QUADRANTS) ...... ERROR! BOOKMARK NOT DEFINED. TABLE 7. 4. AUTOMATED MEASUREMENTS OF SECTORAL PERIPAPILLARY RETINAL NERVE FIBRE LAYER THICKNESS ...... ERROR! BOOKMARK NOT DEFINED.

TABLE 8. 1. CHARACTERISTICS OF INDIVIDUAL SUBJECTS ...... 168 TABLE 8. 2. DIFFUSION TENSOR IMAGING PARAMETERS OF THE VISUAL PATHWAY IN DOGS WITH AND WITHOUT PRIMARY ANGLE CLOSURE GLAUCOMA (PACG)...... 169

TABLE 9. 1. INDIVIDUAL SUBJECT CHARACTERISTICS ...... 180 TABLE 9. 2. PROTEINS IDENTIFIED WITH HIGHEST CONFIDENCE BY LABEL-FREE MASS SPECTROMETRY184 TABLE 9. 3. PROTEINS EXPRESSED MORE ABUNDANTLY IN EYES WITH PRIMARY ANGLE CLOSURE GLAUCOMA (PACG) COMPARED TO OTHER GROUPS ...... 187 TABLE 9. 4. DIFFERENTIALLY EXPRESSED PROTEINS IN THE TEARS OF EYES WITH PRIMARY OPEN ANGLE GLAUCOMA (POAG) ...... 189 TABLE 9. 5. HIGHEST RANKED UPSTREAM REGULATORS IDENTIFIED ON PATHWAY ANALYSIS OF EACH GROUP ...... 190

TABLE 10. 1. OUTCOME FOLLOWING BAERVELDT-350MM2 IMPLANTATION...... 202

TABLE 11. 1. BASELINE DEMOGRAPHIC AND OCULAR CHARACTERISTICS ...... 213 TABLE 11. 2. POSTOPERATIVE VISION AND IOP OUTCOMES ...... 213 P a g e | 25

TABLE 11. 3. DAILY DOSES OF GLAUCOMA MEDICATION ...... 217 TABLE 11. 4. POSTOPERATIVE COMPLICATIONS...... 218 P a g e | 26

LIST OF ABBREVIATIONS

AC Anterior chamber ACD Anterior chamber depth AD Axial diffusivity AH Aqueous humour AMP Amplified fragment length polymorphism ANCOVA Analysis of covariance ANOVA Analysis of variance AOD Angle opening distance ARA Angle recess area ARVO Association for Research in Vision and Ophthalmology ATP Adenosine triphosphate CA Carbonic anhydrase CAI Carbonic anhydrase inhibitors CC Ciliary cleft CI Confidence interval

C02 Carbon dioxide COGS Canine ocular gliovascular syndrome CRP C-reactive protein CT Corneal thickness CVI Content validation index DLCP Distance from the limbus to the first ciliary process DM Descemet’s membrane DNA Deoxyribonucleic acid DTI Diffusion tensor imaging DTT Dithiothreitol ECM Extracellular matrix ECP Endocyclophotocoagulation P a g e | 27

ERG Electroretinography ES Effect size FA Factor analysis FA Fractional anisotropy FDR False discovery rate fERG Flash electroretinography FPL Forced-choice preferential looking FS Female spayed GCC Ganglion cell complex GD Goniodysgenesis GDD Glaucoma drainage device GON Glaucomatous optic neuropathy GLMM Generalised linear mixed model GP Glutathione peroxidase GWA Genome-wide association H+ Hydrogen ion

H20 Water

- HC03 Bicarbonate HRQL Health-related quality-of-life HRUS High resolution ultrasound IAA Iodoacetamide ICA Iridocorneal angle ICC Intraclass correlation coefficient IL4 Interleukin 4 ILM Inner limiting membrane IM Intramuscular IOP Intraocular pressure IPA Ingenuity Pathway Analysis IPL Inner plexiform layer P a g e | 28 ipRGC Intrinsically photosensitive retinal ganglion cell IV Intravenous K+ Potassium ion INL Inner nuclear layer LGN Lateral geniculate nucleus LIU Lens induced uveitis logMAR Logarithm of the minimum angle of resolution LTBP2 Latent transforming growth factor beta binding protein 2 MD Mean diffusivity MDA Malondialdehyde MMC Mitomycin-C mm Millimetres mmHg Millimetres of Mercury MMP Matrix metalloproteinase MN Male neutered Mth Month Na+ Sodium ion NPE Non-pigmented ciliary epithelium NT Nitrotyrosine OCT Optical coherence tomography ONH Optic nerve head ONL Outer nuclear layer OSD Ocular surface disease PACG Primary angle closure glaucoma PAS Peripheral anterior synechia PBVG Petit Basson Griffin Vendéen PCG Primary congenital glaucoma pERG Pattern electroretinography PGA Prostaglandin analogue P a g e | 29

PIFM Pre-iridal fibrovascular membrane PLD Pectinate ligament dysplasia PL Preferential looking PLL Primary lens luxation PLR Pupillary light reflex PNAG Primary narrow angle glaucoma PO Per os POAG Primary open angle glaucoma PSM Peptide spectra matched QOL Quality of life RAPD Random amplified polymorphic DNA RD Radial diffusivity RFLP Restricted fragment length polymorphisms RGC Retinal ganglion cell RNFL Retinal nerve fibre layer ROI Region of interest RPE Retinal pigmented epithelium SARDS Sudden Acquired Retinal Degeneration Syndrome SD-OCT Spectral domain-optical coherence tomography SNP Single-nucleotide polymorphism SID Distances from Schwalbe’s line to the iris surface SLD Distance from Schwalbe’s line to the anterior lens capsule SQI Scan quality index SRBD1 S1 RNA binding domain 1 SRM Standardised response mean SSR Simple sequence repeats ST Scleral thickness STT Schirmer tear test TGF Transforming growth factor P a g e | 30

TGFβ Transforming growth factor beta TID Three times daily TM Trabecular meshwork TNF Tumour necrosis factor tPA Tissue plasminogen activator TrkB Tyrosine kinase receptor B TSCP Trans-scleral cyclophotocoagulation TSD Transsynaptic degeneration UBM Ultrasound biomicroscopy US United States VA Visual acuity VAS Visual Ability Score VIS Vision Impairment Score P a g e | 31

CHAPTER ONE AIM OF THE THESIS AND LITERATURE REVIEW

In this project we aim to identify a systematic paradigm for the assessment of glaucoma in dogs. With a standardised approach to assessment in humans, establishment of an approach in dogs would allow the dog to fulfil a need for an animal model to further advance our understanding of the disease. To determine whether dogs can be used as an effective model of disease of glaucoma in human patients, we must determine how to monitor disease progression and/or response to treatment objectively and determine similarities and differences between the species. Clinical trials are the standard scientific method for assessing the benefits and risks of new therapeutic interventions. ‘Hard’ or ‘true’ endpoints in phase III clinical trials are clinical events relevant to the patient.2 In glaucoma, endpoints are therefore vision loss that affects quality of life. In the context of glaucoma, true endpoints would be significant loss of vision with decrease in quality of vision or quality of life, or development of functional disability.3 As these factors are difficult to identify and measure in companion animals, identification of other indicators of disease – which have potential to act as surrogate endpoints - are of interest. The development and use of validated surrogate endpoints in clinical trials has the potential to offer shorter, less expensive trials with reduced sample size requirements while allowing observation of a greater number of endpoints during follow-up than could be achieved with observation of a true endpoint.3 There are no perfect therapies for prevention of vision impairment in glaucomatous dogs.4,5 Although IOP has traditionally been used as an endpoint in clinical trials, it is an imperfect surrogate for the clinically relevant outcomes of the disease3,6 and has never been validated as a surrogate for any class of IOP-lowering medications (i.e. the IOP lowering effect of the drug reduces vision loss). Disease in human and canine glaucoma patients can progress despite the maintenance of a low IOP7,8 and other cases remain stable despite having IOP measurements that are consistently high9 so the predictive power of IOP to screen for PACG is poor.10 Combining structural and functional surrogate endpoints has shown significantly better abilities to diagnose, stage and detect disease progression compared to isolated structural or functional testing.11,12 We will investigate several clinically relevant potential biomarkers in canine glaucoma. We will look to correlate clinically measurable physical parameters (intraocular pressure, gonioscopy, navigational vision) with imaging techniques and molecular markers to objectively measure disease progression and determine if and when glaucoma and its progression can be detected earlier and more accurately than it currently is. We will be monitoring and managing dogs through all stages of disease, including with the use of a novel bioimplant to treat dogs with glaucoma refractory to medical management. P a g e | 32

LITERATURE REVIEW What is glaucoma? Glaucoma is a term that describes a group of ocular neurodegenerative disorders characterised by progressive retinal ganglion cell (RGC) death and a characteristic glaucomatous optic neuropathy (GON)13-17 due to subsequent loss of optic nerve axons with decrements in visual function.18 Sometimes called ‘the ’ because the disease is not a single entity, glaucoma is one of the leading causes of blindness in humans and veterinary patients worldwide.1,19 Glaucoma is a heterogeneous disease which is usually associated with defects within the trabecular meshwork (TM) and anterior chamber that lead to obstruction of aqueous humour outflow, elevation of intraocular pressure (IOP) and progressive retinal ganglion cell death and optic nerve degeneration.20 The complexities of glaucoma and our incomplete understanding of its aetiopathogenesis make a specific and well-recognised definition of glaucoma difficult to ascertain. To outline the defining characteristics of glaucoma, Casson et al16 formally define the disease as:

“a group of ocular disorders of multi-factorial aetiology united by a clinically characteristic optic neuropathy with potentially progressive, clinically visible changes at the optic nerve head (ONH), comprising focal or generalized thinning of the neuroretinal rim with excavation and enlargement of the optic cup, representing neurodegeneration of retinal ganglion cell axons and deformation of the lamina cribrosa; corresponding diffuse and localized nerve-fibre- bundle pattern loss may not be detectable in early stages; while visual acuity is initially spared, progression can lead to complete loss of vision; the constellation of clinical features is diagnostic.” Based on reports describing pathophysiological changes associated with canine glaucoma,21-24 in dogs glaucoma is defined as13:

“the final common pathway of a group of diseases with increased IOP, decreased RGC sensitivity and function, retinal ganglion cell death, optic nerve axonal loss and concurrent ONH cup enlargement, incremental reduction in visual fields and blindness.” Normotensive glaucoma (GON in the absence of an elevated IOP) is an established entity in humans25 and an elevated IOP does not always result in GON. Elevations in IOP have therefore become a major risk factor, rather than a defining characteristic, of glaucoma in human glaucoma patients in the later half of 20th century.16 Until P a g e | 33 recently, the definition of glaucoma in veterinary medicine has been based on an elevation in IOP that resulted in structural and functional derangements within the eye,26 but this definition has changed to recognise an elevated IOP as a constant risk factor, rather than a sole definition feature.13 Whether normotensive glaucoma does occur in dogs remains unknown. Our inability to detect the presence of glaucoma before the presence of advanced disease is likely related to several factors commonly encountered in veterinary medicine (non-verbal patients, reliance on owner observations and willingness/ability to seek veterinary attention, subjective nature of pain assessment etc) as well as our limited understanding of the disease. The combined prevalence of the canine glaucomas over the past decade is 1.7%,13 which is comparable to the estimated 1-2% prevalence of glaucoma in humans.27 Canine spontaneous glaucoma has been documented as a better animal model for human glaucoma, compared to the rat, rabbit or monkey.28

Classification Due to the varieties of genetic, environmental, structural and functional factors that interact to result in glaucoma, classification of the disease can be made based on aetiology, iridocorneal angle morphology or stage of the disease (Table 1.1).

Table 1. 1. Systems used to classify canine glaucoma

Aetiology Gonioscopic appearance Stage of disease Congenital Open-angle Early noncongestive Primary Narrow-angle Acute or congestive Secondary Closed angle Chronic/end stage

Aetiology Congenital glaucoma Primary congenital glaucoma (PCG) in dogs develops relatively early in life (typically <1 year of age) and is associated with severe abnormalities of the ICA and often other overt anterior segment anomalies.13,29 The inheritance of congenital glaucoma in dogs is not reported though in humans PCG can be sporadic or inherited in an autosomal recessive pattern.30 Primary glaucoma Primary glaucoma is a heritable, breed-related disease with bilateral potential and is a leading cause of blindness in purebred dogs.13,31-38 In dogs, primary glaucoma is most commonly associated with pectinate ligament dysplasia (PLD).13 A higher prevalence P a g e | 34 of primary glaucoma has been reported in several breeds of dog and the incidence of disease has progressively increased with time.32 While primary glaucoma occurs with relative frequency in dogs,32 the genetics have been studied in only a small number of breeds39 and there is limited information known regarding the mode of inheritance.13 Recent advances in the genetics of canine glaucomas have provided a better understanding of molecular and cellular disease mechanisms as well as the resulting clinical signs.39 There is currently no evidence that any of the prominent glaucoma genes associated with human POAG and PACG contribute to the development of canine primary glaucoma.39 Primary open angle glaucoma (POAG) While POAG is the most frequent form of glaucoma in humans,1 in the dog PACG is recognised more frequently than POAG but varies by breed.40 PACG is the second most common form of glaucoma after POAG in humans and is three times more prevalent than POAG in Chinese, Asian Indian and Eskimo populations.41,42 In dogs with POAG, an open ICA narrows as the disease progresses and there is a gradual increase in IOP that eventually leads to retinal ganglion cell death, optic nerve atrophy, and irreversible blindness.43 In the later stages of disease, narrowing of the ciliary cleft contributes to IOP elevation and may lead to secondary lens subluxation. Vision deterioration progresses with disease progression due to optic nerve atrophy and cupping. The retina appears funduscopically normal until the late stages of the disease. Other clinical signs include Haab’s striae, cataract, and buphthalmos.44 In dogs, POAG is a relatively rare condition although its high prevalence in specific breeds suggests an inherited aetiology. Affected breeds include the ,43,45,46 Norwegian Elkhound,44,47,48 and the Petit Basset Griffon Vendeen.46,48-52 In the Beagle and Norwegian Elkhound POAG is an autosomal recessive condition caused by two separate mutations in ADAMTS10.48,50,52 and in the Petit Basset Griffon Vendeen, it is caused by a mutation in ADAMTS17.49

Primary angle closure glaucoma (PACG) In PACG there is anatomical iridotrabecular contact, potentially involving multiple mechanisms, and PACG always involves an elevation in IOP at some stage.16 The vast majority of dogs with PACG have PLD, though only a small proportion of dogs with PLD develop PACG.53 Similar to human PACG, and in contrast to the tendency in POAG, female dogs are twice as likely to be affected as male dogs.1,13,29,32,54-57 The reasons for this difference are not well understood; it is possible that a shorter axial globe length in female compared with male dogs may result in a higher susceptibility for IOP to increase.58 Furthermore, female dogs tend to have a narrower ICA opening, which may predispose them to angle closure.57

A large number of breeds are predisposed to PACG, with the highest prevalence found in breeds such as the American Cocker Spaniel,32,59-61 , 32,62-67 Chow Chow,33 Siberian Husky,32,68 Shiba Inu,69-72 Shih Tzu,71,72, Magyar Vizsla, and P a g e | 35

Newfoundland.13,29,32,55,69 The disease has been studied most thoroughly in the American and ,59,60,64,65,67,68,73-75 Basset Hound,40,65,67,73-78 English and Welsh Springer Spaniels,54,79 Bouvier des Flandres,64,80 Chow Chow,33,68 Siberian husky,68 Great Dane, 81,82 Flat-coated Retriever,83-85 Samoyed,86-88 Shiba Inu,69-72 Shih Tzu,71,72 Eurasier dog (created from Chow Chow and Samoyed),58 ,89 and the Golden Retriever.90

Secondary glaucoma Secondary glaucoma tends to be unilateral and occurs when there is an increase in IOP due to a pre-existing or concomitant (ocular or systemic) disease that alters aqueous humour dynamics.38,91,92 The clinical and pathologic reports of the prevalence of secondary glaucomas in the dog are limited.40 While some of the conditions initiating glaucoma in these cases may be genetic (e.g. cataracts, primary lens luxation), secondary glaucoma is not inherited.13 Anterior and posterior synechiae, formation of preiridal fibrovascular membranes, lens dislocation and pupillary block, and cellular infiltrations (neoplastic or inflammatory) of the trabecular meshwork are all involved in the pathogenesis of secondary glaucoma.91,92 Dogs with underlying goniodysgenesis may be more vulnerable to secondary angle occlusion than anatomically normal dogs.93

The most frequent causes for the secondary glaucomas include lens displacement, anterior uveitis including lens- or cataract-induced uveitis, hyphaema, primary and secondary intraocular neoplasia, postcataract surgery and trauma.13 Secondary glaucomas associated with cataract formation accounted for more than half of the total secondary glaucomas in one report.40 The reported relative prevalence of primary and secondary glaucoma varies from being similar to each other40 to a reported 1:1.78 ratio at one institution.29 The 0.80% prevalence of five different types of secondary glaucomas (cataract formation, lens displacement, following cataract surgery, uveitis, intraocular neoplasia and hyphaema) is comparable to the breed-related or primary glaucomas 0.89% in the same study population.32,40 In pathology reports, secondary glaucoma has been reported in 82% of 211 enucleated globes, and 87% of 167 glaucomatous eyes.34 Abnormal ocular pigment deposition and glaucoma Glaucoma associated with abnormal pigment deposition has been described in middle-aged to older Cairn Terriers,94,95 Boxers and Labrador Retrievers.96 Unlike in ‘pigment dispersal syndrome’ in humans, where pigment originating from the iris deposits within, and obstructs aqueous outflow through the trabecular network,97 in dogs it is suggested this occurs due to abnormally congested melanocytes.96,98 The reason for increased melanin and melanophage proliferation and the slow development of blindness after the onset of glaucoma is unclear. It is suggested that with a relatively low number of melanin-containing cells, initial rises in IOP are slow, allowing the optic nerve and retina to temporarily adapt by altering the perfusion to P a g e | 36 prevent severe ischaemia. Blindness ultimately occurs due to elevated IOPs over a prolonged period resulting in cupping of the optic disc and retinal degeneration.96 Uveitic glaucoma Glaucoma secondary to uveitis accounts for up to 45% of secondary glaucomas in dogs with 17% of dogs with uveitis in one study developing glaucoma within 5 years of diagnosis.93 Potential mechanisms of glaucoma induction include obstruction of the pupil, ICA or TM during active uveitis and/or secondary to peripheral anterior synechia (PAS), pre- iridal fibrovascular membrane (PIFM), and posterior synechia formation.92 The prognosis for vision and for globe retention should be considered guarded in cases of uveitic glaucoma, though depending on the underlying aetiology, early and aggressive intervention and treatment of both the underlying cause and the resultant inflammation, may lead to better outcomes.92

Lens luxation In primary lens luxation (PLL) the lens is spontaneously displaced from its normal position within the patella fossa after breakdown of the lens zonules.99 Acute glaucoma may result through obstruction of the pupil or filtration angle.100 The heritability of PLL has been studied in certain breeds where it is suggested to be an autosomal recessive trait.101,102 with a high incidence of the disease due to the onset of clinical signs after breeding dogs have reproduced.99 Dogs with anterior lens luxation have been reported to be particularly at risk of secondary glaucoma,103 with PLL affecting dogs between 3 and 8 years of age and secondary lens luxations typically in dogs >8 years.101 High IOPs are reported with anterior and posterior as well as subluxated lenses.101,104 Intraocular neoplasia The true incidence of glaucoma secondary to intraocular tumours is unknown with many cases likely mis- or undiagnosed.40 One study reported an incidence of glaucoma secondary to an intraocular tumour in 3.5% of dogs presenting to North American Veterinary teaching hospitals over a 40 year period.40 Unilateral glaucoma has been diagnosed secondary to 28.8 – 42.9% of intraocular melanomas;105-107 and uni- and bilateral glaucoma occur secondary to lymphoma,108 and multiple myeloma.109 Of the globes containing an intraocular mass submitted for histopathology, secondary glaucoma occurred in 20-40% of cases.40 Iridociliary cysts Solitary, thick-walled, round, variably pigmented, usually free-floating or ruptured cysts have been documented in many dog breeds, and are considered to have little clinical significance.110 Angle closure glaucoma secondary to iridociliary cysts is reported in humans111,112 and may be inherited as an autosomal dominant condition.113 Uveal cysts have been suggested to not contribute to ocular pathology in dogs, requiring treatment only if large or numerous enough to impair vision,110 P a g e | 37 multiple attached, sometimes free-floating thin-walled ciliary epithelial cysts are reported in several breeds, and have been linked to the development of glaucoma in Golden Retrievers,38,114 Great Danes115 and American Bulldogs.116 There are several suggested mechanisms of glaucoma resulting from iridociliary cysts38 including inflammatory and morphologic changes resulting in glaucoma when there is existing goniodysgenesis,116 narrowing or closure of the ICA via a ‘posterior-pushing’ mechanism causing anterior displacement of the iris,38 pupil block secondary to the cysts or the complications they cause (e.g. membrane formation, synechia).38 Cataracts and cataract surgery Cataract formation, lens- induced uveitis (LIU) and surgical removal of cataracts in dogs are frequent causes of secondary glaucomas,40,92,117-119 and glaucoma is reportedly the most common postoperative complication resulting in enucleation or evisceration.120 though the frequency of glaucoma as a complication of cataracts varies with 5.1% to 16.8% of dogs developing glaucoma after cataract surgery and up to 20% of dogs with untreated cataracts developing glaucoma.40,121-123 The high prevalence of glaucoma secondary to cataract in the dog is probably related to the high prevalence of cataracts in this species,124

Some of the proposed mechanisms of postoperative glaucoma development include pigment dispersion, trabecular meshwork (TM) atrophy and ciliary cleft (CC) collapse, development of PIFMs or lens epithelial membranes, formation of peripheral anterior or posterior synechiae, pupil occlusion, prolapsed vitreous, and aqueous misdirection.13 Breed-related factors have been proposed to play a role in the development of postoperative glaucoma.123,125,126

Uveitis accounts for many cases of glaucoma in dogs with cataracts or following cataract surgery,92 though the aetiology of LIU remains unclear and may vary between dogs.127 Whilst phacolytic uveitis is very common in dogs with cataracts of all stages, approximately 50% of eyes that develop phacoclastic uveitis following lens capsule rupture develop secondary glaucoma.124

Hyphaema/trauma There are few reports of traumatic glaucoma in dogs which is described as infrequent in dogs.13 Of 16 dogs with hyphaema and retinal detachment or hypertensive retinopathy, secondary glaucoma occurred in 10 dogs (62.5%).128 While only 7.3% of 1593 dogs with hyphaema developed secondary glaucoma,40 recurrent or continuous haemorrhage may be associated with a higher risk. Canine ocular gliovascular syndrome (COGS) Occurring in middle to older aged dogs and with a reported predisposition in the Labrador Retriever, the aetiology of canine ocular gliovascular syndrome (COGS) is unknown.129 COGS is associated with intravitreal spindle cell aggregates, optic nerve or retinal neovascular membranes, intraocular haemorrhage, and neovascular glaucoma.129,130 Diagnosis is difficult and, with limitations in knowledge and P a g e | 38 awareness of the syndrome, may mean this condition remains under-reported. The progression to glaucoma is rapid and there is a poor response to conventional medical therapy. Potentially effective therapies could be similar to those described for human neovascular glaucoma. Based on vitreoretinal abnormalities identified in the contralateral eye and a single histologically confirmed bilateral case of COGS, careful evaluation and follow-up should be performed for the contralateral eye.129

Iridocorneal angle morphology The pectinate ligament (PL) is the anterior-most structure of the ICA, where 85% of aqueous humour drainage occurs in the dog via the conventional aqueous outflow pathway.131 Goniodysgenesis (GD) or pectinate ligament dysplasia (PLD) refers to a range of congenital abnormalities in the development of the pectinate ligament.13 It is a bilateral condition in which there is consolidation of the normally fine pectinate ligament strands into short, broad bands or sheets of tissue, with or without flow holes or fenestration of the ICA, originating during development of the eye.131,132 Although deeper tissues involved in the outflow pathways are of unknown status on gonioscopic examination, the persistence of broad sheets spanning the region of the pectinate ligament is associated with PACG and PNAG in dogs.13 Histologic features described in dogs diagnosed with PLD are presented in Table 1.2. Variations in the development of PLD between littermates and within eyes,133 and the finding of asymmetrical PLD134 highlights the importance of examining the entire drainage angle of both eyes.

PLD has been related to the development of primary narrow and closed- angle glaucomas in a number of breeds, including the Basset Hound,62 Bouvier des Flandres,80 Japanese Shiba Inu,72 Flat-Coated Retriever,83,84 American Cocker and Cocker Spaniel,67 English and ,54,79 Samoyed88 and Leonberger.134 Concurrent narrowing of the ICA has also been reported in affected breeds.79,88

The incidence of glaucoma is not directly associated with either the presence, nor the severity of PLD as glaucoma will not develop in all dogs with PLD,54,63,80,88 and increasing severity of PLD has been associated with ageing in the Basset Hound, English Springer Spaniel and Flat-Coated Retriever58,85 confirming the importance of other factors in the manifestation of disease.80 It remains unclear whether PLD and angle narrowing contribute directly to the development of primary glaucoma or serve as markers for disease at another level of the aqueous humour outflow pathways,79 and use of the term ‘dysplasia’ is queried because in theory, dysplastic tissue must have turnover and undergo remodelling for the dysplasia to resolve.15 It has been suggested that age-related pigment dispersion with tissue remodelling, inflammation and fibrosis are eventually responsible for modification of the trabecular meshwork extracellular matrix and of the outflow pathways.135

Table 1. 2. Histologic features of the filtration angle suggestive of PLD P a g e | 39

Histologic features suggestive of PLD ICA - closed CC - closed Pectinate ligament not well defined Solid uveal tissue spanning from iris root to DM termination TM beams thickened Abnormal termination of DM DM extends along trabecular beams into CC Pigmented spindle cells extend into deep corneal stroma next to DM Uveal tissue extends onto the posterior surface of the peripheral DM Loss of posterior iris epithelium Adapted from132 and 34

Stage of disease The classification of glaucoma according to the stage of disease has more clinical than investigative research relevance.91 The early stages of disease are characterised by noncongestive glaucoma,91 with insidious onset and progression while clinical signs at this stage are potentially overlooked. Sudden and severe spikes in IOP (at times approximating 50-70 mmHg), are typical of acute congestive glaucoma; this presentation is common for PCAG, and it is responsible for the most dramatic clinical signs in affected patients.91 Chronic glaucoma is described in the advanced stages of the disease, when the extensive ocular damage almost invariably results in blindness. Specific clinical signs are characteristic of each described stage of glaucoma, and their identification aids in the choice of treatment and in offering a correct prognosis.91

Aqueous humour dynamics IOP results from a balance between aqueous humour production and its outflow.91 An equation to broadly define AH dynamics is the Goldmann equation:

Flow in = Flow out Rate of formation = Pressure gradient across entire outflow pathway X how easily fluid can exit F = (Pi – Pe) X C

F = rate of AH formation (µl/min) Pi = IOP (mmHg) Pe = episcleral venous pressure (mmHg) C = AH outflow facility (µl/min/mmHg)

However, additional factors as well as complex interactions that are not accounted for in the Goldmann equation do play a role in AH production and outflow.136 Factors P a g e | 40 that are not accounted for in the Goldmann equation, but which reduce the secretion of AH are outlined in Table 1.3.

Table 1. 3 Factors that cause decreased aqueous humour secretion

General Systemic Local Pharmacologic Surgical Age Hypotension Elevated IOP CAI Cyclodestruction Diurnal Hypothermia Uveitis β antagonist cycle Exercise Acidosis Retinal Opioid agonist detachment General Retrobulbar Plasma anaesthesia anaesthesia hyperosmolality Spironolactone Cardiac glycoside Modified from Stamper RL: Aqueous humor: secretion and dynamics. In Tasman W, Jaeger EA (eds): Clinical ophthalmology, vol 2, Philadelphia 1979, Lippincott.

Aqueous humour production AH production occurs in the ciliary body, and involves the processes of diffusion, ultra-filtration and active secretion. Diffusion and ultrafiltration account for approximately 10-20% and 1% of AH production respectively. Together they maintain a ‘reservoir’ of plasma ultrafiltrate. Active secretion is energy dependent and accounts for 80-90% of AH production. Active secretion across the non- pigmented ciliary epithelium results in water movement from the stroma into the posterior chamber, and is mediated by selective ion transport against a concentration gradient.136 Active secretion is relatively constant, being independent of IOP, but is reduced in sleep, while ultrafiltration decreases with increasing IOP.136 The main enzymes essential for the formation of AH are sodium-potassium-activated adenosine triphosphatase (Na+-K+-ATPase) and carbonic anhydrase inhibitor (CAI). Na+-K+-ATPase is located on the basolateral cell membrane of the NPE and results in active secretion of Na+ into the posterior chamber. It is accompanied by Cl- to maintain - 91,136 electroneutrality, and HCO3 can then enter the AH via exchange with chloride. - Carbonic anhydrase is responsible for the formation of HCO3 and its active transport 91 - + across the ciliary epithelium, catalysing the reaction H2O + CO2 -> HCO3 + H . Carbonic anhydrase can account for 60%137 and up to 75% of AH production in dogs.138

Aqueous humour outflow From the posterior chamber, AH flows through the pupil to leave the eye via the P a g e | 41 conventional or uveoscleral (nonconventional) outflow pathway (Fig.1.1).

Figure 1. 1. Schematic diagram of the canine drainage angle depicting aqueous humour outflow pathways in the dog. After passing between pectinate ligament strands, aqueous humour passes through the trabecular meshwork TM), to the angular aqueous plexus (AAP), and then the scleral venous plexus (SVP). The aqueous humour drains (1) to the episcleral and conjunctival veins, (2) into the scleral venous plexus and vortex venous systems. The alternative outflow pathway (uveoscleral outflow) has aqueous humour flow through the ciliary muscle interstitium to the supraciliary or suprachoroidal spaces between the ciliary body (CB), through the sclera (S) and into the orbit. PS: posterior segment; L: lens; PC: posterior chamber; I: iris. (Modified from Martin CL: Glaucoma, in Slatter D, editor: Textbook of small animal surgery, ed 2, Philadelphia, 1993, Saunders.)

Conventional outflow Anatomically, the conventional outflow occurs through the TM. Aqueous humour flows through the pectinate ligament strands into the ciliary cleft. Within the CC, AH passes through the uveal TM, then the corneo-scleral TM, and then through the juxtacanalicular meshwork, which is closely associated with the collector vessels of the angular aqueous plexus. 139 Transport across the juxtacanalicular meshwork is pressure dependent, and AH enters collector channels of the intrascleral venous plexus via the angular aqueous plexus. Aqueous then enters scleral and choroidal veins. Aqueous outflow is pressure dependent with increasing outflow with higher pressures until such pressure is reached that there is collapse of the TM.

Unconventional (uveoscleral) outflow AH drainage via the uveoscleral outflow occurs mostly via the posterior aspect of the uveal meshwork, through the ciliary muscle, along the supraciliary-suprachoroidal spaces and the vortex veins. It also includes AH outflow via the cornea and iris (negligible), retina (small, due to pumping capacity of the RPE). Approximately 15% of P a g e | 42

AH in the dog drains via the uveoscleral pathway, and is influenced by ciliary body muscle contraction and differences in hydrostatic pressure between the anterior chamber and the suprachoroidal spaces.140 P a g e | 43

GENETIC AND BIOCHEMICAL BIOMARKERS IN PRIMARY CANINE GLAUCOMA The following is the re-formatted manuscript published by John Wiley and Sons Ltd: Graham, K.L., McCowan, C. & White, A. Genetic and biochemical biomarkers in canine glaucoma. Veterinary Pathology 2017; 54(2): 427-434. DOI: 10.1177/0300985816666611

Abstract In many health-related fields there is great interest in the identification of biomarkers that distinguish diseased from healthy individuals. In addition to identifying the diseased state, biomarkers have potential use in predicting disease risk, monitoring disease progression, evaluating treatment efficacy, and informing pathogenesis. This review details the genetic and biochemical markers associated with canine primary glaucoma. While there are numerous molecular markers (biochemical and genetic) associated with glaucoma in dogs, there is no ideal biomarker that allows early diagnosis and/or identification of disease progression. Genetic mutations associated with canine glaucoma include those affecting ADAMTS10, ADAMTS17, Myocilin, Nebulin, COL1A2, RAB22A, and SRBD1. With the exception of Myocilin, there is very limited crossover in genetic biomarkers identified between human and canine glaucomas. Mutations associated with canine glaucoma vary between and within canine breeds, and gene discoveries therefore have limited overall effects as a screening tool in the general canine population. Biochemical markers of glaucoma include indicators of inflammation, oxidative stress, serum autoantibodies, matrix metalloproteinases, tumor necrosis factor–α, and transforming growth factor–β. These markers include those that indicate an adaptive or protective response, as well as those that reflect the damage arising from oxidative stress.

Introduction Currently identified biomarkers of canine glaucoma typically lack disease specificity. Combined with our limited understanding of glaucoma aetiopathogenesis, caution must be used to ensure each biomarker is used for the purpose for which it has been developed, investigated, or validated.

In this article, we review molecular and biochemical markers that have been identified in canine primary glaucoma and how these might be used and further developed to enhance our understanding and management of glaucoma (Table 1.3). A biomarker is a measurable anatomic, physiologic, biochemical, or molecular parameter indicative of a normal or pathologic process, disease state, or of a response to an intervention.141,142 In practice, biomarkers include tools and technologies that can aid in understanding the prediction, cause, diagnosis, progression, regression, or outcome of treatment of disease.143 A good biomarker should be precise and reliable, P a g e | 44 distinguish between normal and diseased states, and differentiate between different diseases.144 There are two major types of biomarkers:143 biomarkers of exposure, which are used in risk prediction, and biomarkers of disease, which are used in screening, diagnosis and monitoring of disease progression. The most important criterion for a relevant biomarker is its disease specificity.141

There is no ideal biomarker for glaucoma. Currently available clinical diagnostic tests (e.g. tonometry, gonioscopy) provide a very limited measure of risk of glaucoma and an imprecise measure of disease progression in dogs. While an elevated intraocular pressure (IOP) is considered a constant risk factor in canine glaucoma13 and a major risk factor in human glaucoma, the need for therapeutic interventions that address factors other than IOP is supported by the presence of normal-tension glaucoma and intraocular hypertension without glaucoma in humans, as well as the progression of glaucomatous optic nerve damage and vision loss despite an IOP within the normal range. Other proposed mechanisms involved in glaucoma include neuroinflammation,75 hypoxia/ ischaemia,74,145 increased extracellular glutamate and excitotoxicity,23 oxidative stress,146 obstruction of axoplasmic flow,147 and deprivation of trophic factors.148 By identifying accurate biomarkers of these processes in the canine glaucoma patient, the stage of disease or effect of an intervention might be more precisely identified.

Similar to the use of biomarkers in other medical fields (e.g. oncology), the use of identifying biomarkers to provide a binary ‘yes/no’ response is likely to be ineffective due to the complex, multifactorial and incompletely understood nature of the disease, and the fact that most biomarkers are not unique to glaucoma. Multiple biomarkers that create a molecular ‘phenotype’, or biomarker set that can be used to recognise unique patterns in canine glaucoma patients might allow the earlier detection of disease, lead to additional therapeutic targets and therapeutics, and influence management of human glaucoma through validation of the dog as an animal model of glaucoma.

While genetic markers have been used and developed over time, the methods use just three classes of markers: protein variants (allozymes), DNA sequence polymorphisms, and DNA repeat variation.149 The principle of allozyme markers is that protein variants in enzymes can be distinguished according to differences in size and isolectric point caused by amino-acid substitutions.149 The subsequent development of DNA manipulation techniques led to a shift from enzyme-based to DNA-based markers, providing a more direct and sensitive method of detecting variations. Molecular markers in medicine involve molecules, reactions and/or pathways altered by pathologic conditions.141 P a g e | 45

Table 1. 4. Use of molecular and biochemical biomarkers in canine primary glaucoma

Biomarker Use Strength Weakness Genetic Indicator of risk of disease Identification of mutation Need specific assay for each mutation mutations Minimise or eradicate allows development of Varying sensitivity and specificity of hereditary glaucoma commercial DNA assays different mutations Potential therapeutic target Onset of disease often occurs after Potential target for gene affected animals have been bred therapy

Transforming Target for improving bleb Elevated in serum of Not disease-specific growth factor- survival after filtration glaucomatous humans (less Changes may be causative or a result beta surgery invasive sampling than aqueous of glaucoma Potential therapeutic target humour) Not validated in dogs Matrix Indicator of disease Quantified differences between Not disease-specific metalloproteinase glaucomatous and normal Requires sampling of aqueous humour canine eyes Not validated in dogs Readily identified with established technologies Serum auto- Indicator of disease Minimally invasive sampling Changes may be causative or a result antibodies (blood sample) of glaucoma (against optic Variability of antibodies between dogs nerve antigens) Inconsistent differences between glaucomatous and normal dogs Not validated Inflammatory Indicator of disease Use existing methods to Poor disease specificity mediators May indicate stage of measure specific markers Marked variation in timing and disease Minimally invasive sampling magnitude of changes Potential therapeutic target (blood sample) Range of proteins present in serum Not validated in dogs Tumour necrosis Indicator of disease Established marker of Poor disease specificity factor-alpha inflammatory disease Changes may be causative or a result of glaucoma Oxidative stress Indicator of disease state Investigation of therapeutics in Poor disease specificity markers Potential indicator of risk other diseases have potential Changes may be causative or a result Potential therapeutic target translational benefits of glaucoma Use existing methods to Possible transient change measure specific markers Not validated in dogs

Genetic markers Genetic markers provide information about allelic variation at a given locus,149 and are typically a biomarker of exposure (indicator of risk).143 They exist before the disease/outcome occurs and are generally independent of other exposures. The most common types of genetic markers include restricted fragment length polymorphisms (RFLPs), amplified fragment length polymorphisms (AFLPs), random amplified polymorphic DNA (RAPD), simple sequence repeats (SSRs or microsatellites), and single-nucleotide polymorphisms (SNPs). The identification of specific genetic mutations in several breeds of dog has led to the development of commercial DNA assays that help identify dogs at risk of the disease, and remove them from the breeding population. The value of genetic screening as a tool to minimise or eradicate hereditary glaucoma is reliant on the presence of an assay for the specific mutation, the specificity and P a g e | 46 sensitivity of that mutation for identifying the risk of glaucoma, and use of the test in breeding programs. Canine primary angle closure glaucoma (PACG) is multifactorial, with numerous studies indicating it is not a mere genetic trait.35,70,79,150The fact the pathophysiology of glaucoma is incompletely understood and indeed likely varies between types of glaucoma, between species, and possibly between breeds of dog means that screening dogs and determining their risk of disease will continue to develop with improved understanding of the glaucomas. For example, progressive pectinate ligament abnormalities have been identified in some breeds predisposed to primary glaucoma (Flat Coated Retrievers, Eurasier dogs).58,83,84 The fact that pectinate ligament abnormalities and clinical glaucoma may develop despite prior test results of ‘normal’ or ‘unaffected’ has significant implications on the screening of breeding stock.58,83,84

Although genetic markers are only detected in certain populations (Table 1.4), finding these markers is critical in identifying the cause and pathogenesis of disease144 and research provides potential screening and therapeutic targets in both canine and human glaucoma patients. However, the mutated genes play a very limited role in the actual pathogenesis of disease and typically do not explain the usual clinical picture.141,151 While alteration in expression status of only one gene is almost never disease specific, a biomarker set is particularly valuable for the creation of highly precise diagnostic approaches.141

Table 1. 5. Genetic mutations associated with canine primary glaucoma

Breed Ocular Mutation phenotype Beagle POAG ADAMTS10 (homozygous for Gly661Arg variant) POAG Myocilin Basset Hound POAG ADAMTS17 (homozygous for inversion in intron 12) PACG COL1A2 and RAB22A PACG,normal Nebulin (variant in exon 48) Basset Fauve de POAG ADAMTS17 (homozygous for missense mutation in Bretagne exon 11) Norwegian Elkhound POAG ADAMTS10 (homozygous for missense mutation in exon 9) PBGV POAG ADAMTS17 (homozygous for inversion in intron 12) Shiba Inu POAG Myocilin SRBD1 (SNP in exon 4) Shih Tzu POAG SRBD1 (mutation in intron 1) PACG, primary angle closure glaucoma; POAG, primary open angle glaucoma; PBGV, Petit Basson Griffin Vendéen; SNP, single-nucleotide polymorphism

ADAM Metallopeptidase with Thrombospondin Type 1 Motif 10 A missense mutation (Gly661Arg) in the ADAM Metallopeptidase with Thrombospondin Type 1 Motif P a g e | 47

10 (ADAMTS10) gene was associated with primary open angle glaucoma (POAG) in a research colony of .50,150 This mutation is positioned in the cysteine rich domain and is hypothesised to disrupt protein folding leading to instability.50 This gene is also associated with fibrillin mutations, and is important in microfibril formation and function.50 The mutation has been excluded as a cause of primary glaucoma in the American Cocker Spaniel, Australian Cattle Dog, Chihuahua, , Siberian Husky, Shiba Inu, Shih Tzu, and Yorkshire Terrier.52

Glaucoma in Norwegian Elkhounds was mapped to the known canine POAG locus including the ADAMTS10 candidate gene and subsequently identified as a missense mutation in exon 9.48 The recessively segregating mutation results in an alanine to threonine change (p.A387T) in a highly conserved functional metalloprotease domain of the protein, which likely impairs ADAMTS10 function, leading to POAG in the homozygous dogs.

The Norwegian Elkhound mutation is different from that in Beagles suggesting the pathogenesis may be different in these breeds. Mutations in the Norwegian Elkhound affect a highly conserved residue in the metalloprotease domain, which plays a role in the remodelling of connective tissue.152 Human metalloprotease domain mutations have revealed abnormalities in the cellular cytoskeleton suggesting abnormal interactions with the extracellular matrix. These abnormalities may eventually result in defective microfibrils and glaucoma through alterations in biomechanical properties of tissue and/or through effects on signalling through transforming growth factor-β (TGF-β), which is known to be elevated in the aqueous humour of glaucoma patients.153

Recent investigations have demonstrated significant alterations in the posterior scleral collagen microstructure before the onset of clinical glaucoma in ADAMTS10-mutant dogs154 and significant stiffening of the posterior sclera associated with age in ADAMTS10-mutant dogs that did or did not have glaucoma.155 The rate of stiffening with age was similar between dogs with and without glaucoma despite exposure to progressively elevated IOPs in dogs with glaucoma which might have affected the scleral extracellular matrix.156 Changes to the rate and degree of scleral stiffening with age suggest a weaker capacity for handling an occasional rapid increase in IOP as the eye ages.155

ADAM Metallopeptidase with Thrombospondin Type 1 Motif 17 Glaucoma was found in 38 of 366 Petit Basson Griffon Vendéen (PBGV) dogs and was similar to that described in the Beagle and Norwegian Elkhound.37 All 38 were negative by commercially available DNA testing for the ADAMTS10 mutation (responsible for POAG in the Beagle) and the ADAM Metallopeptidase with Thrombospondin Type 1 Motif 17 (ADAMTS17) mutation (responsible for primary lens luxation in several breeds). However the locus for POAG in the PBGV was successfully mapped in a genome-wide association (GWA) study using exome sequencing.49 Follow up by genome sequencing of an individual case identified an inversion mutation with a breakpoint disrupting the ADAMTS17 gene.

While screening of these PBGVs did not identify mutations known to be associated with POAG in Beagles or primary lens luxation in terrier breeds,37 mutations in these or related genes are possible. As lens zonules consist of mircrofibrils, it seems possible relevant genetic mutations (such as those encoding microfibril proteins) would directly affect them. Further investigations are required to determine whether there is a direct association between POAG and lens luxation and/or whether zonular weakness and breakdown occur as part of the primary disease or is secondary to buphthalmia associated with chronic elevations in IOP. P a g e | 48

In the Basset Hound, a homozygous 19-bp deletion in exon 2 was present in all POAG-affected dogs.20 The deletion is predicted to alter the reading frame of the transcript and introduce a premature stop codon at the 5’ end of exon 3, which would result in a truncated and aberrant protein if the RNA were stable and not subjected to nonsense-mediated decay. If the transcript is translated, the protein is expected to be truncated by 86% including the entire catalytic domain, which would be expected to lead to complete loss of protein function.

All affected Basset Fauve de Bretagne dogs tested were homozygous for a missense mutation in exon 11 causing a glycine to serine amino acid substitution (G519S) in the disintegrin-like domain of ADAMTS17, which is predicted to alter protein function.20 Unaffected Basset Fauve de Bretagne dogs were either heterozygous for the mutation (5/24) or homozygous for the wild-type allele (19/24).

While identifying the ADAMTS17 gene as the site of an identified mutation in these three breeds, these findings demonstrate the breed specificity of specific mutations associated with glaucoma and highlight the difficulties present in undertaking any widespread genetic screening program.

Myocilin Myocilin protein was identified in the anterior chamber of Beagles with POAG, 157 but its role was not clarified. Aqueous myocilin protein levels increased significantly in primary and secondary glaucomatous eyes in many canine breeds including Beagles, while they were low in Beagles with primary cataract, diabetic cataract, and normotensive eyes.158 Changes in myocilin levels in the aqueous humour outflow pathway may be associated with increased IOP and subsequent glaucoma in Beagles.159 There have been no reports on the relationship between myocilin gene mutations and the incidence or pathogenesis of canine PACG.70 Myocilin sequence variants were identified in glaucomatous and normotensive Shiba Inu dogs with both open and closed iridocorneal angles (ICAs). However, these myocilin polymorphisms appear to be independent of the occurrence of glaucoma and the ICA grade. The observation of myocilin mutations/ polymorphisms in Shiba Inu dogs indicated that there were more alterations in exon 1 than in exon 3 and no mutations or polymorphisms detected in exon 2.70 This is similar to glaucoma in humans where many amino acid mutations in exon 3 are known to cause glaucoma,160 but exon 2 mutations are not reported in human POAG or PACG. 161-163 While only 4% of human POAG patients have mutated myocilin, this is the only known genetic mutation associated with both human and canine glaucoma. Given its presence in both species, as well as having been identified in two canine breeds (Shiba Inu and Beagle), further research might be useful to identify this mutation in other breeds predisposed to glaucoma. Collagen Type I Alpha 2 (COL1A2) and RAB22A The gene COL1A2 on chromosome 14 and RAB22A on chromosome 24 were identified as susceptibility loci on a GWA analysis of Basset Hounds with clinically confirmed PACG.76 These results suggested that variation in multiple loci determined the risk for canine PACG in Basset Hounds, in fitting with the complex nature of PACG inheritance.76 P a g e | 49

Nebulin (NEB) Ahram et al77 identified a NEB-based variant g.5588214 A->G in exon 48 (p.2051 Lys->Arg) and confirmed the association of this variant with PACG in Basset Hounds. This gene was predicted to affect protein function due to a possible pathogenic amino acid substitution within a highly phylogenetically conserved region. However, the disease-associated allele appears to be quite common in the US Basset Hound population. Further, homozygosity for the risk allele was observed in 33% of the unaffected animals, and heterozygosity was observed in a small fraction of the affected dogs suggesting additional factors are required for development of glaucoma.

S1 RNA Binding Domain 1 (SRBD1) A significant association of SRBD1 polymorphisms and glaucoma was identified in Shiba Inus and Shih Tzus.71 In Shiba Inus, the strongest association with glaucoma in SRBD1 was observed at rs22018513, which is a synonymous SNP in exon 4. Two other SNPs, rs8655283 and rs22018514, were also significantly associated with glaucoma; however, these significant associations were calculated only secondarily from a strong linkage disequilibrium with rs22018513. In Shih Tzus, however, rs22018513 was not associated with glaucoma. Only rs9172407 in intron 1 of SRBD1 showed a statistically significant association in Shih Tzus, while this association was not found in Shiba Inus. These results suggest that rs22018513 and rs9172407, or a respective neighboring polymorphism, may be a causative factor for glaucoma in Shiba Inus and Shih Tzus, respectively. SRBD1 polymorphisms were associated with canine glaucoma and human glaucoma independent of IOP,71 suggesting that SRBD1 polymorphisms may affect a common disease condition in canine and human glaucoma. Biochemical markers A biochemical marker is a biochemical variable associated with a disease (directly or indirectly) and might be any biochemical compound (antigen, antibody, enzyme, hormone etc) that is sufficiently altered to provide diagnostic or predictive value.164 Biochemical markers may differentiate specific diseases and guide therapy.144 Unlike genetic markers which primarily serve as a marker of disease risk and are independent of other factors, many biochemical markers of glaucoma are non-specific and their presence must be interpreted in context. Biomarkers obtained from diagnostic biological samples (e.g. blood) may indicate subclinical disease, stage or severity of disease. They may be present due to their causative role in glaucoma, or as a result of the disease process. Potential uses of these biomarkers include:143 identification of ‘at risk’ individuals or those in ‘preclinical’ stages; reduction in disease heterogeneity in clinical trials or epidemiologic studies; monitoring the progression of disease including induction, latency and clinically detected phases; and as a direct target for therapeutic interventions.

Transforming growth factor-beta (TGF-β) TGF-β signalling has been implicated in the pathogenesis of ocular, neurodegenerative and vascular diseases, as well as remodelling of extracellular matrix.165,166 The association of microfibril genes Latent Transforming Growth Factor Beta Binding Protein 2 (LTBP2) with primary congenital glaucoma in humans167,168 and ADAMTS10 with POAG in Beagles50,52 led to a theory that microfibril defects cause glaucoma.50,153 Elevated levels of TGF-β1 have been identified in the plasma of POAG humans169 and elevated aqueous humour TGF-β levels found in glaucomatous compared to normal eyes.170-172 Microfibrils are located in the extracellular matrix of a variety of tissues, where they are P a g e | 50 a major reservoir of latent TGF-βs169 and play a central role in TGF-β signalling.173-175 In the eye, microfibrils are found in the trabecular meshwork, inner wall of Schlemm’s canal, iris, ciliary body, and optic nerve head.176-178

The overlap in the cellular and tissue changes in glaucoma and those induced by TGF-β implies that dysfunctional TGF-β signalling could be involved in the pathogenesis of glaucoma in humans, monkeys, pigs and mice.170,179,180 TGF-β plays a role in scarring and fibrosis following glaucoma surgery and is a therapeutic target for improving survival of the conjunctival filtration bleb (minimising fibrosis which would otherwise encapsulate the draining aqueous humour) in glaucoma surgery.181 Preclinical and clinical research supports the role of TGF-β in glaucoma of humans and other animals, 170,179,180 though whether these proteins are always causative, or are induced as a consequence of glaucoma, remains undetermined. A main target of TGF-β activity is the trabecular meshwork. Because TGF-β1 and TGF-β2 are potential modulators of aqueous outflow, extracellular matrix remodelling, and inflammation in glaucomatous eyes, these proteins are currently being investigated as therapeutic targets.170 Thus, knowledge of TGF-β in canine glaucoma patients combined with advances made in other species may be translated to managing the progression of glaucoma in dogs.

Matrix metalloproteinases Intraocular matrix metalloproteinase (MMP) activity and its relationship to the aqueous humour outflow pathway has been quantified in normal and glaucomatous canine eyes182,183 and elevated levels of intraocular MMP-2 and MMP-9 have been documented in dogs with glaucoma.182

Immunohistochemistry has been used to identify MMP-2 and MMP-9 in the ICA of normal and glaucomatous eyes and specifically localised within trabecular meshwork cells, the extracellular matrix, and inflammatory cells.183 Quantitative assessment of immunolabelling revealed greater immunolabelling of MMP-2 and MMP-9 within the ICA tissue, iris and cornea in secondary glaucomas compared to primary glaucomas.183 Nonetheless, there was quantitatively more staining of MMPs in secondary glaucomas,183 despite no difference in MMP activity between primary and secondary glaucomas,182 and it is the activity (rather than the quantity of protein) that is most biologically meaningful. Future prospective investigations are required to determine if these changes directly mediate the glaucomatous changes or arise as a result of the disease process.182,183

Autoantibodies against optic nerve antigens Significant differences in the levels of autoreactivity at certain bands (increased autoreactivity at 40 and 53 kDa and decreased autoreactivity at 48 kDa) were identified in serum autoantibodies against optic nerve antigens between dogs with and without goniodysgenesis-related glaucoma.184 However, the study did not address whether immune-mediated mechanisms are involved in the initial pathogenesis or are a consequence of the disease that may in turn accelerate or intensify the disease process.184 Further studies are required given the population studied (advanced glaucoma versus control dogs), the high variability in autoantibodies among individual dogs, and the considerable overlap between groups.

Inflammatory markers Inflammation can result from a variety of tissue insults including infections, trauma, autoimmunity and malignancy. These stimuli lead to the production of cytokines that up- or down-regulate P a g e | 51 expression of acute-phase proteins.185 There is marked variation in the timing and magnitude of response among different acute-phase proteins.

More marked inflammatory changes have been identified in the earlier stages of glaucoma with substantially decreased inflammatory changes in the more chronically damaged .75 The decreased expression of crystallins in the retina may reflect an early, transient event in glaucoma pathogenesis.143,186-188 In a cohort of dogs considered at risk of developing glaucoma (healthy Eurasier dogs), there was no difference from control dogs in plasma levels of inflammatory markers including C-reactive protein (CRP; biomarker of the early phase of inflammation), haptoglobin (which increases over a few days and remains high for longer) and albumin (decreased synthesis during inflammation).58

It is not definitively known whether photoreceptor cell loss or functional decline is a feature of advanced glaucoma189-192 and several photoreceptor cell specific genes were expressed at decreased levels in glaucomatous retinas,186 while others mediating the neuroinflammatory response are expressed at higher levels in advanced glaucomatous retinopathy. Prominent functional categories of genes with elevated expression in glaucomatous retina compared to normal retina included antigen presentation, complement activation, lysosomal and proteasome activity, and acute phase proteins, inflammation signalling and apoptosis, and decreased expression genes involved in neuronal development and maintenance, cell adhesion, calcium transport and binding, transcriptional regulation and synaptic transmission.186

It has been suggested that glaucoma is not initiated by an immune response to specific retinal antigens, but rather the variety of immunoreactive molecules identified was due to the exposure of numerous epitopes during the rapid neuronal cell death. It is therefore conceivable that once an immune response had occurred, it will further accelerate vision loss independent of IOP.186

Tumour necrosis factor-α Modulation of tumour necrosis factor-α (TNF-α) and its receptors is correlated to the development of glaucoma, with increased immunoreactivity in glaucomatous canine retinas.186 Significantly higher TNF-α levels were detected in the aqueous humour of all dogs with intraocular disease compared to dogs with normal eyes,193 and TNF-α levels were significantly higher in those with acute anterior uveitis compared with those with PACG.193

Oxidative stress markers Oxidative stress is an increase over physiologic values in the intracellular concentrations of reactive oxygen species (ROS).194 This is reflected by changes in the levels of antioxidant defences that can be increased (protective response), or depleted due to ROS action.195

Retinal ganglion cell (RGC) death in glaucoma is associated with the generation and effects of ROS and reactive nitrogen species.196 Oxidative stress (reduced glutathione, decline in superoxide dismutase) occurs in the glaucomatous optic nerve head, trabecular meshwork cells and connective tissue. 197-199 It is likely important in the retinal damage that occurs in glaucoma,200-203 with an apparent transient nature of the oxidative stress suggested and may impact on endothelial cell function.146,196,204 Neurons, especially RGCs, are particularly sensitive to oxidative stress in glaucoma, which results in reduced levels of antioxidants within the neurons.204,205 P a g e | 52

Immunolabeling for malondialdehyde (MDA) and nitrotyrosine (NT), markers of free radical formation and oxidative stress, was identified in acutely but not chronically glaucomatous retinas.146 While the location of the highest level of immunostaining for NT was not described, the highest levels for MDA were seen in the nuclei of neurons including RGCs, photoreceptors of the outer nuclear layer (ONL), and putative neurons of the inner nuclear layer (INL), suggesting an increase in oxidative stress in these cells.

Glutathione peroxidase (GP) activity was significantly lower in a group of ‘at risk’ dogs (Eurasier dogs) compared to control dogs.58 A higher plasma taurine concentration and lower GP activity were observed in adult male Eurasier dogs suggesting an increased vulnerability to oxidative stress with age in this group. Plasma concentrations of methionine and cysteine in this cohort of dogs were similar between groups.58

Glutathione immunostaining was decreased in RGCs, INL neurons, and Müller cell processes in minimally damaged regions of glaucomatous retinas but was increased in severely damaged regions suggesting oxidative stress may occur before the redistribution of glutamate.146 The fact Müller cells and astrocytes were less heavily labelled than neurons, and the finding of glutathione depletion in Müller cell processes suggests these neighbouring glial cells may also undergo oxidative stress, but to a lesser degree than neurons, or that glutathione is upregulated more readily in these cells compared to neurons. In canine glaucoma, the majority of cells remaining in severely damaged regions of retinas appeared to be glial cells. These Müller cells and astrocytes contained high levels of both glutathione and taurine, suggesting they were resistant to further oxidative stress.146

Circulating autoantibodies against heat shock proteins have been identified in humans with glaucoma, and exposure to these antibodies can induce apoptosis in cultured retinal ganglion cells.206-209 Oxidative stress and TGFβ2 expression are both believed to be greater in glaucomatous versus healthy eyes, and are both able to promote expression of heat shock protein-27.207,210 Future work In clinical trials, ‘hard’ or ‘true’ endpoints are clinical events directly relevant to the patient.2 For glaucoma, the true endpoints would be significant loss of vision with decrease in quality of vision or quality of life, or development of functional disability.3 The development and use of validated surrogate endpoints will allow clinical trials and evaluation of interventions targeted at a much earlier stage of disease, with potential for greater effects, compared to what is achieved with current treatments. Surrogate endpoints have the potential to offer shorter, less expensive trials with reduced sample size requirements while allowing observation of a greater number of endpoints during follow-up than could be achieved with observation of a true endpoint.3

While many studies have investigated the genetics of glaucoma by linkage studies, candidate gene reports, GWA investigations, gene discovery has limited population effects211 and specific causal mutations generally differ between humans and dogs. In contrast, biochemical marker candidates have more similarities between human and canine glaucoma patients, suggesting common pathophysiological mechanisms. Therefore, progress in developing biochemical markers of glaucoma in humans will likely have translational benefits in veterinary medicine.

Where candidate biomarkers have been identified in the aqueous humour of glaucomatous compared to normal canine eyes, use of the sampling technique to obtain data from large P a g e | 53 populations has been precluded by the invasive nature of this sampling technique and the requirement for specialised training, which constrains the validation and verification of findings. Body fluid biomarkers have the advantages of accessible collection methods and thus a better by patients or animal owners.212 Despite ease of blood sampling, identifying potential biomarkers in plasma or serum poses difficulties because of the sample complexity and the low concentrations of individual markers, and degradation of biomarkers.142,213,214 Analysis of more proximal samples (e.g. tears or aqueous humour) may present higher sensitivity and specificity for screening and monitoring glaucoma-related biomarkers.215 The tear proteome is a rich source of diagnostic markers and is easily sampled.212 Studies in human glaucoma patients have detected alterations in tear samples related to anti-glaucoma medications,216,217 largely based on analysis of single targeted markers.218-221 With the development of very sensitive proteomic techniques, repeated analyses are possible despite the small volume of tears that can be collected from a patient.212 Development and progress in sample fractionation and mass spectrometry as a major analysis tool for large-scale protein studies has been a focus to help uncover the key molecular processes in human glaucoma.212,215,222

The use of the dog as a model of glaucoma in humans could provide another population to validate and verify techniques and candidate biochemical markers in human glaucoma. Longitudinal studies involving dogs of at-risk breeds or colonies of purpose-bred dogs with an identified predisposition to glaucoma could potentially allow correlation changes in biomarkers with stage of disease.223

While many potential biomarkers have emerged from laboratory and pilot studies, validation in independent experiments, or large-scale clinical studies are lacking.211,212,224 Consideration must be given to the limitations of available techniques.212 While gel-based techniques provide information (e.g. molecular weight, presence of posttranslational protein modifications), they are unable to detect low-abundance proteins, while mass spectrometry techniques cover a larger proteome range compared to gel-based techniques.212 Regardless of the technique used, standardised collection procedures, power analyses to determine appropriate samples sizes, and subsequent verification of findings are required. These limitations in biomarker development are likely associated with the multifactorial nature of the glaucomas themselves, as well as the variation in techniques and procedures available to investigators. Conclusion Challenges in glaucoma research include limitations in our understanding of disease mechanisms, limited knowledge of potential therapeutic targets, a paucity of preclinical models, and limitations in the ability to detect early development and progression of the disease. Novel tools to improve screening, diagnosis and assess disease progression are therefore required. The identification, validation and appropriate use of biomarkers could help address this gap and could inform pathogenesis and identify therapeutic targets. This will require studies that identify candidate biomarkers, evaluate their association with disease, validate the biomarker (or more likely a biomarker set) for the targeted use (e.g. early detection, disease progression), and show effects of interventions on both the biomarker and clinical outcomes. P a g e | 54

DIAGNOSIS Clinical diagnosis Tonometry Tonometry provides an indirect and noninvasive estimation of IOP and is more clinically relevant than direct measurement.225 In clinical veterinary practice, indirect tonometry may be obtained by indentation, applanation or rebound tonometry and overall, both applanation and rebound handheld tonometers are considered useful instruments for clinical measurements of IOP in animals91,226 As application of a force to the cornea is required in performing tonometry, these indirect measurements are subject to the biochemical properties of the corneal surface. In addition to the impact of corneal abnormalities,227 body position and manual restraint also affect IOP.13 Studies of IOP measurements, including comparisons between applanation versus rebound tonometry, across a range of IOPs demonstrate the importance of using the same device when monitoring IOP in individuals.228-236 Reported tonometric findings in dogs are reported in Table 1.5. In addition to differences associated with they type of tonometer used, consideration for several factors that influence IOP readings must be given when monitoring IOPs in dogs. Factors that affect IOP measurements in dogs include 26,227:

• Excessive restraint - pressure on jugulars or eyelids, for example, will increase IOP measurements • Patient’s distress – a greater degree of variation in IOP has been documented in unco operative animals237 • Corneal health - corneal pathology has considerable influence on both tonometers with the degree of over- or underestimation of IOP depending on the alteration of biomechanical properties of the cornea caused by the pathology.227 • Time of the day as diurnal variation in IOP may be up to 3-4mmHg in normotensive eyes, or up to 10mmHg in dogs with POAG.238 • Predictive tool for PCAG in IOP screening of high-risk patients is poor.10 • Age - in several breeds, IOP is reported to decrease with age.58,226,239 and a steady and statistically significant rise in IOP correlated to age in glaucomatous Basset Hounds over a period of 22 months.78

P a g e | 55

Table 1. 6. Tonometers recommended for use in veterinary practice

Applanation Rebound References

How it IOP is inferred from the force required to A magnetized plastic-tipped metal probe (1.4 91 works flatten (applanate) a constant area of the mm diameter) is bounced against the cornea. cornea The velocity at which the probe returns is converted into electrical signals, from which the IOP is calculated Common Tono-Pen® XL; Tono-Pen VetTM TonoVet® device Practical use Requires local anaesthetic No anaesthetic needed 91 Less influenced by examiner/experience Both • Similar IOPs at normal pressures 228,229,234-236 techniques • Increasing differences with increasing IOP • IOP lowering effect (varies for different agents) through unknown mechanism 233,240 Normal IOP 12.9 ± 2.7mmHg – 19.2 ± 5.5mmHg 9.1 ± 3.4mmHg – 10.8mmHg 229,235,239 (mean) Manometry Enucleated eye model – underestimated High accuracy over whole pressure range 228-230 IOP in glaucomatous eyes evaluated (5-80mmHg) in dogs and 25- 50mmHg in cats Normal eyes Lower IOP compared to Mackay-Marg 228 applanation (gold standard for in vivo tonometry) Glaucoma Underestimate elevated IOPs Significantly higher IOP compared to Tono-Pen 228,230- XL in naturally occurring glaucoma 232,234,236 Basset Hounds: rise in IOP correlated to age 78 Corneal 1mmHg elevation in IOP per 100μm 2mmHg elevation in IOP per 100μm increase in 233 disease increase in corneal thickness corneal thickness; more prominent dependency Unaffected by CCT on CCT 226 Lower effect of corneal disease on IOP 227

Increases in IOP are seen with administration of atropine sulfate causes a when given both topically and by intramuscular injection,78,241 after intravenous injections of ketamine, and to less a degree, diazepam.242 An IOP lowering effect (varying degree for different agents) following application of topical anaesthetic agents has been reported in dogs.228,233,235,240,243,244 The mechanism for this finding is unknown, although it is not associated with mechanical repetition of IOP measurements as similar IOP lowering effects reported in humans as measured by a non-contact tonometer.245 A larger degree of IOP lowering associated with topical anaesthetics has been proposed and iris pigmentation based on reported in dogs with darker irides.240,244 Gonioscopy The pectinate ligament and ciliary cleft width have a significant effect on aqueous humour outflow in dogs,246 and gonioscopy allows direct or indirect visualisation of the ICA angle, pectinate ligament, and anterior portion of the ciliary cleft.91 This examination technique facilitates classification of glaucoma on the basis of ICA and anterior sclerociliary cleft morphology. Total internal reflection of light rays at the posterior cornea prohibit direct visualisation of the drainage angle. Use of a goniolens P a g e | 56 facilitates examination of the ICA by replacing air at the corneal surface, therefore reducing the difference in refractive index between these media resulting in a high angle of incidence of the light rays from the ICA.247 A gonioscopic examination should include assessment of the ICA width, depth of the sclerociliary opening and cleft, the length and diameter of the pectinate ligament strands (Fig.1.2), the size of any dysplastic areas, the number of flow holes (by quadrant or degrees of the circumference), and any other abnormalities noted.13 This assessment is difficult to quantify though a grading system has been proposed by Ekesten & Narfstrom (1991) as depicted in Figure 1.3. Goniolenses allow either direct examination of the drainage angle, or indirect examination in which the image is viewed in a mirror. 247

Figure 1. 2. Photographs of pectinate ligament abnormalities (a) Normal pectinate ligament (PL) visible when rarefaction of the initial fibrillar sheet was almost complete, by 2 to 4 weeks after birth, leaving strands of intertwining collagen, progressively encased by attenuate trabecular cells, confluent with the anterior surface of the iris (stage 0). (b-d) The entire ICA and opening of the CC were systematically examined for the presence of a PLA, defined as abnormally broad and thickened pectinate ligament fibres (stage 1) (from Boillot T, et al. Determination of morphological, biometric and biochemical susceptibilities in healthy Eurasier dogs with suspected inherited glaucoma. PLoS One 2014; 9(11): pe111873

P a g e | 57

Figure 1. 3. Assessing the iridocorneal angle using gonioscopy. Top Schematic diagram of grading system proposed in Samoyeds. An estimate is made of the ratio of the ciliary cleft opening width (A) to the distance from the pectinate ligament to the anterior corneal surface (B) (from Ekesten B, Narfstrom K. Correlation of morphologic features of the iridocorneal angle to intraocular pressure in Samoyeds. Am J Vet Res 1991; 52:1875).

Ophthalmoscopy Examination of the optic nerve head, including the neuroretinal rim and cup, is critical for the diagnosis of POAG in humans.248,249 In dogs, myelination of the ONH axons limit early detection of neuroretinal rim narrowing and enlargement of the cup which are characteristic changes associated with glaucoma.13

When assessable, the fundus should be examined for signs of optic nerve head degeneration and cupping, peripapillary areas of chorioretinal atrophy, and diffuse retinal degeneration, especially in comparison to the contralateral healthy eye if applicable.91 These findings are signs of chronic glaucomatous changes, and they are always present with more overt clinical signs, such as buphthalmos, corneal oedema, or Haab’s striae.91 Earlier signs, such as the occasional early mild papilloedema in patients with very high IOP values, are less common and more difficult to detect.91

Direct ophthalmoscopy is used monocularly to provide a direct and upright image of the fundus with approximately 15 – 17 times magnification.247 With indirect ophthalmoscopy, a convex lens lens is used and an inverted virtual image formed with less magnification than in direct ophthalmoscopy, but with a greater field of view. Imaging Ultrasound P a g e | 58

Ultrasound technology allows examination of the trabecular meshwork and sclerociliary cleft that can approximate histological findings.13 The resolution obtained using 10-12MHz probes used in B-scan ultrasonography approximates 300-400μm, using 20MHz probes used in high frequency ultrasound (HFUS) approximates 80μm, and 60MHz probes used in ultrasound biomicroscopy (UBM) achieves a resolution of 20μm and penetration to 5mm depth.13 Similar to findings in PACG in humans, a narrow anterior chamber with increased thickness of the axial lens and vitreous body has been reported in Samoyeds with PNAG or PACG using B-scan ultrasonography.87 However, in Beagles with POAG, a deeper anterior chamber and vitreous were reported with a normal lens axial length and position. High frequency ultrasound findings have been proposed as potential markers of glaucoma in specific breeds,58,250 and structural changes iridocorneal cleft documented in a colony of glaucomatous Basset Hounds over a 12 month period.78 UBM can provide cross-sectional images of ocular tissues at low-power microscopic resolution by using high-frequency ultrasound probes251 and has been used in veterinary ophthalmology to evaluate the anterior segment, including the ICA.252-258 UBM identifies abnormalities deep within the ICA (not visible clinically with gonioscopy), especially on the ciliary cleft and the scleral venous plexus, in glaucomatous eyes as well as eyes predisposed to PACG.251,255 Using UBM, Kawata et al256 identified a significant difference (a larger opening of the ciliary cleft and larger scleral venous plexus) in cases that were responsive to medical therapies compared to dogs that were unresponsive. Further investigation by means of a longitudinal study are required to determine whether these structural parameters might serve as biomarkers or risk factors in the development of glaucoma in this breed.58 Optical coherence tomography (OCT) Resolution of images obtained using OCT approximate 10μm13 though relies on the presence of clear ocular media. OCT can be used to monitor progression of retinal damage in glaucoma or in experimental models of retina and optic damage, structural imaging can also be used to obtain better prognostic information about the status of the optic nerve prior to making any decision whether or not to pursue more aggressive therapeutic options for treatment of glaucoma.259 Morphometric analysis of retinal tissue sections demonstrated significant asymmetry between superior and inferior retina (superior retina is approximately 10–15% thicker than the inferior retina) in healthy canine eyes, and this has been confirmed in vivo using OCT.259

Other diagnostics Provocative testing Provocative tests are performed to assess predisposition to glaucoma. Mydriatic and darkroom provocative testing is commonly performed, while water and corticosteroid provocative testing is performed in POAG.13 Parasympatholytic-induced mydriasis or P a g e | 59 physiologically induced darkroom mydriasis results in pupil dilation so the peripheral iris impairs aqueous humour outflow. It is suggested abrupt elevations in IOP (>5mmHg) within the first hour are suggestive of PNAG.13 Pharmacological pupil dilation with short-acting (tropicamide) and long-acting (atropine) mydriatic agents resulted in a significant increase in IOP in glaucomatous Basset Hounds but not in normal Beagles.78 Atropine was administered after pupil dilation with tropicamide and resulted in a further increase in IOP compared to the increase obtained with tropicamide alone. Electroretinography RGC function evaluated by pattern ERG revealed dogs with glaucoma develop significant and progressive functional deficits as early as 18 months of age before marked increases in IOP and suggests pERG functional deficits are an early indicator of disease progression in canine glaucoma.78 In comparing pERGs between normal and POAG Beagles, significant differences in the amplitude in the peripheral compared to the central retina suggesting the peripheral retina and larger RGCs in POAG Beagles may be more sensitive to sodium thiamylal (possibly due to ischaemic events or decreased perfusion of the peripheral retina) and to elevations in IOP.21,260 In normal dogs with marked acute elevations in IOP, pERG waveforms were more sensitive (compared to fERG) to increases in IOP.259,261 Increasing levels of experimentally induced IOP result in increased latency and reduced amplitude of the a- and b- waves and of the oscillatory potentials of the ERG.262 However, with prolonged IOP elevations, fERGs are typically nonrecordable due to degeneration of both the inner and outer retinal layers. Short-term experimentally induced increases in IOP affected pERG and fERG waveforms with pERG waveforms identified as more sensitive to increases in IOP.246 There was no evidence of histological changes between treated and control eyes in this study (short duration of elevated IOP and single sampling of the retina may obscure histopathological findings). Raman spectroscopy Wang et al18 explored the potential use of Raman spectroscopy for early detection and characterization of glaucoma-like pathological anomalies in dogs. Markers related to altered intracellular proteins were identified to differentiate early disease from healthy tissue18 though this is not used in a typical clinical setting and validation for clinical use has not yet been performed.

P a g e | 60

TREATMENT Although IOP may be successfully lowered during the acute stage of onset of glaucoma, vision is often lost due to rapid retinal and optic nerve degeneration.263 Current treatment of glaucoma is aimed at lowering IOP through either medical or surgical means. IOP is lowered either by reducing the production of AH from the ciliary body, or by increasing AH outflow. Medical treatment Although some ocular hypotensive drugs affect both pathways, they can be categorised by their dominant mechanisms of action as:264 1. those that inhibit the rate of AH production, i.e. carbonic anhydrase inhibitors (CAIs), β-adrenergic antagonists and selective α2-adrenergic agonists; 2. those that increase AH outflow, i.e. nonselective adrenergic agonists, parasympathomimetics and prostaglandin F2α analogues (PGAs); and 3. those that dehydrate the intraocular space via the creation of an osmotic gradient (osmotic agents). Inhibition of aqueous humour production Carbonic anhydrase inhibitors − + − CA catalyses the reversible reaction CO2 + H2O ↔ HCO3 + H . HCO3 is produced in − + the non-pigmented ciliary epithelium (NPE). HCO3 and N pass into the posterior chamber establishing an osmotic gradient.264 AH is formed when water flows into the posterior chamber via osmosis. It is thought CAIs bind to the enzyme’s active site in − 265 the NPE to inhibit HCO3 production. CAIs are non-bacteriostatic sulfonamide-related compounds and due to the frequency and severity of adverse effects with systemic use, topical preparations are typically used in veterinary medicine (Table 1.6). They vary according to their binding affinity to individual CA isozymes, their potency for inhibiting individual isozymes, and their physicochemical properties (influences tissue distribution & scope of activity).266 CAIs 267-270 reduce HCO3 synthesis in the ciliary body, resulting in 20-40% reduction in AH production.264 P a g e | 61

Table 1. 7. Carbonic anhydrase inhibitors used in the treatment of canine glaucoma

Drug Dose Disease Finding state Systemic medications Acetazolamide 10-75mg/kg All Single dose caused IOP reduction for at least 8 hours271 PO Methazolamide 5mg/kg PO Glaucoma No further IOP reduction when topical dorzolamide 2% added272 25mg PO Normal 18% IOP reduction for 3 hours273 50mg PO Normal 25% IOP reduction for 6 hours273 TID dosing Minimal IOP reduction in evening; significant increase after 2-3 days compared with baseline13 Topical medications Dorzolamide 2% topical; Normal 18% IOP reduction (3.1mmHg)274 TID Average 3.1mmHg reduction in IOP275 Glaucoma Max. reduction 5-6.5hrs; return to baseline values by 10.5hrs276 Brinzolamide Topical Normal 3.5% IOP reduction; max. reduction after 5-6.5hrs; return to baseline values by 10.5hrs276 Dorzolamide- 2%/0.5% Glaucoma Greater reduction after 4 days compared to 1 day of timolol topical treatment277

Prostaglandin analogues Prostaglandin analogues (PGAs) reduce IOP by enhancing uveoscleral outflow,278 and possibly through increased trabecular outflow, although the exact mechanism for this is unknown.279 PGAs commonly used in the management of glaucoma in dogs are outlined in Table 1.7. The IOP lowering effect of PGAs is mediated by direct binding to prostaglandin F2α receptor (FP receptor)280 with possible indirect stimulation of E- prostanoid 3 (EP3) receptor.281 In humans, there is the suggestion that PGAs may increase conventional outflow by inducing changes within the TM and Schlemm’s canal.282-284 However, when considering AH dynamics in different species, consideration must be given to differences between species. Topical 0.005% latanoprost increases episcleral venous pressure in dogs and may therefore increase resistance to AH outflow through the TM in dogs.285 Tsai et al286 speculate that the lack of effect of topical latanoprost on the iris sphincter and possibly the ciliary body in humans compared to dogs explains the absence of reports of ICA narrowing associated with latanoprost in humans. It is suggested that PGAs results in several changes in the ciliary muscle including MMP upregulation, ECM degradation and loss of filling spaces between ciliary muscle bundles followed by enlargement of these empty spaces resulting in decreased hydraulic resistance in uveoscleral outflow and IOP reduction.279 This synthesis and activation of MMPs with resultant ECM remodelling in the ciliary muscle is not an acute change, and does not account for the acute reduction in IOP that can result P a g e | 62

when topical PGAs are used to treat acute PACG in dogs where marked elevations (80mmHg) can resolve within 1-2 hours.287 Miller et al 2003287 suggested this acute reduction may be secondary to the miosis which has been demonstrated to result in opening of the ciliary cleft in healthy288 and glaucomatous dogs,287 even though narrowing of the ICA and shallowing of the AC has been documented normal dogs.286

Table 1. 8. Prostaglandin analogues in dogs

Drug Disease Findings Latanoprost Healthy Significant IOP reduction289-291 Miotic & hypotensive effects not altered by concurrent prednisolone treatment (10 clinically normal dogs)292 Decreased ICA and CC entry; increased mid-CC width; no effect on CC length293 Time dependent changes in ciliary body thickness (thinner after 2hrs; rebound thickening after 1 week)293 Significant IOP reduction in beagles when given SID or BID; less IOP fluctuation with BID dosing294 Glaucoma Dramatic reduction rapidly5 11% maintained IOP <22mmHg for ≥1year295 Possible link with anterior uveitis,296 however subsequent studies did not find increased rate of uveitis relapse297,298 Considerations May cause BAB disruption in normal dogs.299 Dogs possibly have more labile BAB300 May increase severity or recurrence of herpetic keratitis301 Conjunctival hyperaemia289 Miosis without evidence of ocular pain276 (axial cataracts with miosis276 Bimatoprost Glaucoma Significant IOP reduction (Beagle)302 Normal Early transient increase in EVP (first 2wks); sustained decrease lasting 4-7 weeks; IOP reduction for minimum 7 weeks303 Considerations Previous study showed INCR EVP with topical latanoprost – possible reason for difference is selective vasodilation of outflow vasculature with IC dosing compared to generalised vasodilation of all surface vessels with topical meds303. Consider additional benefit possible with IC dosing Travoprost Glaucoma Significant IOP reduction (Beagle)304 Considerations Conjunctival hyperaemia305 Unoprostone Glaucoma Significant IOP reduction (Beagle)304 Tafluprost Normal 6 normal beagles; preservative-free306 IOP significantly lower from 4-24hrs Pupil diameter significantly smaller from 30mins-18hours ICA significantly decreased CCW significantly increased No change in CC length, CC area & AC depth Considerations Reported benefits over other PGAs in humans: • enhanced retinal blood flow307 • high potency IOP reduction308 • fewer side effects309 Miosis occurs following topical application of PGAs including bimatoprost,302 latanoprost,294 or travoprost304 in dogs, suggesting the iris sphincter muscle is sensitive to PGAs.310 Conjunctival hyperaemia may also occur as a result of nitric oxide- mediated vasodilation rather than as a proinflammatory effect, although local irritation may also contribute. P a g e | 63

Parasympatholytic drugs Cholinergic agonists (parasympathomimetics e.g. pilocarpine, carbachol, demecarium bromide) result in iris sphincter contraction (miosis) and ciliary muscle contraction (accommodation and expansion of the trabecular meshwork) in humans311 and other species.312,313 The commonly used parasympathomimetics for use in canine glaucoma are outlined in Table 1.8. They are less efficacious in treating acute PACG in dogs though. While miosis and subjective ICA narrowing occurs in dogs, definitive studies documenting this relationship are lacking. One suggested reason for the different efficacy between parasympathomimetics and latanoprost in acute canine PACG is the approximate fivefold greater maximal miosis induced by latanoprost compared with 4% pilocarpine in normotensive beagles.314

Table 1. 9. Parasympathomimetics used in canine glaucoma

Drug Disease state Findings Pilocarpine Normal, POAG Reduce IOP by 30-40% for 6 hours314,315 Increased AH outflow316 Greater reduction in glaucomatous dogs & no difference between concentrations316 TID-QID dosing with solution; SID dosing with gel317 Considerations Local irritation within 0.25-6hrs315,318 alleviated without affecting efficacy by using buffer-tip droptainer to increase the pH319 Miosis and aqueous flare prostaglandin-mediated (probably secondary to neuropeptide release in AH320 Carbachol Post phaco 0.01% solution Intracameral for rapid miosis and inhibit POH321 No protective effect of carbachol on POH322 Normal, POAG 0.75,1.5,2.25,3% IOP reduction within 1hr (peak 2-7hr)323 Considerations Needs to be combined with a surfactant to penetrate corneal epithelium Carbamyl ester of choline13 Carbamate inhibitors: Reversible binding to acetylcholinesterase13 Demecarium Beagles: 0.125%, 0.25% Needs to be compounded bromide Normal, POAG Single drop – up to 77hrs miosis & IOP lowering 55hrs324 (similar effect but longer duration of action) Recommended BID dosing to minimise occasional peaks324 31months until onset of glaucoma in at risk eyes (8 months in control)325 Organophosphorus inhibitors: Irreversibly inhibit the enzyme by forming a stable complex Echothiopate Beagles: 0.3, 0.125, 0.25% 324 iodide Normal, POAG 10mmHg reduction in normotensive 13mmHg reduction in POAG 25-53hrs IOP reduction Associated miosis maximal 1-3hrs; persists 49-55hrs

Beta-adrenergic antagonist (beta-blocker) Beta-blockers decrease AH formation by altering adrenergic neuronal control by blocking β-receptors in the ciliary body processes.326 The most commonly used β- blocker in the treatment of glaucoma is timolol which is non-selective (blocks β1 & 2 P a g e | 64

receptors) (Table 1.9).

A preliminary in vivo study to assess safety and efficacy of a nanoparticle loaded contact lens at treating glaucoma showed a slow and extended release of timolol from contact lenses.327 The study showed lens insertion decreased IOP but the effect was insignificant after day five, and there was no significant effect on the first day.

Table 1. 10. Beta-blockers in canine glaucoma

Drug Disease Findings state Timolol Normal Mixed results showing reduction,328 and no IOP reduction290,329 0.25, 0.5% IOP reduction (dose dependent) with 2-8% concentration330 Decrease in contralateral eye IOP328 0.5% gel-forming prep: 5.4mmHg reduction lasting 24hrs331 POAG 4-5mmHg IOP reduction329 8-14mmHg reduction for up to 6hrs with 4, 6 & 8% concentrations330 Effect enhanced by combining with 2% pilocarpine330 Betaxolol Predisposed Delay onset of PACG325 0.5% Considerations Minimum 5minutes between different medications to avoid reducing ocular bioavailability13 Local intolerance (stinging, burning); photophobia, ptosis, blepharoconjunctivitis, superficial keratitis also possible332 Sig reduction in tear production & turnover333 Reduction in pupil size (cats>dogs)328,330 Significant reduced PR in normal & glaucoma (2-8% formulations)329,330 Systemic absorption via transconjunctival and nasolacrimal duct routes: 0.25% in cats & dogs <10kg; 0.5% in dogs >25pounds5

Osmotic agents Mannitol, the most common osmotic agent, is used on an emergency basis due to its efficacy at rapidly decreasing IOP (Table 1.10).91,334 Osmotic agents increase the osmolality of serum relative to that of intraocular fluids (particularly vitreous). IOP decreases when water diffuses into the chorioretinal vasculature. 335 There is also possible posterior displacement of lens-iris diaphragm with vitreous dehydration which may allow for improved AH outflow through the ICA336. P a g e | 65

Table 1. 11. Hyperosmotic agents in canine glaucoma

Drug Dose Findings Glycerol 1.5g/kg 17% mean maximal IOP reduction 1hr after dose (not significantly different to controls)334 Increased serum osmolality after dosing Significant increase in blood glucose relative to control334 Metabolised to simple carbohydrates (including glucose)337 Occasional nausea, vomit, headache337 1.4g/kg glycerol significantly decreased IOP for up to 10hours in normotensive dogs338 Isosorbide 1.5g/kg 13.5% mean maximal IOP reduction 30mins after dose (not significantly different to controls)334 Decreased serum osmolality after dosing Insignificant increase in blood glucose relative to control334 Not metabolised after systemic absorption, so no effect on glucose homeostasis334,339 Nausea and vomiting less frequent compared to with glycerol340

Prophylactic treatment Given the bilateral nature of PACG, the second eye is at high risk of developing glaucoma within days to years following overt disease in the first eye.13,29,55,60,325,341,342 Several studies have described the effect of prophylactic treatment of the unaffected eyes in dogs deemed at risk of developing PACG (Table 1.11).55,325,342,343 However, these studies all included a variety drugs and treatment frequencies in the treatment arm and surrogacy needs to be evaluated in the context of a particular class of treatment regimens. Other treatments MacKay et al344 investigated the effect of a single dose of varying concentrations (0.005-4%) of topically applied CoherinTM (a complex of peptides isolated from bovine or porcine pituitary gland) on IOP, pupil size and heart rate in a pilot study on glaucomatous Beagles. IOP reductions were identified at certain concentrations (independent of statistically insignificant changes in pupil size) and the authors proposed CoherinTM as an agent with potential use to reduce IOP.344 Topical administration of 2% THC ophthalmic solution three times daily in 21 normal Beagles lowered IOP in the morning (21%) and evening (15%) but did not reduce aqueous humour flow rates.345 This reduction in IOP is less than that reported for commonly used glaucoma medications including latanoprost,290 bimatoprost,302 dorzolamide,274 and dorzolamide-timolol,277 but greater than that reported using timolol alone.290,346 There is inadequate evidence to suggest topical THC for clinical use in dogs as 5 THC responders also had >20% IOP reduction after administration of a control vehicle.345

P a g e | 66

Table 1. 12. Prophylactic glaucoma medications in eyes predisposed to primary glaucoma

Drugs Time to fail Conclusion Paper limitations Ref Demecarium No statistical difference between Retrospective 342 bromide 330days glaucoma meds - 0.125% 143days Tendency to longer time with anti- - 0.25% inflammatory use (324 vs 143 days) Latanoprost 0.005% 284days Dorzolamide 2% 272.5 days Various 19.2 mths Significantly shorter time to failure Retrospective. 29 compared to Miller et al; but longer than Aim of the study was not assessing Miller et al’s control group efficacy of prophylaxis Drugs used varied over time (dipivalyl epinephrine, dorzolamide, brinzolamide, latanoprost, travoprost) Betaxolol 0.5% BID 30.7 mths PACG - Suggest protective effect of The date of onset of glaucoma 325 Demecarium 0.25% 31 mths prophylactic therapy in eyes predisposed determined by careful questioning of /betamethasone/ to glaucoma (control eyes 8 months) the owner, referring veterinarian, or gentamicin SID both, and review of the clinical history Dorzolamide 2% 7/18 at Topical reactions (irritation, KCS, ulcer) 6- and examination findings. (n=18) 15.7mths 8%; proportion of Cockers had facial n. Control group: first examined by Dorzolamide 2% / 8/12 at palsy at or following diagnosis. ophthal at diagnosis of second eye Timolol 0.5% (n=12) 10.4mths Not 100% compliance No fixed follow up – some relied on owner or referring vet No consistent STT measurement during study (patterns noted during analysis; different treatments at different hospitals) Timolol, 10 months Prophylaxis significantly extended time to Retrospective 55 dichlorphenamide, diagnosis in 2nd eye from 5 to 10mths in Not look at individual treatments echothiophate predisposed breeds 4 dogs had prophy sx No sig difference in effect of prophy tx when all breeds were analysed - incr risk 5-10yo - females 2X risk cf male

Surgical treatment The surgical technique used to treat glaucoma in dogs is primarily dependent on the visual status of the eye. Cyclodestructive procedures and filtering techniques are used to control IOP in visual patients while in blind eyes the goal of treatment is managing pain.347 The goal of glaucoma therapy in eyes with vision is to prevent further optic nerve damage and preserve vision.347 Because of the rapidly progressive course of canine primary glaucoma, early surgical intervention has been suggested to improve the surgical outcome.13 P a g e | 67

Sighted eyes (i) Cyclodestructive procedures Cyclodestructive procedures include transscleral and endoscopic cyclophotocoagulation and cycloablation.91 Melanin present in the ciliary body epithelium and stroma is the target of diode laser energy which is absorbed to result in photocoagulation and destruction of the ciliary body, with a resultant decrease in IOP production.91 Trans-scleral cyclophotocoagulation (TSCP) is performed via contact mode application, where scleral indentation by a handpiece probe enhances laser delivery and absorption by the target tissues.91 Various surgical parameters have been described in canine patients. Reports describe adjustment of energy levels to produce a ‘pop’ – the sound produced by photodisruption causing destruction of the ciliary body tissue – in 20-75% of treatment sites.8,348 However, as overtreatment of the ciliary body (recognisable as a ‘pop’) is associated with greater tissue necrosis and inflammation, efforts to minimise the energy delivered to the eye are required to reduce postoperative complications.349,350 Comparing the efficacy of treatments with different settings and number of sites treated is difficult due to different criteria used to define success, and the spectrum of disease that the make up the glaucomas. IOP control is reported between 658 - 92%351 of dogs TSCP, with vision retention in potentially sighted eyes 228 -50%351. Direct comparisons between studies to determine the efficacy of treatment modalities is difficult due to the varying nature of the studies, different criteria used to define success, and the spectrum of disease that the make up the glaucomas. Endocyclophotocoagulation (ECP) differs from TSCP by allowing direct visualisation of the target tissues.91 Energy delivery is optimal, as it can be titrated to reach the desired tissue blanching and destruction, which is visible during the procedure. Reported postoperative complications range from moderate to severe uveitis, ectropion uveae, corneal ulceration, relapsing of glaucoma, and phthisis bulbi.352 (ii) Filtering surgeries Filtering procedures allow diversion of aqueous humour from the anterior chamber. In dogs, procedures that have been described to have uniform failure include iridencleisis, cyclodialysis, corneoscleral trephination, and posterior sclerectomy.13 Placement of a glaucoma drainage device (GDD) is the filtering technique of choice in canine glaucoma at present. However, historically both valved & non valved devices have yielded disappointing results (fibrin occlusion of AC shunt or late-occurring fibrosis around end plate.353

P a g e | 68

Table 1. 13. Glaucoma drainage devices used in dogs

Trial descriptors Quality Items Ref Intervention Population and other content-specific items Follow-up IOP control Vision VALVED Modified Joseph 21 eyes (15 dogs) 9-15 7/21 (33.3%) Not reported 354 GDD Primary glaucoma (14 dogs) months Uveitis and angle closure (1 dog) Krupin-Denver 8 dogs 1-12 4/8 (50%); 2/6 Not reported 355 GDD Glaucoma (n=6); normotensive (n=2; control). months glaucoma & External trabeculectomy in contralateral eyes. 2/2 normal Euthanasia at fail or 52wks Ahmed GDD 1 dog 11 months Yes. Required Yes 356 Secondary glaucoma medication Surgical pseudophakia (10 weeks after cataract surgery) Ahmed GDD 1 dog 5 months Yes No 357 Secondary glaucoma Two weeks after blunt trauma Ahmed GDD 9 dogs 24 months Poorly defined 5/9 (55.6%) 358 Primary glaucoma 2/9 (22.2%) IOP >20mmHg NON-VALVED Baerveldt GDD 5 dogs 8-13 4/5 (80%) 4/5 (80%) 359 Primary glaucoma months after requiring after 3 dogs with clinical glaucoma; 2 predisposed revisional requiring but before the onset of clinically recognisable surgery revisional glaucoma surgery GDD placed in subconjunctival space Baerveldt GDD 13 eyes (11 dogs) 0.5-24 4/13 (30.8%) 4/13 (30.8%) 360 Primary glaucoma months 6 dogs with clinical glaucoma; 7 predisposed but before the onset of clinically recognisable glaucoma Fontal sinus 4 dogs (pilot study) 6-18 weeks 4/4 (100%) Not reported 361 shunt (silicone Clinically normal dogs tubing) Euthanised at 6, 8, 16, 18 weeks Cullen frontal 3 dogs 2 days – 9 1/3 (33.3%) 362 sinus shunt Primary glaucoma months COMBINED PROCEDURE TSCP and Ahmed 51 dogs 2-83 39/51 (76%) 12/29 (41%) 349 GDD Primary glaucoma months Cycloablation 19 eyes (18 dogs) 3 – min. 12 14/19 (73.7%) 11/19 263 and valved GDD No classification of glaucoma aetiology months (57.9%) 2 dogs lost to follow up (included as fails)

Different devices have been trialled in dogs in both clinical and research settings, and as stand-alone or combination procedures (Table 1.12). Factors for consideration when placing a GDD include the size and material of both the tube (lumen and external diameter) and the scleral end-plate, the presence of a valve, the use of adjunctive antifibrotics and antimetabolites, and where the shunt diverts AH. AH is most P a g e | 69 commonly diverted to the subconjunctival space, and other locations for AH diversion include the frontal sinus, suprachoroidal space, and into the sclera (with use of a flap).347

Blind eyes End-stage surgery is recommended for the definitive treatment of eyes that are blind, buphthalmic and a source of pain. Enucleation of the globe is performed to remove chronic glaucomatous globes and should be performed in all cases where glaucoma is secondary to an intraocular tumour. Intraorbital prosthetics (polymethylmethacrylate or silicone) may be placed for cosmesis following enucleation or exenteration.347 If glaucoma is not due to neoplasia and the cornea is healthy and intact, evisceration with placement of an intrascleral prosthesis may be performed.363 Pharmacological ablation (using gentamicin or cidofovir) is an alternative technique for use in cases where general anaesthesia is contraindicated, or where financial constraints preclude other treatment options.364-369

P a g e | 70

CHAPTER TWO DEVELOPMENT OF A VISION IMPAIRMENT SCORE FOR THE ASSESSMENT OF FUNCTIONAL VISION IN DOGS: INITIAL EVIDENCE OF VALIDITY, RELIABILITY AND RESPONSIVENESS

The following is the re-formatted manuscript published by John Wiley & Sons, Pty Ltd: Graham KL, Reid J, Whittaker CJG, Hall EJS, Caruso K, McCowan CI, White A. Development of a vision impairment score for the assessment of functional vision in dogs: initial evidence of validity, reliability and responsiveness. Veterinary Ophthalmology. 5th March 2019 DOI 10.1111/vop.12656

ABSTRACT Aim: To describe the development and initial validation of a questionnaire measuring functional vision in dogs. Methods: A 17-item survey was designed to quantify functional vision in dogs. The Vision Impairment Score (VIS) was determined by summing responses to each question. Questions were assigned to one of five subcategories: overall vision, daily activities, peripheral vision, near vision, and distance vision. Content validity was established during development phases, and construct validity via comparing results of known groups (blind vs sighted; normal vs impaired vision, surgery to improve vision vs non-restorative surgery), and through factor analysis. Concurrent criterion validity was determined with use of a validated health-related quality-of-life (HRQL) assessment tool. Reliability and responsiveness assessments were investigated using intraclass correlation coefficient (ICC) and effect size (ES) respectively. Results: Responses (221) from 201 dog owners were included. Compared to sighted dogs (n=153), blind dogs (n=48) had a higher VIS and greater impairment in all subcategories. Among sighted dogs, a higher VIS was obtained in dogs with low vision compared to those with normal vision (p<0.001). A higher VIS was associated with poorer HRQL (p<0.001). Perfect reliability was obtained for 6/17 questions, and excellent reliability for 11/17 questions (intraclass correlation 1.0 and >0.9 respectively), and the VIS was highly responsive to therapeutic intervention (effect size 1.46). Conclusion: Results suggest the VIS may be clinically useful in assessing and obtaining a quantifiable measure of functional vision in dogs. Ongoing validation of the tool for clinical use is needed. P a g e | 71

INTRODUCTION Vision is a complex process involving light and motion perception,370 visual perspective, field of view, depth perception, visual acuity, colour and form perception.371 The construction and validation of reliable methods to assess vision in dogs are therefore, also complex,372-374 and there remains no gold standard method of measuring vision in dogs.

Vision impairment is reported to affect daily activities and health-related quality of life (HRQL) in people.375-380 Self-reported assessment of functional vision (via questionnaires) is strongly correlated with visual acuity377-379,381,382 and other objective measures of visual function in people with impaired vision.383,384 In the case of infants and young children who cannot self-report, daily visual activities have been correlated with preferential looking visual acuity.385 These activities were subsequently used to calculate a Visual Ability Score (VAS) using predetermined questions for a parent or guardian. We hypothesised that dog owners could provide similar answers about the performance of daily activities in a similar way to a child’s parent/guardian385 and that these could form the basis of a visual impairment scale to inform the measurement of functional vision. A recently published questionnaire for dog owners, the Canine Visual Function Instrument374 contains five sections of which only one concerns activity, but there is no published information relating to the contribution of activity to the total score.

Validity and reliability are key properties of any questionnaire instrument designed to measure health-related aspects and some evidence of these should be demonstrated before clinical use. Three types of validity are described – content, criterion and construct. Content validity, generally established during questionnaire development, ensures that all aspects of the subject being measured are covered without extraneous features included. Criterion validity applies when the new instrument compares well with an existing ‘gold standard’. Where no such gold standard exists, concurrent criterion validity (comparison with a validated related measure) may be sought as an alternative. Construct validity ensures that the test measures what it claims to measure. There are several ways of establishing construct validity including by factor analysis (FA). The FA examines the relationships between the variables - in this case items in a questionnaire - and clusters them into a small number of homogenous groups. Groupings of variables revealed by FA are termed factors and the association between a variable and factor is expressed as a factor loading of the variable (values 0 and 1), where the higher the loading the closer the association. Factorial validity is demonstrated if, after FA, an interpretable factor structure fits the construct (in this case visual impairment) that the questionnaire instrument was designed to measure. Known groups validity is another form of construct validation whereby hypotheses are formulated about how scores generated by the instrument will differ between groups, and these hypotheses are tested.386 A reliable instrument will produce the same score when an unchanging subject is measured at two points in P a g e | 72 time by the same observer (test retest, repeatability/intra-rater reliability), or when two or more people measure the same subject at the same time (reproducibility/inter-rater reliability). Additionally, where an instrument is intended to evaluate clinical change, responsiveness – the ability to detect a clinically significant change – is equally important. The aim of this study was to develop a questionnaire – the Visual Impairment Score (VIS) – to provide a measure of functional vision in dogs by assessing their performance in daily activities. As part of the early development of this tool intended for clinical use, our aim was also to provide initial evidence for its construct validity, concurrent criterion validity, intra-rater reliability and responsiveness. MATERIALS AND METHODS Questionnaire development Item generation A bank of potential questions for inclusion in the VIS was generated from questions typically asked by a veterinary ophthalmologist when assessing vision, and based on questions used in self- and proxy-reports of vision in people, adapted for a canine population.377-379,381-383 Selection of wording for each question was designed to minimise response bias.387,388 The authors (KLG, KC, AW) considered each question. Questions were excluded on the following basis: irrelevant to vision, if the response relied on clinical findings (rather than the owner’s observation), if the intended meaning was not easily interpreted on reading the question, and if factors other than vision were considered to have too major an impact. Content validation Selected questions were presented to a group of evaluators (10 pet owners and 10 veterinarians) via email in order to determine the content validation index (CVI). Evaluators were asked to rate the relevance of each question on a five-point Likert scale (strongly favourable, favourable, neutral, unfavourable, and strongly unfavourable) in specifically assessing vision.389 Evaluators could also provide comments regarding clarity of the question, and the reasoning for the assessment of relevance. Relevance scores were dichotomised into ‘favourable’ (strongly favourable, favourable) and unfavourable (unfavourable, strongly unfavourable) for analysis. If an evaluator rated a question ‘neutral’, it was not included in either group. To determine the CVI for each question, the proportion of evaluators rating the question as ‘favourable’ was calculated.389 Questions were excluded from the VIS if the CVI was ≤0.60. Questions were revised (wording, or grading on Likert scale), or excluded if any evaluator questioned clarity of that question. P a g e | 73

Questionnaire construction Response options for the final questions comprising the questionnaire were numerical (0 to 3- or 4-scale) with higher scores indicative of greater impairment. Each VIS component question was allocated into one of five subcategories exploring aspects of visual function: general visual ability; peripheral vision; performance in daily activities; near vision; and distance vision. Testing for validity, reliability and responsiveness Procedures involving animals were conducted in accordance with the ARVO Statement of the Use of Animals in Ophthalmic and Vision Research. Informed consent was obtained from all participants prior to inclusion in the study which took place between June 2016 and June 2017. Owners of dogs presenting to a veterinarian in one of six veterinary hospitals in Sydney, Australia, including two ophthalmology referral services, and four non- specialty hospitals, for assessment of ocular, eyelid or vision-related concerns were invited to participate in the study. Clinical examination For each dog, a clinical history was obtained to identify concurrent diseases and medical conditions unrelated to the eye(s). An ophthalmic examination was performed including slit lamp biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, UK), indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, NSW Australia), and tonometry (Icare Tonovet, Icare, Finland). Dogs were categorised as having functional vision in both, one or no eyes. An eye was considered blind when there was no menace response. Eyes were considered normal when ophthalmic examination revealed no abnormalities, and there was no history or evidence of reduced vision. Navigational ability was assessed in sighted eyes without clear ocular media, dogs with evidence of retinal disease, and in dogs presenting for assessment of vision loss or impairment. In these cases, binocular and monocular navigational ability was assessed using an improvised obstacle course within the veterinary facility under photopic conditions. A 14mm occlusive contact lens (Capricornia Contact Lens, Slacks Creek, Queensland) was used to assess monocular navigational ability. Sighted eyes were categorised as having low vision if errors were made in navigating obstacles (the dog contacted, or hesitated to approach obstacles compared to assessment under binocular conditions), or if ocular media obscured complete visualisation of structures on examination. Construct validity (i) Factorial validity Factor analysis (FA) was carried out (Minitab v.18) on the first VIS P a g e | 74

questionnaire completed by each participating owner. A principal components method of FA with a varimax rotation was performed. Input variables were all item ratings. Loadings were sorted, and items with loadings of less than 0.3 were excluded.390 Guided by a scree test and the Kaiser criterion, the interpretability of possible factor models was examined. A factor model was sought that accounted for an acceptable amount of the variability in the data, was readily interpretable, and did not include any factors containing only one or two items.391

(ii) Known groups validity Using the first assessment for each dog, box plots and descriptive statistics were used to identify differences between dogs that were blind, those with one sighted eye and those with two sighted eyes, followed by formal statistical analysis using non – parametric Kruskal-Wallis tests. Linear discrimination analysis was used to determine the ability of the VIS to differentiate dogs that were blind, those with one sighted eye, and those with two sighted eyes. The following hypotheses were tested: (1) that the VIS scores will differ between dogs that are blind and those with vision; (2) that the VIS score is higher when both eyes are blind compared to one blind eye, and the score is lowest when both eyes are sighted; (3) that among sighted dogs, the VIS scores will differ between dogs with some degree of vision impairment (‘low vision’), and those without any impairment; and (4) that when surgery had been performed with the aim of improving vision (e.g. cataract removal or retinal reattachment), the resultant VIS would be lower compared to when vision improvement or restoration was not an expected outcome (e.g. corneal or conjunctival grafting procedures). Concurrent criterion validity Participating owners were requested to complete a previously validated health- related quality of life (HRQL) instrument, VetMetrica392 online at the same time as they completed their first VIS questionnaire. VetMetrica is a web-based generic HRQL questionnaire for dogs consisting of 22 questions for the dog owner, which generates scores in four domains of quality of life - energetic/enthusiastic (E/E), happy/content (H/C), active/comfortable (A/C), and calm/relaxed (C/R), with automatic and instantaneous transformation of responses.392 Raw domain scores (0-6) are transformed to a 0-100 scale where 50 represents the average healthy dog. Owners and their dogs were registered on the VetMetrica website* by one of the authors (KLG) and domain scores were downloaded from the website onto an excel spreadsheet for analysis. Box plots and descriptive statistics were used to identify differences between dogs that were blind, those with one sighted eye and those with two sighted eyes, followed by a formal statistical analysis using non-parametric Kruskal-Wallis tests. Pearson correlation coefficient was used to measure the strength of linear associations between the VIS and measures of HRQL. P a g e | 75

Test retest reliability and responsiveness A number of owners of dogs that had stable vision and ophthalmic disease completed two assessments not less than two weeks apart, and test-retest reliability was assessed using the intra-class correlation coefficient (ICC). A one-way random model was assumed where the subjects (dogs) are assumed random.393 This correlation was interpreted as excellent if >0.90, good when 0.80-0.89, adequate when 0.70-0.79 and with limited applicability if <0.70.394 Responsiveness to change was assessed by repeat surveying of dog owners before and after vision-restoring, or vision-improving interventions (surgery performed by a veterinary ophthalmologist to remove a cataract, or surgery to reattach a detached retina). Standardised response mean (SRM) and effect size (ES) were used as measure of responsiveness and calculated using test-retest data as the average difference between the two measurements.395 SRM was calculated as the average difference divided by the standard deviation of the differences between the paired measurements. ES was calculated using baseline standard deviation as the average difference divided by the standard deviation of the first measurement. Effects were considered to be small when the score was between 0.2 and 0.5, moderate between 0.5 and 0.8, and large when above 0.8.396 Statistical analysis of sample population Statistical analyses were conducted using commercially available software (IBM SPSS Statistics version 24, and Minitab 18). All data were considered non parametric for analyses. For all analyses, a p value < 0.05 was accepted as statistically significant. Baseline characteristics were compared between dogs grouped into one of three groups: dogs that were blind, had sight in one eye, and had sight in two eyes. For dichotomous variables (sex, neuter status, purebred status, the presence of an extraocular medical condition, use of medication for extraocular disease, whether there was previous ocular surgery, and whether previous surgery was performed with the goal of improving vision), a Mann-Whitney test was used to identify any statistically significant differences in baseline characteristics between groups based on the number of sighted eyes. Coding was performed to categorise dog age (<4 years, 4-7 years, 8-11 years, 12-14 years, and >14 years). For variables with more than two categories (age, size, number of sighted eyes, quality of vision in sighted eyes, affected part of the eye, concurrent systemic disease, and surgically treated ophthalmic diseases) a Kruskal-Wallis test was used. Vision impairment scores were compared between groups of dogs. When subjects were divided into two groups for comparison (presence of systemic disease, blind versus sighted dogs, impaired versus normal vision in sighted eyes), a Mann-Whitney test was used to compare scores. When there were more than two groups (number of sighted eyes), a Kruskal-Wallis test was used. Pearson correlation coefficient was P a g e | 76

used to measure the strength of linear associations between the VIS and measures of HRQL.

Table 2. 1. Signalment and baseline characteristics of dogs grouped according to the number of visual eyes

Blind One visual eye Two visual eyes P-value* n=48 n=59 n=94 Age (years) 9.7 ± 2.9 10.4 ± 3.3 8.6 ± 4.1 0.006 9 (4-16) 11 (3-16) 8.5 (0.15-17) Sex (male) 44.2% 39.1% 59.0% 0.028 Desexed 96.2% 98.6% 92% 0.144 Purebred dog 59.6% 52.2% 68% 0.114 Size of dog 0.969 - small 65.4% 59.4% 57.0% - medium 17.3% 27.5% 31.0% - large 17.3% 13.0% 12.0% Systemic disease 32.7% 15.9% 39.2% 0.005 - inflammatory/DJ 4.2%% 0 3.0% D 2.1% 6.8% 1.0% - cardiac 0 1.7% 1.0% - neoplastic 16.7% 5.1% 19.0% - diabetes 2.1% 3.4% 5.0% Diseased part of eye - none 0 24.6% 34.2% <0.001 - cornea 5.8% 10.1% 26.1% - lens 36.5% 50.7% 73.2% - retina 42.3% 24.6% 6.5% - glaucoma 25% 56.5% 6.5% - extraocular 0 4.3% 6.5% - other intraocular 1.9% 2.9% 4% - >1 of above 23.1% 63.8% 16% Previous ocular surgery 21.2% 52.2% 52.1% <0.001 - cornea 0 0 17.3% - lens 27.3% 52.8% 78.8% - retina 0 8.3% 1.9% - glaucoma 72.7% 38.9% 1.9% DJD = degenerative joint disease. Results are listed as a percentage of the dogs in that group unless otherwise stated. *p value indicates the result of analysis comparing the variable between groups based on the number of sighted eyes; statistically significant p values (<0.05) are listed in bold

P a g e | 77

RESULTS A total of 258 VIS questionnaires were distributed. There were 221 responses from 201 owners, giving an overall response rate of 85.7%. Fifty-two responses were completed on paper and 169 responses completed online. Completion of the questionnaire by the owner took 10-15 minutes. Data were collected for 48 dogs that were blind, 59 dogs that had vision in one eye only, and 94 dogs that had vision in both eyes. Of these 201 dogs, 98 dogs had no co- morbidities. Signalment and other baseline characteristics of the dogs are summarised in Table 2.1. Among 153 dogs with vision, 65 dogs (42.5%) were considered to have low vision. Eighteen dogs with low vision had one eye (one eye previously enucleated), and dogs with low vision had disease affecting both the anterior and posterior chambers (Table 2.2). No eye that was normal on ophthalmic examination (n=59) was functionally blind and no dog with a normal ophthalmic examination and clear ocular media had difficulties navigating obstacles or clinical signs of reduced vision. Five dogs with no pathology identified on ophthalmic examination were considered to have some degree of vision impairment. Each of these dogs was geriatric (>12 years), and had nuclear sclerosis with or without iris atrophy. One eye with extraocular disease was blind following presumed traumatic proptosis of the globe despite surgical repositioning of the globe at the time of diagnosis.

Table 2. 2. Characteristics of dogs classified with ‘low vision’

DOGS PERCENTAGE NUMBER OF DOGS - TOTAL 65 100 - WITH ONE EYE ONLY 18 27.69 - WITH ONE NORMAL EYE 2 3.08 OPHTHALMIC ABNORMALITY - NUCLEAR SCLEROSIS, IRIS ATROPHY 1 1.54 - KERATOCONJUNCTIVITIS SICCA 1 1.54 - CORNEAL OPACITY 12 18.46 - PREVIOUS CATARACT SURGERY 17 26.15 - CATARACT 8 12.31 - GLAUCOMA 5 7.69 - >1 ABNORMALITY 14 21.54

P a g e | 78

In evaluating potential associations between the VIS and baseline data of the dogs, no statistically significant association was present between VIS and age (p=0.183), sex (VIS: female 23.5, male 20, p=0.747), neutering (VIS: desexed 23, entire: 18, p=0.797), pedigree status (p=0.570), size (VIS: small dog 22.5, medium 18, large 26.5, p=0.071), the presence of concurrent systemic disease (VIS: no systemic disease 21, systemic disease 22.5, p=0.952) or which part of the eye had previous surgery (VIS: cornea 8, lens 11, retina 12.5, p=0.061). Dogs with a concurrent diagnosed medical condition (n=68) had a higher general health impairment score (mean 1.8/4; median 2/4) compared to dogs with no concurrent systemic condition (mean 1.0/4; median 1/4) (p<0.001). Dogs with diabetes mellitus (n=8) had a greater general health impairment score (mean 1.9/4; median 2/4) compared to dogs without diabetes (mean 1.1/4; median 1/4) (p<0.001). Content validation From the original bank of 25 potential questions, six were excluded on the basis of poor clarity when presented as a question on a Likert-scale (n=4), and not specifically assessing vision (n=2). The remaining 19 items were assessed for relevance by 10 pet owners and 10 veterinarians. No evaluator rated any of the questions ‘neutral’ in how that item assessed vision. Two items ‘How much pain does your dog have in or around the eyes?’, and ‘Would you say your dog’s overall health is:’ did not meet content validity criteria (0.55 and 0.40 respectively). Of the remaining 17 questions, 6 were re- worded for clarity (avoid double negatives, and changing the direction of the Likert scale), and all fulfilled the relevance criteria (Table 2.3). Assessment of validity, reliability and responsiveness of the VIS Construct validity (i) Factorial validity FA was performed on responses for the 17 items, obtained from 201 questionnaires. A scree plot and the Kaiser criterion suggested that a model containing one factor was most appropriate. This model accounted for 66% of the variance and was easily interpreted as representing visual impairment (Table 2.4).

P a g e | 79

Figure 2. 1. Boxplot demonstrating lower Vision Impairment Scores (VIS) associated with increasing number of sighted eyes in dogs. The horizontal line within the box represents mean score out of a maximum possible score of 67.

Figure 2. 2. Vision impairment and subcategory scores in dogs according to number of sighted eyes. Statistically significant difference between dogs with 0, 1 and 2 sighted eyes in all categories except for ocular pain (p=0.246). For all categories, including overall VIS, except for ocular pain higher scores (indicating increased impairment) were associated with decreasing number of sighted eyes. The general health question is illustrated, though not used in calculation of the VIS P a g e | 80

Table 2. 3. Questionnaire items* Score Content Validity Test-retest Item range Item-Total Relevance ICC correlation General vision Would you say your dog’s vision is: 0-4 0.818 (<0.001) 0.90 1 Is your dog’s vision worse in bright light (for instance on a sunny day)? 0-3 0.731 (<0.001) 0.75 1 Is your dog’s vision worse in dim light (for instance at dusk)? 0-3 0.508 (0.022) 0.75 1 Peripheral vision Does your dog seem to startle when people/objects approach from the side or 0-4 0.537 (0.015) 0.85 0.975 are in his/her peripheral vision? Performance of daily activities Does your dog have difficulty going down steps or curb? 0-4 0.639 (0.002) 0.85 0.978 Can your dog see hand signals? 0-4 0.549 (0.012) 0.80 0.950 Does your dog have difficulty finding a ball when at play? 0-4* 0.544 (0.013) 0.65 0.979 Does your dog have difficulty navigating strange environments (e.g. vet clinic, 0-3 0.534 (0.015) 0.90 0.952 unfamiliar houses, on walks)? Does your dog’s vision affect him/her from doing his/her normal daily 0-3 0.721 (<0.001) 0.85 1 activities? Does your dog play less (at home), or have less walks/outings because of their 0-3 0.836 (<0.001) 0.90 1 vision? Does your dog see people walking by? 0-4 0.802 (<0.001) 0.90 0.988 Does your dog see other pets/dogs walking by? 0-4 0.559 (0.010) 0.85 0.974 Does your dog look at food/treats when they are presented? 0-4 0.870 (<0.001) 0.75 0.985 Does your dog get on/off beds, couches, car same as he/she used to? 0-4 0.575 (0.008) 0.70 0.981 Near vision Does your dog see insects crawling or flying? 0-4 0.544 (0.013) 0.75 0.984 Does your dog bump into furniture or walls? 0-4 0.508 (0.022) 0.90 1 Distance vision Does your dog not recognise you until you are relatively close? 0-4 0.621 (0.019) 0.70 0.976 P a g e | 81

*Option for the owner to select Never has been interested in balls; aAbsolute difference between the median rating of the top and bottom 25% scores, p value in parentheses

Table 2. 4. Factor analysis: Sorted Unrotated Factor Loadings and Communalities

Variable Factor1 Communality See people 0.915 0.838 See pets 0.912 0.831 Bumping 0.898 0.806 See hand signals 0.893 0.798 Vision rating by owner 0.887 0.786 Difficulties in strange places 0.874 0.764 See insects 0.868 0.753 Recognise owner at a distance 0.847 0.718 Sees ball 0.842 0.709 Looks at treats/food when offered 0.809 0.654 Difficulties down steps 0.790 0.624 Daily activities affected 0.755 0.570 Fewer walks/play 0.754 0.568 Worse in bright light 0.730 0.534 Can get on/off couch 0.730 0.533 Worse in dim light 0.700 0.491 Startles by something in peripheral vision 0.536 0.287 Variance 11.265 11.265 % Var 0.663 0.663

(ii) Known groups validity Figure 3.1 and Table 3.5 present an overview of the VIS and how scores differed between blind (median VIS 48) and sighted eyes (median VIS 14). Figure 2.2 shows the detail in relation to the subcategories included in the VIS. In sighted dogs a higher VIS was recorded for dogs with low vision (median VIS 11) compared to those with normal vision (VIS 21) (p<0.001) (Table 2.5). Compared to dogs with no vision impairment, dogs with low vision had a higher score in the P a g e | 82

following subcategories: performance of daily activities (normal vision: 3/37, low vision: 9/37, p<0.001), near vision (normal vision: 1/8, low vision: 2/8, p<0.001) and distant vision (normal vision: 0/4, low vision: 1/4, p<0.001). When previous vision-improving/restoring surgery had been performed (VIS 13), scores indicated better overall vision compared to when surgery was not performed to improve vision (VIS 24, p<0.001). All four hypotheses were upheld (Figs 2.1, 2.2 and Table 2.5) demonstrating that the VIS can distinguish between degrees of visual impairment as well as the impact of restorative surgery.

Figure 2. 3. Vision impairment score in groups with known characteristics (‘known groups’)

Hypothes Variable Groups compared Number Median VIS p is of dogs (range) value 1 Presence of Blind dogs 48 48 (34-70) <0.001 vision Sighted dogs 153 14 (0-54)

2 Number of 0 48 48 (34-70) <0.001 sighted eyes 1 59 23 (2-54) 2 94 7 (0-51) 3 Quality of Absent 48 48 (34-70) <0.001 vision Low 65 19.5 (0-54) Normal 88 11 (0-48) 4 Goal of ocular Improve vision 70 13 (0-59) <0.001 surgery Unrelated to vision 19 24 (7-59) P value calculated to identify statistically significant differences between the groups for each hypothesis tested; Mann-Whitney test used for hypotheses 1, 4; Kruskal- Wallis test for hypotheses 2,3.

Linear discrimination analysis (using cross validation) showed that the instrument correctly classified 90% of dogs according to whether they were blind or had sight. In the case of discrimination between blind versus sight in one eye versus sight in two eyes, the overall classification rate was 71% with 94%, 50% and 71% correctly classified for blind, one sighted eye, and two sighted eyes respectively. Concurrent criterion validity HRQL questionnaires were completed for 98 dogs with and without visual impairment in the absence of comorbidities. Two dogs had missing data, and so data from 96 dogs were analysed. The HRQL profiles of blind dogs compared with dogs with one or two sighted eyes are shown in Figure 2.3. In general, the profile of all dogs was lower than that of the average healthy dog (score of 50). The median scores for the three groups P a g e | 83 were significantly different for domains Energetic/Enthusiastic, and Happy/Content (p=0.002) but not for Active/Comfortable (p=0.136) or Calm/Relaxed (p=0.165). There was a moderate negative correlation between the VIS and Energetic/ Enthusiastic (R=-0.61, p<0.001), Happy/Content (R=-0.58, p<0.001) and Active/Comfortable (R=-0.45, p<0.001), and a small negative correlation with Calm/Relaxed (R=-0.3, p=0.003) (Fig 2.4).

Figure 2. 4. Plots of scores for four domains of HRQL (Energetic/Enthusiastic, Happy/Content, Active/Comfortable, Calm/Relaxed) generated by owners of 25 blind dogs, 28 dogs with sight in 1 eye and 43 dogs with sight in both eyes, using the web-based VetMetrica generic HRQL questionnaire instrument for dogs. Each blue box represents the scores obtained for between 25% (bottom line) and 75% (top line) of the group with the line in the middle representing the median score.

Table 2. 5. Responsiveness of impairment score to surgical intervention for vision improvement or restoration

VIS General Peripheral Daily task Near Distance vision vision performance vision vision

SRM 1.81 1.83 0.88 1.53 1.13 1.48 ES 1.46 1.43 0.63 1.37 1.05 1.58

SRM Standardised Response Mean; ES Effect Size. Data reported without units where small effects <0.20; moderate effects 0.50-0.80; large effects >0.80. P a g e | 84

Figure 2. 5. The relationship between the HRQL profile and Vision Impairment Score for 96 dogs with vision impairment, but no co-morbidities. Pearson’s correlation coefficients for Energetic/Enthusiastic, Happy/Content, Active/Comfortable and Calm/Relaxed are -0.61, -0.58. -0.45 and -0.30 respectively

Test-retest reliability and responsiveness Eleven owners of 11 dogs with stable ophthalmic disease and vision completed the questionnaire twice to determine test-retest reliability with a median 2.4 months between attempts (range 9-14 weeks). Perfect reliability (ICC=1) was obtained for 6/17 questions, and excellent reliability (ICC>0.9) for the remaining questions (Table 2.3). Responsiveness of the VIS to therapeutic intervention was assessed in 9 dogs that had surgical treatment that improved functional vision as determined by a veterinary ophthalmologist (phacoemulsification, n=7; retinal reattachment, n=2). The median interval between questionnaire attempts was 4.2 months (11 weeks- 6 months). The average score for each question before surgery was higher than after surgery in all dogs (p=0.001). There was a large level of responsiveness (SRM and ES) of the VIS and all subcategory scores to therapeutic intervention (Table 2.6).

DISCUSSION This study describes development of the VIS, a tool to quantify functional vision P a g e | 85 impairment in dogs. The VIS is intended as a clinical tool to facilitate monitoring of gradually progressive vision loss and response to therapeutic interventions that restore sight. In conjunction with clinical examination, existing diagnostic tests and therapeutics, the VIS has potential to support veterinary recommendations and a dog owner’s decision about the need for, or response to treatment. Furthermore initial evidence is presented for the tool’s validity and reliability, both prerequisites for clinical use, and its responsiveness to clinical change. Content validity, generally established during questionnaire development, ensures that all aspects of the subject being measured are covered without extraneous features included. To ensure that the original item bank was comprehensive potential questions for the VIS were derived from 2 sources – similar questionnaires for children with visual disturbance whose parents observe and report on relevant behaviours 380,385 and questions asked during veterinary consultations for ophthalmic conditions in dogs. Initial selection of items from this list was carried out by the investigators on the basis of pre-determined criteria, but thereafter the remaining items were subject to validation by an expert group of veterinarians and veterinary ophthalmologists (10) and a group of 10 dog owners, as the intended users of the tool. A similar approach has been used in the development of structured questionnaires to evaluate pain and HRQL in humans and is considered to be an appropriate and relevant approach.397 Furthermore, the use of a content validity index to quantify content validity is relatively new to veterinary science398 and represents a more objective approach to the demonstration of content validity. Further support for content validity was provided by the similarity between VIS questions and those used in the final section of the recently described canine visual function instrument.374 This section comprises 12 questions, and was “designed to seek information from dog owners about how the quality of vision might interfere with the ability of their dogs to execute specific activities”.374 All VIS questions measure visual impairment, but they were subjectively allocated into subcategories by the authors to facilitate interpretation of the scale. Subcategory impairment scores increased significantly between dogs based on the number of sighted eyes, thus providing additional support for content validity. However further investigations are required to determine whether there is clinical merit to subcategorisation as little correlation between subscales has been shown in equivalent human studies.399 The VIS is modelled on comparable observer (parent) completed questionnaires measuring vision in children that correlate with standard objective measures of vision.380,385 However, the reliance on an observer does introduce bias. Respondent bias can compromise the accuracy of an observational questionnaire where the emotional bond between subject and respondent can influence the responses400 and the human animal bond is known to be strong. Accordingly, in this study, attempts were made to minimise respondent bias by formulating questions relating to specific P a g e | 86 behaviours and by wording questions as objectively as possible. Strategies such as these have achieved good results in the human medical field.380,384,401,402 Field testing was used to provide evidence for construct and criterion validity, reliability and responsiveness, with convenience sampling of dogs attending several veterinary clinics for a variety of ocular, eyelid or vision-related conditions. The response rate for questionnaire completion was very high and may in part be due to the fact that all respondents had presented their dog to a veterinary facility for assessment and were aware of the importance of the study. Whether these same results would be obtained with a random sample of the dog-owning population or a less committed group of dog owners with ophthalmic disease is a matter of conjecture. Furthermore, because of the relatively small sample size, it is unlikely that the full range of ophthalmic problems encountered in veterinary practice was represented. Further testing to confirm applicability of the VIS in a clinical setting where clinical disease is frequently advanced, and occurs with confounding factors, would be appropriate. To provide strong evidence for construct validity two different approaches were employed, namely factorial validity and ‘known groups’. Factorial validity requires the statistical analysis of correlations between responses to the items of an instrument to determine if an underlying factor structure fits the construct upon which the instrument was developed. The simple one factor model revealed in this study was considered to compare well with visual impairment. A good factor model also accounts for a reasonable amount of the variability in the test data. Our model accounted for 66% of the variability which compares favourably with FA of the Vision Impairment Profile, a 32-item questionnaire developed to measure the impact of vision impairment on daily activities in people403 which accounted for 57% of the variability and also with owner reported questionnaires designed to measure attributes of dogs – 57%404 and 63%.405 In the ‘known-groups’ approach to construct validation, predictions are made about how scores obtained with the instrument will differ between groups where a difference is expected to occur, and these are tested. In this study hypotheses were designed to reflect the intended clinical use of the VIS which is to assess functional vision regardless of the aetiology of any impairment. For example, we hypothesised that VIS scores would differ between blind and sighted dogs (hypotheses one and two), and between dogs with some vision impairment compared with those with normal vision (hypothesis three). This is in contrast to the recently reported validation of the Canine Visual Function Instrument, in which dogs were grouped according to their ophthalmic diagnosis (normal, cataract, and progressive retinal atrophy).374 However a clinical diagnosis may not correspond with a dog’s visual capability. For example, dogs with incipient cataracts may have no detectable vision impairment, while a mature cataract renders an eye blind. Compared with the visual discriminant nature of the first three hypotheses (e.g. blind versus sighted, low vision vs normal vision), the fourth hypothesis was concerned with comparing the expected VIS scores P a g e | 87 following surgical treatment to improve vision in one group, but not in the other. This number and range of hypotheses tested, all of which were upheld, provide substantial evidence for known groups construct validity. The ability of the VIS to differentiate blind and sighted dogs was considered to be very good (10% misclassification) and compared favourably with tools to measure chronic pain in dogs (12%)406 and pain in infants (13%),407 but the scale was less able to distinguish blindness from degrees of visual impairment. This could have been the result of measurement error, or because there may be little difference between the impairment of function when one eye is sighted compared to two. Furthermore, 51% of dogs tested had co-morbidities and factors unrelated to vision may affect a dog’s scores on the VIS. For example, a response indicating that a dog no longer jumps on/off couches is not specific for sight, but may indicate the presence of degenerative joint disease. This is a limitation that must be considered when using the test, and why the authors recommend that the VIS is used as an adjunct to clinical judgement, rather than as a standalone tool. Quality of life has been shown to deteriorate with vision loss in people376,379,381,382,399,401 and we hypothesised the same would apply in dogs. As no gold standard measure of vision in dogs exists, evidence for concurrent criterion validity was sought by investigating the correlation between the VIS and a validated HRQL measure for dogs. A generic HRQL tool such as VetMetrica is sensitive to the impact of any condition on QOL and accordingly the impact of vision impairment on QOL was investigated only in dogs with no co-morbidities. The VetMetrica tool produces a HRQL profile for the dog with scores in four domains of QOL allowing evaluation of areas in which QOL might change. In this study there was a significant difference in Energetic/Enthusiastic and Happy/Content between blind dogs, dogs with one-, and dogs with two-sighted eyes. In the remaining domains of QOL (Calm/Relaxed, Active/Comfortable), the HRQL scores were not statistically different between dogs based on the number of sighted eyes. Whether a dog is calm and relaxed can be strongly influenced by the dog’s nature and that may not change with vision impairment. It is perhaps surprising that blindness did not have a statistically significant impact on dogs’ activity levels using this tool. This may be due to the small sample size (25 blind, 28 with sight in one eye and 43 with sight in both), however factors such as the chronicity of vision loss, and whether vision loss was sudden or gradual may affect a dog’s activity levels, but investigating these factors was beyond the scope of this study. Having established that visual impairment does have a measurable impact on canine QOL in two domains, it was deemed appropriate to use the HRQL tool as a measure of concurrent criterion validity. Decreasing vision (increasing VIS) was associated with decreasing QOL in dogs with a moderate correlation for Energetic/Enthusiastic, Happy/Content, Active/Comfortable and a small correlation for Calm/Relaxed, thus confirming concurrent criterion validity of the VIS. In this study reliability of the VIS was investigated using test-retest of dogs with stable P a g e | 88 vision. Using the same owner as observer (intra-rater repeatability), the ICC values >0.9 were considered to be excellent. While a minimum two week period is recommended for test – retest investigations408 to avoid recall bias, the minimum period between assessments in this study was 9 weeks. However inter-rater reliability (use of the VIS by multiple owners of the same dog) was not tested and, until such time as it is, the authors recommend that all assessments are made by the same owner. One of the primary aims in development of the VIS was to identify changes in functional vision with disease progression and in response to therapeutic interventions. Assessment of the ability to evaluate clinically significant change (responsiveness) is therefore important. Responsiveness has been proposed as a criterion, in addition to reliability and validity, to evaluate instruments that are designed as health measures.409 The VIS was found to be responsive to clinical change when surgery was performed to improve vision (cataract or retinal reattachment surgery), demonstrating potential as a clinical tool for measuring functional change. However this was based on a very small cohort of dogs with limited aetiology of visual impairment and further studies are required to confirm its generalisability. Similarly, demonstration of an instrument’s validity is an ongoing process, requiring testing of different populations in alternative situations386 and accordingly further testing of the VIS will be the focus of future studies. The VIS is intended for use as a clinical outcome measure and the results of the studies presented here provide initial evidence of validity, reliability and responsiveness. This validation indicates the potential of the VIS as a clinically useful method to quantify vision in dogs when used to complement ophthalmic examination and veterinary assessment of vision. FOOTNOTES *www.vetmetrica.com

ACKNOWLEDGEMENTS The authors would like to thank the veterinarians, veterinary nurses and the owners for their assistance with recruitment and participation. The authors wish to acknowledge the teams at the Eye Clinic for Animals, Sydney Animal Hospitals and Great Western Animal Hospital in Sydney, Australia for their instrumental work in the conduct of this study. P a g e | 89

CHAPTER THREE A FORCED-CHOICE PREFERENTIAL LOOKING TASK FOR THE ASSESSMENT OF VISION IN DOGS: PILOT STUDY

The following is the re-formatted manuscript published by The British Small Animal Veterinary Association: Graham KL, Byosiere S-E, Feng LC, Bennett PC, Caruso K, McCowan CI, White A. A forced-choice preferential looking task for the assessment of vision in dogs: pilot study. Journal of Small Animal Practice. 2018 Oct 17; DOI: 10.1111/jsap.12965

ABSTRACT Objective: To describe preliminary use of a forced-choice preferential looking task for the clinical assessment of vision in dogs. Design: Pilot study. Animals: Eighteen pet dogs. Methods: The vision of 18 pet dogs was investigated in two separate studies using a forced-choice preferential looking task: multiple observers watched eye, head and body movements on video recordings to identify cues suggesting when a dog had seen the feature of interest. Human observer reliability was determined using eight dogs and computer-generated stimuli. Visual acuity was assessed using computer- generated grating stimuli: in real-time, an observer watched each dog’s eye movement patterns and behaviour to decide whether each grating was seen. Stimuli were presented in a step-wise manner and were controlled by the observer. Acuity was estimated as the highest spatial frequency the dog was determined to have seen. Results: Median estimated visual acuity was better at 1 m compared to that at 3 m. Average test time was longer at a 3-m distance than at 1 m. Inter- and intra-observer reliability was better from 1 m than from 3 m. Conclusion: Preliminary use of a forced- choice preferential looking task for measurement of visual acuity in dogs has potential use as a clinical tool for the assessment of vision in dogs.

P a g e | 90

INTRODUCTION Accurate assessment of vision in veterinary patients is difficult, in part due to an inability to obtain objective and quantifiable measures of vision in a clinical setting. Our incomplete understanding of the components of canine vision, as well as the inability of non-verbal species to provide feedback, means that tools used to assess vision in humans are not directly transferable for use in dogs. Instead, assessment of canine visual performance in a veterinary clinical setting is limited to methods such as testing for a menace response, visual placing, tracking objects and navigating obstacle courses.225 These methods do not readily allow for objective assessments or quantification of visual function in veterinary patients. Visual acuity (VA) refers to the ability to distinguish details of an object with clarity410 and remains unsurpassed as a clinical tool for vision assessment in people.411 VA is determined by an interaction of optical and neural factors based on an animal’s anatomy and physiology.372 While VA is not a complete measure of visual function, a correlation with quality of life and visual function has been demonstrated, and VA has been used as a primary indicator of functional vision impairment in people.377-379,381- 383

Preferential looking (PL) tasks have been developed to assess the VA of infants and preverbal children and are based on the fact than an infant shown a boldly patterned target paired with a blank target of equal luminance will preferentially fixate the patterned target.412 In PL tasks the assessor is masked to the location and spatial frequency of the grating, and uses the child’s eye and/or head movement to judge the location of the stimulus.413,414 PL and operant procedures have been used to assess VA of normal infants and children415-417 as well as in those with known or suspected visual disorders418,419 and have been shown to be reliable and easy to administer in these patients.420,421

Due to a lack of objective measures of vision in animals, PL tasks may be a suitable means of obtaining measurable indicators of vision in these species. The aim of this study was to determine the suitability of a forced-choice PL (FPL) task for use in the clinical assessment of vision in dogs with no previous training or acclimatisation to the testing procedure. MATERIALS AND METHODS Two separate studies were performed to address each of the main variables present in a FPL task. Human observer reliability was initially assessed to determine whether the perception that the dog saw the stimulus was consistent between observers. Assessment of canine visual acuity was subsequently tested. For inclusion in either study an ophthalmic examination was required for all dogs. Examination included slit-lamp biomicroscopy, indirect ophthalmoscopy, tonometry (Icare TonoVet, Icare, Finland), quantitative measurement of tear production (Schirmer Tear Test; Merck Animal Health, NJ, USA), and paddle retinoscopy (Sussex Vision International, West Sussex UK). All procedures and protocols were conducted P a g e | 91 in accordance with the Association for Research in Vision and Ophthalmology statement on the Use of Animals in Ophthalmic and Vision Research and with approval from the University of Sydney Animal Ethics Committee. Observer (human) reliability assessment Animals Eight privately owned Lagotti Romagnoli that had been target trained to touch a stimulus (the larger circle displayed) on a screen for separate studies422-424 were included. Stimulus Stimuli were presented on two second generation Apple iPads (screen size 9.7 inch; screen resolution 1024x768 pixels) mounted adjacent to each other on their long side. Display settings on both devices were set to maximum brightness (device internal settings) throughout the testing procedures. A specialised optotype (letterform used in an eyechart) consisted of the letter ‘O’ in black, displayed on a white background as used in a standard Snellen visual acuity chart (constructed on a 5X5 grid so that the size of the critical detail subtended one fifth of the overall height). Each optotype was composed of lines to subtend a visual angle from the testing distance, conforming to the principles of standard logarithmic visual acuity charts. Acuity was reported on a logarithm of the minimum angle of resolution (logMAR) scale, which is a linear scale providing a measure of vision loss (lower logMAR values indicate better visual acuity). Optotypes were presented in order of decreasing size testing for visual acuities ranging from logMAR 1.6 to logMAR 0.3 (eight trials; one metre testing distance) and logMAR 11.48 to logMAR -0.12 (eleven trials; three metre testing distance) (Fig 3.1).

Figure 3. 1. Relative differences in the size of optotypes and their corresponding visual acuity values, as used in the assessment of human observer reliability. P a g e | 92

Experimental procedure Testing was conducted in the same enclosed room with the same personnel present for all trials. The owner was present but not involved in the testing procedure. Each dog participated in four sessions (two trials at three metre testing distance, followed by two at one metre testing distances) on the same day. An investigator, blinded to the stimulus display location, gently held the dog’s head at a pre-marked distance (one or three metres from the screens). Investigators A and B held a shield (94x55cm) to obstruct the dog’s view of the stimuli between each trial. When the dog looked toward the shield (as identified by investigators A and B), it was lifted so the stimuli were visible to the dog. The handling investigator (investigator C) held the dog stationary for one second before releasing the dog’s collar and allowing the dog to approach and touch the screen displaying the optotype. The dog received a food reward (Good-o treats®) from a remote-controlled food dispensing device (Treat & Train Remote Reward Dog Trainer®, PetSafe, Dubai, UAE) for correct identification of the target and for returning to the start point for the next trial.

For each trial, two investigators (A and B) were obliged to decide as to whether the dog ‘chose’ its target before leaving the testing mark. If the dog did not touch the screen displaying the optotype, or if the dog’s ‘choice’ was determined to have been made after leaving the testing mark, the previous (larger) optotype was presented in the next trial. The experimental session ended if three consecutive errors were made when presented with the same optotype, with correct identification of the previous (larger) optotype between each of these errors.

Each session was recorded using two digital video cameras (Sony HDR-CX405 Memory Stick HD Camcorder), one positioned directly above the screens, facing the dog and the other positioned behind the dog to record the dog and both screens throughout each trial.

Analysis Three investigators independently reviewed video footage of each trial at a later date. Each investigator was required to make a choice on five separate criteria based on their observation of that video clip (Table 3.1). All trials (517) were assessed by one investigator on two separate occasions, two weeks apart, to determine intra-rater reliability. Percentage agreement and a one-way intraclass correlation coefficient (ICC) was used to determine inter- and intra-rater reliability at both testing distances.

P a g e | 93

Table 3. 1. Criteria for investigator decision making in proof of concept trials (determining inter- and intra-rater reliability in a forced-choice looking task)

Question for consideration by observing investigator Possible answers Which screen did the dog choose? Left Right Did the dog make a decision before leaving the marked Yes No position? In which direction did the dog make its first gaze or head Left Right movement? In which direction did the dog make its first move? Left Right After looking at the screens, did the dog later look away Yes No from the stimuli?

Canine vision (estimated visual acuity) assessment Case selection Ten privately owned dogs with no history of ophthalmic disease or vision impairment, and no prior training to interact or respond to a computer-generated stimulus were included. Dogs were excluded if there was any refractive error detected on streak retinoscopy. Stimulus Stimuli were generated using PsychoPy coding software425 on an Apple MacBook Pro (Intel HD Graphics 4000 1536 MB; Apple MacBook Pro 2011) and presented on a 21.5- inch screen (1920 X 1080 pixels) resulting in 72 and 212 pixels per degree, and a horizontal field of view of 27° and 9° when viewed from one and three metres respectively. Testing was conducted in a room with overhead lights on and average luminance on the stimulus 30 candelas/m2. A stimulus panel consisted of one patterned stimulus and one homogenous grey stimulus. The patterned stimulus was a static square sinusoidal grating that varied in spatial frequency (0.28 to 15 cycles/degree) with average screen luminance of 85 candelas/m2 and Michelson contrast of 90% (Fig 3.2). Luminance of the grey stimulus matched the mean luminance of the patterned stimulus, and was close to that of the background screen to minimise contrast at the edge of the stimuli. Two additional ‘anchor’ stimuli, a blank and a 10 cycle/degree grating, were available for use as necessary to provide anchor points for the investigator’s judgment during testing. P a g e | 94

Figure 3. 2. a) Schematic diagram representing stimulus panel used to assess visual acuity. The position of the grating stimulus (top left of screen) and homogenous grey stimulus (bottom right of screen) were randomly generated by software. Contrast and luminance are not measured in these representative diagrams; b) Sample animated diagram displayed on screen (with accompanying sound) to re- focus dog’s attention at the discretion of the investigator (iStock.com/bandian1122); c-d) Altered positions of stimuli on panel displaying gratings of the same spatial frequency. This sequence was selected by the investigator (up arrow key) when the investigator could not make a ‘choice’ on where/whether the dog looked at the stimulus, and could be repeated until the investigator made a choice.

Experimental procedure A rapid-assessment FPL task based on that used in the assessment of visual acuity in human infants421 was designed for use in this study. Grating acuity was tested with the use of a two-alternative forced-choice paradigm in which the grating was oriented vertically in a random order and displayed in a random position on the screen. Dogs were given breaks as deemed necessary by the investigator. Each dog had two sessions at two separate clinic visits. Each testing session consisted of assessment from both one and three metre distances. All sessions were conducted in the same room, with the same equipment and by the same investigator to ensure consistency between conditions. The dog was held at the appropriate viewing distance for that session. The investigator (observer) was shielded from view by a screen behind the display monitor to minimise P a g e | 95 distractions in the dog’s visual field. This observer had a laptop which controlled display of stimuli. Testing was conducted in a manner similar to that reported for the rapid test of infant acuity.421 Briefly, stimuli of low frequency as well as blank anchor cards were initially displayed allowing the investigator to become familiar with the looking style of each dog in the presence and absence of a visible stimulus. Stimuli were then presented in a random order. The investigator could also choose to display anchor cards as a known stimulus, or a separate animation with accompanying sound to catch the dog’s attention (Fig 3.2b). For each grating, the investigator pressed a key according to their choice: left arrow indicating ‘yes’ (the dog had seen the grating) or right arrow indicating ‘no’ (if the dog appeared to not see the grating). If the investigator could not make a choice, the up- arrow key presented a stimulus of the same spatial frequency (Fig 3.2c-d). Gratings of the same spatial frequency were displayed until the left or right arrow key were pressed. The estimated VA was the highest spatial frequency with a ‘yes’ decision in that subset. Data Analysis For analyses of estimated visual acuity, notations were converted to the logMAR equivalent to enable calculation of the geometric mean and standard deviation as previously described.426 A Wilcoxon matched-pairs signed rank test was identify differences in estimated acuity between testing distances. Spearman’s correlation was used to measure the strength and direction of the association between estimated acuity at one and three metre testing distances. Estimates of VA were converted into commonly used units. RESULTS Baseline demographics, ophthalmic findings and refractive errors are presented in Table 3.2. There were no structural ophthalmic abnormalities, history or clinical signs of ocular disease, nor any evidence of vision impairment in any dog. One dog used in proof of concept trials (G) was being treated for diabetes mellitus diagnosed 15 months prior to testing, but had no evidence of cataract formation and physical examination was unremarkable. Observer (human) reliability outcomes A total of 517 trials (243 trials at one metre; 274 trials at three metres) were completed. One dog (M) did not approach the stimuli in sequential trials during the third session and was withdrawn from the study at that point. The remaining seven dogs completed all four sessions. Intra-observer reliability for all trials was 82.4% (426/517); 94.2% (ICC 0.91, 95% CI [0.85, 0.94]) at one metre (229/243), and 79.8% (ICC 0.72, 95% CI [0.67, 0.76]) at three metres (197/247) (p<0.001). Inter-observer reliability was better from one metre (81.3%, ICC 0.84, 95% CI [0.75, 0.83]) than from three metres (60.2%, ICC 0.64, 95% CI [0.58, 0.66]) (p<0.001). Agreement between observer decision and correct stimulus location was highest when the cue was the direction the dog looked before moving P a g e | 96

(67%), and when the cue was any one or more behaviours/movements of the dog (67%). Canine vision (estimated visual acuity) outcomes All dogs completed all trials conducted and a total of 40 estimates of VA were made from each viewing distance. Average test time was longer from a three metre testing distance (6.2 minutes) compared to one metre (4.8 minutes) (p=0.032). The highest estimated VA for each dog from each testing distance, and expressed in frequently used units, is presented in Table 2.3. The median estimated VA was better from one metre (logMAR 0.6; range 0.3-1.1; 95% CI [0.663-0.76]) compared to three metres (logMAR 0.8; range 0.6-1.3; 95% CI [0.78-0.88]) (p<0.001) (Fig 3.3). There was a moderate, positive monotonic correlation between estimated VA from one and three-metres (rs=0.510, n=40, p<0.001) (Fig 3.4). Poorer estimates of acuity were obtained with greater frequency from three metres than from one metre.

Figure 3. 3. Box and whisker plot showing median (range) of estimated visual acuity in dogs. Mean VA represented by cross (+). P a g e | 97

Table 3. 2. Characteristics and baseline data of subjects

Dog Age Breed Sex Neuro- Ocular Refractive Fundus* IOP STT (mm/min) (yrs) ophthal media error (mmHg) 1* 2 (diopters) LEFT RIGHT LEFT RIGHT LEFT RIGHT Proof of concept A 2.1 LR FE WNL Clear WNL 12 14 24 24 X 2.3 LR MN WNL Clear WNL 14 16 28 28 B 4.4 LR FE WNL Clear 0 0 WNL 10 13 18 23 E 3.0 LR FE WNL Clear +2.0 +1.5 WNL 11 11 25 24 G 2.6 LR FS WNL Clear 0 +0.5 WNL 13 10 26 23 H 6.8 LR ME WNL Clear 0 +0.5 WNL 10 8 21 25 L 3.0 LR FE WNL Clear -0.5 -1.0 WNL 13 14 23 23 M 3.8 LR FE WNL Clear -0.5 -0.5 WNL 16 17 26 19 Pilot study 1 6 Pug X MN WNL Clear 0 0 WNL 15 18 17 18 2 4 Terrier X MN WNL Clear 0 0 WNL 12 11 22 24 3 5 Mini FS WNL Clear 0 0 WNL 16 13 19 19 4 2 Mini Dachshund MN WNL Clear 0 0 WNL 15 14 21 24 5 6 Terrier X ME WNL Clear 0 0 WNL 11 13 16 17 6 3 Labrador FS WNL Clear 0 0 WNL 14 17 28 25 7 3 Maltese X FS WNL Clear 0 0 WNL 14 12 23 20 8 4 Poodle X MN WNL Clear 0 0 WNL 17 15 24 21 9 5 Cocker Spaniel FS WNL Clear 0 0 WNL 12 13 17 21 10 2 Poodle X MN WNL Clear 0 0 WNL 9 11 27 25 P a g e | 98

Table 3. 3. Highest estimated visual acuity recorded for each dog expressed in commonly reported notations.

1m testing distance 3m testing distance LogMAR Snellen Snellen Decimal Cycles per LogMAR Snellen Snellen Decimal Cycles per (20 ft) (6m) degree (20 ft) (6m) degree 1 1.00 20/200 6/60 0.10 3.0 1.10 20/240 6/72 0.08 2.5 2 0.83 20/135 6/41 0.15 4.6 0.88 20/152 6/45 0.13 4.0 3 0.73 20/107 6/32 0.19 5.9 0.88 20/152 6/45 0.13 4.0 4 0.50 20/60 6/18 0.33 10.0 0.80 20/120 6/36 0.17 5.0 5 0.50 20/60 6/18 0.33 10.0 0.75 20/112 6/34 0.18 5.25 6 0.83 20/135 6/41 0.15 4.6 0.83 20/135 6/41 0.15 4.6 7 0.63 20/85 6/26 0.24 7.1 0.70 20/100 6/30 0.20 6.25 8 0.83 20/135 6/41 0.15 4.6 0.80 20/120 6/36 0.17 5.0 9 0.75 20/112 6/34 0.18 5.25 1.05 20/224 6/67 0.09 2.75 10 0.50 20/60 6/18 0.33 10 0.60 20/80 6/24 0.25 7.5

P a g e | 99

Figure 3. 4. Correlation between estimates of visual acuity as measured from 1- metre (x-axis) and 3-metres (y-axis)

DISCUSSION In this study, preliminary use of a FPL task was used to obtain measures of visual function (estimated VA) in dogs with clinically normal eyes. Development of a test that provides an objective and quantifiable measure of vision would provide a useful tool in both veterinary ophthalmology and neurology to aid in diagnosis and management of diseases such as glaucoma and retinopathies that result in vision loss, and in assessment of central nervous system pathologies affecting the visual pathway. Because formal two-alternative FPL testing may require a large number of trials,427 such procedures are not practical for use in pet dogs, especially in a clinical setting. The study presented here was modelled on a rapid FPL test for use in infants,421 where limitations in time and co-operation are also encountered. FPL procedures are used clinically and allow accurate subjective estimates of binocular and monocular VA in human infants. 416,420,421 This pilot study suggests clinical use of a FPL task may be of similar use in dogs. The use of optokinetic nystagmus373,428-430 and other preliminary tools374,431 to measure vision have been described, yet, there is no ‘gold standard’ measure of visual acuity or visual function in dogs. Establishing validity of this test as a measure of VA is therefore difficult. The initial trials assessing human observer reliability were performed to identify specific criteria (behaviours or cues) that might be used by the investigators to reliably make a choice about whether the dog saw the stimulus. In this P a g e | 100 study, no specific cue (head turn, eye movement, body movement) achieved greater agreement between observer perception and stimulus location than when one specific cue was used. No defined behaviour or cue was therefore used to determine whether the dog had seen the stimulus in the subsequent pilot study. This lack of definition increases subjectivity in this FPL task, and this limitation is also encountered in the use of these tests with human infants.416,420,421 With a range of behaviours that may indicate the subject has seen the stimulus, limiting identification to one or more specific cues may decrease sensitivity of the tool. To minimise the impact of this variable in this pilot study, personnel experienced in working with animals (rather than utilizing personnel experienced in conducting similar assessments in humans) were used to assess inter- and intra-observer reliability, and a single observer used in assessments of estimated visual acuity. Caution must be used in drawing conclusions as this subjectivity means validation of the FPL task as a clinical tool, requires a large sample population with multiple assessors. It is for this reason, that the results using those dogs previously trained to interact with a screen-display, and approached the stimulus, were discarded in the analysis of VA estimates. In clinical veterinary medicine, it is more appropriate that this limitation be considered rather than avoided, and thus the term ‘estimated visual acuity’ has been used in this study with the acknowledgement of different measures of acuity obtained with behavioural and electrophysiologic testing paradigms in domestic animal species.432

Inter- and intra-rater reliability were assessed using percent agreement in this study. Justification for using this method included the small sample size, prior training of the dogs and experience of each observer, as well as the use of pre-determined ‘rules’ by which observations should be made. Percent agreement does not account for choices where the observer may have guessed the location of the stimulus. Therefore, there is a possibility that inter- and intra-reliability may overestimate the true agreement among observers in this study. Further assessment of this task as a clinical tool requires appropriate sample sizes that allow valid statistical analyses including the assessment of disagreement, ‘no agreement’, and consistency within and between observers. The nature of a FPL task presents two variables that might influence results: the visual capability and performance of the dog and the investigator’s perception of whether the presented stimulus was seen by the dog. Despite the limitations of both dog and observer variables in a FPL task, estimated VAs were comparable to those previously reported using alternative methods.372,373,433,434 To limit the number of variables in this pilot study, only emmetropic dogs were included. In addition, results from those dogs with prior training responding to a computer-generated stimulus that were allowed to approach the stimulus (as opposed to remaining at the testing distance) were excluded from analysis in calculating estimated VA. Ongoing validation of the technique to include dogs with measured refractive errors is indicated as these are reported in association with size,435 breed,436-439 age,438,440,441 and skull shape.435,441 With validation of the technique, some determination of the impact of factors such as hypermetropia, myopia and astigmatism on functional vision, might be made.

Although this testing method was developed based on techniques used in non-verbal people, caution must be used in using VA obtained using these procedures to compare visual capabilities between these species. According to the World Health Organization P a g e | 101 criteria for defining low vision,442 most dogs in this study would be classified as having low vision. Whilst the apparent lack of any visual impairment in all dogs assessed may be attributed to the paucity of appropriate measures of vision in domestic animal species, it is essential to consider the other components of vision, and that the relative importance of these factors varies considerably between species. For this reason, a clinical tool such as this FPL task must be utilised as what it is – a means of obtaining a measurable indicator of one of the components of vision. Other available tests as well as clinical and ophthalmic examinations must be performed in conjunction with any tool such as a FPL task in order to obtain the most well-rounded assessment possible.

Traditional testing of VA involves the use of high contrast optotypes or gratings even though VA is understood to transcend optical resolution.411 The importance of discrimination mechanisms other than resolution (e.g. contrast reduction, noise and contour perturbation) on canine vision remains unknown and was not explored in this study. Measures of various components of vision, including functional vision using obstacle courses,443,444 visual acuity using target detection,434 colour perception,445 brightness discrimination,446 and techniques to assess specific areas of the visual pathway447,448 have been reported in dogs. These techniques require specialised equipment, facilities, anaesthesia and/or prior training, making them unsuitable for widespread clinical use in veterinary patients. In this pilot study, testing of untrained dogs in a veterinary clinic using equipment already available in the clinic (with purpose-developed software) was timely (4.8 – 6.2 minutes per testing interval) and well tolerated by dogs. Validation of the technique, and assessment of responsiveness (the ability to detect change) in a large population with multiple testers (observers) as would be expected in a clinical setting, is therefore warranted.

Differences between testing paradigms make direct comparisons between studies of VA difficult. By converting reported acuities into a common unit of measure, comparisons between studies may be made, though these conversions should be interpreted with caution as different paradigms measure different types of acuity. Studies in humans have shown fairly good agreement between optotype and grating acuity estimates in normal populations, 449,450 and discrepant findings with disease.451,452 This shows a need for a single paradigm that can be implemented on a large scale to allow direct comparisons within and between dogs. This study describes preliminary use of a forced-choice preferential looking task with potential clinical use in the veterinary setting. To justify intervention and treatment of vision impairing ophthalmic diseases in domestic animals and to determine its efficacy, veterinarians and vision researchers need to incorporate objective and/or quantifiable measures of visual function into clinical practice and vision research. The intended use of this test is for the clinical assessment of vision in dogs in conjunction with other tests and clinical examination. This must be considered when interpreting results which are not designed to provide definitive measures of canine visual acuity. The authors suggest these results warrant further investigation to assess the validity, reliability and responsiveness of the technique by different observers using a suitably sized study population, including dogs with and without vision-impairing disease.

P a g e | 102

Online supplementary files

Figure 3. 5. Schematic of frontal view displaying stimulus (as presented to dog) showing display monitor in front of a screen concealing the investigator. The investigator views the dog through a small opening in the screen (black rectangle).

Figure 3. 6. Schematic representation of testing room in which vision assessments were conducting demonstrating layout of apparatus and personnel (investigator in chair behind desk using laptop which controls the stimulus presentation; handler/owner restraining dog on stool at 1m or 3m testing distance)

P a g e | 103

CHAPTER FOUR A MODIFIED PROTOCOL FOR ASSESSMENT OF THE PUPILLARY LIGHT REFLEX IN DOGS PREDISPOSED TO GLAUCOMA

The following is a re-formatted manuscript under review for publication by The British Small Animal Veterinary Association in the Journal of Small Animal Practice: Graham KL, McCowan CI, White AJR. A modified protocol for assessment of the pupillary light reflex in dogs predisposed to glaucoma.

ABSTRACT Objective: To investigate the use of a modified pupillometry technique in dogs without chemical restraint. Animals: Twenty-five pet dogs. Procedures: Following dark adaptation, pupillary light reflexes (PLRs) were assessed in six dogs (12 eyes) with retinal disease (Sudden Acquired Retinal Degeneration; SARDS), the unaffected eye of eight dogs with unilateral primary angle closure glaucoma (PACG), and in 11 healthy dogs (18 eyes). Responses to red, blue and white lights were tested and relative pupil sizes subsequently determined based on video recordings of each test. Results: Mean testing time was 2.3minutes, excluding time for dark adaptation. Baseline pupil size in dogs with SARDS was greater than in normal (p=0.026) and predisposed eyes (p=0.048). Pupil constriction was reduced in predisposed compared to normal eyes when stimulated with high intensity blue (p=0.015) and red light (p=0.005). SARDS- affected eyes had less pupil constriction when stimulated with red light (p<0.001), low intensity blue light (p<0.001) and white light (p=0.005). Conclusion: Objective measures of pupil function were obtained from healthy and diseased eyes without the need for chemical restraint. These results suggest quantitative measures of pupil function may be obtained by clinicians in routine clinical practice for the assessment of neurological and ophthalmic diseases. Further investigations are warranted to validate the technique and evaluate its use in the management of canine glaucoma.

P a g e | 104

INTRODUCTION The pupillary light reflex (PLR) is an autonomic reflex controlling pupil diameter in response to light. For a long time, the PLR was believed to rely only on the activation of rod and cone photoreceptors in dim and daylight, respectively.453 However, measurable PLR activity was identified in the absence of rods and cones454 and the photopigment melanopsin455 found to cause a photo response of intrinsically photosensitive retinal ganglion cells (ipRGCs).456 Based on these findings, we now know the PLR is driven by rod-, cone- and ganglion cell-mediated activity. As the spectral sensitivity of ipRGCs differs from that of rod and cone photoreceptors, assessment of pupil responses to different wavelengths of light permits evaluation of rods, cones and ipRGCs. Pupillometry is a non-invasive, objective method that provides measures of pupil size and response. With the physiological differences between cellular responses within the retina, chromatic pupillometry - measuring pupil responses to different wavelengths and intensities of light - becomes a useful tool in evaluating visual function and has been used in assessment of glaucoma in people,457-460 and pupillometry techniques are also described in dogs.461-467

Chromatic pupillometry has been reported in investigations of the pathophysiology of sudden acquired retinal degeneration syndrome (SARDS). With the need for concerted efforts to improve our ability to detect and manage canine glaucoma,468 investigation of the potential value of chromatic PLR assessments in glaucoma is prudent. Monitoring of pupil responses in canine glaucoma is complicated by changes associated with the disease such as the use of topical prostaglandin-analogues that cause miosis, the development of peripheral anterior synechiae associated with disease, and iris ischaemia resulting from elevated intraocular pressure (IOP).469 These factors limit the use of pupillometry in glaucomatous eyes, however if neuroretinal function is impaired prior to the onset of these changes, chromatic pupillometry may have potential use in the identification of early changes associated with the canine glaucomas.

To the authors’ knowledge, there are no published reports describing quantitative measures of pupil function in dogs without chemical restraint. In this study, we sought to determine whether quantitative assessment of chromatic PLRs in fully conscious dogs was feasible in a veterinary clinical setting. Our secondary aim was to determine whether there were any differences in chromatic PLRs in eyes predisposed to glaucoma, and eyes with and without established disease.

MATERIALS AND METHODS Animals Twenty-five pet dogs registered as patients in NSW veterinary clinics were included. A complete ophthalmic examination, including slit lamp biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, UK), indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, NSW Australia), quantitative tear testing (Schirmer tear test I, Merck Animal Health, NJ, USA), rebound tonometry (Icare Tonovet, Icare, Finland) and P a g e | 105 fluorescein staining, was performed in all cases. Medical records were reviewed and signalment, ophthalmic findings and treatments, concurrent disease and treatments, were recorded for analysis.

Based on clinical examination and history, dogs were allocated into groups: (i) predisposed: eyes with an abnormal drainage angle on gonioscopy (narrowed iridocorneal angle and/or pectinate ligament dysplasia). For inclusion in this group, the contralateral eye had to have been diagnosed with primary glaucoma (clinical and/or histopathological) by a veterinary ophthalmologist; (ii) SARDS: eyes of dogs that presented to a veterinary ophthalmologist for sudden onset of bilateral blindness, with a diagnosis of SARDS made based on the absence of electrical activity on ERG and any visible abnormalities on ophthalmoscopy. Healthy dogs without evidence of clinical or historical ophthalmic or systemic disease were selected as controls.

Procedures All procedures were conducted with approval from the University of Sydney Animal Ethics Committee (2016/1004), and with the dog owner’s consent. PLRs were assessed in a darkened room after 10 minutes of dark adaptation. Each session was recorded using an infra-red video camera set up (1/2" 4-12mm F1.4 aspherical manual CMount infra-red lens, Tamron, Saitama Japan; CMOS, 87.2 fps, 752 x 480, 0.36 MPixel, 1/3", Global shutter, Aptina Imaging Corp. San Jose, California US). The left and right eyes were tested in each case, except in predisposed dogs where only the predisposed eye was assessed. The BPI-50 Precision Illuminator (Retinographics, Inc., CT, USA) was used to stimulate pupillary function under low (1000 lux +/- 5%) and high intensity (10,000 lux +/- 5%) red (660nm) and blue light (465nm) when held 4cm from the dog’s eye. The response to red light was assessed with 5 seconds of low intensity stimulus, then 5 seconds with high intensity light stimulation. After 10 minutes dark adaptation to allow pupil recovery, the PLR was assessed using the same protocol with a blue light. Assessment of PLR under white light conditions was performed using a slit lamp after a further 10 minutes dark adaptation. For each dog, this procedure was repeated in its entirety with a minimum of 2 hours interval, and the average pupil size from both sessions was used for analyses. A proportion of dogs returning for reassessment underwent a repeat of this testing procedure to determine test-retest reliability.

All video recordings were reviewed by the same investigator to determine the point of maximum constriction for each light stimulus. Still images from the video recording were obtained for analysis based on the following criteria: (1) the head and globe were positioned so that the iris plane was parallel to the camera, (2) a minimum of 180° of continuous corneal circumference was visible, and (3) the circumference of the pupil was clearly demarcated (Fig.4.1). Using open source ImageJ software470 the pupil size was measured as a ratio of the corneal diameter at the following time points: resting pupil (after dark adaptation and before stimulation with any light source), and at the point of maximum miosis when stimulated with low intensity red light, high intensity red light, low intensity blue light, high intensity blue light, and white light. Relative pupil constriction was then calculated using the following formula: (relative pupil diameter at maximum constriction) / (relative resting pupil diameter). Results of the relative pupil size are reported as ‘pupil size’ (Fig.4.2). P a g e | 106

Statistical analyses Statistical analyses were performed using commercially available software (GraphPad Prism 7). Kruskal-Wallis tests were used to compare baseline data (age, sex, neuter and purebred status, and the presence or absence of concurrent systemic disease) as well as the relative degree of pupil constriction between groups. Dunn’s multiple comparisons test was then performed to further assess rank differences between groups for which a statistically significant difference was detected. When statistically significant results were identified between normal and predisposed eyes, further evaluation of the variable was performed using a Mann-Whitney test by randomly selecting one eye from each normal dog to eliminate potential within-dog bias. Test- retest reliability was assessed using the ICC. A one-way random model was assumed where the subjects (dogs) are assumed random. This correlation was interpreted as excellent if >0.90, good when 0.80-0.89, adequate when 0.70-0.79 and with limited applicability if <0.70. Spearman’s rho correlation coefficient was used to measure the strength of associations between relative pupil sizes and baseline characteristics including age and IOP. To identify a possible correlation between the degree of pupil constriction with each light stimulus and the duration of disease, predisposed eyes were categorised as acute (<1 week) or chronic (>6 months) based on the time since signs consistent with glaucoma had been noted in the contralateral eye. The degree of pupil constriction for acute versus chronic eyes was compared using a Mann- Whitney test. For eyes with SARDS, the duration of disease was categorized as acute (<1 week), short term (2 months – 1 year) and chronic (>1 year), and a Kruskal-Wallis test used to compare groups. For all analyses, a p value <0.05 was considered statistically significant.

Figure 4. 1. Dog’s head positioned relative to the camera so that the iris plane is parallel to the camera. P a g e | 107

Figure 4. 2. Measurement of relative (a-b) resting and (c-d) stimulated pupil size: the relative size of the pupil (outlined in red) to the cornea (indicated by the peripheral iris margins; outlined in green) was determined by calculating the pupil diameter (yellow line) as a proportion of the corneal diameter (orange line); (c-d) to determine the degree of pupil constriction in response to a light, the same measurements were made, and the resulting relative pupil size measured as a proportion of the resting pupil size.

RESULTS Twenty-five dogs (38 eyes) were included. Eight dogs (eight eyes) were predisposed to glaucoma and six dogs (12 eyes) were diagnosed with SARDS. Eleven dogs (18 eyes) served as the control population. All normal and predisposed dogs had vision, and all dogs with SARDS were blind (Table 4.1). The age, sex, and proportion of purebred dogs were similar between groups (Table 4.2). IOPs were similar between groups, except that predisposed dogs had experienced a significantly higher IOP (64.5mmHg [40-84]) compared to the maximum recorded IOP in normal (12mmHg [8-16]) and SARDS dogs (11.5mmHg [6-17]) in the contralateral (glaucomatous) eye. There was no statistically significant correlation between age and the degree of pupil constriction in this study (overall, nor in individual groups).

Twenty-four of the 25 dogs enrolled in the study tolerated complete assessment of PLRs under red, blue and white light. Recording of the PLR in response to white light P a g e | 108 was not tolerated in one predisposed dog, although all other recordings were suitable for inclusion. Two assessments (one eye in two different dogs) were excluded due to positioning that precluded appropriate pupil measurements (Fig.4.3). Excluding time for dark adaptation (total 30mins), the average time to record PLRs in both eyes was 2.3minutes (1.8-3.1 minutes).

The duration of disease in SARDS-affected dogs was <1 week (n=2), between 2months – 1 year (n=2), and > 1 year (n=2). The estimated duration of glaucoma in the contralateral eye for predisposed dogs was <1 week (n=4), and >1 month (n=4). There was no statistically significant difference in pupil response to any light when comparing acute, short-term, and chronically affected SARDS eyes, and in predisposed dogs when the contralateral eye was either acutely or chronically affected by glaucoma.

Table 4. 1. Comparison of subject characteristics in dogs with normal eyes, eyes predisposed to primary angle closure glaucoma and the eyes of dogs diagnosed with sudden acquired retinal degeneration syndrome (SARDS)

Normal Predisposed SARDS P 11 dogs, 18 8 dogs, 8 eyes 6 dogs, 12 value* eyes eyes Age (years) 6 (3-10) 9 (6-11) 9 (8-11) 0.134 Purebred dog 45.5% (5/11) 37.5% (3/8) 33.3% (2/6) 0.634 Sex (female) 54.5% (6/11) 75% (6/8) 50% (3/6) 0.506 IOP (mmHg) 12 (8-16) 13.5 (12-16) 11.5 (6-17) 0.258 Highest IOP 12 (8-16) 64.5 (40-85)a 11.5 (6-17) <0.001 recorded for that dog (both eyes) Vision present 100% (18/18) 100% (8/8) 0% (0/12) <0.001 Results are reported as the median (range) unless otherwise stated. Results in bold are statistically significantly different from other groups for the variable tested. aHighest IOP documented for dogs in the predisposed group was in the contralateral eye for all cases in this group only. *Result of Kruskal-Wallis test of comparison between three groups.

P a g e | 109

Table 4. 2. Relative pupil diameter after stimulation

Normal Predisposed SARDS p value Red light (low 0.733a 0.748b 0.995a,b <0.001 intensity) 0.538 – 0.995 0.661 – 0.872 0.929 – 1.00 Red light (high 0.527a 0.673b 0.991a,b <0.001 intensity) 0.386 – 0.704 0.501 - 0.788 0.947 - 1.002 Blue light (low 0.497a 0.607 0.763a 0.002 intensity) 0.323 – 0.757 0.555 – 0.947 0.521 – 0.858 Blue light 0.395a 0.519a 0.479 0.036 (high 0.327 – 0.597 0.404 – 0.653 0.360 – 0.554 intensity) White light 0.536a 0.442b 0.792a,b 0.005 0.365 – 0.829 0.458 – 0.656 0.435 – 0.880 Pupil sizes are recorded as relative to baseline pupil size for each eye. Results reported as median and range. Superscripts identify pairs in which there is a statistically significant difference in pupil size between groups. P value of comparison between three groups.

Figure 4. 3. Unsuitable images for pupillometry due to a) eyelid/dazzle reflex obscuring >180° of corneal circumference, and b) head and globe off-centre so that iris plane is not parallel to light source and camera

P a g e | 110

Baseline pupil size Baseline relative pupil size in dogs with SARDS (86.9% of the corneal diameter; 95% CI [80.3%, 88.2%]) was greater compared to both normal (80.7% of corneal diameter; 95% CI [75.9%, 85.2%) (p=0.026) and predisposed eyes (79.8%; 95% CI [79.6%, 88.1%]) (p=0.048).

Red light stimulation A larger relative pupil size was identified in eyes with SARDS compared to both normal and predisposed eyes (p<0.001) when stimulated by low intensity red light (Fig.4).

When stimulated with high intensity red light, there was less pupil constriction was (larger pupil size) in eyes with SARDS compared to both normal (p<0.001) and predisposed eyes (p<0.001), and larger in predisposed compared to normal eyes (p=0.005) (Fig.4.4).

Blue light stimulation There was less pupil constriction in both predisposed and SARDS compared to normal eyes in response to low intensity blue light stimulation (p=0.024 and <0.001 respectively) (Fig.4.5). When stimulated by high intensity blue light, the degree of pupil constriction in predisposed eyes (relative pupil size 0.519 [0.404 – 0.653]) was less than that in normal eyes (0.395 [0.327 – 0.597]) (p=0.015).

To further evaluate the differences in response to high intensity blue light between normal and predisposed dogs, one randomly selected eye from each normal dog was compared with predisposed eyes to eliminate within-dog bias. The statistically significant difference between these groups was also evident on this analysis (normal: 0.360 [0.327 – 0.513]; predisposed: 0.519 [0.404 – 0.653]) (p=0.003) (Fig.4.6).

White light stimulation There was less pupil constriction in dogs diagnosed with SARDS compared to both normal (p<0.001) and predisposed eyes (p=0.004) when stimulated with white light (Fig.4.7).

Test-retest reliability Six dogs (6 eyes) predisposed to glaucoma had the same protocol for assessment of colormetric PLRs repeated between 3-15 days (median 6 days) after the initial assessment. There was excellent test-retest reliability in the degree of pupil constriction achieved with red, blue and white light stimulation (ICC 0.915).

P a g e | 111

Figure 4. 4. Relative baseline pupil sizes in normal, predisposed and SARDS eyes when stimulated with low and high intensity red light. The pupil size after stimulation is calculated relative to the baseline pupil size for each individual. The box represents 95% confidence interval with the mean represented by a horizontal line. The whiskers indicate the range of relative pupil sizes. The groups between which there was a statistically significant difference are indicated with their respective p value. P a g e | 112

Figure 4. 5. Relative baseline pupil sizes in normal, predisposed and SARDS eyes when stimulated with low and high intensity blue light. The pupil size after stimulation is calculated relative to the baseline pupil size for each individual. The box represents 95% confidence interval with the mean represented by a horizontal line. The whiskers indicate the range of relative pupil sizes. The groups between which there was a statistically significant difference are indicated with their respective p value.

P a g e | 113

Figure 4. 6. Pupil constriction after stimulation with high intensity blue light in (a) normal, and (b) predisposed eyes

P a g e | 114

Figure 4. 7. Relative baseline pupil sizes in normal, predisposed and SARDS eyes when stimulated with white light. The pupil size after stimulation is calculated relative to the baseline pupil size for each individual. The box represents 95% confidence interval with the mean represented by a horizontal line. The whiskers indicate the range of relative pupil sizes. The groups between which there was a statistically significant difference are indicated with their respective p value.

DISCUSSION In this study, we have shown that differences in quantitative measures of canine pupil function can be determined without sedation or anaesthesia, using readily available equipment. In doing so, we identified measurable differences in pupil responses between normal eyes and both eyes predisposed to PACG, and those with degenerate retinas. Pupil responses in SARDS cases were consistent with existing reports,461,467,471- 473 providing a form of external validation for our protocol. We suggest quantifying pupil responses with technology that is affordable to veterinarians may facilitate increasing assessment and monitoring of neuroretinal function in veterinary practice. P a g e | 115

We found that the PLR of predisposed eyes differed from that in normal eyes when stimulated by higher intensity compared to lower intensity light. With high intensity blue light stimulation failing to elicit an ‘abnormal’ response from SARDS-affected pupils, our findings suggest melanopsin-drive pupil responses may be altered, and a possible reason for the different response observed in predisposed eyes. Rod-driven pupil responses are purportedly mediated through ipRGCs,474 and attenuated pupil responses irrespective of the stimulus wavelength are reported in people with early POAG, suggesting altered synaptic transmission to ipRGCs, or reduced responsiveness of the outer photoreceptors in these eyes.475 We suggest the reduced pupil responses we identified to high intensity red and blue light in eyes predisposed to glaucoma may indicate ipRGC signalling is affected in these eyes. This change may represent early stages of canine glaucoma in these eyes at a stage we currently do not identify disease.

We identified no statistically significant difference in pupil response between predisposed and normal eyes with low intensity blue light, suggesting no difference in rod function between these groups,471 yet the response to high intensity blue light was suggestive of poorer ipRGC in predisposed compared to normal eyes.

Studies using formal pupillometry protocols that require chemical restraint describe an absent PLR with high intensity red light stimulation, and complete pupil constriction following high intensity blue light stimulation in all dogs with SARDS.461,467,471 We did not replicate these findings, our findings mirrored those reported in another study that did not use chemical restraint during pupillometry.476 Terakado et al476 suggested variation in their cases (dogs with progressive retinal atrophy) from existing reports may be attributed to progressive degeneration of photoreceptor and ganglion cells, and we speculate dogs with SARDS in our study to be affected in such a way. Our study included dogs both acutely and chronically affected by SARDS, where previous reports include only dogs more acutely affected.467,473

There are important limitations for consideration in interpreting these results. In attempting to quantify pupil responses using technologies commonly available in clinical settings and avoiding the risks and costs associated with sedation and/or anaesthesia, our study fails to replicate some conditions used in formal pupillometry assessments. Further studies necessary to validate this technique should investigate how pupil responses correlate with the number of optic nerve axons and whether age related changes to the iris sphincter muscle have any impact on findings if there is no evidence of iris atrophy on ophthalmoscopy.

Conducting pupillometry in conscious dogs is advantageous in that the variability in pupil responses associated with different drugs was avoided.477,478 However, problems including movement (head and eyes) and blinking are introduced. Despite this, only one dog did not tolerate the full protocol, and all dogs tolerated complete assessments with red and blue lights. Relatively short periods of testing (1.8-3.1 minutes) were broken up by 10 minute ‘breaks’ for dark adaptation between stimuli, meaning procedural timing for the dogs did not appear to cause undue stress. With the addition of a total 30 minutes dark adaptation (in 10 minute intervals), this testing P a g e | 116 procedure does take a considerable amount of time, although actual testing is not prolonged and was well tolerated by the dogs.

The sample size in our study was small due to the clinical nature of this prospective pilot study. Although we investigated correlations between variables in each of the groups, the small sample size, and the inability to definitively account for the chronicity of disease in each group limits the information we can obtain about each disease. Furthermore, there is a clear degree of ‘overlap’ in the quantitative measures obtained in this study, precluding the use of such parameters as a stand-alone diagnostic test. Validation with concurrent ERG studies, a larger sample population, with disease states of similar duration/chronicity, and ideally serial assessment of the same dogs/eyes, would further define the primary effect of the disease process, as well as the effect of secondary degenerative changes associated with chronicity, and thus aid in our understanding of the physiology of the PLR in disease states.

In summary, we demonstrated the use of chromatic pupillometry using a protocol modified for use in a heterogenous canine population without the need for chemical restraint. We demonstrated changes in pupillometry that suggested decreased ipRGC function in eyes predisposed to glaucoma, and decreased cone function in eyes with SARDS. We suggest the use of a modified protocol for pupillometry in a veterinary clinical setting has potential to further our understanding of early glaucoma in dogs. With validation, use of pupillometry in dogs without the need for sedation or anaesthesia, may facilitate early diagnosis, and monitoring of the disease and/or response to novel therapeutic interventions in veterinary practice.

ACKNOWLEDGEMENTS The authors would like to thank the veterinarians whose patients were recruited into this study, especially the Eye Clinic for Animals, Sydney who also provided access to the Retinographics chromatic light source used in this study. P a g e | 117

CHAPTER FIVE ANTERIOR SEGMENT ULTRASOUND BIOMICROSCOPY OF THE UNAFFECTED EYE IN DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA

The following is a re-formatted manuscript under review for publication by John Wiley & Sons, Pty Ltd in Veterinary Ophthalmology: Graham KL, McCowan CI, Caruso K, Whittaker CJG, White AJR. Anterior segment ultrasound biomicroscopy of the unaffected eye in dogs with primary angle closure glaucoma.

ABSTRACT Objective: To describe ultrasound biomicroscopic (UBM) findings of the anterior segment in the unaffected (predisposed) eye of dogs with primary angle closure glaucoma (PACG). Methods: UBM examination of the unaffected (predisposed) eye in 19 dogs with PACG, and in 21 healthy controls was performed. Measurements were obtained in the superior and temporal quadrants of each eye. Subject characteristics and measurements were compared between groups, and the effect of group and quadrant was investigated. Reproducibility and intra-rater reliability were evaluated using intra-class correlation coefficients (ICC). Results: There was a statistically significant effect of group on the ciliary cleft (CC) entry (p=0.002), mid-width (p=0.040), angle recess area (p=0.001), angle opening distance (p=0.001), the distance from Descemet’s membrane to the iris [p=0.005], from the limbus to first ciliary process [p=0.020], and peripheral iris thickness (p=0.005) which were smaller in predisposed compared to healthy eyes. There was a statistically significant effect of quadrant on median mid-CC width (superior: 0.19mm, temporal: 0.08mm) in predisposed eyes (p=0.031). When corrected for body size, anterior chamber depth was shallower in predisposed compared to control eyes. Intra-observer reliability was good or excellent for all variables except CC area (ICC 0.754) and geometric iridocorneal angle (ICC 0.712) which were considered adequate. Conclusion: Reproducible and reliable ultrasound biomicroscopic measurements of the canine anterior segment differed between eyes predisposed to PACG and healthy controls. Further investigation is warranted to identify whether these differences reflect early disease, or indicate progressive change to determine the merits of UBM in canine glaucoma. P a g e | 118

INTRODUCTION The identification of early disease has been identified as a critical factor to minimise morbidity associated with canine primary angle closure glaucoma (PACG).468 In people with PACG, quantitative measures of the anterior segment have been used to investigate the disease479-482 and have been presented as a potential population screening test.483 Despite increasing reports describing the use of technologies such as optical coherence technology and ultrasound biomicroscopy (UBM) in reference to canine PACG,58,78,250,251,255,484 the findings described in these reports typically lack validation in larger heterogenous populations. UBM is a high-resolution ultrasound technique allowing in vivo imaging of anterior segment structures at near light microscopic level.485 UBM findings have been reported for the eyes of healthy dogs without ophthalmic disease,255-257 in breeds with PACG58,250,251 and in a research colony of Beagles with PACG.78 Measurements of the anterior segment are affected by pharmacological pupil dilation486 and constriction,293 and by surgical lens removal.258,487 When compared to normal eyes, measures of the ciliary cleft (CC) were smaller both in the fellow eye of dogs with unilateral PACG, and in the eyes of unaffected ‘at risk’ breeds (those with a known hereditary predisposition to PACG) when gonioscopy revealed an abnormal iridocorneal angle (ICA).251 Such measures are proposed as important in the management of canine PACG and validation is indicated to determine the clinical significance of structural changes in eyes before the onset of clinical glaucoma that results in vision loss and discomfort. Indicators of early, and progressive glaucoma are highly desirable as pathophysiology is incompletely understood, and diagnosis typically made when disease is advanced. Anterior segment structures of relevance to the trabecular (conventional) outflow pathway include the ICA and CC.246,488,489 The ocular tunic (cornea and sclera),226,490,491 anterior chamber depth250 and lens position58 also change in diseased eyes. Although it is unclear when these structural changes occur, if detected early, their identification could enhance our understanding of, and ability to treat glaucoma in dogs. For techniques such as UBM to be useful as a clinical tool, reliable and reproducible findings are required.492 Anterior chamber measurements in humans typically use the scleral spur as the point of reference,485 however with no scleral spur in the canine eye, other points of reference are required.139 In this study we evaluate structural measurements of the anterior segment that have relevance to glaucoma, to identify structures that change in eyes predisposed to the development of glaucoma before the onset of clinical disease.

MATERIALS AND METHODS All procedures were conducted with approval from the University of Sydney Animal Ethics Committee (2016/1004) and with informed consent of each owner. P a g e | 119

Case selection Forty pet dogs (19 dogs with unilateral PACG and 21 healthy controls) registered as patients in veterinary hospitals in New South Wales, Australia were included between May 2016 and March 2019. All dogs had a complete ophthalmic examination including slit lamp biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, UK), indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, NSW Australia), quantitative tear testing (Schirmer tear test, Merck Animal Health, NJ, USA), and rebound tonometry (Icare Tonovet, Icare, Finland). PACG was diagnosed by a veterinary ophthalmologist based on a documented elevation in IOP (>25mmHg) in association with clinical signs of glaucoma including retinal degeneration and/or optic nerve degeneration (decreased size, loss of myelin, cupping of the optic nerve head). In all dogs diagnosed with PACG, gonioscopic examination of the contralateral (predisposed) eye revealed an abnormal iridocorneal angle (closed or narrowed angle, with or without pectinate ligament abnormalities) and there was no evidence of ophthalmic or systemic disease that could result in secondary glaucoma in any dog. Exclusion criteria included treatment of the predisposed (unaffected) eye with a topical prostaglandin analogue, and current or previous use of topical atropine for pharmacological mydriasis in either eye. Predisposed eyes being treated with a topical carbonic anhydrase inhibitor (17/19 cases) were not excluded from the study. For inclusion as a healthy control, in each dog there was no evidence of systemic or ophthalmic disease, and a gonioscopically normal (open) iridocorneal angle in each eye. Ultrasound biomicroscopy Image acquisition was performed using a 48mHz probe and high frequency ultrasound unit (Accutome, UBM Plus 24-6300-G-V, Malvern PA, USA). The same investigator performed examinations after topical administration of oxybuprocaine hydrochloride 0.4% (Minims, iNova Pharmaceuticals Pty Ltd, Chatswood NSW Australia) in a room with illumination of 800-1000 lumen/m2. The dog was sitting or lying in sternal recumbency with the eyelids held open manually for imaging. The transducer was positioned on the axial cornea to measure anterior chamber (AC) depth. For assessment of the ICA, the transducer was positioned perpendicular to the limbus at the 12 o’clock position in each eye, and at the 3 and 9 o’clock positions in the left and right eyes respectively to assess the drainage angle in the superior and temporal quadrants. Imaging sequences were recorded on a video loop using the unit’s software and saved for later analysis. Three separate images from sequential scans were obtained for each quadrant in each eye. Data analyses Images for analyses were obtained from the recording of each scan. In dogs with PACG, measurements were obtained of the predisposed eye. In healthy controls, one eye was selected at random using free online software (Random Tool 1.2.0.0, AxGarage). Image selection was based on the following criteria: perpendicular placement of the transducer at the limbus which was ascertained by the corneal P a g e | 120 epithelium, corneal endothelium and the anterior lens capsule being of the same echogenicity; and the image depicts the corneoscleral limbus, the first ciliary process, and the CC at its greatest width. Using internal calipers provided with the UBM device software, measurements were obtained from a single image selected from each video sequence. Anterior segment measurements were obtained as outlined in Table 5.1. Briefly, measures of the ICA included the geometric ICA, angle opening distances (AOD1 and AOD2), and angle recess area (ARA) (Fig.5.1a). Measures of the CC included CC entry, CC length, mid-CC width, CC area, and calculated CC area (Fig.5.1b). Additional measures in the anterior segment included anterior chamber depth (ACD), the distance from Schwalbe’s line (termination of Descemet’s membrane) to the anterior lens capsule (SLD), the distance from Schwalbe’s line to the iris surface (SID), the distance from the limbus to the first ciliary process (DLCP), iridociliary process distance (ICP), iris thickness, corneal thickness (CT) and scleral thickness (ST) (Fig.5.1c-d).256-258,486,492,493 For measurement of ACD, the selected image demonstrated the maximal distance between the corneal endothelium in the region of the vertex, and the most anterior aspect of the anterior lens capsule. P a g e | 121

Table 5. 1. Definition of variables measured

Variable Definition Illustration Drainage angle Geometric iridocorneal angle (ICA) Formed by the iris base in the region of the ciliary cleft entry and the inner corneoscleral junction Fig.1a (A) Angle opening distance1 (AOD1) The distance from a point 500μm from the end of Descemet’s membrane to the iris, on a perpendicular line to the iris Fig.1a (B) Angle opening distance2 (AOD2) Distance from the inner scleral surface to the iris on a perpendicular line from a point 0.45 X the length of DLCP when Fig.1a (C) measuring from the ciliary process Angle recess area (ARA) Enclosed by a line from the end of, and perpendicular to Descemet’s membrane to the iris, then following the iris Fig.1a (D) surface to the trabecular meshwork, and back to Descemet’s membrane Ciliary cleft Ciliary cleft entry (CC entry) Distance between corneoscleral limbus and iris root Fig.1b (A) Ciliary cleft length (CC length) Distance between pectinate ligament (or most anterior visible portion of trabecular meshwork), and the apex of the Fig.1b (B) ciliary cleft Mid-ciliary cleft width (mid- CC width) Distance between inner sclera and ciliary process in the central portion of the ciliary cleft Fig.1b (C) Ciliary cleft width (CC width) Distance from the superior surface of the iris root to the inner scleral surface (measured on a perpendicular line) Fig.1b (D) Ciliary cleft area (CC area) Area bounded by the ciliary cleft width, inner sclera, angle recess, and the superior surface of the iris root Fig.1b (E) Calculated ciliary cleft area (Calc CC area) Average depth of the ciliary cleft multiplied by the ciliary cleft width (formula for a trapezium) General anterior segment Anterior chamber depth (AC depth) Distance from the corneal endothelial surface at the vertex, to the anterior lens capsule Fig.1c (A) Schwalbe’s line to anterior lens capsule Distance between Schwalbe’s line (the termination of Descemet’s membrane) and the anterior lens capsule on a line Fig.1d (B) distance (SLD) perpendicular to the anterior lens capsule Distance from limbus to first ciliary process Distance between the limbus and the ciliary process Fig.1d (C) (DLCP) Schwalbe’s line to iris distance (SID) Shortest distance from the end of Descemet’s membrane to the iris surface Fig.1d (D) Iridociliary process distance (ICP) The distance from the end of Descemet’s membrane to the superior surface of the ciliary process Fig.1d (E) Corneal thickness, peripheral (pCT) Corneal thickness at the level 500μm axial to the termination of Descemet’s membrane Corneal thickness, axial (aCT) Corneal thickness measured at the axial cornea Scleral thickness1 (ST1) Scleral thickness at the level of measurement of the mid-CC width Scleral thickness2 (ST2) Scleral thickness at the level of the end of Descemet’s membrane Iris thickness Distance from the anterior to the posterior surface of the iris measured at the (i) base, (ii) mid-section, and (iii) tip of the iris P a g e | 122

Figure 5. 1. Variables measured (definitions outlined in Table 1) of a) the iridocorneal angle including the [A] geometric iridocorneal angle (ICA) (angle formed by red lines); [B-C] angle opening distances (AOD1, pale blue line and AOD2, yellow line); [D] angle recess area (ARA, enclosed by green dotted lines); b) the ciliary cleft (CC) including [A] CC entry (red line); [B] CC length (yellow line); [C] mid-CC width (green line); [D] CC width (pink line); [E] CC area (enclosed by pale blue dashed line); c) measurement of the axial anterior chamber depth (ACD, red line); and d) general measurements within the anterior chamber including [B] the distance from the end of Descemet’s membrane (Schwalbe’s line) to the anterior lens capsule (SLD, purple line); [C] the distance from the limbus to the first ciliary process (DLCP, green line); [D] the distance from the end of Descemet’s membrane to the iris (SID, red solid line); and [E] the iridociliary process distance (ICP, red dotted line). P a g e | 123

For each eye, image selection and measurement of all variables was repeated three times, and the average of these three measurements was used for analyses. Further calculations were then performed to correct for the variations in body size. Measurements of variables were divided by the distance between Schwalbe’s line and the anterior lens capsule (SLD) for each eye as previously described.257 Reproducibility of images was determined by calculating intra-class correlation coefficient (ICC) for each variable obtained in the three images used to obtain the mean measure used for analyses. Images from 40% of dogs (n=12) that were randomly selected using the same online software described above, were used to determine intra-rater reliability. Images were uploaded for analyses using open source software494 and measurements of all variables were made by the same observer on two occasions, on separate days. The ICC was determined for each variable. Statistical analysis Statistical analyses were conducted using commercially available software (IBM SPSS Statistics version 24; GraphPad Prism 7.0a). Subject characteristics were compared between groups using a student t-test for continuous variables (age, axial globe length, anterior chamber depth) and Fisher’s exact test for categorical data (sex, neuter and purebred status). To investigate the existence of an interaction effect (between group and sex when controlling for age and body weight), a two-way ANCOVA was performed. A two-way ANOVA was performed to determine effect of group (normal versus predisposed), quadrant (superior versus temporal), and the interaction between group and quadrant measurements of the ICA and CC. For variables in which a statistically significant effect on measures was identified, an unpaired t-test was performed to compare measurements of the same quadrant between groups, and a paired t-test performed to compare measurements obtained in different quadrants of the same eye. Data analyses were performed for direct measurements as well as for those corrected for SLD.257 For analyses of anterior segment measurements, a Bonferroni correction was conducted to minimise Type 1 Error. Using an original alpha-value of 0.05, the corrected p value ≤0.004 was accepted as being statistically significant. Intra-class correlation coefficients were calculated using a one-way random model where the subjects (dogs) were assumed random.393 Correlations were considered excellent when >0.90, good when 0.80-0.89, adequate when 0.70-0.79, and of limited applicability when <0.70.394

RESULTS Nineteen dogs predisposed to PACG and 21 healthy controls with no clinical or historical evidence of ophthalmic disease were included (Table 5.2). Measurements in the superior and temporal quadrants were obtained for each dog (Fig.5.2). There was no statistically significant difference in age (p=0.749), sex (p=0.360), nor whether the dog was purebred (p=0.057), or desexed (p=1.00), in axial globe length (p=0.432) or ACD (p=0.189) between predisposed dogs and healthy controls. There was no P a g e | 124 statistically significant interaction between sex and group on any variable measured when controlling for age and body weight.

Figure 5. 2. Sample images of ultrasound biomicroscopic examination of the drainage angle in the superior quadrant of a a) normal/healthy control (case 5); and an eye b) predisposed to developing glaucoma with goniodysgenesis (case 25)

P a g e | 125

Table 5. 2. Individual subject characteristics

I.D Group Age Breed Sex Weight Eye Contralateral eye (years) (kg) 1 Normal 10 Maltese X MN 4.8 R 2 Normal 2 Staffordshire Bull MN 6.9 R Terrier 3 Normal 10 Shih Tzu X FS 7.4 L 4 Normal 9 Shih Tzu X FS 5.1 L 5 Normal 9 Labradoodle FS 9.2 R 6 Normal 6 Labradoodle MN 31 L 7 Normal 3 Terrier X MN 10.1 L

8 Normal 12 Yorkshire Terrier MN 3.5 L Normal; not selected for study inclusion (randomisation) 9 Normal 7 Labradoodle MN 27 L 10 Normal 12 Maltese X FS 4.1 L 11 Normal 10 Labradoodle MN 29.3 L 12 Normal 6 Cocker Spaniel FS 14.3 R 13 Normal 8 Shih Tzu X FS 8.7 R 14 Normal 11 Labradoodle MN 31.2 L 15 Normal 12 Maltese X MN 4.1 L 16 Normal 10 Cocker Spaniel MN 17.3 L 17 Normal 0.5 CKCS FS 4.9 R 18 Normal 13 Poodle X FS 6.8 R 19 Normal 11 Maltese X FS 11.1 R 20 Normal 8 Terrier X FS 7.2 R 21 Normal 6 Cavalier King FS 12.1 L Charles Spaniel 22 Predisposed 14 Maltese MN 4.3 R Blind; PACG (23 months); IOP: 17mmHg, high 64mmHg 23 Predisposed 12 Maltese X FS 5.7 R Blind; PACG (26 months, enucleation, histopathology); IOP: high 51mmHg 24 Predisposed 9 Labradoodle FS 7.9 L Blind; PACG (2 weeks); 1 week post TSCP; IOP: 12mmHg, high 76mmHg 25 Predisposed 9 Labradoodle FS 8.2 L Visual; PACG (<1 week); new diagnosis; IOP: 48mmHg 26 Predisposed 6 GSD FS 28 L Blind; PACG (7 weeks, enucleation, histopathology); IOP high 66mmHg 27 Predisposed 6 Siberian Husky MN 27 L Blind; PACG (16 months, ISP, histopathology); IOP: high 72mmHg 28 Predisposed 12 Manchester Terrier FS 4.8 R Blind; PACG (31 months, buphthalmos); IOP 28mmHg, high 77mmHg 29 Predisposed 12 JRT FS 7.3 L Blind; PACG (13 months, enucleation); IOP high 55mmHg 30 Predisposed 5 Cocker Spaniel MN 15.3 L Blind; PACG (1 months); IOP: 22mmHg, high 68mmHg P a g e | 126

31 Predisposed 12 Poodle FS 6.9 R Blind; PACG (33 months, buphthalmos); IOP 9mmHg, high 39mmhg 32 Predisposed 6 Flat Coated FS 31 R Blind; PACG (7 months, enucleation, histopathology); IOP: high 57mmHg Retriever 33 Predisposed 6 Terrier X MN 8.3 L Blind; PACG (<1 week, scheduled for TSCP); IOP 42mmHg, high 69mmHg 34 Predisposed 9 Cocker Spaniel F 14.2 R Blind; PACG (14 months, buphthalmos); 5 months post TSCP. IOP: 12mmHg 35 Predisposed 10 Maltese MN 6.1 R Blind; PACG (7 weeks, enucleation, histopathology); IOP high 60mmHg 36 Predisposed 5 Siberian Husky FS 31.2 L Blind; PACG (3 months, enucleation, histopathology); IOP: high 54mmHg 37 Predisposed 11 Maltese X FS 7.4 R Blind; PACG (18 weeks, buphthalmos); IOP 20mmHg, high 61mmHg 38 Predisposed 8 Labradoodle MN 9.8 L Blind; PACG (3 weeks); IOP 31mmHg, high 54mmHg 39 Predisposed 8 Cocker Spaniel MN 15 L Visual; PACG (6 weeks); 3 months post TSCP; IOP 9mmHg, high 77mmHg 40 Predisposed 6 Labradoodle FS 9.1 L Visual; PACG (<1 week); new diagnosis; IOP 64mmHg

Table 5. 3. Measurements of the ciliary cleft in eyes predisposed to glaucoma and healthy controls

Variable Normal Predisposed P-value* Reproducibilit Reliabilit y y Superior Temporal Superior Temporal Entry 1.37 1.08 0.66 0.74 S: 0.002 0.901 0.923 (0.87-2.32) (0.52-2.11) (0.19-1.20) (0.36-1.25) T: 0.034 Length 1.90 2.16 1.82 1.41 S: 0.251 0.814 0.899 (1.39-3.41) (1.34-3.15) (1.24-2.95) (0.89-2.66) T: 0.027 Mid- width 0.25 0.27 0.19 0.08 S: 0.007 0.847 0.873 (0.16-0.42) (0.15-0.59) (0.01-0.36) (0.01-0.21) T: 0.433 Width 0.48 0.46 0.47 0.32 S: 0.403 0.804 0.889 (0.32-0.83) (0.30-0.69) (0.12-0.72) (0.1-0.84) T: 0.092 Area 0.38 0.41 0.36 0.10 S: 0.061 0.731 0.754 (0.18-1.03) (0.20-1.17) (0.02-0.55) (0.01-0.48) T: 0.001 Calculated 0.65 0.84 0.98 0.26 S: 0.520 0.572 0.813 area (0.30-1.76) (0.39-1.44) (0.16-1.75) (0.16-1.32) T: 0.100 Results listed as median (range); all measurements in mm (area calculations in mm2); * p value representing statistical comparison of superior (S) and temporal (T) quadrants between normal and predisposed eyes where statistically significant results are indicated in bold P a g e | 127

Group effect There was a statistically significant effect of group on CC entry (p=0.002). The CC entry in predisposed eyes was smaller in both the superior (0.83mm; 95% CI [0.58, 1.18]) and temporal quadrants (0.68mm; 95% CI [0.52, 0.99]) compared to healthy controls (superior: 1.37mm; 95% CI [1.23mm, 1.64], temporal: 1.08; 95% CI [0.90mm, 1.38]) (p=0.002 and 0.034, respectively) (Fig.3).

Figure 5. 3. Measurements of the ciliary cleft. The p values are obtained from two- way ANOVA investigating the effect of group (normal/control versus predisposed) on each variable. All variables measured in millimetres (mm) on the left y-axis, except for ciliary cleft area, which is measured on the right y-axis.

When considering angular measurements, there was a statistically significant effect of group on ARA and AOD2 (p=0.001) (Fig.5.4). The ARA in predisposed eyes was smaller in both the superior (0.56mm2; 95% CI [0.24, 1.09]) and temporal quadrants (0.39mm2; 95% CI [0.14, 0.74]) compared to healthy controls (superior: 1.26mm2; 95% CI [0.96, 1.62], temporal: 0.78mm2; 95% CI [0.66, 1.23]) (p=0.018 and 0.020, respectively). The AOD2 in predisposed eyes was smaller in both the superior (0.40mm; 95% CI [0.22, 0.57]) and temporal quadrants (0.13mm; 95% CI [0.08, 0.39]) P a g e | 128 compared to healthy controls (superior: 0.59mm; 95% CI [0.51, 0.74], temporal: 0.40mm; 95% CI [0.37, 0.63]) (p=0.022 and 0.012, respectively). Quadrant effect There was no statistically significant effect of quadrant on any measurements obtained in this study. Interaction of group and quadrant There was no statistically significant effect of the interaction of group and quadrant on any anterior segment measurements after correction for Type I Error.

Figure 5. 4. Measurement of variables representing the drainage angle in normal (control) and predisposed eyes. The p values are obtained from two-way ANOVA investigating the effect of group (normal/control versus predisposed) on each variable. ICA = geometric iridocorneal angle; ARA = angle recess area; AOD = angle opening distance.

P a g e | 129

Measurements corrected for body weight When corrected for body weight (using SLD), variables on which there was a statistically significant effect of group included ARA (p=0.003) and AOD2 (p=0.001) which were statistically significant without correction for body weight. There was no statistically significant effect of group on mid-iris thickness using raw data, however when corrected for SLD, this effect did reach statistical significance with a greater mid- iris thickness in predisposed (0.27mm, 95% CI [0.24, 0.31]) compared to normal eyes (0.21, 95% CI [0.19, 0.22]) (p<0.001) (Fig.5.5).

Figure 5. 5. Box plot demonstrating median (line within box) and range (whiskers) measurements of angle recess area (ARA), angle opening distance 2 (AOD2), and mid-iris thickness (MID-IRIS) measurements in normal (red boxes with crosshatches) and predisposed (blue box with diagonal lines) eyes using raw and corrected data. P-values showing statistically significant findings (p<0.004) are indicated in red. P a g e | 130

Table 5. 4. Angular measurements associated with the iridocorneal angle in eyes predisposed to glaucoma and healthy controls Variable Normal Predisposed P-value* Reproducibility Reliability Superior Temporal Superior Temporal Geometric 34.3 30.8 34.1 29.3 S: 0.474 0.539 0.712 iridocorneal angle (25.2-47.2) (22.2-59.1) (20.9-49.7) (21.2-53.2) T: 0.728 Angle recess area 1.40 0.78 0.33 0.41 S: 0.018 0.784 0.832 (0.37-2.63) (0.33-2.10) (0.12-0.98) (0.09-1.41) T: 0.020 Angle opening 1.57 1.26 1.17 1.39 S: 0.161 0.832 0.896 distance1 (0.69-2.56) (0.55-2.61) (0.20-1.79) (0.51-1.91) T: 0.549 Angle opening 0.59 0.40 0.38 0.21 S: 0.022 0.861 0.902 distance2 (0.30-0.95) (0.16-1.03) (0.05-0.51) (0.05-0.56) T: 0.012 Results listed as median (range); all measurements in mm; * p value representing statistical comparison of superior (S) and temporal (T) quadrants between normal and predisposed eyes where statistically significant results are indicated in bold P a g e | 131

Table 5. 5. General anterior segment measures in eyes predisposed to glaucoma and healthy controls Variable Normal Predisposed P-value* Reproducibility Reliability Superior Temporal Superior Temporal ICC Descemet’s to anterior lens 2.33 2.29 2.24 2.32 S: 0.348 0.913 0.952 capsule (SLD) (1.98-2.67) (1.96-2.73) (1.56-2.48) (2.16-2.66) T: 0.752 Scleral thickness1 0.75 0.73 0.82 0.78 S: 0.478 0.821 0.899 (0.64-0.96) (0.59-0.93) (0.55-0.97) (0.62-0.93) T: 0.408 Scleral thickness2 0.62 0.66 0.80 0.66 S: 0.041 0.847 0.902 (0.55-0.78) (0.58-0.77) (0.54-0.95) (0.65-0.90) T: 0.025 Corneal thickness1 0.64 0.64 0.71 0.69 S: 0.016 0.885 0.945 (0.53-0.72) (0.49-0.70) (0.58-0.82) (0.63-0.81) T: 0.004 Base iris thickness 0.40 0.45 0.48 0.59 S: 0.326 0.772 0.875 (0.26-0.61) (0.24-0.66) (0.24-0.61) (0.39-0.93) T: 0.073 Mid iris thickness 0.48 0.49 0.49 0.68 S: 0.053 0.739 0.932 (0.26-0.59) (0.33-0.78) (0.39-0.75) (0.47-1.22) T: 0.009 Tip iris thickness 0.28 0.29 0.34 0.47 S: 0.280 0.412 0.831 (0.11-0.70) (0.12-0.92) (0.21-0.78) (0.31-0.72) T: 0.232 Iris-lens contact 1.46 1.85 2.81 4.12 S: 0.341 0.531 0.801 (0.46-5.43) (0.68-5.54) (2.17-3.35) (0.52-4.80) T: 0.108 TID-TM 1.18 0.86 0.66 0.66 S: 0.020 0.894 0.886 (0.63-1.70) (0.51-1.47) (0.17-1.03) (0.25-1.11) T: 0.044 Iridociliary process distance 2.07 1.74 1.53 1.69 S: 0.003 0.713 0.857 (1.48-2.62) (1.08-2.71) (0.83-2.11) (1.37-1.87) T: 0.138 DLCP 1.75 1.46 1.53 1.60 S: 0.026 0.732 0.868 (1.28-2.52) (0.92-2.58) (0.87-1.70) (1.01-1.87) T: 0.418 Anterior chamber depth 2.89 (2.10-3.68) 3.05 (2.04-4.06) 0.189 0.941 0.932 Results listed as median (range); all measurements in mm; * p value representing statistical comparison of superior (S) and temporal (T) quadrants between normal and predisposed eyes where statistically significant results are indicated in bold. P a g e | 132

Reproducibility of images Measurements of all variables were obtained on three separate images for each quadrant in each dog except in four cases (control eyes) where only two images were suitable for inclusion. The reproducibility of CC measures was considered good or excellent for CC length, CC width, CC-mid width and CC entry (ICC >0.804), adequate for CC area (ICC 0.731), and inadequate for calculated area (ICC 0.572) (Table 5.3). Reproducibility of AOD was considered good (ICC 0.832-0.861), was adequate in measures of ARA (ICC 0.784), and was inadequate for geometric measures of the ICA (ICC 0.539) (Table 5.4). Reproducibility was excellent for measures of SLD (ICC 0.913) and ACD (ICC 0.941), good for measures of ST (ICC 0.821-0.847) and SID (ICC 0.894), and adequate for measures of base- and mid-iris thickness (ICC 0.772 and 0.739, respectively), ICP (ICC 0.713), and DLCP (ICC 0.732). For measures of iris tip thickness and ICP, reproducibility was considered of limited applicability (Table 5.5). Intra-rater reliability of measurements Duplicate measurements were obtained of all variables for the scans of 16 dogs. Intra- rater reliability was good or excellent (ICC >0.813) for all measures of the CC except for CC area for which it was considered adequate (ICC 0.754) (Table 5.3). When considering angular measurements, intra-rater reliability was good or excellent (ARA, AOD1 and AOD2) except for measurements of the geometric ICA (ICC 0.712) (Table 5.4). Intra-rater reliability was considered good or excellent for all other variables investigated in this study (ICC 0.801-0.952) (Table 5.5).

DISCUSSION In this study, quantifiable measures of the anterior segment that differed between eyes predisposed to PACG and healthy controls were obtained without the need for sedation or anaesthesia. Our findings support those previously described in a heterogenous population of dogs with PACG,495 suggesting UBM has potential use in assessing the predisposed, but as yet unaffected eye of dogs with PACG. Identification of structural measurements that were significantly affected by group presents a potential means of identifying early disease in dogs with PACG before the onset of clinical signs that result in irreversible damage to the retina and ONH. Both qualitative and quantitative assessments of the CC in dogs have been described58,250,255,486,496 and measures of the CC in this study are comparable to those previously reported (length 2.21-2.6mm; width 0.23-0.41mm).58,486

We identified a smaller angle opening distance (AOD2) in predisposed compared to healthy eyes, that was comparable to measurements previously described.255,258,486 Although our study does not allow any conclusions to be drawn about how the AOD corresponds to gonioscopic findings, or the nature of any association between structural measurements and disease, UBM allows evaluation of the traditional outflow pathways beyond the anterior face of the iridocorneal angle.255 It has been suggested that when combined with gonioscopy, UBM may have predictive value in identifying dogs at risk of developing glaucoma.497 Our study adds to existing literature P a g e | 133 by describing findings in a heterogenous population of dogs with naturally occurring disease. Correcting raw data for body size as previously reported257,498 identified a statistically significant effect of group on some variables for which there was no discernible effect without correction. In people, a shallow AC is considered a cardinal clinical and pathogenic feature predisposing eyes to PACG499-501 and AC depth has been presented as a potential screening test in people.483 Although a shallow AC has not previously been described in a heterogenous population of dogs, it has been described in Samoyeds with PACG.502 It is possible that correction for factors such as body size and skull shape are essential in assessing structural change associated with glaucoma in the dog, and is not surprising given the degree of variation in this species. To the authors’ knowledge, quantitative assessment of the drainage angle in multiple quadrants using UBM has not been investigated in canine glaucoma. We assessed the dorsal and temporal quadrants as ready access is obtained in conscious dogs in a clinical setting. It is possible that more significant changes or variations may be evident with analysis of all quadrants, although this possibility was not explored in this series. These findings support current knowledge that circumferential examination of the drainage angle is required when performing gonioscopy due to variations in pectinate ligament dysplasia within an eye.133,134 On this basis, further investigations and validation of UBM in the management of canine glaucoma should include assessment of multiple quadrants. This may restrict the numbers of dogs in study populations as sedation or general anaesthesia may be required to obtain suitable images of the inferior and nasal quadrants, and consideration should be given to the trade-off between the desire for circumferential assessment and the number of dogs included. As was shown with HRUS,492 we identified variability among structural measures with respect to intra-rater reliability. In this study, a single observer performed the ultrasounds and obtained all measurements for analyses. Intra-observer reliability was promising with most variables considered good or excellent, and all variables had at least ‘adequate’ reliability. Further studies are needed to determine whether more reliable images are obtained using UBM compared to HRUS due to the improved detail obtained with higher frequency transducers.492 Several variables in our study showed reproducibility that was considered inadequate. Factors such as probe position and angle, technique variations between operators, and tolerance levels of individual dogs are expected to produce some degree of variation in the ability to reproduce images of a moving organ in a live animal. Whether use of chemical restraint would improve reproducibility remains undetermined, but consideration for a potential improvement in image reproducibility (and thus a more valid technique) should be weighed against the need for larger sample populations (that might be more readily achieved without the added requirement for sedation or anaesthesia) to account for the degree of variation associated with naturally occurring PACG. The measurements obtained in this study overlapped between groups and such an overlap was also described in measurements of the canine retina derived from OCT imaging.484 We do not suggest technologies such as UBM are necessary nor indicated as a method to distinguish normal from predisposed eyes. We do propose that UBM, P a g e | 134 combined with other information obtained from clinical examination and diagnostic techniques, may allow identification of changes in eyes predisposed to PACG before the onset of overt glaucoma. Our results provide some of the preliminary steps needed to validate UBM as a clinical tool for this purpose. The major limitations in this study are related to the small sample population, and the cross-sectional study design. The use of a single observer to scan eyes and make all measurements limits the ability to draw conclusions about validity and the influence of operator experience.492 Furthermore, with no grading of gonioscopy findings, relationships between gonioscopy and UBM measurements cannot be investigated here, nor compared to existing reports in the medical and veterinary literature. Although conclusions about the clinical significance of UBM findings cannot be made from these data, our findings provide a form of external validation for similar findings reported by Hasegawa et al.495 We propose these findings are an indication for multi- institutional longitudinal studies that investigate whether any of the statistically significant findings reported to date bear clinical significance with respect to canine glaucoma. In this study structural measurements within the anterior segment showed differences between eyes predisposed to glaucoma and those in healthy controls which may represent quantifiable indicators of early disease. UBM as a diagnostic technique is non-invasive and is increasingly available in veterinary ophthalmology. The clinical significance of our findings and any association with aqueous humour outflow in the canine eye remains undetermined, and validation of the technique is warranted. An appropriately powered longitudinal clinical study to determine potential predictive value of biomicroscopic imaging in canine glaucoma is proposed by these authors as the next step.

ACKNOWLEDGEMENTS The authors are grateful to the veterinarians and clients whose patients and dogs were included in this study, and to Dr Mark Billson from the Small Animal Specialist Hospital, for the introduction to veterinary ophthalmology which lay the groundwork for this study. We also wish to acknowledge the financial support of the Canine Research Foundation in the conduct of this study. P a g e | 135

CHAPTER SIX OPTICAL COHERENCE TOMOGRAPHY OF THE RETINA, NERVE FIBRE LAYER AND OPTIC NERVE HEAD IN DOGS WITH GLAUCOMA

The following is the re-formatted manuscript published by John Wiley & Sons, Pty Ltd: Graham KL, McCowan CI, Caruso K, Billson FM, Whittaker CJG, White A. Optical coherence tomography of the retina, nerve fiber layer and optic nerve head in dogs with glaucoma. Veterinary Ophthalmology. 16th June 2019 DOI: 10.1111/vop.12694

ABSTRACT Objective: To evaluate the retina and optic nerve head (ONH) in canine eyes predisposed to glaucoma using optical coherence tomography (OCT). Animals: Twenty-five eyes (24 dogs). Methods: Measures of peripapillary retinal, retinal nerve fibre layer (RNFL), and ganglion cell complex (GCC) thickness and ONH parameters were obtained in vivo by OCT of the unaffected eye in dogs diagnosed with unilateral primary glaucoma (predisposed; n=12) and compared with measures of healthy control eyes (normal; n=13). Repeatability and intra-rater reliability were explored using intra-class correlation coefficients (ICC). Results: Compared to normal eyes, predisposed eyes had a thinner retina in the temporal (p=0.005) and inferior quadrants (p=0.003) and decreased inner retinal thickness (superior: p=0.003, temporal: p=0.001, inferior: p<0.001, nasal: p=0.001). Predisposed eyes had a thinner RNFL compared to normal eyes (p=0.005), and when analysed in quadrants, was thinner in the superior (p<0.001), temporal (p=0.034), and nasal quadrants (p=0.001). Repeatability (ICC 0.763-0.835) and intra-rater reliability (ICC 0.824-0.942) were good to excellent for measures of retinal thickness, and adequate for RNFL measurements (ICC 0.701-0.798). Reliable measurements of optic disc area were obtained and were similar between groups (p=0.597). Measurements of parameters relying on automated software detection (GCC, optic cup, optic rim) had inadequate repeatability and reliability. Conclusion: Statistically significant differences in retinal and RNFL thicknesses were identified in normal and predisposed eyes. Reliable and consistent measurements of variables with manual adjustment of software detected parameters were obtained. Validation of OCT as a diagnostic tool for clinical assessment in canine glaucoma is warranted P a g e | 136

INTRODUCTION Glaucoma is a neurodegenerative disease that results in structural changes to the retina and optic nerve head (ONH) and associated progressive vision loss. These structural and functional changes occur due to retinal ganglion cell (RGC) and axonal loss, although the exact nature of the structure-function relationship in glaucoma remains unknown.1 Experimental and clinical investigations in people with primary open angle glaucoma (POAG) vary with respect to which changes occur first, whether structure and function are indeed correlated, and how these correlations change with disease progression.2-8 Possible reasons for inconsistencies between studies in people include differences in study populations, the presence and type of glaucoma, stage of disease, and the sensitivity and specificity of tests and parameters used to detect change.1,8 Pattern-evoked ERG and automated perimetry provide objective measures of inner retinal and visual function that are not widely available or not practical in veterinary medicine, and despite reports describing tools to measure canine visual function,9-16 vision is difficult to quantify in dogs and visual field deficits are rarely detected in this species. The identification of ONH damage in dogs is complicated by the degree of variation in ONH appearance with myelin extending anterior to the lamina cribrosa in this species.17 Despite the inadequacy of intraocular pressure (IOP) alone for the identification of glaucoma and its progression,18,19 the diagnosis and monitoring of canine glaucoma is primarily based on an elevated IOP and/or clinical features indicative of chronic IOP elevation, and availability of technologies that can identify early structural changes in the posterior segment is limited in veterinary practice. Identification of structural changes before the onset of clinically recognizable disease could facilitate diagnosis of glaucoma at earlier stages than is currently possible. Optical coherence tomography (OCT) is a high-resolution optically based imaging system that uses low-coherence interferometry to provide cross-sectional images of ocular tissues.20 The use of OCT to evaluate the ONH, retinal nerve fibre layer (RNFL), and the macula is common in people with glaucoma,2-4,21,22 and new approaches in the use of OCT in glaucoma management continue to emerge.23-26 Despite widespread use of OCT in clinical ophthalmic and optometry settings for people, the use of OCT for assessment of the canine posterior segment remains limited, although its use has been increasingly reported over the last decade. Veterinary use of OCT for assessment of the posterior segment is reported in normal Beagles,27,28 dogs with Sudden Acquired Retinal Degeneration Syndrome (SARDS),29-31 in research settings investigating diseased retinas,32-36 in cats,37,38 and in case reports describing a variety of retinal conditions.39-42 Grozdanic et al43 reported thinning of the inferior retina as measured by OCT following induction of an acute elevation in IOP in normal Beagles. However, to the authors’ knowledge, there are no published reports of systematic studies of these OCT parameters in eyes with glaucoma or eyes at risk of glaucoma in dogs. In this cross-sectional study, we evaluated the ability of spectral-domain OCT (SD-OCT) to detect structural changes consistent with disease in the predisposed fellow eye of dogs with naturally occurring glaucoma. P a g e | 137

Table 6. 1. Individual subject characteristics

I.D Age Breed Sex Ey Classification Dxa IOPb Dilated Contralateral eye Glaucoma IOP (IOP reported in mmHg) duration yrs e in study mmH mmHg (days) g 1 6 Labrador X MN R Normal C 13 12 Normal*; IOP 12 (15) N/A Poodle 2 9 Poodle MN L Normal C 12 10 Normal*; IOP 10 (14) 3 6 Siberian Husky FS L Normal C 14 14 Entropion (excluded); IOP not measured N/A 4 5 Pomeranian X MN R Normal C 13 13 Normal*; IOP 11 N/A 5 10 Jack Russell FS L Normal C 16 21 Previous trauma (cat scratch) (excluded); IOP 3 N/A Terrier 6 4 Cocker Spaniel FS R Normal C 10 14 Normal*; IOP 13 (16) N/A 7 8 Shih Tzu FS L Normal C 8 10 Normal*; IOP 12 (9) N/A 8 6 Labrador X MN R Normal C 12 12 Normal*; IOP 13 (14) N/A Poodle 9 10 Australian MN L Normal C 11 12 Traumatic cataract (excluded); IOP 3 N/A Cattle Dog 10 9 Labrador X FS R Normal C 16 12 Normal*; IOP 15 (12) N/A Poodle 11 10 Shih Tzu X FS L Normal C 13 19 Surgical pseudophakia (excluded); IOP 9 N/A 12 3 Terrier X MN L Normal C 12 16 N/A. Mean measurement of both eyes used for analyses N/A R 11 18 13 9 Mansfield FS R Predisposedc C, U 15 17 Newly diagnosed primary glaucoma (latanoprost 36 Terrier before/during imaging); IOP 40 14 12 Maltese X M R Predisposed C,U 8 14 Primary glaucoma (on topical latanoprost for months). 90 IOP 64 15 10 Terrier X FS L Predisposed C,U 13 10 Primary glaucoma (on topical latanoprost for 2 weeks). 14 IOP 48 P a g e | 138

16 9 Labrador X FS L Predisposed C,U 12 15 Primary glaucoma (on latanoprost and timolol 65 Poodle /dorzolamide); IOP 62 17 12 Maltese FS L Predisposed C, 13 22 Primary glaucoma (end stage at diagnosis; enucleated 4 145 U, months prior). Histology consistent with primary. Max H IOP recorded 60 18 6 German FS R Predisposed C, 14 19 Primary glaucoma (end stage at diagnosis; enucleated). 350 Shepherd U, Histology consistent with primary. Max IOP recorded 56 H 19 6 Terrier X MN L Predisposed C,U 12 17 Primary glaucoma (end stage at diagnosis, IOP 67; 390 evisceration and placement of ISP) 20 9 Labrador X FS L Predisposedc C,U 16 15 Newly diagnosed primary glaucoma (latanoprost 32 Poodle before/during imaging); IOP 42 (maximum 72) 21 5 Terrier X MN R Predisposed C,U 15 20 Primary glaucoma (on latanoprost and timolol 270 /dorzolamide); IOP 18 (82) 22 9 Maltese FS R Predisposed C,U 16 19 Primary glaucoma (on latanoprost and timolol 420 /dorzolamide >12mths); IOP 6 (85) 23 10 Jack Russell FS R Predisposed C,U 14 17 Primary glaucoma (end stage at diagnosis, IOP 73; 75 Terrier evisceration and placement of ISP) 24 6 Flat-Coated FS R Predisposed C,U, 12 9 Primary glaucoma (end stage at diagnosis; enucleated). 450 Retriever H Histology consistent with primary. Max IOP recorded 56 aDx = Method of diagnosis where C = clinical ophthalmic examination, U = ocular ultrasound, H = ocular histopathology following enucleation of globe; bas measured at time of imaging; ceye naïve to therapy at the time of imaging in a dog with glaucoma. *The mean measurements of this and the contralateral eye were used for analyses. Where a higher IOP was documented in that eye on a separate occasion, the highest documented measurement is recorded in parentheses. Post (tropicamide) dilation IOPs (when obtained) are underlined and italicised; MN = male neutered; FS = female spayed; L = left; R = right; IOP = intraocular pressure; N.R = no record; TSCP = transscleral cyclophotocoagulation. P a g e | 139

MATERIALS AND METHODS Animals Privately owned pet dogs with unilateral primary angle closure glaucoma, as diagnosed by a veterinary ophthalmologist, and healthy control dogs without ophthalmic disease, were included. All dogs had a complete physical and ophthalmic examination, including slit lamp biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, UK), indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, NSW Australia), quantitative tear testing (Schirmer tear test, Merck Animal Health, NJ, USA), rebound tonometry (Icare Tonovet, Icare, Finland) and fluorescein staining. Gonioscopy was performed using a direct (Koeppe, Ocular Instruments, Bellevue, WA, USA) or indirect (G-4 Four Mirror, Volk Optical Inc, Mentor, OH, USA) goniolens according to clinician preference, and findings recorded as either normal or abnormal. Ultrasound biomicroscopy (UBM; 48MHz probe, Accutome, UBM Plus 24-6300-G-V, Malvern PA) of the anterior segment was performed on every eye that was included in analyses, and B-mode ultrasonography (10MHz probe, Philips HD-11, Philips Australia) performed in all dogs with glaucoma to support the absence of any intraocular structural change that may result in secondary glaucoma. All procedures were conducted with approval from the University of Sydney Animal Ethics Committee (2016/1004), and with the informed consent from the dog’s owner.

Eyes were allocated into groups based on the following criteria: (i) normal: eyes with an unremarkable ophthalmic examination, no abnormalities on gonioscopy, and no clinical or historical suggestion of vision impairment; (ii) predisposed: eyes with an abnormal drainage angle on gonioscopy (narrowed iridocorneal angle and/or pectinate ligament dysplasia), a narrowed or collapsed ciliary cleft on UBM examination, and a diagnosis of primary angle closure glaucoma (PACG, clinical diagnosis with or without supportive histopathologic changes) in the contralateral eye made by a veterinary ophthalmologist. Measures of retinal thickness and the ONH were compared between normal eyes and those predisposed to glaucoma. Medical records were reviewed, and clinical data including signalment, histopathology of previously enucleated globes (where available), IOP and duration of clinical signs recorded.

Procedures Image acquisition was performed after physical and ophthalmic examinations. Rebound tonometry was repeated after application of one drop of tropicamide 1.0% (Minims, iNova Pharmaceuticals Pty Ltd, Chatswood, NSW Australia) had resulted in pupil dilation, and before administration of any sedative. The cornea was lubricated throughout imaging procedures with 0.3% hypromellose (GenTeal Eye Gel, Alcon Laboratories Australia Pty Ltd, Frenchs Forest NSW). Imaging was performed under sedation except if the dog was anaesthetized for reasons separate to this study in which case images were obtained at the beginning of the anaesthetic. Sedation typically included medetomidine (2-5µg/kg IV, Domitor, Pfizer Ltd. West Ryde, NSW) and butorphanol (0.1-0.2mg/kg IV, Torbugesic, Zoetis Australia Pty Ltd. Rhodes NSW Australia) administered to effect. After all scans were obtained, sedation was reversed using atipamezole (Antisedan, Pfizer Ltd. West Ryde, NSW). P a g e | 140

Image acquisition All images were obtained using the Optovue iVue SD-OCT unit (Optovue, Inc., Freemont, CA USA) with an image acquisition rate of 25,000 axial scans (A-scans) per second with 5µm depth resolution. Imaging of all eyes was performed by the same investigator and this investigator stabilized the globe and operated the OCT unit. Following pupil dilation, oxybuprocaine hydrochloride 0.4% (Minims, iNova Pharmaceuticals Pty Ltd, Chatswood NSW Australia) was administered topically and Colibri forceps used to hold the conjunctiva near the limbus at 12 o’clock for manipulation and stabilization of the globe. The OCT device was mounted on a tripod with a ball head mount and the height and angle adjusted for imaging of each eye. Scanning was centred over the ONH to capture all sectors of the peripapillary retina (Fig 6.1).

Figure 6. 1. (A) Optic nerve head located at the centre of the scan (left eye) to allow comparison of measurements between subjects; B) schematic overlay demonstrating the sectors of the peripapillary retina that were analysed for assessment of retinal thickness. Retinal thickness was assessed in quadrants (superior, nasal, inferior, temporal); eight sectors of the RNFL were assessed (SN superior, nasal; ST superior, temporal; NU nasal, upper; NL nasal, lower; IN inferior, nasal; IT inferior, temporal; TU temporal, upper; TL temporal, lower; Sup superior; Nas nasal; Inf inferior; Temp temporal).

P a g e | 141

Quality of the scanned image, calculated by the manufacturer’s software as the scan quality index (SQI), was affected by the intensity of reflected light. Data for analyses were obtained from a scan with ‘good’ SQI for each structure investigated (retinal thickness, RNFL, GCC, ONH). Greater light intensities resulted in a higher SQI. These SQI data were classified as ‘good’ or ‘poor’ by the software. Scans classified as ‘poor’ did not allow ocular structures/layers to be visible and easily segmented and were repeated until a scan with a ‘good’ SQI was obtained for analysis. When the signal strength was ‘poor’ due to pathology or miosis affecting light absorption, that eye was excluded from the study. Other factors that resulted in poor images that could not be used included a local weak signal, data that were out of the OCT window boundary, and poor positioning. Weak signals (caused by blinking, poor alignment etc.) resulted in an inability to visualize retinal layers on the image (B- scan). When data were either too high or too low in the OCT window boundary, accurate data acquisition was compromised and affected scans were discarded (Fig 6.2a). Globe position was assessed by evaluating the position of the dorsal retinal vein on en face images, and by assessing horizontal and vertical retinal scans (retina cross line scan). If the dorsal retinal vein, including the detectable thickness associated with that vessel, remained in the superior quadrant, the image was considered satisfactory (Fig 6.2b). If the globe was rotated such that the dorsal retinal vein and its associated retinal thickness was outside the superior quadrant, the scan was discarded (Fig 6.2c). Scans where the retina was tilted in one or both of the horizontal and/or vertical plane (Fig.2d), were either repeated so that the retina was flat (Fig 6.2e), or the eye was excluded from the study. In scans of the ONH (and, consequently, corresponding RNFL measurements), scan quality was affected if not centred on the optic disc, and where the scan was not centred on the ONH, the case was excluded from ONH analyses. Scan adjustments and segmentation For each eye, the series of scans obtained included: retinal scans (retinal map and cross-line), ONH and RNFL assessment (one 3D glaucoma scan to obtain the baseline optic disc for manual adjustment of the optic disc outline, and a subsequent glaucoma scan for measurement of the optic disc and nerve fibre layer parameters), and one scan for ganglion cell complex analysis. To determine retinal thickness, retinal scans centred on the ONH, were obtained for each eye, and thickness of the peripapillary retina recorded for analyses based on a single scan for each subject (Fig 6.3). Total retinal thickness was measured from the inner limiting membrane (ILM) to the retinal pigment epithelium (RPE), inner retinal thickness from ILM to the outer limit of the inner plexiform layer (IPL), and the outer retina from the IPL to the RPE (Fig 6.4). Manual adjustments of software-derived measurements were made to each of the 20 composite B-scan images obtained in each retinal scan. To eliminate the influence of ONH myelination, lines demarcating the boundaries measured (ILM, IPL, and RPE), were manually adjusted to overlay each other over the region of the ONH and myelin (between the ends of the RPE) (Fig 6.5).

P a g e | 142

Figure 6. 2. Selection for appropriate scan quality and position. A) Retinal scan with a poor scan quality index showing distortion of the en face image to the left (dark regions at the top and bottom edges of the image), inadequate definition of retinal layers for manual segmentation, and the retina is not confined to the OCT window. Thickness map of case 1 showing appropriate positioning of the optic nerve head (B) taken after discarding the initial scan (C) due to excessive rotation of the globe shown with the dorsal retinal vein oriented to the temporal quadrant. Retinal crossline scan (scan orientation horizontal or vertical as indicated by the arrow in the top right corner) showing (D) inappropriate scan angle, and (E) repeat scan with appropriate positioning.

Figure 6. 3. Schematic representation of where measurements for determining retinal thickness were obtained (highlighted yellow regions) when the circles are centred over the optic nerve head. The region of each quadrant that was within the inner circle was excluded to minimise the impact of the optic nerve and intraocular myelin on measures of retinal thickness. P a g e | 143

Figure 6. 4. Retinal layer segmentation on B-scan image of the retina from case 1 (normal eye); A) retinal layers identified for manual segmentation of scans to measure retinal thickness including inner, outer and total retina, nerve fibre layer, ganglion cell complex. NFL = nerve fibre layer; GCL = ganglion cell complex layer; IPL = inner plexiform layer; INL = inner nuclear layer; OPL = outer plexiform layer; ONL = outer nuclear layer; ELM = external limiting membrane; PR = photoreceptors; RPE = retinal pigmented epithelium; B) definition of layers for measurement of retinal thickness: inner retina (from inner limiting membrane to inner plexiform layer), outer retina (from inner plexiform layer to retinal pigmented epithelium). P a g e | 144

Figure 6. 5. Three of the 20 B-scan images from case 2 showing manual adjustment of the lines demarcating the boundaries of the inner and outer retina (red arrows in [A]; the level at which retinal thickness is measured in each image is depicted in the en face image to the right). Where the optic nerve head and myelin were present, the lines were aligned to negate any measure of thickness in these regions (yellow arrows in [B-C]). D) Resulting thickness map showing the region of the ONH and intraocular myelin in black, reflecting the manual adjustments. S= superior; T= temporal; I= inferior; N= nasal. The poorer image quality evident in Fig 5a-c is presented here as the only manner to show images of the modified individual B-scans was to take photographs of the computer screen as the analyses were being conducted. Position of temporal and nasal quadrants varies depending on whether the left or right eye is imaged.

P a g e | 145

For calculation of RNFL thickness and ONH parameters, glaucoma scans (3D and standard scans) were centred on the ONH and included a circle with diameter 3.45mm. Each scan pattern consisting of 13 concentric circular scans ranging from 1.3 to 4.9mm diameter with 0.3mm interval, and 12 radial scans 3.4mm in length was used in analysis of the ONH. Automated detection of the optic disc margin, as well as the ends of the RPE/choroid complex, which were adjusted on the 3D ONH images, were manually corrected prior to analysis (Fig 6.6). No manual corrections were made (not possible with the instrument’s proprietary software) to detection or measurement of the optic cup or rim in any study. The GCC consisted of those retinal layers in which retinal ganglion cells were located. This included the RNFL containing ganglion cell axons, the ganglion cell layer containing ganglion cell bodies, and the IPL containing the ganglion cell dendrites. The scan protocol used to obtain measures of GCC consisted of 15,000 points in a 7mm square area within 0.6sec using one horizontal line and 15 vertical lines at 0.5mm intervals. GCC scans were centred 0.75mm temporal to the ONH. Detection and measurement of the GCC was automated for all studies with no manual adjustments possible using the instrument’s proprietary algorithm for this setting. GCC scans were excluded only if not appropriately centred temporal to the ONH.

Figure 6. 6. Optic nerve head analysis. For manual identification of the outline of the ONH, 3D scans were assessed. Markers (on the right and below en face image) were manoeuvred along the 6mm margins to show cross-sectional images of the retina in the horizontal and vertical planes at that level. Identification of the termination of the retinal pigmented epithelium and increased thickness associated with myelination could then be made prior to marking the ONH boundaries on the en face image (case 22). P a g e | 146

Repeatability of scanning technique To evaluate the consistency of measures obtained between scans of the same dog under the same conditions, the entire series of scans were obtained twice (under the same sedation) in 11 dogs (11 eyes). Intra-class correlation coefficient (ICC) was calculated for each variable obtained in sequential scans of the same eye. Reliability of observer measurements To evaluate intra-rater reliability in making manual adjustments to the segmentation of retinal layers and ONH, the same observer performed manual adjustments and obtained a repeat series of measurements for each eye to obtain a new dataset for comparison to the initial measurements obtained, with a 6 month period between analyses to avoid any risk of recall bias. Reliability of the GCC analyses was not assessed as these variables were automated measures that could not be manually adjusted. Statistical Analyses Data were analysed using commercially available software (SPSS v22.0, IBM Corporation; GraphPad Prism 7.0a). Individual characteristics of dogs/eyes in each group were compared using a Mann-Whitney test for continuous variables (age, IOP, duration of disease), and a Fisher exact test was used to compare categorical data between study groups (pure- versus crossbred, sex, eye). A Mann-Whitney test was used to compare measurements of each variable between normal and predisposed eyes. Spearman’s rank-order correlation was used to determine the strength and direction of association of age, IOP, and the duration of glaucoma (when relevant) with measured variables. Regional differences (between peripapillary quadrants) in retinal thickness between eyes in the same group were compared using a Kruskal-Wallis non parametric ANOVA and Dunns multiple comparisons test. The effect of group (ophthalmic diagnosis) on dependent variables was assessed using a Mann-Whitney test as non-parametric data did not satisfy assumptions of a t- test. When two eyes from the same dog were included in the study, the mean of both eyes for each variable was used for analyses. For all analyses, a p value <0.05 was considered statistically significant. In assessing repeatability and intra-observer reliability with the ICC, a one-way random model was assumed where the subjects (dogs) are assumed random.44 Correlations >0.9 were considered excellent, when 0.80-0.89 they were considered good, when 0.70-0.79 correlations were considered adequate, and correlations <0.70 were considered of limited applicability.45

RESULTS Images suitable for analyses were obtained of 25 eyes from 24 dogs to assess for potential changes in the retina and ONH of eyes predisposed to glaucoma (Table 6.1). Twelve dogs (13 eyes) were normal, and 12 dogs (12 eyes) predisposed to primary glaucoma. Of the 12 P a g e | 147 predisposed dogs, the contralateral glaucomatous eye was still present in seven cases. Scanning of eyes with glaucoma was precluded by prostaglandin analogue-induced miosis in five cases. The remaining two glaucomatous eyes (cases 13 and 20) were diagnosed with PACG on the day of scanning. For ethical reasons, treatment with IOP lowering medication (topical latanoprost) was not delayed for imaging. The initial scans for both eyes were unsuitable for analysis, and the development of miosis with lowering of IOP precluded repeat scanning. Scans of six normal eyes were also excluded due to angulation of the scans (n=5) and globe rotation (n=1). There was no statistically significant difference in age (p=0.231), sex (p=0.400), IOP (p=0.366) and the number of purebred dogs (p=1.000) between groups (Table 6.2). Following pharmacologic pupil dilation, the median IOP in eyes predisposed to glaucoma (17mmHg [interquartile range [IQR = 15-19mmHg]) was significantly higher than that in normal eyes (12.5mmHg [IQR 12- 15.5mmHg]) (p=0.023). There was no statistically significant correlation between age and any measure of retinal thickness or ONH parameter in either normal or predisposed eyes. No correlation was identified between any variable measured in predisposed eyes and the duration of glaucoma in that dog (time since diagnosis of glaucoma in the contralateral eye). In eyes predisposed to glaucoma, a moderate positive correlation was identified between the IOP following pharmacologic pupil dilation and RNFL thickness in the superior (r=0.782, p=0.010), and in the inferior retina (r=0.702, p=0.028).

Table 6. 2. Characteristics of dog population studied

Normal Predisposed p value* Dogs 12 12 Age (years) 7.2 (3-10) 8.6 (5-12) 0.231 Purebred dog 5/12 (41.7%) 6/12 (50%) 1.000 Sex (female) 6/12 (50%) 9/12 (75%) 0.400 IOP (mmHg) 12.5 (8-16) 13.5 (8-16) 0.366 Dilated IOP (mmHg) 12.5 (10-21) 17 (10-22) 0.023 Disease durationa (days) N/A 194.8 (14-450) N/A Age, IOP and duration of glaucoma in subjects reported as median (range); IOP reported is that measured at the time of imaging; adisease duration is the time since diagnosis or reports of clinical signs consistent with glaucoma in the contralateral eye were recorded; *p value reported for appropriate statistical comparisons between normal and predisposed dogs/eyes.

Retinal thickness Table 6.3 presents median retinal thickness in each peripapillary region assessed. In normal eyes, total retinal thickness in the superior quadrant was significantly thicker than in both the inferior P a g e | 148

(p=0.010) and nasal quadrants (p=0.025). The inner retina was significantly thicker in the superior compared to the inferior quadrant (p=0.021), but no statistically significant difference was identified in outer retinal thickness between peripapillary quadrants (p=0.120) in normal eyes. When comparing regional differences in predisposed eyes, both total and outer retinal thickness values in the superior quadrant were significantly greater than in the temporal (p=0.016 and 0.012, respectively) and inferior quadrants (p=0.008 and 0.023, respectively), and inner retinal thickness significantly greater in the superior quadrant compared to the inferior quadrant (p=0.020).

Table 6. 3. Measures of retinal thickness in normal and predisposed eyes

Normal Predisposed p value Superior - Total 204 (172-236) 199 (155-221) 0.154 - Inner 107 (74-131) 85 (47-98) 0.003 - Outer 101 (81-157) 116 (86-128) 0.039 Temporal - Total 186 (152-223) 162 (105-188) 0.005 - Inner 92 (70-114) 68 (37-89) 0.001 - Outer 94 (79-131) 90 (57-116) 0.383 Inferior - Total 178 (159 – 198) 158 (115-193) 0.003 - Inner 86 (66-103) 65 (48-85) <0.001 - Outer 92 (80-119) 93 (67-118) 0.618 Nasal - Total 179 (146-231) 164 (129-209) 0.191 - Inner 91 (74-125) 78 (46-92) 0.001 - Outer 90 (64-141) 88 (78-117) 0.639 Units for all measurements reported in μm; results expressed as median (range); p values <0.05 in bold

In predisposed eyes, median total retinal thickness was significantly thinner in the temporal (162µm, 95% CI [134, 173]) and inferior quadrants (158µm, 95% CI [137, 169]) compared to normal eyes (temporal: 186µm 95% CI [175, 202]; inferior: 178µm, 95% CI [172, 187]) (p=0.005 and 0.003 respectively) (Fig 6.7). The inner retina was significantly thinner in predisposed eyes in all quadrants compared to inner retinal thickness in normal eyes (Fig 6.8). There was a statistically significant difference in outer retinal thickness between groups only in the superior quadrant, where the outer retina in predisposed eyes (116µm, 95% CI [102, 121]) was thicker compared to that in normal eyes (101µm, 95% CI [91.5, 113]) (p=0.039). P a g e | 149

Figure 6. 7. B-scan image (left) and corresponding thickness map (right) demonstrating a thinner total retinal thickness in a predisposed left eye (case 17, bottom images) compared to a normal right eye (case 10, top images).

Figure 6. 8. Scatterplot comparing measures of inner retinal thickness of the peripapillary retina in each quadrant showing a statistically significantly thinner inner retina in predisposed (diamonds) compared to normal eyes (circles). P a g e | 150

Nerve fibre layer thickness The median RNFL thickness for all regions assessed is presented in Table 6.4. There was no statistically significant difference in RNFL thickness between regions in either normal (p=0.054) or predisposed eyes (p=0.435). Overall, the median peripapillary RNFL was significantly thinner in predisposed (62.5µm, 95% CI [48.7, 80.7]) compared to normal eyes (92.5µm, 95% CI [80, 109]) (p=0.005) (Fig.9). Predisposed eyes had a thinner RNFL that reached statistical significance in the temporal (normal: 82µm, 95% CI [64.4, 120]; predisposed: 51.5 µm, 95% CI [40.1, 78.1]), superior (normal: 105µm, 95% CI [96.5, 125]; predisposed: 80µm, 95% CI [59.7, 90.3]) and nasal quadrants (normal: 89.5µm, 95% CI [78.1, 110]; predisposed: 60.5µm, 95% CI [44.5, 72.3]) (Fig 6.10).

Table 6. 4. Retinal nerve fibre layer measurements

Normal Predisposed p value TU 67.5 (48-144) 46 (30-91) 0.061 TL 83 (40-150) 51.5 (25-115) 0.022 Temporal 82 (48-206) 51.5 (28-102) 0.034 ST 96 (74-189) 78.5 (31-112) 0.044 SN 107 (86-150) 73.5 (21-123) 0.002 Superior 105 (82-148) 80 (26-104) <0.001 NU 72.5 (51-130) 47 (19-81) 0.007 NL 107 (75-168) 68.5 (26-98) 0.001 Nasal 89.5 (69-149) 60.5 (22-84) 0.001 IN 80.5 (53-238) 61.5 (27-175) 0.111 IT 74 (49-173) 59.5 (27-121) 0.261 Inferior 79.5 (51-144) 64.5 (27-147) 0.234 Av thickn 92.5 (69-138) 62.5 (27-100) 0.005 Av sup 89.5 (72-134) 65.5 (26-87) 0.003 Av inf 90.5 (58-174) 63.5 (29-114) 0.044 Sup-inf 6 (-72 – 31) -4.5 (-27 – 25) 0.734 Units for all measurements reported in μm; results expressed as median (range); ST superior- temporal, SN superior-nasal, NU nasal-upper, NL nasal-lower, IN inferior-nasal, IT inferior- temporal, TL temporal-lower, TU temporal-upper; p values <0.05 in bold

P a g e | 151

Figure 6. 9. Images of the optic nerve head and retinal nerve fibre layer obtained from normal (A. case 3, B. case 8) and predisposed eyes (C. case 15, D. case 19). The software which automatically detects the optic cup, failed to do so in some cases (e.g. A, C). The optic cup is represented in light grey in B and D.

Figure 6. 10. Scatterplot comparing measures of peripapillary retinal nerve fibre layer (RNFL) thickness between normal (circles) and predisposed (diamonds) eyes in each quadrant. P a g e | 152

Optic nerve head analyses The software failed to detect an optic cup in five eyes (normal n=2; predisposed n=3) (Fig.6.9a,c), meaning that only measures of disc area were obtained as part of the ONH analyses in these eyes. There was no statistically significant difference in any ONH parameter measured when comparing normal and predisposed eyes (Table 6.5).

Table 6. 5. Optic nerve head analyses

Normal Predisposed P value Cup: disc area 0.365 (0.05-0.54) 0.26 (0.08-0.84) 0.948 Cup: disc volume 0.66 (0.08-0.79) 0.5 (0.23-0.96) 0.983 Rim area 2.19 (1.08-5.37) 0.33 (4.5) 0.880 Disc area 3.59 (1.66-5.37) 3.01 (1.51-4.5) 0.597 Cup volume 0.11 (0.011-0.399) 0.026 (0.001-0.383) 0.174

Ganglion cell complex Positioning for assessment of the GCC (scan centred temporal to the ONH) was acceptable in 12/13 normal eyes, and in 11/12 predisposed eyes. There was a tendency for a thinner GCC in predisposed compared to normal eyes, although the difference between groups did not achieve statistical significance (Table 6.6).

Table 6. 6. Ganglion cell complex analyses

Normal Predisposed P value Total GCC 85 (57-112) 70 (33-97) 0.164 Superior 83 (63-107) 78 (34-100) 0.354 Inferior 82 (50-118) 65 (32-95) 0.155 Intra-eye difference 8 (-11 – 21) 6 (-5 – 31) 0.705

Repeatability of scanning technique A second complete study that included all scans to assess the retina, RNFL, GCC and ONH was performed in 11 eyes (seven normal; four predisposed). For all measures of total, inner and outer retinal thickness, repeatability of measurements obtained from different scans was adequate P a g e | 153

(ICC 0.748-0.835) while all measures of the GCC were considered inadequate (ICC 0.488 – 0.604) (Table 6.7). In the ONH analysis, the optic disc area had good repeatability (ICC 0.873) (Fig 6.11), but the remaining ONH parameters were considered of limited applicability due to their relatively poor repeatability (ICC 0.592 – 0.689).

Figure 6. 11. Separate scans of the optic nerve head/retinal nerve fibre layer obtained to determine repeatability (consistency) of measurements (case 22, predisposed eye). Good consistency was seen when comparing the optic disc outline (dark grey) which required manual adjustments, but the consistency of automated measures such as the optic cup (light grey) which could not be manually adjusted was of limited applicability.

Reliability of observer measurements Intra-rater reliability was determined for measures of retinal, RNFL thickness and ONH parameters. Intra-rater reliability was good or excellent (ICC>0.824) for all measures of total, inner and outer retinal thickness, and adequate for measures of RNFL in all sectors (ICC 0.713- 0.789) (Table 6.7). Intra-rater reliability of the optic disc area was considered excellent (ICC 0.942) (Fig 6.12), although when the optic cup to disc area and volume ratios were calculated based on automated optic cup volume detection, reliability of these ratios was only considered adequate (ICC 0.742 and 0.791 respectively).

P a g e | 154

Table 6. 7. Intra-class correlation coefficients for determination of repeatability (consistency) and intra-observer reliability

Structure Repeatability Intra-observer (n=11 eyes) reliability (n=25) Temporal retina - Total 0.773 0.824 - Inner 0.822 0.881 - Outer 0.794 0.909 Inferior retina - Total 0.748 0.942 - Inner 0.767 0.869 - Outer 0.753 0.843 Nasal retina - Total 0.787 0.902 - Inner 0.801 0.881 - Outer 0.815 0.893 Superior retina - Total 0.763 0.942 - Inner 0.797 0.878 - Outer 0.835 0.913 Retinal nerve - Temporal, upper 0.781 0.742 fibre layer - Temporal, lower 0.713 0.713 - Temporal, average 0.722 0.729 - Inferior, lateral 0.704 0.729 - Inferior, nasal 0.798 0.738 - Inferior, average 0.749 0.785 - Nasal, lower 0.735 0.727 - Nasal, upper 0.701 0.789 - Nasal, average 0.721 0.770 - Superior, nasal 0.713 0.737 - Superior, lateral 0.741 0.746 - Superior, average 0.720 0.741 - Average 0.742 0.756 - Average Dorsal 0.701 0.763 - Average Ventral 0.733 0.771 Ganglion cell - Total 0.600 N/A complex - Dorsal 0.532 N/A - Ventral 0.604 N/A - Intra-eye difference 0.488 N/A ONH - Cup: disc area 0.689 0.742 - Cup: disc volume 0.592 0.791 - Disc area 0.873 0.942 - Cup volume 0.598 N/A

P a g e | 155

Figure 6. 12. Assessment of intra-rater reliability determined by taking separate measurements (including manual adjustments) on the same optic nerve head/retinal nerve fibre layer scans with an interval of 6 months between analyses (case 6; normal eye). Excellent reliability was identified for measuring optic disc area (dark grey), while reliability of automated measures of optic cup (light grey) was adequate.

DISCUSSION This study was designed to evaluate whether OCT-derived measures of the retina and ONH have potential to aid in the early clinical assessment of canine glaucoma. The ability to detect changes in eyes predisposed to glaucoma prior to the onset of overt clinical signs that indicate late stage, irreversible damage, may allow for timely initiation of therapy at a stage where treatment could be more efficacious than current methods allow. In this study we identified statistically significant differences in structural measures of the retina when comparing normal and predisposed eyes. Based on these findings, the authors suggest further investigation and validation of OCT as a diagnostic tool in canine glaucoma is warranted. To the authors’ knowledge, this is the first report that includes a heterogenous sample of dogs as the control population.27-29,31,43 There are limitations in drawing direct comparisons between existing studies due to the different populations,27-29,31,43 scanning protocols, OCT units,27-29,31,43 and analyses.27-29,31,43 A varying distribution of RGC density associated with nose length,46 and differences in the expression of the visual streak within and between breeds is reported in dogs.47 Furthermore, in humans, SD-OCT measures of retinal and RNFL thickness, and ONH parameters are significantly correlated with age and race,48-50 and myopia is shown to independently influence RNFL thickness. 51 It is therefore important to consider the populations being studied, their relevance to clinical disease in a heterogenous population, and the scant reports describing OCT findings in dogs, when evaluating results of this, and previous studies. We identified a large degree of variability in measures of retinal thickness between sectors, and propose blood vessel orientation and ONH myelin as potential confounding effects. For practical reasons, we did not exclude every scan in which the dorsal retinal vein was not directed straight P a g e | 156 at the 12 o’clock position. A previous investigation into validity of RNFL measures using this device describe poorer reliability and repeatability of the measurements in sectors compared to quadrants.27 We therefore believe considering measurements in quadrants (rather than sectors) more appropriate with this device and using this technique. Despite the differences in populations and techniques, we identified similar general patterns to previous reports, with the inferior retina being thinner than the superior retina. Measures of average retinal thickness in our study fall within the range of what has previously been described in normal Beagles (inferior: 163.929 – 215.6µm28,31; superior: 198.229 – 218.1µm28,31). Using the definitions of inner and outer retinal thickness as recently reported (inner retina: inferior 92.0- 105.0µm, superior 88.3-101.3µm; outer retina: inferior 99.6-107.8µm, superior 113.6- 116.4µm),31 our results (inner retina: inferior 86µm, superior 107µm; outer retina: inferior 92µm, superior 101µm) show a similar pattern despite differences in technique, location and population. Despite the major limitations outlined previously for measurements of RNFL thickness, repeatability between scans and reliability of measurements requiring limited manual adjustments (to the ONH outline and the termination of the RPE), were adequate. Use of this parameter may therefore provide information in clinical disease, even if the measurements obtained are not precise measures of the RNFL in isolation. Compared to the post-processing requirements for measures of retinal thickness (adjusting three lines on each of 20 composite images), RNFL measurements relied on automated software algorithms. The RNFL thickness scans may therefore be more readily accepted by clinicians for whom efficiency, and getting results at the time of testing would be of practical importance. Variations in the location, depth, width and homogeneity of the peripapillary RNFL thickness using OCT occur in people,52 and regional variation was recently described in a small number of healthy Beagles.27,29 Median RNFL thicknesses in normal dogs in this series are similar to the manually adjusted mean values reported in those beagles,27 but were greater than the RNFL thickness reported in the control group of dogs assessed in a study on SARDS using a different OCT unit.29 The substantially thinner RNFL measures in control dogs in the latter study (superior- temporal RNFL thickness: 26.2±0.3µm; inferior-temporal: 24.9±0.4µm)29 likely reflects significant differences in both hardware and software algorithms used to establish these thickness values. Whereas regional differences in RNFL within an eye were described in beagles,27 we did not identify regional differences in RNFL thickness in eyes within either group. This difference may be due to the small sample size in both study populations, the different populations studied (homogenous population of young healthy beagles compared to a heterogenous population of pet dogs), or due to the method of data analysis. The identification of a thinner RNFL in predisposed compared to normal eyes, and the presence of a moderate correlation between RNFL thickness and the IOP following pharmacologic dilation in predisposed, but not in normal eyes is of particular interest. Neurodegeneration of RGCs is considered a hallmark of the diseases that comprise glaucoma, and in people with glaucoma, differing rates of structural change in the RNFL have also been reported between glaucoma progressors and non-progressors. 53-55 Further studies using OCT may provide additional evidence P a g e | 157 for how structural changes of the canine posterior segment in vivo might aid in diagnosis, prognostication and treatment of glaucoma in dogs. A significant limitation of this study is that a single scan (for each structure) was used, with scanning and measurements performed by a single investigator. Our data therefore fail to account for subtle differences in measures related to differences in orientation. Good reliability and repeatability of measurements and scans were obtained for those variables that did not have automated measures (or parts thereof). For measures of RNFL thickness that are automated, yet rely on manual adjustments of the ONH, both reliability and repeatability were considered adequate. When comparing the repeatability and intra-rater reliability achieved in this study with the single existing report describing inter- and intraobserver reliability in assessing RNFL thickness in normal beagles using the same OCT unit,27 it is important to consider the differences in interpretation of the ICC between studies. We used the classic parameters for interpreting ICC45, which are more conservative than those previously described.27 Despite the limitations of having a single observer scan and measure all subjects in this study, from a practical standpoint, we demonstrate the technique as a tool that can be operated by a single person, which can have benefits in clinical practice. Establishing whether alternative techniques, for example having a separate person, or stay sutures to stabilize the globe, would improve repeatability of the scanning procedure, scan quality, or efficiency in the scanning procedure, is beyond the scope of this study. Identification of the optic cup and disc in this series was not possible in all dogs and further investigation is paramount to determine the potential application and significance of these analyses. The only ONH parameter that could be accurately identified and manually delineated in this study was the optic disc area. The disc area measurements in our study were comparable to a previous report,27 achieved good reliability and repeatability, and this parameter is therefore of potential clinical use. However, further studies, and possibly implementation of methods to manually identify, adjust and measure the optic cup and rim, are required before use of these automated measurements can be considered applicable or clinically relevant. Similar to measures of the optic cup and rim, the GCC analysis relied on automated measurements and accurate positioning, with no ability to manually adjust the location or segmentation prior to analysis. This reliance on software, combined with limitations in accurately positioning the eye for follow up scans likely contributed to the poor repeatability and reliability in this study. Based on these findings, we suggest that the GCC analysis using the iVue proprietary software not be used for assessment of the retina without modification. There is a considerable degree of overlap between groups in most parameters studied. This is not an unexpected finding with known differences in RGC density associated with breed and conformation,46,47 variation in the presence and degree of intraocular myelin associated with the canine ONH, and our incomplete understanding of the pathophysiology of glaucoma. The extent of overlap in measurements between groups limits the use of a single OCT study as a stand-alone diagnostic test, but this is not unusual in veterinary diagnostics, and with validation, OCT may be a useful tool in the longitudinal clinical assessment of glaucoma in dogs, as applied in the management of glaucoma in people. P a g e | 158

Limitations in what an OCT unit can achieve should be considered when evaluating studies, and potential uses of this technology in veterinary ophthalmology. The Optovue iVue unit used in this study was selected based on availability and cost, the fact it is portable and simple to use, and still provides the benefit of high scanning speeds associated with SD-OCT. The cost, and the increasing availability of second-hand units that may enter the veterinary market mean this type of device might be more accessible in veterinary clinical practice. Some OCT units, such as the Heidelberg Spectralis (Heidelberg Industries), can detect and make adjustments so that the same area is imaged on serial scans. This feature is not available on the unit used in this study. A visible landmark was therefore imperative so serial measurements of the same structure and location can be made. The ONH has also previously been used as a landmark on which to centre OCT studies because all nerve fibre bundles are directed toward it, and the peripapillary region is therefore indicative of changes throughout the retina despite regional variations.27 However, the effect of retinal vasculature and intraocular myelin when using peripapillary measurements is more considerable, than if retinal scans were, for example, from the area centralis. The primary risk factor56 and only therapeutic target for glaucoma in dogs is an elevated IOP, yet disease progresses despite treatment in this multifactorial and incompletely understood disease. Findings in this series demonstrate that in vivo imaging of the retina and ONH in dogs can quantify structures relevant to glaucoma. With further validation of the technology for use in canine glaucoma patients, improved understanding of pathophysiological mechanisms, identification of early disease, and response to therapeutic interventions may become possible and improve treatment outcomes in dogs with this blinding and potentially painful disease.

ACKNOWLEDGEMENTS The authors would like to thank the Canine Research Foundation for sponsorship that enabled the conduct of this study, and Optimed Pty Ltd, for provision of the Koeppe goniolens that was used in this study. We would also like to acknowledge those veterinarians and owners that assisted with recruitment for this study. Finally we wish to thank the anonymous reviewers whose input made a significant contribution to this manuscript.

P a g e | 159

Table 6. 8. Description of quality and pattern of scans

“Good” scan Scan pattern quality index Retina >40 Raster pattern of 13 horizontal line scans (6mm long) Additional 7 horizontal line scans within central 1.5mm vertical zone) Each horizontal line scan sampled ≥ 5 times and averaged GCC >32 1 horizontal line (7mm scan length) followed by 15 vertical lines (7mm length, 0.5mm interval, centred 1mm temporal to the ONH) RNFL >27 13 concentric rings at diameters: ONH 4.9mm, 4.6mm, 4.3mm 4.0mm, 3.7mm, 3.4mm 3.1mm, 2.8mm, 2.5mm 2.2mm, 1.9mm, 1.6mm, 1.3mm

12 radial line scans (3.4mm length)

P a g e | 160

CHAPTER SEVEN DIFFUSION TENSOR IMAGING OF THE VISUAL PATHWAY IN DOGS WITH PRIMARY ANGLE CLOSURE GLAUCOMA

The following is the re-formatted manuscript currently under review for publication by John Wiley & Sons, Pty Ltd in Veterinary Ophthalmology: Graham KL, Johnson PJ, Barry E, Pérez Orrico M, Soligo DJ, Lawlor M, White A. Diffusion tensor imaging of the visual pathway in dogs with primary angle closure glaucoma.

ABSTRACT Objective To describe in vivo changes in the visual pathway beyond the retina and optic nerve head associated with canine primary angle closure glaucoma (PACG). Methods Magnetic resonance diffusion tensor imaging (DTI) was used to obtain quantitative measures of the optic nerve, chiasm, tract and lateral geniculate nucleus (LGN) in dogs with and without PACG. 3-Tesla DTI was performed on six affected dogs and five breed, age and sex matched controls. DTI indices were compared between healthy controls, dogs with unilateral, and dogs with bilateral PACG. Results Quantitative measurements of the optic nerve, optic tract, optic chiasm and LGN were obtained in all dogs, Dogs with bilateral PACG had a lower fractional anisotropy (FA) of the optic nerve (p=0.009) and lateral geniculate nuclei (LGN) (p=0.029), and higher FA (p<0.001) and axial diffusivity (AD) (p=0.029) of the optic tract compared to controls. Radial diffusivity (RD) in dogs with bilateral PACG was higher compared to both normal (p=0.042) and unilateral PACG (p=0.047). Dogs with unilateral PACG had lower FA of the optic tract compared to controls (p=0.015). DTI indices were similar whether affected eyes remained present (n=3) or had been enucleated/eviscerated (n=3). Conclusion Diffusivity and anisotropy measures provide a quantifiable means to evaluate the visual pathway in dogs. DTI has potential to provide in vivo measures of axonal and myelin injury and transsynaptic degeneration in canine PACG.

P a g e | 161

INTRODUCTION Glaucoma is a heterogenous group of optic nerve disorders with a final common pathway of progressive retinal ganglion cell (RGC) death and a characteristic glaucomatous optic neuropathy.13-18 Glaucomatous damage at the optic nerve head (ONH) results in anterograde degeneration of the RGC, causing changes in the optic nerve, optic chiasm, and optic tracts, and retrograde degeneration affects the retinal nerve fiber layer. Additionally, the process of transsynaptic degeneration (TSD) is proposed as one mechanism associated with glaucoma.503-505 TSD occurs when irreversible injury of the primary neuron proceeds to affect anatomically and/or functionally associated neurons, and occurs independent of inflammation. TSD is well known to occur in neurodegenerative disorders,506-508 and has been demonstrated at the level of the lateral geniculate nucleus (LGN) and in the optic radiation in human and non-human primates with glaucoma using advanced imaging techniques, including diffusion tensor imaging (DTI).503-505,509-517 These neurodegenerative changes have not been demonstrated in vivo in dogs with naturally occurring primary angle closure glaucoma (PACG). A better understanding of the changes within the retino-geniculo-cortical pathway that occur in PACG could further develop our understanding of the underlying pathophysiology in dogs, and inform on optimal treatment strategies, including potential merits of neuroprotective or neuroregenerative techniques. DTI is a magnetic resonance technique used to measure the properties of water diffusion,509 including the orientation and the degree of diffusion along certain axes.518 Water diffusion has specific characteristics within different types of biological tissue519 and when there is cellular pathology, the normal cytoarchitecture that restricts and directs water movement is disrupted and water diffusion is altered. DTI is able to identify these disruptions in diffusion and can therefore be used to evaluate the connectivity, integrity and architecture of neuronal tissue, including white matter.520 DTI indices are generated based on diffusion profiles and correlate with microstructural properties and pathological changes.521 Fractional anisotropy (FA) describes the preferential direction of water’s diffusion within each voxel. FA is measured on a scale from 0-1 where 0 represents maximum isotropy (when water is freely mobilized) and an FA of 1 represents maximum anisotropy, or restriction in the diffusion of water. Mean diffusivity (MD) describes the average mobility of water molecules within a voxel and is proposed to correlate with microstructural disintegration.518 Axial diffusivity (AD) quantifies diffusivity along the principal axis and radial diffusivity (RD) provides an average measure of diffusion along the two minor axes. As different tissue and cell structures affect the direction and form of diffusion, these indices inform on the integrity of neuronal and axonal microstructure and have been suggested as potential early biomarkers of axonal and myelin injury.509-511,522,523 The visual pathway has been investigated using DTI in people with primary open angle glaucoma,524-526 and in experimental models of ophthalmic disease.527-531 When used in the evaluation of human glaucomas, DTI identified a decrease in FA, and increases in MD along the optic tract and optic radiation, as well as increases in RD in the optic radiation, consistent with the presence of TSD.512,514,516,517,532 DTI has been described P a g e | 162 as a feasible technique in the dog and has the potential to identify the presence of neurodegeneration associated with glaucoma in this species.533-536 The aim of our study was to identify if measures of water diffusion are altered within the optic nerve, optic chiasm, optic tract or lateral geniculate nuclei (LGN) in dogs with PACG using DTI. We hypothesize that neurodegenerative changes within the retino- geniculo-cortical pathway of dogs with PACG will cause alterations in diffusivity detectable with DTI.

MATERIAL AND METHODS Subjects Eleven client owned dogs were recruited and informed owner consent was obtained prior to inclusion in the study. All procedures were approved by the University of Sydney Animal Ethics Committee (2017/1156). Recruitment criteria for this study included dogs diagnosed with PACG in one or both eyes by a board certified ophthalmologist and age-, breed- and sex-matched disease-free control subjects. Diagnosis with PACG was based on the following criteria: 1) current or previous IOP ≥25mmHg; 2) changes consistent with glaucoma in the retina (hyperreflective retina, attenuation of retinal vasculature) and/or optic nerve head; and 3) an abnormal iridocorneal angle (closed, narrow, pectinate ligament abnormalities) on gonioscopic examination in the contralateral eye at the time of inclusion, or previously documented. Exclusion criteria included 1) the presence of ocular or systemic disease that might result in secondary glaucoma; and 2) an absence of iridocorneal angle and/or pectinate ligament abnormalities (primary open angle glaucoma). Each side of the visual pathway in affected dogs was then classified as being either affected by glaucoma or as predisposed to the development of PACG based on the clinical diagnosis in the ipsilateral eye. Predisposed eyes included the contralateral eye of dogs diagnosed with unilateral PACG. These eyes had no documented elevation in IOP (IOP <25mmHg), and no evidence of glaucomatous damage to the retina or optic nerve. Unaffected dogs had no ocular abnormality and a normal drainage angle on gonioscopic examination in both eyes, as well as no evidence of systemic disease. The eyes in each unaffected dog were classified as normal. To investigate the effect of optic nerve transection (enucleated globe), and removal of the retina (evisceration) in glaucomatous eyes, DTI indices were compared between cases that had undergone enucleation or evisceration and cases where the glaucomatous eye was still present. Ophthalmic examination All dogs had a complete physical and ophthalmic examination, including slit lamp biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, UK), indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, NSW Australia), quantitative tear testing (Schirmer tear test, Merck Animal Health, NJ, USA), rebound tonometry (Icare Tonovet, Icare, Finland) and fluorescein staining. Gonioscopy was performed P a g e | 163 on all eyes at the time of the study when the cornea was appropriately transparent. In chronically affected cases, the results of gonioscopy performed at/around the time of original diagnoses were used. For gonioscopic examination a direct (Koeppe, Ocular Instruments, Bellevue, WA, USA) or indirect (G-4 Four Mirror, Volk Optical Inc, Mentor, OH, USA) goniolens was used according to clinician preference, and findings recorded as either normal or abnormal. B-mode ultrasonography (10MHz probe, Philips HD-11, Philips Australia) was performed in all dogs with glaucoma to confirm the absence of any intraocular structural change that may result in secondary glaucoma. Imaging protocol For image acquisition, dogs were fasted for a minimum of 8 hours, and premedicated with methadone (0.1-0.4mg/kg IM) with or without acepromazine (0-0.03mg/kg IM depending on the health and temperament of the individual dog. Anaesthesia was induced using propofol (4-6mg/kg IV) and/or thiopentone (4mg/kg IV) to effect via an indwelling intravenous canula. Each patient was intubated, and anaesthesia maintained with inhalational isoflurane and oxygen for the duration of imaging. Imaging was performed on a 3.0T GE Discovery MR750 (GE Healthcare, Milwaukee, WI) whole body scanner using an 8-channel extremity coil (HD Foot Ankle array, Invivo) with the dog positioned in dorsal recumbency. Anatomical data were obtained in a T1-weighted 3D fast spoiled gradient recalled echo (FSPGR) pulse sequence (FOV = 115mm, TR = 6ms, TE = 2.8ms, flip angle = 12°, inversion time = 450ms, acquisition matrix size = 192 x 192, bandwidth = 434Hz/px, number of signal averages = 1, reconstructed axial plans with an acquisition time of 6min 36 seconds and 0.6mm slice thickness). DTI was performed in the coronal plane with 1.5mm slice thickness and no interslice separation using a 2D spin echo-echo planar imaging (SE-EPI) diffusion tensor sequence to cover the whole visual pathway (FOV = 144mm, TR = 6000ms, TE = 60ms, acquisition matrix size = 96 x 96, bandwidth = 1953Hz/px, number of directions = 32, number of b0 =1. b-value 800s/mm2, number of signal averages = 5, and with an acquisition time of 17min 35seconds). An additional SE-EPI sequence with reversed phase encoding polarity was also acquired to correct for magnetic susceptibility related distortion. Data processing Each study was evaluated by a European certified veterinary radiologist (PJJ) for evidence of disease that was unrelated to changes expected in association with glaucoma. Datasets were corrected for noise, Gibb’s artifact, motion and eddy current distortions using FSL (https://fsl.fmrib.ox.ac.uk/fsl/) toolbox.537 Diffusion tensor maps for FA, MD, AD, and RD were calculated using FSL’s Diffusion toolbox.538 Individuals’ T1-weighted images were registered to native diffusion space in order to provide a high resolution anatomical reference for the manual delineation of regions of interest (ROI). Both T1-weighted and white matter landmarks visible on FA maps were used for ROI placement. In order to reduce partial volume averaging, limit differences introduced by variable ROI size and sample multiple sites within each structure three ROIs standardized in size and shape were placed at multiple sites within each structure P a g e | 164 along the optic pathway by a neuroanatomical expert (PJJ) (Figure 7.1).539 The ROIs measured 3.38mm3 for optic nerve sampling, 4mm3 for optic chiasm sampling, 2.25mm3 for optic tract sampling and 5mm3 for lateral geniculate nucleus sampling. Each ROI was applied as a mask to the FA, MD, AD, and RD maps to isolate the region (Figure 7.2). The mean and standard deviation of FA, MD, AD, and RD were then calculated for each ROI. ROI placement at each location was repeated in triplicate and the mean used in statistical analyses.

Figure 7. 1. Magnetic resonance images demonstrating regions of interest (ROI) placement. A) Optic nerve ROIs (pink) in transverse plane, B) Optic chiasm ROI (yellow) in transverse plane, C) Optic tract ROIs (blue) in transverse plane, D) Lateral geniculate nuclei ROIs (green) in transverse plane, E) Opteic nerve (pink), optic chiasm (yellow) and optic tract (blue) ROIs in dorsal plane and F) Lateral geniculate nuclei (green) ROIs in dorsal plane. The images are composed of 3- dimensional T1-weighted images overlaid with fractional anisotropy (FA) heat maps (red = low FA, white = high FA). P a g e | 165

Figure 7. 2. Diffusivity maps in dorsal (top) and transverse (bottom) plane, generated for (from left to right) fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), and axial diffusivity (AD).

Statistical analysis Statistical analyses were performed using commercially available software (GraphPad Prism 7.02, IBM SPSS Statistics v22.0). For all results a p value <0.05 was considered statistically significant. Subjects For analysis of skull shape, each dog was classified as either brachycephalic, small mesocephalic, or large mesocephalic based on conformation (skull shape and size of dog for small versus large mesocephalics) (Table 7.1). A Mann Whitney test was performed to compare subject characteristics (age, body weight, skull shape, IOP at the time of imaging) between dogs with and without PACG. A Pearson correlation was calculated to measure linear dependence between IOP at the time of imaging and both age and the duration of disease in dogs with PACG. A Pearson’s chi-square test was performed to identify statistically significant relationships between categorical variables (disease status of the eye with that of the contralateral eye, whether the dog was pure- or cross-bred, and skull shape). Cramer’s V (φc) was then used to determine the strength of association for any significant finding. P a g e | 166

Intra-observer reliability In order to test the reliability of sampling between placed ROIs within each structure, optic nerve, optic chiasm, optic tract and lateral geniculate nuclei, intra-observer reliability was tested using intra-class correlation coefficients (ICC). This was performed using a two-way mixed model to obtain an assessment of absolute agreement. Intra-observer reliability was assessed as acceptable if the ICC measured 0.7-0.79, good if it measured 0.8-0.89 and excellent if it measured >0.9. DTI measurements For each structure the results from the three ROIs were averaged to obtain a mean for each DTI parameter. A one-way ANOVA was used to investigate differences in mean DTI parameters between groups (healthy controls versus unilateral PACG versus bilateral PACG). Multiple regression analyses were performed to understand whether DTI parameters could be predicted based on independent variables including age, breed (pure- versus cross-bred), the disease status of each eye (normal, unilateral PACG, bilateral PACG), IOP at the time of imaging, and duration of disease. To investigate potential differences between predisposed and affected sides of the visual pathway, a Welch’s t-test was performed to compare DTI measurements ipsilateral to a glaucomatous eye with those ipsilateral to a predisposed eye, and to compare DTI measurements contralateral to a glaucomatous eye with those contralateral to a predisposed eye. Bonferonni Correction was conducted to minimise the influence of Type I error. Welch’s t- tests with Bonferonni Correction were also conducted to compare DTI measures between dogs with PACG that had an enucleation/evisceration of a glaucomatous eye with those where the glaucomatous eye remained in situ.

RESULTS Subjects Six dogs diagnosed with PACG in one (n=3) or both eyes (n=3) and five age-, breed- and sex-matched disease-free control subjects were recruited for this study between February and July 2017. Three dogs with PACG had previously had a glaucomatous eye enucleated (n=2) or eviscerated (n=1) between 3.4 – 4.4 years prior to imaging for this study (Table 7.1). When comparing dogs with (n=6) and without glaucoma (n=5), there was no statistically significant difference in age (p=0.693), body weight (p=0.931) or skull shape (p=0.836) between dogs with glaucoma and healthy controls. The signalment, clinical characteristics, and grouping of subjects is documented in Table 7.1. There was a strong association between disease status of the ipsilateral and contralateral eyes (χ2=20.78, p=0.014, φc=0.638), but no statistically significant P a g e | 167 relationship between disease status of the eye and whether the dog was pure- or cross-bred (p=0.404), and skull shape (p=0.792). There was no statistically significant difference in median IOP at the time of imaging in dogs with glaucoma (median 15mmHg, range 3-32mmHg) compared to control dogs (median 15mmHg, range 11-16mmHg) (p=0.469) (Fig.7.3). In dogs with glaucoma, there was no statistically significant correlation between IOP at the time of imaging and the duration of glaucoma when considering both glaucomatous and predisposed eyes (r=0.288, p=0.443) or predisposed eyes alone (r=0.500, p>0.999). In the control population, no statistically significant correlation between age and IOP was identified (r=-0.277, p=0.436). T1-weighted magnetic resonance imaging No abnormality in the structure of the brain, or in the shape and signal intensity of intracranial structures was identified in dogs with glaucoma or the control group. Intra-observer reliability Intra-observer reliability was excellent for measurements of the optic nerve (ICC 0.92), good for measurements of the optic tract (ICC 0.89) and acceptable for measures of the optic chiasm (ICC 0.71) and lateral geniculate nuclei (ICC 0.76). Diffusion tensor imaging parameters There was no statistically significant difference between paired structures (optic nerves, optic tracts or LGN) within dogs for any DTI parameter assessed in any group (healthy controls, unilateral PACG, bilateral PACG). When comparing DTI measurements of each structure between groups, there was a statistically significant difference in FA of the optic nerve (p=0.012), optic tract (p<0.001), and LGN (p=0.037); in AD of the optic tracts (p=0.027); and in RD of the optic nerve (p=0.041) and LGN (p=0.025) (Table 7.2). FA of the optic nerve was greater in healthy controls (0.4142, 95% CI [0.3198, 0.5086]) compared to dogs with bilateral PACG (p=0.009) (Fig.7.4). FA and AD of the optic tract was higher in healthy controls compared to bilateral PACG (FA: normal 0.5637, 95% CI [0.4457, 0.6717], bilateral PACG 0.3829, 95% CI [0.3377-0.4261], p<0.001; AD: normal 0.0014, 95% CI [0.0013, 0.0016], bilateral PACG 0.0012, 95% CI [0.0010, 0.0014], p=0.029) (Fig.7.5), and FA of the optic tract in unilateral PACG (0.4530, 95% CI [0.4197, 0.4863]) was also less compared to that in healthy controls (p=0.015). In the LGN, FA was greater in healthy controls (0.57, 95% CI [0.4529, 0.6870]) compared to that in bilateral PACG (0.4358, 95% CI [0.3580, 0.5136]) (p=0.029). RD in dogs with bilateral PACG (0.0008, 95% CI [0.005, 0.0012) was greater than both healthy controls (0.0005, 95% CI [0.0004, 0.0006]) (p=0.042) and unilateral PACG (0.0005, 95% CI [0.0004, 0.0006]) (p=0.047) (Fig.7.6).

P a g e | 168

Table 7. 1. Characteristics of individual subjects

I.D Eye Study group Study group Age Breed Sex Glaucoma Glaucoma Treatment IOP (dog) (eye) (yrs) duration (days) (mmHg) HEALTHY CONTROLS 2 L Normal Normal 9 Labrador X FS - N/A 13 Poodle R Normal - N/A 16 5 L Normal Normal 6 Terrier X FS - N/A 14 R Normal - N/A 12 6 L Normal Normal 9 Cocker Spaniel FS - N/A 15 R Normal - N/A 15 9 L Normal Normal 11 Maltese X FS - N/A 11 R Normal - N/A 12 10 L Normal Normal 7 Cattle Dog FS - N/A 16 R Normal - N/A 15 UNILATERAL GLAUCOMA 1 L Unilateral Predisposed 9 Labrador X FS - Dorzolamide/Timolol BID since diagnosis of PACG OD 15 glaucoma Poodle R Glaucoma 484 Latanoprost BID 12 4 L Unilateral Predisposed 6 Terrier X FS - Dorzolamide/Timolol BID since diagnosis of PACG OD 12 glaucoma R Glaucoma 1252 (since ISP placed at time of diagnosis with glaucoma. N/A evisceration) Histopathology consistent with PACG 8 L Unilateral Glaucoma 11 Maltese X FS 1634 Latanoprost and Dorzolamide/timolol BID (buphthalmos) 32 glaucoma R Predisposed - Dorzolamide/Timolol BID since diagnosis with PACG OS 17 BILATERAL GLAUCOMA 3 L Bilateral Glaucoma 10 Labrador X FS 1616 (since Diagnosed with end-stage glaucoma approximately 2 months N/A glaucoma Poodle enucleation) after clinical signs described R Glaucoma 1398 Intermittent use of topical NSAID (ketorolac); >3 years since 3 TSCP (x2) failure; 29 months since shunt placed 7 L Bilateral Glaucoma 11 Maltese X FS 270 Topical ketorolac BID; five months after placement of shunt 9 glaucoma R Glaucoma 1261 (since Diagnosis with end-stage glaucoma (no prior treatment) N/A enucleation) 11 L Bilateral Glaucoma 7 Cattle Dog FS 478 Latanoprost and Dorzolamide/timolol BID 32 glaucoma R Glaucoma 445 Latanoprost and Dorzolamide/timolol BID 26 P a g e | 169

Table 7.1 documents the identification number, study group by subject, study group by eye, age, breed, sex, duration of disease, treatment protocol and intra-ocular pressure at imaging of each subject included in the study. L = left; R= right; FS = female spayed; BID = twice daily; TSCP = transscleral cyclophotocoagulation; OS = left eye; OD = right eye

Table 7. 2. Diffusion tensor imaging parameters of the visual pathway in dogs with and without primary angle closure glaucoma (PACG).

Group Optic nerve Optic tract Optic LGN chiasm FA Normal 0.4142a 0.5637a,b 0.1709 0.5700a (0.3198 - 0.5086) (0.4557 - 0.6717) (0.1042 - 0.2376) (0.4529 - 0.6870) Unilateral PACG 0.3316 0.4530a 0.1830 0.5007 (0.2357 - 0.4275) (0.4197 - 0.4863) (-0.0244 - 0.3903) (0.4371 - 0.5643) Bilateral PACG 0.2519a 0.3819b 0.1288 0.4358a (0.1922 - 0.3116) (0.3377 - 0.4261) (-0.0219 - 0.2795) (0.3580 - 0.5136) Normal 0.0015 0.0008 0.0018 0.0008 MD (0.0012 - 0.0019) (0.0008 - 0.0009) (0.0015 - 0.0022) (0.0007 - 0.0009) Unilateral PACG 0.0014 0.0009 0.0017 0.0007 (0.0012 - 0.0017) (0.0009 - 0.0010) (0.0006 - 0.0027) (0.0006 - 0.0009) Bilateral PACG 0.0016 0.0008 0.0021 0.0008 (0.0014 - 0.0019) (0.0007 - 0.0010) (0.0013 - 0.0029) (0.0007 - 0.0010) AD Normal 0.0023 0.0014a 0.0022 0.0013 (0.0019 - 0.0026) (0.0013 - 0.0016) (0.0017 - 0.0026) (0.0012 - 0.0015) Unilateral PACG 0.0020 0.0014 0.0019 0.0012 (0.0018 - 0.0022) (0.0013 - 0.0015) (0.0011 - 0.0028) (0.0009 - 0.0014) Bilateral PACG 0.0021 0.0012a 0.0023 0.0013 (0.0019 - 0.0023) (0.0010 - 0.0014) (0.0016 - 0.0031) (0.0010 - 0.0016) RD Normal 0.0012 0.0011 0.0005 0.0005a (0.0009 - 0.0015) (0.0010 - 0.0012) (0.0004 - 0.0007) (0.0004 - 0.0006) Unilateral PACG 0.0012 0.0011 0.0007 0.0005b (0.0009 - 0.0014) (0.0005 - 0.0017) (0.0005 - 0.0008) (0.0004 - 0.0006) Bilateral PACG 0.0019 0.0017 0.0009 0.0008a,b (0.0011 - 0.0027) (0.0006 - 0.0027) (-0.0004 - 0.0022) (0.0005 - 0.0012)

Measures of fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD) and radial diffusivity (RD) [mean (lower – upper bound of 95% confidence interval)] are documented. Measures of anisotropy (fractional anisotropy) are scalar (0-1); measures of diffusivity (mean, axial and radial) are reported in mm2/sec. Parameters for which there was a statistically significant difference between groups (p<0.05) are listed in bold, and the groups between which the mean difference was statistically significant are indicated with the same superscript. P a g e | 170

Figure 7. 3. Intraocular pressure (mmHg) at the time of imaging in eyes diagnosed with primary angle closure glaucoma compared to healthy control eyes. There was no statistically significant difference in the median IOP of dogs with glaucoma (median 15mmHg, range 3-32mmHg) compared to control dogs (median 15mmHg, range 11-16mmHg) (p=0.469).

Multiple regression analyses were performed to predict DTI measurements from age, breed, disease status of each eye, IOP at the time of imaging, and disease duration. The only DTI measurement that was predicted with statistical significance by these variables was FA of the optic nerve (F=388.56, p=0.003, adjusted R2=0.997). The variables that contributed to the prediction with statistical significance for FA of the optic nerve included the disease status of the ipsilateral eye (p=0.002), age (p=0.028), whether the dog was pure- or cross-bred (p=0.033), and duration of glaucoma for that dog (p=0.001). When comparing DTI measures of the visual pathway ipsilateral to a glaucomatous eye, FA of the optic nerve was lower (p=0.031), and RD higher (p=0.022) ipsilateral to the predisposed compared to the glaucomatous eye. When measures of the visual pathway contralateral to a glaucomatous eye were compared between these groups, FA of the optic tract (p=0.033) and RD of the LGN (p=0.026) were higher contralateral to a predisposed compared to glaucomatous eye. When comparing the glaucomatous eye of dogs with unilateral glaucoma (unaffected contralateral eye; n=3) with enucleated/eviscerated eyes (n=3), there was no statistically significant difference in DTI indices of the optic nerve.

P a g e | 171

Figures 7.4-7.6 show DTI parameters of the optic nerve, optic tract, and LGN nucleus in control dogs and those with unilateral and bilateral primary angle closure glaucoma (PACG). The box represents the 95% confidence interval with mean AD indicated by a horizontal line, while the whiskers represent the range.

Figure 7. 4. Fractional anisotropy (FA) of the (A) optic nerve, (B) optic tract, and (C) LGN

Figure 7. 5. Axial diffusivity (AD) of the (A) optic nerve, (B) optic tract, and (C) LGN

Figure 7. 6. Radial diffusivity (RD) of the (A) optic nerve, (B) optic tract, and (C) LGN P a g e | 172

DISCUSSION In this study we obtain measures of diffusivity in the visual pathway of dogs with and without angle closure glaucoma. Using this non-invasive technique, in vivo quantitative differences in white matter tracts of the optic nerve, optic tract, and LGN were identified in dogs with PACG when compared to healthy controls. The loss of RGCs and their axons characterizes glaucoma, and reliable markers of these axons are required to further develop our understanding of pathophysiology, and potential advances in management of the disease. Optical coherence tomography (OCT) is routinely used in the assessment of glaucoma in people, where it provides quantitative measures of retinal thickness, including the retinal nerve fibre layer. This technology is not yet routinely available in veterinary practice, however its use is increasingly reported for assessment of the fundus in the veterinary literature467,473,496,540-553 and OCT is proposed as an alternative approach to the detection of early disease recognition.554 Diffusion weighted imaging, provides alternative approach to early evaluation and is a method of obtaining quantitative measures of early cellular damage in glaucoma where OCT cannot be used. For example, the degree of corneal and lens opacities in dogs with naturally occurring disease, but especially the obstruction that results from miosis secondary to topical prostaglandin analogue use, precludes ready assessment of the retina and optic nerve head in this species. In our study, we identified a lower FA in the optic nerve, tract and LGN in dogs with bilateral PACG compared to healthy controls, while FA of the optic nerve was also lower in unilateral PACG compared to healthy controls. These findings are consistent with reports describing lower FA in the optic tract and radiation in people with glaucoma,512,516,517,532 and is consistent with reduced axonal integrity of affected nerves. However, with 78% of RGC axons crossing over at the optic chiasm in dogs,555 and asymmetric onset and progression of PACG in this species, conclusions about the influence of the each eye on DTI measurements of the visual pathway cannot be made from this study. Radial diffusivity is recognized as a relatively specific indicator of myelination556 and its association with neuronal disease557 has been demonstrated in the canine brain (ex vivo536 and in a model of demyelination and brain maturation558) with an increased RD reflecting myelin loss. We identified greater RD in the LGN of dogs with bilateral PACG glaucomatous compared to normal dogs. As two of the three dogs with bilateral PACG in this study had previously had an eye enucleated due to end-stage glaucoma, no conclusion about the role glaucoma alone has on measures of radial diffusivity can be made from these results. However our findings may suggest the presence of axonal loss and demyelination of the LGN in dogs with PACG, and although RD is indicative of myelination, we cannot exclude the potential for dense cell bodies present in the LGN to have influenced the DTI changes reported here. As the first synapse of the RGC axons occurs at the LGN, such a change would provide evidence of TSD in canine glaucoma in vivo which has not previously been reported. Although reports exist suggesting the presence of neurodegeneration separate to that resulting from RGC damage, our results support findings described in people with glaucoma without P a g e | 173 providing any evidence that the changes are not secondary to the anterograde and retrograde degeneration from the RGC that characterizes glaucoma.559 In our study, no statistically significant difference in MD of the optic nerves or optic tracts between any group was identified. This is interesting, especially considering the RD between groups was also statistically similar despite the presence of dogs with different stages of disease. Both MD and RD of the optic nerve in people with open angle glaucoma is greater than in normal eyes.512,515 This lack of a difference may be secondary to our low case numbers and further studies are required to determine the significance of MD and/or RD in evaluating neurodegeneration in the visual pathway in both glaucoma, and when not associated with glaucoma. Identifying whether these measures in dogs are similar, or vary to those in people with POAG, may aid in understanding of physiological differences between the glaucomas in each species. This study involved a small study population. With strict inclusion criteria, recruiting appropriate animals presented some problems. In future studies, imaging dogs at the time of diagnosis and before therapeutic interventions are implemented would be prudent. Restricting dogs with glaucoma to one group (either predisposed or glaucoma, rather than including one eye/side in each group), may provide meaningful data that could further elucidate some of the associations identified here, and other potential factors. The inclusion of dogs with glaucoma that had an eye previously enucleated or eviscerated adds a confounding factor to the measures in affected dogs. Transection of the optic nerve in enucleation of the globe has been shown to result in TSD with resultant changes of the LGN in people and nonhuman primates.560-563 Given the influence optic nerve transection and/or retinectomy is expected to have on DTI parameters, we compared the indices in enucleated/eviscerated eyes with those in dogs with unilateral glaucoma. With no statistical significance identified in any structure assessed, we propose the neurodegeneration and TSD established in people also occurs in dogs with glaucoma. Further studies with adequate numbers of dogs affected by glaucoma without enucleation or evisceration are required to validate this finding and identify the impact of both glaucoma and optic nerve transection/retinectomy on parameters indicative of neurodegeneration. The findings presented in this study describe quantitative measures obtained in dogs at a single time point. Despite grouping of dogs according to clinically identifiable stages of disease, the chronicity of disease in affected dogs still varied within groups. Longitudinal studies may help identify the relationship between the onset and progression of changes in the CNS, and clinically recognizable disease. Although we identified similarities to those reported in people as outlined above, the absence of any statistically significant difference in diffusivities of affected and unaffected optic nerves and tracts in our study should be noted. Whether this is related to the physiology of optic nerve degeneration (comparing POAG in people with PACG in dogs in this series), the small sample size, the variation in chronicity of disease in these subjects, or a combination of these and other factors, is beyond the scope of this preliminary study. No statistically significant difference in DTI indices were obtained in assessment of the optic chiasm in this study. This may be due to the disease process or be associated with the inherent limitations of DTI measures in relation to factors such as non- P a g e | 174

Gaussian diffusion, crossing fibres and imaging noise, the exact impact of which is beyond the scope of this study.564,565 In this study, segmentation of each ROI was manually performed by a single investigator, experienced in the use of DTI and neuroimaging (PJJ), and intra-observer reliability decreased from the proximal to distal visual pathway. This technique is the standard method for DTI analysis in the veterinary literature, however is suggested to be less efficient, more variable, and less accurate than automated segmentation.566,567 Automated techniques have been performed in people,568 non-human primates,569,570 and the cat,523 however despite the presence of several digital atlases for the canine brain,571-573 anatomic priors for the optic pathway that would allow for automated segmentation in this study are lacking. Our cohort had a variable skull and brain shape making registration for automated segmentation more challenging and necessitated manual ROI placement by a veterinary neuroimaging expert. Further assessment of the impact of skull shape on brain registration in dogs is required to validate the use of automated techniques in mixed skull shape cohorts, and future studies with multiple observers are indicated to determine the merits of manual versus automated segmentation should automated options become available. This study evaluated quantitative in vivo changes in white matter tracts of the visual pathway in dogs and adds to the growing research on structural changes associated with canine glaucoma. With validation of objective and measurable changes in the visual pathway, we propose dogs with naturally occurring glaucoma may play an important role in the development of therapeutic strategies, with the potential to progress understanding and management of the disease in both animals and people.

ACKNOWLEDGEMENTS The authors would like to thank the following people for their contribution to this study: Jessica Aalders for assistance with anaesthesia and imaging of the dogs; Dr Jennifer Chau BVSc (Hons) FANZCVS GradDipVetClinStud MVetStud for advice in the preliminary stages; and the dog owners and veterinarians who assisted with recruitment. We would like to thank and acknowledge the contribution of the directors at the Eye Clinic for Animals in Sydney, and the Canine Research Foundation for funding that supported this project.

P a g e | 175

CHAPTER EIGHT SHOTGUN PROTEOMIC ANALYSIS OF THE PRECORNEAL TEAR FILM IN DOGS WITH PRIMARY GLAUCOMA

The following is a re-formatted manuscript currently under review for publication by John Wily & Sons, Pty Ltd in Veterinary Ophthalmology: Graham KL, Diefenbach E, McCowan CI, Pattamatta U, White AJR. Shotgun proteomic analysis of the precorneal tear film in dogs with primary glaucoma.

ABSTRACT Objective: To characterise protein patterns in tears of dogs with primary angle closure glaucoma (PACG) and primary open angle glaucoma (POAG). Animals: Nineteen dogs (25 eyes). Methods: Tear samples were collected using a Schirmer tear strip, from dogs with PACG (PACG-affected eyes, n=8; unaffected eyes predisposed to PACG, n=7), POAG (n=4), and healthy controls (n=6). Protein precipitation and trypsin digestion were performed for analyses via liquid chromatography tandem mass spectrometry. Protein components were characterised, and proteins identified using the SwissProt protein sequence database. Relative protein expression in 17 eyes (15 dogs) was evaluated using Proteome Discoverer 2.0. Pathway analyses were performed to investigate molecular mechanisms associated with primary glaucoma. Results Unique peptides were identified in 505 proteins, with Major allergen Can f 1 and albumin identified with high confidence. Proteins unique to tears from diseased eyes (PACG: n=7; POAG: n=14) were identified. Nucleoside diphosphate was unique to tears in PACG eyes naïve to therapy, while Retinal binding protein and NSFL1 cofactor p47 were unique to medicated PACG eyes. Relative expression of 34 proteins differed between disease states. Pathway analyses identified highest ranked upstream regulatory proteins in each disease state, and that the ‘inflammatory response’ was among the top disease/disorders in dogs with primary glaucoma (PACG and POAG) but not in healthy controls. Conclusion: Tear samples suitable for mass spectrometry were readily obtained from pet dogs without needing specialised equipment. Further studies to validate the findings and explore potential candidate biomarkers for early disease detection and potential therapeutic targets are indicated.

P a g e | 176

INTRODUCTION The need to better understand the mechanisms involved in canine glaucoma has recently been highlighted.468 The search for biomarkers, reflecting molecular and cellular mechanisms of disease, presents an opportunity to achieve this goal, with structural,495,497,574-577 functional,372-374,443,444,448,578-580 and molecular,20,49,52,182,581-590 indicators of disease increasingly reported. The tear film is a readily accessible body fluid which can be sampled in a non-invasive manner, with the potential to provide information about molecular events at the ocular surface, and therefore represents a potential source of biomarkers.591 Comprising proteins produced by the lacrimal gland and serum proteins that leak from conjunctival vessels,212,592,593 the tear film nourishes the ocular surface and removes local waste products, metabolised drugs and inflammatory mediators produced in several ophthalmic diseases.594 In people, the tear proteome has been shown to reflect both ocular591,595,596 and systemic disease,597-601 thus definition of the tear film protein composition may provide insight into the pathophysiology of glaucoma and ocular surface modifications induced by topical therapy.594 Research into the tear film, and its value as a source of biomarkers of disease, has been limited by factors such as the small volumes available for collection and variations in collection methods, but with increasing availability of technologies such as mass spectrometry, exploration of tears as a source of biomarkers is now possible.212 Sampling of the canine tear film, via quantitative measurement of reflex tearing using a Schirmer tear test, is a routine part of clinical ophthalmic examinations. This allows sample collection from sufficient numbers of dogs to enable the conduct of suitably powered studies. Differential expression of proteins in the tears of dogs with and without cancer shows the potential of tears as a source of biomarkers of disease.602 However, despite the identification of unique proteins in the tears of people with different forms of glaucoma, the demonstrated influence of glaucoma medications on the ocular surface and treatment outcomes, as well as the investigation of potential markers of glaucoma in other body fluids182,574 exploration of the canine tear film as it relates to canine glaucoma has not previously been reported. In this study we initially sought to determine whether tear samples obtained during a routine ophthalmic examination using commercially available tear strips could be used to collect tear samples for use in the search for biomarkers of glaucoma in dogs. Our aim was to characterise tear proteins in dogs with and without naturally occurring primary glaucoma as an initial step in determining the feasibility for using tears in the investigation and management of canine glaucoma. MATERIALS AND METHODS Tear samples were obtained from 19 dogs (25 eyes) presenting to veterinary hospitals in NSW, Australia. A complete ophthalmic examination including slit lamp biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, UK), direct and indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, NSW Australia), tonometry (Icare Tonovet, Icare, Finland), and fluorescein staining of the cornea (OptiStrips-FL Sterile Fluorescein Strips, Bausch + Lomb, Australia) was performed in each case. Quantitative tear film assessments were also performed as outlined below. This study was conducted with the approval of the University of Sydney Animal Ethics Committee (AEC 2016/1004) and samples were collected and stored with informed owner consent. P a g e | 177

Subjects Subjects included 13 dogs with primary glaucoma and six healthy controls without evidence of ophthalmic or systemic disease. Based on clinical and ophthalmic examination and medical history, eyes were individually selected for inclusion in the study according to the following criteria: (1) eyes predisposed to primary angle closure glaucoma (PACG; n=7) included those with an abnormal iridocorneal angle (ICA) identified on gonioscopy (narrowed or closed ICA with or without pectinate ligament dysplasia) in dogs where the contralateral eye had been diagnosed with PACG by a veterinary ophthalmologist; (2) eyes with glaucoma were diagnosed by a veterinary ophthalmologist as being PACG (n=8) or primary open angle glaucoma (POAG; n=4). All glaucomatous eyes had a documented elevation in intraocular pressure (IOP) (>25mmHg) in association with clinical signs of glaucoma including one or more of the following: vision impairment or loss, buphthalmia, corneal oedema, retinal degeneration and/or optic nerve degeneration (decreased size, loss of myelin, cupping of the optic nerve head). PACG was differentiated from POAG by gonioscopy as above, and the contralateral eye in each dog with POAG was also diagnosed with POAG of similar severity/chronicity. No subject had current or previous ocular or systemic disease that could result in secondary glaucoma. (3) A tear sample collected from one randomly selected eye in six dogs presenting to a veterinary hospital with no evidence of ophthalmic or systemic disease were selected as controls (normal). Tear collection and storage Sampling of the precorneal tear film was performed using one of two commercially available Schirmer tear test strips (Merck Animal Health, NJ USA; or Haag-Streit UK Ltd, Harlow UK). The strip was placed so that the tip contacted the inferior-temporal corneal surface. The distance of strip wetting (millimetres) was recorded after 60 seconds. The strip was placed into an Eppendorf polypropylene microcentrifuge tube containing 200µL of sterile water and stored at -20°C for up to 24 hours. Samples were centrifuged at -4°C for 30 minutes and then stored at -80°C until preparation for analyses. Tear analyses Variables in collection methods To determine a possible effect of the presence of dye in the testing strip, samples were collected using strips with (Schirmer Tear Test, Merck Animal Health, NJ USA) and without dye (Schirmer tear test strip, Clement Clarke Intl., Ltd.), simultaneously from the same eye without ophthalmic disease. To determine a possible effect of diluent on testing, tear samples obtained using the same brand of STT collected simultaneously from the same eye were compared by one strip being placed in an Eppendorf tube with diluent as described above, and the other strip in a tube without diluent. Protein quantification for each sample was determined using a Bicinchoninic Acid Protein assay (PierceTM BCA Protein Assay Kit, Thermo Fisher Scientific, Rockford IL USA). Briefly, diluted albumin standards and working reagent were prepared according to the manufacturer’s instructions. A calibrated micropipette was used to add 200µL of the working reagent to 25µL of each sample or standard into a microplate well (PierceTM 96-Well plate, Thermo Fisher Scientific, Rockford IL USA). The microplate was placed on a plate shaker for 30 seconds, then covered and incubated for 30 P a g e | 178 minutes at 37°C. After cooling to room temperature, the absorbance at 562nm was determined on a plate reader (PHERAstar® FSX, BMG LABTECH Ortenberg, Germany). Sample preparation Samples were removed from -80°C and allowed to thaw on ice. After ensuring complete immersion of the tear strip in the 200µL of sterile water, the sample tubes were incubated for 30 minutes before centrifuging at 15,000rpm for 30 minutes at 4°C. A micropipette was used to remove 150 µL of sample solution which was then placed into a low protein binding Eppendorf tube. Ice-cold ethanol (1.35ml) was added to the sample and left overnight at - 20°C to precipitate the proteins. The Eppendorf tubes were then centrifuged at 14,000rpm for 30 minutes at 4°C, the supernatant was discarded, and the resulting pellet was air-dried. If the air-dried pellet retained some blue discolouration, the pellet was washed as follows: the sample was agitated using a vortex to dislodge and partially break up the pellet. Ice-cold ethanol (0.75ml) was added to the sample, which was then stored in -80°C for 2 hours. These samples were then also centrifuged at 14,000rpm for 30 minutes at 4°C and the air pellet air dried.

The pellet was solubilised in 20µl of 6M urea and 50mM ammonium bicarbonate (NH4HC03, pH7.5), sonicated in a water bath sonicator for 20 minutes, followed by 1 hour on a platform rocker at room temperature to ensure complete resuspension of the pellet. Sample reduction was achieved by adding 1µl of 200mM dithiothreitol (DTT) (final concentration of 9.5mM; DTT made up in 50mM NH4HC03 pH7.8) which was then incubated for 30 minutes at room temperature. Alkylation was performed with 1µl of 400mM iodoacetamide (IAA) (final concentration of 32mM; IAA made up in 50mM NH4HC03, pH7.8) at room temperature in the dark for 30 minutes. To consume excess IAA, 2µl of 400mM DTT in 50mM NH4HC03 was then added. The sample was then diluted with 50mM NH4HC03 to 1.5M urea by adding 72µl ammonium bicarbonate to the 24µl reduced, alkylated sample. Trypsin digestion at 37°C took place overnight, followed by a further 3 hours of digestion. Formic acid was then added to a final concentration of 1% to stop trypsin digestion and prepare for sample clean-up. Empore™ Solid Phase Extraction 4mm/1ml C18-SD columns (3M Company, Eagan, MN) were used to desalt the samples. Column activation was achieved with 200µl of 90% acetonitrile/0.5% formic acid solution. The column was washed with 500µl of 0.5% formic acid in water, then the sample was loaded in 0.5% formic acid (1ml) to allow the proteins to bind to the column. Unbound non-protein molecules were washed away with 2ml of 0.5% formic acid in water before eluting the bound proteins with 200µl of 70% acetonitrile/0.5% formic acid, followed with 100µl of 90% acetonitrile/0.5% formic acid. The eluted proteins were then vacuum dried. Mass spectrometry Mass spectrometry and data analysis was carried out at the Biomedical Proteomics facility, Children’s Medical Research Institute, Westmead, Australia. Samples were analysed by liquid chromatography tandem mass spectrometry (LC-MS/MS) using a SCIEX tripleTOF 5600 system coupled to an Eksigent NanoLC Ultra 2D Plus HPLC system. The samples, in buffer A (0.1% formic acid) were injected onto a trapping column (PROTECOL C18G, 200 Å, 3 μm, 10 mm × 300 μm) before separation on an analytical column (Acquity UPLC M-class BEHC18 1.7 mm, 300 mm x 150 mm) over a 90 minute multi-step gradient to buffer B (100% acetonitrile, 0.1% P a g e | 179 formic acid) at a flow rate of 5 µl/min for a total run time of 90 minutes. Alternatively, LC-MS instrumentation consisting of a Dionex UltiMate 3000 RSLCnano system (Thermo Scientific) coupled to an LTQ Velos Pro Orbitrap Elite mass spectrometer (Thermo Scientific) was used. The analytical column was made in-house using fused silica tubing (360µm outer and 75µm inner diameter) packed with Dr Maisch Reprosil C18 AQ 1.9µm particles. Peptides were loaded in buffer A (0.1% formic acid) and eluted over a 90 minute multi-step gradient to buffer B (100% acetonitrile, 0.1% formic acid) at 250 nL/min. Mass spectra were acquired with a mass range of 550-2400 m/z, with automatic switch between MS and MS/MS using top 15 method. Data processing and statistical analysis Peptide mass fingerprint spectra were analysed using Proteome Discoverer 2.0™ (Thermo Fisher Scientific, Waltham MA USA) and Proteome Discoverer 2.0. Data were searched against the SwissProt protein sequence database to identify and quantify (label-free) peptides and proteins. Database configurations were as follows: errors at 0.3Da, carbamidomethylation of cysteine residues as static modification, methionine oxidation as dynamic modification, partials at 1, and trypsin enzyme. False discovery rate (FDR) was set at <1% and all shared peptides were excluded from analysis. When using Proteome Discoverer 2.0, MASCOT was the specified search engine, and Percolator was used for FDR calculations. For relative quantitation determination the Events Detector and the Precursor Ion Area Detector options were added to the processing workflow to take the intensities of the three best peptides for each protein in each file. P a g e | 180

Table 8. 1. Individual subject characteristics

ID Group Age Sex Breed Eye Medicationa IOP STT Contralateral eye Quantitative (years) (mmHg) (mm/min) analysisb 1 PACG 10 MN Kelpie L Latanoprost (3) 18 16 End-stage PACG (enucleated) 2 PACG 12 FS Manchester L Nil 38 18 Below (predisposed) Terrier Predisposed R Nil 13 24 Above (PACG) 3 POAG 7 FS Cattle dog L Dorz/Tim (2) 13 20 POAG, IOP 28mmHg Latanoprost (2) 4 POAG 6 FS Beagle R Dorz/Tim 22 22 POAG, IOP 20mmHg Latanoprost 5 Predisposed 6 FS GSD L Dorz/Tim (2) 5 21 End-stage PACG Latanoprost (2) (enucleated) 6 Normal 8 FS Labrador R Nil 15 23 Unremarkable 7 Normal 11 MN Terrier R Nil 7 22 Unremarkable 8 Predisposed 9 FS Labrador X L Dorz/Tim (2) 10 26 Below (PACG) PACG R Dorz/Tim (2) 42 20 Above (predisposed) Latanoprost (2) 9 Predisposed 5 MN Maltese X L Dorz/Tim (2) 10 20 Below (PACG) PACG R Dorz/Tim (3) 54 18 Above (predisposed) Latanoprost (2) 10 Normal 6 FS Poodle L Nil 17 31 Unremarkable 11 Normal 8 FS Poodle R Nil 13 24 Unremarkable 12 Predisposed 12 FS JRT R Dorz/Tim (2) 12 16 Below (PACG) PACG L Dorz/Tim (2) 15 18 Above (predisposed) Latanoprost (2) 13 POAG 5 MN Miniature Fox L Dorz/Tim (3) 11 25 POAG, IOP 60mmHg Terrier Latanoprost (2) Ocuglo, Amlodipine 14 POAG 5 MN Boston Terrier R Dorz/Tim (3) 8 25 POAG, IOP 13mmHg Latanoprost (2) P a g e | 181

15 PACG 13 FS Poodle X R Dorz/Tim (3) 43 18 Below (predisposed) Cocker Predisposed L Dorz/Tim (2) 9 22 Above (PACG) 16 PACG 14 MN Maltese X R Dorz/Tim (2) 9 15 End-stage PACG Latanoprost (2) (enucleated) Pred Forte (2) 17 Normal 8 MN Cocker L Nil 14 26 Unremarkable Spaniel 18 Normal 9 FS Labrador R Nil 11 21 Unremarkable 19 Predisposed 8 MN Cocker L Nil 15 22 Below (PACG) Spaniel PACG R Nil 40 20 Above (predisposed) PACG = primary angle closure glaucoma; POAG = primary open angle glaucoma; normal = healthy control animal; MN = male neutered; FS = female spayed; L = left eye; R = right eye; IOP = intraocular pressure (measured at the time of tear sample collection); STT = Schirmer tear test result (measured at the time of tear sample collection); Dorz/Tim = dorzolomide 2%, timolol 0.05%; Pred Forte = prednisolone acetate 1%, phenylephrine hydrochloride 0.12%; aDaily frequency of topical medication administration listed in parentheses; bTick indicates those subjects where relative quantities of proteins were determined using mass spectrometry. Tear samples from eyes without a tick were included only in analyses to determine whether or not proteins were present in different disease states

P a g e | 182

.

Figure 8. 1. Representation of proteins unique to the tears of dogs with primary glaucoma. All dogs with primary open angle glaucoma (POAG) were being treated with topical ocular hypotensive medications at the time of tear sampling. The eyes of dogs with primary angle closure glaucoma (PACG) either had PACG (affected) that was already being treated (treated) or was newly diagnosed (untreated), while the contralateral (predisposed) eye was on topical ocular hypotensive medication (treated), or yet to be started on prophylactic medication (untreated). P a g e | 183

To identify proteins absent or present in all samples within a group of interest visual inspection was performed of tabulated data for all proteins in each tear sample. The molecular function and biological process for each of these proteins was then searched against the UniProt database. A proportion of tear samples then underwent relative quantitative analyses (n=17). A one-way ANOVA was used to identify proteins that had a differential expression between groups. To investigate possible molecular mechanisms associated with canine glaucoma, pathway analysis was performed on differentially expressed proteins using Ingenuity Pathway Analysis (IPA) software. Established interactions were identified and used by the software to construct protein networks for each group. Statistical analyses were conducted using IBM SPSS Statistics version 24, with graphical representations made using GraphPad Prism 7.0a. Subject characteristics were compared between groups using a two- tailed paired t-test for continuous variables and Fisher’s exact test for categorical data. Results with a p value <0.05 were considered statistically significant. RESULTS Tear samples were obtained from 25 eyes (19 dogs). Thirteen dogs diagnosed with primary glaucoma were included with their eyes allocated to predisposed (n=7), PACG (n=8), and POAG (n=4) groups (Table 8.1). All dogs with POAG were being treated with at least one topical ocular hypotensive medication (carbonic anhydrase inhibitor and/or prostaglandin analogue). Among dogs with PACG, there were four predisposed and four glaucomatous eyes naïve to therapy. Three predisposed eyes and four glaucomatous eyes were on topical medication. When comparing dogs with and without glaucoma, there was no statistically significant difference in age (p=0.621), sex (p=1.000), the number of purebred dogs (p=0.128), whether the left or right eye was affected (p=0.645) or Schirmer tear test results (p=0.920). In this study, dogs with glaucoma had a higher mean IOP (20mmHg) compared to healthy controls (12mmHg) (p=0.010), and were more frequently on a topical ocular hypotensive medication (p=0.001). Protein recovery with and without dye in the tear strip, and diluent in the storage tube In samples obtained simultaneously from the same eye without ophthalmic disease, there was 70.4µg (355µg/ml) protein in the sample obtained using a strip with dye, and 72.7µg (363µg/ml) in the sample obtained using a strip without dye. When samples obtained simultaneously from the same eye in a different dog without ophthalmic disease were placed into one Eppendorf tube with diluent and one without diluent, protein measurements were again similar between samples (with diluent: 66.2µg [331µg/ml]; without diluent: 60.6µg [303µg/ml]). Proteins unique to disease states Component peptides and proteins from 25 eyes (19 dogs) were identified. The proteins identified with most confidence based on the total number of identified peptide spectra matched for the protein (PSMs) included Major allergen Can f 1, Serum albumin, Actin, Aldehyde dehydrogenase, and Minor allergen Can f2 (Table 8.2). The disease groups were compared to identify proteins uniquely present or absent. Fourteen proteins were present in all tear samples from dogs with POAG and absent in all other tear samples (Fig.8.1). There P a g e | 184 were seven proteins identified only in dogs with PACG. Apolipoprotein was present only in tears from medically treated eyes with PACG, and Transforming protein RhoA in tears from untreated eyes with PACG. Nucleoside diphosphate kinase was present in the tears from both eyes of dogs with PACG (both predisposed and PACG-affected eyes) naïve to therapy. Retinol binding protein and NSFL1 cofactor p47 were identified only in tears from PACG-affected eyes being treated with topical medication.

Table 8. 2. Proteins identified with highest confidence by label-free mass spectrometry

Protein PSMs Number of Number Unique of Peptides Peptides Major allergen Can f 1 5696 17 17 Serum albumin 1798 44 66 Actin, cytoplasmic 1 994 17 32 Aldehyde dehydrogenase, dimeric NADP-preferring 780 31 32 Actin, alpha cardiac muscle 1 678 3 18 Minor allergen Can f 2 593 14 14 Haptoglobin 517 22 29 Keratin, type II cytoskeletal 71 515 23 36 Keratin, type II cytoskeletal 1 472 24 32 Alpha-actinin-4 463 45 45 Keratin, type I cytoskeletal 10 462 14 28 Heat shock cognate 71 kDa protein 433 22 31 Lysozyme C, milk isozyme 414 7 7 Glyceraldehyde-3-phosphate dehydrogenase 356 3 16 Clusterin 352 20 20 Heat shock 70 kDa protein 1 346 23 26 Peroxiredoxin-1 327 18 19 14-3-3 protein sigma 321 12 17 Keratin, type I cytoskeletal 25 304 8 23 Trefoil factor 2 298 11 11 PSMs = peptide spectra matched for the protein

P a g e | 185

Figure 8. 2. Illustration of the number of proteins unique to, and common between disease states when comparing (A) dogs with primary angle closure glaucoma (PACG) and healthy controls; and (B) dogs with primary glaucoma and healthy controls. NORM = normal/healthy controls; PD = predisposed to PACG; PACG = primary angle closure glaucoma; POAG = primary open angle glaucoma

Comparative proteomics analysis by label-free mass spectrometry Protein expression levels were explored in 17 eyes (15 dogs) including six healthy controls, four eyes predisposed to PACG, four eyes with PACG, and three eyes with POAG. A total of 2334 unique peptides comprising 505 proteins with distinct peptides were identified using FDR of 1.0%. We identified 390 proteins in healthy control eyes, 318 proteins in eyes predisposed to PACG, 245 protein in eyes with PACG, and 262 proteins in eyes with POAG. The number of proteins shared, and unique to, eyes with and without disease is represented in Figure 8.2. The majority of proteins had an intracellular location, with the relative proportion of proteins located in the nucleus, cytoplasm, plasma membrane, and extracellular space relatively consistent between groups (Fig.8.3). There was a statistically significant difference between groups in the level of expression of 34 proteins. Eyes with PACG had a statistically significant increase in the expression of Annexin A1, S100-A6, S100-A2, Prostaglandin reductase 1 and Calpain-2 catalytic subunit compared to other groups (Table 8.3). The levels of Ubiquitin-40S ribosomal protein S27a were greater in predisposed compared to all other eyes (normal: p<0.001; PACG: p=0.006; POAG: p=0.016), and Coagulation factor V had greater expression in predisposed compared to PACG eyes (p=0.017), but was not expressed differently to other groups with statistical significance (Fig.8.4). There were 15 proteins whose expression was significantly greater in eyes with POAG compared to all other groups (Table 8.4). Higher levels of Alpha-actinin-4 and Heat shock cognate 71kDa protein were identified in tears from POAG compared PACG eyes. Higher levels of Ezrin was found in tears from POAG eyes compared to PACG (p=0.004) and normal eyes (p=0.012), but the difference with levels in predisposed eyes was not statistically significant (p=0.361).

P a g e | 186

Figure 8. 3. Pie charts showing the cellular location of proteins identified in (A) normal eyes; (B) eyes predisposed to PACG; (C) eyes with PACG; and (D) eyes with POAG were similar between disease states. The percentage listed represents the number of proteins in each cellular location as a proportion of all proteins identified in that group.

Figure 8. 4. Relative expression of Ubiquitin-40s ribosomal protein S27a (white circles) and Coagulation factor V (black circles) in the tears of eyes predisposed to primary angle closure glaucoma. The circles represent mean relative expression and the lines indicate 95% confidence interval. P values are listed for all comparisons where there was a statistically significant difference in the amount of protein between groups. P a g e | 187

Table 8. 3. Proteins expressed more abundantly in eyes with primary angle closure glaucoma (PACG) compared to other groups

Protein Mean relative protein levels* P Function Biological process PACG Other groups value Annexin A1 610 Normal 100 <0.001 Phospholipase A2 Adaptive and innate immune Predisposed 100 <0.001 inhibitor response; POAG 140 <0.001 Cytokine-mediated signalling Inflammatory response Chemotaxis Phagocytosis Apoptosis Protein S100- 270 Normal 100 0.001 Effector of glucocorticoid- Axonogenesis A6 Predisposed 110 0.006 mediated responses Positive regulation of POAG 50 0.003 (innate immunity); fibroblast proliferation regulator of the Signal transduction inflammatory process/anti inflammatory action Protein S100- 1760 Normal 100 <0.001 Calcium, protein and Endothelial cell migration A2 Predisposed 40 <0.001 transition metal ion POAG 70 <0.001 binding Prostaglandin 950 Normal 100 <0.001 Key step in inactivation of Leukotriene metabolic process reductase 1 Predisposed 60 <0.001 leukotriene 4 POAG 90 <0.001

Calpain-2 720 Normal 100 <0.001 Calcium ion, protein and Cellular response to amino catalytic Predisposed 80 <0.001 enzyme binding; acid stimulus; positive subunit POAG 130 <0.001 peptidase activity regulation of neuron death; proteolysis; regulation of interleukin-6 production; response to hypoxia *Mean relative protein levels are expressed relative to the amount of that protein identified in the tears of normal dogs (healthy control group); p values indicate the result of comparison with the protein levels in PACG tears P a g e | 188

Figure 8. 5. Network pathway generated by Ingenuity Pathway Analysis for the proteins identified in the tears of dogs with primary angle closure glaucoma. The two major networks identified by Ingenuity Pathway Analysis in this disease group included infectious diseases (grey lines) and inflammatory disease (pink lines). The network pathway demonstrates direct (solid line) and indirect relationships (broken line) between molecules. An arrow between the proteins indicates the relationship between molecules as: activation; causation; expression; localisation; membership; modification; molecular cleavage; phosphorylation; protein-DNA interactions; protein-RNA interactions; regulation of binding; or transcription. A line with no arrowhead indicates the relationship between molecules as: chemical-chemical interactions; chemical-protein interaction; correlation; protein-protein interaction; or RNA-RNA interaction. Increasing brightness of colour indicates increased relative expression of the marker. P a g e | 189

Table 8. 4. Differentially expressed proteins in the tears of eyes with primary open angle glaucoma (POAG)

Protein Mean relative protein levelsa Normal Predisposed PACG POAG Alpha-actinin-4 100 86 23* 61 Fructose-bisphosphate aldolase A 100* 118* 49* 363 Aldehyde dehydrogenase, dimeric NADP- 100* 182* 89* 520 preferring Plastin-3 100* 233* 112* 459 Heat shock cognate 71 kDa protein 100 76 31^ 160 Transgelin-2 100^ 88^ 66^ 205 Fatty acid-binding protein 5 100* 91* 67* 201 Macrophage-capping protein 100* 106* 80* 332 Actin, cytoplasmic 1 100^ 86^ 38* 218 Cofilin-1 100* 122* 75* 352 Ezrin 100^ 136 59* 173 Elongation factor 1-alpha 1 100* 143 58* 273 Heat shock protein beta-1 100* 104^ 53* 213 Gelsolin 100* 105^ 41* 205 Transforming protein RhoA 100* 2* 79* 316 Calpain-1 catalytic subunit 100* 123* 84* 303 Adenylyl cyclase-associated protein 1 100* 88* 47* 286 Ig heavy chain V region GOM 100* 141* 61* 601 aMean relative protein levels are expressed relative to the amount of that protein identified in the tears of normal dogs (healthy control group); *indicates p<0.01 when compared with the protein levels in POAG tears; ^indicates 0.01

Table 8. 5. Highest ranked upstream regulators identified on pathway analysis of each group Rank Normal Predisposed Primary angle closure glaucoma Primary open angle glaucoma 1 Estrogen-related receptor alpha (Esrra) Transforming growth factor beta 1 Transcription factors 1/3/4 (Tcf 1/3/4) Protein-serine O- (TGFβ-1) palmitoleoyltransferase porcupine (POR) 2 Transcription factor Sp1 (SP1) Tumour protein p53 (TP53) Peroxisome proliferator-activated Myoblast determination protein 1 receptor gamma (PPARG) (MYOD1) 3 K-Ras (KRAS) K-Ras (KRAS) Transcription factor Sp3 (SP3) DNA methyltransferase 3 alpha (DNMT3A) 4 Vascular endothelial growth factor A Transcription factors 1/3/4 (Tcf 1/3/4) Transforming growth factor beta 1 Protein kinase C (Pkc) (VEGFA) (TGFβ-1) 5 Transforming growth factor beta 1 Transcription factor Sp1 (SP1) Transcription factor Sp1 (SP1) Rho GTPase-activating protein (TGFβ-1) (RhoGap) 6 Insulin-like growth factor 1 receptor Fibroblast growth factor 19 (FGF19) Leptin (LEP) Calcineurin (Pp2b) (IGF1R) 7 Serine/threonine kinase 11 (STK11) Transcription factor Sp3 (SP3) NRG1; Neuregulin 1 Tensin-1 (TNS1) 8 Tumour protein p53 (TP53) HRas Proto-Oncogene, GTPase (HRAS) CCAAT/enhancer-binding protein alpha Cerebral cavernous malformations 2 (CEBPA) protein (CCM2) 9 Glucokinase (GCK) Receptor tyrosine-protein kinase erbB-2 Peroxisome proliferator-activated Myosin phosphatase Rho-interacting (CD340) (ERBB2) receptor alpha (PPARA) protein (MPRIP) 10 Peroxisome proliferator-activated Tumour necrosis factor (TNF) Fms Related Tyrosine Kinase 1 (FLT1) Rho GTPase activating protein 26 receptor gamma (PPARG) (ARHGAP26 11 Cytochrome P450 Family 19 Subfamily Forkhead box protein O1 (FOX01) Fibroblast growth factor 19 (FGF19) A-kinase anchor protein 11 (AKAP11) A Member 1 (CYP19A1) 12 Rapamycin-insensitive companion of Interleukin 4 (IL4) Tumour necrosis factor (TNF) Rhophilin-2 (RHPN2) mTOR (RICTOR) 13 Forkhead box protein O1 (FOXO1) CCAAT/enhancer-binding protein alpha Tumour protein p53 (TP53) Disheveled-associated activator of (CEBPA) morphogenesis 1 (DAAM1) 14 Fibroblast growth factor 19 (FGF19) Peroxisome proliferator-activated MYC Proto-Oncogene (MYC) Myosin IXB (MYO9B) receptor alpha (PPARA) 15 Leptin (LEP) MYC Proto-Oncogene (MYC) Neural Precursor Cell Expressed, Synaptotagmin binding, cytoplasmic Developmentally Down-Regulated 9 RNA interacting protein (SYNCRIP) (NEDD9) P a g e | 191

Molecular function and pathway analysis of differential proteins Pathway analyses were used to identify functional networks from normal, predisposed, PACG (Fig.8.5) and POAG eyes. The inflammatory response was among the highest ranked disease/disorders in tears from dogs with PACG (predisposed and PACG eyes), but not in normal and POAG eyes, while infectious disease was among the top 5 processes in all dogs with glaucoma (predisposed, PACG and POAG eyes), but not in normal dogs. The highest ranked upstream regulatory proteins for each disease state that was identified on pathway analyses are presented in Table 8.5. The highest ranked canonical pathways according to IPA software for normal eyes were: RhoA signalling; thrombin signalling; integrin signalling; and Actin cytoskeleton signalling. In predisposed eyes, the highest three ranked pathways were: the role of cytokines in mediating communication between immune cells; T- helper cell differentiation; and HMGB1 signaling. In the tears from eyes with PACG the highest ranked pathways were: thrombin signalling; Phospholipase C signalling; Sphingosine-1- phosphate signalling; RhoA signalling; and Semaphorin signalling; and in POAG eyes were: RhoA signalling; Integrin signalling; Actin cytoskeleton signalling; tight junction signalling; and the regulation of actin-based motility by Rho.

DISCUSSION There is a recognised need to improve our understanding of the molecular and cellular mechanisms of the canine glaucomas.468 Although every protein derives from a specific gene, and numerous studies describe genetic factors associated with the disease in dogs,20,49,52,581- 590,603 there remains a need to investigate different proteins associated with the canine glaucomas. In this study we identified 34 proteins with differential expression levels between the tears from eyes with POAG, PACG, those predisposed to the development of PACG, and healthy control eyes. In addition, we identified a unique protein in PACG eyes naïve to therapy (Transforming protein RhoA), in PACG eyes on topical medication (Apolipoprotein), and two proteins unique to eyes predisposed to PACG that were naïve to therapy (Thymosin beta-4, WD repeat-containing protein 1), and on topical medication (Retinol binding protein, NSFL1 cofactor p47). In this study, we collected samples suitable for analysis using commercially available tear strips as used routinely in clinical practice. Among the proteins identified with most confidence in this study, Major Canine Allergen, albumin and keratin type II cytoskeletal were identified with the highest degree of confidence in a previous proteomic analysis of pooled tears from healthy dogs.604 Factors such as collection method (absorbent material versus microcapillary tubes), the type of tears (reflex, open or closed eye), and the use of low protein tubes for storage of samples instead of the standard Eppendorf used in this study, are important considerations.604-608 Tear fluid samples collected using absorbent materials are reported to contain higher proportions of serum and cellular proteins and result in larger volume and reduced protein losses during protein extraction.212 The need for a standard method of tear collection is therefore apparent. Our results demonstrate the value in collection with absorbent tear strips. These are generally well tolerated and routinely performed for reasons separate to sample collection, meaning collection of samples from large populations should be readily achievable.

P a g e | 192

In this study a label-free LC-MS/MS proteomics approach was used to identify differential proteins between diseased and disease-free eyes, and between types of glaucoma. In this discovery experiment the bottom-up approach was used, where complex protein samples are digested into peptides by trypsin and the resulting peptide mixtures are analysed by mass spectrometry. Molecule identification relies on protein databases to identify unique peptides and proteins using m/z peak lists. The database of canine proteins is currently less complete than that for humans thus it is possible that proteins have been missed. Also, as the number of peptides present in the sample is very large, only a subset of the peptides can be analysed in a single run in discovery proteomic experiments, limiting the number of proteins identified. However, we identified a larger number of proteins in our samples compared to an existing description of the tear proteome of healthy dogs.604 In part, this may be due to factors such as differences in collection methods, the type of tears collected as previously discussed, but also due to the type of analysis. Future studies should include different techniques (e.g. ELISA or Western blots) to validate findings.

With reports describing tear protein patterns in medicated609,610 and untreated594 human glaucoma patients, it is well known that ocular hypotensive agents preserved with benzalkonium chloride are strongly associated with topical inflammation.611-615 The tear film is crucial in maintaining a healthy ocular surface, and in people, it has been explored to understand the impact of topical medications on ocular surface disease (OSD),616 and on bleb fibrosis associated with filtration surgeries.617,618 With increasing use of glaucoma drainage devices reported in dogs,349,619-623 the effect of topical medications on surgical outcomes warrants consideration in future investigations. In our study 20 dogs were on topical medication to lower the IOP. For ethical reasons, withholding or discontinuing medication is not a realistic option in clinical investigations such as this. This means the effect of disease versus medication cannot be determined using these data, and obtaining samples from newly diagnosed patients prior to the initiation of treatment is imperative. Despite limitations in validation and clinical application, potential biomarkers of POAG in people have been reported in blood, aqueous humour, the trabecular meshwork, optic nerve and retina associated with oxidative stress, immune response and apoptosis.624,625 Although our findings were not expected to replicate those in people with POAG, we identified numerous proteins in dogs with glaucoma that have been investigated or identified in people with the glaucomas. The cytokines interleukin-4 (IL4), transforming growth factor beta (TGFβ), and tumour necrosis factor (TNF) which have been implicated in OSD in people,626 and we identified them in dogs with PACG. We found heterogeneous nuclear ribonucleoprotein L and Cytosols to be unique to dogs with POAG, and they have been identified in the retinas of people with POAG.625 We also found increased levels of inflammatory proteins Protein S100 and immunoglobulins which have been identified in medically treated people with POAG and pseudoexfoliative glaucoma.609 Several of the proteins that have been evaluated for their role in glaucoma in people were identified in our study, including TNF, TP53,627 Apolipoprotein,628 Cytochrome P450, family 46, subfamily A, polypeptide 1,629 Heat-shock 70kD protein 1A,630 TGFβ,624,631,632 and tyrosine kinase receptor B (TrkB).627,633 In our study, no protein unique to POAG was expressed differentially between people with POAG naïve to therapy and other groups.594 However, we did identify greater amounts of Actin cytoplasmic 1 in the tears of POAG (treated) dogs compared to all other groups, and this

P a g e | 193 protein also differed between people with POAG naïve to therapy and healthy controls.594 Although the specific proteins differed between species, we also identified heat shock proteins and immunoglobulins that were differentially expressed in tears from POAG compared to other eyes, and this has been described in people with POAG naïve to therapy.594 In addition to identifying individual proteins, we performed pathway analyses to investigate the functional and biological processes governed by groups of proteins. The benefit of pathway analyses is the ability to look at networks and potential interactions rather than relying solely on the identification of individual proteins. There is a recognised need for improved understanding of molecular processes involved in the canine glaucomas.468 Glaucoma represents a group of disorders rather than a single disease entity. By identifying factors such as upstream regulatory molecules and functional or molecular group patterns, there is a greater ability to draw on existing knowledge of physiological processes, evaluate the relevance (or potential relevance), and identify areas for more focussed investigations. A major limitation in veterinary clinical research is the ability to include sample sizes that provide appropriately powered studies. When the complex nature of the glaucomas is considered as well as the low specificity of identified proteins for individual disease states, a suitably powered study investigating tear biomarkers should have a large number of subjects, and the small number of cases included in this study is one of the limitations that must be considered when drawing conclusions from our findings. However, we propose this method of tear collection, storage and analysis has potential use in multi-centre studies, facilitating access to both the heterogenous population of affected dogs as well as the technology and consistency required for analyses of comparable results. With this approach, the influence and significance of variables such as topical medication, stage of disease, the use of pooled versus individual samples, the sub-type of glaucoma, as well as other factors that remain as yet unknown, might be identified. For example, serum Ig concentrations in dogs have been reported to be poor indicators of mucosal secretion,634 and in addition to known diurnal variations in tear production,635 large diurnal and day-to-day variations in Ig concentrations in serial tear samples exist.634 It is not known whether such variations occur with other tear proteins, however the fact that these variables may exist to different degrees in different proteins, is important to consider, yet was beyond the scope of this preliminary study. The tear film has been explored as a potential source of biomarkers for ophthalmic591,594,636- 645 and systemic597-599 disease in people and in dogs.602 Not only are biomarkers of glaucoma required to understand causative factors in the disease, but also in the assessment of response to therapies. Potential clinical benefits of improving our understanding of molecular and cellular mechanisms associated with glaucoma include: identifying patients that might be better (or poorer) responders to therapeutic intervention; identifying patients that may benefit from newly emerging cell-based and viral-gene-transfer therapies;646 identifying potential therapeutic targets (for example, potential neuroprotective targets such as TrkB which we identified in our study;633 and identifying disease and initiating therapy at a stage before irreversible damage of retinal ganglion cells. With this range of potential uses of biomarkers, patient follow-up should be used to determine the value of the marker (or set of biomarkers) for its (their) intended use. The search for biomarkers associated with glaucoma is more advanced in the medical compared to the veterinary literature. Although preliminary in nature, and without suggesting these results represent any specific biomarker of the disease, we suggest the identification of

P a g e | 194 proteins unique to eyes with one of the two most common types of canine hereditary glaucoma, as well as the differential expression of 34 proteins between dogs with primary glaucoma, dogs predisposed to the development of PACG, and normal dogs, indicate further exploration of the tear film in dogs with glaucoma has potential to identify biomarkers of the disease.

ACKNOWLEDGEMENTS The authors would like to thank the veterinarians and veterinary hospitals who assisted with sample collection, and the owners of the dogs included in the study. We also acknowledge the funding support from the Canine Research Foundation for which we remain grateful.

P a g e | 195

CHAPTER NINE USE OF A BAERVELDT-350MM2 GLAUCOMA DRAINAGE DEVICE TO MAINTAIN VISION AND CONTROL INTRAOCULAR PRESSURE IN DOGS WITH GLAUCOMA: A RETROSPECTIVE STUDY (2013- 2016)

The following is the re-formatted manuscript published by John Wiley & Sons, Pty Ltd: Graham KL, Donaldson D, Billson FA, Billson FM. Use of a Baerveldt-350mm2 glaucoma drainage device to maintain vision and control intraocular pressure in dogs with glaucoma: a retrospective study (2013-2016). Veterinary Ophthalmology. 2017 Sep 1; 20 (5): 427-434. DOI 10.1111/vop.12443

ABSTRACT Objective: To evaluate the Baerveldt-350mm2 glaucoma drainage device (GDD) in dogs with refractory glaucoma when modifications to address postoperative hypotony (extraluminal ligature; intraluminal stent) and the fibroproliferative response (intraoperative Mitomycin-C; postoperative oral colchicine and prednisolone), are implemented as reported in human ophthalmology. Design: Retrospective case series. Animals: Twenty-eight client-owned dogs (32 eyes) including seven dogs (nine eyes) with primary glaucoma and 21 dogs (23 eyes) with secondary glaucoma. Methods: The medical records of all dogs undergoing placement of a Baerveldt-350mm2 GDD at a veterinary ophthalmology referral service between 2013 and 2016 were reviewed. Signalment, diagnosis, duration and previous treatment of glaucoma, previous intraocular surgery, IOP, visual and surgical outcomes were recorded. Results: IOP was maintained <20mmHg in 24/32 (75.0%) eyes. Fourteen eyes (43.8%) required no adjunctive treatments to maintain this IOP control. Fewer doses of glaucoma medication were required following surgery. Vision was retained in 18/27 (66.7%) eyes with vision at the time of surgery. No eyes that were blind at the time of surgery (n=5) had restoration of functional vision. Complications following surgery included hypotony (26/32; 81.3%), intraocular hypertension (24/32; 75.0%) and fibrin formation within the anterior chamber (20/32; 62.5%). The average follow up after placement of the GDD was 361.1 days (median 395.6 days). Conclusion: Efforts to minimise post-operative hypotony and address the fibroproliferative response following placement of a Baerveldt-350mm2 GDD showed an increased success rate to other reports of this device in dogs and offers an alternative surgical treatment for controlling intraocular pressure in dogs with glaucoma.

P a g e | 196

INTRODUCTION Glaucoma is a neurodegenerative disease characterised by progressive retinal ganglion cell death and optic nerve degeneration and is one of the leading causes of blindness in humans and dogs worldwide.1,13,19,647 Glaucoma in dogs is consistently associated with an elevated IOP with both primary and secondary causes recognised.13 Therapeutic interventions used to manage glaucoma are aimed at reducing IOP to a level which is compatible with comfort and halts progressive vision loss.647 Initial glaucoma therapy uses ocular hypotensive medications which lower IOP either by reducing production or by increasing the rate of aqueous humour outflow from the eye.647 Unfortunately failure to control IOP often occurs over time at which stage surgical techniques are implemented. Surgical management is aimed at either facilitating aqueous humour outflow or decreasing aqueous humour production through ciliary body destruction. Historically, ciliary body cyclodestructive procedures (cyclophotocoagulation or cyclocryotherapy) have been used to manage IOP inadequately controlled with medication in dogs. Limitations associated with ciliary body destruction include the inability to directly assess treatment targets (with transscleral techniques) and the susceptibility of adjacent tissues to damage. Decreasing aqueous humour production can have ramifications for eye health as aqueous humour is required for intraocular nutrition, metabolism and the removal of retinal byproducts.648,649 These techniques are therefore associated with relatively high ocular morbidity (pain, decreased vision, inflammation, hypotony and phthisis bulbi) as well as cataract formation and corneal decompensation which can affect visual outcome.648-650 While many cyclodestructive procedures may result in IOP within the reference range in the short term, long term control is more variable with a normal IOP, retinal and optic nerve function often not retained for >6 months after surgery.8 Reports in dogs following transsceral cyclophotocoagulation (TSCP) with varying periods of follow up show control of IOP in up to 92% of cases,351 however retention of vision in potentially visual eyes was only 22%8 - 50%.351 Based on these results, increasing outflow of aqueous humour should be a more physiologically appropriate treatment of elevated IOP.360 Glaucoma drainage devices (GDDs), or aqueous shunts, allow for aqueous humour to bypass the iridocorneal angle facilitating outflow and are becoming the primary surgical option for management of raised IOP in human glaucoma patients.651 The most commonly used GDDs in human glaucoma surgery include the Ahmed, Baerveldt, Krupin and Molteno devices. These devices differ in their endplate surface area, shape, plate thickness, material and the presence or absence of a valve.652,653 Efforts to control post-operative hypotony have always been considered important to success of a GDD. While valved GDDs are reported to have a lower incidence of early postoperative hypotony compared to non-valved devices, hypotony was still reported indicating that the valves often do not function as anticipated (with a theoretical closing at 8–9 mm Hg).654 Reports in the medical literature describing restrictions to aqueous flow through non-valved devices in the early post-operative period demonstrate improved results and decreased hypotony related complications.654 There has been some success described using valved

P a g e | 197

Ahmed devices in dogs (with263,349 and without356,358 adjunctive TSCP), but there are few reports on the use of non-valved devices.359,360,362 The Baerveldt GDD consists of a non-valved silicone tube and endplate which, when encapsulated, creates a potential space into which aqueous humour drains.652,655 The primary resistance to flow is via passive diffusion through the capsule that forms around the endplate.653 Clearance of aqueous from the periocular tissues is presumed to be primarily via venous capillaries. Use of Baerveldt GDDs in human glaucoma patients has been associated with lower adjunctive medications and rates of additional surgery compared to Ahmed devices.651,655-658 The relatively larger size of the scleral endplate has been shown clinically and experimentally in humans to correlate with better drainage capacity compared to smaller devices.652,653,659 Investigations into the pathogenesis of implant failure in humans have focused on managing the postoperative inflammatory response, minimising the risk and duration of hypotony, as well as managing the fibroproliferative response to help maintain a functional ‘draining bleb’ around the implant.656 Pro-inflammatory factors present in glaucomatous aqueous accelerate fibrosis associated with the scleral endplate.656 Therefore, in addition to controlling post- operative inflammation with anti-inflammatory medication, recommendations to minimise exposure of the developing bleb to glaucomatous aqueous by delaying aqueous flow through the shunt have been proposed.656 Using this information, we revisited the use of non-valved implants in dogs with the aim of minimising hypotony by obstructing aqueous flow through the implant tubing and using oral medications postoperatively to minimise the fibrous response associated with the implant. Placement of the GDD in this study was undertaken in dogs diagnosed with naturally occurring glaucoma inadequately controlled with medical management.

MATERIALS AND METHODS Selection criteria Medical records of all dogs (28 dogs; 32 eyes) treated with surgical placement of a Baerveldt- 350mm2 GDD between September 2013 and February 2016 at the Small Animal Specialist Hospital were reviewed. Nine eyes (seven dogs) with primary glaucoma and 23 eyes (21 dogs) diagnosed with glaucoma following phacoemulsification were included. Diagnosis of primary glaucoma was made if goniodysgenesis, a narrow or closed iridocorneal angle was identified on gonioscopy (in the affected or contralateral eye), and when physical and ophthalmic examinations revealed no evidence of disease that might result in secondary glaucoma. Signalment, eye(s) affected, IOP, previous intraocular surgery, medical and surgical interventions, complications, visual status, IOP control and post-operative medication were recorded.

P a g e | 198

Figure 9. 1. Baerveldt-350mm2 drainage device

Surgical procedure Surgery on each eye was performed by a veterinary ophthalmologist with one surgeon treating 29 eyes and the other surgeon treating three eyes. Dogs were premedicated with methadone (0.1-0.5mg/kg IM; Physeptone, Aspen Pharma Pty Ltd, St Leonards NSW} with or without acepromazine (0.005-0.03mg/kg IM; ACP-2, Ceva Animal Health Pty Ltd, Glenorie NSW). Anaesthesia was induced using propofol (4-6mg/kg IV; Propofol, Sandoz Pty Ltd, Pyrmont NSW) and/or thiopentone (4mg/kg IV; Pentothal, Link Medical Products Pty Ltd, Warriewood NSW) to effect. Each patient was intubated and anaesthesia maintained with inhalational isoflurane and oxygen. Atracurium besylate (0.2mg/kg IV; Hospira Australia Pty Ltd, Mulgrave Victoria) was administered to achieve neuromuscular blockade with additional doses (0.1mg/kg increments) administered to allow appropriate globe positioning as required. A fornix-based conjunctival flap was made in the dorsolateral conjunctiva and the Baerveldt device (Abbott Medical Optics Pty Ltd, Australia) (Fig.9.1) placed beneath the dorsal and lateral rectus muscles (Fig.9.2). A scleral site was prepared and in 23 cases, 0.4mg/ml Mitomycin-C (MMC; Baxter Healthcare, Australia) soaked swabs were applied to the site for 2-4 minutes. The implant was secured to the sclera using 9/0 nylon (Johnson & Johnson Medical Pty Ltd, North Ryde NSW, Australia). An intraluminal suture (partial stent) was placed within the tube using 4/0, 5/0 or 6/0 nylon with one or two extraluminal ligatures tied around the tubing using 6/0 or 8/0 polyglactin 910 (Vicryl; Johnson & Johnson Medical Pty Ltd, Australia). In addition, up to three through-and-through stab incisions were made in the tubing anterior to the ligature (using either 8/0 or 9/0 gauge suture needles) (Fig.9.3). The anterior chamber was entered either via a scleral tunnel using a 21-gauge needle (five cases) or by creation of a scleral flap and a 23-gauge needle (27 cases) (Fig.9.4). Viscoelastic (sodium hyaluronate 10mg/ml [Provisc, Alcon Laboratories, Inc, or Acrivet Biovisc 1.2%, Bausch & Lomb Inc., Warszawa Poland]) and/or air were used to maintain the anterior chamber and facilitate placement of the tube into the anterior chamber (Fig.9.5). The scleral flap and

P a g e | 199 conjunctiva were closed with 9/0 polyglycolic acid suture (Safil; B.Braun Pty Ltd Australia). Triamcinolone 4mg (Kenacort 40mg/ml; Aspen Pharmacare Australia [all cases]) and dexamethasone (0.1mg, Dexafort; Intervet Australia [3 cases]) were injected subconjunctivally. Intracameral injections of tissue plasminogen activator (tPA; Actilyse, Boehringer Ingelheim Pty Ltd, Australia) were used postoperatively to manage fibrinous reactions and intraocular hypertension. Postoperative management consisted of oral amoxicillin/clavulinic acid (15-25mg/kg PO q12 hr; Clavulox, Pfizer Australia Pty Ltd, West Ryde), oral prednisolone (0.25-1mg/kg PO q12 hr; Apex Laboratories Pty Ltd, Somersby NSW) and colchicine (0.02-0.03mg/kg PO q24 hr; Aspen Pharmacare Australia Pty Ltd, St Leonards NSW). Topical medications included prednisolone 1% drops (Prednefrin forte; Allergan Australia Pty Ltd, Gordon), ketorolac trometamol 5mg/ml [Acular; Allergan Australia Pty Ltd, Gordon] and a topical antibiotic preparation (chloramphenicol drop 5mg/ml [Chlorsig; Aspen Pharmacare Pty Ltd, St Leonards] or ointment 10mg/g [Opticin; Troy Laboratories Australia Pty Ltd, Glendenning]). In addition a topical carbonic anhydrase inhibitor (dorzolamide hydrochloride 2%/timolol maleate 0.5% [Cosopt; Merck Sharp & Dohme (Australia) Pty Ltd, Macquarie Park NSW)] or brinzolamide 1% [Azopt; Alcon Laboratories (Australia) Pty Ltd, Frenchs Forest NSW]) and/or a prostaglandin analogue (travaprost 0.004% [Travatan; Alcon Laboratories (Australia) Pty Ltd, Frenchs Forest NSW], latanoprost 50µg/ml [Xalatan; Pfizer Australia Pty Ltd, West Ryde NSW] or bimatoprost 0.03% [Lumigan; Allergan Australia Pty Ltd, Gordon NSW]) were used to help control IOP in the post-operative period. Medications for concurrent diseases were continued as necessary.

Figure 9. 2. The dorsal (a) and lateral (b) rectus muscles are identified and isolated; (c) A MMC-soaked sponge is used to soak the scleral bed at the surgeon’s discretion; (d) The implant is prepared with an intraluminal suture and extraluminal ligature before being placed beneath the rectus muscles; (e) The implant is secured to the sclera.

P a g e | 200

Figure 9. 3. Stab incisions are made Figure 9. 4. Entry into the anterior through the tubing anterior to the chamber using a 23-gauge needle under a extraluminal ligature. scleral flap.

Figure 9. 5. a) Left: Mattress sutures may be used to anchor the tubing to the sclera with caution not to direct the tip of the tubing toward the corneal endothelium; b) Right: Tubing positioned within the anterior chamber without contacting corneal endothelium or iris.

Follow up IOP was measured via rebound tonometry (Icare Tonovet, Icare, Finland) postoperatively at intervals (q1-6hr) for the duration of hospitalisation with frequency determined by the post- operative IOP. Clinical examination by a veterinary ophthalmologist was performed daily until discharge, and then weekly for the first month before decreasing the frequency dependent on progress. Outcomes evaluated Successful control of IOP was defined as IOP <20mmHg. Vision was considered present if there was an intact menace response on clinical exam with the ability to navigate an unfamiliar environment (veterinary clinic) under photopic conditions.

P a g e | 201

RESULTS Clinical findings Breeds represented included two Miniature ; one case in a Basset Hound, Bichon Frise, Cavalier King Charles Spaniel, Cocker Spaniel, Miniature Daschund, Miniature Pinscher, Shar Pei, Tenterfield Terrier, West Highland White Terrier; and 17 mixed breed dogs. There were 19 female and nine male dogs. The age range of dogs at the time of surgery was 4.1- 14.1 years (average 9.1; median 9.9 years). The left eye only was operated on in 9 cases and the right in 15 cases. Four dogs had a Baerveldt implant placed in both eyes within the study period. Seven dogs (nine eyes) had been diagnosed with primary glaucoma while remaining dogs were diagnosed with secondary glaucoma following phacoemulsification. Glaucoma surgery was performed on average 58.5 days (median 21; range 4-264) after onset of glaucoma and 206.6 days (median 46; range 15-2135) after phacoemulsification surgery. Cyclodestructive procedures had been used without success to treat four eyes prior to implantation of the Baerveldt device with one dog receiving two treatments with transscleral cyclophotocoagulation (TSCP) and one dog receiving two treatments with TSCP and one cryosurgery treatment. One eye had undergone surgical placement of an Ex-PRESS glaucoma shunt (Alcon, Inc, Israel) two days prior to placement of the Baerveldt device and IOP remained uncontrolled. The average time until discharge was 8.3 days (median 6; range 1-45 days). The average duration of follow up was 361.1 days (median 395.6; range 8-890 days). Implant and technique modifications An intraluminal suture was placed in all cases. The first 24 eyes had 6/0 nylon placed within the lumen of the tube extending from the anterior chamber to the scleral endplate. After initial results and due to ongoing concerns for hypotony, 5/0 nylon was used in six eyes. The use of 4/0 nylon in two eyes resulted in intraocular hypertension and was no longer used. Extraluminal ligatures placed around the implant tubing were placed in all cases. Initially 6/0 absorbable suture was used, but with difficulties controlling IOP in the first 2-3 weeks postoperatively, 8/0 absorbable suture was used in latter cases. The number of venting holes placed through the tubing anterior to this ligature at the time of surgery varied at the surgeon’s discretion depending on how refractory each eye was to medication, as well as the degree of IOP elevation. MMC was applied intraoperatively in 23 cases. Cases where MMC was not used included the initial cases in this series (n=7), and those where surgery was performed on an emergency basis, precluding waiting the period of time for ordering of the compounded medication (n=2). Postoperative courses of oral colchicine and prednisolone were started in all cases. Colchicine was continued for a minimum of six weeks postoperatively unless adverse side effects were

P a g e | 202 reported (n=6). Prednisolone was tapered according to individual cases, with more rapid tapering and lower doses used in diabetic dogs due to the implications on diabetic control. Surgical outcome Following GDD implantation, IOP was maintained <20mmHg in 24/32 eyes (75.0%) (Table 9.1). Five eyes (15.6%) required an additional surgical procedure to maintain adequate IOP control. Surgeries required to help control IOP included surgical breakdown of iris adhesions (n=4) and phacoemulsification to deepen the anterior chamber to facilitate appropriate tube positioning (n=1).

Table 9. 1. Outcome following Baerveldt-350mm2 implantation

Outcome Definition Eyes affected (%) IOP control IOP <20mmHg 24/32 (75.0%) IOP <20mmHg without medication 14/32 (43.8%) IOP <20mmHg with medication 5/32 (15.6%) IOP <20mmHg with additional surgery 5/32 (15.6%) No control (IOP >20mmHg) 8/32 (25.0%) Visual outcome Vision retained 18/27 (66.7%) Loss of vision 9/27 (33.3%) Return of vision in eyes blind before surgery 0/5 (0%) Complications Hypotony (IOP <5mmHg) 26/32 (81.3%) Intraocular hypertension (IOP >25mmHg) 24/32 (75.0%) Fibrin formation in anterior chamber 20/32 (62.5%) Corneal ulceration 14/32 (43.8%) Corneal degeneration, dystrophy, pigment 12/32 (37.5%) Vision loss 8/32 (25%) Hyphaema 6/30 (18.75%) Cataract formation 4/32 (12.5%) Endophthalmitis 3/32 (9.4%) Tube protrusion through conjunctiva 1/32 (3.1%) Exophthalmos 1/32 (3.1%) Absolute keratoconjunctivitis sicca 1/32 (3.1%) Gastrointestinal signs (vomit, diarrhea) 6/32 (18.8%)

At the time of surgery, 27 eyes had vision (positive menace response) and five eyes were functionally blind (no menace response). Of eyes that had vision at the time of surgery, 18/27 (66.7%) retained functional vision at the time of censor (or at the time of death when death was not related to implant failure). Of the five eyes that were blind prior to surgery, two regained some vision for a short period (up to 48 hours) in the immediate postoperative period but this was not maintained. Eleven out of 32 (34.4%) eyes were enucleated following placement of a Baerveldt device either due to inadequate IOP control (8/32) or endophthalmitis (3/32). Endophthalmitis was suspected when there was rapidly progressive corneal oedema, blepharospasm and hyopyon,

P a g e | 203 and was confirmed on histopathology. This loss of IOP control and/or endophthalmitis occurred on average 82.7 days after surgery (median 65, range 3-242 days). An average of 6.3 (median 6) doses of glaucoma medication were administered on a daily basis prior to glaucoma surgery. At all time points measured through the postoperative period, there were fewer daily doses of glaucoma medication administered in all eyes that were not censored (mean 1.2 doses/day at one month [n=29]; 0.3 doses/day at 3months [n=24]; 0.6 doses/day at 12 months [n=10]). Additional surgery was performed on ten eyes (31.3%) following Baerveldt implantation. Four eyes required surgery to separate iris adhesions to the tube (n=3) and to the anterior lens capsule (n=1). One eye underwent irrigation/aspiration for management of suspected endophthalmitis and one eye for treatment of hyphaema. Other surgeries performed on one eye included placement of a buccal mucosal graft over an extruded tube; phacoemulsification to create a deeper anterior chamber and allow breakdown of tube-iris adhesions; placement of an ExPRESS shunt and adjunctive TSCP; and placement of a corneoconjunctival transposition for treatment of progressive stromal corneal mineralisation. Of these cases, both the eye diagnosed with presumed endophthalmitis and the eye undergoing placement of an additional shunt with adjunctive TSCP were considered surgical failures as IOP was not controlled in either case with the use of a Baerveldt GDD. Complications Complications noted throughout the postoperative period are outlined in Table 9.1. Adverse gastrointestinal side effects were noted in six cases in the immediate postoperative period. Oral colchicine was discontinued when adverse side effects were noted (within seven days of surgery) in these cases.

DISCUSSION This series demonstrates control of IOP and preservation of vision following placement of a Baerveldt-350mm2 glaucoma drainage device in dogs. An important difference between this series and previous reports in dogs using the Baerveldt device was a consistent attempt to control postoperative hypotony and the fibroproliferative response by modifying previous protocols. Postoperative hypotony was minimised with placement of an extraluminal ligature and an intraluminal suture in all cases. Efforts to minimise the fibroproliferative response were made with the use of intraoperative MMC and a prolonged oral course of colchicine and prednisolone. The small number of cases, and the number of variables with which each eye was treated in this retrospective study mean definitive conclusions as to the significance of factors such as the type of glaucoma, use of MMC, the size of the intraluminal suture, adjunctive TSCP and postoperative medications, cannot be determined. A randomised prospective study evaluating surgical techniques would allow evaluation of the significance of these factors.

P a g e | 204

The definition of success following glaucoma surgery varies between reports in both human and veterinary literature making direct comparisons between studies difficult. For this reason in this series a successful surgical outcome was defined as control of IOP (<20mmHg) with the outcome of vision reported as no change in vision status following surgery. Control of IOP (75.0%) and maintenance of functional vision (66.7%) in this series is comparable with other studies using GDDs in dogs where IOP control is reported 22-80% and maintenance of vision in 41- 88.0%.263,349,358-360 Functional vision was not restored in any eyes which were blind at the time of implant placement. Based on these results, and on intensive postoperative management requirements of these patients, it is suggested that eyes that are blind prior to surgery have a grave prognosis for return of vision, and these cases should be considered poor surgical candidates. The authors suggest the improved results reported in this series are potentially a result of modifications to previous protocols. In humans most early complications following placement of a GDD occur as a result of postoperative hypotony and include choroidal effusions and/or haemorrhages, shallow anterior chambers with or without aqueous misdirection or maculopathy.660 Placement of an intraluminal suture and ligation of the tubing of non-valved GDDs are used to manage hypotony in the immediate post-operative period for non-valved implants in humans and therefore a similar approach was used in this series. Oral colchicine and prednisone were used postoperatively in all cases in this series to minimise the fibroproliferative response. The anti-inflammatory effects of prednisone are well established. Colchicine binds to the subunits of fibroblast microtubules, thereby obstructing their assembly.661 Colchicine has been reported to reduce subconjunctival fibrosis by reducing the number of fibroblasts and collagen fibres in the filtering wound.662,663 Molteno et al663 showed this anti-inflammatory combination resulted in blebs with thinner walls and improved results. In addition to post-operative medications, MMC was used in 23 eyes in this series. MMC has been shown to be helpful in promoting bleb formation and duration664 and to decrease bleb capsule thickness in dogs.665-667 Further investigation and modifications to minimise postoperative hypotony are considered essential due to the degree of fibrin production that accompanies hypotony in dogs and given that problems with hypotony were still the most common complication encountered in the current report. Endophthalmitis was documented (confirmed on histopathology) in three cases with one case considered a result of a conjunctival rent. The other two dogs were diabetic and this may have predisposed these cases to infection compared to non-diabetic patients.668 No adverse effects were directly attributable to MMC, however the influence of MMC and other anti-fibrotic agents on conjunctival wound healing should be considered and these agents should be used with caution. In some cases where an elevated IOP was documented despite maximum tolerated medical management, additional doses of medication (typically a prostaglandin analogue) administered on an emergency basis resulted in a reduction in IOP. We suggest it is these cases (rather than those where aqueous centesis is the only intervention that would lower IOP) that may require less intervention with potentially better outcomes, although case numbers are too low to draw definitive observations. A possible reason for this is exposure of

P a g e | 205 the bleb to glaucomatous aqueous at an early stage if the extraluminal ligature was removed early. Compared to the aqueous of normal eyes the aqueous of glaucomatous eyes has an irritant action (lasting for up to 9 weeks) on the episcleral tissues overlying the implants.669 We therefore suggest that selecting cases for surgery while the IOP still responds to glaucoma medication would allow for better restriction of aqueous flow by occluding the implant tube with an absorbable suture, thereby protecting the filtration bleb during its formation. The requirement for aqueous centesis and/or intracameral tPA injections to manage fibrin and IOP, typically resulted in hypotony and further fibrin production and should be avoided if possible. Implant extrusion has been reported with the use of Baerveldt359 and Ahmed358 implants in dogs. In this series there was no extrusion of scleral endplates which were positioned directly over the sclera, beneath Tenons capsule and conjunctiva, and under the extraocular muscles, as opposed to the subconjunctival space as previously reported.360 In the first case in this series there was conjunctival erosion over the implant tube which has been reported in humans.670 A buccal mucosal graft prevented further problems in this case and care was taken with placement of the tube in future cases. A significant limitation of the findings reported here are the issues that arise from the retrospective nature of our study. All dogs undergoing glaucoma surgery were included rather than selecting only dogs with primary glaucoma. The impact of this aspect of case selection on the surgical outcomes reported here cannot be determined. Two dogs underwent placement of a Baerveldt device 15 days after phacoemulsification was performed (on the same day for both dogs). Both dogs had markedly inflamed, painful eyes with corneal oedema and required daily aqueous centesis in an effort to control IOP prior to glaucoma surgery. With no infectious agents identified on culture of aqueous humour, toxic anterior segment syndrome was suspected. Were it possible to manage the intraocular inflammation successfully, control of IOP may have been obtained without glaucoma surgery, however delaying intervention would have resulted in loss of vision. Surgery did not control IOP or maintain vision in either of these eyes. In the authors’ opinion, the ocular morbidity seen with the use of the Baerveldt GDD appears improved through the follow up period to date compared to eyes treated with cyclodestructive procedures. Eyes appeared generally comfortable with most complications in the early post-operative period. Further investigation is indicated to compare the use of the Baerveldt GDD to other reported surgical treatments of glaucoma in dogs, as well as whether there is any difference in surgical outcome dependent on glaucoma type (primary versus secondary). The success of glaucoma filtering surgery is ultimately reliant on the presence of a functional, filtering bleb. Blebs in human glaucoma patients are often visualised and ultrasound is used to assess for the presence of a bleb in cases of implant failure when direct visualisation is not possible. Visualisation and monitoring of a bleb was inadequate in this series. This is most likely due to the posterior location of the implant as a result of the large corneal diameter in dogs in association with the posterior location of the scleral extraocular muscle insertions under which the implant was placed. However there was a bleb noted in several cases.

P a g e | 206

Further investigations are warranted to better understand the process of aqueous drainage when using Baerveldt implants in dogs. CONCLUSION The present series shows control of IOP and maintenance of vision in dogs with glaucoma following surgical placement of a Baerveldt glaucoma drainage device. The authors consider the additional techniques and medications to minimise postoperative hypotony and fibrosis essential in achieving this outcome. The success rate and follow up reported is comparable to existing reports in the veterinary literature describing the surgical treatment of glaucoma in dogs. With the surgical outcomes described, the Baerveldt drainage device is proposed to be a more physiologically appropriate treatment for refractory canine glaucoma compared to cyclodestructive techniques. However further studies to refine the techniques, evaluate the significance of modifications and variables, and minimise early postoperative hypotony, as well as compare longer term follow up with existing techniques are indicated.

P a g e | 207

CHAPTER TEN COMPARISON OF DIODE LASER TRANSSCLERAL CYCLOPHOTOCOAGULATION VERSUS IMPLANTATION OF A 350MM2 BAERVELDT GLAUCOMA DRAINAGE DEVICE FOR THE TREATMENT OF GLAUCOMA IN DOGS (A RETROSPECTIVE STUDY: 2010-2016)

The following is the re-formatted manuscript published by John Wiley & Sons Pty Ltd: Graham KL, Hall EJ, Caraguel C, White A, Billson FA, Billson FM. Comparison of diode laser transscleral cyclophotocoagulation versus implantation of a 350mm2 Baerveldt glaucoma drainage device for the treatment of glaucoma in dogs: a retrospective study (2010-2016). Veterinary Ophthalmology. 2018; 21(5): 487-497. DOI 10.1111/vop.12536

ABSTRACT Objective: To compare outcomes following transscleral cyclophotocoagulation (TSCP) and 350mm2 Baerveldt implantation in the treatment of canine refractory glaucoma. Design: Retrospective case study. Case selection: Client owned dogs undergoing surgical treatment of glaucoma within a veterinary referral hospital. Procedures: Eighty-six glaucoma surgeries were performed on 83 eyes (69 dogs) diagnosed with primary or secondary glaucoma. Medical records were retrieved and baseline data, surgery, medications, intraocular pressures (IOPs), vision and complications were extracted. Results: Fifty-four eyes (44 dogs) were treated with TSCP and placement of an anterior chamber suture shunt; 28 eyes (24 dogs) were implanted with a Baerveldt glaucoma drainage device (GDD); and 4 eyes (4 dogs) underwent GDD implantation after failure of TSCP to manage IOP. Following TSCP, IOP control (<20mmHg) and vision retention occurred in 81.5% and 42.6% respectively for 16.1±1.36 months. Following GDD implantation, 71.4% maintained IOP <20mmHg and 69.6% maintained vision for 11.0±0.94 months. IOP control without loss of vision was more likely following Baerveldt implantation (17/28; 60.7%) than TSCP (19/54; 35.2%) (p=0.027). One eye had functional vision restored following GDD placement. IOP control without adjunctive medications was more likely following Baerveldt implantation (p=0.02). Conclusions: In this study, eyes treated with Baerveldt GDD implantation were more likely to maintain IOP control and retain vision compared to eyes treated with TSCP and placement of an anterior chamber suture shunt. Lack of formal randomisation, inconsistencies in surgical techniques and TSCP protocols, and potential unmeasured confounders must be considered when extrapolating from this retrospective study.

P a g e | 208

INTRODUCTION The surgical management of glaucoma in human patients primarily aims to lower intraocular pressure (IOP), with ongoing developments of new filtering techniques, devices and wound modulation facilitating continued progress in the field. In contrast, limited progress has been made in surgical techniques used for the management of canine glaucoma which primarily rely on the visual status of the eye. Published studies describing progress in glaucoma surgery in veterinary medicine are limited,353 yet cyclodestructive procedures8,263,349,351,671-676 and implantation of glaucoma drainage devices (GDDs) to facilitate aqueous humour outflow263,347,349,354-356,358-360,362,665,671,677 have been described by several authors in the treatment of canine glaucomas. Surgical attempts to divert aqueous flow into the subconjunctival space in dogs have been complicated by the postoperative inflammatory response resulting in premature closure of the surgical site359,678 largely due to fibrin occlusion or fibrosis around the scleral end plate.353 The use of aqueous shunts has therefore not been widespread and cyclodestructive procedures such as transscleral cyclophotocoagulation (TSCP), endoscopic cyclophotocoagulation, and cyclocryotherapy have predominantly been used in the management of canine glaucoma inadequately controlled by medication alone. Despite reports of adequate IOP control in 658 -92%351 of dogs following TSCP, retention of vision in potentially sighted eyes is relatively low (228-50%351). Direct comparisons between studies to determine the efficacy of treatment modalities is difficult due to the varying study designs, including different definitions of success, as well as the spectrum of diseases that make up the glaucomas. Various GDDs have been used experimentally355,359,361 and in case reports involving single dogs.356,357,679 To the authors’ knowledge, of the published reports involving multiple cases treated with a GDD in a clinical setting both with263,349 and without354,358,360,362 adjunctive cyclodestruction, successful control of IOP is reported in 22.2%358 -80.9%354, and maintenance of vision in 30.8%360 -57.9%263. Complications reported with cyclodestructive procedures include cataract formation, corneal decompensation,649 inflammation, hypotony, phthisis bulbi650 and damage to retinal and optic nerve function, each of which might each result in loss of vision despite a normal IOP.8 Recently the use of a 350mm2 Baerveldt GDD was reported in a series of dogs (32 eyes) with glaucoma.677 While the use of this non-valved filtration device is not without complications, an improved surgical outcome was described using modifications reported in equivalent surgeries in human patients.677 On this basis, and similar to what has previously been proposed, diversion of aqueous humour outflow is considered a more physiologically appropriate treatment of elevated IOP.360 To further investigate outcomes achieved with filtration surgery compared to TSCP techniques within this veterinary ophthalmology service, a retrospective inferential case series was conducted. Postoperative outcomes were compared between those eyes treated with implantation of a 350mm2 Baerveldt GDD as previously described677 and eyes treated with diode laser TSCP and placement of an anterior chamber suture shunt in dogs with glaucoma inadequately controlled by medications alone.

P a g e | 209

MATERIALS AND METHODS Patient selection The medical record of every dog that underwent surgical treatment of glaucoma at the Small Animal Specialist Hospital (Sydney, Australia) between May 2010 and February 2016 was retrieved. All eyes that underwent diode laser TSCP (54 eyes; 44 dogs) and/or implantation of a GDD (28 eyes; 24 dogs) during the study period were included. Eyes that had additional de novo glaucoma surgery (4 eyes, 4 dogs) after the initial glaucoma surgery failed to adequately control IOP were included only in the analysis of the first glaucoma surgery performed on that eye. One dog (1 eye) had TSCP and cyclocryotherapy performed at a separate institution before referral for management of refractory glaucoma that has required daily aqueous centesis to manage the IOP. This eye was included only in the ‘combined’ group for analysis. Eyes treated with end-stage salvage procedures (enucleation, evisceration and chemical ablation) as the primary surgical treatment were excluded. Data extracted and analysed included age at the time of surgery, breed, sex, ophthalmic findings, affected eye, history of ocular disease and surgery, glaucoma surgeries, IOP, ocular hypotensive medications used, presence of vision, and complications reported during the postoperative and follow up period at set intervals of 1, 2, 3, 6, 12, 18 and 24 months postoperatively. A complete ophthalmological examination was performed in all cases and included biomicroscopy (Keeler PSL Classic Portable Slit Lamp, Keeler Ltd, Windsor, SL, UK), direct and indirect ophthalmoscopy (Welch Allyn Australia Pty Ltd, Rydalmere, NSW, Australia), tonometry (Icare Tonovet, Icare, Finland), Schirmer tear test (STT; Merck Animal Health, NJ, USA) and fluorescein staining of the cornea (OptiStrips-FL Sterile Fluoroscein Strips, Bausch & Lomb Pty Ltd, Macquarie Park, NSW, Australia). Ocular ultrasound (10MHz probe, Philips HD- 11, Philips Australia) after the application of topical anaesthesia (oxybuprocaine hydrochloride 0.4%; Minims, Bausch & Lomb Australia Pty Ltd, Macquarie Park, NSW, Australia) was performed when intraocular structures were not clearly visible. All eyes were diagnosed with either primary or secondary glaucoma based on a documented elevation in IOP ≥25mmHg with or without signs of current or previous intraocular hypertension (changes to the optic nerve head and/or retina). In all cases, surgical treatment of glaucoma was deemed necessary by a veterinary ophthalmologist to maintain intraocular normotension and preserve comfort and/or vision in the affected eye. Surgical Procedures All dogs were premedicated with methadone (0.1-0.5mg/kg IM; Physeptone, Aspen Pharma Pty Ltd, St Leonards, NSW, Australia) with or without acepromazine (0.005-0.03mg/kg IM; ACP-2, Ceva Animal Health Pty Ltd, Glenorie, NSW, Australia) dependent on age, health and demeanour of the patient. Anaesthesia was induced using propofol (4-6mg/kg IV; Propofol, Sandoz Pty Ltd, Pyrmont, NSW, Australia) and/or thiopentone (4mg/kg IV; Pentothal, Link Medical Products Pty Ltd, Warriewood, NSW, Australia) to effect. Each patient was intubated and anaesthesia maintained with inhalational isoflurane and oxygen. Atracurium besylate (0.2mg/kg IV; Hospira Australia Pty Ltd, Mulgrave, Victoria, Australia) was administered to achieve neuromuscular blockade in all eyes where a GDD was placed, and in TSCP cases at the

P a g e | 210 surgeon’s discretion to allow appropriate globe positioning after placement of surgical 6/0 silk stay sutures (Ethicon, Johnson & Johnson Medical Pty Ltd, Australia). Additional doses (0.1mg/kg increments) were administered as needed. Intermittent positive pressure ventilation using a mechanical ventilator was performed when neuromuscular blockade was achieved. Diode laser TSCP was performed using an 810-nm diode laser (Diovet, Iridex Medical, Mountain View, CA, USA) and a G-probe. The probe was positioned 3-4mm posterior to the limbus with care taken to avoid the three o’clock and nine o’clock positions. Between 13-31 sites (mean 26; range 13-31 sites), were treated in all eyes with power settings of 1000mW for 5000ms and/or 1500mW for 1000-1500ms at the discretion of the surgeon. A 6/0 nylon (Ethilon; Alcon Laboratories Inc., Australia) was then placed into the anterior chamber dorsally and secured under a superficial conjunctival flap to allow some shunting of aqueous humour into the subconjunctival space. Subconjunctival triamcinolone 4mg (Kenacort 40mg/ml; Aspen Pharmacare, St Leonards NSW, Australia) was injected at the end of surgery in 19 eyes. Implantation of the 350mm2 Baerveldt GDD (Abbott Medical Optics Pty Ltd, Pymble, NSW, Australia) was performed as previously described.677 Briefly, a 350mm2 Baerveldt GDD (Abbott Medical Optics Pty, Ltd, Australia) was placed beneath the dorsal and lateral rectus muscles. It was then secured to the sclera under a fornix-based conjunctival flap using 9/0 nylon (Ethilon; Alcon Laboratories Inc., Australia). Mitomycin-C 0.4mg/ml (MMC; Baxter Healthcare, Old Toongabbie, NSW, Australia) was applied to the scleral bed before placement of the implant in 23 eyes. An intraluminal stent (4/0 – 6/0 nylon) was placed within the implant tube. Partial ligation of the tube, anterior to the scleral endplate, using 6/0 or 8/0 polyglactin 910 (Vicryl, Johnson & Johnson Pty Ltd, Australia) was performed to minimise postoperative hypotony. In 23 eyes, a scleral flap was created, beneath which a 23-gauge needle was used to enter the anterior chamber. In 5 eyes a 21-gauge needle was used to form a scleral tunnel into the anterior chamber without a scleral flap. Viscoelastic (sodium hyaluronate 10mg/ml [Provisc, Alcon Laboratories, Inc, or Acrivet Biovisc 1.2%, Bausch & Lomb Inc., Warszawa Poland]) and/or air were used to maintain the anterior chamber and facilitate placement of the tube into the anterior chamber. The scleral and conjunctival flaps were closed using 9/0 polyglycolic acid suture (Safil; B.Braun Pty Ltd Australia) in a simple continuous pattern. Outcomes evaluated IOP control, vision status, the number of ocular hypotensive medications used, and complications reported during the postoperative and follow up period were evaluated at 1, 2, 3, 6, 12, 18 and 24 months after surgery. Adequate IOP control was defined as IOP <20mmHg. The presence of vision was determined by an intact menace response and the ability to navigate an unfamiliar environment under photopic conditions. Surgical failure was defined as IOP >20mmHg, if additional glaucoma surgery was required to maintain IOP <20mmHg, and/or when vision loss occurred. The final outcome for each case was determined at the last follow-up time, time of surgical failure or censor (conclusion of the

P a g e | 211 study period, or death when this occurred during the study period for reasons unrelated to the eye). Twelve eyes (TSCP: 7; GDD: 5) were functionally blind before glaucoma surgery. In these eyes, surgery was performed either if vision loss or ocular hypertension was relatively acute, and the owner wished to pursue all potential avenues to restore vision in that eye; or if the owner declined other surgical (salvage) procedures, and the surgical goal was comfort via ocular normotension. Postoperative management Postoperative management of all cases involved the use of topical prednisolone 1% (Prednefrin forte; Allergan Australia Pty Ltd, Gordon, NSW, Australia) and/or diclofenac sodium 1mg/ml (Voltaren; Alcon Laboratories Australia Pty Ltd, Frenchs Forest, NSW, Australia) or ketorolac trometamol 5mg/ml (Acular; Allergan Australia Pty Ltd, Gordon, NSW, Australia) and a topical antibiotic preparation (chloramphenicol drop 5mg/ml [Chlorsig; Aspen Pharma Pty Ltd, St Leonards, NSW, Australia] or ointment 10mg/g [Opticin; Troy Laboratories Australia Pty Ltd, Glendenning, NSW, Australia]). A topical carbonic anhydrase inhibitor (dorzolamide hydrochloride 2%/timolol maleate 0.5% [Cosopt; Merck Sharp & Dohme Pty Ltd, Macquarie Park, NSW, Australia)] or brinzolamide 1% [Azopt; Alcon Laboratories Pty Ltd, Frenchs Forest, NSW, Australia]) and/or a prostaglandin analogue (travaprost 0.004% [Travatan; Alcon Laboratories Pty Ltd, Frenchs Forest, NSW, Australia], latanoprost 50µg/ml [Xalatan; Pfizer Australia Pty Ltd, West Ryde, NSW, Australia] or bimatoprost 0.03% [Lumigan; Allergan Australia Pty Ltd, Gordon, NSW, Australia]) were used to help control IOP in the post- operative period. Oral medications used in the TSCP group included prednisolone (0.25- 1mg/kg q12-24hr; Apex Laboratories Pty Ltd, Somersby, NSW, Australia) or meloxicam (0.1mg/kg q24hr; Metacam 1.5mg/ml, Boehringer Ingelheim Pty Ltd, North Ryde, NSW, Australia). Dogs undergoing implantation of the Baerveldt device were treated with oral amoxicillin/clavulinic acid (15-25mg/kg twice daily; Clavulox, Pfizer Australia Pty Ltd, West Ryde, NSW, Australia). Prolonged courses of oral prednisolone (0.25-1mg/kg twice daily) and colchicine (0.02-0.03mg/kg daily; Lengout, Aspen Pharmacare Australia Pty Ltd, St Leonards, NSW, Australia) were used for a minimum of 6 weeks (unless premature discontinuation was indicated due to clinical signs of gastrointestinal upset) for their anti inflammatory and anti fibrotic effects respectively. Aqueous centesis, intracameral tissue plasminogen activator (tPA; Actilyse, Boehringer Ingelheim Pty Ltd, Macquarie Park, NSW, Australia), intravenous mannitol bolus (0.5-1.5g/kg), adjunctive surgery and de novo glaucoma surgeries were performed postoperatively in cases where elevations in IOP (>25mmHg) did not respond to maximum tolerated medical management. Additional surgeries performed after TSCP or GDD implantation were considered de novo glaucoma surgeries when performed to reduce an elevated IOP. These surgeries included placement of a GDD; TSCP; combination GDD placement and TSCP; breakdown of synechia when synechia formation had caused, or contributed to an elevated IOP; and phacoemulsification. Surgical procedures performed when the eye was normotensive and the goal was not reduction of IOP were not considered a separate/additional glaucoma surgery.

P a g e | 212

Follow up time was calculated as the time from surgery until failure in cases that lost vision or where IOP control was not achieved and/or maintained. In cases where surgery successfully maintained vision and IOP control, eyes were censored at the conclusion of the study period or at the time of death if this occurred during the study period and was unrelated to the eyes. Follow up was obtained through ongoing examinations at the referral institution except where distance precluded these visits. In these cases follow up was obtained through communication with the treating veterinary ophthalmologist or general practitioner when no specialist was seen. The referring veterinarian and/or the owner was contacted in all cases to confirm the final status of the dog/eye. Data analysis For data analysis the cases were divided into three groups; all eyes treated with TSCP at this hospital (54 eyes) were included in the ‘TSCP’ treatment group. Placement of a suture into the anterior chamber was performed in each TSCP treated case in this series. Due to variations in surgical technique within the TSCP group, the significance of these variables on surgical outcomes and complications between TSCP treated eyes, was determined. Eyes treated with ciliary body cyclodestructive procedures that subsequently required GDD implantation to control IOP were included in the TSCP group until the time of TSCP failure, and were subsequently analysed as part of the ‘combined’ treatment group. Eyes treated with implantation of a Baerveldt GDD as the initial glaucoma surgery (28 eyes) were included in the ‘Baerveldt’ treatment group. Statistical analyses were conducted in GenStat (17th Edition, VSN International). Baseline and postoperative outcomes were compared between treatment groups. A natural logarithm of variables not meeting assumptions for normality was made for analysis. Modeling was undertaken using Restricted Maximum Likelihood (REML) for continuous outcomes, Generalised Linear Mixed Models (GLMM) for binary and poisson distributed outcomes and Ordinal Logistic Regression (OLR) for categorical outcomes. Each model included a fixed effect of Treatment and a random effect of Dog. For all results, a p value < 0.05 was considered statistically significant. RESULTS Eighty-six glaucoma surgeries were performed on 83 eyes (69 dogs) between May 2010 and February 2016. This included 54 eyes (44 dogs) treated with TSCP between May 2010 and May 2014 and 28 eyes (24 dogs) implanted with a 350mm2 Baerveldt device between September 2013 and February 2016. Three eyes (3 dogs) in the TSCP group as well as 1 additional eye (1 dog) that had been treated with TSCP and ciliary body cryoablation at a separate referral practice, had a GDD implanted after cryodestruction failed to maintain IOP <25mmHg. Fifteen dogs had both eyes surgically treated for glaucoma within the study period. Baseline demographic and ocular characteristics are presented in Table 10.1. Treatment groups were similar though dogs implanted with a Baerveldt GDD and those in the combined

P a g e | 213

treatment group received a higher number of daily doses of glaucoma medication prior to surgery compared to the TSCP group (p <0.001).

Table 10. 1. Baseline demographic and ocular characteristics

Overall TSCP Baerveldt Combined P Value for TSCP (86 surgeries; 69 (54 eyes; 44 (28 eyes; 24 TSCP/GDD (4 vs GDD (3 group dogs) dogs) dogs) eyes; 4 dogs) comparison)

Age at surgery* 9.8±0.31 (10.1) 10.4±0.32 (10.6) 9.3±0.33 (10.0) 7.6±0.33 (7.7) 0.064 (0.136) Breed (pedigree) 45/69 (65.2%) 34/44 (77.3%) 11/24 (45.8%) 0/4 0.334 (0.364) Sex (female) 39/69 (56.5%) 24/44 (54.5%) 15/24 (62.5%) 4/4 (100%) 0.356 (0.701) Glaucoma diagnosis - Primary 29/69 (42.0%) 24/44 (54.5%) 5/24 (20.8%) 2/4 (50%) 0.308 (0.263) - Phacoemulsification 38/69 (55.1%) 19/44 (43.2%) 19/24 (79.2%) 2/4 (50%) - Uveitic 2/69 (2.9%) 2/44 (4.5%) 0/24 0/4 - Lens luxation 1/69 (1.4%) 1/44 (2.3%) 0/24 0/4 Highest IOP (mmHg)* 63.7±1.77 (60) 61.6±1.76 (59.5) 67.5±1.83 (66) 67±1.01 (65) 0.102 (0.202) Glaucoma medications* 5.4±0.15 (5) 4.8±0.12 (5) 6.4±0.15 (6) 6±0.15 (6.5) <0.001 (<0.001) (doses/day) Prior lens removal (%) 51/86 (59.3%) 28/54 (51.9%) 21/28 (75%) 2/4 (50%) 0.08 Vision before surgery (%) 74/86 (86.0%) 47/54 (87.0%) 23/28 (82.1%) 4/4 (100%) 0.560 (0.946) Duration of glaucoma 132.7±30.48 (26) 175.4±37.19 (26) 62.4±7.67 (20) 30.8±1.42 (35) 0.257 (0.541) (days)

*Results reported as mean ± standard error (median); IOP = intraocular pressure; mmHg = millimetres of Mercury

The average time until discharge from hospital following surgery was significantly longer in eyes treated with a GDD (p<0.001; Table 10.2). The mean follow-up time was similar between groups (TSCP: 16.1 months (range 10-1644 days); Baerveldt: 11.0 months (range 8-826 days); combined: 19.0 months (range 265-841 days); p=0.066; Table 10.2). Relatively short follow up periods (<3 months) were available in 17 eyes (TSCP 10 eyes; Baerveldt 7 eyes). In 10 of these eyes (TSCP 5 eyes; Baerveldt 5 eyes), surgical failure occurred within this period while the remaining dogs (TSCP 5 eyes; Baerveldt 2 eyes) died for reasons unrelated to the eye.

Table 10. 2. Postoperative vision and IOP outcomes

Overall TSCP Baerveldt Combined P Value♯ Maintain IOP and visual status 40/86 (46.5%) 19/54 (35.2%) 17/28 (60.7%) 4/4 (100%) 0.027 (0.018) Vision retained 39/74 (52.7%) 20/47 (42.6%) 16/23 (69.6%) 4/4 (100%) 0.056 (0.065) IOP controlled 68/86 (79.1%) 44/54 (81.5%) 20/28 (71.4%) 4/4 (100%) 0.06 (0.099) - without medications 27/86 (31.4%) 11/54 (20.4%) 13/28 (46.4%) 3/4 (75%) 0.02 (0.021) - required medication 26/86 (30.2%) 23/54 (42.6%) 3/28 (10.7%) 0/4 0.008 (0.028) - required surgery 15/86 (17.4%) 10/54 (18.5%) 4/28 (14.3%) 1/4 (25%) 0.017 (<0.001) Time to discharge (days)* 6.8±0.69 (5) 5.8±0.57 (4) 7.0±0.42 (6) 17.5±2 (9.5) <0.001 (<0.001) Time until blind (months)* 5.3±1.27 (3.1) 6.2±1.46 (4.4) 3.0±0.43 (2.2) N/A 0.558 Follow up (months)* 14.6±1.3 (13.3 16.1±1.4 (14.4 11±0.94 (10.5) 19.6±1.1 (20 0.066 (0.201) *Reported as mean± standard error (median); ♯for TSCP vs Baerveldt comparison (three group comparison)

P a g e | 214

Outcomes following TSCP Complete records describing the number of sites and laser power used were available in 31 cases (57.4%). Nine eyes were treated with a low power/longer duration protocol (1000mV/5000ms), 11 eyes with a higher power/shorter duration (1500mV/1000ms) protocol, and 11 eyes were treated with a combination of these settings. Eyes in which laser settings were not reported (21) were more likely to lose vision compared to eyes treated with a combination of low and higher power protocols (p=0.02), and were more likely to have IOP controlled postoperatively compared to eyes treated with the high power protocol (p=0.02). Eyes treated with the high power protocol for shorter durations (p=0.03) and those for which laser settings were not recorded (p=0.01) were more likely to lose vision compared to eyes treated with a combination of high and low power protocols. Eyes treated with the combined protocol were more likely to have IOP controlled with additional surgery compared to eyes treated with the low power TSCP protocol (p=0.02). There was otherwise no statistical difference in the frequency of reported complications and outcomes between TSCP protocols in this series. The average time each site was treated was 3758mS (range 1500-5000mS) and the total treatment time for each globe 80,000mS (range 37500-140000mS). In eyes where laser settings were reported, an elevated postoperative IOP was more likely to occur in eyes with a larger number of sites were treated (p=0.042). Vision loss was more likely to occur in eyes with a longer maximum duration of individual laser doses (p=0.035). There was otherwise no statistically significance between the number of sites treated and duration of treatment with surgical outcomes or complications reviewed. Of 54 eyes treated with TSCP, subconjunctival triamcinolone used in 19 eyes (35.2%) at the time of surgery and intracameral tPA injected in the postoperative period in 30 eyes (55.6%). Corneal ulceration was more likely to occur in eyes treated with subconjunctival triamcinolone (p=0.010) and the development of phthisis bulbi more likely in eyes where anterior chamber sutures were placed (p=0.028). There was otherwise no statistically significant association between outcomes or the occurrence of complications with the use of intraoperative triamcinolone, postoperative tPA, or whether laser settings were recorded. Surgical outcome Postoperative outcomes are presented in Table 10.2. The outcome as a proportion of the eyes in each group are depicted in Figure 10.1. Eyes implanted with a Baerveldt device were significantly more likely to maintain both IOP control and pre-operative visual status (17/28, 60.7%) compared to eyes treated with TSCP (19/54, 35.2%) (p=0.027). Eyes treated with combined, sequential procedures were more likely to maintain vision and IOP control than GDD placement or TSCP alone (p=0.018). There was no statistically significant difference in the number of eyes that retained sight, or in which sight was restored through the postoperative period between groups (p=0.07) (TSCP: 20/47, 42.6%; Baerveldt: 16/23, 69.6%). The 4 eyes treated with GDD after TSCP failed, were sighted with IOP <20mmHg at the last follow-up (19.6±1.12 months). There was no significant difference in the time until vision loss between eyes treated with TSCP (mean 188.5; range 2-1244 days) and those

P a g e | 215 undergoing GDD implantation (mean 90.7; range 13-242 days) (p=0.558). In 1 eye that was functionally blind before surgery, after implantation of the Baerveldt device, IOP was controlled allowing discontinuation of glaucoma medications. With restoration of pupil motility and resolution of prostaglandin analogue-induced miosis, function (albeit abnormal) vision was achieved in this eye.

Figure 10. 1. Comparison of postoperative outcomes between treatment groups expressed as a proportion of the dogs in that treatment group. *Outcomes with a statistically significant difference between treatment groups; **Not including de novo glaucoma surgery.

Overall maintenance of IOP <20mmHg was similar between groups (TSCP 81.5%; Baerveldt 71.4%; combined 100%) (p=0.06) (Table 10.2). Eyes treated with both procedures were more likely to maintain IOP control without additional glaucoma medications (75%; p=0.021), while eyes treated with Baerveldt implantation were more likely to maintain IOP control without additional medical intervention than TSCP treated eyes (Baerveldt 46.4%; TSCP 20.4%)(p=0.02). When IOP was inadequately controlled following surgery, topical glaucoma medications were used as first-line therapy. Eyes in the TSCP group (42.6%) were more likely to require use of glaucoma medication compared to those in the Baerveldt group (10.7%) (p=0.008). IOP control (IOP <20mmHg) was similar between groups throughout the postoperative period except at 7 days (p=0.023) and 1 month (p=0.010) postoperatively where Baerveldt treated

P a g e | 216

eyes had a lower IOP (Fig.10.2). All eyes required multiple doses of daily ocular hypotensive medications prior to surgery with dogs undergoing Baerveldt implantation treated with more doses per day (p<0.001). The number of daily medication doses was lower in both groups after surgery, and was similar between groups at different time points postoperatively except at 1 month and 3 months after surgery where eyes treated with Baerveldt implantation required fewer doses of glaucoma medication to maintain intraocular normotension (p=0.019 and p=0.018 respectively) (Table 10.3).

14 80 Postoperative intraocular pressures

70 12

60 10

50

8 ) 40

mmHg 6 (

30 pressures pressures (mmHg)

4 20

2

10

Median intraocular pressure (coloured bars) Median pressure intraocular Mean (cross) and range Mean (cross) (black line) range and intraocular

0 0 1 2 3 6 12 18 24 Time after surgery (months)

Figure 10. 2. Intraocular pressures (median, mean, range) at follow up points during the postoperative period.

Additional surgery Following glaucoma surgery, 11 eyes underwent de novo glaucoma surgery due to inadequate control of IOP despite surgery and postoperative glaucoma medications. This included 9 eyes (16.7%) in the TSCP group of which 6 eyes were treated with additional TSCP only, 2 eyes with a Baerveldt GDD, and 1 eye with both repeat TSCP and a Baerveldt implant. Of these cases, the cases undergoing GDD implantation all retained vision whilst no cases treated with repeat TSCP alone retained vision.

P a g e | 217

Two eyes (7.1%) in the Baerveldt group underwent de novo glaucoma surgery following Baerveldt implantation. One case was treated with TSCP and placement of an Ex-PRESS aqueous shunt. The second case underwent phacoemulsification. Neither eye retained vision. Additional surgery that was not considered de novo glaucoma surgery was performed in 9 eyes and was more likely to be required in eyes in the combined treatment group (25%) compared to either TSCP or GDD treated eyes (p<0.001). TSCP treated eyes were more likely to require additional surgery than those implanted with a GDD (TSCP 18.5%; Baerveldt 14.3%)(p=0.017}) (Table 10.2). Surgeries included irrigation and aspiration of the anterior chamber and breaking down of iris adhesions to the lens capsule or tube in 3 eyes (5.6%) in the TSCP group. All of these eyes were subsequently deemed surgical failures. Seven eyes (25%) in the Baerveldt group underwent additional surgery. Procedures included grafting to treat and overcome partial extrusion of the implant tubing (1), irrigation and aspiration of the anterior chamber (3), and breakdown of iris adhesions (3). Only 1 of these eyes was subsequently deemed a surgical failure.

Table 10. 3. Daily doses of glaucoma medication

TSCP Baerveldt Time post surgery (days) n mean±SE (median) n mean±SE (median) P-value

Pre-op 54 4.8±0.14 (5) 28 6.4±0.26 (6) <0.01

7 days 53 3.4±0.23 (3) 28 3.8±0.60 (3.5) 0.809

1 month 49 2.4±0.23 (2) 26 1.3±0.41 (0) 0.019

2 months 45 1.7±0.24 (2) 24 1.5±0.53 (0) 0.223

3 months 44 1.6±0.26 (1.5) 22 0.8±0.38 (0) 0.018

6 months 42 1.2±0.20 (1) 19 0.7±0.34 (0) 0.191

12 months 35 1.0±0.19 (1.0) 14 0.5±0.24 (0) 0.222

18 months 29 0.8±0.15 (0) 8 0.9±0.35 (1) 0.706 24 months 24 0.7±0.22 (0) 2 0.5±0.49 (0.5) 0.989 n = total number of eyes remaining at that time point; SE = standard error

Complications Complications recorded through the postoperative period are listed in Table 10.4. The most commonly reported complications included intraocular hypertension (TSCP 77.8%; Baerveldt 75%; combined 75%); hypotony (TSCP 53.7%, Baerveldt 75%, combined 100%); cataract formation (TSCP 85.7%, Baerveldt 28.6%, combined 100%); corneal ulceration (TSCP 50%, Baerveldt 39.3%, combined 50%); and fibrinous anterior uveitis (TSCP 29.6%, Baerveldt 60.7%, combined 75%). Anterior chamber collapse was reported in 1 dog in each of the TSCP and Baerveldt groups. Dogs treated with Baerveldt implantation were more likely to develop fibrin within the anterior chamber whether or not they had previous TSCP while phthisis bulbi was more likely following TSCP (p=0.004).

P a g e | 218

Vision loss was seen in 54.7% (27 eyes) of the cases treated with TSCP, 34.8% (8 eyes) in cases treated with GDD. All cases (4/4 eyes) that had both TSCP and GDD implantation maintained sight during the study period, though these differences were not statistically significant (p=0.097).

Table 10. 4. Postoperative complications

TSCP Baerveldt Combined P Value* Cataract formation (%) 12/14 (85.7) 2/7 (28.6%) 2/2 (100%) 0.106 (<0.001) Cataract progression in phakic eyes 2/14 (14.3%) 2/8 (25%) 0/2 N/A (0.858) Corneal ulceration (%) 27/54 (50%) 11/28 (39.3%) 2/4 (50%) 0.396 (0.832) Endophthalmitis 3/54 (5.6%) 3/28 (10.7%) 0/4 0.411 (0.588) Fibrin 16/54 (29.6%) 17/28 (60.7%) 3/4 (75%) 0.014 (0.017) Hyphaema 7/54 (13.0%) 4/28 (14.3%) 2/4 (50%) 0.892 (0.223) Hypotony 29/54 (53.7%) 21/28 (75.0%) 4/4 (100%) 0.059 (0.158) Intraocular hypertension 42/54 (77.8%) 21/28 (75.0%) 3/4 (75%) 0.110 (0.009) Keratoconjunctivitis sicca 9/54 (16.7%) 4/28 (14.3%) 2/4 (50%) 0.12 (<0.001) Loss of vision 27/47 (57.4%) 8/23 (34.8%) 0/4 0.07 (0.097) Phthisis bulbi 7/54 (13.0%) 0/28 (0%) 0/4 0.004 (0.013) Retinal detachment 4/54 (7.4%) 1/28 (3.6%) 0/4 0.558 (0.818) Tube obstruction - 8/28 (28.6%) 1/4 (25%) N/A Tube exposure - 0/28 (0%) 1/4 (25%) N/A

DISCUSSION Refractory glaucoma is a subset of the disease where conventional treatments fail to satisfactorily control IOP.650 In this case series we investigated a clinical suspicion that dogs with glaucoma had improved postoperative outcomes when treated with implantation of a Baerveldt GDD compared to ciliary body destruction using TSCP. Surgical treatment of glaucoma in dogs becomes the only remaining option to maintain comfort and vision in eyes where ocular hypotensive medications fail to maintain satisfactory IOP control yet currently available techniques have limited success. Of critical importance to note when extrapolating findings in this series, is the routine placement of an anterior chamber suture in TSCP treated eyes at this institution. To the authors’ knowledge, the routine use of an anterior chamber suture shunt in dogs undergoing TSCP has not been reported previously. Earlier reports on the use of Baerveldt GDDs in dogs describe limited success359,360 which we suggest can be improved using the techniques and treatment modifications used in human glaucoma surgeries, that we have recently described.677 In this series eyes with IOP controlled and vision retention that required adjunctive therapies, not including de novo glaucoma surgeries, were not considered surgical failures. This definition makes direct comparison of reported results with other studies difficult, but allows direct comparison between these treatment groups. Eyes treated with anti glaucoma medications postoperatively were not necessarily dependent on these medications, but may have been prescribed due to a trending increase in IOP, or maintained due to preoperative difficulties in obtaining IOP control, and

P a g e | 219 thus attributing these eyes as a surgical ‘failure’ was considered without merit. The authors suggest that the improved outcome in eyes undergoing Baerveldt implantation in this study compared to both TSCP treated eyes and previous outcomes using Baerveldt devices359,360 is highly dependent on technical modifications. These modifications included initial limitations to aqueous outflow through the implant with the use of an external ligature around the tubing and an intraluminal suture. Concerted efforts to control the postoperative fibroproliferative response using intraoperative MMC and a prolonged postoperative course of oral colchicine and prednisolone were also considered paramount in obtaining the results described here.677 As the mainstay of canine glaucoma surgery in eyes with potential for vision, cyclodestructive procedures are reported with relatively good IOP control in up to 92% of cases.351 In this series, where an anterior chamber suture shunt was placed in an effort to minimise the impact of elevated IOP in the immediate postoperative period, IOP control (81.5%) was consistent with previous reports of TSCP when used alone. This series also includes a larger number of eyes and more prolonged follow up than previously described. However, it is important to note the limitations of drawing conclusions from this study as there were some statistically significant differences between eyes treated with different TSCP protocols, including those where laser settings were not recorded. This variation in TSCP settings used within this series, limits the ability to draw definitive conclusions in this retrospective study. The authors believe retention of vision is an important goal in management of canine glaucoma for quality of life, the human-animal bond, and owner expectations, and therefore warrants consideration when evaluating therapeutic interventions. Retention of vision in TSCP treated eyes in this series (42.6%) is consistent with previous reports (22%8 - 50%351 of potentially visual eyes) where TSCP treated eyes had no concurrent shunting procedure (anterior chamber suture) performed. This difference between TSCP and Baerveldt treated eyes did not reach statistical significance, though comparison between TSCP treated eyes and those treated with both procedures showed significantly improved vision in the eyes receiving both treatments. While the decreasing IOP trends following surgery were similar in both groups receiving single surgical interventions, the shorter duration of follow up, and smaller case numbers in the Baerveldt group mean assessment of IOP control in the longer term (>12 months) requires further study. There was no significant difference in the rate of vision loss in eyes treated with TSCP versus Baerveldt implantation, though the difference was significant when comparisons between the three treatment groups were analysed (Table 8.4). Retention of vision following TSCP has been reported with slightly greater success than identified in our TSCP group.8,351 This may be due to a combination of factors. The use of a shunting mechanism (via the anterior chamber suture) in all TSCP treated eyes in this series is unique to this study when comparing to previous reports of TSCP treated canine eyes with glaucoma. Direct comparisons between studies are therefore difficult to make, however the technique reported here may play an essential role in reported outcomes. Eyes with both primary and secondary glaucoma were included in our study. It has been reported that the longevity of vision retention is poorer in aphakic eyes treated with TSCP672 possibly due to previous intraocular surgery and/or pre- existing uveitis which would not be present in the primary glaucomas.

P a g e | 220

At this institution, TSCP with placement of an anterior chamber suture shunt was the only glaucoma surgery from May 2010, when TSCP was first available, until September 2013, when the first Baerveldt was placed. Therefore, while all eyes were recruited to the study, and allocated to treatment groups according to the time of presentation, eyes treated with TSCP had a longer period for follow up compared to eyes treated with Baerveldt implantation. While there was no statistically significant difference in the duration of follow up between groups, it is possible that an increase in the number of surgical failures may yet occur in the Baerveldt group, where a larger number of dogs were censored at the cut-off date, compared to eyes treated with TSCP. Analysis of procedural variables in TSCP treated eyes revealed statistical significance associated with IOP control and the number of treatment sites, the laser settings used, and the use of subconjunctival triamcinolone. Due to the retrospective nature of this analysis, whether these factors have clinical significance, and how this might alter comparisons between TSCP and GDD treated eyes, cannot be determined without further investigation in a well designed prospective study. The most frequently reported complication in the TSCP group was cataract formation (85.7%). The development of cataracts in eyes treated with TSCP in this series is considerably higher than previous reports describing cataract formation following TSCP in 68-25%351 of cases. These differences may be due to differences in surgical technique, varying follow up periods, with more extensive follow up in this series, or due to the laser energy levels used. There was no significant difference between TSCP and GDD treated eyes in the type of glaucoma (primary versus secondary; p=0.308) or whether the eye had undergone surgical lens removal (p=0.08), but consideration should be given to the discrepancies in sample sizes and follow up times when drawing conclusions from this finding. A well designed prospective study would be required to explore whether either surgical technique is more likely to induce cataract formation, and therefore potentially less desirable to perform in phakic eyes. Collapse of the anterior chamber is a complication reported in humans following filtering surgeries, typically due to postoperative hypotony.660 In one dog with primary glaucoma treated with GDD implantation, there was extensive anterior chamber collapse resulting in tube-iris and lens contact, and subsequent cataract formation. When tube obstruction resulted in an elevated IOP, phacoemulsification was performed to remove the cataract, deepen the anterior chamber and facilitate appropriate tube positioning. IOP control was subsequently maintained until death due to unrelated neurological disease, though functional vision was lost and the eye considered a surgical failure. The shallower anterior chamber in primary glaucoma cases when compared to pseudophakic dogs means lens damage due to direct contact with implant tubing should be considered and precautions taken. Two cataracts reported in this series were consistent with contact of the tube with the lens. Elevated IOP (≥25mmHg) and hypotony (IOP ≤5mmHg) were common complications observed in cases following TSCP as well as in cases following GDD implantation. Aggressive management of intraocular hypertension included topical medications, anterior chamber paracentesis and/or intracameral tPA injections, and was implemented in both treatment groups though overall, eyes undergoing GDD implantation were managed more aggressively following surgery. The significantly longer period of postoperative hospitalisation in the GDD

P a g e | 221 group arose with more care taken to ensure stable intraocular normotension in eyes before discharge from hospital. While similar strategies were implemented following TSCP, dogs with elevated IOPs were discharged from hospital on glaucoma medication whilst waiting for the effects of ciliary body destruction. In comparison, when this institution started placing GDDs, dogs were hospitalised until IOP appeared stable with either medical or surgical intervention performed as required. These differences in the intensity of management, in addition to the significant finding of increased daily doses of glaucoma medication in the GDD group before surgery, indicate more aggressive management in these eyes which may be an important factor in the surgical success through minimising IOP elevations at levels that would cause irreversible retinal ganglion cell and optic nerve damage. Because IOP control is essential to preserve the visual field in eyes with glaucoma,680 the differences in IOP before surgery, in the perioperative period, and postoperatively might be responsible for the increased incidence of vision loss noted in the TSCP group. An additional consideration in dogs being treated with prostaglandin analogues is the resultant miosis typical of these drugs in dogs. In this series, functional vision was restored to 1 eye that was functionally blind before surgery. This pseudophakic eye developed secondary glaucoma after phacoemulsification and became functionally blind despite IOP control when an axial posterior capsular plaque obstructed all vision through the miotic (prostaglandin- induced) pupil. When this dog presented with an IOP of 72mmHg whilst on medications, the decision was made to proceed with glaucoma surgery in an attempt to discontinue glaucoma medication, and allow the pupil to dilate to see if any vision could be restored. This outcome suggests glaucoma surgery as a therapeutic option to improve vision warrants further investigation for dogs on prostaglandin analogues where miosis can be extreme and itself, impair vision. Mild aqueous flare has been reported following TSCP,8,351,672 and is considered a normal postoperative occurrence in eyes treated with cyclophotoablation.673 To the authors’ knowledge, the relatively high rate of ocular morbidity (pain, decreased vision, inflammation, hypotony and phthisis bulbi) reported in humans following TSCP650 has not been described in dogs. We report a TSCP technique with a concurrent form of aqueous humour shunting. We suggest this may explain the higher incidence of phthisis bulbi in this study, though acknowledge the longer follow up times may also contribute to this finding. Whether the same degree of morbidity does or does not occur in dogs, or whether these factors are under- or not recognised or reported, is difficult to determine. It is possible that ocular complications such as hypotony and phthisis bulbi, may not be painful or adversely affect a dog’s quality of life and so veterinary assessment of these dogs is not sought. These complications also may develop with time, and follow up periods reported in the literature may be inadequate in recognising long-term complications. The authors’ overall clinical impression is that eyes undergoing Baerveldt implantation are significantly less inflamed with improved postoperative recovery and with fewer long-term complications compared to TSCP treated eyes. Postoperative fibrin formation was significantly more likely to occur in eyes treated with a GDD. Rather than being related to postoperative inflammation itself, fibrin is suspected to be related to the sudden onset and degree of hypotony in the immediate postoperative period. This is consistent with the authors’ overall opinion that the hypotony and postoperative complications tended to occur in the early postoperative period in the GDD

P a g e | 222 group compared to the more delayed onset in the TSCP treated eyes. Hypotony causes most early postoperative complications in humans,660 and while concerted efforts were made to minimise hypotony in this series, further work to investigate techniques, devices and modifications to more adequately control hypotony in the immediate postoperative period remain imperative. Limited conclusions regarding surgical techniques can be made from results obtained from the small group of dogs where Baerveldt GDDs were implanted after cyclodestructive procedures failed to control IOP. The excellent results in these 4 dogs (100% vision retention and IOP control), in combination with relatively good success rates reported in a combination procedure involving TSCP and implantation of an Ahmed device349 suggest investigation of combined, whether simultaneous and/or sequential, procedures is warranted. Longer follow-up is required to determine whether complications such as decreased vision, cataracts and endophthalmitis occur at different frequencies between treatment groups. The availability of longer follow-up periods in the TSCP group was due to this being the surgical technique used at this institution prior to introducing management with GDDs. A large proportion of eyes had undergone previous intraocular surgery or had concurrent ocular disease. It is therefore difficult to determine the effect of either TSCP or Baerveldt implantation on the development of postoperative complications including KCS, corneal ulceration, retinal detachment and cataract progression. Consideration must be given in drawing conclusions about the time until blindness and follow up times we report. While these factors did not achieve statistically significant differences between groups, the raw data do suggest both these factors were of greater duration in TSCP treated eyes. The differences in the number of eyes in each group, and the longer period of follow up time available for TSCP treated eyes were important in statistical analysis, but the limitations in this series may highlight discrepancies in the statistical and clinical significance of factors analysed. A limitation of this study is that all canine glaucoma patients were included regardless of glaucoma type, severity of disease or other prognostic factors. While some eyes were therefore likely poor surgical candidates, patients were selected in a similar manner between groups with TSCP being the sole method of surgical treatment of glaucoma at this facility until implantation of Baerveldt devices were undertaken at which point all dogs undergoing primary surgical treatment of glaucoma were treated with placement of a GDD. Due to the retrospective nature of this study, the impact of factors such as the type of glaucoma, surgical techniques and surgeon experience on the surgical outcome could not be appropriately determined. Given the financial commitments, requirements for ongoing medication and monitoring of the eyes, and the importance of optimising the health and welfare of the patient, the possibility/likelihood of retention of vision is an important consideration. The major limitation in comparing treatment outcomes in our study arises due to the retrospective nature of the study. A prospective trial randomly allocating surgical patients to treatment with either TSCP or GDD implantation would provide more conclusive data comparing the outcomes of each technique.

P a g e | 223

CONCLUSION This study shows a better overall surgical outcome in dogs with glaucoma when treated with 350mm2 Baerveldt GDD implantation when techniques to limit early aqueous flow, and manage the postoperative fibroproliferative response are made, compared to diode laser TSCP with a placement of an anterior chamber suture shunt. These findings demonstrate that the use of a Baerveldt device to divert the outflow of aqueous humour can be used successfully as a surgical treatment of glaucoma in dogs, contrary to historical reports on the use of these devices.359,360 While further studies are required to evaluate the use of Baerveldt GDDs and other aqueous shunts, their use with or without concurrent and/or sequential ciliary body destruction for the treatment of canine glaucoma, warrants consideration.

ACKNOWLEDGEMENTS The authors would like to thank the veterinarians whose case material was included in this report; the referring veterinary practices and their staff for assistance with case follow up; Chendong Ma, Jean Yee Hwa Yang and Jennifer Chan for assistance with statistical analysis; and the anonymous reviewers whose input improved the manuscript.

P a g e | 224

CHAPTER ELEVEN CONCLUSION

The dog has been suggested as a model of ophthalmic related diseases in humans based on reports of dogs in both research colonies and those with naturally occurring disease. This project was designed to investigate the potential for dogs with naturally occurring glaucoma to serve as a model of the primary glaucomas in people. We did not aim, nor were the studies designed to determine whether the diseases were the same in each species. Nor was the objective to present the dog as a model of disease on which experimentation can or should be conducted. Instead, our objective was to present a paradigm for the assessment of glaucoma in dogs. By assessing glaucoma in a more complete and standardised manner, there is greater potential for veterinary and medical advances to progress. Funding and support for veterinary clinical research is desperately lacking. At the same time, the number of laboratory trials investigating novel approaches and drugs for the management of the glaucomas in people far exceed the number that ever reach clinical trial phases. In this thesis, our aim was to establish how the dog might be used as a model of glaucoma in humans with the long-term objective of combining veterinary and medical research efforts to benefit people as well as dogs affected by the disease. The One Health initiative is recognised internationally with increasing awareness of the need for collaborative efforts on local and global scales to optimise the health of humans, animals, and the environment. Health and medical research, especially for conditions such as glaucoma, have great potential to benefit from carefully constructed collaborative efforts between veterinarians, physicians, medical researchers, and industry. The development and use of validated surrogate endpoints in clinical trials has the potential to offer shorter, less expensive trials with reduced sample size requirements while allowing observation of a greater number of endpoints during follow-up than could be achieved with observation of a true endpoint.3 Assessing both structure and function in people with glaucoma is significantly better at diagnosing, staging and detecting glaucoma progression compared to isolated structural or functional testing,11,12 and so our project investigated structural and functional markers associated with canine glaucoma. We present preliminary studies describing methods by which functional vision can be measured in veterinary practice without the need for expensive and specialised equipment, and demonstrate the use of existing technologies routinely used in the assessment and management of glaucoma in people, in a canine population. By using materials (Schirmer tear test strip) and biological fluids (tears) which are routinely assessed whenever a complete ophthalmic examination is performed, we introduce the precorneal tear film as a potential source of biomarker(s) of disease. The use of minimally or non-invasive techniques removes ethical dilemmas for veterinary practitioners and the general public to participate in investigational clinical research, thereby increasing the chance of identifying and validating a potential biomarker that may be useful for both veterinary and human populations affected by glaucoma.

P a g e | 225

The use and presentation of animals as models of disease is not uncommon in veterinary and medical literature. As any model of disease must be validated for the specific purpose for which it is proposed to model disease, appropriate validation of animal models of disease is often incomplete, or even absent. Important, and often limiting factors associated with the use of animals in research include the ethics of breeding, using and destroying animals for use in research; the failure to validate the animal as a model of the disease; and the inability to account for factors associated with naturally occurring disease in experimental models. In this project, we present preliminary validation of the methodologies presented, yet a full understanding that ongoing validation of the techniques for clinical use in identifying disease and disease progression is also presented throughout this body of work, but is beyond the scope of this thesis. The next step in this research is to proceed with the conduct of longitudinal studies. With serial measurements of structural, functional and molecular indicators of disease, we hope to determine how the information obtained in individual cross-sectional studies presented in this thesis, is associated with glaucoma, and seek to validate the techniques described as methods to measure progressive disease. If progressive changes are identified in structural parameters in dogs considered predisposed to the development of glaucoma, and if functional vision tests show progressive deterioration (indicating sensitivity to gradual vision impairment), the combination of structural and functional changes provide an objective measure by which progressive disease can be monitored in the dog in a way that is not possible. In addition to confirming the presence of disease before the onset of signs of end- stage disease, this paradigm presents measurable and objective targets for therapeutic trials.

P a g e | 226

REFERENCES

1. Quigley HA, Broman AT. The number of people with glaucoma worldwide in 2010 and 2020. British Journal of Ophthalmology. 2006;90(3):262-267.

2. Medeiros FA. Biomarkers and surrogate endpoints in glaucoma clinical trials. Br J Ophthalmol. 2015;99(5):599-603.

3. Lesko LJ, Atkinson Jr A. Use of biomarkers and surrogate endpoints in drug development and regulatory decision making: criteria, validation, strategies 1. Annual Review of Pharmacology and Toxicology. 2001;41(1):347-366.

4. Martin C, Carmichael KP, Vygantas K, Whitley R. Glaucoma. In: Martin C, ed. Ophthalmic Disease in Veterinary Medicine. London: Manson Publishing; 2010:227-268.

5. Willis AM. Ocular hypotensive drugs. Vet Clin North Am Small Anim Pract. 2004;34(3):755- 776.

6. Lisboa R, Weinreb RN, Medeiros FA. Combining structure and function to evaluate glaucomatous progression: implications for the design of clinical trials. Curr Opin Pharmacol. 2013;13(1):115-122.

7. Kass MA, Heuer DK, Higginbotham EJ, Johnson CA, Keltner JL, Miller JP, Parrish RK, Wilson MR, Gordon MO. The Ocular Hypertension Treatment Study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open- angle glaucoma. Archives of ophthalmology. 2002;120(6):701-713.

8. Cook C, Davidson M, Brinkmann M, Priehs D, Abrams K, Nasisse M. Diode laser transscleral cyclophotocoagulation for the treatment of glaucoma in dogs: results of six and twelve month follow-up. Veterinary and Comparative Ophthalmology. 1997;7(3):148-154.

9. Drance SM. The collaborative normal-tension glaucoma study and some of its lessons. Can J Ophthalmol. 1999;34(1):1-6.

10. Sandberg C, Miller T. Intraocular pressure screening in dogs predisposed to primary angle closure glaucoma. Paper presented at: Annual Meeting of the American College of Veterinary Ophthalmologists2005.

11. Medeiros FA, Lisboa R, Weinreb RN, Girkin CA, Liebmann JM, Zangwill LM. A combined index of structure and function for staging glaucomatous damage. Archives of ophthalmology. 2012;130(9):1107-1116.

P a g e | 227

12. Medeiros FA, Zangwill LM, Anderson DR, Liebmann JM, Girkin CA, Harwerth RS, Fredette MJ, Weinreb RN. Estimating the rate of retinal ganglion cell loss in glaucoma. Am J Ophthalmol. 2012;154(5):814-824 e811.

13. Plummer CE, Regnier A, Gelatt KN. The Canine Glaucomas. In: Gelatt KN, Gilger BC, Kern TJ, eds. Veterinary Ophthalmology. Vol 2. Iowa: John Wiley & Sons, Inc.; 2013:1050-1145.

14. Nickells RW. Retinal ganglion cell death in glaucoma: the how, the why, and the maybe. Journal of glaucoma. 1996;5(5):345-356.

15. Pizzirani S. Definition, Classification, and Pathophysiology of Canine Glaucoma. Vet Clin North Am Small Anim Pract. 2015;45(6):1127-1157, v.

16. Casson RJ, Chidlow G, Wood JP, Crowston JG, Goldberg I. Definition of glaucoma: clinical and experimental concepts. Clin Exp Ophthalmol. 2012;40(4):341-349.

17. Almasieh M, Wilson AM, Morquette B, Cueva Vargas JL, Di Polo A. The molecular basis of retinal ganglion cell death in glaucoma. Prog Retin Eye Res. 2012;31(2):152-181.

18. Wang Q, Grozdanic SD, Harper MM, Hamouche K, Hamouche N, Kecova H, Lazic T, Hernandez-Merino E, Yu C. Detection and characterization of glaucoma-like canine retinal tissues using Raman spectroscopy. Journal of biomedical optics. 2013;18(6):067008-067008.

19. Ofri R, Narfström K. Light at the end of the tunnel? Advances in the understanding and treatment of glaucoma and inherited retinal degeneration. The Veterinary Journal. 2007;174(1):10-22.

20. Oliver JA, Forman OP, Pettitt L, Mellersh CS. Two Independent Mutations in ADAMTS17 Are Associated with Primary Open Angle Glaucoma in the Basset Hound and Basset Fauve de Bretagne Breeds of Dog. PLoS One. 2015;10(10):e0140436.

21. Ofri R, Samuelson DA, Strubbe DT, Dawson WW, Brooks DE, Gelatt KN. Altered retinal recovery and optic nerve fiber loss in primary open-angle glaucoma in the beagle. Exp Eye Res. 1994;58(2):245-248.

22. Gelatt-Nicholson KJ, Gelatt KN, MacKay EO, Brooks DE, Newell SM. Comparative Doppler imaging of the ophthalmic vasculature in normal Beagles and Beagles with inherited primary open-angle glaucoma. Vet Ophthalmol. 1999;2(2):97-105.

23. Brooks DE, Garcia GA, Dreyer EB, Zurakowski D, Franco-Bourland RE. Vitreous body glutamate concentration in dogs with glaucoma. Am J Vet Res. 1997;58(8):864-867.

P a g e | 228

24. Miller P, Poulsen G, Nork T, Galbreath E, Dubielzig R. Photoreceptor cell death by apoptosis in spontaneous acute glaucoma in dogs. Paper presented at: Invetigative Ophthalmology and Visual Science1997.

25. Lee BL, Bathija R, Weinreb RN. The definition of normal-tension glaucoma. J Glaucoma. 1998;7(6):366-371.

26. Gelatt KN. Canine Glaucomas. In: Gelatt KN, ed. Veterinary Ophthalmology. 2nd ed. Philadelphia: Lea & Febiger; 1991:396-428.

27. Ritch R, Shields MB, Krupin T. The glaucomas. St Louis: Mosby; 1996.

28. Brooks DE, Komaromy AM, Kallberg ME. Comparative optic nerve physiology: implications for glaucoma, neuroprotection, and neuroregeneration. Vet Ophthalmol. 1999;2(1):13-25.

29. Strom AR, Hassig M, Iburg TM, Spiess BM. Epidemiology of canine glaucoma presented to University of Zurich from 1995 to 2009. Part 1: Congenital and primary glaucoma (4 and 123 cases). Vet Ophthalmol. 2011;14(2):121-126.

30. Ko F, Papadopoulos M, Khaw PT. Primary congenital glaucoma. Prog Brain Res. 2015;221:177-189.

31. Mackay EO, Kallberg ME, Gelatt KN. Aqueous humor myocilin protein levels in normal, genetic carriers, and glaucoma Beagles. Vet Ophthalmol. 2008;11(3):177-185.

32. Gelatt KN, MacKay EO. Prevalence of the breed-related glaucomas in pure-bred dogs in North America. Vet Ophthalmol. 2004;7(2):97-111.

33. Corcoran K, Koch S, Peiffer R. Primary glaucoma in the chow chow. Ophthalmic Literature. 1995;1(48):62.

34. Smith R, Peiffer R, Wilcock B. Some aspects of the pathology of canine glaucoma. Veterinary Clinical Pathology. 1993;22:16-16.

35. Ekesten B, Torrang I. Heritability of the depth of the opening of the ciliary cleft in Samoyeds. Am J Vet Res. 1995;56(9):1138-1143.

36. Boillot T, Boderiou Y, Maller D, Rosolen SG. Bilateral glaucoma affecting a population of adult Eurasier dogs in France: a possible inherited disease. Paper presented at: European Society of Veterinary Ophthalmology2012.

37. Bedford PG. Open-angle glaucoma in the Petit Basset Griffon Vendeen. Vet Ophthalmol. 2017;20(2):98-102.

P a g e | 229

38. Deehr AJ, Dubielzig RR. A histopathological study of iridociliary cysts and glaucoma in Golden Retrievers. Vet Ophthalmol. 1998;1(2-3):153-158.

39. Komaromy AM, Petersen-Jones SM. Genetics of Canine Primary Glaucomas. Vet Clin North Am Small Anim Pract. 2015;45(6):1159-1182, v.

40. Gelatt KN, MacKay EO. Secondary glaucomas in the dog in North America. Vet Ophthalmol. 2004;7(4):245-259.

41. Coleman AL. Glaucoma. Lancet. 1999;354(9192):1803-1810.

42. He M, Foster PJ, Johnson GJ, Khaw PT. Angle-closure glaucoma in East Asian and European people. Different diseases? Eye (Lond). 2006;20(1):3-12.

43. Gelatt KN, Peiffer RL, Jr., Gwin RM, Gum GG, Williams LW. Clinical manifestations of inherited glaucoma in the beagle. Invest Ophthalmol Vis Sci. 1977;16(12):1135-1142.

44. Ekesten B, Bjerkas E, Kongsengen K, Narfstrom K. Primary glaucoma in the Norwegian Elkhound. Veterinary and comparative ophthalmology. 1997;7:14-18.

45. Gelatt K. Familial glaucoma in the beagle dog. Journal of the American Animal Hospital Association. 1972.

46. Gelatt K, Peiffer Jr R, Gwin R, Sauk Jr J. Glaucoma in the beagle. Transactions Section on Ophthalmology American Academy of Ophthalmology and Otolaryngology. 1975;81(4 Pt 1):OP636-644.

47. Oshima Y, Bjerkas E, Peiffer RL, Jr. Ocular histopathologic observations in Norwegian Elkhounds with primary open-angle, closed-cleft glaucoma. Vet Ophthalmol. 2004;7(3):185- 188.

48. Ahonen SJ, Kaukonen M, Nussdorfer FD, Harman CD, Komaromy AM, Lohi H. A novel missense mutation in ADAMTS10 in Norwegian Elkhound primary glaucoma. PLoS One. 2014;9(11):e111941.

49. Forman OP, Pettitt L, Komáromy AM, Bedford P, Mellersh C. A Novel Genome-Wide Association Study Approach Using Genotyping by Exome Sequencing Leads to the Identification of a Primary Open Angle Glaucoma Associated Inversion Disrupting ADAMTS17. PLoS One. 2015;10(12):e0143546.

50. Kuchtey J, Olson LM, Rinkoski T, MacKay EO, Iverson T, Gelatt KN, Haines JL, Kuchtey RW. Mapping of the disease locus and identification of ADAMTS10 as a candidate gene in a canine model of primary open angle glaucoma. PLoS Genet. 2011;7(2):e1001306.

P a g e | 230

51. Kato K, Sasaki N, Gelatt K, MacKay E, Shastry B. Autosomal recessive primary open angle glaucoma (POAG) in beagles is not associated with mutations in the myocilin (MYOC) gene. Graefe's Archive for Clinical and Experimental Ophthalmology. 2009;247(10):1435-1436.

52. Kuchtey J, Kunkel J, Esson D, Sapienza JS, Ward DA, Plummer CE, Gelatt KN, Kuchtey RW. Screening ADAMTS10 in dog populations supports Gly661Arg as the glaucoma-causing variant in beagles. Investigative Ophthalmology & Visual Science. 2013;54(3):1881-1886.

53. Miller PE, Bentley E. Clinical Signs and Diagnosis of the Canine Primary Glaucomas. Vet Clin North Am Small Anim Pract. 2015;45(6):1183-1212, vi.

54. Cottrell BD, Barnett K. Primary glaucoma in the Welsh springer spaniel. Journal of Small Animal Practice. 1988;29(3):185-199.

55. Slater MR, Erb HN. Effects of risk factors and prophylactic treatment on primary glaucoma in the dog. J Am Vet Med Assoc. 1986;188(9):1028-1030.

56. Vajaranant TS, Nayak S, Wilensky JT, Joslin CE. Gender and glaucoma: what we know and what we need to know. Current opinion in ophthalmology. 2010;21(2):91.

57. Tsai S, Bentley E, Miller PE, Gomes FE, Vangyi C, Wiese A, Almazan A, Li H, Conforti P, Lee SS, Robinson MR. Gender differences in iridocorneal angle morphology: a potential explanation for the female predisposition to primary angle closure glaucoma in dogs. Veterinary Ophthalmology. 2012;15 Suppl 1(s1):60-63.

58. Boillot T, Rosolen SG, Dulaurent T, Goulle F, Thomas P, Isard PF, Azoulay T, Lafarge-Beurlet S, Woods M, Lavillegrand S, Ivkovic I, Neveux N, Sahel JA, Picaud S, Froger N. Determination of morphological, biometric and biochemical susceptibilities in healthy Eurasier dogs with suspected inherited glaucoma. PLoS One. 2014;9(11):e111873.

59. Magrane WG. Canine glaucoma. II. Primary classification. J Am Vet Med Assoc. 1957;131(8):372-374.

60. Lovekin LG. Primary Glaucoma in Dogs. J Am Vet Med Assoc. 1964;145:1081-1091.

61. Lovekin LG, Belhorn R. Clinicopathologic changes in primary glaucoma in Cocker Spaniel. American Journal of Veterinary Research. 1968;29(2):379-&.

62. Martin CL, Wyman M. Glaucoma in the Basset Hound. J Am Vet Med Assoc. 1968;153(10):1320-1327.

P a g e | 231

63. Wyman M, Ketring K. Congenital glaucoma in the basset hound: a biologic model. Transactions Section on Ophthalmology American Academy of Ophthalmology and Otolaryngology. 1975;81(4 Pt 1):645-652.

64. Boeve MH, Stades FC. [Glaucoma in dogs and cats. Review and retrospective evaluation of 421 patients. I. Pathobiological background, classification and breed predisposition]. Tijdschr Diergeneeskd. 1985;110(6):219-227.

65. Bedford P. The aetiology of primary glaucoma in the dog. Journal of Small Animal Practice. 1975;16(1‐12):217-239.

66. Bedford PG. The aetiology of canine glaucoma. Veterinary Record. 1980;107(4):76-82.

67. Bedford PG. A gonioscopic study of the iridocorneal angle in the English and American breeds of Cocker Spaniel and the Basset Hound. J Small Anim Pract. 1977;18(10):631-642.

68. Reilly CM, Morris R, Dubielzig RR. Canine goniodysgenesis‐related glaucoma: a morphologic review of 100 cases looking at inflammation and pigment dispersion. Veterinary Ophthalmology. 2005;8(4):253-258.

69. Kato K, Sasaki N, Matsunaga S, Nishimura R, Ogawa H. Incidence of canine glaucoma with goniodysplasia in Japan : a retrospective study. J Vet Med Sci. 2006;68(8):853-858.

70. Kato K, Sasaki N, Matsunaga S, Nishimura R, Ogawa H. Cloning of canine myocilin cDNA and molecular analysis of the myocilin gene in Shiba Inu dogs. Vet Ophthalmol. 2007;10 Suppl 1(s1):53-62.

71. Kanemaki N, Tchedre KT, Imayasu M, Kawarai S, Sakaguchi M, Yoshino A, Itoh N, Meguro A, Mizuki N. Dogs and humans share a common susceptibility gene SRBD1 for glaucoma risk. PLoS One. 2013;8(9):e74372.

72. Kato K, Sasaki N, Matsunaga S, Mochizuki M, Nishimura R, Ogawa H. Possible association of glaucoma with pectinate ligament dysplasia and narrowing of the iridocorneal angle in Shiba Inu dogs in Japan. Vet Ophthalmol. 2006;9(2):71-75.

73. Whiteman AL, Klauss G, Miller PE, Dubielzig RR. Morphologic features of degeneration and cell death in the neurosensory retina in dogs with primary angle-closure glaucoma. Am J Vet Res. 2002;63(2):257-261.

74. Alyahya K, Chen CT, Mangan BG, Gionfriddo JR, Legare ME, Dubielzig RR, Madl JE. Microvessel loss, vascular damage and glutamate redistribution in the retinas of dogs with primary glaucoma. Vet Ophthalmol. 2007;10 Suppl 1(s1):70-77.

P a g e | 232

75. Mangan BG, Al‐Yahya K, Chen CT, Gionfriddo JR, Powell CC, Dubielzig RR, Ehrhart EJ, Madl JE. Retinal pigment epithelial damage, breakdown of the blood–retinal barrier, and retinal inflammation in dogs with primary glaucoma. Veterinary Ophthalmology. 2007;10(s1):117- 124.

76. Ahram DF, Cook AC, Kecova H, Grozdanic SD, Kuehn MH. Identification of genetic loci associated with primary angle-closure glaucoma in the basset hound. Mol Vis. 2014;20:497- 510.

77. Ahram DF, Grozdanic SD, Kecova H, Henkes A, Collin RW, Kuehn MH. Variants in Nebulin (NEB) Are Linked to the Development of Familial Primary Angle Closure Glaucoma in Basset Hounds. PLoS One. 2015;10(5):e0126660.

78. Grozdanic SD, Kecova H, Harper MM, Nilaweera W, Kuehn MH, Kardon RH. Functional and structural changes in a canine model of hereditary primary angle-closure glaucoma. Investigative Ophthalmology & Visual Science. 2010;51(1):255-263.

79. Bjerkas E, Ekesten B, Farstad W. Pectinate ligament dysplasia and narrowing of the iridocorneal angle associated with glaucoma in the English Springer Spaniel. Veterinary Ophthalmology. 2002;5(1):49-54.

80. van der Linde-Sipman JS. Dysplasia of the pectinate ligament and primary glaucoma in the Bouvier des Flandres dog. Vet Pathol. 1987;24(3):201-206.

81. Barnett K, Mason IK. Primary Glaucoma in the Great Dane. In. Vol 112. 24th Transactions of the American College of Veterinary Ophthalmologists (Abstract)1993.

82. Wood JL, Lakhani KH, Mason IK, Barnett KC. Relationship of the degree of goniodysgenesis and other ocular measurements to glaucoma in Great Danes. Amerincan Journal of Veterinary Research. 2001;62(9):1493-1499.

83. Read RA, Wood JL, Lakhani KH. Pectinate ligament dysplasia (PLD) and glaucoma in Flat Coated Retrievers. I. Objectives, technique and results of a PLD survey. Veterinary Ophthalmology. 1998;1(2-3):85-90.

84. Wood JL, Lakhani KH, Read RA. Pectinate ligament dysplasia and glaucoma in Flat Coated Retrievers. II. Assessment of prevalence and heritability. Veterinary Ophthalmology. 1998;1(2-3):91-99.

85. Pearl R, Gould D, Spiess B. Progression of pectinate ligament dysplasia over time in two populations of Flat-Coated Retrievers. Vet Ophthalmol. 2015;18(1):6-12.

86. Ekesten B. Correlation of intraocular distances to the iridocorneal angle in Samoyeds with special reference to angle-closure. Prog Vet Comp Ophthalmol. 1992;2:67-73.

P a g e | 233

87. Ekesten B, Narfstrom K. Age-related changes in intraocular pressure and iridocorneal angle in Samoyeds. Progress in Veterinary and Comparative Ophthalmology. 1992;2:37-40.

88. Ekesten B, Narfstrom K. Correlation of morphologic features of the iridocorneal angle to intraocular pressure in Samoyeds. American Journal of Veterinary Research. 1991;52(11):1875-1878.

89. Ahonen SJ, Pietila E, Mellersh CS, Tiira K, Hansen L, Johnson GS, Lohi H. Genome-wide association study identifies a novel canine glaucoma locus. PLoS One. 2013;8(8):e70903.

90. Spiess, Bolliger, Borer G, Murisier, Richter, Pot, Walser R, Watte, Hassig. [Prevalence of pectinate ligament dysplasia in golden retrievers in Switzerland]. Schweiz Arch Tierheilkd. 2014;156(6):279-284.

91. Maggio F. Glaucomas. Top Companion Anim Med. 2015;30(3):86-96.

92. Pumphrey S. Canine Secondary Glaucomas. Vet Clin North Am Small Anim Pract. 2015;45(6):1335-1364, vii.

93. Johnsen DA, Maggs DJ, Kass PH. Evaluation of risk factors for development of secondary glaucoma in dogs: 156 cases (1999–2004). Journal of the American Veterinary Medical Association. 2006;229(8):1270-1274.

94. Petersen‐Jones S. Abnormal ocular pigment deposition associated with glaucoma in the . Journal of Small Animal Practice. 1991;32(1):19-22.

95. Covitz D, Barthold S, Diters R, Riis R. Pigmentary glaucoma in the cairn terrier. Transactions of the Fifteenth Annual Scientific Program of the American College of Veterinary Ophthalmologists. 1984:246-250.

96. De Sandt V, Roswitha R, Boevé MH, Stades FC, Kik MJ. Abnormal ocular pigment deposition and glaucoma in the dog. Veterinary ophthalmology. 2003;6(4):273-278.

97. Farrar SM, Shields MB. Current concepts in pigmentary glaucoma. Surv Ophthalmol. 1993;37(4):233-252.

98. Peterson-Jones S, Mentzer A, Dubielzig R, Render J, Steficek B, Kiupel M. Ocular melanosis in the Cairn Terrier: histopathological description of the condition, and immunohistological and ultrastructural characterization of the characteristic pigment-laden cells. Veterinary Ophthalmology. 2008;11(4):260-268.

P a g e | 234

99. Gould D, Pettitt L, McLaughlin B, Holmes N, Forman O, Thomas A, Ahonen S, Lohi H, O’Leary C, Sargan D. ADAMTS17 mutation associated with primary lens luxation is widespread among breeds. Veterinary ophthalmology. 2011;14(6):378-384.

100. Davidson MG, Nelms SR. Diseases of the Lens and Cataract Formation. In: Gelatt KN, Gilger BC, Kern TJ, eds. Veterinary Ophthalmology. Vol II. Iowa: John Wiley & Sons, Inc; 2013:1199- 1233.

101. Curtis R, Barnett KC, Lewis SJ. Clinical and pathological observations concerning the aetiology of primary lens luxation in the dog. Vet Rec. 1983;112(11):238-246.

102. Willis MB, Curtis R, Barnett KC, Tempest WM. Genetic aspects of lens luxation in the Tibetan terrier. Vet Rec. 1979;104(18):409-412.

103. Formston C. Observations on subluxation and luxation of the crystalline lens in the dog. Journal of Comparative Pathology and Therapeutics. 1945;55:168IN167-184IN168.

104. Glover TL, Davidson MG, Nasisse MP, Olivero DK. The intracapsular extraction of displaced lenses in dogs: a retrospective study of 57 cases (1984-1990). Journal of the American Animal Hospital Association. 1994;31(1):77-81.

105. Diters RW, Dubielzig RR, Aguirre GD, Acland GM. Primary ocular melanoma in dogs. Vet Pathol. 1983;20(4):379-395.

106. Wilcock BP, Peiffer RL, Jr. Morphology and behavior of primary ocular melanomas in 91 dogs. Vet Pathol. 1986;23(4):418-424.

107. Bussanich N, Dolman P, Rootman J, Dolman C. Canine uveal melanomas: series and literature review. The Journal of the American Animal Hospital Association (USA). 1987.

108. Cello RM, Hutcherson B. Ocular changes in malignant lymphoma of dogs. Cornell Vet. 1962;52:492-523.

109. Hendrix D, Gelatt K, Smith P, Brooks D, Whittaker C, Chmielewski N. Ophthalmic disease as the presenting complaint in five dogs with multiple myeloma. Journal of the American Animal Hospital Association. 1998;34(2):121-128.

110. Corcoran KA, Koch SA. Uveal cysts in dogs: 28 cases (1989-1991). J Am Vet Med Assoc. 1993;203(4):545-546.

111. Thomas R, Mulligan N, Billson FA. Angle closure glaucoma due to iris and ciliary body cysts. Australian and New Zealand journal of ophthalmology. 1989;17(3):317-319.

P a g e | 235

112. Tanihara H, Akita J, Honjo M, Honda Y. Angle closure caused by multiple, bilateral iridociliary cysts. Acta Ophthalmol Scand. 1997;75(2):216-217.

113. Vela A, Rieser JC, Campbell DG. The heredity and treatment of angle-closure glaucoma secondary to iris and ciliary body cysts. Ophthalmology. 1984;91(4):332-337.

114. Sapienza J, Simo F, Prades‐Sapienza A. Golden Retriever uveitis: 75 cases (1994–1999). Veterinary Ophthalmology. 2000;3(4):241-246.

115. Spiess BM, Bolliger JO, Guscetti F, Haessig M, Lackner PA, Ruehli MB. Multiple ciliary body cysts and secondary glaucoma in the Great Dane: a report of nine cases. Vet Ophthalmol. 1998;1(1):41-45.

116. Pumphrey SA, Pizzirani S, Pirie CG, Needle DB. Glaucoma associated with uveal cysts and goniodysgenesis in American Bulldogs: a case series. Vet Ophthalmol. 2013;16(5):377-385.

117. Rubin L, Gelatt K. Spontaneous resorption of the cataractous lens in dogs. Journal of the American Veterinary Medical Association. 1968;152(2):139-153.

118. van der Woerdt A, Nasisse MP, Davidson MG. Lens-induced uveitis in dogs: 151 cases (1985- 1990). J Am Vet Med Assoc. 1992;201(6):921-926.

119. Van Der Woerdt A. Lens-induced uveitis. Vet Ophthalmol. 2000;3(4):227-234.

120. Moore DL, McLellan GJ, Dubielzig RR. A study of the morphology of canine eyes enucleated or eviscerated due to complications following phacoemulsification. Veterinary Ophthalmology. 2003;6(3):219-226.

121. Biros DJ, Gelatt KN, Brooks DE, Kubilis PS, Andrew SE, Strubbe DT, Whigham HM. Development of glaucoma after cataract surgery in dogs: 220 cases (1987-1998). J Am Vet Med Assoc. 2000;216(11):1780-1786.

122. Klein HE, Krohne SG, Moore GE, Stiles J. Postoperative complications and visual outcomes of phacoemulsification in 103 dogs (179 eyes): 2006–2008. Veterinary ophthalmology. 2011;14(2):114-120.

123. Sigle KJ, Nasisse MP. Long-term complications after phacoemulsification for cataract removal in dogs: 172 cases (1995–2002). Journal of the American Veterinary Medical Association. 2006;228(1):74-79.

124. Wilcock BP, Peiffer RL, Jr. The pathology of lens-induced uveitis in dogs. Vet Pathol. 1987;24(6):549-553.

P a g e | 236

125. Lannek EB, Miller PE. Development of glaucoma after phacoemulsification for removal of cataracts in dogs: 22 cases (1987-1997). J Am Vet Med Assoc. 2001;218(1):70-76.

126. Scott EM, Esson DW, Fritz KJ, Dubielzig RR. Major breed distribution of canine patients enucleated or eviscerated due to glaucoma following routine cataract surgery as well as common histopathologic findings within enucleated globes. Veterinary Ophthalmology. 2013;16(s1):64-72.

127. Fischer CA. Lens-induced uveitis in dogs. J Am Anim Hosp Assoc. 1972;8:39-48.

128. Nelms SR, Nasisse MP, Davidson MG, Kirschner SE. Hyphema associated with retinal disease in dogs: 17 cases (1986-1991). J Am Vet Med Assoc. 1993;202(8):1289-1292.

129. Treadwell A, Naranjo C, Blocker T, Zarfoss M, Dubielzig RR. Clinical and histological characteristics of canine ocular gliovascular syndrome. Vet Ophthalmol. 2015;18(5):371-380.

130. Zeiss CJ, Dubielzig RR. A morphologic study of intravitreal membranes associated with intraocular hemorrhage in the dog. Vet Ophthalmol. 2004;7(4):239-243.

131. Samuelson D, Gum G, Gelatt K, Barrier K. Aqueous outflow in the beagle: unconventional outflow, using different-sized microspheres. American journal of veterinary research. 1985;46(1):242-248.

132. Bauer B, Sandmeyer L, Philibert H, Feng C, Grahn B. Chronic Glaucoma in Dogs: Relationships Between Histologic Lesions and the Gonioscopic Diagnosis of Pectinate Ligament Dysplasia. Veterinary Pathology. 2016;53(6):1197-1203.

133. Martin C. Development of pectinate ligament structure of the dog: study by scanning electron microscopy. American Journal of Veterinary Research. 1974;35(11):1433-1439.

134. Fricker GV, Smith K, Gould DJ. Survey of the incidence of pectinate ligament dysplasia and glaucoma in the UK Leonberger population. Veterinary Ophthalmology. 2016;19(5):379-385.

135. Pizzirani S, Carroll V, Pirie C. Pathologic factors involved with the late onset of canine glaucoma associated with goniodysgenesis. Preliminary study. Paper presented at: ACVO 39th Annual Conference. Boston2008.

136. Gabelt BAT, Kaufman PL. Production and flow of aqueous humor. In: Levin LA, Nilsson SF, Ver Hoeve J, Wu SM, eds. Adler's Physiology of the Eye. Eleventh ed. Edinburgh: Saunders Elsevier; 2011.

137. Shahidullah M, Wilson WS, Yap M, To CH. Effects of ion transport and channel-blocking drugs on aqueous humor formation in isolated bovine eye. Invest Ophthalmol Vis Sci. 2003;44(3):1185-1191.

P a g e | 237

138. O'Rourke J, Macri FJ, Berghoffer B. Studies in uveal physiology. I. Adaptation of isotope clearance procedures for external monitoring of anterior uveal bloodflow and aqueous humor turnover in the dog. Arch Ophthalmol. 1969;81(4):526-533.

139. Samuelson DA. Ophthalmic Anatomy. In: Gelatt KN, Gilger BC, Kern TJ, eds. Veterinary Ophthalmology. Vol I. Iowa: John Wiley & Sons, Inc; 2013:39-170.

140. Gum GG, MacKay EO. Physiology of the Eye. In: Gelatt KN, Gilger BC, Kern TJ, eds. Veterinary Ophthalmology. Vol I. Iowa: John Wiley & Sons, Inc; 2013:171-207.

141. Golubnitschaja O, Flammer J. What are the biomarkers for glaucoma? Surv Ophthalmol. 2007;52 Suppl 2(6):S155-161.

142. Bhattacharya SK, Lee RK, Grus FH, Seventh APORICWG. Molecular biomarkers in glaucoma. Invest Ophthalmol Vis Sci. 2013;54(1):121-131.

143. Mayeux R. Biomarkers: potential uses and limitations. NeuroRx. 2004;1(2):182-188.

144. Rachakonda V, Pan TH, Le WD. Biomarkers of neurodegenerative disorders: how good are they? Cell Research. 2004;14(5):349-358.

145. Savagian CA, Dubielzig RR, Nork TM. Comparison of the distribution of glial fibrillary acidic protein, heat shock protein 60, and hypoxia-inducible factor-1α in retinas from glaucomatous and normal canine eyes. American Journal of Veterinary Research. 2008;69(2):265-272.

146. Chen T, Gionfriddo JR, Tai PY, Novakowski AN, Alyahya K, Madl JE. Oxidative stress increases in retinas of dogs in acute glaucoma but not in chronic glaucoma. Veterinary Ophthalmology. 2015;18(4):261-270.

147. Anderson DR, Hendrickson A. Effect of intraocular pressure on rapid axoplasmic transport in monkey optic nerve. Invest Ophthalmol. 1974;13(10):771-783.

148. Quigley HA, Nickells RW, Kerrigan LA, Pease ME, Thibault DJ, Zack DJ. Retinal ganglion cell death in experimental glaucoma and after axotomy occurs by apoptosis. Invest Ophthalmol Vis Sci. 1995;36(5):774-786.

149. Schlötterer C. The evolution of molecular markers—just a matter of fashion? Nature Reviews Genetics. 2004;5(1):63-69.

150. Gelatt KN, Gum GG. Inheritance of primary glaucoma in the beagle. Am J Vet Res. 1981;42(10):1691-1693.

P a g e | 238

151. Wiggs JL, Auguste J, Allingham RR, Flor JD, Pericak-Vance MA, Rogers K, LaRocque KR, Graham FL, Broomer B, Del Bono E, Haines JL, Hauser M. Lack of association of mutations in optineurin with disease in patients with adult-onset primary open-angle glaucoma. Arch Ophthalmol. 2003;121(8):1181-1183.

152. Dagoneau N, Benoist-Lasselin C, Huber C, Faivre L, Megarbane A, Alswaid A, Dollfus H, Alembik Y, Munnich A, Legeai-Mallet L, Cormier-Daire V. ADAMTS10 mutations in autosomal recessive Weill-Marchesani syndrome. Am J Hum Genet. 2004;75(5):801-806.

153. Kuchtey J, Kuchtey RW. The microfibril hypothesis of glaucoma: implications for treatment of elevated intraocular pressure. J Ocul Pharmacol Ther. 2014;30(2-3):170-180.

154. Boote C, Palko JR, Sorensen T, Mohammadvali A, Elsheikh A, Komaromy AM, Pan X, Liu J. Changes in posterior scleral collagen microstructure in canine eyes with an ADAMTS10 mutation. Mol Vis. 2016;22:503-517.

155. Palko JR, Morris HJ, Pan X, Harman CD, Koehl KL, Gelatt KN, Plummer CE, Komaromy AM, Liu J. Influence of Age on Ocular Biomechanical Properties in a Canine Glaucoma Model with ADAMTS10 Mutation. PLoS One. 2016;11(6):e0156466.

156. Girard MJ, Suh JK, Bottlang M, Burgoyne CF, Downs JC. Biomechanical changes in the sclera of monkey eyes exposed to chronic IOP elevations. Invest Ophthalmol Vis Sci. 2011;52(8):5656-5669.

157. MacKay E, Hart H, Gelatt K. Aqueous humor and trabecular meshwork myocilin in normal and POAG Beagles. Veterinary Ophthalmology. 2004;7:453.

158. Gelatt K, MacKay E, Kallberg M. Aqueous humor myocilin in dogs with glaucoma and cataract. Veterinary Ophthalmology. 2005;8:439.

159. Hart H, Samuelson DA, MacKay E, Lewis P, Sherwood M, Gelatt K. Immunolocalization of MYOC Protein Within the Anterior Eye of Normal and Primary Open–Angle Glaucomatous Dogs. Investigative Ophthalmology & Visual Science. 2005;46(13):3697-3697.

160. Fingert JH, Heon E, Liebmann JM, Yamamoto T, Craig JE, Rait J, Kawase K, Hoh ST, Buys YM, Dickinson J, Hockey RR, Williams-Lyn D, Trope G, Kitazawa Y, Ritch R, Mackey DA, Alward WL, Sheffield VC, Stone EM. Analysis of myocilin mutations in 1703 glaucoma patients from five different populations. Hum Mol Genet. 1999;8(5):899-905.

161. Gong G, Kosoko-Lasaki O, Haynatzki GR, Wilson MR. Genetic dissection of myocilin glaucoma. Hum Mol Genet. 2004;13 Spec No 1(suppl 1):R91-102.

162. Ray K, Mukhopadhyay A, Acharya M. Recent advances in molecular genetics of glaucoma. Mol Cell Biochem. 2003;253(1-2):223-231.

P a g e | 239

163. Aung T, Yong VH, Chew PT, Seah SK, Gazzard G, Foster PJ, Vithana EN. Molecular analysis of the myocilin gene in Chinese subjects with chronic primary-angle closure glaucoma. Invest Ophthalmol Vis Sci. 2005;46(4):1303-1306.

164. Biochemical Marker. 2009; 8th:http://medical- dictionary.thefreedictionary.com/biochemical+marker. Accessed 14 May 2016, 2016.

165. Das P, Golde T. Dysfunction of TGF-β signaling in Alzheimer’s disease. The Journal of Clinical Investigation. 2006;116(11):2855-2857.

166. Mehta JL, Attramadal H. The TGFβ superfamily in cardiovascular biology. Cardiovascular Research. 2007;74(2):181-183.

167. Ali M, McKibbin M, Booth A, Parry DA, Jain P, Riazuddin SA, Hejtmancik JF, Khan SN, Firasat S, Shires M, Gilmour DF, Towns K, Murphy AL, Azmanov D, Tournev I, Cherninkova S, Jafri H, Raashid Y, Toomes C, Craig J, Mackey DA, Kalaydjieva L, Riazuddin S, Inglehearn CF. Null mutations in LTBP2 cause primary congenital glaucoma. Am J Hum Genet. 2009;84(5):664- 671.

168. Jelodari-Mamaghani S, Haji-Seyed-Javadi R, Suri F, Nilforushan N, Yazdani S, Kamyab K, Elahi E. Contribution of the latent transforming growth factor-beta binding protein 2 gene to etiology of primary open angle glaucoma and pseudoexfoliation syndrome. Mol Vis. 2013;19:333-347.

169. Kuchtey J, Kunkel J, Burgess LG, Parks MB, Brantley MA, Kuchtey RW. Elevated Transforming Growth Factor β1 in Plasma of Primary Open-Angle Glaucoma PatientsElevated Plasma TGFβ1 in POAG. Investigative Ophthalmology & Visual Science. 2014;55(8):5291-5297.

170. Prendes MA, Harris A, Wirostko BM, Gerber AL, Siesky B. The role of transforming growth factor beta in glaucoma and the therapeutic implications. Br J Ophthalmol. 2013;97(6):680- 686.

171. Ochiai Y, Ochiai H. Higher concentration of transforming growth factor-beta in aqueous humor of glaucomatous eyes and diabetic eyes. Jpn J Ophthalmol. 2002;46(3):249-253.

172. Fuchshofer R, Tamm ER. The role of TGF-beta in the pathogenesis of primary open-angle glaucoma. Cell Tissue Res. 2012;347(1):279-290.

173. Horiguchi M, Ota M, Rifkin DB. Matrix control of transforming growth factor-beta function. J Biochem. 2012;152(4):321-329.

174. Ramirez F, Rifkin DB. Extracellular microfibrils: contextual platforms for TGFβ and BMP signaling. Current opinion in cell biology. 2009;21(5):616-622.

P a g e | 240

175. Ramirez F, Sakai LY. Biogenesis and function of fibrillin assemblies. Cell Tissue Res. 2010;339(1):71-82.

176. Wheatley HM, Traboulsi EI, Flowers BE, Maumenee IH, Azar D, Pyeritz RE, Whittum-Hudson JA. Immunohistochemical localization of fibrillin in human ocular tissues. Relevance to the Marfan syndrome. Arch Ophthalmol. 1995;113(1):103-109.

177. Hann CR, Fautsch MP. The elastin fiber system between and adjacent to collector channels in the human juxtacanalicular tissue. Investigative Ophthalmology & Visual Science. 2011;52(1):45-50.

178. Schlotzer-Schrehardt U, von der Mark K, Sakai LY, Naumann GO. Increased extracellular deposition of fibrillin-containing fibrils in pseudoexfoliation syndrome. Invest Ophthalmol Vis Sci. 1997;38(5):970-984.

179. Bhattacharya SK, B’ann TG, Ruiz J, Picciani R, Kaufman PL. Cochlin expression in anterior segment organ culture models after TGFβ2 treatment. Investigative Ophthalmology & Visual Science. 2009;50(2):551-559.

180. Birke MT, Birke K, Lutjen-Drecoll E, Schlotzer-Schrehardt U, Hammer CM. Cytokine- dependent ELAM-1 induction and concomitant intraocular pressure regulation in porcine anterior eye perfusion culture. Invest Ophthalmol Vis Sci. 2011;52(1):468-475.

181. Lama PJ, Fechtner RD. Antifibrotics and wound healing in glaucoma surgery. Surv Ophthalmol. 2003;48(3):314-346.

182. Weinstein WL, Dietrich UM, Sapienza JS, Carmichael KP, Moore PA, Krunkosky TM. Identification of ocular matrix metalloproteinases present within the aqueous humor and iridocorneal drainage angle tissue of normal and glaucomatous canine eyes. Veterinary Ophthalmology. 2007;10(s1):108-116.

183. Weinstein WL. Identification and localization of matrix metalloproteinases (MMPs) present within the aqueous humor and iridocorneal drainage angle tissue of normal and glaucomatous canine eyes, uga; 2009.

184. Pumphrey SA, Pizzirani S, Pirie CG, Anwer MS, Logvinenko T. Western blot patterns of serum autoantibodies against optic nerve antigens in dogs with goniodysgenesis-related glaucoma. Am J Vet Res. 2013;74(4):621-628.

185. Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest. 2003;111(12):1805-1812.

186. Jiang B, Harper MM, Kecova H, Adamus G, Kardon RH, Grozdanic SD, Kuehn MH. Neuroinflammation in advanced canine glaucoma. Mol Vis. 2010;16:2092-2108.

P a g e | 241

187. Ahmed F, Brown KM, Stephan DA, Morrison JC, Johnson EC, Tomarev SI. Microarray analysis of changes in mRNA levels in the rat retina after experimental elevation of intraocular pressure. Invest Ophthalmol Vis Sci. 2004;45(4):1247-1258.

188. Piri N, Song M, Kwong JM, Caprioli J. Modulation of alpha and beta crystallin expression in rat retinas with ocular hypertension-induced ganglion cell degeneration. Brain Res. 2007;1141:1-9.

189. Nork TM, Ver Hoeve JN, Poulsen GL, Nickells RW, Davis MD, Weber AJ, Vaegan, Sarks SH, Lemley HL, Millecchia LL. Swelling and loss of photoreceptors in chronic human and experimental glaucomas. Arch Ophthalmol. 2000;118(2):235-245.

190. Wygnanski T, Desatnik H, Quigley HA, Glovinsky Y. Comparison of ganglion cell loss and cone loss in experimental glaucoma. Am J Ophthalmol. 1995;120(2):184-189.

191. Holopigian K, Greenstein VC, Seiple W, Hood DC, Ritch R. Electrophysiologic assessment of photoreceptor function in patients with primary open-angle glaucoma. J Glaucoma. 2000;9(2):163-168.

192. Velten IM, Korth M, Horn FK. The a-wave of the dark adapted electroretinogram in glaucomas: are photoreceptors affected? Br J Ophthalmol. 2001;85(4):397-402.

193. Durieux P, Etchepareborde S, Fritz D, Rosolen SG. Tumor necrosis factor-alpha concentration in the aqueous humor of healthy and diseased dogs: a preliminary pilot study. J Fr Ophtalmol. 2015;38(4):288-294.

194. Halliwell B, Gutteridge J. Free Radicals in Biology and Medicine. 3rd ed. New York: Oxford University Press Inc.; 1989.

195. Sies H. Oxidative Stress. San Diego: Academic Press; 1985.

196. Moreno MC, Campanelli J, Sande P, Sanez DA, Keller Sarmiento MI, Rosenstein RE. Retinal oxidative stress induced by high intraocular pressure. Free Radic Biol Med. 2004;37(6):803- 812.

197. Sacca SC, Pascotto A, Camicione P, Capris P, Izzotti A. Oxidative DNA damage in the human trabecular meshwork: clinical correlation in patients with primary open-angle glaucoma. Arch Ophthalmol. 2005;123(4):458-463.

198. Alvarado J, Murphy C, Juster R. Trabecular meshwork cellularity in primary open-angle glaucoma and nonglaucomatous normals. Ophthalmology. 1984;91(6):564-579.

P a g e | 242

199. Izzotti A, Bagnis A, Sacca SC. The role of oxidative stress in glaucoma. Mutat Res. 2006;612(2):105-114.

200. Tezel G. The immune response in glaucoma: a perspective on the roles of oxidative stress. Exp Eye Res. 2011;93(2):178-186.

201. Aslan M, Cort A, Yucel I. Oxidative and nitrative stress markers in glaucoma. Free Radic Biol Med. 2008;45(4):367-376.

202. Ko ML, Peng PH, Ma MC, Ritch R, Chen CF. Dynamic changes in reactive oxygen species and antioxidant levels in retinas in experimental glaucoma. Free Radic Biol Med. 2005;39(3):365- 373.

203. Ferreira SM, Lerner SF, Brunzini R, Reides CG, Evelson PA, Llesuy SF. Time course changes of oxidative stress markers in a rat experimental glaucoma model. Invest Ophthalmol Vis Sci. 2010;51(9):4635-4640.

204. Carter-Dawson L, Shen FF, Harwerth RS, Crawford M, Smith EL, Whitetree A. Glutathione content is altered in Müller cells of monkey eyes with experimental glaucoma. Neuroscience Letters. 2004;364(1):7-10.

205. Arnal E, Miranda M, Johnsen-Soriano S, Alvarez-Nolting R, Diaz-Llopis M, Araiz J, Cervera E, Bosch-Morell F, Romero FJ. Beneficial effect of docosahexanoic acid and lutein on retinal structural, metabolic, and functional abnormalities in diabetic rats. Curr Eye Res. 2009;34(11):928-938.

206. Tezel G, Hernandez R, Wax MB. Immunostaining of heat shock proteins in the retina and optic nerve head of normal and glaucomatous eyes. Arch Ophthalmol. 2000;118(4):511-518.

207. Tezel G, Seigel GM, Wax MB. Autoantibodies to small heat shock proteins in glaucoma. Invest Ophthalmol Vis Sci. 1998;39(12):2277-2287.

208. Wax MB, Tezel G, Kawase K, Kitazawa Y. Serum autoantibodies to heat shock proteins in glaucoma patients from Japan and the United States. Ophthalmology. 2001;108(2):296-302.

209. Wax MB, Tezel G, Yang J, Peng G, Patil RV, Agarwal N, Sappington RM, Calkins DJ. Induced autoimmunity to heat shock proteins elicits glaucomatous loss of retinal ganglion cell neurons via activated T-cell-derived fas-ligand. J Neurosci. 2008;28(46):12085-12096.

210. Alice LY, Fuchshofer R, Birke M, Kampik A, Bloemendal H, Welge-Lüssen U. Oxidative stress and TGF-β2 increase heat shock protein 27 expression in human optic nerve head astrocytes. Investigative ophthalmology & visual science. 2008;49(12):5403-5411.

P a g e | 243

211. Gonzalez-Iglesias H, Alvarez L, Garcia M, Escribano J, Rodriguez-Calvo PP, Fernandez-Vega L, Coca-Prados M. Comparative proteomic study in serum of patients with primary open-angle glaucoma and pseudoexfoliation glaucoma. J Proteomics. 2014;98:65-78.

212. You J, Willcox MD, Madigan MC, Wasinger V, Schiller B, Walsh BJ, Graham PH, Kearsley JH, Li Y. Tear fluid protein biomarkers. Advances in Clinical Chemistry. 2013;62:151-196.

213. Liotta LA, Petricoin EF. Cancer biomarkers: closer to delivering on their promise. Cancer Cell. 2011;20:279-280.

214. Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects. Mol Cell Proteomics. 2002;1(11):845-867.

215. Tezel G. A decade of proteomics studies of glaucomatous neurodegeneration. Proteomics Clin Appl. 2014;8(3-4):154-167.

216. Wong TT, Zhou L, Li J, Tong L, Zhao SZ, Li XR, Yu SJ, Koh SK, Beuerman RW. Proteomic profiling of inflammatory signaling molecules in the tears of patients on chronic glaucoma medication. Invest Ophthalmol Vis Sci. 2011;52(10):7385-7391.

217. Pieragostino D, Bucci S, Agnifili L, Fasanella V, D'Aguanno S, Mastropasqua A, Ciancaglini M, Mastropasqua L, Di Ilio C, Sacchetta P, Urbani A, Del Boccio P. Differential protein expression in tears of patients with primary open angle and pseudoexfoliative glaucoma. Mol Biosyst. 2012;8(4):1017-1028.

218. Pavlenko T, Chesnokova N, Davydova H, Okhotsimskaia T, Beznos O, Grigor'ev A. [Level of tear endothelin-1 and plasminogen in patients with glaucoma and proliferative diabetic retinopathy]. Vestnik Oftalmologii. 2012;129(4):20-23.

219. Slepova O, Frolov M, Morozova N, Frolov A, Lovpache D. [Markers of Fas-mediated apoptosis in primary open-angle glaucoma and opportunities of their pharmacological correction]. Vestnik Oftalmologii. 2011;128(4):27-31.

220. Openkova YY, Korobeiynikova EN, Rykin VS, Vinkova GA. [The analysis of status of biochemical indicators in blood serum and lacrimal fluid in patients with primary open-angle glaucoma]. Klin Lab Diagn. 2013(5):8-11.

221. Ghaffariyeh A, Honarpisheh N, Shakiba Y, Puyan S, Chamacham T, Zahedi F, Zarrineghbal M. Brain-derived neurotrophic factor in patients with normal-tension glaucoma. Optometry. 2009;80(11):635-638.

222. Tyers M, Mann M. From genomics to proteomics. Nature. 2003;422(6928):193-197.

P a g e | 244

223. Iwabe S, Moreno-Mendoza NA, Trigo-Tavera F, Crowder C, Garcia-Sanchez GA. Retrograde axonal transport obstruction of brain-derived neurotrophic factor (BDNF) and its TrkB receptor in the retina and optic nerve of American Cocker Spaniel dogs with spontaneous glaucoma. Vet Ophthalmol. 2007;10 Suppl 1(s1):12-19.

224. Pinazo-Duran MD, Zanon-Moreno V, Garcia-Medina JJ, Gallego-Pinazo R. Evaluation of presumptive biomarkers of oxidative stress, immune response and apoptosis in primary open-angle glaucoma. Curr Opin Pharmacol. 2013;13(1):98-107.

225. Ollivier FJ, Plummer CE, Barrie KP. Ophthalmic Examination and Diagnostics, Part 1: The Eye Examination and Diagnostic Procedures. In: Gelatt K, ed. Veterinary Ophthalmology. 4th ed. Iowa: Blackwell Publishing; 2007:438-476.

226. Kato K. Comparison of two handheld applanation tonometers and the association of central corneal thickness, age, and intraocular pressure in normal and diseased canine eyes. Veterinary Ophthalmology. 2014;17(6):417-425.

227. Spiessen L, Karck J, Rohn K, Meyer‐Lindenberg A. Clinical comparison of the TonoVet® rebound tonometer and the Tono‐Pen Vet® applanation tonometer in dogs and cats with ocular disease: glaucoma or corneal pathology. Veterinary ophthalmology. 2015;18(1):20-27.

228. Görig C, Coenen RT, Stades FC, Djajadiningrat-Laanen SC, Boevé MH. Comparison of the use of new handheld tonometers and established applanation tonometers in dogs. American journal of veterinary research. 2006;67(1):134-144.

229. Knollinger AM, La Croix NC, Barrett PM, Miller PE. Evaluation of a rebound tonometer for measuring intraocular pressure in dogs and horses. J Am Vet Med Assoc. 2005;227(2):244- 248.

230. Rusanen E, Florin M, Hässig M, Spiess BM. Evaluation of a rebound tonometer (Tonovet®) in clinically normal cat eyes. Veterinary ophthalmology. 2010;13(1):31-36.

231. Miller PE, Pickett JP, Majors LJ, Kurzman ID. Evaluation of two applanation tonometers in cats. Am J Vet Res. 1991;52(11):1917-1921.

232. Priehs DR, Gum GG, Whitley RD, Moore LE. Evaluation of three applanation tonometers in dogs. Am J Vet Res. 1990;51(10):1547-1550.

233. Park YW, Jeong MB, Kim TH, Ahn JS, Ahn JT, Park SA, Kim SE, Seo K. Effect of central corneal thickness on intraocular pressure with the rebound tonometer and the applanation tonometer in normal dogs. Vet Ophthalmol. 2011;14(3):169-173.

234. Slack JM, Stiles J, Moore GE. Comparison of a rebound tonometer with an applanation tonometer in dogs with glaucoma. Vet Rec. 2012;171(15):373.

P a g e | 245

235. Leiva M, Naranjo C, Pena M. Comparison of the rebound tonometer (ICare®) to the applanation tonometer (Tonopen XL®) in normotensive dogs. Veterinary ophthalmology. 2006;9(1):17-21.

236. Nagata N, Yuki M, Hasegawa T. In vitro and in vivo comparison of applanation tonometry and rebound tonometry in dogs. J Vet Med Sci. 2011;73(12):1585-1589.

237. Miller P, Pickett J, Majors L, Kurzman I. Clinical comparison of the Mackay-Marg and Tono- Pen applanation tonometers in the dog. Prog Vet Comp Ophthalmol. 1991;1(3):171-176.

238. Gelatt K, Gum G, Barrie K, Williams L. Diurnal variations in intraocular pressure in normotensive and glaucomatous Beagles. Glaucoma. 1981;3(2):121-124.

239. Gelatt KN, MacKay EO. Distribution of intraocular pressure in dogs. Vet Ophthalmol. 1998;1(2-3):109-114.

240. Boillot T, Gauvin M, Rosolen SG. Effect of topical application of tetracaine on intraocular pressure in dogs: preliminary results. J Fr Ophtalmol. 2013;36(5):402-407.

241. Kovalcuka L, Birgele E, Bandere D, Williams DL. Comparison of the effects of topical and systemic atropine sulfate on intraocular pressure and pupil diameter in the normal canine eye. Vet Ophthalmol. 2015;18(1):43-49.

242. Kovalcuka L, Birgele E, Bandere D, Williams DL. The effects of ketamine hydrochloride and diazepam on the intraocular pressure and pupil diameter of the dog’s eye. Veterinary ophthalmology. 2013;16(1):29-34.

243. Ghaffari MS, Rezaei MA, Mirani AH, Khorami N. The effects of ketamine-midazolam anesthesia on intraocular pressure in clinically normal dogs. Vet Ophthalmol. 2010;13(2):91- 93.

244. Taylor NR, Zele AJ, Vingrys AJ, Stanley RG. Variation in intraocular pressure following application of tropicamide in three different dog breeds. Vet Ophthalmol. 2007;10 Suppl 1(s1):8-11.

245. Baudouin C, Gastaud P. Influence of topical anesthesia on tonometric values of intraocular pressure. Ophthalmologica. 1994;208(6):309-313.

246. Morrison JC, Van Buskirk EM. The canine eye: pectinate ligaments and aqueous outflow resistance. Investigative Ophthalmology & Visual Science. 1982;23(6):726-732.

247. Maggs DJ. Diagnostic techniques. In: Maggs DJ, Miller PE, Ofri R, eds. Slatters Fundamentals of Veterinary Ophthalmology. 5th ed. St Louis, Missouri: Elsevier Saunders; 2013:79-109.

P a g e | 246

248. Gordon MO, Kass MA. The Ocular Hypertension Treatment Study: design and baseline description of the participants. Arch Ophthalmol. 1999;117(5):573-583.

249. Leske MC, Heijl A, Hyman L, Bengtsson B. Early Manifest Glaucoma Trial: design and baseline data. Ophthalmology. 1999;106(11):2144-2153.

250. Dubin AJ, Bentley E, Buhr KA, Miller PE. Evaluation of potential risk factors for development of primary angle-closure glaucoma in Bouviers des Flandres. Journal of the American Veterinary Medical Association. 2017;250(1):60-67.

251. Hasegawa T, Kawata M, Ota M. Ultrasound biomicroscopic findings of the iridocorneal angle in live healthy and glaucomatous dogs. Journal of Veterinary Medical Science. 2016;77(12):1625-1631.

252. Aubin M, Powell CC, Gionfriddo JR, Fails A. Ultrasound biomicroscopy of the feline anterior segment. Veterinary Ophthalmology. 2003;6:15-17.

253. Bentley E, Miller PE, Diehl KA. Use of high-resolution ultrasound as a diagnostic tool in veterinary ophthalmology. Journal of the American Veterinary Medical Association. 2003;223(11):1617-1622.

254. Bentley E, Miller PE, Diehl K. Evaluation of intra- and interobserver reliability and image reproducibility to assess usefulness of high-resolution ultrasonography for measurement of anterior segment structures of canine eyes. American journal of veterinary research. 2005;66:1775-1779.

255. Gibson TE, Roberts SM, Severin GA, Steyn PF, Wrigley RH. Comparison of gonioscopy and ultrasound biomicroscopy for evaluating the iridocorneal angle in dogs. Journal of the American Veterinary Medical Association. 1998;213(5):635-638.

256. Kawata M, Tsukizawa H, Nakayama M, Hasegawa T. Rectification of width and area of the ciliary cleft in dogs. Journal of Veterinary Medical Science. 2010;72(5):533-537.

257. Kawata M, Hasegawa T. Evaluation of the distance between Schwalbe’s line and the anterior lens capsule as a parameter for the correction of ultrasound biomicroscopic values of the canine iridocorneal angle. Veterinary Ophthalmology. 2013;16:169-174.

258. Crumley W, Gionfriddo JR, Radecki SV. Relationship of the iridocorneal angle, as measured using ultrasound biomicroscopy, with post‐operative increases in intraocular pressure post‐ phacoemulsification in dogs. Veterinary Ophthalmology. 2009;12(1):22-27.

259. Grozdanic SD, Matic M, Betts DM, Sakaguchi DS, Kardon RH. Recovery of canine retina and optic nerve function after acute elevation of intraocular pressure: implications for canine glaucoma treatment. Vet Ophthalmol. 2007;10 Suppl 1(s1):101-107.

P a g e | 247

260. Ofri R, Dawson W, Gelatt K. Visual resolution in normal and glaucomatous dogs determined by pattern electroretinogram. Progress in Veterinary and Comparative Ophthalmology. 1993;3:111-116.

261. Sims MH, Brooks DE. Changes in oscillatory potentials in the canine electroretinogram during dark adaptation. Am J Vet Res. 1990;51(10):1580-1586.

262. Hamor RE, Gerding Jr PA, Ramsey DT, Whiteley HE, Benson GJ, Schaeffer DJ. Evaluation of short-term increased intraocular pressure on flash-and pattern-generated electroretinograms of dogs. American journal of veterinary research. 2000;61(9):1087-1091.

263. Bentley E, Miller PE, Murphy CJ, Schoster JV. Combined cycloablation and gonioimplantation for treatment of glaucoma in dogs: 18 cases (1992-1998). J Am Vet Med Assoc. 1999;215(10):1469-1472.

264. Maślanka T. A review of the pharmacology of carbonic anhydrase inhibitors for the treatment of glaucoma in dogs and cats. The Veterinary Journal. 2015;203(3):278-284.

265. Maren TH. Carbonic anhydrase: chemistry, physiology, and inhibition. Physiol Rev. 1967;47(4):595-781.

266. DeSantis L. Preclinical overview of brinzolamide. Surv Ophthalmol. 2000;44 Suppl 2:S119- 129.

267. Becker B, Constant MA. Experimental tonography; the effect of the carbonic anhydrase inhibitor acetazoleamide on aqueous flow. AMA Arch Ophthalmol. 1955;54(3):321-329.

268. Bloom JN, Levene RZ, Thomas G, Kimura R. Fluorophotometry and the rate of aqueous flow in man. I. Instrumentation and normal values. Arch Ophthalmol. 1976;94(3):435-443.

269. Holm O, Wiebert O. The effect of systemically given acetazolamide (Diamox®) upon the formation of aqueous humour in the human eye, measured with a new photogrammetric method. Acta Ophthalmologica. 1968;46(6):1243-1250.

270. Linnér E, Friedenwald JS. The appearance time of fluorescein as an index of aqueous flow. American Journal of Ophthalmology. 1957;44(2):225-229.

271. Gelatt KN, Gum G, Williams LW, Gwin RM. Ocular hypotensive effects of carbonic anhydrase inhibitors in normotensive and glaucomatous Beagles. Am J Vet Res. 1979;40(3):334-345.

272. Gelatt KN, MacKay EO. Changes in intraocular pressure associated with topical dorzolamide and oral methazolamide in glaucomatous dogs. Vet Ophthalmol. 2001;4(1):61-67.

P a g e | 248

273. Skorobohach BJ, Ward DA, Hendrix DV. Effects of oral administration of methazolamide on intraocular pressure and aqueous humor flow rate in clinically normal dogs. Am J Vet Res. 2003;64(2):183-187.

274. Cawrse MA, Ward DA, Hendrix DV. Effects of topical application of a 2% solution of dorzolamide on intraocular pressure and aqueous humor flow rate in clinically normal dogs. Am J Vet Res. 2001;62(6):859-863.

275. Whelan N, Welch P, Pace A, Brienza C. A comparison of the efficacy of topical brinzolamide and dorzolamide alone and in combination with oral methazolamide in decreasing normal canine intraocular pressure. Paper presented at: Proceedings of the 30th Annual meeting of the American College of Veterinary Ophthalmologists, Chicago1999.

276. Willis AM, Diehl KA, Robbin TE. Advances in topical glaucoma therapy. Vet Ophthalmol. 2002;5(1):9-17.

277. Plummer CE, MacKay EO, Gelatt KN. Comparison of the effects of topical administration of a fixed combination of dorzolamide–timolol to monotherapy with timolol or dorzolamide on IOP, pupil size, and heart rate in glaucomatous dogs. Veterinary Ophthalmology. 2006;9(4):245-249.

278. Gum GG, Kingsbury S, Whitley RD, Garcia A, Gelatt KN. Effect of topical prostaglandin PGA2, PGA2 isopropyl ester, and PGF2 alpha isopropyl ester on intraocular pressure in normotensive and glaucomatous canine eyes. J Ocul Pharmacol. 1991;7(2):107-116.

279. Maślanka T. Pharmacology of topical prostaglandin F2α analogs and their place in the treatment of glaucoma in small animals. Journal of Veterinary Pharmacology and Therapeutics. 2015;38(2):105-112.

280. Ota T, Aihara M, Narumiya S, Araie M. The effects of prostaglandin analogues on IOP in prostanoid FP-receptor–deficient mice. Investigative Ophthalmology & Visual Science. 2005;46(11):4159-4163.

281. Ota T, Aihara M, Saeki T, Narumiya S, Araie M. The effects of prostaglandin analogues on prostanoid EP1, EP2, and EP3 receptor-deficient mice. Invest Ophthalmol Vis Sci. 2006;47(8):3395-3399.

282. Bahler CK, Howell KG, Hann CR, Fautsch MP, Johnson DH. Prostaglandins increase trabecular meshwork outflow facility in cultured human anterior segments. Am J Ophthalmol. 2008;145(1):114-119.

283. Bulin C, Albrecht U, Bode JG, Weber AA, Schror K, Levkau B, Fischer JW. Differential effects of vasodilatory prostaglandins on focal adhesions, cytoskeletal architecture, and migration in human aortic smooth muscle cells. Arterioscler Thromb Vasc Biol. 2005;25(1):84-89.

P a g e | 249

284. Oh D-J, Martin JL, Williams AJ, Russell P, Birk DE, Rhee DJ. Effect of latanoprost on the expression of matrix metalloproteinases and their tissue inhibitors in human trabecular meshwork cells. Investigative Ophthalmology & Visual Science. 2006;47(9):3887-3895.

285. Tsai S, Miller PE, Struble C, Howard S, Almazan A, Burke JA, Hughes PM, Li H, Conforti P, Lee SS, Robinson MR. Topical application of 0.005% latanoprost increases episcleral venous pressure in normal dogs. Vet Ophthalmol. 2012;15 Suppl 1(s1):71-78.

286. Tsai S, Almazan A, Lee SS, Li H, Conforti P, Burke J, Miller PE, Robinson MR. The effect of topical latanoprost on anterior segment anatomic relationships in normal dogs. Vet Ophthalmol. 2013;16(5):370-376.

287. Miller P, Bentley E, Diehl K, Carter R. High resolution ultrasound imaging of the anterior segment of dogs with spontaneous primary angle-closure glaucoma prior to, and following the topical application of 0.005% latanoprost. Investigative Ophthalmology & Visual Science. 2003;44(13):4408-4408.

288. Miller P, Bentley E, Croft MB, Whiteman A, Kaufman P. The effect of 0.005% latanoprost on the canine iridocorneal angle: an ultrasound biomicroscopic study. Investigative Ophthalmology & Visual Science. 2002;43(13):4073-4073.

289. Studer ME, Martin CL, Stiles J. Effects of 0.005% latanoprost solution on intraocular pressure in healthy dogs and cats. Am J Vet Res. 2000;61(10):1220-1224.

290. Smith LN, Miller PE, Felchle LM. Effects of topical administration of latanoprost, timolol, or a combination of latanoprost and timolol on intraocular pressure, pupil size, and heart rate in clinically normal dogs. American Journal of Veterinary Research. 2010;71(9):1055-1061.

291. Sarchahi AA, Abbasi N, Gholipour MA. Effects of an unfixed combination of latanoprost and pilocarpine on the intraocular pressure and pupil size of normal dogs. Vet Ophthalmol. 2012;15 Suppl 1(s1):64-70.

292. Kahane N, Bdolah-Abram T, Raskansky H, Ofri R. The effects of 1% prednisolone acetate on pupil diameter and intraocular pressure in healthy dogs treated with 0.005% latanoprost. Vet Ophthalmol. 2016;19(6):473-479.

293. Park S, Kang S, Lee E, Kwak J, Park E, Lim J, Seo K. Ultrasound biomicroscopic study of the effects of topical latanoprost on the anterior segment and ciliary body thickness in dogs. Veterinary Ophthalmology. 2016;19(6):498-503.

294. Gelatt KN, MacKay EO. Effect of different dose schedules of latanoprost on intraocular pressure and pupil size in the glaucomatous Beagle. Vet Ophthalmol. 2001;4(4):283-288.

P a g e | 250

295. Focht T, Bentley E, Miller P. Can initial response to latanoprost therapy be used as a prognostic indicator in canine patients with primary closed angle glaucoma. Paper presented at: 37th Annual Meeting of the American College of Veterinary Ophthalmologists2006.

296. Fechtner RD, Khouri AS, Zimmerman TJ, Bullock J, Feldman R, Kulkarni P, Michael AJ, Realini T, Warwar R. Anterior uveitis associated with latanoprost. Am J Ophthalmol. 1998;126(1):37- 41.

297. Chang JH, McCluskey P, Missotten T, Ferrante P, Jalaludin B, Lightman S. Use of ocular hypotensive prostaglandin analogues in patients with uveitis: does their use increase anterior uveitis and cystoid macular oedema? British Journal of Ophthalmology. 2008;92(7):916-921.

298. Markomichelakis NN, Kostakou A, Halkiadakis I, Chalkidou S, Papakonstantinou D, Georgopoulos G. Efficacy and safety of latanoprost in eyes with uveitic glaucoma. Graefes Arch Clin Exp Ophthalmol. 2009;247(6):775-780.

299. Johnstone McLean NS, Ward DA, Hendrix DV. The effect of a single dose of topical 0.005% latanoprost and 2% dorzolamide/0.5% timolol combination on the blood-aqueous barrier in dogs: a pilot study. Vet Ophthalmol. 2008;11(3):158-161.

300. Bito LZ. Species differences in the responses of the eye to irritation and trauma: a hypothesis of divergence in ocular defense mechanisms, and the choice of experimental animals for eye research. Exp Eye Res. 1984;39(6):807-829.

301. Alm A, Grierson I, Shields MB. Side effects associated with prostaglandin analog therapy. Surv Ophthalmol. 2008;53 Suppl1(6):S93-105.

302. Gelatt KN, Mackay EO. Effect of different dose schedules of bimatoprost on intraocular pressure and pupil size in the glaucomatous Beagle. J Ocul Pharmacol Ther. 2002;18(6):525- 534.

303. Lee SS, Burke J, Shen J, Almazan A, Orilla W, Hughes P, Zhang J, Li H, Struble C, Miller PE, Robinson MR. Bimatoprost sustained-release intracameral implant reduces episcleral venous pressure in dogs. Vet Ophthalmol. 2018;10.1111/vop.12522.

304. Gelatt KN, MacKay EO. Effect of different dose schedules of travoprost on intraocular pressure and pupil size in the glaucomatous Beagle. Vet Ophthalmol. 2004;7(1):53-57.

305. Carvalho AB, Laus JL, Costa VP, Barros PS, Silveira PR. Effects of travoprost 0.004% compared with latanoprost 0.005% on the intraocular pressure of normal dogs. Vet Ophthalmol. 2006;9(2):121-125.

P a g e | 251

306. Kwak J, Kang S, Lee ER, Park S, Park S, Park E, Lim J, Seo K. Effect of preservative-free tafluprost on intraocular pressure, pupil diameter, and anterior segment structures in normal canine eyes. Vet Ophthalmol. 2017;20(1):34-39.

307. Akaishi T, Kurashima H, Odani-Kawabata N, Ishida N, Nakamura M. Effects of repeated administrations of tafluprost, latanoprost, and travoprost on optic nerve head blood flow in conscious normal rabbits. J Ocul Pharmacol Ther. 2010;26(2):181-186.

308. Traverso CE, Ropo A, Papadia M, Uusitalo H. A phase II study on the duration and stability of the intraocular pressure-lowering effect and tolerability of Tafluprost compared with latanoprost. J Ocul Pharmacol Ther. 2010;26(1):97-104.

309. Uusitalo H, Pillunat LE, Ropo A, Phase IIISI. Efficacy and safety of tafluprost 0.0015% versus latanoprost 0.005% eye drops in open-angle glaucoma and ocular hypertension: 24-month results of a randomized, double-masked phase III study. Acta Ophthalmol. 2010;88(1):12-19.

310. Yoshitomi T, Ito Y. Effects of indomethacin and prostaglandins on the dog iris sphincter and dilator muscles. Invest Ophthalmol Vis Sci. 1988;29(1):127-132.

311. Abramson DH, Chang S, Coleman J. Pilocarpine therapy in glaucoma: effects on anterior chamber depth and lens thickness in patients receiving long-term therapy. Arch Ophthalmol. 1976;94(6):914-918.

312. Carreras FJ, Porcel D, Gonzalez-Caballero F. Expanding forces in aqueous outflow pathways of a nonaccommodating mammal: an approach via comparative dynamic morphology. Comp Biochem Physiol A Physiol. 1997;117(2):197-209.

313. Van Alphen GW, Robinette SL, Macri FJ. Drug effects on ciliary muscle and choroid preparations in vitro. Arch Ophthalmol. 1962;68(1):81-93.

314. Gwin RM, Gelatt KN, Gum GG, Peiffer RL, Jr., Williams LW. The effect of topical pilocarpine on intraocular pressure and pupil size in the normotensive and glaucomatous beagle. Invest Ophthalmol Vis Sci. 1977;16(12):1143-1148.

315. Whitley R, Gelatt K, Gum G. Dose-response of topical pilocarpine in the normotensive and glaucomatous Beagle. American Journal of Veterinary Research. 1980;41(3):417-424.

316. Gum G, Metzger K, Gelatt K, Gilley R, Gelatt K. Tonographic effects of pilocarpine and pilocarpine/epinephrine in dogs. Journal of Small Animal Practice. 1993;34(3):112-116.

317. Carrier M, Gum G. Effects of 4% pilocarpine gel on normotensive and glaucomatous canine eyes. American Journal of Veterinary Research. 1989;50(2):239-244.

P a g e | 252

318. Martin CL, Wyman M. Primary glaucoma in the dog. The Veterinary Clinics of North America. 1978;8(2):257-286.

319. Gelatt K, Mackay E, Gelatt J, Stengard-Ollies K, Aza J. Effects on intraocular pressure and pupil size in glaucomatous beagles after topical pilocarpine instilled with standard (pH 5) and buffer-tip (pH 7) droptainers. Journal of Ocular Pharmacology and Therapeutics. 1997;13(2):95-104.

320. Krohne S, Gionfriddo J, Morrison E. Inhibition of pilocarpine-induced aqueous humor flare, hypotony, and miosis by topical administration of anti-inflammatory and anesthetic drugs to dogs. American Journal of Veterinary Research. 1998;59(4):482-488.

321. Stuhr CM, Miller PE, Murphy CJ, Schoster JV, Thomas CB. Effect of intracameral administration of carbachol on the postoperative increase in intraocular pressure in dogs undergoing cataract extraction. Journal of the American Veterinary Medical Association. 1998;212(12):1885-1888.

322. Crasta M, Clode AB, McMullen Jr RJ, Pate DO, Gilger BC. Effect of three treatment protocols on acute ocular hypertension after phacoemulsification and aspiration of cataracts in dogs. Veterinary Ophthalmology. 2010;13(1):14-19.

323. Gelatt K, Gum G, Wolf E, White M. Dose response of topical carbamylcholine chloride (carbachol) in normotensive and early glaucomatous beagles. American Journal of Veterinary Research. 1984;45(3):547-554.

324. Gum G, Gelatt K, Gelatt J, Jones R. Effect of topically applied demecarium bromide and echothiophate iodide on intraocular pressure and pupil size in beagles with normotensive eyes and beagles with inherited glaucoma. American Journal of Veterinary Research. 1993;54(2):287-293.

325. Miller PE, Schmidt GM, Vainisi SJ, Swanson JF, Herrmann MK. The efficacy of topical prophylactic antiglaucoma therapy in primary closed angle glaucoma in dogs: a multicenter clinical trial. Journal of the American Animal Hospital Association. 2000;36(5):431-438.

326. Frishman WH, Fuksbrumer MS, Tannenbaum M. Topical Ophthalmic β‐Adrenergic Blockade for the Treatment of Glaucoma and Ocular Hypertension. The Journal of Clinical Pharmacology. 1994;34(8):795-803.

327. Jung HJ, Abou-Jaoude M, Carbia BE, Plummer C, Chauhan A. Glaucoma therapy by extended release of timolol from nanoparticle loaded silicone-hydrogel contact lenses. J Control Release. 2013;165(1):82-89.

328. Wilkie D, Latimer C. Effects of topical administration of timolol maleate on intraocular pressure and pupil size in dogs. American Journal of Veterinary Research. 1991;52(3):432- 435.

P a g e | 253

329. Gum G, Larocca R, Gelatt K, Mead J, Gelatt J. The effect of topical timolol maleate on intraocular pressure in normal beagles and beagles with inherited glaucoma. Progress in Veterinary and Comparative Ophthalmology. 1991;1(3):141-149.

330. Gelatt K, Larocca R, Gelatt J, Strubbe D, MacKay E. Evaluation of multiple doses of 4 and 6% timolol, and timolol combined with 2% pilocarpine in clinically normal beagles and beagles with glaucoma. American Journal of Veterinary Research. 1995;56(10):1325-1331.

331. Takiyama N, Shoji S, Habata I, Ohba S. The effects of a timolol maleate gel-forming solution on normotensive beagle dogs. Journal of Veterinary Medical Science. 2006;68(6):631-633.

332. Juzych MS, Zimmerman TJ. Beta-blockers. Philadelphia: Lippincott-Raven; 1997.

333. Shimazaki J, Hanada K, Yagi Y, Yamagami J, Ishioka M, Shimmura S, Tsubota K. Changes in ocular surface caused by antiglaucomatous eyedrops: prospective, randomised study for the comparison of 0.5% timololv 0.12% unoprostone. British Journal of Ophthalmology. 2000;84(11):1250-1254.

334. Wasserman NT, Kennard G, Cochrane ZN, Felchle LM. Effects of oral isosorbide and glycerol on intraocular pressure, serum osmolality, and blood glucose in normal dogs. Vet Ophthalmol. 2013;16(1):20-24.

335. Duncan Jr L, Ellis P, Paterson C. Effect of hyperosmotic agents on vitreous osmolality. Experimental Eye Research. 1970;10(1):129-132.

336. Galin MA, Davidson R. Influence of Osmotic Agents on Outflow Resistance. Am J Ophthalmol. 1965;59(6):1057-1063.

337. D'Alena P, Ferguson W. Adverse effects after glycerol orally and mannitol parenterally. Arch Ophthalmol. 1966;75(2):201-203.

338. Lorimer DW, Hakanson NE, Pion PD, Merideth RE. The effect of intravenous mannitol or oral glycerol on intraocular pressure in dogs. Cornell Vet. 1989;79(3):249-258.

339. Mauger TF, Craig LE. Havener's Ocular Pharmacology. 6th ed. St Louis: Mosby; 1994.

340. Mehra KS, Singh R, Char JN, Rajyashree K. Lowering of intraocular tension. Effects of isosorbide and glycerin. Arch Ophthalmol. 1971;85(2):167-168.

341. Brooks DE. Glaucoma in the dog and cat. Vet Clin North Am Small Anim Pract. 1990;20(3):775-797.

P a g e | 254

342. Dees DD, Fritz KJ, MacLaren NE, Esson DW, Gaerig S, Annora M, Atkins RM, Knollinger AM. Efficacy of prophylactic antiglaucoma and anti‐inflammatory medications in canine primary angle‐closure glaucoma: a multicenter retrospective study (2004–2012). Veterinary Ophthalmology. 2014;17(3):195-200.

343. Stavinohova R, Newton JR, Busse C. The effect of prophylactic topical carbonic anhydrase inhibitors in canine primary closed-angle glaucoma. J Small Anim Pract. 2015;56(11):662- 666.

344. MacKay EO, Gelatt KN. Effect of Coherin™ on intraocular pressure, pupil size and heart rate in the glaucomatous Beagle: a pilot study. Veterinary Ophthalmology. 2013;16(3):198-203.

345. Fischer KM, Ward DA, Hendrix DV. Effects of a topically applied 2% delta-9- tetrahydrocannabinol ophthalmic solution on intraocular pressure and aqueous humor flow rate in clinically normal dogs. Am J Vet Res. 2013;74(2):275-280.

346. Wilkie DA, Latimer CA. Effects of topical administration of timolol maleate on intraocular pressure and pupil size in dogs. Am J Vet Res. 1991;52(3):432-435.

347. Maggio F, Bras D. Surgical Treatment of Canine Glaucoma: Filtering and End-Stage Glaucoma Procedures. Vet Clin North Am Small Anim Pract. 2015;45(6):1261-1282, vi-vii.

348. Cook CS. Surgery for glaucoma. Vet Clin North Am Small Anim Pract. 1997;27(5):1109-1129.

349. Sapienza JS, van der Woerdt A. Combined transscleral diode laser cyclophotocoagulation and Ahmed gonioimplantation in dogs with primary glaucoma: 51 cases (1996–2004). Veterinary Ophthalmology. 2005;8(2):121-127.

350. Rebolleda G, Munoz FJ, Murube J. Audible pops during cyclodiode procedures. J Glaucoma. 1999;8(3):177-183.

351. Hardman C, Stanley RG. Diode laser transscleral cyclophotocoagulation for the treatment of primary glaucoma in 18 dogs: a retrospective study. Vet Ophthalmol. 2001;4(3):209-215.

352. Francis BA, Kawji AS, Vo NT, Dustin L, Chopra V. Endoscopic cyclophotocoagulation (ECP) in the management of uncontrolled glaucoma with prior aqueous tube shunt. J Glaucoma. 2011;20(8):523-527.

353. Gelatt K, Brooks D, Miller T, Smith P, Sapienza J, Pellicane C. Issues in ophthalmic therapy: the development of anterior chamber shunts for the clinical management of the canine glaucomas. Progress in Veterinary and Comparative Ophthalmology. 1992;2:59-64.

354. Bedford P. A clinical evaluation of a one‐piece drainage system in the treatment of canine glaucoma. Journal of small animal practice. 1989;30(2):68-75.

P a g e | 255

355. Gelatt KN, Gunn G, Samuelson D, Mandelkorn R, Olander K, Zimmerman T. Evaluation of the Krupin-Denver Valve Implant in the normal and glaucomatous beagle. Journal of the American Veterinary Medical Association. 1987;191(11):1404-1409.

356. Garcia-Sanchez GA, Whitley RD, Brooks DE, Trigo F, Pinon A. Ahmed valve implantation to control intractable glaucoma after phacoemulsification and intraocular lens implantation in a dog. Vet Ophthalmol. 2005;8(2):139-144.

357. Lew M, Lew S, Brzeski W. Short-term results of Ahmed glaucoma valve implantation in the surgical treatment of angle-recession glaucoma in dog. Polish journal of veterinary sciences. 2007;11(4):377-383.

358. Westermeyer HD, Hendrix DV, Ward DA. Long-term evaluation of the use of Ahmed gonioimplants in dogs with primary glaucoma: nine cases (2000–2008). Journal of the American Veterinary Medical Association. 2011;238(5):610-617.

359. Tinsley D, Betts D. Clinical experience with a glaucoma drainage device in dogs. Veterinary and Comparative Ophthalmology. 1994;4(2):77-84.

360. Bentley E, Nasisse M, Glover T, Nelms S. Implantation of filtering devices in dogs with glaucoma: preliminary results in 13 eyes. Veterinary and Comparative Ophthalmology. 1996;6(4):243-246.

361. Cullen CL, Allen AL, Grahn BH. Anterior chamber to frontal sinus shunt for the diversion of aqueous humor: a pilot study in four normal dogs. Vet Ophthalmol. 1998;1(1):31-39.

362. Cullen CL. Cullen frontal sinus valved glaucoma shunt: preliminary findings in dogs with primary glaucoma. Vet Ophthalmol. 2004;7(5):311-318.

363. Naranjo C, Dubielzig RR. Histopathological study of the causes for failure of intrascleral prostheses in dogs and cats. Vet Ophthalmol. 2014;17(5):343-350.

364. Rankin AJ, Lanuza R, KuKanich B, Crumley WC, Pucket JD, Allbaugh RA, Meekins JM. Measurement of plasma gentamicin concentrations postchemical ciliary body ablation in dogs with chronic glaucoma. Vet Ophthalmol. 2016;19(1):57-62.

365. Low MC, Landis ML, Peiffer RL. Intravitreal cidofovir injection for the management of chronic glaucoma in dogs. Vet Ophthalmol. 2014;17(3):201-206.

366. Duke FD, Strong TD, Bentley E, Dubielzig RR. Canine ocular tumors following ciliary body ablation with intravitreal gentamicin. Vet Ophthalmol. 2013;16(2):159-162.

P a g e | 256

367. Marchione B, da Silva Curiel JM, Zhou Y, Seetao J. Effectiveness of gentamicin for pharmacologic ablation to treat end stage glaucoma in dogs. Veterinary Ophthalmology. 2011;14(6):421.

368. Goldenberg R, Evans P, da Silva Curiel J, Marchione B. Retrospective evaluation of success rates of pharmacological ciliary body ablation with gentamicin for the treatment of uncontrolled glaucoma in dogs. Veterinary Ophthalmology. 2014;17(6):E49.

369. Bingaman D, Lindley D, Glickman N, Krohne S, Bryan G. Intraocular gentamicin and glaucoma: a retrospective study of 60 dog and cat eyes (1985-1993). Ophthalmic Literature. 1995;1(48):61.

370. Miller PE, Murphy CJ. Vision in dogs. Journal of the American Veterinary Medical Association. 1995;207:1623-1634.

371. Duke-Elder S. System of Ophthalmology Vol. 1 The Eye in Evolution. St Louis: CV Mosby Co; 1958.

372. Odom J, Bromberg N, Dawson W. Canine visual acuity: retinal and cortical field potentials evoked by pattern stimulation. American Journal of Physiology-Regulatory, Integrative and Comparative Physiology. 1983;245(5):R637-R641.

373. Ezeh P, Myers L, Cummins K, Whitley R. Utilizing an optokinetic device in assessing the functional visual acuity of the dog. Progress in Veterinary Neurology. 1990;1(4):427-432.

374. Miller WW, Parisi D. Development and validation of the canine visual function instrument. Veterinary Ophthalmology. 2018;21(6):586-594.

375. Keeffe JE. Assessment of low vision in developing countries: Assessment of functional vision. Vol Book 2. University of Melbourne, Australia: World Health Organization; 1995.

376. Scott IU, Schein OD, West S, Bandeen-Roche K, Enger C, Folstein MF. Functional status and quality of life measurement among ophthalmic patients. Archives of Ophthalmology. 1994;112(3):329-335.

377. Steinberg EP, Tielsch JM, Schein OD, Javitt JC, Sharkey P, Cassard SD, Legro MW, Diener- West M, Bass EB, Damiano AM. The VF-14: an index of functional impairment in patients with cataract. Archives of Ophthalmology. 1994;112(5):630-638.

378. Mangione C, Lee P, Berry S, Spritzer K, Janz N, Klein R, Owsley C, Hays R. The NEI-VFQ 51- item test version and its relationship with visual acuity across 5 diseases. Investigative Ophthalmology & Visual Science. 1996;37(3):873-873.

P a g e | 257

379. Fletcher AE, Ellwein LB, Selvaraj S, Vijaykumar V, Rahmathullah R, Thulasiraj R. Measurements of vision function and quality of life in patients with cataracts in southern India: report of instrument development. Archives of Ophthalmology. 1997;115(6):767-774.

380. Katsumi O, Chedid SG, Kronheim JK, Henry RK, Jones CM, Hirose T. Visual Ability Score‐A new method to analyze ability in visually impaired children. Acta Ophthalmologica Scandinavica. 1998;76(1):50-55.

381. Brenner MH, Curbow B, Javitt JC, Legro MW, Sommer A. Vision change and quality of life in the elderly: response to cataract surgery and treatment of other chronic ocular conditions. Archives of Ophthalmology. 1993;111(5):680-685.

382. Gutierrez P, Wilson MR, Johnson C, Gordon M, Cioffi GA, Ritch R, Sherwood M, Meng K, Mangione CM. Influence of glaucomatous visual field loss on health-related quality of life. Archives of Ophthalmology. 1997;115(6):777-784.

383. Carta A, Braccio L, Belpoliti M, Soliani L, Sartore F, Gandolfi SA, Maraini G. Self-assessment of the quality of vision: association of questionnaire score with objective clinical tests. Current Eye Research. 1998;17(5):506-512.

384. Gothwal VK, Lovie-Kitchin JE, Nutheti R. The development of the LV Prasad-Functional Vision Questionnaire: a measure of functional vision performance of visually impaired children. Investigative Ophthalmology & Visual Science. 2003;44(9):4131-4139.

385. Katsumi O, Chedid SG, Kronheim JK, Henry RK, Denno S, Hirose T. Correlating preferential looking visual acuity and visual behavior in severely visually handicapped children. Acta Ophthalmologica Scandinavica. 1995;73(5):407-413.

386. Streiner DL, Norman GR, Cairney J. Validity. In: Health measurement scales: A practical guide to their development and use. 5th ed. Oxford, United Kingdom: Oxford University Press; 2014:227-253.

387. Oppenheim AN. Questionnaire design, interviewing and attitude measurement. 2nd ed. 2000 ed. London: Bloomsbury Academic; 2000.

388. Foddy W, Foddy WH. Constructing questions for interviews and questionnaires: Theory and practice in social research. Cambridge, UK: Cambridge University Press; 1994.

389. Lynn MR. Determination and quantification of content validity. Nursing Research. 1986;35:382-385.

390. Floyd FJ, Widaman KF. Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment. 1995;7(3):286-299.

P a g e | 258

391. Norman GR, Streiner DL. Principal components and factor analysis. In: Biostatistics the Bare Essentials St. Louis: Mosby-Year Book Inc; 1994:129-142.

392. Reid J, Wiseman-Orr M, Scott EM, Nolan A. Development, validation and reliability of a web- based questionnaire to measure health-related quality of life in dogs. Journal of Small Animal Practice. 2013;54:227-233.

393. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin. 1979;86(2):420-428.

394. Lohr KN, Aaronson NK, Alonso J, Burnam MA, Patrick DL, Perrin EB, Roberts JS. Evaluating quality-of-life and health status instruments: development of scientific review criteria. Clinical Therapeutics. 1996;18(5):979-992.

395. Husted JA, Cook RJ, Farewell VT, Gladman DD. Methods for assessing responsiveness: a critical review and recommendations. Journal of Clinical Epidemiology. 2000;53(5):459-468.

396. Kocher MS, Horan MP, Briggs KK, Richardson TR, O'holleran J, Hawkins RJ. Reliability, validity, and responsiveness of the American Shoulder and Elbow Surgeons subjective shoulder scale in patients with shoulder instability, rotator cuff disease, and glenohumeral arthritis. The Journal of Bone and Joint Surgery. 2005;87(9):2006-2011.

397. Armstrong FD, Toledano SR, Miloslavich K, Lackman‐Zeman L, Levy JD, Gay CL, Schuman WB, Fishkin PE. The Miami pediatric quality of life questionnaire: parent scale. International Journal of Cancer. 1999;83(S12):11-17.

398. Noble CE, Wiseman-Orr LM, Scott ME, Nolan AM, Reid J. Development, initial validation and reliability testing of a web-based, generic feline health-related quality-of-life instrument. Journal of Feline Medicine and Surgery. 2018;doi- org.ezproxy1.library.usyd.edu.au/10.1177/1098612X18758176:1098612X18758176.

399. Swamy BN, Chia E-M, Wang JJ, Rochtchina E, Mitchell P. Correlation between vision- and health-related quality of life scores. Acta Ophthalmologica. 2009;87:335-339.

400. Todorov A, Kirchner C. Bias in proxies' reports of disability: data from the National Health Interview Survey on disability. American Journal of Public Health. 2000;90(8):1248.

401. Frost N, Sparrow J, Durant J, Donovan J, Peters T, Brookes S. Development of a questionnaire for measurement of vision-related quality of life. Ophthalmic Epidemiology. 1998;5(4):185- 210.

402. Addington-Hall J, Kalra L. Measuring quality of life: Who should measure quality of life? BMJ: British Medical Journal. 2001;322(7299):1417.

P a g e | 259

403. Weih LM, Hassell JB, Keeffe JE. Assessment of the impact of vision impairment. Investigative Ophthalmology & Visual Science. 2002;43(4):927-935.

404. Serpell JA, Hsu Y. Development and validation of a novel method for evaluating behavior and temperament in guide dogs. Applied Animal Behaviour Science. 2001;72(4):347-364.

405. Hsu Y, Serpell JA. Development and validation of a questionnaire for measuring behavior and temperament traits in pet dogs. Journal of the American Veterinary Medical Association. 2003;223(9):1293-1300.

406. Wiseman-Orr ML, Nolan AM, Reid J, Scott EM. Development of a questionnaire to measure the effects of chronic pain on health-related quality of life in dogs. American Journal of Veterinary Research. 2004;65(8):1077-1084.

407. Manificat S. A new instrument to evaluate infant quality of life. MAPI Research Institute Quality of Life Newsletter. 1999;23:7-8.

408. Streinmer DL. A checklist for evaluating the usefulness of rating scales. The Canadian Journal of Psychiatry. 1993;38(2):140-148.

409. Tuley MR, Mulrow CD, McMahan CA. Estimating and testing an index of responsiveness and the relationship of the index to power. Journal of Clinical Epidemiology. 1991;44(4-5):417- 421.

410. Walls GL. The vertebrate eye and its adaptive radiation. New York: Hafner Pub. Co.; 1963.

411. Westheimer G. Optotype recognition under degradation: comparison of size, contrast, blur, noise and contour‐perturbation effects. Clinical and Experimental Optometry. 2016;99:66- 72.

412. Fantz RL. Pattern vision in young infants. The Psychological Record. 1958;8:43-47.

413. Teller DY. The forced-choice preferential looking procedure: A psychophysical technique for use with human infants. Infant Behavior and Development. 1979;2:135-153.

414. Gwiazda J, Brill S, Mohindra I, Held R. Infant visual acuity and its meridional variation. Vision Research. 1978;18(11):1557-1564.

415. Dobson V, Teller DY. Visual acuity in human infants: a review and comparison of behavioral and electrophysiological studies. Vision Research. 1978;18(11):1469-1483.

416. Atkinson J, Braddick O, Pimm-Smith E. 'Preferential looking'for monocular and binocular acuity testing of infants. British Journal of Ophthalmology. 1982;66(4):264-268.

P a g e | 260

417. Boothe RG, Dobson V, Teller DY. Postnatal development of vision in human and nonhuman primates. Annual Review of Neuroscience. 1985;8(1):495-545.

418. Jacobson SG, Mohindra I, Held R. Visual acuity of infants with ocular diseases. American Journal of Ophthalmology. 1982;93(2):198-209.

419. Mayer D, Dobson V. Assessment of vision in young children: a new operant approach yields estimates of acuity. Investigative Ophthalmology & Visual Science. 1980;19(5):566-570.

420. Teller DY, McDonald MA, Preston K, Sebris SL, Dobson V. Assessment of visual acuity in infants and children; the acuity card procedure. Developmental Medicine & Child Neurology. 1986;28(6):779-789.

421. McDonald M, Dobson V, Sebris S, Baitch L, Varner D, Teller D. The acuity card procedure: a rapid test of infant acuity. Investigative Ophthalmology & Visual Science. 1985;26(8):1158- 1162.

422. Byosiere S-E, Feng LC, Woodhead JK, Rutter NJ, Chouinard PA, Howell TJ, Bennett PC. Visual perception in domestic dogs: susceptibility to the Ebbinghaus–Titchener and Delboeuf illusions. Animal Cognition. 2016;20(3):435-448.

423. Byosiere S, Feng L, Rutter N, Woodhead J, Chouinard P, Howell T, Bennett P. Do dogs see the Ponzo illusion. Animal Behavior and Cognition. 2017;44(4):396-412.

424. Byosiere S-E, Feng LC, Chouinard PA, Howell TJ, Bennett PC. Relational concept learning in domestic dogs: Performance on a two-choice size discrimination task generalises to novel stimuli. Behavioural Processes. 2017;145:93-101.

425. Peirce JW. Generating stimuli for neuroscience using PsychoPy. Frontiers in Neuroinformatics. 2009;2:10.

426. Holladay JT. Proper method for calculating average visual acuity. Journal of Refractive Surgery. 1997;13(4):388-391.

427. Teller DY. Measurement of visual acuity in human and monkey infants: the interface between laboratory and clinic. Behavioural Brain Research. 1983;10(1):15-23.

428. Schmid KL, Wildsoet CF. Assessment of visual acuity and contrast sensitivity in the chick using an optokinetic nystagmus paradigm. Vision research. 1998;38(17):2629-2634.

429. Cameron DJ, Rassamdana F, Tam P, Dang K, Yanez C, Ghaemmaghami S, Dehkordi MI. The optokinetic response as a quantitative measure of visual acuity in zebrafish. Journal of Visualized Experiments. 2013;10.3791/50832(80):e50832.

P a g e | 261

430. Bloom M, Berkley M. Visual acuity and the near point of accommodation in cats. Vision Research. 1977;17(6):723-730.

431. Graham KL, Reid J, Whittaker CJG, Hall EJS, Caruso K, McCowan CI, White A. Development of a vision impairment score for the assessment of functional vision in dogs: initial evidence of validity, reliability and responsiveness. Veterinary Ophthalmology. 2018(Under review).

432. Clark DL, Clark RA. The effects of time, luminance, and high contrast targets: Revisiting grating acuity in the domestic cat. Experimental Eye Research. 2013;116:75-78.

433. Murphy CJ, Mutti DO, Zadnik K, Ver Hoeve J. Effect of optical defocus on visual acuity in dogs. American Journal of Veterinary Research. 1997;58(4):414-418.

434. Tanaka T, Ikeuchi E, Mitani S, Eguchi Y, Uetake K. Studies on the visual acuity of dogs using shape discrimination learning. Animal Science Journal. 2000;71(6):614-620.

435. Gaiddon J, Bouhana N, Lallement P. Refraction by retinoscopy of normal, aphakic, and pseudophakic canine eyes: advantage of a 41-diopter intraocular lens? Veterinary and Comparative Ophthalmology. 1996;6:121-124.

436. Black J, Browning SR, Collins AV, Phillips JR. A canine model of inherited myopia: familial aggregation of refractive error in labrador retrievers. Investigative Ophthalmology & Visual Science. 2008;49(11):4784-4789.

437. Mutti DO, Zadnik K, Murphy CJ. Naturally occurring vitreous chamber-based myopia in the Labrador retriever. Investigative Ophthalmology & Visual Science. 1999;40(7):1577-1584.

438. Kubai MA, Bentley E, Miller PE, Mutti DO, Murphy CJ. Refractive states of eyes and association between ametropia and breed in dogs. American Journal of Veterinary Research. 2008;69(7):946-951.

439. Williams LA, Kubai MA, Murphy CJ, Mutti DO. Ocular components in three breeds of dogs with high prevalence of myopia. Optometry and Vision Science. 2011;88(2):269-274.

440. Murphy CJ, Zadnik K, Mannis M. Myopia and refractive error in dogs. Investigative Ophthalmology & Visual Science. 1992;33(8):2459-2463.

441. Ofri R. Optics and Physiology of Vision. In: Gelatt K, Gilger BC, Kern TJ, eds. Veterinary Ophthalmology. Vol I. 5th ed. Iowa: John Wiley & Sons Inc; 2013:208-269.

442. Virgili G, Acosta R, Bentley SA, Giacomelli G, Allcock C, Evans JR. Reading aids for adults with low vision. Cochrane Database of Systemic Reviews. 2018;18(4):CD003303.

P a g e | 262

443. Garcia MM, Ying G-s, Cocores CA, Tanaka JC, Komáromy AM. Evaluation of a behavioral method for objective vision testing and identification of achromatopsia in dogs. American Journal of Veterinary Research. 2010;71(1):97-102.

444. Annear MJ, Gornik KR, Venturi FL, Hauptman JG, Bartoe JT, Petersen‐Jones SM. Reproducibility of an objective four‐choice canine vision testing technique that assesses vision at differing light intensities. Veterinary Ophthalmology. 2013;16(5):324-328.

445. Neitz J, Geist T, Jacobs GH. Color vision in the dog. Visual Neuroscience. 1989;3(02):119-125.

446. Pretterer G, Bubna-Littitz H, Windischbauer G, Gabler C, Griebel U. Brightness discrimination in the dog. Journal of Vision. 2004;4(3):10-10.

447. Ekesten B, Komáromy AM, Ofri R, Petersen-Jones SM, Narfström K. Guidelines for clinical electroretinography in the dog: 2012 update. Documenta Ophthalmologica. 2013;127(2):79- 87.

448. Willis CK, Quinn RP, McDonell WM, Gati J, Parent J, Nicolle D. Functional MRI as a tool to assess vision in dogs: the optimal anesthetic. Veterinary Ophthalmology. 2001;4(4):243-253.

449. Dobson V, Quinn GE, Tung B, Palmer EA, Reynolds JD. Comparison of recognition and grating acuities in very-low-birth-weight children with and without retinal residua of retinopathy of prematurity. Cryotherapy for Retinopathy of Prematurity Cooperative Group. Investigative ophthalmology & visual science. 1995;36(3):692-702.

450. Mayer DL, Dobson V. Visual acuity development in infants and young children, as assessed by operant preferential looking. Vision Research. 1982;22(9):1141-1151.

451. Kushner BJ, Lucchese NJ, Morton GV. Grating visual acuity with Teller cards compared with Snellen visual acuity in literate patients. Archives of Ophthalmology. 1995;113(4):485-493.

452. Mash C, Dobson V. Long-term reliability and predictive validity of the Teller Acuity Card procedure. Vision Research. 1998;38(4):619-626.

453. Loewenfeld I. The Pupil: Anatomy, Physiology, and Clinical Applications. Ames, Iowa: Iowa State University Press; 1993.

454. Lucas R, Douglas R, Foster R. Characterization of an ocular photopigment capable of driving pupillary constriction in mice. Nature Neuroscience. 2001;4(6):621-626.

455. Hattar S, Liao H, Takao M, Berson D, Yau K. Melanopsin-containing retinal ganglion cells: architecture, projections, and intrinsic photosensitivity. Science. 2002;295:1065-1070.

P a g e | 263

456. Berson D, Dunn F, Takao M. Phototransduction by retinal ganglion cells that set the circadian clock. Science. 2002;295:1070-1073.

457. Feigl B, Mattes D, Thomas R, Zele AJ. Intrinsically photosensitive (melanopsin) retinal ganglion cell function in glaucoma. Investigative Ophthalmology & Visual Science. 2011;52(7):4362-4367.

458. Nissen C, Sander B, Milea D, Kolko M, Herbst K, Hamard P, Lund-Andersen H. Monochromatic pupillometry in unilateral glaucoma discloses no adaptive changes subserved by the ipRGCs. Frontiers in Neurology. 2014;5(15):1-5.

459. Kankipati L, Girkin CA, Gamlin PD. The post-illumination pupil response is reduced in glaucoma patients. Investigative Ophthalmology & Visual Science. 2011;52(5):2287-2292.

460. Adhikari P, Zele AJ, Thomas R, Feigl B. Quadrant field pupillometry detects melanopsin dysfunction in glaucoma suspects and early glaucoma. Scientific Reports. 2016;6:33373.

461. Grozdanic SD, Matic M, Sakaguchi DS, Kardon RH. Evaluation of retinal status using chromatic pupil light reflex activity in healthy and diseased canine eyes. Investigative Ophthalmology & Visual Science. 2007;48(11):5178-5183.

462. Whiting RE, Yao G, Narfström K, Pearce JW, Coates JR, Dodam JR, Castaner LJ, Katz ML. Quantitative assessment of the canine pupillary light reflex. Investigative Ophthalmology & Visual Science. 2013;54(8):5432-5440.

463. Whiting RE, Narfström K, Yao G, Pearce JW, Coates JR, Castaner LJ, Katz ML. Pupillary light reflex deficits in a canine model of late infantile neuronal ceroid lipofuscinosis. Experimental Eye Research. 2013;116:402-410.

464. Whiting RE, Pearce JW, Castaner LJ, Jensen CA, Katz RJ, Gilliam DH, Katz ML. Multifocal retinopathy in Dachshunds with CLN2 neuronal ceroid lipofuscinosis. Experimental Eye Research. 2015;134:123-132.

465. Kim J, Heo J, Ji D, Kim M-S. Quantitative assessment of pupillary light reflex in normal and anesthetized dogs: a preliminary study. Journal of Veterinary Medical Science. 2015;77(4):475-478.

466. Grozdanic SD, Kecova H, Lazic T. Rapid diagnosis of retina and optic nerve abnormalities in canine patients with and without cataracts using chromatic pupil light reflex testing. Veterinary Ophthalmology. 2013;16(5):329-340.

467. Grozdanic SD, Lazic T, Kecova H, Mohan K, Kuehn MH. Optical coherence tomography and molecular analysis of sudden acquired retinal degeneration syndrome (SARDS) eyes suggests the immune‐mediated nature of retinal damage. Veterinary Ophthalmology. 2019;22(3):305- 327.

P a g e | 264

468. Komáromy AM, Bras D, Esson DW, Fellman RL, Grozdanic SD, Kagemann L, Miller PE, Moroi SE, Plummer CE, Sapienza JS, Storey ES, Teixeira LB, Toris CB, Webb TR. The future of canine glaucoma therapy. Veterinary Ophthalmology. 2019;10.1111/vop.12678.

469. Miller PE, Bentley E. Clinical signs and diagnosis of the canine primary glaucomas. Veterinary Clinics: Small Animal Practice. 2015;45(6):1183-1212.

470. WS R. ImageJ. 1997-2018; https://imagej.nih.gov/ij/.

471. Yeh CY, Koehl KL, Harman CD, Iwabe S, Guzman JM, Petersen-Jones SM, Kardon RH, Komáromy AM. Assessment of rod, cone, and intrinsically photosensitive retinal ganglion cell contributions to the canine chromatic pupillary response. Investigative Ophthalmology & Visual Science. 2017;58(1):65-78.

472. Komáromy AM, Abrams KL, Heckenlively JR, Lundy SK, Maggs DJ, Leeth CM, MohanKumar PS, Petersen‐Jones SM, Serreze DV, van der Woerdt A. Sudden acquired retinal degeneration syndrome (SARDS)–a review and proposed strategies toward a better understanding of pathogenesis, early diagnosis, and therapy. Veterinary Ophthalmology. 2016;19(4):319-331.

473. Young WM, Oh A, Williams JG, Foster ML, Miller WW, Lunn KF, Mowat FM. Clinical therapeutic efficacy of mycophenolate mofetil in the treatment of SARDS in dogs—a prospective open‐label pilot study. Veterinary Ophthalmology. 2018;21(6):565-576.

474. Güler AD, Ecker JL, Lall GS, Haq S, Altimus CM, Liao H-W, Barnard AR, Cahill H, Badea TC, Zhao H. Melanopsin cells are the principal conduits for rod–cone input to non-image-forming vision. Nature. 2008;453(7191):102.

475. Najjar RP, Sharma S, Atalay E, Rukmini AV, Sun C, Lock JZ, Baskaran M, Perera SA, Husain R, Lamoureux E. Pupillary responses to full-field chromatic stimuli are reduced in patients with early-stage primary open-angle glaucoma. Ophthalmology. 2018;125(9):1362-1371.

476. Terakado K, Yogo T, Nezu Y, Harada Y, Hara Y, Tagawa M. Efficacy of the use of colorimetric pupil light reflex device in the diagnosis of fundus disease or optic pathway disease in dogs. Journal of Veterinary Medical Science. 2013;75(11):1491-1495.

477. Gunderson EG, Lukasik VM, Ashton MM, Merideth RE, Madsen R. Effects of anesthetic induction with midazolam-propofol and midazolam-etomidate on selected ocular and cardiorespiratory variables in clinically normal dogs. American Journal of Veterinary Research. 2013;74(4):629-635.

478. Stephan DD, Vestre W, Stiles J, Krohne S. Changes in intraocular pressure and pupil size following intramuscular administration of hydromorphone hydrochloride and acepromazine in clinically normal dogs. Veterinary Ophthalmology. 2003;6(1):73-76.

P a g e | 265

479. Marchini G, Pagliarusco A, Toscano A, Tosi R, Brunelli C, Bonomi L. Ultrasound biomicroscopic and conventional ultrasonographic study of ocular dimensions in primary angle-closure glaucoma. Ophthalmology. 1998;105(11):2091-2098.

480. Sihota R, Dada T, Gupta R, Lakshminarayan P, Pandey RM. Ultrasound biomicroscopy in the subtypes of primary angle closure glaucoma. Journal of Glaucoma. 2005;14(5):387-391.

481. Mochizuki H, Takenaka J, Sugimoto Y, Takamatsu M, Kiuchi Y. Comparison of the prevalence of plateau iris configurations between angle-closure glaucoma and open-angle glaucoma using ultrasound biomicroscopy. Journal of Glaucoma. 2011;20(5):315-318.

482. Wang B, Narayanaswamy A, Amerasinghe N, Zheng C, He M, Chan Y, Nongpiur M, Friedman D, Aung T. Increased iris thickness and association with primary angle closure glaucoma. British Journal of Ophthalmology. 2011;95(1):46-50.

483. Devereux JG, Foster PJ, Baasanhu J, Uranchimeg D, Lee P-S, Erdenbeleig T, Machin D, Johnson GJ, Alsbirk PH. Anterior chamber depth measurement as a screening tool for primary angle-closure glaucoma in an East Asian population. Archives of Ophthalmology. 2000;118(2):257-263.

484. Graham KL, McCowan CI, Caruso K, Billson FM, Whittaker CJG, White A. Optical coherence tomography of the retina, nerve fiber layer and optic nerve head in dogs with glaucoma. Veterinary Ophthalmology. 2019;10.1111/vop.12694

485. Pavlin CJ, Foster FS. Ultrasound Biomicroscopy of the Eye. New York: Springer-Verlag; 1995.

486. Dulaurent T, Goulle F, Dulaurent A, Mentek M, Peiffer RL, Isard PF. Effect of mydriasis induced by topical instillations of 0.5% tropicamide on the anterior segment in normotensive dogs using ultrasound biomicroscopy. Veterinary Ophthalmology. 2012;15(s1):8-13.

487. Rose MD, Mattoon JS, Gemensky-Metzler AJ, Wilkie DA, Rajala-Schultz PJ. Ultrasound biomicroscopy of the iridocorneal angle of the eye before and after phacoemulsification and intraocular lens implantation in dogs. American Journal of Veterinary Research. 2008;69(2):279-288.

488. Samuelson DA, Gelatt KN. Aqueous outflow in the beagle. I. Postnatal morphologic development of the iridocorneal angle: pectinate ligament and uveal trabecular meshwork. Current Eye Research. 1984;3(6):783-794.

489. Samuelson DA, Gelatt KN. Aqueous outflow in the beagle. II. Postnatal morphologic development of the iridocorneal angle: corneoscleral trabecular mesh work and angular aqueous plexus. Current Eye Research. 1984;3(6):795-808.

P a g e | 266

490. Park Y-W, Jeong M-B, Lee ER, Lee Y, Ahn J-S, Kim S-H, Seo K. Acute changes in central corneal thickness according to experimental adjustment of intraocular pressure in normal canine eyes. The Journal of Veterinary Medical Science. 2013;75(11):1479-1483.

491. Garzón‐Ariza A, Guisado A, Galán A, Martín‐Suárez E. Diurnal variations in intraocular pressure and central corneal thickness and the correlation between these factors in dogs. Veterinary Ophthalmology. 2017;DOI:10.1111/vop.12533:1-7.

492. Bentley E, Miller PE, Diehl KA. Evaluation of intra-and interobserver reliability and image reproducibility to assess usefulness of high-resolution ultrasonography for measurement of anterior segment structures of canine eyes. American Journal of Veterinary Research. 2005;66(10):1775-1779.

493. Pavlin CJ, Harasiewicz K, Foster FS. Ultrasound biomicroscopy of anterior segment structures in normal and glaucomatous eyes. American Journal of Ophthalmology. 1992;113(4):381- 389.

494. ImageJ [computer program]. Version November 2018. https://imagej.net/ImageJ2018.

495. Hasegawa T, Kawata M, Ota M. Ultrasound biomicroscopic findings of the iridocorneal angle in live healthy and glaucomatous dogs. Journal of Veterinary Medical Science. 2016;77(12):1625-1631.

496. Grozdanic SD, Matic M, Betts DM, Sakaguchi DS, Kardon RH. Recovery of canine retina and optic nerve function after acute elevation of intraocular pressure: implications for canine glaucoma treatment. Veterinary Ophthalmology. 2007;10(s1):101-107.

497. Dubin A, Bentley E, Buhr K, Miller PE. Evaluation of potential risk factors for development of primary angle-closure glaucoma in Bouviers des Flandres. Journal of the American Veterinary Medical Association. 2017;250(1):60-67.

498. Kawata M, Tsukizawa H, Nakayama M, Hasegawa T. Rectification of width and area of the ciliary cleft in dogs. Journal of Veterinary Medical Science. 2010;72(5):533-537.

499. Nongpiur ME, Khor CC, Jia H, Cornes BK, Chen L-J, Qiao C, Nair KS, Cheng C-Y, Xu L, George R. ABCC5, a gene that influences the anterior chamber depth, is associated with primary angle closure glaucoma. PLoS Genetics. 2014;10(3):e1004089.

500. Alsbirk PH. Primary angle-closure glaucoma. Oculometry, epidemiology, and genetics in a high risk population. Acta Ophthalmologica Supplementum. 1976(127):5-31.

501. Congdon NG, Quigley HA, Hung PT, Wang T, Ho T. Screening techniques for angle‐closure glaucoma in rural Taiwan. Acta Ophthalmologica Scandinavica. 1996;74(2):113-119.

P a g e | 267

502. Ekesten B. Correlation of intraocular distances to the iridocorneal angle in Samoyeds with special reference to angle-closure glaucoma. Progress in Veterinary and Comparative Ophthalmology. 1993;3:67-73.

503. Yücel YH, Zhang Q, Weinreb RN, Kaufman PL, Gupta N. Atrophy of relay neurons in magno- and parvocellular layers in the lateral geniculate nucleus in experimental glaucoma. Investigative Ophthalmology & Visual Science. 2001;42(13):3216-3222.

504. Yücel YH, Zhang Q, Gupta N, Kaufman PL, Weinreb RN. Loss of neurons in magnocellular and parvocellular layers of the lateral geniculate nucleus in glaucoma. Archives of Ophthalmology. 2000;118(3):378-384.

505. Weber AJ, Chen H, Hubbard WC, Kaufman PL. Experimental glaucoma and cell size, density, and number in the primate lateral geniculate nucleus. Investigative Ophthalmology & Visual Science. 2000;41(6):1370-1379.

506. Su JH, Deng G, Cotman CW. Transneuronal degeneration in the spread of Alzheimer's disease pathology: immunohistochemical evidence for the transmission of tau hyperphosphorylation. Neurobiology of Disease. 1997;4(5):365-375.

507. Kiernan JA, Hudson AJ. Changes in sizes of cortical and lower motor neurons in amyotrophic lateral sclerosis. Brain. 1991;114(2):843-853.

508. Conti AC, Raghupathi R, Trojanowski JQ, McIntosh TK. Experimental brain injury induces regionally distinct apoptosis during the acute and delayed post-traumatic period. Journal of Neuroscience. 1998;18(15):5663-5672.

509. Sidek S, Ramli N, Rahmat K, Ramli N, Abdulrahman F, Tan L. Glaucoma severity affects diffusion tensor imaging (DTI) parameters of the optic nerve and optic radiation. European Journal of Radiology. 2014;83(8):1437-1441.

510. Zhang Q-J, Wang D, Bai Z-L, Ren B-C, Li X-H. Diffusion tensor imaging of optic nerve and optic radiation in primary chronic angle-closure glaucoma using 3T magnetic resonance imaging. International Journal of Ophthalmology. 2015;8(5):975.

511. Michelson G, Engelhorn T, Wärntges S, El Rafei A, Hornegger J, Doerfler A. DTI parameters of axonal integrity and demyelination of the optic radiation correlate with glaucoma indices. Graefe's Archive for Clinical and Experimental Ophthalmology. 2013;251(1):243-253.

512. Garaci FG, Bolacchi F, Cerulli A, Melis M, Spano A, Cedrone C, Floris R, Simonetti G, Nucci C. Optic nerve and optic radiation neurodegeneration in patients with glaucoma: in vivo analysis with 3-T diffusion-tensor MR imaging. Radiology. 2009;252(2):496-501.

P a g e | 268

513. El-Rafei A, Engelhorn T, Wärntges S, Dörfler A, Hornegger J, Michelson G. Glaucoma classification based on visual pathway analysis using diffusion tensor imaging. Magnetic Resonance Imaging. 2013;31(7):1081-1091.

514. Engelhorn T, Michelson G, Waerntges S, Hempel S, El-Rafei A, Struffert T, Doerfler A. A new approach to assess intracranial white matter abnormalities in glaucoma patients: changes of fractional anisotropy detected by 3T diffusion tensor imaging. Academic Radiology. 2012;19(4):485-488.

515. Engelhorn T, Michelson G, Waerntges S, Otto M, El-Rafei A, Struffert T, Doerfler A. Changes of radial diffusivity and fractional anisotopy in the optic nerve and optic radiation of glaucoma patients. The Scientific World Journal. 2012;2012:849632.

516. Michelson G, Engelhorn T, Waerntges S, Doerfler A. Diffusion tensor imaging for in vivo detection of degenerated optic radiation. ISRN Ophthalmology. 2011;2011:648450.

517. Murai H, Suzuki Y, Kiyosawa M, Tokumaru AM, Ishii K, Mochizuki M. Positive correlation between the degree of visual field defect and optic radiation damage in glaucoma patients. Japanese Journal of Ophthalmology. 2013;57(3):257-262.

518. Johnson PJ. Advanced neuroimaging in veterinary science. In: Hagen R, Martig S, Raw M-E, Allison Z, eds. EAVDI Yearbook 2015. Reviews in Veterinary Diagnostic Imaging. Cambridge, United Kingdom: EAVDI Ltd.; 2015:1-14.

519. Takahashi M, Ono J, Harada K, Maeda M, Hackney DB. Diffusional anisotropy in cranial nerves with maturation: quantitative evaluation with diffusion MR imaging in rats. Radiology. 2000;216(3):881-885.

520. Assaf Y, Pasternak O. Diffusion tensor imaging (DTI)-based white matter mapping in brain research: a review. Journal of Molecular Neuroscience. 2008;34(1):51-61.

521. Kuchling J, Brandt AU, Paul F, Scheel M. Diffusion tensor imaging for multilevel assessment of the visual pathway: possibilities for personalized outcome prediction in autoimmune disorders of the central nervous system. EPMA Journal. 2017;8(3):279-294.

522. Wang M-Y, Wu K, Xu J-M, Dai J, Qin W, Liu J, Tian J, Shi D. Quantitative 3-T diffusion tensor imaging in detecting optic nerve degeneration in patients with glaucoma: association with retinal nerve fiber layer thickness and clinical severity. Neuroradiology. 2013;55(4):493-498.

523. Barry EF, Cerda‐Gonzalez S, Luh WM, Daws RE, Raj A, Johnson PJ. Normal diffusivity of the domestic feline brain. Journal of Comparative Neurology. 2019;527(5):1012-1023.

524. Zhang Y-y, Tang W-j, Song X-y, Puyang Z, Chen A, Zhao J, Li X-j, Chen Y-y. Diffusion Tensor Imaging Detects Microstructural Differences of Visual Pathway in Patients with Primary

P a g e | 269

Open Angle Glaucoma and Ocular Hypertension. Frontiers in Human Neuroscience. 2018;12:426.

525. Xu Z, Sun J, Zhang X, Feng Y, Pan A, Gao M, Zhao H. Microstructural visual pathway abnormalities in patients with primary glaucoma: 3 T diffusion kurtosis imaging study. Clinical Radiology. 2018;73(6):591. e599-591. e515.

526. You Y, Joseph C, Wang C, Gupta V, Liu S, Yiannikas C, Chua BE, Chitranshi N, Shen T, Dheer Y. Demyelination precedes axonal loss in the transneuronal spread of human neurodegenerative disease. Brain. 2019;142(2):426-442.

527. Yang X-L, van der Merwe Y, Sims J, Parra C, Ho LC, Schuman JS, Wollstein G, Lathrop KL, Chan KC. Age-related changes in eye, brain and visuomotor behavior in the DBA/2J mouse model of chronic glaucoma. Scientific Reports. 2018;8(1):4643.

528. Chan KC, Yu Y, Ng SH, Mak HK, Yip YW, van der Merwe Y, Ren T, Yung JS, Biswas S, Cao X. Intracameral injection of a chemically cross-linked hydrogel to study chronic neurodegeneration in glaucoma. Acta Biomaterialia. 2019;94:219-231.

529. Zhang X, Sun P, Wang J, Wang Q, Song S-K. Diffusion tensor imaging detects retinal ganglion cell axon damage in the mouse model of optic nerve crush. Investigative Ophthalmology & Visual Science. 2011;52(9):7001-7006.

530. Ho LC, Wang B, Conner IP, van der Merwe Y, Bilonick RA, Kim S-G, Wu EX, Sigal IA, Wollstein G, Schuman JS. In vivo evaluation of white matter integrity and anterograde transport in visual system after excitotoxic retinal injury with multimodal MRI and OCT. Investigative Ophthalmology & Visual Science. 2015;56(6):3788-3800.

531. Zhong Y-F, Tang Z-H, Qiang J-W, Wu L-J, Wang R, Wang J, Jin L-X, Xiao Z-B. Changes in DTI parameters in the optic tracts of macaque monkeys with monocular blindness. Neuroscience Letters. 2017;636:248-253.

532. Chen Z, Lin F, Wang J, Li Z, Dai H, Mu K, Ge J, Zhang H. Diffusion tensor magnetic resonance imaging reveals visual pathway damage that correlates with clinical severity in glaucoma. Clinical and Experimental Ophthalmology. 2013;41(1):43-49.

533. Anaya García MS, Hernández Anaya JS, Marrufo Meléndez O, Velázquez Ramírez JL, Palacios Aguiar R. In Vivo study of cerebral white matter in the dog using diffusion tensor tractography. Veterinary Radiology & Ultrasound. 2015;56(2):188-195.

534. Jacqmot O, Van Thielen B, Fierens Y, Hammond M, Willekens I, Schuerbeek PV, Verhelle F, Goossens P, De Ridder F, Clarys JP. Diffusion tensor imaging of white matter tracts in the dog brain. The Anatomical Record. 2013;296(2):340-349.

P a g e | 270

535. Jacqmot O, Van Thielen B, Michotte A, Willekens I, Verhelle F, Goossens P, De Ridder F, Clarys JP, Vanbinst A, Peleman C. Comparison of Several White Matter Tracts in Feline and Canine Brain by Using Magnetic Resonance Diffusion Tensor Imaging. The Anatomical Record. 2017;300(7):1270-1289.

536. Leong D, Calabrese E, White LE, Wei P, Chen S, Platt SR, Provenzale JM. Correlation of diffusion tensor imaging parameters in the canine brain. The Neuroradiology Journal. 2015;28(1):11-18.

537. Alfaro-Almagro F, Jenkinson M, Bangerter NK, Andersson JL, Griffanti L, Douaud G, Sotiropoulos SN, Jbabdi S, Hernandez-Fernandez M, Vallee E. Image processing and quality control for the first 10,000 brain imaging datasets from UK Biobank. Neuroimage. 2018;166:400-424.

538. Behrens TE, Woolrich MW, Jenkinson M, Johansen‐Berg H, Nunes RG, Clare S, Matthews PM, Brady JM, Smith SM. Characterization and propagation of uncertainty in diffusion‐weighted MR imaging. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2003;50(5):1077-1088.

539. Van Hecke W, Emsell L, Sunaert S. Diffusion tensor imaging: A practical handbook.10.1007/978-1-4939-3118-7. New York: Springer; 2016.

540. Bemis AM, Pirie CG, LoPinto AJ, Maranda L. Reproducibility and repeatability of optical coherence tomography imaging of the optic nerve head in normal beagle eyes. Veterinary Ophthalmology. 2017;20(6):480-487.

541. Hernandez‐Merino E, Kecova H, Jacobson SJ, Hamouche KN, Nzokwe RN, Grozdanic SD. Spectral domain optical coherence tomography (SD‐OCT) assessment of the healthy female canine retina and optic nerve. Veterinary Ophthalmology. 2011;14(6):400-405.

542. Osinchuk SC, Leis ML, Salpeter EM, Sandmeyer LS, Grahn BH. Evaluation of retinal morphology of canine sudden acquired retinal degeneration syndrome using optical coherence tomography and fluorescein angiography. Veterinary Ophthalmology. 2018;doi.org/10.1111/vop.12602.

543. Beltran WA, Cideciyan AV, Guziewicz KE, Iwabe S, Swider M, Scott EM, Savina SV, Ruthel G, Stefano F, Zhang L. Canine retina has a primate fovea-like bouquet of cone photoreceptors which is affected by inherited macular degenerations. PloS One. 2014;9(3):e90390.

544. Iwabe S, Ying G-S, Aguirre GD, Beltran WA. Assessment of visual function and retinal structure following acute light exposure in the light sensitive T4R rhodopsin mutant dog. Experimental Eye Research. 2016;146:341-353.

545. Beltran W, Acland G, Aguirre G. Morphologic analysis of retinal disease in the area centralis of RPGR-ORF15 mutant dogs. Investigative Ophthalmology & Visual Science. 2007;48(13):3718-3718.

P a g e | 271

546. Rodarte‐Almeida ACV, Petersen‐Jones S, Langohr IM, Occelli L, Dornbusch PT, Shiokawa N, Montiani‐Ferreira F. Retinal dysplasia in American pit bull terriers–phenotypic characterization and breeding study. Veterinary Ophthalmology. 2016;19(1):11-21.

547. Teixeira LB, Ver Hoeve JN, Mayer JA, Dubielzig RR, Smith CM, Radcliff AB, Duncan ID. Modeling the chronic loss of optic nerve axons and the effects on the retinal nerve fiber layer structure in primary disorder of myelin. Investigative Ophthalmology & Visual Science. 2016;57(11):4859-4868.

548. Gekeler F, Gmeiner H, Völker M, Sachs H, Messias A, Eule C, Bartz‐Schmidt KU, Zrenner E, Shinoda K. Assessment of the posterior segment of the cat eye by optical coherence tomography (OCT). Veterinary Ophthalmology. 2007;10(3):173-178.

549. Espinheira Gomes F, Parry S, Ledbetter E. Spectral domain optical coherence tomography evaluation of the feline optic nerve and peripapillary retina. Veterinary Ophthalmology. 2019; 10.1111/vop.12633(Epub ahead of print):1-10.

550. Gonzalez‐Alonso‐Alegre EM, Rodriguez‐Alvaro A, Esteban‐Martín J. Atypical chorioretinal coloboma in a Golden Retriever: a retinographic, fluoroangiographic, and optical coherence tomography study. Veterinary Ophthalmology. 2016;19(6):525-530.

551. Gornik KR, Pirie CG, Duker JS, Boudrieau RJ. Canine multifocal retinopathy caused by a BEST1 mutation in a Boerboel. Veterinary Ophthalmology. 2014;17(5):368-372.

552. Hoffmann I, Guziewicz KE, Zangerl B, Aguirre GD, Mardin CY. Canine multifocal retinopathy in the Australian Shepherd: a case report. Veterinary Ophthalmology. 2012;15:134-138.

553. Schaefer EA, Whiting RE, Pearce JW, Grahn BH, Hamm CW, Moore CP, Narfström KL. Bilateral retinoschisis in a dog: A veterinary clinical application for optical coherence tomography. Veterinary Ophthalmology. 2018;21(6):668-674.

554. Komáromy AM, Bras D, Esson DW, Fellman RL, Grozdanic SD, Kagemann L, Miller PE, Moroi SE, Plummer CE, Sapienza JS, Storey ES, Teixeira LB, Toris CB, Webb TR. The future of canine glaucoma therapy. Veterinary Ophthalmology. 2019;Early View.

555. Lee I, Kim J, Lee C. Anatomical characteristics and three‐dimensional model of the dog dorsal lateral geniculate body. The Anatomical Record: An Official Publication of the American Association of Anatomists. 1999;256(1):29-39.

556. Song S-K, Sun S-W, Ramsbottom MJ, Chang C, Russell J, Cross AH. Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. Neuroimage. 2002;17(3):1429-1436.

P a g e | 272

557. El-Rafei A. Diffusion Tensor Imaging Analysis of the Visual Pathway with Application to Glaucoma [PhD]. Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU): Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU); 2012.

558. Wu Y-C, Field AS, Duncan ID, Samsonov AA, Kondo Y, Tudorascu D, Alexander AL. High b- value and diffusion tensor imaging in a canine model of dysmyelination and brain maturation. Neuroimage. 2011;58(3):829-837.

559. Lawlor M, Danesh-Meyer H, Levin LA, Davagnanam I, De Vita E, Plant GT. Glaucoma and the brain: trans-synaptic degeneration, structural change, and implications for neuroprotection. Survey of Ophthalmology. 2018;63(3):296-306.

560. Goldby F. A note on transneuronal atrophy in the human lateral geniculate body. Journal of Neurology, Neurosurgery, and Psychiatry. 1957;20(3):202.

561. Matthews MR, Cowan W, Powell T. Transneuronal cell degeneration in the lateral geniculate nucleus of the macaque monkey. Journal of Anatomy. 1960;94(Pt 2):145.

562. Matthews MR. Further observations on transneuronal degeneration in the lateral geniculate nucleus of the macaque monkey. Journal of Anatomy. 1964;98(Pt 2):255.

563. Rossiter JP. Trans-synaptic degeneration of lateral geniculate nuclei following remote loss of right eye. JAMA Ophthalmology. 2015;133(1):e141789-e141789.

564. Pierpaoli C, Jezzard P, Basser PJ, Barnett A, Di Chiro G. Diffusion tensor MR imaging of the human brain. Radiology. 1996;201(3):637-648.

565. Alexander AL, Hasan KM, Lazar M, Tsuruda JS, Parker DL. Analysis of partial volume effects in diffusion‐tensor MRI. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2001;45(5):770-780.

566. Jbabdi S. Imaging structure and function. In: Johansen-Berg H, Behrens T, eds. Diffusion MRI. Amsterdam: Elsevier Science; 2014:585-605.

567. Ashburner J, Friston KJ. Unified segmentation. Neuroimage. 2005;26(3):839-851.

568. Johansen-Berg H, Behrens TE. Diffusion MRI: from quantitative measurement to in vivo neuroanatomy. Amsterdam: Elsevier Science; 2013.

569. Liu C, Frank QY, Yen CC-C, Newman JD, Glen D, Leopold DA, Silva AC. A digital 3D atlas of the marmoset brain based on multi-modal MRI. Neuroimage. 2018;169:106-116.

P a g e | 273

570. Seidlitz J, Sponheim C, Glen D, Frank QY, Saleem KS, Leopold DA, Ungerleider L, Messinger A. A population MRI brain template and analysis tools for the macaque. Neuroimage. 2018;170:121-131.

571. Datta R, Lee J, Duda J, Avants BB, Vite CH, Tseng B, Gee JC, Aguirre GD, Aguirre GK. A digital atlas of the dog brain. PLoS One. 2012;7(12):e52140.

572. Milne ME, Steward C, Firestone SM, Long SN, O'Brien TJ, Moffat BA. Development of representative magnetic resonance imaging–based atlases of the canine brain and evaluation of three methods for atlas-based segmentation. American Journal of Veterinary Research. 2016;77(4):395-403.

573. Nitzsche B, Boltze J, Ludewig E, Flegel T, Schmidt MJ, Seeger J, Barthel H, Brooks OW, Gounis MJ, Stoffel MH. A stereotaxic breed-averaged, symmetric T2w canine brain atlas including detailed morphological and volumetrical data sets. Neuroimage. 2019;187:93-103.

574. Boillot T, Rosolen SG, Dulaurent T, Goulle F, Thomas P, Isard P-F, Azoulay T, Lafarge-Beurlet S, Woods M, Lavillegrand S, Ivkovic I, Neveux N, Sahel J-A, Picaud S, Froger N. Determination of morphological, biometric and biochemical susceptibilities in healthy Eurasier dogs with suspected inherited glaucoma. PLoS One. 2014;9(11):e111973.

575. Gibson T, Roberts S, Severin G, Steyn P, Wrigley R. Comparison of gonioscopy and ultrasound biomicroscopy for evaluating the iridocorneal angle in dogs. Journal of the American Veterinary Medical Association. 1998;213(5):635-638.

576. Grozdanic SD, Kecova H, Harper M, Nilaweera W, Kuehn M, Kardon R. Functional and structural changes in a canine model of hereditary primary angle-closure glaucoma. Investigative Ophthalmology & Visual Science. 2010;51(1):255-263.

577. Graham KL, McCowan CI, Caruso K, Billson FM, Whittaker CJ, White AJ. Optical coherence tomography of the retina, nerve fiber layer and optic nerve head in dogs with glaucoma. Veterinary Ophthalmology. 2019;10.1111/vop.12694.

578. Tanaka T, Ikeuchi E, Mitani S, EGUCHI Y, UETAKE K. Studies on the visual acuity of dogs using shape discrimination learning. Nihon Chikusan Gakkaiho. 2000;71(6):614-620.

579. Graham KL, Reid J, Whittaker CJ, Hall EJ, Caruso K, McCowan CI, White AJ. Development of a vision impairment score for the assessment of functional vision in dogs: initial evidence of validity, reliability, and responsiveness. Veterinary Ophthalmology. 2019;10.1111/vop.12656.

580. Graham K, Byosiere SE, Feng L, Sanders M, Bennett P, Caruso K, McCowan C, White A. A forced‐choice preferential looking task for the assessment of vision in dogs: pilot study. Journal of Small Animal Practice. 2019;60(6):340-347.

P a g e | 274

581. Ahonen SJ, Kaukonen M, Nussdorfer FD, Harman CD, Komáromy AM, Lohi H. A novel missense mutation in ADAMTS10 in Norwegian Elkhound primary glaucoma. PLoS One. 2014;9(11):e111941.

582. Ahram DF, Cook AC, Kecova H, Grozdanic SD, Kuehn MH. Identification of genetic loci associated with primary angle-closure glaucoma in the basset hound. Molecular VIsion. 2014;20:497-510.

583. Boote C, Palko JR, Sorensen T, Mohammadvali A, Elsheikh A, Komáromy AM, Pan X, Liu J. Changes in posterior scleral collagen microstructure in canine eyes with an ADAMTS10 mutation. Molecular Vision. 2016;22:503-517.

584. Kanemaki N, Tchedre K, Imayasu M, Kawarai S, Sakaguchi M, Yoshino A, Itoh N, Meguro A, Mizuki N. Dogs and humans share a common susceptibility gene SRBD1 for glaucoma risk. PLoS One. 2013;8(9):e74372.

585. Kuchtey J, Olson LM, Rinkoski T, MacKay EO, Iverson T, Gelatt K, Haines J, Kuchtey R. Mapping of the disease locus and identification of ADAMTS10 as a candidate gene in a canine model of primary open angle glaucoma. PLoS Genetics. 2011;7(2):e1001306.

586. Mackay E, Källberg M, Gelatt K. Aqueous humor myocilin protein levels in normal, genetic carriers, and glaucoma Beagles. Veterinary Ophthalmology. 2008;11(3):177-185.

587. Oliver JA, Ricketts S, Kuehn MH, Mellersh CS. Primary closed angle glaucoma in the Basset Hound: Genetic investigations using genome-wide association and RNA sequencing strategies. Molecular Vision. 2019;25:93-105.

588. Oliver JA, Rustidge S, Pettitt L, Jenkins CA, Farias FH, Giuliano EA, Mellersh CS. Evaluation of ADAMTS17 in Chinese Shar‐Pei with primary open‐angle glaucoma, primary lens luxation, or both. American Journal of Veterinary Research. 2018;79(1):98-106.

589. Palko JR, Iwabe S, Pan X, Agarwal G, Komáromy AM, Liu J. Biomechanical properties and correlation with collagen solubility profile in the posterior sclera of canine eyes with an ADAMTS10 mutation. Investigative Ophthalmology & Visual Science. 2013;54(4):2685-2695.

590. Pugh C, Farrell L, Carlisle A, Bush S, Ewing A, Trejo-Reveles V, Matika O, de Kloet A, Walsh C, Bishop S, Prendergast J, Rainger J, Schoenebeck J, Summers K. Arginine to Glutamine Variant in Olfactomedin Like 3 (OLFML3) Is a Candidate for Severe Goniodysgenesis and Glaucoma in the Border Collie Dog Breed. G3. 2019;9(3):943-954.

591. Grus FH, Sabuncuo P, Augustin AJ. Analysis of tear protein patterns of dry‐eye patients using fluorescent staining dyes and two‐dimensional quantification algorithms. Electrophoresis. 2001;22(9):1845-1850.

P a g e | 275

592. Willcox M, Morris C, Thakur A, Sack R, Wickson J, Boey W. Complement and complement regulatory proteins in human tears. Investigative Ophthalmology & Visual Science. 1997;38(1):1-8.

593. Ohashi Y, Dogru M, Tsubota K. Laboratory findings in tear fluid analysis. Clinica Chimica Acta. 2006;369(1):17-28.

594. Pieragostino D, Agnifili L, Fasanella V, D'Aguanno S, Mastropasqua R, Di Ilio C, Sacchetta P, Urbani A, Del Boccio P. Shotgun proteomics reveals specific modulated protein patterns in tears of patients with primary open angle glaucoma naive to therapy. Molecular BioSystems. 2013;9(6):1108-1116.

595. Tiffany J. Tears in health and disease. Eye. 2003;17(8):923.

596. Sahay P, Rao A, Padhy D, Sarangi S, Das G, Reddy MM, Modak R. Functional activity of matrix metalloproteinases 2 and 9 in tears of patients with glaucoma. Investigative Ophthalmology & Visual Science. 2017;58(6):BIO106-BIO113.

597. Herber S, Grus FH, Sabuncuo P, Augustin AJ. Two‐dimensional analysis of tear protein patterns of diabetic patients. Electrophoresis. 2001;22(9):1838-1844.

598. Evans V, Vockler C, Friedlander M, Walsh B, Willcox MD. Lacryglobin in human tears, a potential marker for cancer. Clinical & Experimental Ophthalmology. 2001;29(3):161-163.

599. Böhm D, Keller K, Pieter J, Boehm N, Wolters D, Siggelkow W, Lebrecht A, Schmidt M, Kölbl H, Pfeiffer N. Comparison of tear protein levels in breast cancer patients and healthy controls using a de novo proteomic approach. Oncology Reports. 2012;28(2):429-438.

600. Tomosugi N, Kitagawa K, Takahashi N, Sugai S, Ishikawa I. Diagnostic potential of tear proteomic patterns in Sjögren's syndrome. Journal of Proteome Research. 2005;4(3):820- 825.

601. Salvisberg C, Tajouri N, Hainard A, Burkhard PR, Lalive PH, Turck N. Exploring the human tear fluid: discovery of new biomarkers in multiple sclerosis. Proteomics -Clinical Applications. 2014;8(3-4):185-194.

602. de Freitas Campos C, Cole N, Van Dyk D, Walsh B, Diakos P, Almeida D, Torrecilhas A, Luiz Laus J, Willcox M. Proteomic analysis of dog tears for potential cancer markers. Research in Veterinary Science. 2008;85(2):349-352.

603. Palko JR, Morris HJ, Pan X, Harman CD, Koehl KL, Gelatt KN, Plummer CE, Komáromy AM, Liu J. Influence of age on ocular biomechanical properties in a canine glaucoma model with ADAMTS10 mutation. PLoS One. 2016;11(6):e0156466.

P a g e | 276

604. Winiarczyk M, Winiarczyk D, Banach T, Adaszek L, Madany J, Mackiewicz J, Pietras-Ozga D, Winiarczyk S. Dog tear film proteome in-depth analysis. PLoS One. 2015;10(12):e00144242.

605. Roberts S, Erickson OF. Dog tear secretion and tear proteins. Journal of Small Animal Practice. 1962;3(1):1-5.

606. Green-Church KB, Nichols KK, Kleinholz NM, Zhang L, Nichols JJ. Investigation of the human tear film proteome using multiple proteomic approaches. Molecular Vision. 2008;14:456- 470.

607. Sebbag L, McDowell E, Hepner P, Mochel J. Effect of tear collection on lacrimal total protein content in dogs and cats: a comparison between Schirmer strips and ophthalmic sponges. BMC Veterinary Research. 2018;14(1):61.

608. Martinez P, Storey E, Pucheu-Haston C. Survey of cytokines in normal canine tears by multiplex analysis: a pilot study. Veterinary Immunology and Immunopathology. 2018;201:38-42.

609. Pieragostino D, Bucci S, Agnifili L, Fasanella V, D'Aguanno S, Mastropasqua A, Ciancaglini M, Mastropasqua L, Di Ilio C, Sacchetta P. Differential protein expression in tears of patients with primary open angle and pseudoexfoliative glaucoma. Molecular BioSystems. 2012;8(4):1017-1028.

610. Wong TT, Zhou L, Li J, Tong L, Zhao SZ, Li XR, Yu SJ, Koh SK, Beuerman RW. Proteomic profiling of inflammatory signaling molecules in the tears of patients on chronic glaucoma medication. Cornea. 2011;52(10):7385-7391.

611. Servat JJ, Bernardino CR. Effects of common topical antiglaucoma medications on the ocular surface, eyelids and periorbital tissue. Drugs & Aging. 2011;28(4):267-282.

612. Baudouin C. Detrimental effect of preservatives in eyedrops: implications for the treatment of glaucoma. Acta Ophthalmologica. 2008;86(7):716-726.

613. Ichijima H, WM P, JV J, HD C. Confocal microscopic studies of living rabbit cornea treated with benzalkonium chloride. Cornea. 1992;11(3):221-225.

614. Baudouin C, Labbé A, Liang H, Pauly A, Brignole-Baudouin F. Preservatives in eyedrops: the good, the bad and the ugly. Progress in Retinal and Eye Research. 2010;29(4):312-334.

615. Baudouin C, Pisella P, Fillacier K, Goldschild M, Becquet F, De Saint Jean M, Béchetoille A. Ocular surface inflammatory changes induced by topical antiglaucoma drugs: human and animal studies. Ophthalmology. 1999;106(3):556-563.

P a g e | 277

616. Kaštelan S, Tomić M, Soldo KM, Salopek-Rabatić J. How ocular surface disease impacts the glaucoma treatment outcome. BioMed Research International. 2013;2013(696328):1-7.

617. Johnson D, Yoshikawa K, Brubaker R, Hodge D. The effect of long-term medical therapy on the outcome of filtration surgery. American Journal of Ophthalmology. 1994;117(2):139-148.

618. Baudouin C. Ocular surface and external filtration surgery: mutual relationships. Developments in Ophthalmology. 2012;50:64-78.

619. Westermeyer H, Hendrix D, Ward DA. Long‐term evaluation of the use of Ahmed gonioimplants in dogs with primary glaucoma: nine cases (2000–2008). Journal of the American Veterinary Medical Association. 2011;238(5):610-617.

620. Graham KL, Donaldson D, Billson FA, Billson FM. Use of a 350‐mm2 Baerveldt glaucoma drainage device to maintain vision and control intraocular pressure in dogs with glaucoma: a retrospective study (2013–2016). Veterinary Ophthalmology. 2016;20(5):427-434.

621. Graham KL, Hall EJ, Caraguel C, White A, Billson FA, Billson FM. Comparison of diode laser trans‐scleral cyclophotocoagulation versus implantation of a 350‐mm2 Baerveldt glaucoma drainage device for the treatment of glaucoma in dogs (a retrospective study: 2010–2016). Veterinary Ophthalmology. 2018;21(5):487-497.

622. Bentley E, Miller PE, Murphy CJ, Schoster JV. Combined cycloablation and gonioimplantation for treatment of glaucoma in dogs: 18 cases (1992–1998). Journal of the American Veterinary Medical Association. 1999;215(10):1469-1472.

623. Cullen C, Allen A, Grahn B. Anterior chamber to frontal sinus shunt for the diversion of aqueous humor: a pilot study in four normal dogs. Veterinary Ophthalmology. 1998;1(1):31- 39.

624. Takai Y, Tanito M, Ohiro A. Multiplex cytokine analysis of aqueous humor in eyes with primary open-angle glaucoma, exfoliation glaucoma, and cataract. Investigative Ophthalmology & Visual Science. 2012;53(1):241-247.

625. Funke S, Perumal N, Beck S, Gabel-Scheurich S, Schmelter C, Teister J, Gerbig C, Gramlich OW, Pfeiffer N, Grus FH. Proteomic alterations in human retina samples. Nature Scientific Reports. 2016;6(29759).

626. Hagan S, Tomlinson A. Tear fluid biomarker profiling: a review of multiplex bead analysis. The Ocular Surface. 2013;11(4):219-235.

627. Fan BJ, Liu K, Wang DY, Tham CC, Tam PO, Lam DS, Pang CP. Association of polymorphisms of tumor necrosis factor and tumor protein p53 with primary open angle glaucoma: A replication study of ten genes in a Chinese population Investigative Ophthalmology & Visual Science. 2010;51:4110-4116.

P a g e | 278

628. Lam C, Fan B, Wang D, Tam P, Tham CY, Leung D, Fan DP, Lam DC, Pang C. Association of apolipoprotein E polymorphisms with normal tension glaucoma in a Chinese population. Journal of Glaucoma. 2006;15(3):218-222.

629. Fourgeux C, Martine L, Björkhem I, Diczfalusy U, Joffre C, Acar N, Creuzot-Garcher C, Bron A, Bretillon L. Primary open-angle glaucoma: association with cholesterol 24S-hydroxylase (CYP46A1) gene polymorphism and plasma 24- hydroxycholesterol levels. Investigative Ophthalmology & Visual Science. 2009;50(12):5712-5717.

630. Ayub H, Khan MI, Micheal S, Akhtar F, Ajmal M, Shafique S, Benish Ali SH, den Hollander AI, Ahmed A, Qamar R. Association of eNOS and HSP70 gene polymorphisms with glaucoma in Pakistani cohorts. Molecular Vision. 2010;16:18-25.

631. Ochiai Y, Ochiai H. Higher concenetration of transforming growth factor-β in aqueous humor of glaucomatous eyes and diabetic eyes. Japan Journal of Ophthalmology. 2002;46:249-253.

632. Fuchshofer R, Tamm ER. The role of TGF-β in the pathogenesis of primary open-angle glaucoma. Cell and Tissue Research. 2011;347(1):279-290.

633. Pasutto F, Matsumoto T, Mardin CY, Sticht H, Brandstätter JH, Michels-Rautenstrauss, Weisschuh N, Gramer E, Ramdas WD, van Koolwijk LM, Klaver CC, Vingerling JR, Weber BH, Kruse FE, Rautenstrauss B, Barde Y-A, Reis A. Heterozygous NTF4 mutations impairing neurotrophin-4 signaling in patients with primary open-angle glaucoma. American Journal of Human Genetics. 2009;85(4):447-456.

634. German A, Hall E, Day M. Measurement of IgG, IgM and IgA concentrations in canine serum, saliva, tears and bile. Veterinary Immunology and Immunopathology. 1998;64(2):107-121.

635. Giannetto C, Piccione G, Giudice E. Daytime profile of the intraocular pressure and tear production in normal dog. Veterinary Ophthalmology. 2009;12(5):302-305.

636. Li K, Liu X, Chen Z, Huang Q, Wu K. Quantification of tear proteins and sPLA2-IIa alteration in patients with allergic conjunctivitis. Molecular Vision. 2010;16:2084-2091.

637. Hida RY, Ohashi Y, Takano Y, Dogru M, Goto E, Fujishima H, Saito I, Saito K, Fukase Y, Tsubota K. Elevated levels of human α-defensin in tears of patients with allergic conjunctival disease complicated by corneal lesions: Detection by SELDI ProteinChip system and quantification. Current Eye Research. 2005;30(9):737-744.

638. Matheis N, Okrojek R, Grus FH, Kahaly GJ. Proteomics of tear fluid in thyroid-associated orbitopathy. Thyroid. 2012;22(10):1039-1045.

P a g e | 279

639. Koo B-S, Lee D-Y, Ha H-S, Kim J-C, Kim C-W. Comparative analysis of the tear protein expression in blepharitis patients using two-dimensional electrophoresis. Journal of Proteome Research. 2005;4(3):719-724.

640. Acera A, Suárez T, Rodríguez-Agirretxe I, Vecino E, Durán JA. Changes in tear protein profile in patients with conjunctivochalasis. Cornea. 2011;30(1):42-49.

641. Nichols JJ, Green-Church KB. Mass spectrometry-based proteomic analyses in contact lens- related dry eye. Cornea. 2009;28(10):1109-1117.

642. Zhou L, Beuerman RW, Ang LP, Chan CM, Li SF, Chew FT, Tan DT. Elevation of human α- defensins and S100 calcium-binding proteins A8 and A9 in tear fluid of patients with pterygium. Investigative Ophthalmology & Visual Science. 2009;50(5):2077-2086.

643. Zhou L, Beuerman RW, Chan CM, Zhao SZ, Li XR, Yang H, Tong L, Liu S, Stern ME, Tan D. Identification of tear fluid biomarkers in dry eye syndrome using iTRAQ quantitative proteomics. Journal of Proteome Research. 2009;8(11):4889-4905.

644. Srinivasan S, Thangavelu M, Zhang L, Green KB, Nichols KK. iTRAQ quantitative proteomics in the analysis of tears in dry eye patients. Investigative Ophthalmology & Visual Science. 2012;53(8):5052-5059.

645. Saijyothi AV, Angayarkanni N, Syama C, Utpal T, Shweta A, Bhaskar S, Geetha IK, Vinay PS, Thennarasu M, Sivakumar RM. Two dimensional electrophoretic analysis of human tears: collection method in dry eye syndrome. Electrophoresis. 2010;31(20):3420-3427.

646. Fan BJ, Wiggs J. Glaucoma: genes, phenotypes, and new directions for therapy. Journal of Clinical Investigation. 2010;120(9):3064-3072.

647. Marquis RE, Whitson JT. Management of glaucoma: focus on pharmacological therapy. Drugs Aging. 2005;22(1):1-21.

648. Caprioli J. The Ciliary Epithelia and Aqueous Humor. In: Hart WJ, ed. Adler's Physiology of the Eye. 9th ed. St Louis: Mosby Year Book Inc,; 1992:228-247.

649. Bito LZ. Glaucoma: A physiologic perspective with Darwinian overtones. Journal of Glaucoma. 1992;1(3):193-205.

650. Lima FE, Magacho L, Carvalho DM, Susanna Jr R, Ávila MP. A prospective, comparative study between endoscopic cyclophotocoagulation and the Ahmed drainage implant in refractory glaucoma. Journal of glaucoma. 2004;13(3):233-237.

651. Patel S, Pasquale LR. Glaucoma drainage devices: a review of the past, present, and future. Paper presented at: Seminars in ophthalmology2010.

P a g e | 280

652. Minckler DS, Shammas A, Wilcox M, Ogden T. Experimental studies of aqueous filtration using the Molteno implant. Transactions of the American Ophthalmological Society. 1987;85:368.

653. Prata JA, Jr., Mermoud A, LaBree L, Minckler DS. In vitro and in vivo flow characteristics of glaucoma drainage implants. Ophthalmology. 1995;102(6):894-904.

654. Hong CH, Arosemena A, Zurakowski D, Ayyala RS. Glaucoma drainage devices: a systematic literature review and current controversies. Surv Ophthalmol. 2005;50(1):48-60.

655. Minckler DS, Francis BA, Hodapp EA, Jampel HD, Lin SC, Samples JR, Smith SD, Singh K. Aqueous shunts in glaucoma: a report by the American Academy of Ophthalmology. Ophthalmology. 2008;115(6):1089-1098.

656. Molteno AC, Suter AJ, Fenwick M, Bevin TH, Dempster AG. Otago glaucoma surgery outcome study: cytology and immunohistochemical staining of bleb capsules around Molteno implants. Invest Ophthalmol Vis Sci. 2006;47(5):1975-1981.

657. Tsai JC, Johnson CC, Kammer JA, Dietrich MS. The Ahmed shunt versus the Baerveldt shunt for refractory glaucoma II: longer-term outcomes from a single surgeon. Ophthalmology. 2006;113(6):913-917.

658. Syed HM, Law SK, Nam SH, Li G, Caprioli J, Coleman A. Baerveldt-350 implant versus Ahmed valve for refractory glaucoma: a case-controlled comparison. J Glaucoma. 2004;13(1):38-45.

659. Heuer DK, Lloyd MA, Abrams DA, Baerveldt G, Minckler DS, Lee MB, Martone JF. Which is better? One or two?: a randomized clinical trial of single-plate versus double-plate Molteno implantation for glaucomas in aphakia and pseudophakia. Ophthalmology. 1992;99(10):1512-1519.

660. Sharkawi E, Artes PH, Oleszczuk JD, Bela C, Achache F, Barton K, Bergin C. Systematic Occlusion of Shunts: Control of Early Postoperative IOP and Hypotony-related Complications Following Glaucoma Shunt Surgery. J Glaucoma. 2016;25(1):54-61.

661. Borisy GG, Taylor EW. The mechanism of action of colchicine. Binding of colchincine-3H to cellular protein. J Cell Biol. 1967;34(2):525-533.

662. Yu YS, Youn DH. The effect of colchicine on fibroblast proliferation after glaucoma filtering surgery. Korean J Ophthalmol. 1987;1(2):59-71.

663. Molteno AC, Straughan JL, Ancker E. Control of bleb fibrosis after glaucoma surgery by anti- inflammatory agents. S Afr Med J. 1976;50(23):881-885.

P a g e | 281

664. Duan X, Jiang Y, Qing G. [Long-term follow-up study on Hunan aqueous drainage implantation combined with mitomycin C for refractory glaucoma]. Yan Ke Xue Bao. 2003;19(2):81-85.

665. Glover TL, Nasisse MP, Davidson MG. Effects of topically applied mitomycin-C on intraocular pressure, facility of outflow, and fibrosis after glaucoma filtration surgery in clinically normal dogs. Am J Vet Res. 1995;56(7):936-940.

666. Palmberg P. The failing filtering bleb. Ophthalmology Clinics. 2000;13(3):517-529.

667. Wilcox MJ, Barad JP, Wilcox CC, Peebles EL, Minckler DS. Performance of a new, low-volume, high-surface area aqueous shunt in normal rabbit eyes. J Glaucoma. 2000;9(1):74-82.

668. Peleg AY, Weerarathna T, McCarthy JS, Davis TM. Common infections in diabetes: pathogenesis, management and relationship to glycaemic control. Diabetes Metab Res Rev. 2007;23(1):3-13.

669. Teng CC, Chi HH, Katzin HM. Histology and mechanism of filtering operations. Am J Ophthalmol. 1959;47(1 Part 1):16-33.

670. Smith MF, Doyle JW, Ticrney JW, Jr. A comparison of glaucoma drainage implant tube coverage. J Glaucoma. 2002;11(2):143-147.

671. Bras D, Maggio F. Surgical Treatment of Canine Glaucoma: Cyclodestructive Techniques. Vet Clin North Am Small Anim Pract. 2015;45(6):1283-1305, vii.

672. O'Reilly A, Hardman C, Stanley R. The use of transscleral cyclophotocoagulation with a diode laser for the treatment of glaucoma occurring post intracapsular extraction of displaced lenses: a retrospective study of 15 dogs (1995–2000). Veterinary ophthalmology. 2003;6(2):113-119.

673. Nasisse MP, Davidson MG, MacLachlan NJ, Corbett W, Tate LP, Newman HC, Hardie EM. Neodymium:yttrium, aluminum, and garnet laser energy delivered transsclerally to the ciliary body of dogs. Am J Vet Res. 1988;49(11):1972-1978.

674. Lutz E, Webb T, Bras I, Sapienza J, Wilkie D, Gemensky-Metzler A. Diode endoscopic cyclophotocoagulation in dogs with primary and secondary glaucoma: 292 cases (2004- 2013). Vet Ophthalmol. 2013;16(6):40.

675. Bras I, Robbin T, Wyman M, Rogers A. Diode endoscopic cyclophotocoagulation in canine and feline glaucoma. Veterinary Ophthalmology. 2005;8(6):449.

676. Lutz E, Sapienza J. Diode endoscopic cyclophotocoagulation in pseudophakic and aphakic dogs with secondary glaucoma. Veterinary Ophthalmology. 2009;12:398.

P a g e | 282

677. Graham KL, Donaldson D, Billson FA, Billson FM. Use of a 350-mm(2) Baerveldt glaucoma drainage device to maintain vision and control intraocular pressure in dogs with glaucoma: a retrospective study (2013-2016). Vet Ophthalmol. 2017;20(5):427-434.

678. Bartholomew R. Glaucoma implants. Their use in difficult cases of glaucoma. Transactions of the ophthalmological societies of the United Kingdom. 1977;98(4):482-485.

679. Bedford P. Use of a one‐piece drainage system in the treatment of a closed angle glaucoma in a dog. Journal of Small Animal Practice. 1988;29(4):231-237.

680. Investigators A. The relationship between control of intraocular pressure and visual field deterioration. Advanced Glaucoma Intervention Study (AGIS): 7. Am J Ophthalmol. 2000;130(4):429-440.