Analysis of Morbidity and Mortality of Wild Koalas in South-East Queensland Using Passive Surveillance Data

Analysis of Morbidity and Mortality of Wild Koalas in South-East Queensland using Passive Surveillance Data

Viviana Gonzalez-Astudillo

BVSc, MNR

A thesis submitted for the degree of Doctor of Philosophy at

The University of Queensland in 2018

School of Veterinary Science

Abstract

The koala is an iconic marsupial threatened by habitat loss, trauma, droughts, and diseases. The Australian state of Queensland was estimated to hold the largest koala population during the 1990s; however, this number has nearly halved in two decades, with declines of up to 80%, particularly in South-East Queensland (SE QLD). Despite this alarming reduction, the relative contribution of threats causing this decline in SE QLD had not been quantitatively assessed. In addition, information on the spatial- temporal distribution of morbidity and mortality cases is lacking for SE QLD, along with competing injury and disease risks. To address this issue, a retrospective epidemiological study from cases submitted to hospitals (1997-2013) was performed. The analysis included N=20,250 records, highlighting that most koalas arrived dead (48%), followed by other being euthanized (35%), and a smaller number being released (17%). Trauma by motor vehicles, urogenital chlamydiosis and low body conditions were the leading causes of admission and frequently co-occurred. Of concern was that 37% of koalas were injured but otherwise healthy, and that urogenital chlamydiosis, which causes infertility, overwhelmingly affected females. Multinomial logistic regression models identified year of admission, age class, sex and season as risk factors significantly influencing koala outcomes for main clinical syndromes assessed. Exploratory space-time permutation scans detected significant clusters of trauma, chlamydiosis and low body condition in specific local government areas. In this study, passively-acquired data provided insight into the complex interplay of threats impacting koalas in SE QLD.

Although hospitals have been providing specialised care to koalas for decades, there is no standard nomenclature to record admissions, complicating the estimation of causes of mortality and the incidence of comorbidities. Additionally, autopsies are frequently not conducted in a systematic manner, hampering the identification of comorbidities and disease interactions in the SE QLD population decline. A literature search for the commonly documented morbidity and mortality causes for koalas was used to categorise standardised nomenclature. This was followed by a large scale autopsy survey (2013-2016) including N=519 koala autopsies from carcasses derived from main SE QLD hospitals. Demographically, there was a higher number of mature koalas of breeding age admitted with infections and trauma compared to young/subadult or senescent koalas. Trauma by motor vehicles, urogenital and ocular

2 chlamydiosis were the leading causes of admission, frequently presenting as comorbidities. About 19% of koalas suffering injuries were otherwise healthy. Of concern was the high numbers of koalas with chronic diseases leading to low body condition and the disproportionate number of females diagnosed with reproductive disease. No observable evidence was found to explain low body condition in 6% of young/subadult males. A low body condition was found to significantly influence the likelihood of animal attacks verses normal body condition injured by motor vehicles or animals (p = 0.007). Novel and rare conditions were characterised grossly and histologically for koalas. Trauma and chlamydiosis have been maintained as main causes of admission to hospitals over two decades of koala population surveys, stressing the relative importance of these conditions in the koala decline. Chronic diseases dominated the study, and there is evidence debilitation increases susceptibility to succumb to other causes. The demographic distribution of admission could reflect population structure and cumulative exposure risk to trauma. The overrepresentation of female koalas with chlamydiosis reflects the surveillance by hospitals to curb the release of permanently infertile carriers. Although there are limitations to the interpretation of hospital-acquired data, both studies present comprehensive and quantifiable evidence that aetiologies are acting jointly as multifactorial elements in the continuing koala decline.

Koala disease monitoring is usually based on opportunistic sampling, which may lead to variation in diagnosis for the same koala between clinical and pathological examination. Thus, understanding the accuracy when making a diagnosis is essential to understand the causes for the koala decline. Mixed models were used to estimate the probability of agreement of producing the same diagnosis for the same koalas by veterinary hospitals or pathologists. Almost 80% of the variation in the agreement was explained by the diagnosis effect, highlighting that some diagnoses may be easier to make than others. There was a high probability of agreement for commonly diagnosed conditions, and low probabilities of agreement for historically underdiagnosed conditions, subclinical parasitosis, and resolved tissue injuries. Variation in the agreement on producing the same diagnosis was influenced to a lesser degree by koala-specific characteristics such as carcass preservation. This highlights that clinical

3 examination in wildlife hospitals is accurate for common koala conditions, while autopsies increased detection sensitivity to identify specific aetiologies, diagnoses and comorbidities. Multiple factors can influence the quality of an autopsy. In spite of this, autopsy remains an important method to confirm or overturn a diagnosis, and clarify or correct clinical signs in cases with diagnostic uncertainty, such as with emerging or novel diseases in wildlife.

Declaration by Author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, financial support and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my higher degree by research candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis and have sought permission from co-authors for any jointly authored works included in the thesis.

Publications during Candidature

• Peer-reviewed papers

Gonzalez-Astudillo V., Leon-Alvarado O.D., Ossa-Lopez, P.A., Rivera-Paez, F.A., Ramirez-Chaves, H.E. 2018. Sarcoptic mange in wild Quichua Porcupines (Coendou quichua Thomas, 1899) in Colombia. International J for Parasitology: Parasites and Wildlife. 7: 95-98.

Gonzalez-Astudillo V., Allavena R., McKinnon A., Larkin R., Henning J. 2017. Decline causes of Koalas in South East Queensland, Australia: a 17-year retrospective study of mortality and morbidity. Scientific Reports 7:42587.

Gonzalez-Astudillo V., Hernández S.M., Peters V., Yabsley M.J., Keel K.M., Munk B., Ruder M.G., Brown J., Fischer J.R., and Nemeth M.N. 2016. Mortality of passerines and relatives submitted to a wildlife diagnostic laboratory (Southeastern Cooperative Wildlife Disease Study, USA): a 36-year retrospective analysis. Journal of Wildlife Diseases 52: 441-458.

Gonzalez-Astudillo V., Bustamante-Rengifo J.A., Bonilla A., Castillo-Giraldo A.O., Lehmicke A.J.J., Astudillo-Hernandez M. 2016. Synanthropic Cockroaches (Blattidae: Periplaneta spp.) Harbor Pathogenic Leptospira in Colombia. Journal of Medical Entomology. 53:177-82.

Gonzalez-Astudillo V., Ramirez H.E., Henning J., Gillespie T.R. 2016. Current knowledge of studies of pathogens in Colombian mammals. MANTER: Journal of Parasite Biodiversity. Occasional papers, Number 4, September 1. doi:10.13014/K2057CV6.

Nemeth, N.M., Gonzalez-Astudillo V., Oesterle, P., Howerth, E.W. 2016. A 5-Year Retrospective Review of Avian Diseases Diagnosed at the Department of Pathology, University of Georgia. Journal of Comparative Pathology 155:105-120.

González Astudillo V., Schaffer-White A., Allavena R., Palmieri C. 2015. Multiple Intra- abdominal Serosal Myxosarcomas in Two Koalas (Phascolarctos cinereus). Journal of Comparative Pathology 152: 283-286.

González-Astudillo V., Peña-Stadlin J.E., Bustamante J.A., Astudillo Hernández M. 2015. Terminal leptospirosis in a woolly monkey (Lagothrix lagotricha) in Colombia. Revista MVZ Cordoba 20: 4505-4510.

Gonzalez-Astudillo V., Peña-Stadlin J.E., Astudillo-Hernández M. 2015. Anti- leptospiral agglutinins in marmosets (Saguinus oedipus and Saguinus leucopus) from illegal trade. Revista MVZ Cordoba 20: 4790-4799.

González Astudillo V., Aguirre A. 2014. News from the IAEH: Achieving multidisciplinary EcoHealth education in early professional development. EcoHealth 11: 1-2.

• Conference abstracts Gonzalez-Astudillo V, Henning J, Allavena R. Analysis of morbidity and mortality of wild koalas in south-east Queensland using passive surveillance data. 66th Wildlife Disease Association Annual International Conference. San Cristobal de Las Casas, Chiapas, Mexico, July 23-28, 2017.

Gonzalez-Astudillo V, Kidd L, Henning J, Sadowski P, Allavena R. Applying state-of- the-art technology to investigate koala bone disease. Australian and New Zealand College of Veterinary Scientists. College Science Week. Surfers Paradise, Queensland, July 6-8, 2017.

Gonzalez-Astudillo V, Henning J, Allavena R. Can passive surveillance aid in the understanding of decline of an iconic marsupial? The koala in SE QLD. Wildlife Society & Wildlife Disease Association Conference (Australasian Section). Harewood Golf Club, Christchurch, New Zealand, November 23-30, 2016.

Gonzalez-Astudillo V, Allavena R, Larkin R, McKinnon A, Henning J. The koala (Phascolarctos cinereus) in SE QLD: using passive surveillance to understand the decline of an iconic species. II School of Veterinary Science Research Conference. The University of Queensland, Australia, October 27th, 2016.

Gonzalez-Astudillo V, Allavena R, Larkin R, McKinnon A, Henning J. The koala (Phascolarctos cinereus) in SE QLD: using passive surveillance to understand the

decline of an iconic species. Society for Conservation Biology (Oceania conference). Brisbane Convention and Exhibition Centre, Brisbane, Australia, July 6-8, 2016.

Gonzalez-Astudillo V, Henning J, Allavena R. The koala (Phascolarctos cinereus) in SE QLD: using passive surveillance to understand the decline of an iconic species. Inaugural School of Veterinary Science Research Conference. The University of Queensland, Australia, November 4, 2015.

Henning J, Gonzalez-Astudillo V, Allavena R. Investigating koala declines with online tools. International Symposia on Veterinary Epidemiology and Economics. Merida, Mexico. November 3-7, 2015.

Gonzalez-Astudillo V, Henning J, Allavena R. Analysing morbidity and mortality in wild koalas from SE QLD using passive surveillance data. American College of Veterinary Pathologists Annual Conference. Minnesota, USA. October 17-21, 2015.

Gonzalez-Astudillo V, Schaffer-White A, Allavena R, Palmieri C. Serosal myxosarcoma in two koalas (Phascolarctos cinereus): Case Series. American College of Veterinary Pathologists Annual Conference. Atlanta, USA. November 8-11, 2014.

Gonzalez-Astudillo V, Henning J, Allavena R. Retrospective analysis of fracture epidemiology, with pathological analysis of fracture healing processes in koalas (Phascolarctos cinereus). American College of Veterinary Pathologists Annual Conference. Atlanta, USA. November 8-11, 2014.

Publications included in this thesis

Gonzalez-Astudillo V., Allavena R., McKinnon A., Larkin R., Henning J. 2017. Decline causes of koalas in south-east Queensland, Australia: a 17-year retrospective study of mortality and morbidity. Scientific Reports. 7:42587 – incorporated as Chapter 2.

Contributor Statement of contribution Viviana Gonzalez-Astudillo (Candidate) Conception and design (40%) Analysis and interpretation (60%) Drafting and production (50%) Rachel E. Allavena Conception and design (20%) Analysis and interpretation (15%) Drafting and production (10%) Allan McKinnon Conception and design (0%) Analysis and interpretation (0%) Drafting and production (5%) Rebecca Larkin Conception and design (0%) Analysis and interpretation (0%) Drafting and production (5%) Joerg Henning Conception and design (40%) Analysis and interpretation (25%) Drafting and production (30%)

Contributions by others to the thesis

No contributions by others.

Statement of parts of the thesis submitted to qualify for the award of another degree

None.

Research Involving Human or Animal Subjects

ANRFA/SVS/193/13/EHP valid from 12/08/2013 to 12/08/2016 given by The University of Queensland’s Native and Exotic Wildlife and Marine Mammals Animal Ethics Committee (Appendix 1)

WISP13247813, valid from 07/08/2913 to 06/08/2018 given by the Queensland Government Department of Environment and Heritage Protection (Appendix 2).

Acknowledgements

I am in debt with the Queensland Government, Department of Environment and Heritage Protection for the financial support during my doctoral studies. Embarking on that Australia-bound flight four years ago has turned into an incredible journey. The University of Queensland has been a great institution to be a part of so inevitably, I leave with a strong sense of belonging to the UQ community. Thank you for supporting students during our scholarly endeavours.

I thank and dedicate this thesis to my mother, who has given me the necessary tools to be where I am today. Gracias inmensas, mama! I also dedicate my work to my father, who taught me since my early ages to love and respect All Creatures Great And Small. I also thank my partner for always having something inappropriate to say at the appropriate time.

Inevitably, I am and will always be grateful for being surrounded by marvellous wildlife to investigate on, so the never-ending fight for the conservation for all species continues.

Many thanks to A/Prof. Rachel Allavena and to A/Prof. Joerg Henning for their advice in conducting the work hereby mentioned. I have met many people that have made the doctoral journey easier, lighter, smoother at times of desperation. Far too many names to type which makes me realize how lucky I am to have been surrounded by such amazing human beings. I will reach out to express my gratitude to each of you individually.

To all, Thanks heaps! Muchas gracias!

Financial Support

The study presented in this doctoral thesis, living expenses and fees were financially supported by a grant provided by the Queensland Government, Department of Environment and Heritage Protection to The University of Queensland.

I acknowledge the reception of a UQ Research Training Tuition Fee Scholarship during my doctoral studies.

Keywords

Autopsy, Chlamydia, Cause of death, Epidemiology, Marsupialia, Methateria, Mortality, Pathology, Phascolarctidae, Trauma.

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code 070704: Veterinary Epidemiology, 40%

ANZSRC code 070709: Veterinary Pathology, 50%

ANZSRC code 050202: Conservation and Biodiversity, 10%

Fields of Research (FoR) Classification

FoR code: 0707 Veterinary Sciences, 80%

FoR code: 0502 Environmental Science and Management, 20%

Table of Contents

Abstract ...... 2

Declaration by Author ...... 5

Publications during Candidature ...... 6

Publications included in this thesis ...... 9

Contributions by others to the thesis ...... 10

Statement of parts of the thesis submitted to qualify for the award of another degree10

Research Involving Human or Animal Subjects ...... 10

Acknowledgements ...... 11

Financial Support ...... 12

Keywords ...... 13

Australian and New Zealand Standard Research Classifications (ANZSRC) ...... 13

Fields of Research (FoR) Classification ...... 13

Table of Contents ...... 1

List of Figures and Tables ...... 7

List of Abbreviations ...... 9

Chapter 1 Introduction, Literature Review and Scope of the Study ...... 10

1.1 Pressures on Koala Populations in South-East Queensland ...... 11

1.2 Biology and Population Dynamics ...... 12

1.2.1 Taxonomic classification and DNA diversity ...... 12

1.2.2 Natural diet ...... 13

1.2.3 Home ranges ...... 13

1.2.4 Reproductive cycle ...... 14

1.3 Koala Distribution in Queensland ...... 14

1.4 Habitat-related threats to koala populations...... 15

1.4.1 Habitat loss ...... 15

1.4.2 Climate change, droughts and bushfires ...... 16

1.5 Conservation actions ...... 16

1.6 Main Koala Threats from disease and injury ...... 17

1.6.1 Diseases ...... 17

1.6.2 Traumatic Injuries ...... 22

1.7 Data Collection Systems for Animal Diseases ...... 24

1.7.2 Disease surveillance in wildlife populations ...... 26

1.7.3 Role of Autopsy in Wildlife Passive Surveillance ...... 28

1.7.4 Surveillance of disease and mortalities in koalas ...... 29

1.8 Study Objectives and Thesis Structure ...... 32

Chapter 2 Decline causes of Koalas in South East Queensland, Australia: a 17-year retrospective study of mortality and morbidity...... 34

2.1 Abstract...... 34

2.2 Introduction ...... 34

2.3 Materials and Methods ...... 35

2.3.1 Study population ...... 35

2.3.2 Quantitative analysis ...... 36

2.3.3 Koala submissions ...... 36

2.4 Results ...... 38

2.4.1 Study population ...... 38

2.4.2 Aetiologies ...... 38

2.4.3 Clinical syndromes ...... 40

2.4.4 Koala outcomes ...... 41

2.4.5 Trauma by motor vehicles ...... 43

2.4.6 Chlamydia-like signs...... 44

2.4.7 Trauma by animal attack ...... 45

2.5 Discussion ...... 45

Chapter 3 Pathology and Trends in Morbidity and Mortality in the Koala (Phascolarctos cinereus) in South-East Queensland, Australia...... 53

3.1 Introduction ...... 53

3.2 Materials and Methods ...... 53

3.2.1 Survey Population ...... 53

3.2.2 Autopsies and Sample Collection ...... 54

3.2.3 Data Analysis ...... 54

3.2.4 Clinical Disease Grading Criteria ...... 55

3.3 Results ...... 55

3.3.1 Summary of Causes of Admission ...... 55

3.3.2 Urinary tract Disease ...... 59

3.3.3 Reproductive Disease ...... 61

3.3.4 Ocular Disease ...... 63

3.3.5 Respiratory Disease ...... 64

3.3.6 Neoplastic Disease ...... 64

3.3.7 Trauma ...... 65

3.3.8 Loss of Body Condition...... 66

3.3.9 Idiopathic conditions ...... 67

3.4 Discussion ...... 69

3.4.1 Benefits of autopsies and passive surveillance for detecting disease in wildlife populations ...... 69

3.4.2 Infectious diseases ...... 71

3.4.3 Chlamydial diseases ...... 72

3.4.4 Respiratory disease ...... 76

3.4.5 Trauma ...... 76

3.4.6 Idiopathic loss of body condition & KoRV ...... 78

3.4.7 Idiopathic conditions ...... 81

3.4.8 Neoplastic disease ...... 82

3.4.9 Comorbidities in koala studies and Seasonality of Admissions ...... 82

3.4.10 Limitations of the study & Conclusions ...... 84

Chapter 4 Accuracy of diagnoses of koala disease and injury in South-East Queensland 86

4.1 Introduction ...... 86

4.2 Materials and Methods ...... 89

4.2.1 Diagnosis of disease and injury of koalas in South-East Queensland ...... 89

4.2.2 Submission of koalas for pathological examination and autopsy-derived diagnosis ...... 91

4.2.3 Statistical analysis ...... 91

4.3 Results ...... 93

4.3.1 Study population ...... 93

4.3.2 Diagnostic agreement...... 94

4.4 Discussion ...... 96

4.4.1 Agreement between clinical and autopsy-derived diagnoses ...... 96

4.4.2 Clinical versus autopsy-derived diagnoses: opportunities to improve passive surveillance systems...... 101

Chapter 5 Discussion ...... 104

5.1 Main goals driving doctoral thesis and Chapters ...... 104

5.2 Epidemiology of Koala Morbidity and Mortality in QLD ...... 106

5.3 Trauma ...... 108

5.4 Chlamydia & loss of body condition ...... 110

5.5 Comorbidities ...... 112

5.6 Integrated Conservation Strategy for SEQLD Koalas ...... 114

List of References ...... 117

Appendix 1 Copy of Animal Ethics Approval Certificate ANRFA/SVS/193/13/EHP by UQ Native and Exotic Wildlife and Marine Mammals Animal Ethics Committee...... 141

Appendix 2 Copy of Approval of Scientific Purposes Permit WISP 13247813 by the Department of Environment and Heritage Protection...... 142

Appendix 3 Trauma by motor vehicles (N=5,047): Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 143

Appendix 4 Chlamydia-like signs (N=3,730). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized), released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 144

Appendix 5 Chlamydia-like signs & Wasting (N=3,477). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 145

Appendix 6 Trauma by animal attacks (N=1,537). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 146

Appendix 7 Chlamydia-like signs & trauma by motor vehicle (N=811). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 147

Appendix 8 Trauma by other causes (N=686). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 148

Appendix 9 Chlamydia-like signs & Senescence & Wasting (N=517). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: euthanized koalas)...... 149

Appendix 10 Chlamydia-like signs & Trauma animal attack (N=271). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas)...... 150

Appendix 11 Wasting (N=259). Final multinomial logistic regression of risk factors associated with outcome of diagnosis (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013 (reference category: released koalas). 151

Appendix 12 Trauma by motor vehicle: Predicted probabilities (with 95% confidence intervals) derived from multinomial logistic regression models of risk factors associated with the diagnostic outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013...... 152

Appendix 13 Chlamydia-like signs: Predicted probabilities (with 95% confidence intervals) derived from multinominal regression models of risk factors associated with the

5 diagnostic outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013...... 153

Appendix 14 Chlamydia-like signs & wasting: Predicted probabilities (with 95% confidence intervals) derived from multinominal regression models of risk factors associated with the diagnostic outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013...... 154

Appendix 15 Trauma by animal attack: Predictive probabilities (with 95% confidence intervals) derived from multinominal regression models of risk factors associated with the diagnostic outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013...... 155

Appendix 16 Pathology of Chlamydiosis: Ocular disease with clinical grading .. 156

Appendix 17 Pathology of Chlamydiosis: Lower urinary tract disease with clinical grading 157

Appendix 18 Pathology of Chlamydiosis: Upper urinary tract disease...... 158

Appendix 19 Pathology of Neoplasms ...... 159

Appendix 20 Pathology of Parasitosis ...... 161

Appendix 21 Pathology of Putative AIDS-like Syndrome ...... 162

Appendix 22 Pathology of Idiopathic Diseases ...... 163

Appendix 23 Pathology of Congenital Diseases ...... 164

Appendix 24 Pathology of Degenerative Diseases ...... 165

Appendix 25 Probability of agreement (including 95% credible intervals) for diagnoses of koala diseases and injuries between Wildlife Hospitals (WH1, WH2, WH3) and The University of Queensland (UQ) in South-East Queensland, Australia (2013-2016).166

Appendix 26 Excerpt from Basic Field Autopsy Guide for Koalas ...... 169

Appendix 27 Excerpt from Basic Field Autopsy Guide for Koalas: Heart Dissection 170

List of Figures and Tables

Figure Title Page 2-1 Total counts and proportions of aetiologies Chlamydia-like signs, trauma caused by motor vehicles, and wasting occurring in koalas 40 submitted to wildlife hospitals in SEQLD from 1997–2013. 2-2 Locations of significant clusters of aetiologies as determined by the 42 space-time permutation scans. 2-3 Proportion of top four koala diagnoses (Trauma by motor vehicle, Chlamydia-like signs, clinical syndrome combination of Chlamydia- like signs & wasting, and trauma by animal attacks) affecting koalas 44 submitted to wildlife hospitals in SEQLD from 1997–2013 with a CI 95%. 2-4 Outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals by year of submission between from 44 1997–2013. 3-1 Demographic distribution of the frequency of etiologies for South- 57 East Queensland koalas autopsied at The University of Queensland. 3-2 Seasonal distribution of the frequency of top 4 final diagnoses (Chlamydia-like/Wasting N=98, Trauma motor vehicle N=51, Chlamydia-like N=36, Chlamydia-like/Senescence N=33) diagnosed 59 during autopsy examinations in koalas admitted to wildlife hospitals in South-East Queensland from 2013 through to 2016. 3-3 Cystitis, urinary bladder. Hyperplasia causing mucosal infolding. 60 3-4 Cystitis, urinary bladder. Discolouration of fur surrounding the 60 common vestibule. 3-5 Hydronephrosis and hydroureter, kidney (asterisk), and ureter 61 (arrowhead). 3-6 Organized infarct, kidney. 61 3-7 Bursitis, ovarian bursa, and bladder (asterisk). 62 3-8 Frequency of infection sites with gross and histopathological lesions compatible with chlamydiosis observed during autopsies conducted 62 in koalas admitted to South-East Queensland hospitals from 2013- 2016 (N=304). 3-9 Demographic distribution of the frequency of main body part or system affected by putative chlamydiosis as diagnosed in South- 63 East Queensland koalas autopsied from 2013-2016. 3-10 Keratoconjunctivitis, eye and periorbita (clinical scoring = 3). 64 3-11 Proportion of South-East Queensland koalas affected by lethal trauma in relation to their age class, autopsied at The University of 65 Queensland from 2013-2016. 3-12 Severe fibromatosis, spleen. 68 3-13 Fibromatosis, spleen. Masson’s trichrome. 68 3-14 Fibromatosis, liver (same koala as Figure 15). 69 3-15 Fibromatosis, liver. Masson’s trichrome. 69 Table Title Page 1-1 Comparison between the body mass and body length in northern 13 and southern wild koalas. 1-2 Koala population density distribution in states and territories based 15 on reported government and non-governmental records.

1-3 Examples of global morbidity and mortality wildlife studies utilising 30 passive surveillance methodology. 1-4 Examples of global morbidity and mortality wildlife studies utilising 31 an active surveillance methodology. 2-1 Frequency of occurrences of diagnoses (with a CI 95%) in koalas 39 submitted to wildlife hospitals from 1997 through 2013. 2-2 Significant space-time clusters identified through case-bases space- 41 time permutation. Koalas were submitted with trauma by motor vehicle, Chlamydia-like infection and wasting to wildlife hospitals in South East Queensland, from 1997 through 2013 2-3 Frequency of the top 10 (>1%) clinical syndromes in koalas 43 submitted to wildlife hospitals (with a CI 95%) in SEQLD from 1997 through 2013 3-1 Distribution of aetiologies diagnosed in each age class in koalas 56 admitted to South-East Queensland hospitals and receiving autopsies at The University of Queensland from 2013 through to 2016. 3-2 Distribution of co-morbidities and their association with traumatic 66 injuries in koalas presented for autopsies at The University of Queensland from 2013-2016. 3-3 Body Condition (BC) score of koalas admitted to wildlife hospitals 67 from 2013 through to 2016 in South-East Queensland receiving autopsies at The University of Queensland grouped by age class (young/subadult, mature, aged; N=506). 4-1 Probabilities of agreement (including the 95% credible intervals) for 94 diagnoses of diseases and injuries in South-East Queensland, Australia (2013-2016). 4-2 Posterior summaries for the model parameters: posterior mean with 96 95% credible intervals and standard deviation (SD)

List of Abbreviations

Symbol Meaning

% Percent

μL microlitre

AIDS Acquired Immune Deficiency Syndrome cm centimetre

CO2 Carbon dioxide ha hectares kg kilograms

KoRV Koala retrovirus

NSW New South Wales ompA Outer membrane protein A

PCR Polymerase chain reaction

SD Standard deviation

SEQLD South-East Queensland

VIC Victoria

Chapter 1 Introduction, Literature Review and Scope of the Study

The Koala (Phascolarctos cinereus) is an iconic marsupial, a potent icon of Australia’s natural heritage worldwide. The koala has a fragmented but broad distribution, extending from north Queensland (QLD) to the south-east portion of South Australia (SA). The species geographic distribution is tied to its primary food source, the Eucalyptus forest and woodlands. Koalas have a set of important values stemming from its significance as a symbol of cultural identification, contributions to the tourism industry, and their ecological niche (Martin and Handasyde, 1999). Economically, koalas were estimated to be worth AUD$1.1 billion to the Australian tourism industry according to a 1997 study (Hundloe and Halmilton, 1997). Despite their popularity and perceived value, the conservation of koalas is menaced by a suite of threats including but not limited to, habitat loss from the continued urbanisation of coastal areas overlapping with viable koala habitat, domestic dog attacks, motor vehicle collisions, climate change, as well as infectious diseases (Gonzalez-Astudillo et al., 2017, Dique, Preece, de Villiers, 2003, Seabrook et al., 2011b, Lane, 2008).

QLD was a key state in the conservation of koalas as a vast portion of its area was considered to have ideal habitat for the species, in addition to holding the largest population of koalas during the 1990s, with an estimated of 295,000 animals (Threatened Species Scientific Committee, 2012). The South-East (SE) portion of the state in particular holds the most abundant population in the region (Phillips, 2000). Since the 1990s, SEQLD koalas have experienced a severe decline (Department of Environment and Heritage Protection, 2015a, Rhodes et al., 2015). There have been considerable governmental efforts to monitor koalas in SEQLD; in particular, to improve the current understanding of their abundance, distribution, and population dynamics. Population monitoring surveys were initially focused at the Koala Coast (considered as a single genetic management unit (Sherwin et al., 2000)) and Pine Rivers Shire; however, surveyed regions were later extended to other Local Government Areas (LGAs) within SEQLD. These surveys have resulted in one of the largest datasets on wild koala populations, forming the basis of analyses to determine and quantify the causes of low SEQLD koala numbers (Rhodes et al., 2015). In contrast to their southern counterparts, recent population modelling surveys in QLD have documented rapid declines that have increased over time between 1996-2014 in the Koala Coast and Pine Rivers populations region of up to 80% and 54% respectively (Rhodes et al., 2015). Nonetheless, koala populations were not considered for vulnerable listing in the southern portion of their distribution (Victoria – VIC, SA), because of their reported stable and abundant status.

Despite low SEQLD koala populations numbers, effective application of conservation measures has been hampered by the lack of consensus regarding the species conservation status and threats koala are facing, primarily owing to lack of uniformity in the threats faced across their range (Phillips, 2000). Responding to these rising pressures on SEQLD koala populations, the species was conveyed protection as vulnerable in 2012 across the entire state according to the Commonwealth Environmental Protection Biodiversity Conservation Act 1999 (EPBC Act) (EPBC, 1999) and under the Queensland Nature Conservation Act 1992 (Government, 1992) in 2015.

1.1 Pressures on Koala Populations in South-East Queensland

In QLD, major threats to the species are infectious diseases and traumatic injuries. The prevalence of diseases and trauma from motor vehicles and dog predation is influenced by factors such as season, and demographics (Jackson et al., 1999). Because trauma directly results from anthropogenic activities, spatial modelling and koala mapping play a key role in collecting data to assess the viability of SEQLD and identify management options (Government, 2015). In regards to infectious disease, chlamydiosis, the disease caused by the bacterium Chlamydia spp., has deleterious effects in the species, mainly owed to its chronic sequelae of irreversible infertility, progressive debilitation, and blindness. However, chlamydial infection does always not equal to disease expression (Polkinghorne et al., 2013), leaving many unanswered questions in the epizootiology of this pathogen. Often, chlamydiosis is diagnosed via clinical examination if hallmark signs are overt. Subclinical infections can be detected via molecular methods such as PCR, which offer a much higher sensitivity to chlamydia infections but are not available for field testing. Enzyme immunoassays such as Clearview® are routinely used in koalas at point-of-care facilities or in the field, offering a high specificity but less sensitivity than the gold-standards (Hanger et al., 2013). In regards to koala retrovirus (KoRV), high levels of viraemia are significantly correlated with tumours (Tarlinton et al., 2006). More recently, reports document that KoRV infection is not only a predisposing factor for overt chlamydiosis (Waugh et al., 2017), but also acts as a primary pathogen in other important koala immunosuppressive-related illnesses (Tarlinton et al., 2005, Oliveira et al., 2007). The effects of maladaptive, chronic stress leading to immunosuppression cannot be underestimated, particularly when considering populations inhabiting isolated fragments or affected by extreme weather events (Davies et al., 2013, Narayan et al., 2013). Given that funding for wildlife conservation is generally limited and threats are commonly multifactorial, plans to recover the species need

11 to be based on identifying threats and quantitative assessments for efficient resource allocation.

Owing to the multiple threats affecting koalas it is clear that the establishment of a long-term surveillance system for the species is needed to obtain vital population statistics. One of the most efficient ways to do this is by utilising data on koala admissions to wildlife hospitals. When koalas are found injured, sick or orphaned by members of the public, wildlife ambulances established by volunteer groups across QLD collect and submit koalas to wildlife hospitals. Multidisciplinary teams at these institutions are specialised in koala care and provide treatment and rehabilitation. The information resulting from these hospital admissions has inadvertently resulted in highly valuable long-term morbidity and mortality data for the species. Utilising these koala hospital records of admissions for research offers a multiple-season, broad-coverage data set of koala diseases and injuries. Thus these wildlife hospital records can be considered as a passive surveillance system as they are a by-product of other health related activities, such as clinical records kept for koalas (Sergeant and Perkins, 2015). Data of this type can also be used to identify disease risk factors in time and space, serve as an ‘early outbreak’ warning system (Cox-Witton et al., 2014), and assist in the subjective estimation of baseline levels of common diseases at the population level (Maas et al., 2016).

1.2 Biology and Population Dynamics 1.2.1 Taxonomic classification and DNA diversity

The Koala (Phascolarctos cinereus, Goldfuss 1817) is the only extant species within the taxonomic family Phascolarctidae (Sherwin et al., 2000); the only remaining representatives within this family are six-well known fossil records and others of debatable phylogeny (Benton, 2004). Following the taxonomy ladder, koalas belong to the Infraclass Marsupialia, meaning their young are carried inside a pouch. Other anatomical distinctive characteristics of the koala are due to being diprodonts (Order Diprodontia, Suborder Vombatiformes). Diprodonts, functionally, have a single pair of incisors in their mandible, and display syndactylia: a fusion of their second and third hind digits (Martin and Handasyde, 1999). Although koalas are currently accepted as a single species, clinal variations along a broad geographic range rather than true DNA divergence led to morphological differences in body size and to the past description of three sub-species. Specifically, southern koala populations have larger body sizes and darker fur colour patterns compared to northerner individuals (Martin and Handasyde, 1999, Briscoe et al., 2014) Table 1-1). However, the 12 evidence was limited to support these three subspecies as evolutionarily significant units. The recommendation was made that local koala populations are treated as management units, instead of evolutionarily significant units for population management purposes (Houlden et al., 1999).

Table 1-1. Comparison between the body mass and body length in northern and southern wild koalas.

Weight Head & Body Length Location (kg range) (cm range) Queensland Male 6.5 (4.2-9.1) 70.5 (67.4-73.6) Female 5.1 (4.1-7.3) 68.7 (64.8-72.3) Victoria Male 12.0 (9.5-14.9) 78.2 (75.0-82.0) Female 8.5 (7.0-11.0) 71.6 (68.0-73.0) Extracted from (Martin and Handasyde, 1995)

1.2.2 Natural diet

Ecologically, koalas have a highly specialised diet based primarily on Eucalyptus leaves and rarely on other plant species (Melaleuca sp.). Thus koalas are termed ‘habitat specialists’ as they primarily feed from around 50 of the up to 700 Eucalyptus (Blanshard, 1994, Melzer, 1995). Eucalyptus spp. leaves are high in materials that are toxic and indigestible for other mammalian species (secondary plant metabolites, lignin, cellulose) in addition to holding a low nutritional value (proteins, minerals, digestible carbohydrates). Generally, Eucalyptus leaves contain ca. 50% water, 18% fibre, 13% tannins, 8% lipids, 5% carbohydrates (sugars and starches), 4% proteins and 2% minerals (Blanshard, 1994)

1.2.3 Home ranges

Koala home ranges can vary according to many factors, including population density, sex, season, foliage and habitat quality (Moore et al., 2010). In QLD, the literature reports highly variable koala home ranges, from 8-16 ha in Brisbane, 15-34 in the South-East, to 101-135 in the central part of the state (Ellis et al., 2002a, White, 1999, Thompson, 2006), with males typically occupying larger territories than females. In SEQLD, the estimated distance of dispersal between natal and breeding home ranges for both sexes averaged 3.5 km and it is usually performed by sub-adult koalas during the early breeding season (August-January). Koalas maintain a relative structural population stability across time (Mitchell, 1989), with

13 males remaining predominantly solitary (Mitchell, 1989) and dispersing during specific times of the year for breeding. Dispersal can trigger high mortality rates in urbanised areas particularly in males, as they disperse in a higher proportion covering longer distances compared to females (Dique, Thompson, Preece, de Villiers et al., 2003). Subadults are the primary source of gene flow between koala populations in Queensland (Tucker et al., 2007); thus, their unnatural removal compromises the long-term viability of subgroups and could display population-level effects.

1.2.4 Reproductive cycle

Koalas are seasonally polyestrous (Johnston et al., 2004).The breeding season of koalas extends from September to February, with births occurring from October to April. Because they are primarily nocturnal, mating occurs at night. Koalas are generally uniparous (Blanshard and Bodley, 2008) with pregnancy lasting for 34-36 days. Juvenile koalas remain in the pouch for around 6 months, time after which the joey emerges and starts riding in the back until a year of age (Martin and Handasyde, 1995).

1.3 Koala Distribution in Queensland

The koala’s geographic range has contracted markedly since European settlement (Gordon et al., 2006) leading to marked variation in koala densities across the country (Table 1-2). In QLD calculating an accurate estimate is complicated by multiple factors influencing koala densities (e.g. climate, deforestation), as well as by the scarcity of data regarding abundance and distribution over far-reachign areas (Threatened Species Scientific Committee, 2012; Seabrook et al., 2011, Gordon et al., 2006). Currently, koalas occur in north-eastern, central, and SEQLD with fragmented populations in the western part of the state. The most recent estimates predict the highest koala densities in coastal regions of SEQLD, more precisely in the central and southern areas, with densities of 0.04 koalas per hectare across all regions (Rhodes et al., 2015). The SEQLD koala population inhabits forest and woodland fragments primarily in Moreton Bay, Noosa, Ipswich, Gold Coast, Redland, Logan and Brisbane LGAs (Department of Environment and Heritage Protection, 2015b). Studies utilising faecal pellet surveys have estimated that throughout time, distribution has not varied as much as koala numbers, which have dropped from 59,000 in 1995 to 11,600 in 2009 in south-west QLD (Seabrook et al., 2011). Koala population densities in the state are varied, and include 0.02- 2 koalas per hectare in Springsure, 0.4 per hectare in Redlands (coastal), up to 1.6 in Oakey (Gordon et al., 1990, White and Kunst, 1990). The koala decline in QLD, particularly in the

SE has a direct correlation with the amount of cleared habitat especially that of eucalypt/acacia ecosystems , with an estimated decline in occurrence of ~27% and 31% in occupancy (Gordon et al., 2006). Koala densities and distribution due to changes in dispersal are affected by the contraction of koalas to riparian habitats, particularly during droughts (Seabrook et al., 2011).

Table 1-2 Koala population density distribution in states and territories based on reported government and non-governmental records. High Low State Densities Densities Reference (koalas/ha) (koalas/ha) Queensland 1-3 0.01 (Melzer et al., 2000b) New South Wales 3 0.006 (Jurskis and Potter, 1997) Australian Capital Territory - very low Victoria <1 (Menkhorst, 2004) South Australia 5.5 (Masters et al., 2004)

1.4 Habitat-related threats to koala populations 1.4.1 Habitat loss

Maintaining good quality habitat is key for the preservation of koalas. This factor is the primary determinant of koala presence (e.g. foliage cover) in a given area as well as securing gene flow (Rhodes et al., 2006, Dudaniec et al., 2013). Notwithstanding the conservation actions by federal and state governments as well as legislative protection, the habitat of the koala has continued to deteriorate. Habitat degradation is still ongoing in several locations (Maxwell et al., 1996) specifically in QLD and NSW, for urban expansion and densification (coastal) from an increasing human population, as well as for agriculture (inland).

Habitat loss influences koalas directly through loss of shelter and browse, leading to a patchy distributions, and indirectly by forcing dispersal to substantial non-forested habitat causing environmental stress and increased trauma from fragmentation by linear infrastructure (e.g. roads, freeways) (Brown and Grice, 1984, Canfield, 1987d, Ellis et al., 1993, White and Timms, 1994, de Oliveira et al., 2014, Department of Environment, Water, Heritage and the Arts, 2014). How koalas respond to these indirect sources of stress from habitat loss have been previously ignored. A prolonged activation of the adrenocortical response may lead to a loss of body condition and indirectly result in increased disease

15 incidence (White and Timms, 1994). Ecosystem degradation also leads to less-publicized threats in urbanised landscapes (e.g. drowning in backyard pools, and overcrowding leading to greater changes of pathogen transmission (Protection, 2017, Lee et al., 2010).

1.4.2 Climate change, droughts and bushfires

The impacts of a changing climate on koalas include alteration of the structure and quality of habitat from increased frequency of droughts, bushfires, alteration of the chemical composition and browse structure in plant communities, changes in weather patterns (e.g. temperature, rainfall, humidity), and altered sea level levels affecting coastal populations (Department of Environment, Water, Heritage and the Arts, 2014, Lunney et al., 2014, Lunney et al., 2012). Low rainfall results in decreased moisture and leads to higher faecal cortisol levels, a sign of physiological stress in koalas inhabiting the arid edges of their range (Davies et al., 2013). Bushfires are events of random occurrence that directly impact koalas by causing mortality in high numbers (Lunney et al., 2007). Surviving koalas in unburnt zones may repopulate burnt areas following vegetation recovery. Climate change also influences plant nutritional quality as increased atmospheric CO2 levels lead to elevated tannins and lower protein levels reducing the nutritional quality of Eucalyptus species, resulting in nutrient shortages, malnutrition and starvation (IUCN, 2009). As extrinsic stressors, droughts can promote juvenile koala mortality if maintained over a sustained period further worsening declines either from mortality or increased dispersal (Department of Environment, Water, Heritage and the Arts, 2014, Lunney et al., 2014). For instance, significant changes in QLD koala occupancy have been associated with drier and hotter than average summers, worsened by vegetation clearance (Seabrook et al., 2011b).

1.5 Conservation actions

Each state currently lists conservation plans in place to secure koala conservation and management. Because habitat loss and fragmentation are currently recognized as main driver of mortality and injury in koalas in QLD, the Department of Environment and Heritage Protection – DEHP, has responded by having in place the South East Queensland Koala Conservation State Planning Regulatory Provisions (SPRP). The legislation controls development during the assessment stage, in particular, covering areas where koalas are at most risk – Koala Coast and Pine Rivers areas (Government, 2010). In QLD, each council may also have their own koala conservation plans, particularly in those with known koala populations (Redlands City Coucil, 2017). Some of these plans may include: protect and

16 reinstate koala habitat by planting food trees and acquisition of environmentally significant land, public education and community awareness campaigns, protection of native vegetation, construction of fencing and underpasses to enhance koala movement across roads (Department of Transportation and Main Roads, 2010), encouraging research into koala movements and population statistics, or habitat restoration and linkage strategies (Brisbane City Council, 2017).

1.6 Main Koala Threats from disease and injury 1.6.1 Diseases 1.6.1.1 Chlamydiosis

Disease is considered the largest contributor to koala mortality in SEQLD, followed by mortality by natural deaths, vehicles and dog predation (Rhodes et al., 2011). Few disease- causing agents in wildlife have been described as thoroughly as Chlamydia pecorum/Chlamydia pneumoniae, the causative agents of koala chlamydiosis. C. pecorum is thought to have been introduced by cattle and sheep crossing during the last 300 years to koalas (Polkinghorne et al., 2013), whereas C. pneumoniae is believed to have been in Australia an unknown period of time as multiple wildlife reservoirs have been found (Bodetti et al., 2002). Chlamydiosis causes four well-described clinical syndromes in koalas comprising infections in the ocular, urinary, and reproductive systems (Blanshard and Bodley, 2008), in addition to a rhinitis/pneumonia complex in which the role of Chlamydia is starting to be unveiled (Mackie et al., 2016).

Chlamydia spp. have undergone significant taxonomic reclassifications (Glassick et al., 1996). During 1999 another taxonomic revision was recommended, following the sequence analysis of the 16S and 23S rRNA genes of the Chlamydiales order (Everett et al., 1999), suggesting that the two species in koalas be moved to a new proposed genus, Chlamydophila. The new nomenclature was not thoroughly adopted, resulting in the merging of all current nine species back into the Chlamydia genus (Stephens et al., 2009).

In general, C. pecorum, the most dominant infecting species, is considered the most pathogenic of the two (Jackson et al., 1999), particularly if there is a chronic active infection in the female reproductive tract which leads to high shedding (Wan et al., 2011). Thus, the main concern with Chlamydia is that urogenital infection results in impairment of reproduction in koalas (i.e. chronic bursitis/metritis and orchitis/epididymitis with associated secondary infections potentially causing local extinctions (Johnston et al., 2015, Obendorf,

1983)). Nonetheless, some Chlamydia-positive koala populations have actually reached over-abundant numbers ( Department of Environment, Water, Heritage and the Arts, 2014). In fact, recent population modelling estimates conducted in SEQLD determined that annual birth rates were reasonably high and that chlamydiosis does not appear to be limiting the regional growth rate (Rhodes et al., 2011).

Severe manifestations of chlamydiosis are thought to be low in wild koalas and thus not perceived as a threat to koala populations. Nonetheless, the potential for extinction of sexually-transmitted pathogens such as Chlamydia is considered high, particularly because a reproductively suppressed host can remain in the population spreading the pathogen (McCallum, 2012). Although infected individuals may remain in populations enhancing transmission, there are discrepancies in chlamydial disease occurrence. New insights into the Chlamydia and KoRV relationship indicate that overt disease appears to be enhanced by co-infection with subtype-specific (exogenous) KoRV (Chaban et al., 2017, Waugh et al., 2017). Opinion on the frequency of severe disease manifestations or the true wild population impact are varied; some studies in QLD indicate indirect relationships between the prevalence of cystitis, population density and fecundity rate (drop from a mean of 84% to 41%) leading to shorter generation times, lower recruitment and immigration (Gordon et al., 1990, Martin and Handasyde, 1990). Others consider that overt chlamydiosis is a weak predictor of population change, highlighting the role of co-occurring stressors (co-infections, overcrowding, habitat fragmentation) as better predictors (McCallum, 2012, Martin, 1985).

A major cause of morbidity from chlamydia is urinary tract infection which manifests as cystitis, urethritis, ureteritis, or nephritis (Brown et al., 1987). Chlamydial infection can result in mucosal hyperplasia or progress to mucosal and submucosal fibrosis (Speight et al., 2016, Burach et al., 2014). Intramural fibrosis in the urinary bladder commonly leads to incontinence and urine staining of the common vestibule, resulting in what is colloquially known as ‘wet bottom’. Upper urinary tract disease commonly results in hydroureter or hydronephrosis, ureteritis and chronic interstitial nephritis (Canfield, 1989).

Chlamydial infection in the eyes may be unilateral or bilateral, with chronic disease being frequently bilateral. Acute cases can display a serous discharge and may be partly closed with chronic cases showing a markedly inflamed conjunctiva with a granular surface. If there is development of keratitis, it may extend centrally making the cornea appear slightly blue. Pannus and corneal opacity will develop later severely impairing vision (Hirst et al., 1992, Cockram and Jackson, 1981). 18

The last clinical syndrome appears to be associated to chlamydial disease is the rhinitis/pneumonia complex, recognised as a spectrum of manifestations. Diseases within this clinical syndrome can manifest in the upper or lower respiratory tract and be caused by primary pathogens or secondary opportunists. Historically, autopsy surveys have detected pulmonary conditions more often in QLD and NSW than VIC (Pratt, 1937), only to be reported as a common post-mortem finding in the same state four decades later (Butler, 1978). A possible association with the order Chlamydiales was suggested following the isolation of ‘Chlamydia psittaci’ from one case out of 12 koalas suffering from rhinitis (Brown and Grice, 1984).To date, only a single case report has been able to detect a direct association of C. pecorum infecting bronchiolar epithelial cells in a koala dying from respiratory disease (Mackie et al., 2016).

1.6.1.2 Other bacterial infections

Although chlamydiosis remains the most studied bacterial disease in koalas, other multiple infectious agents are capable of inducing disease but are considered less significant. B. bronchiseptica is recognized as an important pathogen for captive koalas (McKenzie et al., 1979). Pulmonary autopsy findings include consolidation and pus affecting the lobes, with multifocal solid grey masses in major airways microscopically characterised by suppurative bronchopneumonia and fibrinopurulent pleuritis, sometimes with necrosis along intralesional Gram-negative bacteria (McKenzie, 1981, Oxenford and Canfield, 1986, McKenzie et al., 1979). The Moggill Koala Hospital has vaccinated orphaned koalas for the past two decades utilising Canvac CCI BB (inactivated B. bronchiseptica cell free extract) which may have contributed to the decline in free-ranging Bordetella cases for SEQLD koalas (Pers. Comm. A. McKinnon).

Many bacteria have been implicated as common opportunistic agents in the koala reproductive tract, ,occasionally leading to seasonal reproductive failures (Munday, 1978), including E. coli, Corynebacterium, Bacterioides spp. and staphylococcal infections resulting in suppurative vaginitis likely initiated by chlamydial infection (Obendorf, 1993, Osawa et al., 1992, Canfield et al., 1983, Munday, 1978). Corynebacterium can be sexually transmitted as it has been isolated in the semen and prepuce of clinically-healthy males, a cause of concern of ascending infection and infertility in nboth sexes (Johnston et al., 1998). Despite being known as a hostile environment for bacteria, infections in the pouch epithelium leading to dermatitis can also affect koala reproductive success. Bacteria such as E. coli and Pseudomonas fluorescens have been isolated from a large pustules in the teats (Blanshard, 19

1993). Koalas with pouch young may get more readily stressed if handled, triggering disease from stress, as it has occurred in some melioidosis cases (Hirst et al., 1992, Ladds et al., 1990).

Due to the close relationship the urinary and reproductive tract have, bacteria affecting urinary organs are of interest, such as Leptospira, despite the low seroprevalence in koalas (Milner et al., 1981). In mammals, leptospirosis causes multifocal interstitial suppurative nephritis, resembling findings of renal chlamydiosis, although no reports of clinical disease or leptospire-associated lesions have been documented in koalas. However, leptospirosis lesions tend to be cortical and the arboreal lifestyle of the koala may preclude significant infection from maintenance hosts.

Despite the disproportionate size of the digestive system and in particular, the disproportionate size of the intestinal tract of koalas, it is not frequently affected by pathogens or diseases. Digestive bacterial diseases documented in koalas include hepatic Tyzzer’s disease (Clostridium piliforme) (Canfield and Hartley, 1991) and enterotoxaemia causing a generalised enteritis in SEQLD (Weigler et al., 1987). Although disease has not been demonstrated, Helicobacter spp. DNA has been detected from different portions of the gastrointestinal tract in koalas, possibly its detection being influenced by diet in this species (Coldham et al., 2013).

Scattered reports of mycobacterial infections are found in the literature for koalas. Infections by non-tuberculous Mycobacterium ulcerans characterised by indolent ulceration and underrunning of the cutaneous tissue and coagulative necrosis of the subcutaneous tissue have been identified in the skin and respiratory system in koalas associated with moist habitats (McOrist et al., 1985, Mitchell et al., 1984). Social behaviour causing skin abrasions from during fights or tree climbing has been documented as a risk factor (Mitchell et al., 1987). Infection by both M. ulcerans and M. scrofulaceum has been associated with a granulomatous reaction extended into the underlying bone (Mitchell et al., 1984).

1.6.1.3 Viral diseases

The most well-studied viral pathogen in koalas is the koala retrovirus – KoRV, a gammaretrovirus (Chappell et al., 2017) with five subtypes (A, B/J, C, D, E) closely related to the gibbon ape leukaemia virus (GALV) and the woolly monkey virus (WMV). KoRV has been found in the majority of free-ranging and captive koalas and it is through to have derived from interspecies transmission from an undetermined source. The research on this 20 pathogen has focused on providing a unique insight into a real-time endogenisation event into the host’s (the koala) genome, in addition to existing in both endogenous (transmitted through the germ line) and exogenous forms (contact from infected individuals). KoRV has been detected in northern koala populations with an estimated 100% prevalence in QLD, contrasting with the low prevalence in their southern counterparts (Simmons et al., 2012). Studies have suggested an association of KoRV with important diseases in koalas, specifically, of leukaemia and lymphoma in wild and captive koalas (Xu et al., 2013, Tarlinton et al., 2005), thought to occur via KoRV-induced immunomodulation (Denner and Young, 2013, Waugh et al., 2017). More recently, the association of chlamydial disease and neoplasia was established with an exogenous form (KoRV-B) in wild QLD koalas (Chaban et al., 2017).

Koalas with haemorrhagic enteritis have been described to display intranuclear inclusions resembling those seen in cases of feline enteritis; however their clinical significance was uncertain (Ladds, 2009c).

1.6.1.4 Fungal diseases

Cryptococcosis exists in two species – Cryptococcus neoformans and C. gatti, a species complex recently divided into seven species (Hagen et al., 2015). Cryptococcus is highly prevalent in the environment, resulting in a high rate of infection, primarily subclinical in nature in the koala (Krockenberger et al., 2002). All the clinical infections in the koala have been attributed to C. gattii, however both C. neoformans and C. gattii have been detected in the nasal cavity in koalas (Schmertmann et al., 2017, Martinez et al., 2017). Although cryptococcosis has been documented in other marsupials and mammals, the majority of reports relate to koalas. Cryptococcus is carried asymptomatically as evidenced in serological studies (Connolly et al., 1999, Krockenberger et al., 2002). Koalas are good amplifiers of cryptococci under certain conditions without succumbing to clinical disease (Payne et al., 2005). Clinically apparent infection appear to be opportunistic and presumably associated with underlying factors such as immunodeficiency, stress, retroviral infection or respiratory co-infection (Canfield et al., 1986, Worley et al., 1993, Hanger, 2003). Clinical forms identified in the koala are respiratory, disseminated and one with central nervous system (CNS) involvement, with clinical signs varying according to the system affected (Connolly et al., 1999).

Aspergillus sp. has sporadically been reported in koalas causing disease. In one case, it was cultured from lesions of a hand-raised koala with generalised severe seborrhoea, characterised microscopically by parakeratotic hyperkeratosis (Lucas, 2000). The second case was an ulcerative granulomatous enteritis (Lenghaus, 1990). Both cases had in common minimal inflammatory response, potentially indicating an underlying immunocompromise. Candidiasis has also been documented in scattered reports causing disease in koalas (Blanshard, 1994, Connolly, 1999), with a description of a stomatitis case (Connolly, 1999). Koalas are also susceptible to dermatomycoses (ringworm (Wood, 1978) from infection by Trichophyton mentagrophytes and Microsporum gypseum, primarily in captive locations (Canfield et al., 1992).

1.6.1.5 Parasitic diseases

Toxoplasmosis has been reported in koalas primarily in captivity (Hartley et al., 1990, Dubey et al., 1991) but is more important in macropods as it causes fulminant disease. Clinical signs can be unspecific involving fever and shallow breathing (Dubey et al., 1991) to sudden death (Canfield et al., 1990, Hartley et al., 1990). Grossly, lesions are also varied and range from no visible lesions to pulmonary congestion and oedema. Histopathology may reveal myocardial necrosis, pulmonary congestion with alveolar histiocytosis, germinal centre hypertrophy and macrophage hyperplasia in lymph nodes and spleen and congestion in adrenal glands (Canfield et al., 1990).

Koalas carry infections in their blood, mostly subclinically, by trypanosomes, which have been identified as Trypanosma irwini, T. gilletti, and T. copemani (McInnes et al., 2009, McInnes, Hanger, Simmons et al., 2011). Further research is needed to determine if a low number of infected koalas can manifest clinical signs compatible with trypanosomiasis (McInnes et al., 2009). Studies have demonstrated the disease potentiating role trypanosomes can have in cases of immunosuppression or concurrent disease, making Trypanosoma spp. exhibit condition-dependent virulence as reported in other species (McInnes, Gillett, Hanger et al., 2011).

1.6.2 Traumatic Injuries 1.6.2.1 Motor vehicle collisions

Trauma caused by collisions with motor vehicles (e.g. private vehicles, buses, trains) is an important cause of mortality in koalas throughout their range (Canfield, 1991, Obendorf, 1983), but more frequently associated with koalas inhabiting urbanised habitats (Phillips, 22

2000) resulting in an estimated 83% fatality rate (Dique, Thompson, Preece, Penfold et al., 2003). Koalas are colliding more frequently with motor vehicles during breeding periods or post-natal dispersal (Mitchell, 1990a, Logan and Sanson, 2002). Other factors known to influence the frequency of vehicle collisions are the density of koalas in the area, the traffic volume and number of roads, and the frequency by which koalas attempt to cross ( Department of Environment, Water, Heritage and the Arts, 2014). Increased road- associated mortality results from road building based on design and network planning not accounting for koala or other wildlife occupancy ( Department of Environment, Water, Heritage and the Arts, 2014). A study conducted in NSW determined that increasing the volume of the roads to accommodate a growing number of vehicles had less impact than building new roads irrespective of traffic volume (Rhodes et al., 2014). Alternatives to reduce the mortality of species roaming during dawn and dusk have been proposed in QLD, such as adopting daylight saving time, to move commuting times away from sunrise and sunset (Ellis et al., 2016). Despite the potential benefits of the implementation of this measure on koala conservation, careful consideration of the impact on diurnal wildlife is required.

1.6.2.2 Animal attacks

Mortality from animal attacks is considered one of the main contributors to koala mortality in SEQLD (Rhodes et al., 2011, Gonzalez-Astudillo et al., 2017). Because koalas have few natural predators, it is currently accepted that the primary predation perpetrators are introduced domestic and feral dogs (Martin and Handasyde, 1999). Koala habitat rapidly being urbanised leads to increased contact with and attacks by domestic dogs, whereas wild dog attacks occur while koalas are moving through fragmented habitat in rural regions (White, 1999, McAlpine et al., 2007) or even in peri-urban areas in QLD where dingoes are known to roam (Department of Employment, Economic Development and Innovation, 2014). Attacks to koalas can also occur in protected land (e.g. reserves, national parks) where dogs are known to roam ( Department of Environment, Water, Heritage and the Arts, 2014). The relative frequency of feral versus domestic dogs to koala trauma remains undetermined.

Koala-dog encounters mainly result in moderate to terminal trauma that requires either intensive care or euthanasia ( Department of Environment, Water, Heritage and the Arts, 2014). Stress induced prior to or by a prolonged hospitalization is thought to also play a role in the mortality of koalas while in care following a traumatic event (Obendorf, 1983). If a koala survives an attack, complications may ensue that require medical attention such as 23 secondary infections (Wigney et al., 1989). In addition, young koalas are particularly vulnerable to predation by multiple species. Some reports have documented back-young or recently weaned koalas being taken by Powerful owls (Ninox strenua), Wedge-tailed Eagles (Aquila audax), lace monitors (Varanus varius) and large goannas (Lee and Martin, 1988, Lunney et al., 1990, Phillips, 1990, Jurskis and Potter, 1997, Melzer et al., 2003). Cases of snake envenomation (Stewart, 2012) or predation by carpet pythons (pers. comm. R. Larkin) are known and likely are underreported causes of koala deaths.

1.7 Data Collection Systems for Animal Diseases

1.7.1. Overview of animal health surveillance

Animal health surveillance is a process in which data are collected, collated, and analysed to disseminate this information to authorities or other parties, so that targeted action can be carried out to control a disease or outbreak if it is found beyond a certain threshold (Salman, 2003b, Thrusfield, 2007, Sergeant and Perkins, 2015). Thus, disease surveillance is conducted for various purposes: as an early warning system to identify disease incursions, to provide evidence for the absence of animal diseases, for a population health assessment to describe extent or distribution of diseases, to identify new, exotic, emerging diseases or unusual epidemiological events or to evaluate disease control strategies (Salman, 2003b, Thrusfield, 2007, OIE, 2015). Disease monitoring is often used interchangeably with surveillance; however, monitoring only detects disease levels in a population and does not take action (Salman, 2003b). To carry out surveillance, suitable resources including personnel, equipment, legal frameworks, communication, and transportation networks are required. During the planning phase, determining the presence of the disease control infrastructure is also important. A disease control infrastructure takes into consideration two main components: trained personnel to recognize clinical signs in animals, control disease, and collect field samples; and diagnostic testing facilities, availability of local or regional laboratories, and ability to transport samples. The consideration of animal ethics and welfare holds relevance not only from a legal standpoint, but also from how it may impact components that are central to the study, such as sample size, which is usually subjected to justification. In regards to data collection, the recruitment of skilled personnel in computer database design and management is relevant to organize and provide maintenance on a database that will record data generated during surveillance. Lastly, the utilisation of specific communication channels between relevant parties needs to be discussed once the data collected has been reviewed and analysed (OIE, 2015). Surveillance system have been 24 established for major livestock diseases, but establishing disease surveillance in wildlife is challenging.

1.7.1.1 Passive and active surveillance

In passive surveillance, data collection does not follow a pre-defined plan so data is considered a by-product of other health-related investigative activities. In other words, passive surveillance data is the result of routine collection of information generated for another purpose (Sergeant and Perkins, 2015). For example, passive surveillance of livestock diseases might comprise data derived from the reporting of clinical or subclinical (suspect) cases to animal health authorities (Lilienfeld and Stolley, 1994). Other data for passive surveillance can be obtained from farmers, routine abattoir inspection, statutory notification of disease by private veterinarians, laboratory reports and findings from research projects (Thrusfield, 2007, Sergeant and Perkins, 2015).

Active surveillance is designed and commenced by the primary user of the data and it involves the collection of information on a disease or pathogen in an animal population (Sergeant and Perkins, 2015). This data collection method involves a deliberate and comprehensive effort to collect data on specific diseases, usually involving clinically healthy animals, being relevant in the surveillance for conditions in which subclinical carriers (e.g. maintenance hosts) predominate (Thrusfield, 2007). In some instances, it can also target specific diseases or pathogens, whereas in others it can be non-specific and efforts are directed at any significant disease that can be detected (Sergeant and Perkins, 2015). Active surveillance is also utilised to demonstrate disease free status (Thrusfield, 2007).

1.7.1.2 Strengths and limitations of passive and active surveillance

Passive surveillance is a good tool to provide a general picture of a disease in a population (Sergeant and Perkins, 2015) and baseline levels for diseases (Maas et al., 2016). Data collection involves a lesser cost than active surveillance (Thrusfield, 2007). By maintaining disease surveillance through routine data collection, authorities could be alerted to the presence of new and emerging pathogens. For this reason, passive surveillance is often described as the first step to novel and emerging disease identification an ‘early disease or pathogen warning system’, facilitating subsequent follow up with epidemiological studies to determine sources of infection (Russell and Franson, 2014). Active surveillance cannot target these conditions as the pathogen or disease is unknown (Thrusfield, 2007, Sergeant and Perkins, 2015). With passive surveillance, disease detection maybe more rapid, as the 25 data has already been generated, allowing veterinary authorities to act promptly. However, a disease or mortality event should be followed up with well-designed epidemiological studies (Russell and Franson, 2014). Nonetheless, passive surveillance can still provide information on trends and distribution of disease overtime, change in a disease pattern (Salman, 2003b), or frequency of occurrence of disease but this depends on the type of data available and the specific disease (Sergeant and Perkins, 2015). The frequency of data collected in passive surveillance is dependent on willingness, awareness level, or knowledge of health care professionals, farmers, and animal owners, to report o and allow the flow of data for the particular disease (Salman, 2003b). As passive surveillance does not actively survey the total population, there is a lack of denominator values, impeding the estimation of disease occurrence (Thrusfield, 2007). Additionally, because data collection is not consistent across different communities and for different diseases (Salman, 2003b), the development of comparative studies is hampered.

Active surveillance is planned and designed for a specific purpose, so data collection is controlled and of higher quality than passive surveillance (Sergeant and Perkins, 2015). Disease frequency with confidence intervals (or the confidence that disease would be detected if present) can be calculated as the population structure can be calculated from representative sampling in active surveillance program (Thrusfield, 2007, Sergeant and Perkins, 2015). With active surveillance, the best and most accurate information of known diseases or pathogens is provided. Sampling is generally carried out by veterinary authorities and guided by a statistical or probability-based sampling plan (Maas et al., 2016), but this is a costly approach (Nsubuga et al., 2006). An active surveillance system is considerably more time-consuming, and more expensive as it requires that the user of the data (veterinary authorities, government, industry representatives) carries the entire cost of the operation (Maas et al., 2016, Sergeant and Perkins, 2015). These costs will inevitably increase if the occurrence of the target disease is rare (Salman, 2003b). Because of its structure and the expenses incurred, active surveillance generally does not allow large numbers of animals to be processed. Additionally, this sampling method cannot be structured or planned to target new or emerging pathogens or diseases (Thrusfield, 2007).

1.7.2 Disease surveillance in wildlife populations

Disease surveillance of wildlife has usually focussed on diseases that are of major public concern, like those in iconic or threatened wildlife species or diseases of public health significance. Unfortunately, major obstacles in implementing wildlife disease surveillance 26 programs are limited political interest, as well as the difficulty of implementing such programs in free-ranging wildlife (Grogan et al., 2014). In Australia, some of the wildlife diseases that are actively researched include koala chlamydiosis (Quigley et al., 2018), chytrid fungus (Berger et al., 1998), devil facial tumour disease (McCallum, 2008), and the koala retrovirus (Tarlinton et al., 2006). Other pathogens causing infections in wildlife that receive considerable attention and resources hold public health and economic relevance, such as the Australian bat lyssavirus, Menangle virus, and Hendra virus (Black et al., 2008, Bunn and Woods, 2005).

Passive (Table 1-3) and active (Table 1-4) surveillance methods have been utilised in wildlife to collect data to monitor for disease or wildlife mortality dynamics. Wildlife disease surveillance is often based on opportunistic sampling of diseased or moribund animals, and does not follow a study design or implementation of the conventional techniques utilised for livestock or companion species (Duncan et al., 2008). Regardless of the sporadic nature of data collection, these opportunities provide a chance to investigate morbidity and mortality at the population level (Wobeser, 2005).

Certain aspects of disease surveillance in wildlife are equal or similar to those in domestic animals: from the pathogens, to the sampling techniques, or laboratory tests utilised to determine infection status. Nonetheless, there are challenges in conducting surveillance in wildlife compared to livestock. For instance, wildlife are not owned, an observation is limited, thus cases are not accompanied by clinical histories. Secondly, selecting the sampling size to evaluate is critical, as it must be representative of the population of interest, to avoiding random error or bias (Wobeser, 2007, Cooper, 2002). Thirdly, another limiting factor is laboratory tests utilised for diagnosis are validated for domestic species, so their suitability should be ensured prior to being applied to a wild species (Wobeser, 2007). Often, population data and other factors inherent to wildlife biology have to be provided by wildlife professionals so the outcomes of a wildlife disease surveillance program can be adequately analysed, interpreted and disseminated (OIE, 2015). Last but not least, funding is primarily structured to target impactful diseases for public health or the livestock industry (Ladds, 2009b). Funding for threats directed solely to biodiversity conservation are thus often restricted.

For wildlife, the organisational models to coordinate and run surveillance programs tend to be unique: Thus, examples of programs of disease surveillance in humans or livestock are not applicable because the networks are considerably larger and require different personnel 27

(OIE, 2015). Because undomesticated species are usually government-managed resources, the responsibility and success of a disease surveillance program is typically a shared coalition between many government (and occasionally non-governmental) sectors: agriculture, public health, tourism, veterinary services, border services, environment, etc. (OIE, 2015). Thus, a clear definition of responsibility is important during the planning stage. Due to the protected status of wildlife, approval of animal ethics and plausibly, other governmental permits are required. Lastly, the perceptions and support of the community need to be considered for the adoption and implementation of wildlife disease surveillance, particularly when private land is involved.

1.7.3 Role of Autopsy in Wildlife Passive Surveillance

Post-mortem or autopsy investigations play a vital role in general disease surveillance. In most instances, autopsy is the cornerstone of wild animal disease diagnosis as it is often the only investigative method that can be applied following wildlife deaths, playing an important role identifying novel or exotic pathogens and diseases. During an autopsy, all causes of disease and death identified in a wild animal are recorded. Gross and histopathological findings in combination with specialised testing (microbiological culture, PCR, toxicological testing (OIE, 2015)) would ideally lead to a final diagnosis, as in only in a small number of cases background clinical data is available to correlate with the epizootiology of the disease event. There are various factors influencing outcomes of wildlife autopsies, including acquiring a representative sample size, logistics, autopsy technique skill set, or funding, amongst others. Despite these challenges, there are many benefits in conducting autopsies: there is increased sensitivity in detection of disease and injury compared to standard clinical care, and it improves identification rates of primary cause of disease and injury if koalas are admitted dead. Autopsies allow the generation of ‘health statistics’ (basic health data) that would be otherwise unattainable, and the determination of comorbidities, as the goal is to evaluate all organs for disease and injury, not just confirm the presence of one aetiology. Lastly, the characterisation of novel conditions or pathogens expands the current body of knowledge for the species, and leads to subsequent studies into the pathogenesis of such conditions to determine potential population-level impact. Without a structure to capture surveillance data, the rationale behind the submission of any wild animal for an autopsy is often tied to significance of the event the animal is affected by, for instance, exposure to rabies, or a communicable disease or unknown cause of death (Duncan et al., 2008). Nonetheless, zoological institutions in Australia have been incorporated into a surveillance program so disease and trauma information can be included 28 into a national database (Cox-Witton et al., 2014), recording cases for many species, including koalas. Analysis of data generated by autopsy surveillance studies provides opportunities to obtain an exclusive population snap shot, determining multifactorial causes of diseases, co-occurring conditions and injuries, and the relative contribution and quantification of risk factors impacting wildlife diseases and mortalities.

1.7.4 Surveillance of disease and mortalities in koalas

Sick, injured or orphaned koalas from all local government areas in SEQLD are admitted into hospitals, sanctuaries (Currumbin, 2017), or zoo clinics by members of the public, biologists or representatives of wildlife agencies (Department of Environment and Heritage Protection, 2015a). Since 1997, these koala admissions have been entered in a large clinical database managed by the Moggill Koala Hospital, supervised by the Department of Environment and Heritage Protection – DEHP since 1991. Two other main wildlife hospitals (Australia Zoo and Currumbin Wildlife Sanctuary) report their data on koala admissions to MKH. Wildlife hospitals have recorded valuable information about causes of disease and injury for SE QLD koalas when providing specialised treatment for decades, amassing valuable koala health data useful for epidemiological analyses. In 2014, the clinical database was transformed into a modern online SQL database denominated ‘KoalaBASE’, improving reporting features (http://www.ava.com.au/node/27181). Koala hospital accessions, along diagnostic laboratory data can be utilised for passive surveillance to conduct inexpensive monitoring, offering unprecedented detail of data with ample population coverage, derived from a strong network of specialised koala health care providers for the SEQLD. Establishing a disease surveillance system provides many benefits for koala conservation, and a key factor for the success of any wildlife surveillance program is the incorporation of a strong pathological component through systematic autopsies with standardised terminology so consistent investigations of koala mortality and morbidity can be carried out.

Table 1-3. Examples of global morbidity and mortality wildlife studies utilising passive surveillance methodology.

Institution Species Location Submitter generating Study goal Timeframe Reference data 2000 - Monotremes Australia Not specified Sanctuary Categorise major causes of admission (Scheelings, 2016) 2014 Aquatic Temporal & demographic disease 2007 - Australia Not specified Sanctuary (Hoque et al., 2012) birds trends 2010 Retrospective evaluation of diagnostic Wild Licensed Health data 1992 - (MacDonald et al., Canada Turkeys hunters Centre to determine population health status 2014 2016) following reintroduction efforts Characterise anthropogenic impact Factors influencing morbi-mortality risks 2005 - Turtles USA Public Clinic (Sack et al., 2017) Mortality causes, treatment outcomes 2014

Rwanda Causes morbidity & mortality in infant Mountain - 1967 - Uganda Trackers mortality and correlate histological (Hassell et al., 2017) Gorillas 2014 DRC results with clinical morbidity State biologists Avian Diagnostic Retrospective study of mortality causes 1976 - (Gonzalez-Astudillo et USA species Other Laboratory & assess mortality patterns over time 2012 al., 2016) stakeholders Mortality causes across flyways Diagnostic 1975 – (Russell and Franson, Raptors USA Not specified Laboratory Effect of lead bullet ban on proportion 2013 2014) of raptors diagnosed with lead poisoning Public Game wardens Identify risk factors, frequencies, and (Richards et al., 2005) Law- Institutional 1993 - Raptors USA common species affected by gunshot enforcement Network 2002 injuries Licensed hunters

Table 1-4. Examples of global morbidity and mortality wildlife studies utilising an active surveillance methodology.

Species Location Submitter Study goal Timeframe Reference

Investigate seroprevalence of Bat EBLV-1 among native bats Bats France specialists 2004 - 2009 (Picard-Meyer et al., 2011) to determine which bat sp Others? were involved in transmission Examine fresh avian faeces and urine for the presence of Wild avian influenza virus & Not specified aquatic compare 2006 - 2007 (Pannwitz et al., 2009) Germany Researchers? birds results to cloacal and orophagyngeal swabs in another cohort Estimate incidence of Community suspected and confirmed village human and animal rabies Mammals Kenya 1992 - 1993 (Kitala et al., 2000) leaders cases & compare estimates to government existing passive surveillance system Describe abundance shifts in New Not rabbit populations & assess Rabbits 1998-2001 (Henning et al., 2006) Zealand Specified serological haemorrhagic virus status

1.8 Study Objectives and Thesis Structure

The aims of this thesis are to determine the primary causes of morbidity and mortality in koalas in South-East Queensland using passive surveillance data. Chapter 1 covered an extensive review of the current knowledge of koala identifying knowledge gaps. This study has three broad aims with the hope to address the aforementioned shortfalls as follows:

1) Retrospectively summarize the main causes of submission of koalas to South-East Queensland wildlife hospitals during 1997-2013, and establish significant associations between the koala hospital outcome with the major clinical diagnoses to determine competing risks of morbidity and mortality using a government-managed database.

Despite the large number of records and timespan of the database, a lack of standardisation in autopsies and recording terminology needed to be addressed to accurately determine the causes of morbidity and mortality in koalas using this dataset. A retrospective analysis utilising epidemiological tools that modelled demographic and temporal risk factors in detail was conducted to determine the major drivers of decline affecting koalas in this geographic region during a period of major population decline.

2) Conduct a prospective morbidity and mortality analysis involving standardised post- mortem procedures and systematic data recording to accurately report causes of death in koalas admitted to wildlife hospitals in South-East Queensland from 2013-2016.

Large databases establish the foundation for passive surveillance and provide detailed information for wildlife studies. However, when data recording is performed in a non- standardised manner, as in the retrospective study, the identification of true disease causes, severity and burden is impeded. We applied systematic autopsies and standardised terminology to koala carcasses routed from three major hospitals to identify the pattern and type of comorbidities affecting koalas admitted to hospitals, as well as to determine the causes of mortality affecting koalas during the period following the population decline analysed in the retrospective study.

3) Utilise a model-based approach to estimate agreement of diagnoses made by wildlife clinicians and pathologists on causes of koala disease and injury in South-East Queensland from 2013-2016.

We propose a framework to evaluate the accuracy in diagnosis made in koalas by wildlife clinicians and pathologists as a way to implement quality control. Both quality control 32 mechanisms and clear case definitions are key factors involved in the maintenance of a proper surveillance system. Additionally, measuring the discrepancy in diagnoses made by wildlife clinicians and pathologists is essential to determine the causes of the koala decline, as well as providing a form of validation of clinical diagnoses.

Chapter 2 Decline causes of Koalas in South East Queensland, Australia: a 17-year retrospective study of mortality and morbidity.

2.1 Abstract

Koala populations are in catastrophic decline in certain eastern Australian regions. Spanning from 1997–2013, a database derived from wildlife hospitals in southeast Queensland with N = 20,250 entries was classified by causes of morbidity and mortality. A total of 11 aetiologies were identified, with chlamydiosis, trauma, and wasting being most common. The clinical diagnosis at submission varied significantly over the observation period. Combinations of aetiologies were observed in 39% of koalas submitted, with chlamydiosis frequently co-occurring. Urogenital (cystitis 26.8%, bursitis 13.5%) and ocular (conjunctivitis 17.2%) chlamydiosis were the most frequently diagnosed representations of the infection. Approximately 26% of submissions comprised koalas involved in vehicle accidents that were otherwise healthy. Age and sex of the koala as well as season and submission period were compared for the case outcomes of ‘dead on arrival’, ‘euthanized’, or ‘released’ for the four most common clinical diagnoses using multinomial logistic regression models. Exploratory space-time permutation scans were performed and overlapping space-time clusters for chlamydiosis, motor vehicle traumas and wasting unveiled high risk areas for koala disease and injury. Our results suggest that these aetiologies are acting jointly as multifactorial determinants for the continuing decline of koalas.

2.2 Introduction

The koala (Phascolarctos cinereus) is a folivorous marsupial whose distribution is tied to its food source, the Eucalyptus forests in Australia. Eastern Australia has ca. 40% of the ideal habitat for koalas (Phillips, 2000), corresponding to tropical, subhumid and semi-arid regions in the state of Queensland (QLD). Since the 1990s, koalas of South East Queensland (SEQLD) have experienced a severe population reduction. The most recent estimate in the Koala Coast reported a population decline of 80% (Rhodes et al., 2015). Similar population declines experienced in New South Wales (NSW) and Australian Capital Territory (ACT), have contributed to listing the species as “threatened” according to the Australian Federal Environment and Biodiversity Protection Act 2012. In general, marsupials have declined substantially in Australia: 25% of their original range has been cleared and 24 marsupial species have been declared extinct (Johnson and Isaac, 2009). As a charismatic animal, the koala faces infectious (Legione et al., 2016) and non-infectious (Speight et al., 2014b,

Griffith et al., 2013) challenges that need to be closely monitored. In particular, the long-term conservation of the koala is threatened by habitat encroachment and rapid urbanization (Vitousek et al., 1997, Schipper et al., 2008). Other threats include over-browsing of habitat resulting starvation, extreme droughts likely to be linked to climate change and events of stochastic occurrence such as bushfires (Wallis, 2013), trauma related to anthropogenic activities, such as injury and mortality resulting from predation by domestic animals (Obendorf, 1983, Canfield, 2008, Department of Environment, Water, Heritage and the Arts, 2014) and motor vehicle (MV) collisions (Narayan et al., 2013). Furthermore inbreeding depression and diseases (mainly by infection with Chlamydia spp. and potentially, koala retrovirus- KoRV) (Narayan et al., 2013, Environment, 2014, Melzer et al., 2000b, Dickens, 1975) also jeopardize the survival of this species. Studies aiming at improving our understanding of wildlife diseases are recognized as a powerful way of monitoring ecosystem health and recent developments in host-pathogen ecology, transdisciplinarity, and involvement of endangered species in outbreaks has generated more attention for wildlife health (Daszak et al., 2000). Public concern and investment in koalas is high, however there are discrepancies in estimated and reported conservation status (Woinarski and Burbidge, 2016), which affect overall population management efforts (Martin et al., 2008). Despite the listing of koalas as vulnerable to extinction in QLD, NSW and ACT20, there is ambiguity regarding QLD population estimates (Threatened Species Scientific Committee, 2011), particularly in inland bioregions.

2.3 Materials and Methods 2.3.1 Study population

Data for this study originated from clinical database managed by the MKH, which has been under the supervision of the Department of Environment and Heritage Protection (DEHP) since 1991. In this database, sightings and submissions of koalas to wildlife hospitals are recorded. Sightings are defined observations of free-ranging koalas made by the public or rangers, but that were never admitted to the hospital. The MKH receives submissions from the public encompassing an area of 349,729 km2 in Queensland, grouped by local government areas (e.g. councils). Two other major wildlife hospitals in SEQLD (AZWH and CSWH, amongst others) report their koala submissions to MHK. Data entry in the clinical database commenced in 1997. In 2014, the clinical database, originally developed in Microsoft Access 2007, was converted into a state-of-art online SQL database called

‘KoalaBASE’, with significantly improved data entry and reporting features (http://www.ava.com.au/node/27181).

For submitted individuals, upon reception, mandatory information is gathered in a record sheet, including the exact koala collection location, demographic data, and previous identification information (if any). A physical examination is conducted, in which any clinical signs, injuries and identification features, including microchip number are noted. Koalas are aged according to the wear of the last pre-molar and first molar, or by identifying the individual with previous submissions. Any koalas that were rehabilitated were microchipped prior to release. Sightings, uncaptured, or captured and instantly released, relocated, submissions with missing information, healthy, healthy orphans, and koalas from states other than Queensland were excluded from the analysis.

2.3.2 Quantitative analysis

Data mining was performed using Microsoft Excel® 2013. Re-categorization of outcome (‘dead on arrival’, ‘euthanized’, ‘released’) and aetiology of submission into pre-determined categories was necessary prior to quantitative analysis. A classification chart for the following aetiological causes was developed: 1) MV Trauma, 2) Chlamydia-like signs, 3) Chlamydia-like signs & Wasting, 4) Trauma by animal attack, 5) Chlamydia-like signs & MV trauma, 6) Undetermined, 7) Trauma by other causes, 8) Chlamydia-like signs & Senescence & Wasting, 9) Chlamydia-like signs & Trauma by animal attack, 10) Wasting and 11) Other.

2.3.3 Koala submissions

Koala submissions were then classified in two ways, a) aetiologies b) clinical syndromes. An ‘aetiology’ describes a discrete (single) cause of submission of a koala to a hospital and is used to quantify the overall frequency of aetiologies, thereby highlighting the temporal and spatial importance of individual aetiologies.

Since koalas can be exposed to several aetiologies, multiple occurrences of different aetiologies can be recorded for the same koala (e.g. a koala can have occurrences of the aetiology ‘Chlamydia’ and of the aetiology ‘Trauma’). Therefore the term ‘clinical syndrome’ summarizes what ultimately caused the disease/injury and death of koalas (which could be a single aetiology or a combination of aetiologies). The term ‘clinical syndrome’ is used to describe the frequency of the conditions that lead to the disease/injury and death of koalas

36 and to model risk factors associated with outcomes (‘dead on arrival’, ‘euthanized’, ‘released’) for of each diagnosis.

Data analysis was conducted using the software package STATA v. 14.0 (StataCorp, USA). Data were arranged in a contingency tables to describe the frequency of occurrences of aetiologies, clinical syndromes and outcome (‘dead on arrival’, ‘euthanized’, ‘released’) across koala-specific (age class, sex) and temporal (year period, season) explanatory variables. Explanatory variables were categorised as following: dichotomized age [young (baby or joey, young and subadult), adult], sex [female, male], season [January-March, April- June, July-September, October-December], and submission year periods [1997-2001, 2002-2005, 2006-2009, 2010-2013]).

We explored the association of koala-specific and temporal risk factors with the outcome (‘dead on arrival’, ‘euthanized’ and ‘released’) for the four most frequent clinical syndromes using multinomial logistic regression models. ‘Released’ koalas were considered the base category used for comparison for koalas ‘dead on arrival’ and koalas that were ‘euthanized’. Approximately 2% of records had missing risk factor information had to be excluded from the analysis. A multiple Wald test was computed to evaluate the statistical significance of all categories together for any categorical risk factor variable (Doohoo et al., 2009). Variables for which p < 0.10 in the univariate analysis were considered for multivariable analysis.

A backward and forward model selection process was run, and p-values, RRR with 95% confidence intervals for each explanatory variable were calculated. The stepwise selection process was stopped once all covariates were significantly (p < 0.05) contributing to the model. First order interactions between explanatory variables were also explored. Only main effects were tested and explanatory variables significant at p<0.05 were maintained. If a variable was not selected for the initial multivariable model, it was added back to the model, and significant variables (if so) were retained. This process allowed the identification of explanatory variables that may not influence the outcome by themselves, but contribute significantly to it when combined with other explanatory variables(Bursac et al., 2008). The Bayesian information criterion was used compare to non-nested models for the same clinical syndrome. Goodness-of-fit statistics were calculated to assess how well the final model fit the data. Predicted probabilities for each outcome were calculated using the ‘margins’ command in STATA, holding the remaining variables in the model at their means.

A case-based space-time permutation model (Kulldorff et al., 2005) using Kulldorff’s SaTScan™(Kulldorff and Information Management Services, 2009) software was developed to identify spatial-temporal clustering or ‘hotspots’ of occurrences of aetiologies. Since population-at-risk data was not available, a large number of random permutations of the spatial and temporal attributes of each koala case in the dataset was produced to calculate the scan statistics. Most likely clusters were calculated for each simulated dataset and statistical significance was evaluated using Monte Carlo hypothesis testing (Dwass, 1957).

2.4 Results 2.4.1 Study population

A dataset of 42,257 records from three hospitals (Moggill Koala Hospital - MKH, Australia Zoo Wildlife Hospital - AZWH, Currumbin Wildlife Sanctuary Hospital - CWSH) in SEQLD spanning from January 1, 1997 through December 31, 2013 was utilized for analysis. A total of N=22,007 koala records were excluded from the analysis to reach a study population of N=20,250. Exclusions included sightings (N=15,113), uncaptured (N=1,044) or instantly released (N=1,127) koalas, relocations (N=650), records with missing information (N=3,234), healthy (N=403), healthy orphaned (N=137) koalas and submissions from states other than QLD (NSW, N=299). Demographically, there were similar numbers of males (N=10,232, 50.9%) and females (N=9,605, 47.8%), and there was no sex information for N=252 (1.25%) records. Adults (N=16,094, 79.5%) were present in much greater proportion than young koalas (N=4,156, 20.5%).

2.4.2 Aetiologies

A total of 11 aetiologies were identified, which occurred 41,606 times in the 20,250 koalas monitored over the study period. The most common aetiology was Chlamydia-like signs (N=21,619, 52.0%), followed by MV trauma (N=6,432, 15.5%), and wasting (N=5,935, 14.3%), amongst others (Table 2-1).

Within the Chlamydia-like signs, we found cystitis to be the most common sign (N=5,422, 26.8%), followed by conjunctivitis (N=3,485, 17.2%), bursitis (N=2,741, 13.5%), pneumonia (N=2,493, 12.3%), nephritis (N=2,432, 12%), and metritis (N=1,022, 5%). Due to the discrepancy in overall body condition in acute vs chronic chlamydiosis, koalas within the Chlamydia-like signs category were grouped separatedly from Chlamydia-like signs & Wasting. Male-specific Chlamydia-like signs were diagnosed at a very low frequency overall (e.g. prostatitis N =9, orchitis N =3). ‘Other diseases’ observed included: dermatitis (N =135), 38 septicaemia (N =68), ectoparasitosis (ticks, N =49), and hepatitis (N =47), amongst others. A total of N=900 occurrences of undetermined cause were found.

There were variations in the occurrences of all aetiologies over time (Figure 2-1). The occurrence of Chlamydia-like signs has maintained an overall proportion above 50% across all years of the database, despite major variations in nearly two decades. Koalas are being presented less commonly with MV trauma (2009-2013) compared to a decade ago (1997- 2001), and current admissions include more emaciated individuals compared to the mid- 1990s.

Table 2-1 Frequency of occurrences of diagnoses (with a CI 95%) in koalas submitted to wildlife hospitals from 1997 through 2013 (No. of occurrences of aetiological causes = 41,606).

Diagnoses Frequency Proportion CI 95% Chlamydia-like 21,619 52.0 51.5-52.4 Trauma Motor Vehicle 6,432 15.5 15.1-15.8 Wasting 5,935 14.3 13.9-14.6 Trauma animal attack 2,154 5.2 5.0-5.4 Trauma other causes 1,168 2.8 2.7-3.0 Senescence 1,160 2.8 2.6-3.0 Other diseases 1,099 2.6 2.5-2.8 Undetermined 900 2.1 2.0-2.3 Blind 678 1.6 1.5-1.8 Cancer 420 1.0 0.9-1.1 Congenital/Hereditary 41 0.1 0.1-0.1 Total 41,606 100

To identify spatial-temporal clusters of occurrences of aetiologies, we conducted a case- based space-time permutation for the three most prevalent occurrences (Chlamydia-like signs, MV trauma, and wasting) for the study period 01/01/1997 to 31/12/2013. Koala cases derived from 345 local government areas grouped within other regions such as the Wide Bay Burnett, Central QLD, and Darling Downs South West, apart from SEQLD, the region from which most observations originated (Table 2-2). For Chlamydia-like signs, the number of locations recorded was 8,170 for 8,331 cases. Five significant Chlamydia-like signs clusters were identified. MV trauma was recorded for 4,779 locations totalling 4,998 cases and four significant clusters were identified. Wasting cases were recorded in 4,327 locations comprising of 4,371 cases and three significant clusters were identified. Interestingly, there was an overlap for case-based space-time clusters for occurrences of MV trauma, Chlamydia-like signs, and wasting for the 2010-2013 year period (blue clusters in Figure 2- 2) and for the year period 1997-2001 (green clusters in Figure 2-2) and 2010-2013. 39

Additional clusters were observed for Chlamydia-like signs for the year periods 1997-2001 and 2002-2005 (Figure 2-2).

Figure 2-1 Total counts and proportions of aetiologies Chlamydia-like signs, trauma caused by motor vehicles, and wasting occurring in koalas submitted to wildlife hospitals in SEQLD from 1997–2013 (No. of occurrences of aetiologies = 41,606).

2.4.3 Clinical syndromes

The identified 11 clinical syndromes were diagnosed as a single cause at submission or in combination, hence a total of 159 single or clinical syndromes combinations were made over the study period for the 20,250 records. Clinical syndromes involved single aetiologies in 52.5% (N=10,637), two in 15.8% (N=3,195), three in 14.4% (N=2,914) and 17.3% with four or more aetiologies (N=3,504). We focused subsequent analysis on the clinical syndromes with a prevalence >1%, which excluded 14.8% (N=3,007) koalas. MV trauma was the most frequent diagnosis (N=5,183, 30.1%), followed by Chlamydia-like signs, and clinical syndrome Chlamydia-like signs & wasting (Table 2-3). Lesser frequency was seen for trauma by animal attacks, trauma by other causes, wasting, and the clinical syndrome combination of Chlamydia-like signs & MV trauma, Chlamydia-like signs, senescence & wasting, and Chlamydia-like sings & trauma by animal attacks. 40

Table 2-2 Significant space-time clusters identified through case-bases space-time permutation. Koalas were submitted with trauma by motor vehicle, Chlamydia-like infection and wasting to wildlife hospitals in South East Queensland, from 1997 through 2013 (No. koalas with trauma by motor vehicle, Chlamydia-like infection and wasting and spatial information =15,524).

Interestingly, although Chlamydia-like signs were the most frequent occurring aetiology (Table 1-1), it was primarily presented in combination with other aetiologies. On the other hand, submissions due to MV trauma, the leading clinical syndrome, occurred in largely ‘healthy’ animals. The proportion of koalas diagnosed with MV trauma, Chlamydia-like signs, and trauma by animal attacks has declined over time, in contrast to the increasing proportion diagnosed with Chlamydia-like signs & wasting (Figure 2-3).

2.4.4 Koala outcomes

Most koalas were classified as ‘dead on arrival’ (utilising a total number of N=20250 records) N=9,690, 47.8%; CI 95%: 47.2 – 48.5), followed by ’euthanized’ (N=7,068, 34.9%; CI 95%:34.3 – 35.6) and ’released’ koalas (N=3,492, 17.2%; CI 95%: 16.7 – 17.8). Over time, the proportion of ‘released’ and ‘dead on arrival’ koalas has increased, in contrast to the decreasing trend of euthanized koalas (Figure 2-4a). There is also a seasonal pattern influencing koala outcomes (Figure 2-4), largely driven by more dead koalas submitted during August (N=1,249, 6.2%), September (N=1,306, 6.5%), and October (N=1,174, 5.8%), 41 overlapping with QLD’s koala breeding period. There was a slight variation in the proportion of ‘euthanized’ and ‘released’ koalas during this period. Associations of koala-specific (age class, sex) and temporal (year period, season) risk factors with the outcome (‘dead on arrival’, ‘euthanized’ and ‘released’) were explored for the four most frequent clinical syndromes using multinomial logistic regression models. The outcome ‘released’ was used as the reference category.

Figure 2-2 Locations of significant clusters of aetiologies as determined by the space- time permutation scans. (a–c) The area comprised within the rectangle includes the South East Queensland study area for the space-time permutation. The following map inserts display the significant clusters by the space-time permutation scans: (a) trauma by motor vehicle, (b) Chlamydia- like signs, (c) wasting. Each cluster is labelled with its p-value, number of observations, and time frame. Maps generated using ESRI ArcGIS Desktop v10.2.158. Centre points for each cluster generated using SaTScanTM v8.056. SaTScanTM is a trademark of Martin Kulldorff. The SaTScanTM software was developed under the joint auspices of (i) Martin Kulldorff, (ii) the National Cancer Institute, and (iii) Farzad Mostashari of the New York City Department of Health and Mental Hygiene.

Table 2-3 Frequency of the top 10 (>1%) clinical syndromes in koalas submitted to wildlife hospitals (with a CI 95%) in SEQLD from 1997 through 2013 (No. koalas with diagnosis at prevalence of >1% = 17,243).

Clinical syndrome Frequency Proportion CI 95% Trauma motor vehicle 5,183 30.1 29.3-30.8 Chlamydia-like signs 3,744 21.7 20.9-22.5 Chlamydia-like signs-Wasting 3,483 20.2 19.4-21.0 Trauma animal attack 1,561 9.1 8.3-9.8 Chlamydia-like signs-Trauma motor vehicle 811 4.7 3.9-5.5 Undetermined 728 4.2 3.5-5.0 Trauma other causes 686 4.0 3.2-4.7 Chlamydia-like signs-Senescence- Wasting 517 3.0 2.2-3.8 Chlamydia-like signs-Trauma animal attack 271 1.6 0.8-2.3 Wasting 259 1.5 0.7-2.3 Grand Total 17,243 100.0

2.4.5 Trauma by motor vehicles

MV trauma submissions were at decreased relative risk ratios (RRR) of being ‘dead on arrival’ and ‘euthanized’, compared to released koalas, if submitted between 2005 and 2013 compared to 1997-2001 period (Appendix 3), also highlighted by the higher predicted probability of ‘release’ in those years (Appendix 12). Adults submitted due to MV trauma were more likely than young koalas to be ‘euthanized’ (compared to ‘released’).

Figure 2-3 Proportion of top four koala diagnoses (Trauma by motor vehicle, Chlamydia- like signs, clinical syndrome combination of Chlamydia-like signs & wasting, and trauma by animal attacks) affecting koalas submitted to wildlife hospitals in SEQLD from 1997–2013 with a CI 95% (No. koalas with diagnosis = 20,250).

Figure 2-4 Outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals by year of submission between from 1997–2013 (No. koalas with clinical syndromes diagnosed = 20,250). The following figures display the (a) Proportion of outcomes by year of submission and (b) Counts of outcomes by month of submission.

2.4.6 Chlamydia-like signs

A similar temporal trend was observed for koalas submitted with Chlamydia-like signs (Appendix 4) and the clinical syndrome combination of Chlamydia-like signs & wasting (Appendix 5), although risk of being ‘dead on arrival’ and ‘euthanized’, compared to ‘released’ koalas, declined significantly only during the last four-year period (2009-2013).

Females with Chlamydia-like signs alone or in combination with wasting had higher RRR for being ‘dead on arrival’ or ‘euthanized’ compared to males, also highlighted by the higher predicted probability of ‘release’ for males (Appendix 13 and Appendix 14). Similar to MV trauma, adults submitted with Chlamydia-like signs were more likely than young to be ‘euthanized’ than ‘released’.

2.4.7 Trauma by animal attack

For submissions due to trauma caused by animal attacks, the RRR of being ’euthanized’ or ’dead on arrival’ was significant lower between 2005 and 2013 compared to period 1997- 2001 (Appendix 6) – this is similar to MV trauma. Adults compared to young koalas were found to be more likely to be ‘dead on arrival’ and ‘euthanized’ compared to ‘released’, and males were more likely to be ‘euthanized’ compared to ‘released’. Predictive probabilities for this model are shown in Appendix 15. Additional models for Chlamydia-like signs & trauma by motor vehicle (Appendix 7), trauma by other causes (Appendix 8), Chlamydia- like signs & senescence & wasting (Appendix 9), Chlamydia-like signs & animal attacks (Appendix 10) and Wasting (Appendix 11) are included.

2.5 Discussion

The current study is the most comprehensive analysis of koala morbidity and mortality to date, and the largest koala database ever explored. Further, this study presents a detailed analysis of factors impacting a specific wildlife population during a time of major population contraction, and serves as a model of other wild species facing extinction events. The koala, although geographically isolated from many other threatened vertebrate species in the world, faces similar perils derived from anthropogenically-induced changes to habitats to which wild species have little time to adapt. The weight of data derived from long-term studies is compelling and aids in the determination of the risks to the future conservation of wild species.

The hospital records analysed here represent a passive surveillance that offers unique opportunities to detect novel or rare pathogens, assist in the establishment of baseline levels of certain diseases, and provide insights into multifactorial causes of injury and disease of wild populations across time and space, and offer a rare opportunity to explore large-scale, multi-seasonal, long-term data that would be otherwise challenging to acquire. This dataset

45 permitted a unique retrospective view into the SEQLD koala populations over a notable period of rapid population decline. Previous passive surveillance studies utilizing hospital records have had considerably smaller sample sizes and had not modelled demographic and temporal risk factors in detail. It would be conceivable to assume wild koalas succumb to these aetiologies at equal rates and this was reflected in the observed hospital submissions.

Trauma was one of the main causes of admission to SEQLD hospitals, affecting a quarter of admitted koalas that were apparently free of comorbidities. The results of the current study are in agreement with previous studies from SEQLD (Weigler et al., 1987) and those conducted in other Australian states (Speight et al., 2014a, Obendorf, 1983, Griffith et al., 2013, Narayan and Williams, 2016, Canfield, 1987a). The high proportion of koalas presented to hospitals with traumatic injuries by MV and their high temporal variability may correlate with infrastructural changes across the koala range (Dique, Thompson, Preece, Penfold et al., 2003, de Oliveira et al., 2014), particularly during 2001-2005. Although the total number of submissions varied over time, there was a larger number of koalas excluded in 2005 from the data analysis, due to being submitted as healthy animals or with missing information. Of particular concern is that 37% of injured koalas had no underlying disease, which suggests potential for impact on local populations by healthy breeding koalas being prematurely removed. Virtually all koala trauma cases were derived from anthropogenic sources (Griffith et al., 2013, Obendorf, 1983), directly (MV) or indirectly (dogs, livestock, drowning), except intraspecific aggression. Habitat encroachment due to land clearing and urbanisation, shows a strong association with road vehicle and dog attack trauma (Dique et al., 2003, Ellis et al., 2016, Mifsud, no year). However, residents in these areas may report injured or sick charismatic fauna to hospitals, thus increasing location and selection bias.

Seasonality has been reported (Narayan and Williams, 2016) to be a contributing factor to mortality, directly influencing koala behaviour, particularly, dispersal patterns during breeding periods (Gifford et al., 2002), from August to January in QLD. Although we did not find an association between trauma and season, breeding periods enhance opportunities for sexually-active males to encounter different sources of hazard in search for mates, resources or inbreeding avoidance (Weigler et al., 1987). Other reports have found young koalas to be more likely to be affected by MV trauma (Griffith et al., 2013,), potentially triggered by the forced dispersal following weaning (Griffith et al., 2013) or that the habitat in those sites is fragmented but stable, causing a more natural population dispersal. Dispersal across age classes can be triggered by different drivers across sites (Dique, 46

Thompson, Preece, de Villiers et al., 2003); thus, it becomes difficult to pinpoint reasons behind trauma and age class. However, we could speculate that ongoing clearing for development in QLD triggers a uniform age dispersal, enhancing opportunities for all age classes but as there are more adults present, more will encounter roads. In general, there was a higher proportion of adults compared to young koalas in the database, possibly explained by adults comprising a higher population proportion, adults dispersing more readily, are easier to spot, or are removed at a slower rate from the environment by scavengers (Dique, Thompson, Preece, de Villiers et al., 2003).

Although previous reports from SEQLD (Weigler et al., 1987) have found no relationship between injuries caused by animals and sex, the demographics of trauma by animal attacks can reflect koala biological traits. Mortality surveys from other states (Canfield, 1987a, Canfield, 1987b) have determined that the majority of attacks by domestic carnivores occur in koalas with poor body condition, suggesting an increased vulnerability to trauma when debilitated as they roam more frequently on the ground. This finding contrasts with the present study, in which the top two traumatic injuries (MV and dogs) occurred in apparently healthy individuals. Although detailed post-mortem and histological examinations that would exclude any underlying disease were not performed in every dead individual, an ongoing prospective mortality survey conducted by our research team indicates most koalas affected by trauma are indeed healthy (unpublished data). Overall, trauma by animal attacks displays a slight but steady decrease across the years, which may reflect the efforts from the QLD State sponsored Koala Plan (Protection, 2016) to identify threats, incorporate careful planning of koala sensitive areas in development, and educate regarding responsible pet ownership (Melzer et al., 2000a). In general, the ‘released’ koala proportion for all clinical syndromes has increased, suggesting increased success in rehabilitation efforts. Additional threat mitigation efforts currently under consideration or being carried out to aid include differential speed signs (Dique, Thompson, Preece, Penfold et al., 2003), installation of Intelligent Transportation System devices for driver awareness (Birdlife Australia, 2016, Protection, 2013), dog registrations (Obendorf, 1983), to cite a few. Additional traffic-calming measures such as daylight savings (Ellis et al., 2016) can be adopted in QLD to decrease road-associated injury and mortality, following careful consideration of its impact on diurnal species. These measures could aid in the alleviation of current road and dog associated mortality for wild crepuscular or nocturnal species in Australia and elsewhere.

Risk factor analysis for trauma by other causes (e.g. drowning, intraspecific aggression, tree falls - data not shown here) highlight that younger koalas were at higher risk for being 47 submitted ‘dead on arrival’. One plausible explanation is that young and inexperienced koalas may suffer lethal trauma following falls whilst climbing or during fights (Narayan and Williams, 2016). Adults are perhaps more likely to survive misadventure, arriving to hospitals with terminal injuries, resulting in euthanasia.

The relationship of koala population decline with disease has been less clear. Pathogens relying on frequency-dependent transmission such as Chlamydia spp. can influence population dynamics (McCallum, 2012), affecting koalas by increasing mortality from wasting and blindness and decreasing population recruitment through impairment of reproduction (Jackson et al., 1999). However, common diseases such as chlamydiosis and potentially KoRV-driven immunodepression seem to only play an important role at the population when other extrinsic factors inducing physiologic stress (Munday, 2012), stemming from urbanisation such as habitat fragmentation (Munday, 2012, Gordon et al., 2006, Dique, Preece, de Villiers, 2003) are present. This is particularly interesting with KoRV, as there is a 100% prevalence of this endogenous virus in QLD koalas (Simmons et al., 2012); however, according to this study, the leading submission causes do not correspond to immunosuppressive disease, lymphoma, or leukaemia which are thought to be linked to KoRV-B infection (Denner and Young, 2013).

Chlamydia-like signs made up the second and third (in combination with wasting) largest groups of submissions, which agrees with past reports (Griffith et al., 2013). Chlamydia can affect multiple body systems, leading to chronic illness and euthanasia on welfare grounds. In koalas, chlamydial infection is a disease complex represented by four syndromes (rhinitis- pneumonia, ocular, reproductive, urinary tract infection). Immunohistochemistry for Chlamydia has demonstrated positive labelling for other sites such as the rectum, cloaca, spleen and lung, demonstrating a multisystemic spread of the infectious organism (Higgins et al., 2005b). The two species known to date to infect koalas are C. pneumoniae and C. pecorum (Burach et al., 2014). C. pecorum has been isolated from ocular, urogenital and rectal sites, and is commonly associated with overt clinical disease and thus thought to be more virulent (Jackson et al., 1999). Whereas the usual site for infection for C. pneumoniae is the ocular system; with infection at other sites infrequent (Jackson et al., 1999, Burach et al., 2014). Briefly, cystitis, or ‘wet bottom’, a primary cause of submission in other states (Griffith et al., 2013), leads to other ascending urogenital disease manifestations (Burach et al., 2014, Higgins et al., 2005b) further compromising multiple organ systems (Burach et al., 2014). Ovarian bursitis and orchitis cause infertility (Obendorf, 1981, Weigler et al., 1987, Burach et al., 2014, Brown et al., 1987, Girjes et al., 1988, Stalder et al., 2003) and although 48 they may not cause direct mortality, bursitis is a reason for euthanasia due to infertility according to QLD governmental guidelines. Chlamydial urogenital disease continues to be the most prevalent chlamydiosis manifestation (Canfield, 1987a, Canfield, 1987b, Obendorf, 1983, Backhouse and Bolliger, 1961a, McKenzie, 1981), likely contributing to a gradual deterioration, in turn increasing detection in comparison to other diseases or chlamydial manifestations, leading to overrepresentation. Conjunctivitis leads to blindness and thus increased predation and starvation due to affecting foraging capacity. Other signs such as the rhinitis-pneumonia complex have been reported previously in higher numbers(Backhouse and Bolliger, 1961a); we report a declining proportion for nephritis and pneumonias. Lower proportion of pneumonias correlates with observations reported by MKH, which vaccinated incoming vulnerable (e.g. orphaned) koalas in previous years (pers. comm. P. Theilemann).

Studies in free-ranging koalas in QLD have reported variability in prevalence of chlamydiosis in the wild (Jackson et al., 1999), and the existence of healthy carriers, suggesting that infection prevalence does not equal manifestation of disease. Subclinical infections may explain the underrepresentation of chlamydiosis in our study in comparison to trauma- associated cases, particularly as road-killed koalas will be invariably more visible to the public than koalas affected by chlamydiosis or other chronic, debilitating disease associated with forested settings. Another source of underrepresentation of chlamydiosis may be due to healthy carriers particularly if infected with less virulent species (e.g. C. pneumoniae (Jackson et al., 1999)). The demographics of chlamydiosis diagnoses will likely shift in the future in response to the introduction of more accurate, cost-effective diagnostic tools (Obendorf, 1981).

For some causes, particularly chlamydiosis, the female sex was a risk factor associated with poorer clinical outcomes, highlighting the impact of the disease on the female koala population in hospitals. Most studies have focused on female reproductive pathology from the standpoint of burden of disease and infertility sequelae (Cunningham and Beagley, 2008). Female chlamydiosis is diagnosed more readily due to overt disease expression, leading to poor clinical outcomes as infertile individuals are euthanased. Further, the increased incidence in females maybe a confounder derived from the active surveillance performed with ultrasound on females, masking and subsequently underdiagnosing the less-evident male reproductive disease. Regarding the increased number of female euthanasia in some year groups, such as 2010-2013, no particular factor other than number of admissions of diseased female koalas was found. 49

Acute or chronic chlamydiosis results in poor prognosis for SEQLD koalas, particularly if combined with wasting. Progressive wasting, or emaciation in a wild koala can be due to chronic debilitation due to disease (Obendorf, 1983) or dental attrition caused by advanced age. Adults are more likely to be wasted due to cumulative risk of exposure to pathogens, sources of injury, dental attrition from advanced age and environmental stressors, compared to young koalas. Adults were also more likely to be euthanized in cases of disease/injury, plausibly because there may be more willingness to treat younger individuals. Dental attrition jeopardizes the koala’s masticatory effectiveness, leading to progressive body condition loss as for hindgut fermenters, food particle size and cell wall breakdown is necessary to maximize microbial fermentation (Logan and Sanson, 2002). However, most submitted koalas were not senescent, indicating the increased proportion of wasted koalas in QLD over time is due to undiagnosed chronic diseases. Further research, with detailed post mortem examinations will help elucidate the causes of the increased proportion of wasting, and add to the weight of data in regards to the associations between the different aetiologies and outcomes found in this study. Similar efforts conducted in other states with comparable population structure, pressures, climate and geography (NSW) (McAlpine et al., 2012) paired with continuing studies addressing aspects of dynamics of the SEQLD population are needed to project conclusions concerning the future of contracting koala populations. Findings from previous post-mortem surveys report virtually the same frequency and type of aetiologies (Weigler et al., 1987, Obendorf, 1983, Speight et al., 2014a, Canfield, 1987a), with minor variations concerning certain diseases, which indicates that the pressures that koalas face across their range have remained the same across space and time, are unlikely to change in the near future, and are likely capable of causing local extinctions (Biolink, 2005).

Besides chlamydiosis, there were also highly variable temporal patterns in other less frequent disease manifestations, highlighting the overall variability of different aetiologies across time. The ‘other diseases’ category included dermatitis, septicaemia, ectoparasitosis, hepatitis, amongst others; however, their significance in SEQLD koala subpopulations could not be discerned in the present study.

In the exploratory spatial analysis overlapping clusters between MV trauma, Chlamydia-like signs, and wasting were identified for two year periods (2010-2013 and 1997-2001). The fact that wasting and Chlamydia-like signs cases were observed in the same spatial- temporal regions could unveil clusters in SEQLD related to undiagnosed infectious disease. The intense, long-term habitat clearing in QLD (Gordon et al., 2006, Dique, Thompson, 50

Preece, Penfold et al., 2003, de Oliveira et al., 2014) could be leading to starvation in koalas, an issue that has surprisingly not generated much discussion. Most of these clusters also are located within densely populated councils. Thus, this could reflect 1) a direct observant effect from increased numbers of reporting residents; or 2) a true disease (stressed koalas expressing overt disease) and injury (higher traffic or road density) cluster derived from increased risk. Future spatial-temporal modelling would be required to identify environmental and climatic risk factors for morbidity and mortality.

Medical records or passive surveillance data for research is an increasingly attractive area, particularly for wildlife investigations, offering a rare and cost-effective opportunity to explore large-scale, multi-seasonal, long-term data that would be otherwise challenging to acquire. Unlike with active surveillance, some bias is expected with passive as it does not comprise specific standardised data collection to assess populations for evidence of disease. For example, people may be more likely to report a koala than the more commonly-urbanized Australian possums (taxonomic bias) or adults and males would be easier to spot than young koalas (demographic bias). Nonetheless, this data set is valuable due to the large number of records over time and space. Some dead koalas submitted to hospitals receive no or cursory necropsies due to financial constraints, limited staff time or expertise, potentially leading to mis- or underdiagnosis of true causes of mortality and morbidity, and failure to identify comorbidities. Thus data extracted from hospitals or rehabilitation centres are inherently flawed (e.g. misdiagnosis due to lack of validated testing in certain pathological conditions; varying public engagement attitudes towards wildlife in a certain areas; lack of consistency in post-mortem assessments, submission of carcasses with limited diagnostic value due to a suboptimal preservation, etc.). However, these institutions remain an important source of passive surveillance data that would be otherwise challenging to actively collate.

Our analysis permitted a unique retrospective view into the SEQLD koala populations over a notable period of rapid population decline, offering the possibility to infer prospective dynamics and risk factors. Results may serve to determine risk factors in other wild species impacted in this geographic area and more broadly by urbanisation threats, allowing for a holistic wildlife management and conservation policy to be developed in threatened species. This is the first comprehensive study to analyse retrospectively the complex nature of the koala decline in the SEQLD, spanning nearly two decades. Future directions needed to curb the ongoing decline of koalas in QLD and adjacent states and territories include: reduction of euthanasia rates, further research to identify environmental risk factors of morbidity and 51 mortality, and development of science-based mitigation strategies including instating restricted clearing for development, building road capacity, encouragement of responsible pet ownership, standardised national surveys and detailed mortality studies.

Chapter 3 Pathology and Trends in Morbidity and Mortality in the Koala (Phascolarctos cinereus) in South-East Queensland, Australia.

3.1 Introduction

The koala is a medium-sized, arboreal folivorous marsupial, with a broad but patchy distribution tied to woodlands and forests of Eucalyptus spp., its primary food source (Martin and Handasyde, 1999). The koala is nationally listed as vulnerable in response to population declines in the states of Queensland and New South Wales (NSW), and the Australian Capital Territory (Department of Environment, 2014, Department of Sustainability, 2013). Despite being overly abundant in the south extension of their distribution, koalas in South- East Queensland (SEQLD) are threatened by the population limiting effects of disease (Waugh et al., 2017, Rhodes et al., 2011) and habitat clearing for urbanization (McAlpine et al., 2006) exposing koalas to trauma from vehicle collisions (Dique, Thompson, Preece, Penfold et al., 2003) and animal attacks (McAlpine et al., 2004). Despite investments in medical care, and species management, there has been a rapid population decline from 1996 through to 2014, with declines of 80% in the Koala Coast and 54% Pine Rivers sites are recorded in surveyed populations (Rhodes et al., 2015).

A recent retrospective epidemiological study of koala mortality spanning 17 years determined that the reason for koala hospital submissions was multifactorial. Vehicle collisions, debilitation and infertility resulting from chlamydiosis were the major causes of mortality and morbidity (Gonzalez-Astudillo et al., 2017). However, this study was based on retrospective medical records and autopsies were not conducted in all animals, impeding the interrogation of comorbidity and disease interaction. The purpose of this prospective pathology study was to determine the causes of mortality and comorbidities driving koala submissions to hospitals in SEQLD, and to characterize their gross and histological lesions.

3.2 Materials and Methods 3.2.1 Survey Population

Koalas in SEQLD are presented to 3 major wildlife hospitals (Moggill Koala Hospital, Currumbin Wildlife Sanctuary Hospital and Australia Zoo Wildlife Hospital). Koala admissions derive from councils in SEQLD, and also from cities in NSW in close proximity to the QLD border. Koalas were either found dead, died during hospitalization, or were euthanized. All samples were collected under a Scientific Purposes Permit from the Department of Environment and Heritage Protection, Queensland Government (WISP 53

13247813, generated 07/08/2013), and under ethical review of The University of Queensland Animal Welfare Unit (ANRFA/SVS/193/13/EHP)

3.2.2 Autopsies and Sample Collection

Autopsies were conducted at the School of Veterinary Science, The University of Queensland, Gatton campus, between April 2013 and July 2016. The majority of carcasses were frozen at -20oC due to logistical reasons, and thawed between 0-4ºC for 4-7 days prior to examination. Koalas were identified by ear tags, microchip numbers, or by the hospital accession numbers used for clinical histories. Aging was done according to premolar tooth wear (Gordon, 1991) designating koalas to an age class (young/subadult, mature, and aged) (Ellis et al., 2014, Gordon, 1991). Body condition (BC) score was recorded from 1 through 10 by palpating the infraspinatus and supraspinatus muscles and ranged from emaciated 1- 2 , poor 3-4, fair 5-6, good 7-8 and excellent >8 BC (Blanshard and Bodley, 2008, Jackson et al., 2003), with score then correlated to the presence of chronic illness or advanced age.

Histopathology samples were fixed in 10% NBF, sectioned at 4µm and stained with hematoxylin and eosin by routine methods. Special stains performed included: Periodic Acid Schiff’s (PAS), for mucins, mucopolysaccarides and fungal agents, Gram’s for infectious agents, Masson’s Trichrome for collagen and smooth muscle, and Ziehl-Neelsen for acid- fast organisms. Ancillary testing included bacterial and fungal culture by routine methods at Veterinary Laboratory Services, The University of Queensland (Gatton, Australia). PCR (panfungal, bacterial) assays were conducted at Pathology West (Sydney, Australia) for the molecular identification of bacteria and fungi. Panfungal PCR targeted fragments of the rDNA gene cluster, specifically, those internally transcribed in spacer region located between the 18S and 28S (Lau et al., 2007). The bacterial PCR assay targeted conserved regions U1 and U3 for the 16S rRNA gene (Relman, 1993) using broad range primers. Variable portions within the 16S rRNA gene provide unique signatures of any bacterium. Identification was conducted via sequence analysis of the amplified product.

3.2.3 Data Analysis

Lesions in koalas were categorized by a standard system that separated body part, organ system, type of pathogen or etiology (if determined), pathological process, and final diagnosis(es) (Gonzalez-Astudillo et al., 2017). Analysis of relationships between koala records, final diagnoses, and variables of interest (e.g. demographics) were performed using contingency tables. We hypothesized that low BC increased the likelihood of koalas in being 54 injured by animal attacks (livestock, dogs). To test this relationship, a two-tailed Fisher’s Exact test was used to compare the status of loss of BC (low BC defined as BCS ≤4 versus normal BC) in koalas diagnosed with trauma from motor vehicle collisions and in koalas with trauma caused by animal attacks. WIN PEPI v. 11.44 (Brixton Health, 2015) was used to conduct the statistical analysis.

3.2.4 Clinical Disease Grading Criteria

Chlamydial conjunctivitis and cystitis was subjectively scored by the submitting hospitals based on the degree of tissue damage and disease progression. The conjunctivitis scoring ranged from 0-3 where 1 is mild (acute membrana nictans inflammation lacking proliferative change with epiphora); 2 moderate (mild proliferation of conjunctiva and membrana nictans with serous exudate, with or without keratitis or periorbital alopecia); and 3 severe (chronic inflammation of conjunctiva and membrana nictans with mucopurulent discharge causing adhesion of eyelids, possible conjunctival bleeding, active keratitis leading to partial or total corneal opacity, focally extensive periorbital alopecia; Pers. Comm. A. McKinnon and (Loader, 2010)). The cystitis scoring ranged from 0-3 where 1 is mild (common vestibule is slightly or moderately soiled); 2 moderate (soiled common vestibule, incontinence evident from dribbling micturition, urine odor present, tacky fur); and 3 severe (soiled common vestibule with matted fur, fur loss, excoriation, incontinence, hemorrhagic dysuria, associated with low BC). Extreme cases can occur with cloacal prolapse, and nephritis (Pers. Comm. A. McKinnon). Chlamydial bursitis was not clinically scored as it causes permanent infertility regardless of the degree (Obendorf, 1981), with affected females being culled.

3.3 Results 3.3.1 Summary of Causes of Admission

The predominant etiology diagnosed corresponded to infectious diseases (63%, 329 of 519, Table 3-1); over half (58.5%) of koalas with infections (304 of 329) had hallmark lesions indicating Chlamydia spp. infection

Comorbidities including trauma were diagnosed in 57.6% (299 of 519) of koalas, with chlamydiosis the most frequent comorbidity (250 of 299). Primary comorbidity combinations observed included multicentric infections (96 of 299) leading to loss of BC (50 of 299). The vast majority (93 of 96) of multicentric infections being from chlamydiosis affecting the urogenital tract or urogenital and ocular (27 of 96) systems. The second most common 55 combination of comorbidities were found in koalas admitted with injuries frequently caused by motor vehicle collisions but also suffering from infections (55 of 299) caused by urogenital chlamydiosis and a few other agents. Surprisingly over half of these animals with trauma and multicentric infection (29 of 55) were still in good BC. The third most frequent comorbidity combination was koalas diagnosed with infections and senescence (45 of 299). Infections in this group were commonly caused by urogenital and ocular chlamydiosis. The last group included the combination of infections and other diseases (23 of 299). In this group, most infections included urogenital chlamydiosis (18 of 23) over half occurring in debilitated animals (14 of 23). Other diseases included congenital diseases (renal hypoplasia), and non-infectious diseases (e.g. degenerative changes in myocardium). Female koalas were more often diagnosed with comorbidities (N=154, 60.39%) compared to male koalas (N=101, 39.60%). Similar comorbidity proportions were found across all age classes, although mature koalas (76 of 204) were diagnosed with slightly higher numbers compared to young/subadults (61 of 204) and aged (67 of 204) koalas.

Table 3-1. Distribution of aetiologies diagnosed in each age class in koalas admitted to South-East Queensland hospitals and receiving autopsies at The University of Queensland from 2013 through to 2016. Koalas may be counted more than once due to comorbidities.

Age class n total (n total %) Diagnostic Young/Subadult Mature Aged Unknown Subtotal Category Infectious 77 (23) 160 (49) 86 (26) 6 (2) 329 (44) Trauma 79 (42) 76 (40) 30 (16) 4 (2) 189 (25) Dental attrition 0 (0) 0 (0) 84 (100) 0 (0) 84 (11) Other diseases 19 (28) 23 (33) 25 (36) 2 (3) 69 (9) Neoplasia 8 (20) 22 (54) 9 (22) 2 (5) 41 (5) Idiopathic wasting 18 (62) 14 (48) 0 (0) 0 (0) 32 (4) Undetermined 2 (25) 6 (75) 0 (0) 0 (0) 8 (1) Total 202 (27) 299 (40) 234 (31) 14 (2) 751 (100)

Male-to-female ratio was 1.056:1 (266 males, 51%; 250 females, 48%; and 1 hermaphrodite, <1%) with sex not recorded in 2 instances. Mature females were most often diagnosed with infectious diseases, whilst male koalas presented more frequently with trauma in the young/subadult and mature age classes, and from idiopathic low BC in the young/subadult age class (Figure 3-1).

Figure 3-1. Demographic distribution of the frequency of etiologies for South-East Queensland koalas autopsied at The University of Queensland. Only koalas with sex (N=743) and age class information (N=750) are displayed. In this figure, koala data is counted more than once to account for co-morbidities. Not included in the graph N=3 where sex and N=15 age were not recorded, and N=1 hermaphrodite.

Koala hospital records were recovered for 361 koalas (data not shown). Most records (351 of 361) included incident dates, revealing a marked monthly variation in admissions throughout the study period, with most koalas admitted in late winter (June-August) through summer (December-February). The koala fate at hospital was variable. The majority of koalas were euthanized (63%, 325 of 519), compared to 14% being dead on arrival or dying during hospitalization (75 of 519). Information was not recovered for fate in 23% of koalas (119 of 519). In these instances, the koala’s fate was deduced based on the severity of lesions observed for natural death (i.e. massive skull fracture), or by observing evidence of barbiturate crystals within tissues for euthanasia. When considering animals which were dead on presentation (14%, 75 of 519), trauma was the cause of death for 25% of koalas (19 of 75), infectious diseases were the cause of death in 29% of cases (22 of 75), 95% of these due to putative chlamydiosis.

In 11% of cases (8 of 75), other disease categories were the cause of death, including bilateral renal fibrosis, multifocal necrotizing or lymphoplasmacytic hepatitis (2 cases), interstitial pulmonary fibrosis, unidentified crystalluria and snake envenomation (1 case

57 each). Additional koala gross and histology images can be found in Appendix 16 through to Appendix 24.

There was noticeable increase in the number of admission during July-September and October-December. However, to facilitate comparison with the retrospective epidemiological analysis, the top four causes of admission (denominated ‘clinical syndromes’ in the previous report) were grouped by 3-month periods and analysed temporally (Figure 3-2). Specifically, for the top four final diagnoses which included Chlamydia-like signs/Wasting, Trauma by motor vehicle, Chlamydia-like signs, and Chlamydia-like signs/Senescence, periods of high admissions in the autopsy survey extended from late winter to early summer. When only considering the month of admission to hospital, the highest proportions for these four final top diagnoses occurred in August (15%) through to September (11%), and November (18.4%). Trauma by motor vehicles displayed an elevated number of admissions from June through November. November was the month with the highest number of admissions diagnosed with the combination Chlamydia-like signs/Senescence. The final diagnoses of Chlamydia-like signs/Wasting and Trauma by motor vehicles both displayed a biphasic trend. Peaks for Chlamydia-like signs/Wasting occurred in summer through early spring (January-March) and spring and early summer (October-December), this same period displayed a peak for Trauma by motor vehicle as well as late winter and spring (July- September).

18 16 14 12 10 8 6 4 2 0

Seasons No. ofkoalas admittedinto hospitals

Chlamydia-like/Wasting Trauma motor vehicle Chlamydia-like Chlamydia-like/Senescence

Figure 3-2. Seasonal distribution of the frequency of top 4 final diagnoses (Chlamydia- like/Wasting N=98, Trauma motor vehicle N=51, Chlamydia-like N=36, Chlamydia- like/Senescence N=33) diagnosed during autopsy examinations in koalas admitted to wildlife hospitals in South-East Queensland from 2013 through to 2016. This figure does not include the N=58 koalas for which date of incident was not recovered.

3.3.2 Urinary tract Disease

The most frequent form of putative chlamydiosis was urogenital infection, diagnosed in 51% (267 of 519) of the total number of koalas. Chlamydia frequently manifested as cystitis (Figure 3-3, Figure 3-4) affecting 34% (176 of 519) of koalas. Cystitis grossly caused thickening and firmness of the urinary bladder wall, with mucosal infolding and occasional petechiae. Soiling around the common vestibule (Canfield et al., 1989), colloquially referred to as ‘wet bottom’ (Weigler et al., 1987) was a common observation in chronic cases. Clinical chlamydial scoring was recovered for 43 cases of cystitis, most corresponding to severe (60%, 26 of 43) or moderate (19%, 8 of 43). The predominant histological characteristic was a moderate-to-severe fibrosis accompanied by detrusor muscle hypertrophy, or chronic, multifocal, transmural lymphoplasmacytic infiltrate, and occasional urothelial ulceration. Chronic cystitis resulted in obstruction of the bladder outlet due to chronic damage leading to excessive detrusor muscle activity, evidenced by increased collagen deposition and myohypertrophy in the bladder wall (Canfield et al., 1983, Tubaro et al., 2010), (Stalder et al., 2003, Weigler et al., 1987).

Figure 3-3. Cystitis, urinary bladder. Hyperplasia causing mucosal infolding.

Figure 3-4. Cystitis, urinary bladder. Discolouration of fur surrounding the common vestibule.

Renal complications associated with chlamydial infection were observed in 20% of all koalas (106 of 519) and included obstructive uropathy such as hydronephrosis (9 of 106) (Figure 3-5) and renal infarcts (2%, 13 of 106) (Figure 3-6). The predominant histological lesion was lymphocytic interstitial nephritis or chronic lymphocytic segmental pyelonephritis. Severe renal disease was infrequent, and included ascending urinary tract infection - UTI (4 cases), end stage kidney disease (4 cases), and pyelonephritis (10 cases). A total of N=55 mature male koalas were shared with another research group within the School of Veterinary Science. From these, N=2 had urogenital gross and histopathological lesions compatible with chlamydiosis. One koala submitted with mild chronic cystitis displayed a severe chronic lymphoplasmacytic interstitial prostatitis and urethritis. The other koala was submitted with a severe hemorrhagic cystitis and dislayed a mild lymphoplasmacytic urethritis. Both koalas were positive to C. pecorum via real-time PCR (Palmieri et al., 2018).

Figure 3-5. Hydronephrosis and hydroureter, kidney (asterisk), and ureter (arrowhead).

Figure 3-6. Organized infarct, kidney.

Additional renal complications such as crystal precipitations were observed in 4% of koalas (21 of 519), concomitant with chlamydiosis in other tissues. Crystal precipitations were mostly composed of struvite (48%, 10 of 21) and calcium oxalate (24%, 5 of 21).

3.3.3 Reproductive Disease

Ovarian bursitis (bursitis) occurred in 23% (122 of 519) of the total number of koalas, and in 49% of females admitted for autopsy. Bursitis was characterized by chronic and moderate- to-severe cystic dilations of the ovarian bursa (Figure 3-7). Exudate within the bursa was often clear and straw-colored, but varied from purulent to hemorrhagic. The bursa was bilaterally compromised in most cases (56%, 68 of 122). The primary histological feature was the replacement of ovarian and fallopian structures with mature collagen, occasionally with lymphocytic infiltration. Involvement in other sites in the reproductive tract co-occurred with bursitis in 22% of cases (27 of 122), including metritis or endometritis, fibrotic metritis, salpingitis, suppurative vaginitis, or pyometra. In the male, 9 of 14 cases of urogenital infections were consistent with chlamydiosis, with cystitis and urethritis being the most 61 common presentations. Grossly when infected, the prostate (10 of 14) appeared normal or enlarged and firm, with histological evidence of chronic multifocal prostatitis and cystitis with occasional prostatic intraglandular pustules.

Figure 3-7. Bursitis, ovarian bursa, and bladder (asterisk). Bilateral cystic dilation with severe fibrosis, symmetrical distortion, thickening and dilation of the ovarian bursa with hemorrhagic cystitis.

Multi-systemic infection in suspected chlamydiosis was common, particularly involving both the urinary and reproductive tract (Figure 3-8). Mature female koalas outnumbered males in cases compatible with reproductive chlamydiosis (Figure 3-9). Additionally, female koalas presented co-occurring extra-urogenital chlamydiosis more frequently than male koalas (20% vs 14% of the time, respectively). The majority of co-occurring chlamydiosis in the male corresponded to infections in the urogenital tract and conjunctivitis (2 cases).

Figure 3-8. Frequency of infection sites with gross and histopathological lesions compatible with chlamydiosis observed during autopsies conducted in koalas admitted to South-East Queensland hospitals from 2013-2016 (N=304).

Figure 3-9. Demographic distribution of the frequency of main body part or system affected by putative chlamydiosis as diagnosed in South-East Queensland koalas autopsied from 2013-2016. Koalas in this figure are counted more than once as chlamydiosis affecting several tissues was a frequent finding. Not included in the figure are koalas for which age class (N=9) and sex (N=3) was not recorded or non-dichotomous (hermaphrodite).

3.3.4 Ocular Disease

Ocular chlamydiosis affected 17% (88 of 519) of koalas (Figure 3-10). Chlamydial clinical scoring was recovered for half the cases (45 of 88) and frequently was high-grade (50%, score ≥2). Ocular chlamydiosis was often diagnosed grossly, and described as a chronic, bilateral (78%) mucopurulent conjunctivitis or keratoconjunctivitis, with periocular fur matting or alopecia and adhesion of the palpebrae. Lesions observed with minor frequency were corneal ulcers, blindness and cataracts. In koalas with only ocular chlamydiosis there was concurrent low BC in only 9% (8 of 88). However, in 67% (59 of 88) of koalas with low BC, ocular chlamydiosis presented in combination with multi-systemic conditions.

Figure 3-10. Keratoconjunctivitis, eye and periorbita (clinical scoring = 3).

3.3.5 Respiratory Disease

Four percent (21 of 519) of koalas were found to have pulmonary disease including suppurative pneumonias or other pneumonias, and bronchopneumonias. Focal or interstitial fibrosis was found in 33% (7 of 21) of cases with respiratory lesions. Cases with rare or novel pathogens included a focal pleuritis by Psynchrobacter spp., a pulmonary abscess by Mycobacterium abscessus, and a pulmonary phaeohyphomycosis by Cladosporium spp., all PCR-confirmed.

3.3.6 Neoplastic Disease

Eight percent (41 of 519) of koalas had neoplasms, occurring mostly in mature or aged koalas (31 of 41), without an obvious sex distribution. Round cell tumors affected 58% (24 of 41) of koalas with neoplasms, including lymphomas (20 cases) and leukemias. Mesenchymal neoplasms affected 39% (16 of 41), including cranial (Canfield et al., 1987c) and costal osteochondromas (7 cases) and primary serosal myxosarcoma (Gonzalez Astudillo et al., 2015) (2 cases). The least common neoplasms (15%) were epithelial, including colonic and hepatic adenocarcinomas, squamous cell carcinoma, and a bronchioloalveolar carcinoma. Despite the overall low occurrence of neoplastic disease, neoplasms often co-occurred with chlamydiosis (16 of 41), but infrequently with trauma (2 of 41).

3.3.7 Trauma

Trauma was diagnosed in 36% (189 of 519) of cases. From these, 58% (109 of 189) were from motor vehicles collisions, 21% (40 of 189) from animal attacks (dogs, livestock), 7% (13 of 189) falls, and 14% (27 of 189) from other causes. Males (110 of 189, Figure 3-1) in the young/subadult (44%) and mature (38%) age classes (Figure 3-11) were predominantly affected by trauma compared to females. Acute bone fractures affected nearly half (46%, 87 of 189), mostly located in the head (55%) or long bones (35%; data not shown). Severe peracute hemoabdomen and bilateral hemothorax from hepatic and pulmonary lacerations were frequent. Healed bone fractures were infrequent but two cases of proliferative osteophytic periosteal masses compatible with fracture calluses were observed. No signs of underlying pathological conditions were found in 50% (94 of 189) of koalas injured (Table 3-2), most (94.7%) of which had good BC. However, in those koalas presenting with a comorbidity, chlamydiosis was the most common condition, affecting 19% (63 of 189). A total of 15 females affected by trauma, most injured by motor vehicles (10 of 17), had evidence of carrying pouch or back young, and 2 had evidence of carrying offspring (elongated nipples or mammary adenomegaly).

Figure 3-11. Proportion of South-East Queensland koalas affected by lethal trauma in relation to their age class, autopsied at The University of Queensland from 2013-2016 (Proportion based on N=189 observations).

Table 3-2. Distribution of co-morbidities and their association with traumatic injuries in koalas presented for autopsies at The University of Queensland from 2013-2016. Trauma by other causes includes categories such as intraspecific trauma, heat stress, and undetermined causes of trauma.

Type of trauma n total (n total %) Motor vehicles Healthy 62 (32.8) Diseased (senescence, 43 (22.2) illness) Idiopathic wasting 4 (2.1) Subtotal 109 (57.1) Animal attacks Healthy 17 (9.0) Diseased (senescence, 19 (10.1) illness) Idiopathic wasting 4 (2.1) Subtotal 40 (21.2) Other causes Healthy 12 (6.3) Diseased (senescence, 13 (7.4) illness) Idiopathic wasting 2 (1.1) Subtotal 27 (14.8) Tree falls Healthy 3 (1.6) Diseased (senescence, 6 (3.2) illness) Idiopathic wasting 4 (2.1) Subtotal 13 (6.9) Total 189 (100.0)

3.3.8 Loss of Body Condition

Low BC was typically disease-related (61%, 180 of 294), due to chlamydiosis (142 of 180). Dental attrition from senescence accounted for 16% cases (84 of 519) (Table 3-3). Low BC was commonly observed in males in the young/subadult (Figure 3-1) as well as in the mature (Table 3-3) age class. There was no observable evidence of an underlying pathological condition that could account for low BC found in 6% (32 of 519) of koalas, most of which were males (87%).

To explore the relationship of disease-induced low BC as a predisposing factor to trauma by animal attacks, the number of koalas with low BC suffering from significant trauma from

66 both vehicles (N=108) and domestic animals (N=40) was analyzed. A total of 38% of koalas (15 of 40) injured by domestic animals such as dogs or livestock had a chronic disease or were aged causing low BC compared to 16% (17 of 108) with low BC from the same causes involved in motor vehicle collisions. Low BC was found to significantly influence the likelihood of koalas being attacked by animals (p = 0.007) compared to koalas with normal BCs injured by motor vehicles and animals.

Table 3-3. Body Condition (BC) score of koalas admitted to wildlife hospitals from 2013 through to 2016 in South-East Queensland receiving autopsies at The University of Queensland grouped by age class (young/subadult, mature, aged; N=506).

Young/Subadult Mature Aged N total (N total %) N total (N total %) N total (N total %) BC BC BC BC BC BC score Score Score Score Score Score LOW HIGH LOW HIGH LOW HIGH Total Female 28 (18) 39 (25) 52 (22) 55 (23) 46 (39) 27 (23) 247 (49) Male 45 (29) 42 (27) 75 (32) 52 (22) 32 (27) 13 (11) 259 (51) 127 506 Subtotal 73 (47) 81 (53) 107 (46) 78 (66) 40 (34) (54) (100)

3.3.9 Idiopathic conditions

Seven koalas (1%) presented with multinodular masses in the spleen (Figure 3-12), termed by the authors the “koala fibromatosis syndrome”. Nodules were formed by well differentiated fibrocytes which effaced and infiltrated the spleen (Figure 3-13) and liver. The fibromatosis lesions also affected the liver in 3 cases (Figure 3-14, Figure 3-15). No specific comorbid pattern was observed in these koalas; 3 of 7 had chlamydiosis, and an intestinal fibrosarcoma, senescence, and trauma was diagnosed in 1 of 7. Four of 7 of the koalas affected were males.

Lympholysis with moderate-to-marked white pulp depletion, was present in 9 (2%) koalas. Endocrine lesions were rare; the most striking being the histological evidence of mild-to- severe loss of adrenal chromaffin cells in 4% of koalas (21 of 519), often observed in older low BC females (61%, 13 of 21). Gross appearance of the adrenal gland was unremarkable.

Figure 3-12. Severe fibromatosis, spleen. The entire splenic architecture is markedly effaced by 1-3 cm firm, multinodular masses of collagenous tissue.

Figure 3-13. Fibromatosis, spleen. Masson’s trichrome. Multifocally there are collagen fibers and fibroblasts forming nodules of variable sizes, admixed with multiple vascular caverns and channels. Despite the infiltrative characteristic of the condition, the vascular endothelium and fibrocytes were quiescent and well-differentiated.

Figure 3-14. Fibromatosis, liver (same koala as Figure 15). Multifocally, the liver has focal, white-pink, multinodular, raised, well-demarcated, and non-infiltrative nodules unlike their splenic counterparts.

Figure 3-15. Fibromatosis, liver. Masson’s trichrome. The nodules are comprised by fibrocytes and collagenous fibres with well-defined margins.

3.4 Discussion 3.4.1 Benefits of autopsies and passive surveillance for detecting disease in wildlife populations

The overall goal of the study was to determine the causes of mortality in South-East Queensland koalas, particularly the presence of and interplay of comorbidities, using a passive surveillance approach. This koala population has undergone a major decline in the last 20 years, approaching functional extinction in certain areas affected by severe habitat 69 fragmentation and elevated mortality (Rhodes et al., 2011). There is a need to understand the complex interplay of threats koalas are facing in this region in order to develop mitigation and population management strategies which will allow the species to recover. There is also a need to assess comorbidities in koalas because they lead to debilitation, limiting their breeding potential and increasing their vulnerability to disease or predation (Canfield, 1987d), further exacerbating population decline. Lastly, understanding the interplay and relationship between comorbidities and trauma provides insights into some of the most significant drivers of the population decline in koalas in the region.

Koala carcasses for the present study were obtained via a passive surveillance approach, defined as the opportunistic gathering of data resulting from other health-related activities (Sergeant and Perkins, 2015). There are many benefits in incorporating systematic autopsies to koala disease surveillance. Some of these benefits include monitoring primary causes of death, generation of koala basic health statistics and baseline levels of disease from increased sensitivity in disease detection and comorbidity pattern compared to a previous retrospective epidemiological analysis using medical records also acquired through passive surveillance (Gonzalez-Astudillo et al., 2017), and characterization of novel diseases (Sergeant and Perkins, 2015, Thrusfield, 2007). Koala studies utilizing active surveillance may produce more accurate data to determine true disease and infection prevalence and incidence data from the population as it will identify healthy individuals (Sergeant and Perkins, 2015, Chaban et al., 2017). However, active surveillance poses logistical and financial challenges as well as ethical, and animal welfare implications as it involves the disturbance of wild animals in their natural habitat (Sergeant and Perkins, 2015).

A systematic nomenclature and organizational system for disease classification had been previously developed for the retrospective data analysis. In this system, disease was classified by organ system, body part, etiological agent and final diagnosis (es) for each koala (Gonzalez-Astudillo et al., 2017). This classification scheme was utilized in the present study to record autopsy findings, permitting logical, accurate, and rapid data recording, and facilitating analysis of both primary disease and associated comorbidities. The use of systematic nomenclature is encouraged for the future monitoring of koala diseases, and to facilitate comparisons with databases at a national scale such as the recently-established surveillance program with zoological institutions in Australia (Cox-Witton et al., 2014).

Historically, the investigation of koala diseases via autopsy has been often overlooked, hampering the understanding of the pattern and type of comorbidities affecting koalas. A previous retrospective epidemiological study in the region (Gonzalez-Astudillo et al., 2017) had not been able to thoroughly interrogate comorbidities because systematic autopsies were not consistently conducted in every deceased koala. Additionally, in many instances only the obvious cause of death was recorded without further detailed investigations. Autopsy technique should be both consistent and thorough to ensure the examination of the entire koala carcass and organs by both gross examination and histopathology. Failure to inspect a koala carcass or inconsistent technique may yield an incomplete diagnoses for subclinical conditions or diseases during early stages, affecting overall knowledge of koala diseases.

A higher sensitivity to disease and comorbidity detection in the present study can be attributed to the application of a consistent systematic autopsy methodology, standardization of recording terminology to construct a pathology database to clarify or corroborate a clinical diagnosis, and the availability of specialized testing. Additionally, autopsies helped collect data to determine the true health status of koalas and quantity how they may be impacted by trauma. For example koalas killed by animal attacks were included in a statistical assessment investigating how BC would influence the likelihood of animal attacks in koalas. Consistency in autopsy technique, level of expertise of supervising pathologists, medical record and ancillary testing availability were factors leading to the establishment of a high (>90%) proportion of conclusive autopsy diagnoses in this study, a high rate for wildlife. Other free-ranging wildlife autopsy studies report variable rates of determination of primary causes of death, ranging from 46% in manatees (Shea et al., 1985), 75% in belugas (Lair et al., 2016), 92% in common loons (Stone and Carrick, 1996), and 95% in common doves (Gerhold et al., 2007). Main factors influencing the achievement of a diagnosis in wildlife autopsy include quality of lesions or carcass preservation, freezing artefact, collection method, and species.

3.4.2 Infectious diseases

Infectious diseases were the predominant diagnostic category detected in this study, and this was similar to other SEQLD cross-sectional studies (Burton and Tribe, 2016, Loader, 2010, Weigler et al., 1987)., While most koalas readily acquire infections, the numbers recovering (Mitchell, 1991, Cockram and Jackson, 1981) and perishing are not consistent (Cockram and Jackson, 1981), with disparity in disease manifestation in populations with 71 high levels of infection (Jackson et al., 1999). This suggests the existence of additional host or environmental factors driving the manifestation of disease in koalas that need to be determined. In the present survey, as in previous studies, the main infectious disease diagnosed in koalas was chlamydiosis.

3.4.3 Chlamydial diseases

Chlamydiosis, the disease caused by the bacterium Chlamydia spp., is the most frequently reported disease in koalas, reaching prevalence of over 80% in certain eastern koala populations (Polkinghorne et al., 2013). Generally, chlamydial infection is classified as subclinical i.e. lacking of clinical signs but present on culture or molecular testing, overt disease i.e. severe external signs or inapparent overt disease, requiring ultrasonography or an autopsy to identify structural changes ( Brown et al., 1987). Two species of Chlamydia spp. has been isolated in koalas, Chlamydia pneumoniae causing conjunctivitis and pneumonia, and C. pecorum, documented as more frequent and pathogenic, leading to conjunctivitis, cystitis and reproductive infection (cystic dilation of ovarian bursae, oviducts and uteri, vaginitis and pyometron), resulting in infertility (Higgins et al., 2005b, Polkinghorne et al., 2013, Burach et al., 2014, Blanshard and Bodley, 2008). Renal infarcts have been known to occur in koalas (Canfield, 1989). Despite the low proportion of koalas with renal infarcts in the study, these were included in the putative chlamydial urinary group because of the known association C. pneumoniae has with atherosclerosis/thrombosis and the bacterium’s ability to induce systemic vasculitis affecting low and high caliber blood vessels, occasionally leading to myocardial infarction in humans (Ljungstrom et al., 1997, Vainas et al., 2002). The classical descriptions of chlamydial infections in koalas such as urogenital and ocular infection were all detected in this survey, and frequently represented chronic disease stages. Each of the classical chlamydia forms affecting the urogenital and ocular system correspond to the hallmark lesions associated with chlamydiosis in wild koalas (Polkinghorne et al., 2013, Obendorf, 1981, Hirst et al., 1992, Brown et al., 1987, Canfield et al., 1983, Canfield, 1989, Canfield, 1991, Canfield and Spencer, 1993). Therefore, the gross and histological presentation of these lesions is sufficient evidence to support a diagnosis, making molecular testing complementary. Nonetheless, N=2 mature male koalas clinically diagnosed with cystitis were examined via histopathology and real-time PCR. Histopathology revealed lymphoplasmacytic inflammations affecting the urethra and/or prostate and were positive to C. pecorum, partially supporting our findings. In koalas, chlamydial detection without assessment for associated pathology has limited prognostic value as due to subclinical or carrier states, the presence of the organism by PCR testing does not equate with clinical disease (Weigler et al., 1988). Further, as recent reports have shown in QLD (Nyari et al., 2017), koalas with significant chlamydia-associated structural pathology, such as the chronic 72 disease stages predominate in our dataset, may yield a PCR-negative result. Intermittent shedding may be the mechanism influencing the discrepancy in disease manifestation versus molecular results, as demonstrated by koalas infected with high chlamydial loads from C. pecorum but lacking clinical signs, or koalas with marked disease but with negative PCR results (Nyari et al., 2017). Determinants of the expression of chlamydial clinical disease can be associated with host-pathogen (KoRV subgroup (Waugh et al., 2017, Chaban et al., 2017)), or environmental factors (habitat loss, fragmentation) (Lunney et al., 2012). Nonetheless, elements predisposing koalas to chlamydial disease expression, versus subclinical infection were beyond the scope of this analysis.

The vast majority (>90%) of infections found in this study through systematic autopsy method were due to chlamydiosis. Establishing comparison to other studies is complicated by the variety of study design and approaches utilized. For instance, early studies applied serology which proved unreliable (Hirst et al., 1992), and only nowadays accurate data is being produced via molecular methods. Nonetheless these studies indicate that the bacterium is present in virtually every population assessed with infection prevalence ranging from 0% found in koalas inhabiting islands, to over 80% in continental populations (Devereaux et al., 2003,Jackson et al., 1999). Studies conducted in QLD have reported variable levels of chlamydial infection rates (10-85%) and disparity in clinical disease compared to disease status (Loader, 2010, Jackson et al., 2003). The generally higher proportion of chlamydiosis in our study (58.5% 304 of 519) could also reflect the bias in the submission of animals with diseases at chronic and end-stages, versus examination of wild populations which include carriers and healthy animals. Elevated frequencies of milder chlamydial infections reported by other studies (Loader, 2010) compared to ours are plausibly due to discrepancies from sampling method (active surveillance). Thus, our methodology will overestimate the prevalence of severe end-stage chlamydiosis in wild populations, and underestimate subclinical carrier states, but was appropriate for identifying the overall impact of chlamydiosis as a cause of mortality and infertility in the population. Further, as the systematic autopsies investigated all body systems, the multicentric nature and the distribution of chlamydiosis throughout various target organ systems in the koalas was accurately determined.

3.4.3.1 Urinary disease

The main clinical manifestation of chlamydiosis in koalas is lower UTI, detectable visually during chronic, moderate to severe stages as ‘wet bottom’ in trees or clinically during an external examination (Blanshard and Bodley, 2008). Gross inspection during an autopsy combined with histopathology detects mild, subclinical, externally inapparent cystitis and

73 conjunctivitis cases, thus had increased sensitivity (Speight et al., 2016, Funnell et al., 2013). Cystitis was found in similar proportion in the current study (34%) to other reports (Canfield, 1989, Canfield, 1987d, Obendorf, 1983) but higher compared to the proportion reported in an epidemiological study conducted by the authors (27%),(Gonzalez-Astudillo et al., 2017) likely from the increased sensitivity of systematic autopsies. During chronic stages, cystitis can result in variable forms of renal complications ranging from obstructive disease to chronic or active renal inflammation causing fibrosis (Canfield and Spencer, 1993). The differential diagnosis for chlamydia associated renal asymmetrical scarring in koalas include chronic pyelonephritis, prolonged nephrotoxic damage or chronic renal ischemia (Canfield and Spencer, 1993) as in domestic species. Obstructive uropathies in koalas have been documented (Canfield and Spencer, 1993, Connolly, 1999) presumably due to ureteral intramural fibrosis when passing through the bladder wall (Polkinghorne et al., 2013) at the ureteral papilla. However, given the increased prevalence of chlamydia as a cause of urinary tract disease, it is considered to be the most likely cause of kidney changes in a majority of the koalas in this study. Occasional ascending UTI are known to cause chronic renal infections (Canfield, 1989), nearly half (47%) of cases in this study associated with other classical lesions of chlamydiosis. The proportion of renal lesions following cystitis reported by other autopsy surveys is lower (5-34%)(Canfield, 1989, Canfield and Spencer, 1993), possibly due to differences in regional risk factors altering susceptibility to infection or simply study design, impeding accurate comparisons.

3.4.3.2 Reproductive infections

A disproportionate number of females of breeding age were diagnosed with urinary and reproductive chlamydial infection in comparison to males, despite the documented similarity in subclinical infection rates for both sexes (Speight et al., 2016, Polkinghorne et al., 2013). This finding reflects the documented high proportion of females with reproductive disease (Stalder et al., 2003, Loader, 2010, Canfield et al., 1983).,Factors accountable for the increased rate of reproductive chlamydiosis diagnosed clinically in females versus males include urogenital anatomy (Obendorf, 1981), increased rate of opportunistic infections resulting in complications (Connolly, 1999), and lack of sonographic changes in the male (Loader, 2010) unless severely affected. Additionally, increased rates of female diagnoses may be influenced by the effectiveness of palpation in detecting chronic chlamydial bursitis without specialized equipment. In koalas, ultrasonography can be utilized in both sexes to diagnose reproductive disease (Blanshard and Bodley, 2008). Nonetheless, an increased number of affected females would be detected and graded (Stalder et al., 2003) via 74 ultrasonography when triaged during surveillance at hospitals to identify irreversibly infertile individuals (Canfield et al., 1983, Johnston et al., 2015).,

In females, infertility results from chronic chlamydial bursitis or inflammatory and degenerative sequelae (e.g. tubal infertility due to obstructive fibrosis) and cystic dilation of the ovarian bursa, which surprisingly, did not affect the ovaries (Obendorf, 1981). Koalas with occlusive fibrosis (Higgins et al., 2005a) tend to lack inflammatory infiltrate or in the presence of inflammation, lack inclusion bodies with routine staining (Hemsley and Canfield, 1997). Besides the impact of infertility, infected females with chronic, multi-systemic disease have been associated with high chlamydial shedding rates (Wan et al., 2011) via macrophage dissemination (Higgins et al., 2005b, Burach et al., 2014). Thus, the detection of persistently-infected koalas results in euthanasia as a population management strategy as they become chronic carriers (Wan et al., 2011), actively transmitting chlamydiosis as they continue to cycle and mate but are unable to produce young (Griffith and Higgins, 2012). Detecting male chlamydiosis is relevant because urogenital infection can adversely impact fertility (Johnston et al., 2015). Supporting our findings, increased detection rates of subclinical disease with histopathology versus clinical exam can explain the historical underreporting of male koala urogenital disease documented for the region (Gonzalez- Astudillo et al., 2017, Loader, 2010).

3.4.3.3 Ocular chlamydiosis

Ocular chlamydiosis was predominantly chronic and correlated with low BC, plausibly compromising survival. However, in this study conjunctivitis was a common manifestation with co-occurring, debilitating, multicentric chlamydial disease. In animals with solely ocular chlamydiosis, only 9% had low BC, suggesting solely ocular pathology does not significantly affect the systemic health of the koala, rather it is multicentric disease that causes the severe debilitation and BC loss. Previous QLD reports questioned the relationship between low BC and ocular disease (Hirst et al., 1992), but only the present study has correlated it with comorbidities. In this study, >60% of koalas with low BC had ocular chlamydiosis concomitant with comorbidities. Thus it is plausible that general systemic disease rather than sole visual impairment (Hirst et al., 1992) is more likely to cause systemic deterioration in koalas reducing their BC and compromising survivorship. The low proportion of koalas with acute or milder ocular chlamydiosis in our sample set is likely a reflection of the bias in the sampling method towards severe or chronic infections and animals with permanently compromised vision (Hirst et al., 1992). As milder cases of ocular chlamydiosis were likely 75 to be successfully treated and released at veterinary hospitals. An active surveillance methodology or mark and re-capture study would assist in monitoring proportion of subclinical carriers (Studdert, 2011) and milder ocular infections in the general population.

3.4.4 Respiratory disease

In this study 4% of koalas had pulmonary disease, including suppurative pneumonias and bronchopneumonias. Our investigative sampling approach permitted the description of novel koala respiratory pathogens leading to acute focal or diffuse lung infections, increasing the knowledge of primary and secondary pathogens of the so-called koala rhinitis/pneumonia complex (Blanshard and Bodley, 2008). These included Psychrobacter sp. which can cause lung infections in sheep (Vela et al., 2003), Cladosporium spp. and non-tuberculous M. abscessus are known to infect lungs in immunocompromised hosts (Castro et al., 2013, Stout et al., 2016)., Certain koala chlamydial isolates have been suggested to have a close association with ocular and respiratory infection (Polkinghorne et al., 2013, Jackson et al., 1999, Burach et al., 2014). Despite this, no causal relationship between chlamydial organisms and lung disease in koalas has been established under experimental conditions to date. However, one report documented chlamydial organisms infecting the bronchiolar epithelium in a koala causing pneumonia (Mackie et al., 2016). Although of minor occurrence in the present study, resolved pulmonary insults have been previously reported in QLD koalas as idiopathic pneumonia initiated as a bronchitis (McKenzie, 1981).

3.4.5 Trauma

Historically, trauma has been the major contributor to koala mortality in SEQLD (Gonzalez- Astudillo et al., 2017). Koala-motor vehicles collisions have a high fatality rate (>80%) (Dique, Thompson, Preece, Penfold et al., 2003) from a combination of factors (Henning et al., 2015, Canfield, 1991), primarily because collisions result in comminuted fractures (Canfield, 1987d, Canfield, 1991) and associated severe soft tissue trauma. Another important cause of trauma was animal attack, particularly from dogs. In our previous retrospective data analysis, overall trauma was detected in 48% (9,703 of 20,250) individual koala admissions, compared to 36% in the present study, constituting a drop in the proportion of lethal or terminally injured koalas admitted to hospitals.

Trauma sources identified in the retrospective data analysis were varied and included motor vehicles 66% (N=6,430), animal attacks 22% (N=2,252), and miscellaneous sources 12% 76

N=1,161), with 40 animals with multiple traumas being excluded from these numbers. The proportions in the current study are lower (motor vehicles 58%, animal attacks 21%) except those for miscellaneous sources (21%).

In the same retrospective study a considerable portion (37% 7,430 of 20,250) of koalas killed by trauma were otherwise presumed healthy. Thus, approximately one third of animals killed by trauma, could be considered unnaturally removed, reducing the population of healthy breeding stock. Although the combined proportion of diseased and healthy (36% 189 of 519) and of exclusively healthy and fatally injured koalas (18% 94 of 519) in the present study was lower, it demonstrates trauma remains a frequent cause of mortality in QLD, particularly by motor vehicles (Burton and Tribe, 2016, Gonzalez-Astudillo et al., 2017). We found that over half of koalas with chronic sequelae from infections also affected by trauma where still in good BC. This result likely reflects the high levels of infection and disease in the population. The removal of sick and debilitated koalas may be potentially beneficial due to reduced competition for resources (Dique, Thompson, Preece, de Villiers, 2003), opening territory for inter-group migration, as well as eliminating potential chlamydial- exogenous KoRV carriers.

The demographics of trauma are explained by the species biology, where dispersing animals encounter sources of trauma during the post-natal period, displacement by territorial koalas, or breeding season (Mitchell, 1990b, Logan and Sanson, 2002). During post- weaning, both young female and male koalas attempt to disperse (Dique, Thompson, Preece, de Villiers, 2003). Male koalas also increase ground travel time enhancing opportunities for trauma when searching for mates during breeding season or when displaced by territorial dominant koalas (Mitchell, 1990b, Logan and Sanson, 2002). In QLD, there is evidence that male subadult koalas dispersing between groups contributes to gene flow (Ellis et al., 2002b). Thus, the premature removal of healthy younger male koalas from trauma can disrupt gene flow negatively influencing koala population heterogeneity, a concerning pattern as 45 of 94 of the injured but healthy koalas in the present survey were in the young/subadult age class (Fowler et al., 2000).

Trauma remains one of the main threats to the stability of koala populations, particularly in the presence of fragmentation. However, other primary threats such as disease and its association to KoRV infection (Chaban et al., 2017), as well as the rapid urbanization and clearing of koala habitat, need to be tackled in integrated koala conservation strategies considering habitat loss, disease, and trauma (Rhodes et al., 2011). When koalas inhabit 77 forested land in close proximity to urbanized centres, there is high mortality from trauma, demonstrating poor adaptability when anthropogenically-related threats permeate into their habitat (Preece et al., 2017). A marked reduction of koala habitat clearing is needed to protect koala viability in SE QLD from urbanization threats. Additionally, road mitigation strategies are needed, such as over or underpasses (Dexter et al., 2016) which may not only benefit koalas – especially those that are fit to reproduce, but other wildlife as well. Trauma by animal attacks affected 8% (40 of 519) of koalas in the present survey, the majority caused by dog predation. Although wild dogs are known to predate on native species (Department of Employment, Economic Development and Innovation, 2014), the relative contribution of wild versus domestic dog predation in koalas is yet to be quantified. Elevated rates of predation are another inevitable consequence of urbanization of koala habitat particularly in coastal regions (Department of Sustainability, 2013). Domestic dog predation can be lessened by responsible ownership via educational campaigns. Koalas are crepuscular species (Martin and Handasyde, 1999) so the key factor is to restrict dog movements on private land during dawn and dusk, the peak koala activity times.

3.4.6 Idiopathic loss of body condition & KoRV

Koalas live on a highly specialized, nutrient-restricted diet, and have limited visceral fat reserves even with ideal BC compared to eutherian mammals (Martin and Handasyde, 1999). As a result, koalas are prone to debilitating and sometimes irreversible BC loss with the onset of mild illness, temporary climatic phenomena, or even alterations in plant chemistry resulting in nutrient deprivation (Moore and Foley, 2005, Gordon et al., 1988, Seabrook et al., 2011a, Gonzalez-Astudillo et al., 2017). Thus, BC is a relevant parameter to assess in koalas (Gonzalez-Astudillo et al., 2017) as it may impact successful rehabilitation, population management, or interact with comorbid states and reflect disease severity or chronicity. In order to assess this phenomenon, BC scores were determined for each koala to later evaluate their association with disease and injury. Further, we were interested in investigating if BC as a parameter could predict susceptibility to causes of disease and injury. Previous work by Canfield (1987a) suggested the presumptive role of poor BC in predisposing koalas to carnivore predation (Canfield, 1987d). Our study demonstrated low BC was significantly correlated with increased likelihood of animal attacks, the majority of which are caused by dog predation. A statistically significant association of low BC with animal attacks may be the result of increased time spent on the ground when debilitated by disease in urbanized landscapes. Although there is evidence that most animal attacks on koalas come from domestic dogs (Griffith et al., 2013, Lunney 78 et al., 2007, Dique, Thompson, Preece, de Villiers et al., 2003), the threat of carnivore predation could also permeate into protected forests or bushland habitat (Queensland, 2016) when carried out by carnivores roaming freely (feral dogs, dingoes) (Canfield, 1987d).

The retrospective epidemiological study suggested but could not confirm, that the high number of koalas with poor BC corresponded to undiagnosed chronic disease (Gonzalez- Astudillo et al., 2017). The uncertainty regarding a possible relation of disease and poor BC derived from a failure of application of systematic autopsies to every deceased koala in this retrospective dataset. Our study found that virtually half of diseased koalas (48%, 249 of 519) in the study had poor BC, somewhat supporting the original suggestion made by the previous report. Conversely, and also providing some support to previous findings, 94.7% (89 of 94) of koalas only affected by trauma were in adequate BC.

Besides chronic disease, other factors leading koalas to progressive BC deterioration including those associated with aging. Dental attrition was included as a condition because it compromises mastication in koalas, reducing digestion efficiency, and further lowering the BC (Ellis et al., 2014, Lanyon and Sanson, 1986). Other causes of low BC include stressors such as prolonged veterinary treatment, or those derived from biotic (social dynamics, resource availability, predation) (Booth et al., 1990) and climatic factors (Davies et al., 2013).

Idiopathic loss of body condition was noted in a small number of koalas in this survey. Notably, males were diagnosed more frequently with disease-associated and idiopathic loss of BC. The idiopathic loss of BC found in males primarily affected young/subadult koalas, and typically was associated with non-specific clinical histories (e.g. dehydration) and autopsy findings. The inability to establish a home range (including consumption of appropriate browse) could explain a progressive BC loss primarily in males. Additionally, a NSW study reported limited or minimal pathology in 12% (15 of 127) of koalas with poor BC, affecting individuals of varied ages but mostly females, reporting lymphatic depletion as a common finding. Canfield (1987a) Obendorf (1983) noted that 29% (13 of 44) of koalas presented for autopsy had non-specific findings including loss of muscle bulk, dehydration, and atrophy of the lymphoid tissue and of the adrenal cortex amongst others. This was later termed the ‘koala stress syndrome’, and ill-defined entity, and affected wild koalas of both sexes (Obendorf, 1983). In response to systemic stress, marsupials activate the hypothalamic-pituitary-adrenal axis to secrete glucocorticoids, such as cortisol (Stewart, 2003), known to have lympholytic effects (Selye, 1950). Although both Canfield (1987a) and Obendorf (1983) report lymphoid-related issues, evidence of lymphoid depletion was not a 79 consistent finding across koalas with idiopathic loss of BC in the present study. Previous autopsy surveys in QLD have reported mixed non-specific findings, from putative immune- suppressive related disease to unexplained chronic ill-thrift in 6% (Loader, 2010) and 11% of koalas (McKenzie, 1981). Small-scale autopsy surveys in NSW also documented 25% of koalas autopsied with non-specific findings lacking severity to justify death (Backhouse and Bolliger, 1961b). The lack of consistency and nonspecific nature of autopsy findings across all koalas affected by idiopathic loss of BC limits our ability to determine the cause of this condition. Nonetheless, these studies note the importance of gathering meaningful gross and histopathological data from inconclusive autopsies so future correlations can be made with clinical results. The diagnostic rule outs for these cases include diseases inducing metabolic derangement and biochemical changes lacking structural modifications, putative microbiome alterations, starvation induced by clearing or unfavorable changes in plant chemistry (IUCN, 2009, Barker et al., 2013, Moore and Foley, 2005).

Wildlife clinicians have noted a suite of pathological conditions causing premature deaths in koalas regardless of medical treatment (Chaban et al., 2017). Upon autopsy examination, findings are non-specific, prompting the grouping of a category termed the ‘AIDS-like syndrome'. This syndrome is defined by the occurrence of two or more of the following clinical signs or diseases: unexplained chronic ill-thrift, severe dermatopathies, candidiasis, periodontal disease and stomatitis, and gastrointestinal disease (Loader, 2010, Chaban et al., 2017). Koala diseases commonly associated with the AIDS-like syndrome include candidiasis, chlamydiosis and lymphosarcoma (Loader, 2010, Xu et al., 2013). Studies in wild SE QLD koalas by Loader (2010) demonstrated these conditions in 6% of koalas surveyed, higher than the 3.5% koalas found affected by these koala AIDS-like conditions in the present study, discrepancies likely arising from a combination of clinical examination and autopsy as sampling approaches. QLD koalas carry a diverse exogenous and endogenous KoRV variants (Chappell et al., 2017, Fabijan et al., 2017). Prevalence to endogenous KoRV variants is 100% in QLD and infected populations tested have mean high proviral loads (Simmons et al., 2012). The literature discusses a lower disease incidence in KoRV-free populations, such as those in southern Australia, compared to KoRV-positive populations, like Queensland, due to immunosuppression (Tarlinton et al., 2006, Waugh et al., 2017). A possible association of the AIDS-like syndrome and exogenous KoRV had been proposed (Tarlinton et al., 2006, Xu et al., 2013); and more evidence has been recently gathered for wild QLD koalas to further establish a link between the virus and this disease 80 complex (Chaban et al., 2017). Chaban (2018) found that in SE QLD koalas 43% of those diagnosed with the AIDS-like syndrome were in fact infected with KoRV-B, an exogenous subgroup linked to disease manifestation in northern populations (Waugh et al., 2017). Although KoRV status was beyond the scope of the present study, all the aforementioned conditions described by several reports as koala ‘AIDS-like syndrome’ were observed in this study, strongly suggesting an immunocompromised status in these individuals. If disease manifestation by chlamydiosis and other opportunistic infections depends on KoRV-induced immunodeficiency, then the best conservation efforts should be directed on limiting the spread of exogenous KoRV through immunization (Waugh, Gillett, Pokinghorne et al., 2016, Chappell et al., 2017), not that of chlamydial infection.

3.4.7 Idiopathic conditions

The application of systematic autopsies broadens our understanding of the spectrum of diseases affecting koalas, and provides pathological descriptions of novel diseases, such as serosal myxosarcomas (Gonzalez Astudillo et al., 2015) and a better opportunity to investigate pathogenesis. Several idiopathic conditions were noted in our study including chromaffin cell loss, and fibromatosis. Although infrequent, some koalas had evidence of loss of chromaffin cells, the neuroendocrine cells populating the adrenal medulla that release catecholamines. The koala adrenal gland has been investigated (Booth, 1987, Davies et al., 2013),due to the koala’s apparent vulnerability to stress during hospitalization and standard veterinary care triggering maladaptive stress responses (Narayan et al., 2013), linked to the ‘koala stress syndrome’(Marik and Levitov, 2010). Despite the interest in investigating the koala adrenal gland function, chromaffin cell loss has not been reported in these studies. Cellular death is likely the mechanism involved in chromaffin cell loss, although the causes behind it need to be further studied. Autophagy, a proposed distinctive type of cellular death, is a frequent reaction of sublethally injured cells, also evidenced in atrophy by glandular cells undergoing cyclic physiologic regression (McGavin and Zachary, 2006). Autophagy is already documented in the post-natal regression of extra-adrenal chromaffin cells innate to most mammals (Schober et al., 2013). Evidence of autophagy genes would need to be demonstrated in koala intra-adrenal chromaffin cells to establish an association with autophagocytosis. Another mechanism could be apoptosis or necrosis, but in the present study there was no evidence of either process (cell shrinkage, karyolysis, karyorrhexis, inflammation, haemorrhage, etc.). Although establishing cause or effect with stress cannot be discerned from this finding, losing intra-medullary chromaffin cells could

81 compromise neuroendocrine secretion. The physiological impact of this morphologic change needs further study.

To our knowledge, the present study documents the first cases fibromatoses in koalas. The “koala fibromatosis syndrome” primarily affected the spleen, and occasionally the liver and was characterized histologically by an infiltrative growth of well differentiated fibroblastic cells and production of collagen in large quantities. In macaques, retroperitoneal fibromatosis has been associated with simian retrovirus type 2 (Yanai et al., 2003). Despite lymphoreticular organs not being affected in non-human primates, the histopathological description of fibromatosis in mesenteric lymph nodes mirrors the description of lesions in the liver and spleen in seven koalas in the present study. The histological evidence suggests the fibromatous is benign, however it has the potential to impact individual animal medical management, as these growths can be misdiagnosed as malignant neoplasms on abdominal ultrasound, which is routinely used in assessing females for bursitis. Further research is needed to determine the cause, pathogenesis, and impact of fibromatoses on koala populations.

3.4.8 Neoplastic disease

By conducting systematic autopsies in all deceased koalas, the malignant nature and specific organs affected by neoplastic disease were accurately detected. A higher proportion of koalas with neoplasms were described compared to previous SEQLD (Gonzalez-Astudillo et al., 2017) and NSW studies (Canfield, 1987a), supporting the documented increased disease occurrence in northern koalas (Simmons et al., 2012). Of all neoplasms, lymphosarcomas were frequent (Hemsley et al., 1996, McKenzie, 1981), often displaying multiple organ involvement, suggesting admissions at advanced disease stages.

3.4.9 Comorbidities in koala studies and Seasonality of Admissions

One of the greatest benefits of our study was the ability to accurately and specifically interrogate the presence and interplay of comorbidities in the SEQLD koala population. The study demonstrated a high rate of comorbidity, and thus reinforced that the threats facing this population are complex and multiple. An elevated proportion of koalas were found with comorbidities that led to debilitation, and potentially affecting their viability and breeding ability. Often, comorbidities were caused by the multi-centric occurrence of chlamydiosis primarily affecting the urogenital (Obendorf, 1981, Canfield, 1987c, Weigler et al., 1987,

Obendorf, 1983) or ocular system (Canfield, 1987c, Weigler et al., 1987, Canfield, 1991), resulting in loss of body condition (Weigler et al., 1987, McKenzie, 1981).

Because of the impact comorbidities have in koalas, the simultaneous occurrence of conditions has been previously investigated in other autopsy studies, documenting variable comorbidity rates. The literature was explored to determine the most frequent co-occurring conditions diagnosed in koalas in QLD and elsewhere. We found that the most common comorbidities reported are multi-centric infections and infections in koalas suffering from trauma. Specifically, koalas were found to be traumatically injured and diseased with chlamydiosis (Canfield, 1987c, Weigler et al., 1987, Canfield, 1991), found injured with poor BC (Canfield, 1991, Canfield, 1987c), or injured and affected by other diseases in less frequency (sepsis, various infections in specific organs, complications with fracture healing) (Canfield, 1987c, Backhouse and Bolliger, 1961b, Canfield, 1991, Obendorf, 1983). Multi- centric infections affected the urinary (lower and upper tract) and reproductive system primarily female koalas (Obendorf, 1981, Canfield, 1987c). Although loss of BC as a consequence of comorbidity was reported in VIC (Obendorf, 1983) as well as in NSW (Canfield, 1991, Canfield, 1987c), the reports derived from QLD generally documented a higher frequency of koalas with loss of BC from comorbidities such as chronic infections or neoplastic disease (Weigler et al., 1987, McKenzie, 1981).

Direct comparison of comorbidity rates with these reports is complicated by the study design used and focus of research. Nonetheless, the determination of discrepancies or similarities in comorbidity categories assists in establishing a historical pattern for disease interaction in koalas. These studies stress the interplay of trauma and chlamydiosis as main threats for koalas across states, and their effect on the long-term stability of fragmented koala populations in SEQLD.

The differences in the comorbidities rates in these studies could be attributed to elevated rates of disease in the koala population in northern koala populations (Waugh et al., 2017) or spatial-temporal threat variability. If discrepancies are related to method, these could be explained by prolonged sampling timeline (> 2 years), utilization of fresh samples, selection of smaller sample sizes (Backhouse and Bolliger, 1961b, Weigler et al., 1987, Canfield, 1991) and selection of a targeted surveillance sampling method, where monitoring effort is directed to high-risk populations to measure disease levels or absence (Thrusfield, 2007, Salman, 2003a).

Seasonality is an important driver in admissions of wild koalas to hospitals, although the retrospective study only found season to be a risk factor only influencing aged koalas with low BC and with chlamydiosis (Gonzalez-Astudillo et al., 2017). Admissions of koalas with this diagnosis peaked during the July-September period. In the autopsy survey, the peaks observed in admissions for the final diagnosis Trauma by motor vehicles occurred during winter and early autumn likely from post-natal dispersion, and during spring and early summer, plausibly due to increased ground transportation time during breeding periods. Biphasic seasonal peaks in hospital admissions were observed for the final diagnosis combination of Chlamydia-like signs/Wasting during summer through early autumn and for spring and early summer. Although seasonal trends have been reported for acute ocular chlamydia infection with increased frequency of cases during summer (Cockram and Jackson, 1981), that would not explain the high prevalence of chronic chlamydiosis cases found in this study. Other factors influencing the dynamics of seasonal disease manifestation need to be investigated further. Increased numbers of disease cases during autumn are thought to be due to cold stress as documented in captive marsupials (Canfield and Cunningham, 1993). Generally, there was a correlation of temporal trends of hospital admission in the retrospective data analysis and prospective autopsy survey. Both studies noted peak submissions in August through to October, and the autopsy survey recording high admissions through November. Higher number of admissions for all cases during these months can be linked to post-natal dispersal, whereas increased hospital admissions during August to November are more likely to result from territory or mate search in QLD (Canfield, 1991, Griffith et al., 2013, Henning et al., 2015). Identification of factors driving variance in disease is difficult due to the complex nature of infection, although factors driving koala disease manifestation such as KoRV-related immunodepression, are finally being unveiled (Chaban et al., 2017).

3.4.10 Limitations of the study & Conclusions

There are restrictions to data interpretation and in producing an accurate diagnosis when the investigative approach is passive surveillance. Some of these limitations are minimal control over data quality, lack of denominator values to estimate disease frequency, and overrepresentation of certain demographic groups or diseases (Thrusfield, 2007) (i.e. incurable, terminal). For instance, chronic conditions were observed in high frequency grossly and histologically across the main diagnostic categories as well as low BC, reflecting a biased selection of animals that could not be saved by the clinics. Chronically affected koalas often had complications arising from opportunistic infections worsening chlamydial 84 disease (e.g. pyometron, suppurative conjunctivitis), and the combination of severity and comorbidity accounts for the decision to euthanize at wildlife hospitals. Acute pathology was infrequent across infectious causes of death in our study. Specifically, there was a low proportion of koalas at acute infectious disease stages, particularly chlamydiosis reflecting our sampling method not selecting for individuals where overall health is not significantly compromised (Hirst et al., 1992). Those animals with acute and less severe disease, especially chlamydial conjunctivitis or cystitis, and other bacterial infections would likely be successfully rehabilitated and released by the koala hospitals.

Sampling free ranging wildlife involves many challenges, in particular carcass preservation. In koalas, the quality of tissues available for study may have sporadically compromised diagnostic effort or induce histological artefact when koala carcasses were admitted eviscerated or not following the optimal freezing/thawing protocol. For instance, occasional evisceration, either from manner of death, partial autopsies or pap harvest, caused poor preservation of abdominal organs, and impeded ancillary testing due to contamination. In spite of these restrictions, utilizing hospital-admitted koala bodies and systematic terminology permitted the acquisition of significant amount of data; a rare and cost-effective opportunity to obtain information at the population level over multiple seasons.

Autopsies proved to be an effective method for koala disease surveillance as they detected subclinical conditions and identified comorbidities, as well as increasing the detection rate of diseases and pathogens affecting koala conservation compared to previous studies (Gonzalez-Astudillo et al., 2017). To our knowledge, this is the one of the most extensive pathological studies applied to a declining wild species in Australia to date, allowing the identification of major causes of death, comorbidity trends and permitting the statistical evaluation of variables influencing trauma. Habitat loss and fragmentation remains the primary cause of trauma so instating restrictive clearing for development in SEQLD should be considered, as well as building road capacity and eco-passages to avoid further disruption to metapopulation dynamics. Disease needs to be addressed by science-based approaches that identify risk factors driving occurrence, vaccination, increase accuracy of diagnostic testing and improve therapeutic options.

Chapter 4 Accuracy of diagnoses of koala disease and injury in South-East Queensland

4.1 Introduction

The main goal of a disease surveillance program is to gather and analyse data so it can be disseminated to stakeholders, who take action to control diseases (Thrusfield, 2007). Data collection can be carried out via active surveillance, defined as a systematic or regular recording of cases where pre-planned activities are designed and commenced by the primary user of the data (Salman, 2003b). Another method used to monitor diseases is passive surveillance, which involves the collection of data that is a by-product of other health-related investigative activities (Sergeant and Perkins, 2015).

There are advantages and disadvantages to the use of passive and active surveillance methods. In active surveillance, data collected is of higher quality and is controlled because sampling strategies are planned and designed to fulfil a specific purpose (Sergeant and Perkins, 2015). Because surveys are tailored for a specific goal, a more precise estimation of disease frequency can be achieved (Thrusfield, 2007, Sergeant and Perkins, 2015). However, some of the downsides of active surveillance are that, while timeframes of data collection might be shorter, it often requires more costly and time consuming procedures, requiring that the interested party (government, industry representatives, farmers) carries the entire cost of the data collection and restricting the number of animals to be processed (Maas et al., 2016, Sergeant and Perkins, 2015). With active surveillance sampling, expenses will inevitably rise if the occurrence of the target disease is rare necessitating a larger sample sizes (Salman, 2003b) and because of restrictions in costs, coverage and timeframe, it is difficult to target unknown or emerging pathogens with active surveillance (Thrusfield, 2007).

Passive surveillance, on the other hand, is often the first step in novel and emerging disease identification, and is often referred to as an ‘early disease or pathogen warning system’(Thrusfield, 2007, Sergeant and Perkins, 2015). This method provides generic data on diseases in a population (Sergeant and Perkins, 2015), and allows for the measurement of disease occurrence at a much lesser cost (Maas et al., 2016), and with greater population coverage (Thrusfield, 2007). Occasionally, disease trends or changes in disease patterns can be assessed with passive surveillance (Salman, 2003a) depending on the type of data available, and the disease of concern (Sergeant and Perkins, 2015); although follow-up

86 epidemiological investigations are often required (Russell and Franson, 2014). The main disadvantages of passive surveillance include the inability to estimate disease frequency from lacking denominator values (Thrusfield, 2007), and inability to conduct comparative studies (Salman, 2003b) because of little control over data quality and the lack of consistency in data collection. Additionally, because initial data collection is not guided by the purpose of passive surveillance, the information available for use may not always relevant. Further, the validity of the data collected by passive surveillance is entirely dependent on the data collectors (health care professionals, farmers, animal owners) - their disposition, awareness level, and knowledge of the particular disease (Salman, 2003b).

Wildlife disease monitoring can be carried out through passive or active surveillance. Passive surveillance is commonly used in wildlife because often, diseased or moribund wildlife are found opportunistically, offering chances to investigate morbidity and mortality (Wobeser, 2005). As funding is limited to conduct active wildlife monitoring, passive surveillance of wildlife provides an alternative for disease monitoring with a wider population coverage and at a reduced cost.

A data source for passive surveillance are often the existing veterinary or laboratory records. A study utilised 22-years of data from a pathology laboratory to investigate disease and trauma data from N=56 autopsied wild turkeys following a reintroduction program in Canada (MacDonald et al., 2016). In this report, laboratory admissions relied on reports of sick or dead animals by wildlife agencies and members of public. Another study utilised 37-years of data equivalent to N=2,583 avian accessions to a pathology laboratory in the US, to retrospectively determine mortality factors and patterns of mortality of certain avian families over time (Gonzalez-Astudillo et al., 2016). Admissions were made mainly by members of the public and field biologists.

Because of the higher investments in effort and associated costs, active surveillance of wildlife is often only utilised to monitor diseases of high economic impact or those at the animal-human interface. One study carried out in New Zealand, illustrates the use of active surveillance by conducting lethal trapping of possums, having been identified as an infection source of tuberculosis (Davidson, 1976), to measure and ultimately decrease the incidence of the disease in cattle (Livingstone et al., 2015). In another study, bat specialists trapped N=308 bats from 2004-2009 to investigate the circulation of the European Bat Lyssavirus type 1 between native bat species (Picard-Meyer et al., 2011).

A tool to diagnose, monitor and investigate animal diseases as part of active and passive surveillance programs is through clinical examination of live animals. Live and healthy wildlife may be examined after trapping or hunting (Hubalek et al., 1993, Pannwitz et al., 2009). Sick or injured wildlife might be admitted to wildlife hospitals (Department of Environment and Heritage Protection, 2016), sanctuaries (Currumbin, 2017), or zoo clinics (Zoo, 2018, Steele et al., 2000) by members of the public (Sack et al., 2017). In Australia, members of the public report sick or injured wildlife via a 24-hour emergency animal disease hotline to animal carer organisations, or directly to local veterinarians. Live wild animals admitted to wildlife hospitals undergo a physical examination to assess the type of lesions and body system affected, severity of disease, and to identify and record risk factors associated with disease occurrence (e.g. demographics, species, seasonality). During clinical appraisal, a list of differential diagnoses or ‘rule outs’ is obtained, ultimately leading to a final diagnosis (Jackson and Cockcroft, 2007), which will estimate the progression or resolution of disease (prognosis) and determine the appropriate clinical management strategy such as treatment or euthanasia.

Another tool in active and passive wildlife surveillance is through pathological examination of dead animals (Wobeser, 2007, Cooper, 2002) namely an autopsy. Autopsies can be conducted from actively-sourced field-collected tissues (Gonzalez-Astudillo et al., 2016), lethal trapping (Warburton and Livingstone, 2015)) or through submissions by members of the public. Wildlife found dead by private citizens (Sack et al., 2017) or harvested during hunting seasons (MacDonald et al., 2016) can then be examined and sampled. The collection of good quality data during autopsies to generate a correct diagnosis that is useful for disease surveillance can only be achieved if a systematic autopsy technique is followed to make sure the entire carcass is examined, and case definitions are standardised. Common factors that may affect an autopsy diagnosis are poor carcass preservation, histological or sample processing artefacts (McInnes, 2009), lack of clinical histories, and the technical skill and knowledge of personnel. Carcass preservation is a particular challenge with wildlife autopsy as animals are frequently found dead, and only examined after a prolonged post-mortem interval. Further, the logistics of acquiring cadavers from geographically large habitat areas may impact of post-mortem interval and carcass preservation strategy and thus quality. However, autopsies are a valuable tool for understanding disease mechanisms, effectiveness of therapeutic regimes (Tejerina et al., 2012), as well as helping increase the knowledge of anatomy and physiology in wildlife species (Skerratt et al., 2005). An autopsy also serves as a quality control mechanism for

88 clinicians to confirm or correct their diagnosis and provide feedback on the impact of case management (McGavin and Zachary, 2007, Shojania and Burton, 2008). Lastly, autopsy in wildlife can be used to discover and describe existing pathogens or detect emerging diseases (Nation, 2015).

Only a few studies exist in the veterinary literature comparing diagnostic agreement between clinical and pathological examination, and they focus on companion animals (Kent et al., 2004, Dank et al., 2012). Diagnostic agreement in causes of disease and injury across veterinary facilities handling wildlife may be influenced by the methods employed to make these diagnoses (e.g. clinical versus pathological examination) as well as the specific disease and injury identified. Thus, the estimation of agreement between clinical and pathology-derived diagnoses becomes essential to understand the correct causes of wildlife morbidity and mortality, as well as providing guidance to improve and evaluate the effectiveness of a surveillance program. To achieve this, our primary interest with this study was to develop a statistical tool to assess quality control of a passive surveillance system by estimating the probability of agreement for different clinical and autopsy-derived diagnoses of disease and injury, and by comparing probability of agreement between pairs of wildlife and pathology veterinary facilities. We utilised hospital and autopsy data collected on koalas, a marsupial declining sharply in parts of eastern Australia (Rhodes et al., 2015).

4.2 Materials and Methods 4.2.1 Diagnosis of disease and injury of koalas in South-East Queensland 4.2.1.1 Submission of koalas to wildlife hospitals and clinical diagnosis

Members of the public are the main stakeholder, who usually find sick, injured or orphaned koalas and then contact wildlife carer groups, the RSPCA or directly the Queensland Department of Environment and Heritage Protection (DEHP), the state government entity responsible for wildlife. Koalas are the collected by wildlife officers or the koala rescue ambulance operated by DEHP (https://www.ehp.qld.gov.au/wildlife/koalas/care/koala- ambulance.html) and transported to the DEHP Moggill Koala Hospital, the main veterinary facility for koalas in SEQLD. Two other private veterinary facilities also accept admission of injured or sick koalas and provide specialised veterinary care, namely Currumbin Wildlife Hospital (Currumbin, QLD) and the Australia Zoo Wildlife Hospital (Beerwah, QLD). Staff of all three wildlife hospitals is comprised by wildlife veterinarians, and veterinary technicians.

Koalas receive an external clinical exam to determine the cause of admission and to generate a presumptive diagnosis (es). First-time koala admissions receive DEHP’s (e.g. microchip) or hospital (e.g. ear tag) identification. A DEHP koala record sheet is generated for every koala admission, listing information on date and time of submission, demographics, clinical diagnoses, hospital outcome (euthanised, dead on arrival, dying whilst on treatment, or released). Ancillary laboratory testing aids might be used in predicting clinical outcome, and establishing a final diagnosis. Basic laboratory testing for all koalas includes a blood smear to perform leucocyte counts, a packed cell volume and total protein. If there is suspicion of a metabolic or infectious disease, clinical laboratory testing may include complete blood counts (CBC), a biochemistry panel, and urinanalysis. A CBC tells the status of the main blood cell lines (white and red), and hydration. The (basic) biochemistry panel informs liver and kidney function. Urinalysis might be attempted for koalas with signs of urinary disease. Koalas of both sexes (particularly those with urinary disease) might receive an ultrasound by conducting a bilateral dorsal hair clipping to remove the dense fur to assess the status of the urogenital tract. Any sign of chlamydial bursitis in the female reproductive tract will render that female a candidate for euthanasia due to irreversible infertility (Canfield et al., 1983). Depending on disease stage, males can be treated; however, euthanasia will be elected if the prognosis is poor (Johnston et al., 2015).

Sick koalas with treatable conditions are hospitalised in treatment wards and orphaned koalas are raised by foster carers. Koalas with poor prognosis such as severe loss of body condition from chronic disease or injuries are humanely euthanized and an autopsy will be performed to determine the cause/extent of lesions.

Koalas found dead are submitted following the same procedure and a presumptive diagnosis is recorded in the koala record sheet. Data in the record sheets are entered in KoalaBASE – a digital database for koala admissions to wildlife hospitals. KoalaBASE is an online SQL database developed by the University of Queensland with standardised fields for koala data entry and reporting (http://www.ava.com.au/node/27181). KoalaBASE contains fields for single or multiple diagnoses made by wildlife clinicians on each koala admitted. These diagnoses were used as the clinical diagnoses provided by wildlife hospitals in the current research study.

4.2.2 Submission of koalas for pathological examination and autopsy-derived diagnosis

The School of Veterinary Science, The University of Queensland (UQ) is the reference centre for pathological examination of koalas in QLD. Usually only cases of dead koalas with unusual manifestations of disease are submitted to UQ. However, for the purpose of this research project, wildlife hospitals were advised to route all koala carcasses to UQ from April 2013 through to May 2016. Autopsies were conducted following standard procedures (Appendix 26 and Appendix 27). Although occasional autopsies were conducted in freshly dead bodies (within 3 hours since death), most carcasses were frozen before the autopsy was conducted. Koalas were identified during autopsy from ear tags, microchips or accession numbers on accompanying medical records and this information was used to retrieve their clinical histories. A general external examination was conducted to determine sex and age (Gordon, 1991), and record carcass preservation status. Body condition was scored using a non-conventional scoring method from 1 through 10 by palpating the amount of muscle mass above the scapular bone. Tissue samples collected during autopsy varied according to findings, and were preserved in 10% neutral buffered formalin for histological processing following routine procedures. In a case-by-case basis, ancillary testing was done and consisted of microbiological culture and PCR testing. A final diagnosis was reached by taking into consideration the gross autopsy examination and histological findings and/or ancillary testing results if available. A standardised method of recording single or multiple final diagnoses for each koala based was on pathology data. 4.2.3 Statistical analysis

The quantitative assessment of comparing clinical and autopsy-derived diagnoses of koala diseases and injuries focused on estimating the probability of agreement between diagnoses. Thereby we wanted to estimate 1) the probability of agreement for individual koala diseases and injuries, 2) the probability of agreement for individual koala diseases and injuries by pairs of veterinary facilities (UQ-University of Queensland; WH1, WH2, WH3 –Wildlife Hospital 1 to 3). Thus we developed a Bayesian mixed model to estimate 1) the probability of agreement for individual koala diseases and injuries and 2) the probability that pairs of veterinary facilities obtained the same type of diagnosis for a koala disease or injury.

We considered three random effects in the analysis: 1) unmeasured koala-specific characteristics that may have affected the diagnosis outcome, 2) unmeasured factors that may have affected the diagnosis outcome for a koala when seen by different veterinary facilities and 3) diagnosis effects. 91

The following notations were used:

• i = koala

• j = veterinary facility

• k = type of diagnosis

Yijk• = binary outcome; if Yijk = 1 then the i-th koala has been diagnosed with k by veterinary facility j, or 0 otherwise.

• Ui = unmeasured koala-specific characteristics that may have affected the diagnosis

outcome

Vij• = unmeasured factors that may have affected the diagnosis outcome for koala i when seen by veterinary facility j (this is to account for the fact that the same koala may have been presented in different conditions to the different veterinary facility)

Zk = diagnosis effect (this is to account for the fact that some diseases or injuries may • be easier to detect than others)

• µj = propensity of veterinary facility j to provide diagnoses.

We assume that conditionally on Ui, Vij and Zk, the Yijk are mutually independent Bernoulli

variables with probability pijk that koala i is diagnosed with a k by veterinary facility j, such

that

−1 Φ (pijk) = µj + Ui + Vij + Zk, (1) where Φ−1(·) is the quantile function of a standard Gaussian variable, also known as probit function.

We also assume that Ui, Vij and Zk are independent zero mean Gaussian variables with variances σ2, τ2and ν2, respectively.

We fit the model in (1) using Bayesian methods with the following set of independent priors.

4 • µj N (0, 10 ), j (UQ, WH1, WH2, WH3)

∼ ∈ 92

• σ2 N (1, 104) left-truncated in 0.

∼ • τ 2 N (1, 104) left-truncated in 0.

∼ • ν2 N (1, 104) left-truncated in 0.

∼ We use a Markov chain Monte Carlo (MCMC) algorithm to fit the model using a data-

augmentation technique based on auxiliary variables (Holmes and Held, 2011).

We define the probability of agreement between veterinary facility j and veterinary

facility j! for a diagnosis k based an individual koala i as

P Ajj’ | ik = P [Yijk = 1, Yij’k= 1] + P [Yijk = 0, Yij’k = 0]. (2)

Our approach allows us to deal with missing data by borrowing strength of information across all observations and provide estimates of probability of agreement also for veterinary facilities who have not examined some of the koalas.

All probabilities are presented with 95% credible intervals.

The statistical analyses were conducted using the software program R.

4.3 Results 4.3.1 Study population

A total of N=267 koalas that were admitted between February 16 2014 and 21 November 2015 to wildlife hospitals were analysed, where a clinical diagnosis was made and then forwarded to UQ for an autopsy-based diagnosis. A total of 88 individual clinical diagnoses were provided across the three wildlife hospitals, with 26, 17, and 45 individual clinical diagnoses provided by WH1, WH2, WH3 respectively for the N=267 koalas admitted. UQ provided 41 individual autopsy-based diagnoses for the N=267 koalas admitted. Of the N=267 koalas autopsied at UQ, N=230 were admitted with chlamydial-related syndromes, N=153 were in poor body condition, N=64 with trauma caused by motor vehicle collisions and N=25 with trauma from animal attacks.

4.3.2 Diagnostic agreement

The mean probability of agreement in diagnosis of koala diseases and injuries between clinical and autopsy-based examination was 0.532. Highest probability of agreement was found for “trauma due to hit by car” (0.789), followed by “trauma due to animal attack” (0.683) and “emaciation/poor body condition” (0.617) while the lowest probability of agreement was identified for ectoparasitosis (mites or ticks, 0.489; Table 4-1). All six pairs of veterinary facilities showed high probabilities of agreement for the diagnoses “trauma due to hit by car” and “trauma due to animal attack” (Appendix 25). We observed that UQ and WH2 are the pair of veterinary facilities with the highest probabilities of agreement, in particular for “trauma due to hit by car”, “trauma due to animal attack”, “chlamydial cystitis-urogenital” and “emaciation-poor body condition”.

Table 4-1. Probabilities of agreement (including the 95% credible intervals) for diagnoses of diseases and injuries in South-East Queensland, Australia (2013-2016).

Probability of agreement Diagnostic Category (95% credible interval) Infectious Disease Abscessation 0.511 (0.397-0.651) Chlamydia 0.557 (0.441-0.769) Chlamydial conjunctivitis 0.579 (0.46-0.802) Chlamydial cystitis/urogenital 0.612 (0.49-0.835) Chlamydial female reproductive 0.602 (0.48-0.825) Chlamydial male reproductive 0.5 (0.38-0.636) Chlamydial nephritis/pyelonephritis 0.544 (0.426-0.756) Dermatitis 0.532 (0.419-0.723) Ectoparasitosis (mite, tick) 0.489 (0.366-0.616) End stage kidney disease 0.513 (0.395-0.659) Fungal infection 0.521 (0.405-0.669) Hydronephrosis/hydroureter 0.517 (0.402-0.673) Infection 0.546 (0.438-0.717) Lymphadenitis/lymphamegaly 0.517 (0.4-0.657) Parasitosis other 0.53 (0.419-0.706) Pneumonia/bronchopneumonia 0.504 (0.387-0.667) Renal complication 0.494 (0.365-0.626) Septicaemia 0.494 (0.368-0.607) Environmental/Miscellaneous Disease Anaemia 0.515 (0.398-0.679) Blind 0.536 (0.421-0.739) Bone marrow disease 0.507 (0.377-0.621) 94

Diabetes 0.521 (0.408-0.673) Disease 0.504 (0.384-0.624) Facial oedema 0.519 (0.406-0.67) Heat stress 0.524 (0.408-0.677 ) Hypoglycaemia 0.529 (0.418-0.694) Intersex 0.529 (0.416-0.692) Lung non-inflammatory 0.497 (0.372-0.615) Lymphatic - other 0.497 (0.372-0.611) Mammary gland mass 0.527 (0.412-0.686) Membrane injury 0.514 (0.395-0.635) Mouth - non-inflammatory 0.574 (0.461-0.769) Nephrosis 0.508 (0.393-0.675) Omentum fibrosis 0.519 (0.403-0.667) Organ aplasia/hypoplasia 0.573 (0.463-0.774) Peritonitis 0.516 (0.4-0.656) Phlebitis 0.508 (0.392-0.647) Prolapse 0.526 (0.411-0.686) Pulmonary fibrosis 0.511 (0.389-0.666) Snake bite 0.543 (0.433-0.717) Urinary male/female non-chlamydia 0.511 (0.394-0.671) Degenerative Disease Cardiac/pericardial disease 0.502 (0.379-0.638) Dermatosis 0.501 (0.381-0.628) Emaciation/poor body condition 0.617 (0.494-0.839) Musculoskeletal disease-other 0.526 (0.413-0.715) Senescence 0.535 (0.421-0.711) Putative AIDS-like syndrome Gastrointestinal - inflammatory 0.51 (0.394-0.675) Gastrointestinal - noninflammatory 0.546 (0.433-0.747) Mouth - inflammatory 0.509 (0.393-0.677) Suspected retrovirus 0.529 (0.409-0.699) Benign/Malignant Neoplasms Koala fibromatosis syndrome 0.505 (0.382-0.633) Neoplasia craniofacial tumour 0.515 (0.397-0.653) Neoplasia leukaemia 0.503 (0.392-0.624) Neoplasia lymphoma 0.545 (0.433-0.739) Neoplasia other 0.501 (0.383-0.661) Trauma Trauma animal attack 0.683 (0.548-0.887) Trauma hit by car 0.789 (0.652-0.945) Trauma other 0.511 (0.397-0.636) Trauma tree fall 0.528 (0.416-0.71) Trauma unknown 0.519 (0.405-0.71)

The propensity of veterinary facilities and the random effects are displayed in Table 4-2. The unmeasured diagnosis effect v2 corresponds to most of the variation observed, accounting 95 for about 80% of the total random effects variation (0.158 / (0.018+0.017+0.158) =81.8%). This highlights that some diseases or injuries may be easier to detect than others by different veterinary facilities.

Table 4-2. Posterior summaries for the model parameters: posterior mean with 95% credible intervals and standard deviation (SD)

Model Posterior mean (95% SD parameters credible interval) µWH1 -0.357 (-0.554, -0.163) 0.099 µWH2 0.922 (0.447, 1.420) 0.246 µWH3 0.031 (-0.143, 0.202) 0.087 µUQ 0.702 (0.53, 0.875) 0.088 σ2 0.018 (0.001, 0.057) 0.015 τ 2 0.017 (0.000, 0.061) 0.017 ν2 0.158 (0.082, 0.275) 0.049

4.4 Discussion The overarching goal of this study was to present a mechanism to improve passive surveillance in koalas, a declining wildlife species. Good surveillance systems depend on many factors to function properly. In the present study, we determined and applied two main aspects that will assist with the perpetuity of koala disease surveillance: standardisation of recording terminology or ‘case definitions’, and mechanisms to assess the quality of a surveillance system.

The development of a standardised method to record disease and injury proved to be a useful tool allowing the estimation of probabilities of agreement for diagnoses across veterinary facilities. Standardisation of data entry had been incorporated into KoalaBASE, which reduces data input errors through the use of drop-down menus, facilitating standardised data entry amongst institutions and subsequent use and collation of data for quantitative analysis. We used standardised clinical diagnoses data compiled into KoalaBASE and compared it with standardised autopsy-based data.

4.4.1 Agreement between clinical and autopsy-derived diagnoses

We used a unique method to assess the quality of data recording on koala disease and injury from clinical and pathological examinations, thus we estimated the likelihood of different observers in reaching the same diagnostic conclusion. In veterinary and medical

96 science, the Kappa coefficient is usually used to evaluate of how well two observers or raters agree, either between observers using the same methodology (e.g. radiology) or different approaches (e.g. rate of clinical diagnostic errors in human medicine corroborated by autopsies). The Kappa Statistic is a popular test utilised to rule out random judgement by at least one observer and is applicable in any situation where two or more observers are evaluating the same item. Kappa returns a single value rating based on the degree to which each of the observers ratings randomly coincided by pure chance. The Kappa value is then interpreted on a nominal scale as poor, slight, fair, moderate, substantial, or almost perfect depending on the value (0 through to 0.99)(Viera and Garrett, 2005). Other tests utilised in the human medical literature include Pearson’s Chi-square test (Pakis et al., 2010, Kotovicz et al., 2008), or Fisher exact tests (Olsen et al., 2014, Schaub et al., 2016) when comparing categorical variables, the Cochran-Armitage test (Thurnheer et al., 2009, Sonderegger-Iseli et al., 2000), the Shapiro-Wilk’s test (Dank et al., 2012), the Barnard’s unconditional test of superiority (Kent et al., 2004) and logistic regression (Kotovicz et al., 2008).

Using a Bayesian mixed model approach we were able to produce a meaningful quantitate, rather than a dichotomous assessment of agreement which has a number of additional benefits 1) we were able to consider unmeasured variation in form of random effects and 2) we were able to estimate probabilities which are easier to communicate than an artificial Kappa coefficients interpreted on a nominal scale. Because of these reasons, probabilities of agreement proved to be more advantageous and allowed the communication of results in a more efficient and practical manner. In addition, a Kappa statistic is restricted to a single coefficient limiting the understanding of how other factors may influence the outcome of interest, and does not provide information regarding the reliability of the assessment.

Probability of agreement was averaged by diagnosis. This revealed a high probability of agreement for diagnoses ‘trauma motor vehicle’, ‘trauma animal attack’, ‘emaciation/poor body condition’, ‘chlamydial cystitis’, and ‘chlamydia female reproductive’; all common diagnoses for koalas. Animal attacks, mainly inflicted in the study by domestic dogs have characteristic markings of puncture wounds and lesions in the neck, ribs, and torso (Henning et al., 2015, Wong et al., 1999). Trauma caused by motor vehicles often affects the upper body and results in open or closed comminuted fractures of long bones, skull and mandible (Henning et al., 2015, Canfield, 1991). No other cause of trauma would lead to lesions with such characteristics so these two causes of trauma are morphologically diagnosed with 97 accuracy, explaining the high probability of agreement across all six observers. In regards to chlamydial disease, cystitis is the most frequent clinical presentation, and has an obvious staining of the fur surrounding the common vestibule during chronic disease (Canfield et al., 1989) an observable sign even in the distance. A common secondary finding of cystitis is reproductive disease in females koalas, so in SEQLD hospitals, female koalas with urinary disease are triaged to determine existence of infertility due to chlamydiosis, known as bursitis. Typically, bursitis is confirmed by palpation or abdominal ultrasound, which is commonly used in specialist koala hospitals. Compromise of the reproductive tract would render affected females candidates for euthanasia as they are potential sources of infection and are permanently infertile, thus euthanasia is applied as a population management strategy. Lastly, a high probability of agreement existed for diagnosing low body condition (emaciation/poor body condition). Low body condition would reflect either serious or chronic disease with a poor prognosis, and would indicate a poor prognosis for rehabilitation due to the low tolerance of koalas to hospitalisation, their lack of metabolic reserve as they have minimal fat stores and possibly their highly-specialised, nutrient-depleted diet. The commonality of all the diagnoses with high probability of agreement were that these aetiologies had overt, significant and generally externally visible highly distinctive morphologic features. They also represented some of the most common and significant diseases and threats affecting koalas. The one exception being ovarian bursitis in females, but the high correlation in this disease would stem from the specific monitoring utilised in hospitals who target affected animals for euthanasia, combined with the pathognomonic and overt pathology seen at autopsy for this disease.

A low probability of agreement was found for diagnosing ‘chlamydia male reproductive’, ‘ectoparasitosis’, ‘lung fibrosis’, and ‘renal complications’. Chlamydial infections in the male reproductive tract have been historically underreported in Australia mainly due to shadowing from the heavy focus on female infection and sequelae, and lack of sonographic changes in the male unless there is severe disease (Loader, 2010). Ectoparasitosis, vastly represented by hard ticks (Mesostigmata: Ixodida) at low loads, was rarely included in koalas clinical records because it is rendered as an incidental finding, and its quantification is further complicated at the clinicians and pathologist end by post-engorgement detachment (Stone and Carrick, 1996). Both lung fibrosis or any previous resolved insult to the pulmonary parenchyma, and renal complications (renal fibrosis or atrophy, renal infarcts), if subclinical, could only have been detected via laparoscopy (not routinely performed in wildlife) or 98 autopsy. Auscultation or blood test would help in detecting the few severe conditions affecting either body system. These diseases with the lowest probabilities of agreement were either incidental, subclinical, or showed little external visible manifestations.

The human medical literature reports low agreement between clinical and autopsy diagnoses for aged individuals (25%-40% (Lee, 1994)), and for circulatory disorders (Pinto- Carvalho et al., 2008). Low agreement with anaemia or conditions in which circulation would be compromised such as septicaemias were also observed in the present study. Depending on the stage, septicaemias in koalas can manifest grossly as severe multifocal haemorrhages in multiple internal sites (Dickens, 1978) to being a peracute illness only manifesting as sudden death (Blanshard and Bodley, 2008), which was not obvious in the N=4 cases routed for autopsy, N=2 of which arrived eviscerated from previous autopsies. Grossly, anaemia would be an unspecific finding observed as generalised pale mucous membranes (Loader, 2010). Both of these conditions were infrequently reported by hospitals but unsupported by morphological findings during koala autopsies, possibly masked by the post-mortem movement of blood during freezing and thawing, and/or poor carcass preservation mainly from evisceration or previous partial autopsies. The low detection of blood disorders at necropsy likely reflects the fact that standard clinical pathology assessments like CBC and biochemistry cannot be performed post-mortem, reducing the sensitivity of detection after death.

Variation between veterinary facilities in agreement on diagnoses was mainly due to the unmeasured diagnosis effect, accounting for the fact that some diseases or injuries may be easier to detect than others. In general, koalas suffer from a few but well-described diseases and traumatic threats that lead to specific clinical signs and gross lesions during clinical or pathological examination. Wildlife clinicians and pathologists would have been accustomed to koalas suffering from extreme debilitation from chronic cystitis or females with reproductive sequelae particularly because chlamydiosis is often a chronic condition where re-infections and recrudescence are common resulting in outward clinical signs (Griffith et al., 2013). In addition, the high frequency of koala admissions displaying skull fractures with a marked seasonal pattern (Dique, Thompson, Preece, Penfold et al., 2003) would point wildlife clinicians and pathologist to motor vehicle collisions. Conversely, without an autopsy and histopathology performed, it would be difficult to determine resolved renal or pulmonary

99 injuries represented by fibrosis, unless the extent of previous damage to the organ was a cause for hospital submission e.g. laboured breathing.

Trauma by motor vehicle and by animal attacks had the highest agreement across all facilities. Nonetheless, UQ-WH2 had the highest probabilities of agreements for most diagnoses except for septicaemias, non-inflammatory diseases of the lung, leukaemias, other diseases of the lymphatic system (lympholysis, red pulp atrophy), and ectoparasitosis and other diagnoses of ambiguous terminology e.g. disease or bone marrow disease. The high level of agreement found between these two facilities may be due to the fact that WH2 was the only institution conducting partial autopsies prior to routing each koala carcass to UQ, increasing their sensitivity to disease detection. The routing of eviscerated bodies with partial autopsies was considered a factor negatively influencing diagnostic agreement due to its effect in body preservation. Nonetheless, the koala cases submitted for autopsies by this institutions suffered severe, chronic diseases, as well as being the institution contributing with the lowest submissions, lessening their chances of variance.

WH1-WH3 had evidence of a high probability of agreement in diagnosing koalas with renal complications (renal fibrosis, renal infarct, renal atrophy). It is possible that this result reflected the combination of blood testing and ancillary testing capabilities (e.g. clinical imaging) of WH1 as well as the communication of results for koalas with additional clinical findings to WH3 from patient referral.

The pairs UQ-WH2, as well as UQ-WH3 and WH2-WH3 had high probabilities of agreement in diagnosing koalas with low body conditions. Discrepancies in the result may reflect the variation in the scale utilised at hospitals. General koala husbandry guidelines recommend a 1-5 scale (Jackson et al., 2003, Department of Environment and Heritage Protection, 2004); nonetheless, slight modifications to this scale are done by hospitals using a 9 or 10 point system instead, the latter utilised by both UQ and WH3. The majority of veterinary pairs assessed, namely UQ-WH2, UQ-WH3 and WH2-WH3 had a high level of agreement in diagnosing chlamydial cystitis and chlamydial reproductive disease in females, the most common manifestations of chlamydiosis in koalas. Lastly, UQ-WH2 and WH2-WH3 had high probability of agreement for chlamydial conjunctivitis, closely followed by UQ-WH3. A general high frequency of agreement was expected for koala chlamydiosis across all facilities assessed based on the ample experience of wildlife clinicians and pathologists working with Australian native species.

100

4.4.2 Clinical versus autopsy-derived diagnoses: opportunities to improve passive surveillance systems

Many factors drive diagnostic discrepancies between clinicians and pathologists, including diagnostic testing availability, funding, and experience. The estimation of pathological and clinical diagnostic agreement is rarely performed in veterinary medicine compared to human medicine. In veterinary medicine, only two studies looking at the rate of clinical versus pathological agreement were found in the literature, a low number despite the significant body of information produced in recent decades in veterinary pathology (Kent et al., 2004, Dank et al., 2012). The standardisation method utilised by these studies to measure diagnostic agreement was a broad pathophysiologic categorisation (degenerative, anatomic, metabolic, neoplastic, infectious, inflammatory, toxic, and traumatic). Our diagnostic categorisation method was more specific (60 diagnostic categories) increasing the opportunities for variance and thus of discrepancy. In addition, our analysis provided a more specific and useful interpretation for each diagnosis.

The discrepancy rate between clinical and pathological diagnosis in the human medical literature is improving, despite the considerable decline in non-forensic human autopsies overtime, from 30-40% during the 1960s to <6% in 1994 (Shojania et al., 2003, Sonderegger-Iseli et al., 2000). This may reflect the marked improvement in ante-mortem diagnostic techniques, particularly medical imaging, available to the medical profession. The veterinary studies report a similar pattern: decreased autopsy rates were documented in a veterinary teaching hospital between 1989 (59%) and 1999 (48%). Nonetheless the differences reported between the clinical (37%) and pathological (39.8%) diagnoses in veterinary assessments did not show a significant variation in a study conducted by Kent et al., (2004). A follow-up study was conducted, determining an improvement in the rate and accuracy of clinical diagnoses in 2009 compared to 1989 and 1999, as well as a significant decrease in the number of necropsies performed in 2009 (21.4%) compared to 1999 (48.4%) and 1989 (58.9%) (Dank et al., 2012). A limiting factor is that both studies were derived from teaching institutions where autopsy costs are subsidised improving submission rates, compared to the number of those carried out at veterinary practices.

Reasons for declining interest in autopsies can be attributed to different factors: decrease in clinical interest but increased effort placed in ante-mortem diagnoses from utilisation of diagnostic tools, malpractice fear, and costs (Shojania and Burton, 2008). Clinicians argue that the cases selected for autopsies are those of the highest diagnostic uncertainty, 101 affecting interpretation (Shojania and Burton, 2008). Technological advancement has brought novel laboratory testing methods and therapeutic approaches to conditions; however no tests are perfect (Thurnheer et al., 2009). This likely reflects, the complexity of veterinary medicine with multiple species, less resource for detailed biomedical research, significant cost restrictions on clients and institutions limiting ante-mortem work ups, and a reduction in the availability of sophisticated ante-mortem diagnostic modalities such as advanced medical imaging.

Autopsies of wildlife are considered a gold standard method in diagnosing disease because ante-mortem condition of the animal is usually not available. Despite the challenges posed with wildlife for autopsies (e.g. preservation issues), pathological methods are often the only tool available to determine cause of death. Autopsies should be carried out whenever possible to help improve knowledge for clinical practice, confirm or generate diagnoses, or detect reporting novel conditions. Although autopsies cannot be performed in every case or ensure 100% accuracy, studies report that there is considerable improvement of diagnostic capability in institutions conducting autopsies. For example, a study in the human medicine field determined that autopsies disclosed additional important lesions in 14% of cases (Lee, 1994). Autopsies are considered confirmatory for the cause of death, however, it is important to understand the potential limitations in generating accurate diagnoses in wildlife. The quality of data collected for this study was reliant on state of carcass preservation. Ideally, the approach for data collection could have been strengthened by fresh autopsies conducted on intact carcasses. Thus, the collection of samples from fresh or unfrozen wildlife carcasses is recommended if logistically feasible. However, committing to conduct autopsies and analyse samples in every deceased wild animal carries the use of multiple resources for any institution handling wildlife.

The approach of measuring diagnostic agreement utilising probabilities should be applied in general for veterinary examinations where multiple rates or multiple methods are used. This serves as quality control mechanism and is in particular important in surveillance procedures. We aimed at determining if disease and injury causes could be accurately obtained via clinical examination alone and used the koala as an example. Notably, diseases with easily externally visible, classic or pathognomonic features showed a strong trend for accurate diagnosis and agreement between pathologists and clinicians. This study demonstrated that although clinical diagnoses are accurate for well-described diseases in koalas, autopsies are necessary to detect specific diseases and comorbidities which are of common occurrence in this species. Consequently, the application of autopsies is needed 102 to corroborate clinical diagnoses and identify those obscure lesions that may be missed during clinical examination, also to broaden our understanding of the spectrum of diseases that can pose a threat to koalas, an iconic faunal species in Australia.

103

Chapter 5 Discussion

5.1 Main goals driving doctoral thesis and Chapters

The main goal of this doctoral thesis was to determine the major causes of morbidity and mortality in SEQLD koala populations because the relative contribution and interaction of threats causing koala population decline had not been quantitatively assessed on a large scale to date. This thesis addressed these gaps by analysing data on archived Queensland government koala hospitals records for over 15 years. The resulting retrospective epidemiological analysis on the cause-specific morbidity-mortality drivers from 1997-2013 is shown in Chapter 2. This epidemiological study resulted in the exploration of the largest passive surveillance dataset on koala submissions to wildlife hospitals and the most comprehensive analysis to date on major multifactorial causes of admission and associated risk factors in SEQLD. The analysis performed on this dataset allowed an examination of the SEQLD koala population over a critical and significant period of population decline. Analysing hospital records offers a unique opportunity to establish baseline disease levels and provide insights into causes of injury and disease across space and time. Other studies utilising hospital records have had considerably smaller sample sizes, focused on single diseases, or did not modelled risk factors in detail. Because recording terminology utilised at wildlife hospitals was not standardised, our understanding of the true causes of morbidity and mortality playing part in the current decline of SEQLD koalas was limited. In this study, the standardisation of terminology into specific pre-defined data categories resulted in a more organised, less error-prone methodology for data entry and future analyses. Some studies in the past analysed geographical risk or ‘hotspots’ for koala mortality in SEQLD (Preece, no year), but were restricted to a single location such as the Koala Coast, one of the most significant koala populations in SEQLD and Australia. However, the present study identified spatial-temporal clusters for causes of disease and trauma in koalas over a broader geographical area and over one and half decades. This unique approach using a retrospective view into the population allowed the determination of complex multifactorial drivers related to the koala decline in SEQLD. The identification of these factors is paramount to tailor future management strategies.

Despite the specialised care koalas have received for decades, autopsy techniques and recording nomenclature had not been consistent or standardised, affecting the interrogation of comorbidities and application of effective interventions to tackle threats. Comorbidities are important in koalas due to their ability to lead to loss of body condition. Based on hospital 104 data, the retrospective study in Chapter 2 interrogated causes of morbidity and mortality, but was limited in sensitivity to identify comorbidity and combination of diseases by just using clinical examinations. This gap was addressed by the large-scale autopsy survey to detect comorbidities described in Chapter 3. The systematic autopsies along with case-based ancillary testing conducted were necessary to allow an accurate and specific detection of a high rate of comorbidities in the population sampled. Chapter 3 concluded that the interaction of diseases is a significant threat for SEQLD koala population and a major cause of debilitation for mature koalas likely impairing their ability to breed and survive.

The work generated in Chapter 3 gave us the opportunity to address significant gaps in the koala pathology literature, where the main body of information is restricted to a few common diseases. Despite the iconic status of the koala, no autopsy guidelines were found for the species. In certain cases, domestic species autopsy guidelines can be extrapolated to wildlife. However, unfamiliarity with certain koala anatomical features can lead to misinterpretation when conducting an autopsy. To address this, a koala field autopsy guide (excerpts in Appendix 26 and Appendix 27) was developed for stakeholders as an additional product of the work conducted during this doctoral thesis. The pathological assessments conducted in Chapter 3 permitted the collection of a significant amount of photographic material from classical, and novel koala conditions. A koala pathology catalogue (Appendix 16-24) was designed and included in this thesis containing reference gross and histopathology images that will be shared with stakeholders and other interested parties. These two documents complement each other and can be utilised in future workshops when upskilling hospital personnel in basic post-mortem techniques. Conducting an in-depth pathology diagnostic investigation in every deceased animal is unusual in wildlife, particularly on a declining species compared to the usual experience in veterinary medicine, where typically cases of the most diagnostic uncertainty are chosen for autopsy. This work resulted in the identification of novel koala diseases (e.g. koala fibromatosis or koala primary serosal myxosarcoma (Gonzalez Astudillo et al., 2015), as well as undocumented host- parasite relationships (data not shown; (Spratt and Beveridge, 2016)).

The different methods utilised by wildlife clinicians and pathologists when making a diagnosis do have inherently different sensitivities and specificities. Understanding the differences in the diagnoses generated by clinicians and pathologists and the factors influencing diagnostic accuracy is essential to control the quality and sensitivity of 105 surveillance systems. Chapter 4 presents a study in which a statistical approach to measure agreement between assessors or observers was developed. Data from koala diagnoses was compared to evaluate the probabilities of agreement in diagnosing causes of diseases and injury based on clinical and pathological examination. A high probability of agreement was determined between diagnoses made through clinical and pathology examinations for frequently-diagnosed conditions in koalas in SEQLD, namely common chlamydial syndromes and trauma and their interaction, or comorbidities. In contrast, a low probability was found for resolved tissue injuries and subclinical parasitoses. This study also highlighted that autopsies were a necessary method to detect comorbidities which frequently affect koalas. By measuring diagnostic agreement, this study highlighted the importance of systematic autopsies in identifying an unusual previously undescribed disease such as a koala fibromatosis case that was misdiagnosed as a malignant neoplasm by one hospital, and not detected by another. Autopsies also provide answers in cases of diagnostic uncertainty such as a case of unexplained chronic ill-thrift in a koala with albinism which further pathological investigation is pointing to a novel disease for the species (data not shown). Due to the different factors influencing wildlife medicine, there may be discrepancies in the accuracy of diagnoses generated by clinicians and pathologists. Other factors possibly influencing variation in agreement include financial limitations of clinical workups in wildlife for certain indications, or plausibly the presence of incidental or subclinical diseases that were not clinically obvious for workups or worthy of investigation. The present study determined that in order to secure the perpetuity of a diagnostic investigation program in koalas or wildlife, clear case definitions and the establishment of quality control mechanisms need to be applied. In addition, this study concluded that clinical examination in SEQLD wildlife hospitals is precise for common koala conditions displaying classical signs that are visibly externally or that their diagnosis is feasible with medical imaging equipment. Nonetheless, autopsies provided enhanced sensitivity to specific diagnoses such as resolved organ injuries or subtle morphologic lesions, novel or subclinical diseases and comorbidities.

5.2 Epidemiology of Koala Morbidity and Mortality in QLD

The spatial-temporal distribution of morbidity and mortality locations demonstrated that the majority of admissions derived from the most densely-populated councils. This suggests greater chances of detection in urbanised areas by members of the public and a potential 106 spatial bias in koala admission associated with areas that are easily accessible by members of the public.

Studying koala morbidity and mortality requires strong collaborative effort by various institutions (wildlife hospitals, universities, and consulting groups). This multi-institutional approach may result in the inconsistent recording of information, potentially affecting data quality. There were certain logistical and technical limitations experienced in this study. These included the lack of carcass identification, and clinical histories, as well as suboptimal sample preservation from evisceration either from previous autopsies or manner of death. Despite the formation of crystals causing cellular membrane disruption impeding histological assessment, freezing was selected as a carcass preservation method to partially offset the logistic limitations of sample transportation. Additionally, extending carcass preservation was an important factor in the subtropical SEQLD weather (Santos et al., 2011) particularly during periods of high koala mortality to allow sufficient time for cases to be autopsied. Despite the challenges inherent to the collection and preservation of wild koala carcasses prior to routing and autopsy, a high rate of final diagnoses were obtained, indicating that this method was appropriate to detect the great majority of conditions affecting koalas. Thus autopsies are recommended to remain a regular diagnostic investigative method for koalas. Nonetheless, for unusual disease aetiologies, a short (3 hour) death-time to autopsy interval is suggested to improve diagnostic sensitivity and sample quality.

This study focused exclusively on disease and injury resulting in mortality. However, future koala population monitoring must incorporate data on causes of admission, comorbidities and post-release survival to better understand patterns in those able to recover. Follow-up studies would be vital to review efforts and better direct resources, because hand-rearing, hospitalising and releasing koalas is an important welfare and conservation outcome that drives considerable financial and human resources.

There are strengths and weaknesses when choosing passive or active surveillance (Thrusfield, 2007); a full review of each method relative to wildlife research is included in Chapter 1. In general, passive surveillance is preferred to monitor threats in wildlife and was the chosen method for data collection in this thesis. Nonetheless, active surveillance could still be performed for koala disease investigation if a government fund is set to reactively respond to new hotspots. In doing so, limited resources can be accurately targeted to specific locations where prodromes for major koala diseases or major trauma mortality are 107 being detected. This study concludes that the proper functioning of a disease surveillance system for koalas needs the application of pathology methods utilising standardisation of recording terminology for every koala case, and the usage of mechanisms to evaluate quality control, such as the model-approach proposed in this thesis. In doing so, the effectiveness of future intervention strategies for koalas can be monitored at population level to control regional threats such as road eco-passages, immunisation or treatments. Because of the sensitivity in disease detection, autopsies remain the gold standard method to investigate animal disease with passive surveillance. In addition, when carried out in a systematic manner, autopsies allow the identification of novel diseases and offer opportunities to investigate pathogenesis.

5.3 Trauma

Trauma by motor vehicles and animal attacks was identified as a major threat for koalas in QLD, typically resulting in death or in poor prognosis following hospital admission. Trauma causes typical gross morphological lesions, especially those caused by motor vehicles and animal attacks. These classical lesions influenced the high probability of agreement found between clinical and pathological examination for trauma. Additionally, this study determined that animal attacks occurred with significantly more frequency in koalas displaying marked loss of body condition, a pattern of concern for koalas presenting chronic disease inhabiting a rapidly urbanising landscape. Another factor that influenced trauma was demography, displaying a (non-statistically significant) seasonal distribution that may be explained by the biological traits of koalas. Higher number of admissions between August- October period were noted in the retrospective epidemiological study and autopsy survey, which in QLD can be linked to mate search as spring is the peak season for koala breeding (Martin and Handasyde, 1999), post-natal dispersal (Mitchell, 1990a), or forced dispersal by established territorial male koalas.

Half of injured koalas were healthy individuals prematurely removed by contact with anthropogenically-derived trauma sources (cars, dogs), nearly half of which were young or subadult individuals. Younger male koalas have been identified as the primary dispersers within SEQLD koala metapopulations, thus their premature removal impacts gene flow (Tucker et al., 2007) which could lead to the formation of genetically discrete groups. Interruption of gene flow may carry consequences that are difficult to measure and may have to be considered in future management. The epidemiology study detected a declining

108 temporal trend in the proportion of koalas diagnosed with trauma which may be perceived as beneficial. However, koalas colliding with motor vehicles were still admitted to hospitals in high numbers overtime (data not shown) maintaining over >350 koala admissions under this category per annum in 14 of the 17 years of the dataset.

Koala admissions are influenced by factors that need to be taken into account before inferring koala population level conclusions. Despite the increasing body of evidence supporting the population decline in SEQLD koalas since 1990s (Department of Environment and Resource Management, 2010), hospital admissions have remained stable in SEQLD from the continuation of threats. According to the epidemiological study (data not shown) which only includes records of ‘dead on arrival’, ‘euthanized’ and ‘released’ koalas, koala accessions averaged N=1,191 per annum (year of minimum accessions 2005, N=356, year of maximum accessions 2009 N=1,504) from 1997 through to 2013. In KoalaBASE, a total of N=1,234 cases for 2014 and N=912 cases for 2015 were found for the three SEQLD hospitals involved in the study. This demonstrates that despite the koala decline across SEQLD, koala hospital admissions have remained constant. Nonetheless, hospital admissions will inevitably decline once population numbers are low enough to reduce chances of detection.

A major factor influencing koala trauma admissions was found to be location. Previously, a study determined threat hotspots in relation to spatial patterns in the Koala Coast (Preece, no year), detecting a high rate of habitat fragmentation areas already impacted by focal extinctions. The present study detected multiple spatial-temporal hotspots for trauma from data from a larger area of the SEQLD. Although the motor vehicle trauma hotspots identified could be in response to increased submission of injured charismatic fauna to hospitals, it is more likely that higher traffic and road density in highly-populated areas is a more plausible cause. Trauma in koalas reflects the rapid conversion of koala habitat to urbanised zones (Rhodes et al., 2015, Rhodes et al., 2011). Likely, trauma will remain a threat because these risks are capable of infiltrating into koala bushland habitat adjacent to suburban territories.

In koalas, virtually all causes of trauma are derived from anthropogenic sources. Nonetheless, the underlying driver for trauma is the overlapping of koala habitat with areas converted to fit human needs. Future mitigation efforts in place or being considered in SEQLD that would reduce the mortality of healthy koalas must include the reduction of koala habitat loss with the forthcoming urban expansion and protection of habitat remnants in 109

SEQLD, potentially including private land (Lunney et al., 2000). Different interventions have been proposed to prevent koala road-associated mortalities. However, any intervention applied seeking a population-level impact needs to be accompanied by a major commitment from the government to halt the further loss of koala habitat. An evaluation of management options would be needed including the exclusion of koala road passes, providing koala safe crossing, and compulsory traffic calming strategies (Dique, Thompson, Preece, Penfold et al., 2003, Smith, 1990, Ellis et al., 2016) based on detailed spatial analyses in areas where koala populations are most fragmented by road networks. Specifically, if future spatial- temporal analyses could incorporate background population-at-risk data obtained from the hotspots identified, this would aid in determining if these hotspots correspond to real risk for koalas or to confounders (e.g. geographical changes in koala distribution). Mortality caused by dog predation needs to be approached by managing responsible pet ownership and wild dog control. Domestic dog-associated mortality should to be tackled via educational and awareness campaigns for owners, inviting them to restrain dog movements during dawn and dusk, the peak koala activity times. Wild dog-associated mortality is more difficult to manage, and their predation rate has not been clearly determined on koalas. Wild dogs are present in all areas of QLD (Department of Employment, Economic Development and Innovation, 2014) where they may be posing a threat to koalas in bushland habitat (Mifsud, no year). As wild dogs are considered pest animals (Department of Employment, Economic Development and Innovation, 2014), density reduction will benefit koalas and other wildlife. However, any of these interventions to reduce motor vehicle collisions or animal attacks need to be done along a long-term data-capturing system that provides some sort of assessment as to the success of the intervention, as every design may not be suitable for all arboreal mammals (Goldingay and Taylor, 2016). The continuation of the koala hospital-based surveillance system will facilitate a robust data-capturing system to monitor the effectiveness of interventions applied at the population level.

5.4 Chlamydia & loss of body condition

Chlamydiosis or disease consistent with chlamydiosis was identified as the main bacterial disease in wild koalas manifesting as urogenital and ocular syndromes and was a frequent cause of loss of body condition. The relative contribution of chlamydiosis to the koala decline has been questioned (Penn et al., 2000) as previously, Chlamydia spp. was proposed as a commensal organisms in koalas (Carrick, 1996). The present study determined that disease 110 consistent with chlamydiosis was a leading cause of chronic clinical syndromes often resulting in loss of body condition in SEQLD koalas. Although there was no evidence that chlamydiosis resulted in direct mortality in every case, the complications arising from multicentric nature of the infection were evident. Specifically because it was frequent to find koalas with chlamydiosis affected with comorbidities, and these led to debilitation, likely affecting survival. The most common chlamydial syndrome detected in this study was urogenital infection, often in females. Ovarian bursitis and orchitis are frequent sequelae of chlamydia-driven infection resulting in infertility (Girjes et al., 1988, Obendorf, 1981, Weigler et al., 1987) and although it may not be a direct cause of mortality in koalas, it is a reason for euthanasia, due to permanent infertility, according to current QLD hospital koala population management guidelines. These findings are further supported by the high probability of agreement found across wildlife hospitals in SEQLD diagnosing cystitis, female reproductive infection and loss of body condition. The high discrepancy in agreement detected for male reproductive disease demonstrates that most diagnostic effort in koala reproductive infection diagnosis is still focused on females, stressing the need for improved methods of male chlamydial testing. Both the retrospective analysis and autopsy survey demonstrate that sex has remained as a relevant risk factor influencing the outcome of koalas with chlamydial disease at hospitals resulting from hospital management of the disease. A large proportion of females are euthanized at hospitals following triage if found to be infertile. The proportion of male koalas found with prostatic disease in the autopsy survey was higher than the one reported in the retrospective study, despite the vast difference in sample size, revealing historical under diagnosis. This is an impactful finding because it is known that Chlamydia spp. can infect several parts of the male genital tract (Johnston et al., 2015). Despite infertility sequelae of chlamydiosis for both male and female koalas, annual koala birth rates in SEQLD appear not to be impacted (Rhodes et al., 2011). According to the present study, the main disease-associated conservation concern is the chronic chlamydial disease deteriorating koala health leading to mortality from loss of body condition. The overall goal of this study was not to assess total chlamydiosis prevalence in SEQLD due to the study design and sampling methodology. However, the study did determine the impact of chlamydia on morbidity and mortality. To detect chlamydial infection status molecular methods need to be used because there is evidence that subclinical infection may be common in SEQLD koalas (Jackson et al., 1999).

Despite hospital admission numbers remaining relatively constant across the study period, koalas admitted with low body condition from chlamydiosis have shown a steady increase. 111

In fact, the proportion of koalas losing body condition from chronic chlamydial infections increased overtime and became the top diagnosis in the autopsy survey. Northern koalas are reported to manifest more overt disease compared to their southern counterparts. This likelihood of overt disease was linked to the presence of certain chlamydial ompA genotypes (Legione et al., 2016). However, recent studies in wild koalas are confirming a link between chlamydial disease manifestation, neoplasia, and the KoRV (Chaban et al., 2017). The KoRV-disease relationship is inherently complex and is only starting to be unveiled in recent studies (Chaban et al., 2017). Evidence points to KoRV as a determinant in the manifestation of overt chlamydiosis due to an immunosuppressive state induced by exogenous KoRV variants (Lifson, 2014). Interventions being developed for chlamydiosis in koalas are vaccination (Waugh, Khan, Carver et al., 2016), but once developed, its delivery method at the population-level including boosters needs to be resolved. Additionally, more work in the interplay of Chlamydia with viruses like KoRV and other co-infections is required. KoRV infection status was not interrogated in the present study, as it was outside of the scope of the major questions in this doctoral thesis.

5.5 Comorbidities

Although the retrospective epidemiological analysis interrogated combination of disease occurrence in koalas, it was unable to provide accurate comorbidity frequencies because autopsies were not systematically conducted in every koala in the database. Because of this, a major goal of the autopsy survey was to interrogate comorbidities in SEQLD koalas because disease interaction was poorly understood. Autopsies are a relatively low cost but highly sensitive procedure that can be performed by clinicians and trained technical staff. The limiting factors that may hamper an accurate diagnosis are the ability to conduct histopathology or ancillary testing due to lack of funding or access to diagnostic facilities. Other aspects include logistics, carcass preservation status as wild animals are seldom found in optimal condition, and lack of veterinary support (Duncan et al., 2008). In Australia the main constraint to perform diagnostic investigation in koalas is veterinary and laboratory assistance due to limited access to facilities from distant locations, or financial restrictions (Gillett, 2014), resulting in under diagnosis or diagnostic exclusion of certain koala conditions.

The sensitivity of the autopsy methodology was corroborated by the higher proportion of koalas with comorbidities found in the autopsy survey compared to the retrospective study

112 that was based on clinical examinations study, despite the vast difference in sample size. Previous studies have descriptively interrogated comorbidities in koalas (Canfield, 1987c, Weigler et al., 1987, Canfield, 1991, Hirst et al., 1992) reporting either multi-centric or underlying infections in koalas affected by injuries. (Hirst et al., 1992) documented low survival rates in individuals with systemic disease, but only descriptively investigated their association with ocular disease and did not quantify disease interaction in detail. In contrast, the work conducted in this thesis unveiled high comorbidity rates thanks to the standardisation of terminology and sensitivity of systematic autopsies.

A high rate of comorbidities were diagnosed in koalas frequently displaying severe loss of body condition compromising overall health (e.g. chlamydiosis in multiple sites, opportunistic infections or neoplasms) compared to those koalas diagnosed with a single disease, regardless of its severity. Interpreting an increase in the proportion of koalas admitted with low body condition has to be done with an understanding of how koalas were submitted to hospitals. Koalas that were electively euthanased by hospitals were often received due to the extent of lesions from injuries or diseases, frequently in combination with low body condition. Thus, the comorbidity detection rate that may not be representative in severity or chronicity of the disease interaction affecting the source koala population in SEQLD. A small proportion of primarily younger and apparently healthy koalas were found to be losing body condition without an observable cause. It is possible that the loss of body condition in younger koalas is due to malnourishment from consumption of browse from fast-growing but low-nutrient producing Eucalyptus, a secondary effect of increased atmospheric CO2 from climate change (IUCN, 2009). Exogenous KoRV subgroups have also been linked as a putative cause of unexplained loss of body condition comprising the “AIDS-like syndrome” in QLD koalas (Chaban et al., 2017).

A great proportion of koalas with comorbidities were mature individuals. Despite recent ecological modelling indicating annual birth rates are actually high (Rhodes et al., 2011), due to the affected age class, and evidence of KoRV geographical expansion, population dynamics monitoring is needed to ensure SEQLD koala declines are not further exacerbated by reduced birth rates. The impact of comorbidities detected in this study impacts current koala rehabilitation efforts because only 1 of 4 koalas are successfully rehabilitated in SEQLD wildlife hospitals and released (Burton and Tribe, 2016). This stresses the need of conducting future monitoring of post-release survival rates due to the efforts invested in this activity. The clinical and rehabilitation outcome for each koala can improve if diagnoses are 113 made at earlier disease stages so treatment can be more effective. Due to the therapeutic and prophylactic action of vaccination, if immunising koalas against chlamydiosis or KoRV is ever a reality, a greater challenge ahead would be to determine effective vaccine delivery methods. According to the results of this thesis, wildlife clinicians, managers and policy makers will continue facing the challenge of managing a koala population with a concerning proportion of individuals debilitated by comorbidities

5.6 Integrated Conservation Strategy for SEQLD Koalas

Koalas are facing multiple threats that need to be tackled simultaneously and immediately. This situation highlights the need for an integrated koala surveillance strategy in SEQLD. Firstly, a multi-institutional approach in which clear case definitions are utilised for clinical and pathology case recordings must be established. Pathology methods would be a core component of this strategy and for this to be accomplished, necessary upskilling of hospital personnel in autopsy techniques would be needed. Data collated would utilise the same standardisation method to reduce data input errors, and facilitate analyses. When multiple observers are involved in providing diagnoses, quantifying diagnostic agreement is essential. Statistical tools developed in this research can be used beyond clinical and autopsy data and for any measures of agreement relating to diagnoses, such as veterinary diagnostic imaging, enhancing the applicability of quality data collected. The standardised approaches developed for diagnoses and data entry also allow the faster processing and analysis of data and therefore provide almost-real time outputs when being incorporated in a flexible database system with automated reporting functions. Monitoring in real time can be achieved by KoalaBASE, a robust, state-of-the-art tool for passive surveillance in koalas developed by our research team that incorporates standardised data entry to allow the recording of disease and mortality trends. The database offers opportunities to monitor management measures to lower koala mortalities and disease occurrence. Results of the investigation of morbidity and mortality should not only be communicated government and interested parties, but also to members of the public to increase their awareness and provide appreciation of their efforts to report sick and injured koalas.

A wildlife disease surveillance system relies on an exact diagnosis, but obtaining a correct and consistent diagnosis with multiple observers is challenging. Factual knowledge of prevalence patterns in population, experience, intuition and most importantly the use of

114 diagnostic tests (clinical versus pathological) determine the art of identifying the nature of the patient’s disease. To support these efforts, the autopsy guide and pathology reference catalogue gathers visual material of conditions commonly found in koalas which can be shared with stakeholders to assist with the upskilling of technical personnel across institutions assisting koalas in SEQLD. In this thesis a method was developed to measure diagnostic agreement that assessed how well observers agree and which factors influence the outcome of interest to improve the quality of surveillance outcomes in koalas. The results are expressed in probabilities of agreement, making the communication of results more practical compared to that of a single coefficient. The aim of estimating probabilities of diagnostic agreement is not to focus on error rates, but to provide an understanding on how to interpret results and improve disease diagnostic investigation for koalas. Another key component proposed for an integrated conservation strategy is the improvement of rehabilitation rates for koalas as well as an increased recognition of the value of post-release survival studies. The epidemiological analysis reported improvement overtime of koala releases in SEQLD (Gonzalez-Astudillo et al., 2017). Nonetheless, there were no studies found in the literature measuring the impact of the contribution of rehabilitation and release to QLD koala conservation.

The incorporation of education campaigns to raise awareness for members of the public to improve koala admissions would be essential, as well as incorporation of citizen science. For example, technology, such as smart phone applications that allow capturing information on injured (or at-risk koalas) by members of the public. These could produce routing alerts to the nearest wildlife rescuers resulting in faster responses and earlier transfer of koalas to hospitals ultimately improving survival rates and decreasing euthanasia rates. Major causes of disease and injury for SEQLD koalas were identified determining that threats to this population are multifactorial and complex. The interaction of disease and disease and trauma was explored in detail, unveiling a concerning future for koalas in SEQLD. Using field investigations and modelling approaches, risk factors associated with major diseases and injuries highlighted in this research need to be identified in wild koalas to guide the development and testing of appropriate interventions to overcome threats in SEQLD. Various tools and approaches were developed to improve passive surveillance of koalas that can help with the planning, implementation, and monitoring the interventions to tackle threats as well as in making the decisions on disease control and prevention.

115

Despite the documented shortcomings of utilising hospital-derived data, it is conceivable to assume that the identified causes of admission and risk factors are affecting the SEQLD koala populations plausibly at similar rates. Preventing further mortality from disease and trauma is needed to stabilise the declining SEQLD koala population and assist in the retention of viable populations within their natural ranges. Tackling multiple threats at once can be achieved through an integrated conservation strategy for koalas (Rhodes et al., 2011). This approach should aim at two main aspects: firstly, carry out a structured surveillance program with case definitions and quality control mechanisms so a data collection system is maintained to monitor koalas in real time allowing stakeholders respond to threats. Secondly, monitor the effectiveness of interventions applied to koalas at the population level (vaccination, eco-passages or road exclusion structures (Dique, Thompson, Preece, Penfold et al., 2003). Also, the protection of koala habitat ahead of the forthcoming urban densification and expansion, as well as raising community awareness of common koala threats. Habitat protection would not only have a positive effect in reducing trauma, but also in connecting metapopulations ensuring gene flow (Tucker et al., 2007), reducing stress-triggered disease (Lunney et al., 2012) as well as safeguarding koalas from the effects of heatwaves (Gordon et al., 1988) and climate change (IUCN, 2009). The future of SEQLD koalas is dire unless a major commitment is made to protect habitat remnants and to tackle the major threats causing significant population impact such as trauma, diseases and comorbidities.

116

List of References

BACKHOUSE, T. C. & BOLLIGER, A. 1961. Morbidity and mortality in the koala (Phascolarctos cinereus). Australian Journal of Zoology 9, 24-37.

BARKER, C. J., GILLETT, A., POLKINGHORNE, A. & TIMMS, P. 2013. Investigation of the koala (Phascolarctos cinereus) hindgut microbiome via 16S pyrosequencing. Vet Microbiol, 167, 554-64.

BENTON, M. J. 2004. Vertebrate Palaeontology, Wiley-Blackwell.

BERGER, L., SPEARE, R., DASZAK, P., GREEN, D. E., CUNNINGHAM, A. A., GOGGIN, C. L., SLOCOMBE, R., RAGAN, M. A., HYATT, A. D., MCDONALD, K. R., HINES, H. B., LIPS, K. R., MARANTELLI, G. & PARKES, H. 1998. Chytridiomycosis causes amphibian mortality associated with population declines in the rain forests of Australia and Central America. Proc Natl Acad Sci U S A, 95, 9031-6.

BIOLINK. 2005. Conserving koalas in the Coomera–Pimpama koala habitat area: A view to the future. In: Goldcoast City Countil. (ed.). Gold Coast, Queensland, Australia,.

Birdlife Australia. 2016. News. Australian Birdlife. Carlton, VIC: Birdlife Australia.

BLACK, P. F., MURRAY, J. G. & NUNN, M. J. 2008. Managing animal disease risk in Australia: the impact of climate change. Revue scientifique et technique (International Office of Epizootics), 27, 563-80.

BLANSHARD, W. Interesting cases at Long Pine Koala Sanctuary. Proceedings of the Wildlife Disease Association (Australasian Section), 1993 North Stradbroke Island. 31-33.

BLANSHARD, W. Medicine and Husbandry of koalas. In: SYDNEY, U. O., ed. Proceedings #233: Wildlife. The TG Hungerford Refresher Course for Veterinarians. Postgraduate Foundation in Veterinary Science. , 1994. 547-626.

BLANSHARD, W. & BODLEY, K. 2008. Koalas. In: VOGELNEST, L. & WOODS, R. (eds.) Medicine of Australian Mammals. CSIRO Publishing.

BODETTI, T. J., JACOBSON, E., WAN, C., HAFNER, L., POSPISCHIL, A., ROSE, K. & TIMMS, P. 2002. Molecular evidence to support the expansion of the hostrange of Chlamydophila pneumoniae to include reptiles as well as humans, horses, koalas and amphibians. Syst Appl Microbiol, 25, 146-52.

BOOTH, R. J. Adrenal weight: Body weight ratios in relation to cause of death in the koala (Phascolarctos cinereus). In: SCHULTZ, D. J., ed. Wildlife Disease Association Meeting (Australasian Section), 1987 Flinders Ranges, South Australia. Wildlife Disease Association, 8.

BOOTH, R. J., CARRICK, F. N. & ADDISON, P. A. 1990. The structure of the koala adrenal gland and the morphological changes associated with the stress of disease. In: LEE, K. A., HANDASYDE, K. & SANSON, G. D. (eds.) Biology of the Koala. Sydney: Surrey Beatty & Sons. 117

BRISCOE, N. J., KROCKENBERGER, A., HANDASYDE, K. & KEARNEY, M. R. 2014. Bergmann meets Scholander: geographical variation in body size and insulation in the koala is related to climate. Journal of Biogeography, 42, 791–802.

BROWN, A. S., GIRJES, A. A., LAVIN, M. F., TIMMS, P. & WOOLCOCK, J. B. 1987. Chlamydial disease in koalas. Australian Veterinary Journal, 64, 346-50.

BROWN, A. S. & GRICE, R. G. 1984. Isolation of Chlamydia psittaci from koalas (Phascolarctos cinereus). Australian Veterinary Journal, 61, 413.

BUNN, C. & WOODS, R. 2005. Emerging wildlife diseases - impact on trade, human health and the environment. Microbiology Australia, 26, 53-55.

BURACH, F., POSPISCHIL, A., HANGER, J., LOADER, J., PILLONEL, T., GREUB, G. & BOREL, N. 2014. Chlamydiaceae and Chlamydia-like organisms in the koala (Phascolarctos cinereus)--organ distribution and histopathological findings. Veterinary Microbiology, 172, 230-40.

BURSAC, Z., GAUSS, C. H., WILLIAMS, D. K. & HOSMER, D. W. 2008. Purposeful selection of variables in logistic regression. Source Code for Biology and Medicine, 3, 17.

BURTON, E. & TRIBE, A. 2016. The Rescue and Rehabilitation of Koalas (Phascolarctos cinereus) in Southeast Queensland. Animals (Basel), 6.

BUTLER, R. Patterns in koala mortality. In: BERGIN, T. J., ed. The Koala, Proceedings of the Taronga Symposium on Koala Biology, Management and Medicine. , 1978 Sydney. The Zoological Parks Board of NSW, 174-175.

CANFIELD, P. J., OXENFORD, C. J., LOVE, D. N. & DICKENS, R. K. 1983. Pyometra and pyovagina. Australian Veterinary Journal, 60, 337-338.

CANFIELD, P. J., OXENFORD, C. J., LOMAS, G. R. & DICKENS, R. K. 1986. A disease outbreak involving pneumonia in captive koalas. Australian Veterinary Journal, 63, 312-3.

CANFIELD, P. 1987a. A mortality survey of free range koalas from the north coast of New South Wales. Australian Veterinary Journal, 64, 326-328.

CANFIELD, P. 1987b. A study of koala deaths. Australian Science Magazine. Darling Downs Institute Press.

CANFIELD, P. J., PERRY, R., BROWN, A. S. & MCKENZIE, R. A. 1987c. Cranio-facial tumours of mixed cartilage and bone in koalas (Phascolarctos cinereus). Australian Veterinary Journal, 64, 20-22.

CANFIELD, P. J. 1989. A survey of urinary tract disease in New South Wales koalas. Australian Veterinay Journal, 66, 103-6.

CANFIELD, P. J., GEE, D. R. & WIGNEY, D. I. 1989. Urianalysis in captive koala. Australia Veterinary Journal, 66, 376-377.

118

CANFIELD, P. J., HARTLEY, W. J. & DUBEY, J. P. 1990. Lesions of Toxoplasmosis in Australian Marsupials. Journal of Comparative Pathology, 103, 159-167.

CANFIELD, P. J. 1991. A survey of koala road kills in New South Wales. Journal of Wildlife Diseases, 27, 657-60.

CANFIELD, P. J. & HARTLEY, W. J. 1991. Tyzzer's disease (Bacillus piliformis) in Australian marsupials. Journal of Comparative Pathology, 105, 167-73.

CANFIELD, P. J., SPENCER, A. J., HARTLEY, W. J., SPIELMAN, D., VOGELNEST, L. & HULST, F. 1992. Disorders of Keratinization in a Group of Related, Captive Koalas (Phascolarctos cinereus), with a Review of Other Skin Conditions in Koalas. Journal of Zoo and Wildlife Medicine, 23, 414-421.

CANFIELD, P. J. & CUNNINGHAM, A. A. 1993. Disease and mortality in australasian marsupials held at london-zoo, 1872-1972. Journal of Zoo Wildlife Medicine, 24, 158-167.

CANFIELD, P. J. & SPENCER, A. J. 1993. Renal complications of cystitis in koalas. Australian Veterinary Journal, 70, 310-311.

CARRICK, F. 1996. A review of disease in koalas. In: GORDON, G., ed. Koalas - Research for management: Proceedings of the Brisbane Koala Symposium, Brisbane, Queensland. World Koala Research Inc., 37-41.

CASTRO, A. S., OLIVEIRA, A. & LOPES, V. 2013. Pulmonary phaeohyphomycosis: a challenge to the clinician. Eur Respir Rev, 22, 187-8.

CHABAN, B., ONG, V. A., HANGER, J. & TIMMS, P. 2017. Molecular dynamics and mode of transmission of Koala Retrovirus (KoRV) as it invades and spreads through a wild Queensland koala population. Journal of Virology.

CHAPPELL, K. J., BREALEY, J. C., AMARILLA, A. A., WATTERSON, D., HULSE, L., PALMIERI, C., JOHNSTON, S., HOLMES, E. C., MEERS, J. & YOUNG, P. R. 2017. Phylogenetic Diversity of Koala Retrovirus within a Wild Koala Population. Journal of Virology, 91, 1-11.

COCKRAM, F. A. & JACKSON, A. R. 1981. Keratoconjunctivitis of the koala, Phascolarctos cinereus, caused by Chlamydia psittaci. Journal of Wildlife Diseases, 17, 497-504.

COLDHAM, T., ROSE, K., O'ROURKE, J., NEILAN, B. A., DALTON, H., LEE, A. & MITCHELL, H. 2013. Detection of Helicobacter species in the gastrointestinal tract of ringtail possum and koala: Possible influence of diet, on the gut microbiota. Veterinary Microbiology, 166, 429-437.

CONNOLLY, J. 1999. Emerging diseases of koalas and their medical management. Proceedings #327: Wildlife in Australia: Healthcare and management, Postgraduate Foundation in Veterinary Science, The University of Sydney. Dubbo, NSW. 15-40.

119

CONNOLLY, J. H., KROCKENBERGER, M. B., MALIK, R., CANFIELD, P. J., WIGNEY, D. I. & MUIR, D. B. 1999. Asymptomatic carriage of Cryptococcus neoformans in the nasal cavity of the koala (Phascolarctos cinereus). Medical Mycology, 37, 331- 8.

COOPER, J. E. 2002. Diagnostic pathology of selected diseases in wildlife. Revue Scientifique Et Technique De L’Office International Des Epizooties, 21, 77-89.

Brisbane City Council. 2017a. Habitat Brisbane [Online]. Brisbane. Available: https://www.brisbane.qld.gov.au/environment-waste/natural- environment/biodiversity-brisbane/habitat-brisbane. [Accessed 05 February 2017].

Redlands City Council. 2017b. Koala [Online]. Redland City Council. Available: https://www.redland.qld.gov.au/info/20254/wildlife_in_the_redlands/623/koala [Accessed 24 March 2017].

COX-WITTON, K., REISS, A., WOODS, R., GRILLO, V., BAKER, R. T., BLYDE, D. J., BOARDMAN, W., CUTTER, S., LACASSE, C., MCCRACKEN, H., PYNE, M., SMITH, I., VITALI, S., VOGELNEST, L., WEDD, D., PHILLIPS, M., BUNN, C. & POST, L. 2014. Emerging infectious diseases in free-ranging wildlife-Australian zoo based wildlife hospitals contribute to national surveillance. PLoS One, 9, e95127.

CUNNINGHAM, K. A. & BEAGLEY, K. W. 2008. Male genital tract chlamydial infection: implications for pathology and infertility. Biology of Reproduction, 79, 180-9.

CURRUMBIN. 2017. Currumbin Wildlife Sanctuary [Online]. Currumbin, QLD. Available: https://currumbinsanctuary.com.au/ [Accessed 16 February 2018].

DANK, G, SEGEV, G., MOSHE, D., & KENT, M.S. 2012. Follow‐up study comparing necropsy rates and discrepancies between clinical and pathologic diagnoses at a veterinary teaching hospital: 2009 versus 1989 and 1999. Journal of Small Animal Practice, 53: 679-683.

DASZAK, P., CUNNINGHAM, A. A. & HYATT, A. D. 2000. Emerging infectious diseases of wildlife - Threats to biodiversity and human health. Science, 287, 443-449.

DAVIES, N. A., GRAMOTNEV, G., MCALPINE, C., SEABROOK, L., BAXTER, G., LUNNEY, D., RHODES, J. R. & BRADLEY, A. 2013. Physiological Stress in Koala Populations near the Arid Edge of Their Distribution. PLoS ONE, 8, e79136.

DE OLIVEIRA, S. M., MURRAY, P. J., DE VILLIERS, D. L. & BAXTER, G. S. 2014. Ecology and movement of urban koalas adjacent to linear infrastructure in coastal south-east Queensland. Australian Mammalogy, 36, 45-54.

Department of Employment, Economic Development and Innovation (DEEDI), 2014. Wild Dog Management Strategy 2011–16 [Online]. Queensland Government, Biosecurity Queensland. Brisbane, Australia. Available: https://www.daf.qld.gov.au/__data/assets/pdf_file/0016/62431/Wild-dog- strategy-2011-16.pdf [Accessed January 10, 2018].

120

Department of Environment and Heritage Protection (DEHP), 2015a. Koalas [Online]. Queensland Government. Brisbane, Australia. Available: http://www.ehp.qld.gov.au/wildlife/koalas/index.html [Accessed February 14, 2015].

Department of Environment and Heritage Protection (DEHP). 2017. Koala threats [Online]. Queensland Government. Brisbane, Australia. Available: https://www.ehp.qld.gov.au/wildlife/koalas/koala- threats.html#koalas_and_swimming_pools. [Accessed September 12, 2017].

Department of Environment and Heritage Protection (DEHP). 2016. Living with wildlife: Koalas. Queensland Government. Brisbane, Australia. Available: https://ehp.qld.gov.au/wildlife/koalas/pdf/koalas-and-cars.pdf . [Accessed October 7, 2017].

Department of Environment and Heritage Protection (DEHP), 2015b. SEQ koala survey data. [Online]. Queensland Government. Brisbane, Australia. Available: https://data.qld.gov.au/dataset/seq-koala-survey-data-2010-2015 . [Accessed December 15, 2017].

Department of Environment and Heritage Protection (DEHP). 2013. Koala Threats [Online]. Queensland Government. Brisbane, Australia. Available: https://www.ehp.qld.gov.au/wildlife/koalas/koala-threats.html. [Accessed March 12, 2016].

Department of Environment and Energy (DEE) , 2013. Koala populations in Queensland, New South Wales and the Australian Capital Territory and national environment law [Online]. Australian Government. Canberra. Available: http://www.environment.gov.au/biodiversity/threatened/publications/factsheet- koala-populations-queensland-nsw-act-national [Accessed June 5 2017]

Department of Environment and Resource Management (DERM), 2010. Koala Coast: Koala Population Report 2010 [Online]. Queensland Government. Brisbane, Australia. Available: http://www.derm.qld.gov.au/wildlife- ecosystems/wildlife/koalas/research/pdf/koala-coast-report2010.pdf. [Accessed December 15, 2017]

Department of the Environment. 2014. EPBC Act referral guidelines for the vulnerable koala (combined populations of Queensland, New South Wales and the Australian Capital Territory) [Online]. Commonwealth of Australia. Available: http://www.environment.gov.au/system/files/resources/dc2ae592-ff25-4e2c- ada3-843e4dea1dae/files/koala-referral-guidelines.pdf. [Accessed October 13, 2017].

Department of Transportation and Main Roads. 2010. Fauna Management [Online]. Brisbane: Queensland Government. Available: http://www.tmr.qld.gov.au/Community-and-environment/Environmental- management/Land/Fauna-management.aspx [Accessed April 2, 2017 2017].

DENNER, J. & YOUNG, P. R. 2013. Koala retroviruses: characterization and impact on the life of koalas. Retrovirology, 10, 108. 121

DEVEREAUX, L. N., POLKINGHORNE, A., MEIJER, A. & TIMMS, P. 2003. Molecular evidence for novel chlamydial infections in the koala (Phascolarctos cinereus). Systematic and Applied Microbiology, 26, 245-53.

DEXTER, C. E., APPLEBY, R. G., EDGAR, J. P., SCOTT, J. & JONES, D. N. 2016. Using complementary remote detection methods for retrofitted eco-passages: a case study for monitoring individual koalas in south-east Queensland. Wildlife Research, 43, 369-379.

DICKENS, R. K. 1975. The koala (Phascolarctos cinereus): past, present and future. Australian Veterinary Journal, 51, 459-463.

DIQUE, D., PREECE, H. & DE VILLLIERS, D. 2003. Koalas in Pine Rivers Shire: distribution, abundance and management. Brisbane: Queensland Parks and Wildlife Service.

DIQUE, D., THOMPSON, J., PREECE, H., DE VILLIERS, D. & CARRICK, F. N. 2003. Dispersal patterns in a regional koala population in south-east Queensland. Wildlife Research, 30, 281-290.

DIQUE, D., THOMPSON, J., PREECE, H., PENFOLD, G. C., DE VILLIERS, D. & LESLIE, R. S. 2003. Koala mortality on roads in south-east Queensland: the koala speed-zone trial. Wildlife Research, 30, 419-426.

DOOHOO, L., MARTIN, W. & STRYHN, H. 2009. Veterinary Epidemiologic Research, Prince Edward Island, Canada, VER Inc.

DUBEY, J. P., HEDSTROM, O., MACHADO, C. R. & OSBORN, K. G. 1991. Disseminated Toxoplasmosis in a Captive Koala (Phascolarctos cinereus). Journal of Zoo and Wildlife Medicine, 22, 348-350.

DUDANIEC, R. Y., RHODES, J. R., WORTHINGTON-WILMER, J., LYONS, M., LEE, K. E., MCALPINE, C. & CARRICK, F. N. 2013. Using multilevel models to identify drivers of landscape-genetic structure among management areas. Molecular Ecology, 22, 3752-3765.

DUNCAN, C., BACKUS, L., LYNN, T., POWERS, B. & SALMAN, M. 2008. Passive, opportunistic wildlife disease surveillance in the Rocky Mountain Region, USA. Transboundary Emerging Diseases, 55, 308-14.

DWASS, M. 1957. Modified Randomization Tests for Nonparametric Hypotheses. The Annals of Mathematical Statistics, 28, 181-187.

ELLIS, B., MELZER, A., CARRICK, F. & HASEGAWA, M. 2002a. Tree use, diet and home range of the koala (Phascolarctos cinereus) at Blair Athol, central Queensland. Wildlife Research, 29, 303-311.

ELLIS, W. A., HALE, P. T. & CARRICK, F. 2002b. Breeding dynamics of koalas in open woodlands. Wildlife Research, 29, 19-25.

122

ELLIS, W., ATTARD, R., JOHNSTON, S., THEILEMAN, P., MCKINNON, A. & BOOTH, D. 2014. Faecal particle size and tooth wear of the koala (Phascolarctos cinereus). Australian Mammalogy, 36, 90.

ELLIS, W. A., FITZGIBBON, S. I., BARTH, B. J., NIEHAUS, A. C., DAVID, G. K., TAYLOR, B. D., MATSUSHIGE, H., MELZER, A., BERCOVITCH, F. B., CARRICK, F., JONES, D. N., DEXTER, C., GILLETT, A., PREDAVEC, M., LUNNEY, D. & WILSON, R. S. 2016. Daylight saving time can decrease the frequency of wildlife-vehicle collisions. Biology Letters, 12.

ELLIS, W. A. H., GIRJES, A. A. & CARRICK, F. N. 1993. Chlamydial infection in koalas under relatively little alienation pressure. Australia Veterinary Journal, 70, 427- 428.

Environment Protection and Biodiversity Conservation (EPBC) Act 1999. 1999. Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) [Online]. Department of the Environment and Energy. Australian Government, Canberra. Available: http://www.environment.gov.au/epbc. [Accessed April 4, 2017].

EVERETT, K. D., BUSH, R. M. & ANDERSEN, A. 1999. Emended description of the order Chlamydiales, proposal of Parachlamydiaceae fam. nov. and Simkaniaceae fam. nov., each containing one monotypic genus, revised taxonomy of the family Chlamydiaceae, including a new genus and five new species, and standards for the identification of organisms. International Journal of Systematic and Evolutionary Microbiology, 2, 415-440.

FABIJAN, J., WOOLFORD, L., LATHE, S., SIMMONS, G., HEMMATZADEH, F., TROTT, D. J. & SPEIGHT, N. 2017. Lymphoma, Koala Retrovirus Infection and Reproductive Chlamydiosis in a Koala (Phascolarctos cinereus). Journal of Comparative Pathology, 157, 188-192.

FOWLER, E. V., HOULDEN, B. A., HOEBEN, P. & TIMMS, P. 2000. Genetic diversity and gene flow among southeastern Queensland koalas (Phascolarctos cinereus). Molecular Ecology, 9, 155-164.

FUNNELL, O., JOHNSON, L., WOOLFORD, L., BOARDMAN, W., POLKINGHORNE, A. & MCLELLAND, D. 2013. Conjunctivitis associated with Chlamydia pecorum in three koalas (Phascolarctos cinereus) in the Mount Lofty Ranges, South Australia. Journal of Wildlife Diseases, 49, 1066-9.

GERHOLD, R. W., TATE, C. M., GIBBS, S. E., MEAD, D. G., ALLISON, A. B. & FISCHER, J. R. 2007. Necropsy findings and arbovirus surveillance in mourning doves from the southeastern United States. Journal of Wildlife Diseases, 43, 129- 35.

GIFFORD, A., FRY, G., HOULDEN, B. A., FLETCHER, T. P. & DEANE, E. M. 2002. Gestational length in the koala, Phascolarctos cinereus. Animal Reproduction Science, 70, 261-6.

123

GILLETT, A. 2014. An Examination of Disease in Captive Australian Koalas (Phascolarctos cinereus) and Potential Links to Koala Retrovirus (KoRV). In: PYE, G. W., JOHNSON, R. N. & GREENWOOD, A. D. (eds.) Technical Reports of the Australian Museum, Online. Sydney: Australian Museum.

GIRJES, A. A., HUGALL, A. F., TIMMS, P. & LAVIN, M. F. 1988. Two distinct forms of Chlamydia psittaci associated with disease and infertility in Phascolarctos cinereus (koala). Infection and Immunity, 56, 1897-900.

GLASSICK, T. V., GIFFARD, P. & TIMMS, P. 1996. Outer membrane protein 2 gene sequences indicate that two chlamydial species, Chlamydia pecorum and Chlamydia pneumoniae cause infections in koalas. Systemic Applied Microbiology, 19, 457-464.

GOLDINGAY, R. L. & TAYLOR, B. D. 2016. Targeted field testing of wildlife road- crossing structures: koalas and canopy rope-bridges. Australian Mammalogy, 39, 100-104.

GONZALEZ-ASTUDILLO, V., ALLAVENA, R., MCKINNON, A., LARKIN, R. & HENNING, J. 2017. Decline causes of Koalas in South East Queensland, Australia: a 17-year retrospective study of mortality and morbidity. Scientific Reports, 7, 42587.

GONZALEZ-ASTUDILLO, V., HERNANDEZ, S. M., YABSLEY, M. J., MEAD, D. G., KEEL, K. M., MUNK, B. A., FISCHER, J. R., RUDER, M. G., BROWN, J. D., PETERS, V. E. & NEMETH, N. M. 2016. Mortality of Selected Avian Orders Submitted to a Wildlife Diagnostic Laboratory (Southeastern Cooperative Wildlife Disease Study, USA): A 36-Year Retrospective Analysis. Journal of Wildlife Diseases, 52, 441-458.

GONZALEZ ASTUDILLO, V., SCHAFFER-WHITE, A., ALLAVENA, R. & PALMIERI, C. 2015. Multiple Intra-abdominal Serosal Myxosarcomas in Two Koalas (Phascolarctos cinereus). Journal of Comparative Pathology, 152, 283-286.

GORDON, G. 1991. Estimation of the age of the Koala, Phascolarctos cinereus (Marsupialia: Phascolarctidae) from tooth wear and growth. Australian Mammalogy, 14, 5-12.

GORDON, G., BROWN, A. S. & PULSFORD, T. 1988. A koala (Phascolarctos cinereus Goldfuss) population crash during drought and heatwave conditions in south‐ western Queensland. Australian Journal of Ecology, 13, 451-461.

GORDON, G., MCGREEVY, D. G. & LAWRIE, B. C. 1990. Koala populations in Queensland: major limiting factors. In: LEE, A. K., HANDASYDE, K. A. & SANSON, G. D. (eds.) Biology of the Koala. Sydney: Surrey Beatty & Sons.

GORDON, G., HRDINA, F. & PATTERNSON, R. 2006. Decline in the distribution of the Koala Phascolarctos cinereus in Queensland. Australian Zoologist, 33, 345-358.

GRIFFITH, J. E., DHAND, N. K., KROCKENBERG, M. B. & HIGGINS, D. P. 2013. A retrospective study of admission trends of koalas to a rehabilitation facility over 30 years. Journal of Wildlife Diseases, 49, 18-28. 124

GRIFFITH, J. E. & HIGGINS, D. 2012. Diagnosis, treatment and outcomes for koala chlamydiosis at a rehabilitation facility (1995-2005). Australia Veterinary Journal, 90, 457-463.

GROGAN, L. F., BERGER, L., ROSE, K., GRILLO, V., CASHINS, S. D. & SKERRATT, L. F. 2014. Surveillance for Emerging Biodiversity Diseases of Wildlife. Plos Pathogens, 10.

HAGEN, F., KHAYHAN, K., THEELEN, B., KOLECKA, A., POLACHECK, I., SIONOV, E., FALK, R., PARNMEN, S., LUMBSCH, H.T., & BOEKHOUT, T. 2015. Recognition of seven species in the Cryptococcus gattii/Cryptococcus neoformans species complex. Fungal Genetics and Biology, 78: 16-48.

HANGER, J. LOADER, J., WAN, C., BEAGLEY, K.W., TIMMS, P., & POLKINGHORNE, A. 2013. Comparison of antigen detection and quantitative PCR in the detection of chlamydial infection in koalas (Phascolarctos cinereus). The Veterinary Journal, 195: 391-393.

HANGER, J. Cancer and haematological disease in koalas: a clinical and virological update. In: TRIBE, A. & BOOTH, R. J., eds. Proceedings of the Australian Association of Veterinary Conservation Biologists Annual Conference, 2003 Cairns, QLD. 19-30.

HARTLEY, W. J., DUBEY, J. P. & SPIELMAN, D. S. 1990. Fatal Toxoplasmosis in Koalas (Phascolarctos cinereus). Journal of Parasitology, 76, 271-272.

HASSELL, J. M., ZIMMERMAN, D., CRANFIELD, M. R., GILARDI, K., MUDAKIKWA, A., RAMER, J., NYIRAKARAGIRE, E. & LOWENSTINE, L. J. 2017. Morbidity and mortality in infant mountain gorillas (Gorilla beringei beringei): A 46-year retrospective review. American Journal of Primatology, 79.

HEMSLEY, S. & CANFIELD, P. J. 1997. Histopathological and immunohistochemical investigation of naturally occurring chlamydial conjunctivitis and urogenital inflammation in koalas (Phascolarctos cinereus). Journal of Comparative Pathology, 116, 273-90.

HEMSLEY, S. W., CANFIELD, P. J. & HUSBAND, A. J. 1996. Histological and immunohistological investigation of alimentary tract lymphoid tissue in the koala (Phascolarctos cinereus), brushtail possum (Trichosurus vulpecula) and ringtail possum (Pseudocheirus peregrinus). Journal of Anatomy, 188 ( Pt 2), 279-88.

HENNING, J., HANNON, C., MCKINNON, A., LARKIN, R. & ALLAVENA, R. 2015. The causes and prognoses of different types of fractures in wild koalas submitted to wildlife hospitals. Preventive Veterinary Medicine, 122, 371-8.

HENNING, J., PFEIFFER, D., DAVIES, P., MEERS, J. & MORRIS, R. 2006. Temporal dynamics of rabbit haemorrhagic disease virus infection in a low-density population of wild rabbits (Oryctolagus cuniculus) in New Zealand. Wildlife Research, 33, 293-303.

HIGGINS, D. P., HEMSLEY, S. & CANFIELD, P. J. 2005a. Association of uterine and salpingeal fibrosis with chlamydial hsp60 and hsp10 antigen-specific antibodies 125

in Chlamydia-infected koalas. Clinical and Diagnostic Laboratory Immunology, 12, 632-9.

HIGGINS, D. P., HEMSLEY, S. & CANFIELD, P. J. 2005b. Immuno-histochemical demonstration of the role of chlamydiaceae in renal, uterine and Salpingeal disease of the koala, and demonstration of chlamydiaceae in novel sites. Journal of Compartive Pathology, 133, 164-74.

HIRST, L., BROWN, A. S., KEMPSTER, R., HALL, J. E. & WOOLCOCK, J. B. 1992. Keratitis in free-ranging koalas (Phascolarctos cinereus) on Magnetic Island, Townsville. Journal of Wildlife Diseases, 28, 424-427.

HOQUE, M. A., BURGESS, G. W., GREENHIL, A. R., HEDLEFS, R. & SKERRATT, L. F. 2012. Causes of Morbidity and Mortality of Wild Aquatic Birds at Billabong Sanctuary, Townsville, North Queensland, Australia. Avian Diseases, 56, 249- 256.

HOULDEN, B. A., COSTELLO, B. H., SHARKEY, D., FOWLER, E. V., MELZER, A., ELLIS, W., CARRICK, F., BAVERSTOCK, P. R. & ELPHINSTONE, M. S. 1999. Phylogeographic differentiation in the mitochondrial control region in the koala, Phascolarctos cinereus (Goldfuss 1817). Molecular Ecology, 8, 999-1011.

HUNDLOE, T. & HALMILTON, C. 1997. Koalas and Tourism: An Economic Evaluation, The Australia Institute.

International Union for the Conservation of Nature (IUCN). 2009. Koalas and climate change: Hungry for CO2 cuts. In: JOHNSON, C., DEGABRIEL, J. & MOORE, B. (eds.) IUCN Red List.

JACKSON, M., WHITE, N., GIFFARD, P. & TIMMS, P. 1999. Epizootiology of Chlamydia infections in two free-range koala populations. Veterinary Microbiology, 65, 255- 64.

JACKSON, S., REID, K., SPITTAL, S. D. & ROMER, L. 2003. Koalas. In: JACKSON, S. (ed.) Australian Mammals: Biology and Captive Management. Melbourne: CSIRO Publishing.

JOHNSON, C. N. & ISAAC, J. L. 2009. Body mass and extinction risk in Australian marsupials: The ‘Critical Weight Range’ revisited. Australian Ecology, 34, 35-40.

JOHNSTON, S. D., DEIF, H. H., MCKINNON, A., THEILEMANN, P., GRIFFITH, J. E. & HIGGINS, D. P. 2015. Orchitis and epididimytis in koalas (Phascolarctos cinereus) infected with Chlamydia pecorum. Veterinary Pathology, 1-4.

JOHNSTON, S.D., O’CALLAGHAN, P., NILSSON, K., TZIPORI, G., & CURLEWIS, J.D. 2004. Semen-induced luteal phase and identification of a LH surge in the koala (Phascolarctos cinereus). Reproduction, 128, 629-34.

JOHNSTON, S. D., O'BOYLE, D., FROST, A. J., MCGOWAN, M. R., TRIBE, A. & HIGGINS, D. 1998. Antibiotics for the preservation of koala (Phascolarctos cinereus) semen. Australian Veterianary Journal, 76, 335-8.

126

JURSKIS, V. & POTTER, M. 1997. Koala surveys, ecology and conservation at Eden, State Forests of New South Wales, Sydney, Forest Research and Development Division.

KENT, M.S., LUCROY, M.D., DANK, G., LEHENBAUER, T.W., & MADEWELL, B.R. 2004. Concurrence between clinical and pathologic diagnoses in a veterinary medical teaching hospital: 623 cases (1989 and 1999). Journal of the American Veterinary Medical Association. 224: 403-406.

KITALA, P. M., MCDERMOTT, J. J., KYULE, M. N. & GATHUMA, J. M. 2000. Community-based active surveillance for rabies in Machakos District, Kenya. Preventive Veterinary Medicine, 44, 73-85.

KROCKENBERGER, M. B. Koala Disease - An overview. Sydney: Taronga Registry of Wildlife Health.

KROCKENBERGER, M. B., CANFIELD, P. J., BARNES, J., VOGELNEST, L., CONNOLLY, J., LEY, C. & MALIK, R. 2002. Cryptococcus neoformans var. gattii in the koala (Phascolarctos cinereus): serological evidence for subclinical cryptococcosis. Medical Mycology, 40, 273-82.

KULLDORFF, M., HEFFERNAN, R., HARTMAN, J., ASSUNCAO, R. & MOSTASHARI, F. 2005. A space-time permutation scan statistic for disease outbreak detection. PLoS Med, 2, e59.

KULLDORFF, M. & INFORMATION MANAGEMENT SERVICES, I. 2009. SaTScanTM v8.0: Software for the spatial and space-time scan statistics. [Online]. Available: http://www.satscan.org/ [Accessed].

LADDS, P. 2009a. Mycotic and algal-associated diseases in terrestrial mammals. In: LADDS, P. (ed.) Pathology of Australian Native Wildlife. CSIRO.

LADDS, P. 2009b. Pathology of Australian Native Wildlife, Colligwood, VIC, CSIRO.

LADDS, P. 2009c. Viral Diseases in Terrestrial Mammals. In: LADDS, P. (ed.) Pathology of Australian Native Wildlife. Collingwood, Victoria: CSIRO Publishing.

LADDS, P. W., THOMAS, A. D., SPEARE, R. & BROWN, A. S. 1990. Melioidosis in a koala. Aust Vet J, 67, 304-5.

LAIR, S., MEASURES, L. N. & MARTINEAU, D. 2016. Pathologic Findings and Trends in Mortality in the Beluga (Delphinapterus leucas) Population of the St Lawrence Estuary, Quebec, Canada, From 1983 to 2012. Veterinary Pathology, 53, 22-36.

LANE, C. 2008. Caboolture, Pine Rivers & Redcliffe Councils Report for Koala Habitat Survey and Mapping: Final Report May. Brisbane: GHD Pty Ltd.

LANYON, J. M. & SANSON, G. D. 1986. Koala (Phascolarctos cinereus) dentition and nutrition. II. Implications of toothwear in nutrition. Journal of Zoology, 209, 169- 181.

127

LAU, A., CHEN, S., SORRELL, T., CARTER, D., MALIK, R., MARTIN, P. & HALLIDAY, C. 2007. Development and clinical application of a panfungal PCR assay to detect and identify fungal DNA in tissue specimens. Journal of Clinical Microbiology, 45, 380-5.

LEE, A. & MARTIN, R. 1988. The koala: a natural history., Sydney, New South Wales University Press.

LEE, K. E., SEDDON, J. M., CORLEY, S. W., ELLIS, W. A. H., JOHNSTON, S. D., DE VILLIERS, D. L., PREECE, H. J. & CARRICK, F. N. 2010. Genetic variation and structuring in the threatened koala populations of Southeast Queensland. Conservation Genetics, 11, 2091-2103.

LEGIONE, A. R., AMERY-GALE, J., LYNCH, M., HAYNES, L., GILKERSON, J. R., SANSOM, F. M. & DEVLIN, J. M. 2016. Chlamydia pecorum Infection in Free- ranging Koalas (Phascolarctos cinereus ) on French Island, Victoria, Australia. Journal of Wildlife Diseases, 52, 426-9.

LENGHAUS, C. 1990. Deaths in koalas. Veterinary Pathology Report, 27, 13.

LIFSON, J. D. 2014. Prevention and Treatment of Koala Retrovirus (KoRV) Infection: Lessons from Studies of AIDS Viruses in Nonhuman Primate Models. Technical Reports of the Australian Museum, Online, 24, 65-69.

LILIENFELD, D. E. & STOLLEY, P. D. 1994. Foundations of Epidemiology, New York, Oxford University Press.

Ljungström, L., Franzén, C., Schlaug, M., Elowson, S., & Viidas, U. 1997. Reinfection with Chlamydia pneumoniae may induce isolated and systemic vasculitis in small and large vessels. Scandinavian Journal of Infectious Diseases. Supplementum. 104: 37-40.

LOADER, J. 2010. An Investigation of the Health of Wild Koala Populations in South- East Queensland. Bachelor of Applied Science, The University of Queensland

LOGAN, M. & SANSON, G. D. 2002. The effect of tooth wear on the feeding behaviour of free-ranging koalas (Phascolarctos cinereus Goldfuss). Journal of Zoology, 256, 63-69.

LUCAS, J. 2000. Superficial dermatitis due to Aspergillus in a hand reared koala. Veterinary Pathology Report, 59, 29.

LUNNEY, D., STALENBERG, E., SANTIKA, T., & RHODES, J.R. 2014. Extinction in Eden: identifying the role of climate change in the decline of the koala in south- east NSW. Wildlife Research, 41:22-34.

LUNNEY, D., CROWTHER, M. S., WALLIS, I. R., FOLEY, W. J., LEMON, J. 2012. Koala populations and climate change: a case study on the Liverpool Plains, north-west NSW. In: LUNNEY, D. & HUTCHINS, P. (eds.) Wildlife and Climate Change: Towards Robust Conservation Strategies for Australian Fauna. Mosman, NSW: Royal Zoological Society of NSW.

128

LUNNEY, D., GRESSER, S., O'NEILL, L. E., MATTHEWS, A. & RHODES, J. 2007. The impact of fire and dogs on koalas at Port Stephens, New South Wales, using population viability analysis. Pacific Conservation Biology, 13, 189-201.

LUNNEY, D., MATTHEWS, A., MOON, C. & FERRIER, S. 2000. Incorporating Habitat Mapping into Practical Koala Conservation on Private Lands. Conservation Biology, 14, 669-680.

LUNNEY, D., URQUHART, C. A. & REED, P. 1990. Koala summit: managing koalas in New South Wales., Hurstville, New South Wales National Parks and Wildlife Service.

MAAS, M., GRONE, A., KUIKEN, T., VAN SCHAIK, G., ROEST, H. I. & VAN DER GIESSEN, J. W. 2016. Implementing wildlife disease surveillance in the Netherlands, a One Health approach. Revue Scientifique et Technique, 35, 863- 874.

MACDONALD, A. M., JARDINE, C. M., CAMPBELL, G. D. & NEMETH, N. M. 2016. Mortality and Disease in Wild Turkeys (Meleagris gallopavo silvestris) in Ontario, Canada, from 1992 to 2014: A Retrospective Review. Avian Diseases, 60, 644-8.

MACKIE, J. T., GILLETT, A., PALMIERI, C., FENG, T. & HIGGINS, D. 2016. Pneumonia due to Chlamydia pecorum in a Koala (Phascolarctos cinereus). Journal of Comparative Pathology, 155, 356-360.

MARIK, P. E. & LEVITOV, A. 2010. The "koala stress syndrome" and adrenal responsiveness in the critically ill. Intensive Care Medicine, 36, 1805-1806.

MARTIN, R. W. 1985. Overbrowsing, and decline of a population of the koala, Phascolarctos cinereus, in Victoria. III. Population dynamics. Australian Wildlife Research, 12, 377–385.

MARTIN, R. W. & HANDASYDE, K. 1990. Population dynamics of the koala (Phascolarctos cinereus) in southeastern Australia. In: LEE, K. A., HANDASYDE, K. & SANSON, G. D. (eds.) Biology of the Koala. Sydney: Surrey Beatty & Sons.

MARTIN, R. W. & HANDASYDE, K. 1995. Koala (Phascolarctos cinereus). In: STRAHAN, R. (ed.) The Mammals of Australia. Sydney: Reed Books.

MARTIN, R. W. & HANDASYDE, K. 1999. The koala: natural history, conservation and management. , Sydney, UNSW Press.

MARTINEZ-NEVADO, E., GONZALEZ ALONSO-ALEGRE, E., JIMENEZ-MARTINEZ, A., RODRIGUEZ-ALVARO, A., MARTINEZ DE MERLO, E., GARCIA-GARCIA, J., & GARCIA-REAL, I., 2017. Atypical presentation of Cryptococcus neoformans in a koala (Phascolarctos cinereus): a Magnetic Resonance Imaging and Computed Tomography study. Journal of Wildlife Diseases, 48, 250-4.

MASTERS, P., DUKA, T., BERRIS, S. & MOSS, G. 2004. Koalas on Kangaroo Island: from introduction to pest status in less than a century. Wildlife Research, 31.

129

MAXWELL, S., BURBRIDGE, A. A. & MORRIS, K. 1996. The 1996 action plan for Australian marsupials and monotremes. Project 500. Environment Australia. Canberra, Australian Capital Territory: IUCN/SSC Australasian Marsupial and Monotreme Specialist Group of the World Conservation Union/Species Survival Commission.

MCALPINE, C., CALLAGHAN, J. G., BOWEN, M. E., LUNNEY, D., RHODES, J., MITCHELL, D. L., PULLAR, D. V. & POSSINGHAM, H. P. 2004. Conserving koala (Phascolarctos cinereus) populations in a fragmented landscape (Noosa Queensland): Technical Report.

MCALPINE, C., MELZER, A., LUNNEY, D., FOLEY, B., ADAMS-HOSKING, C., LAWLER, I., WHISSON, D., PHILLIPS, S., KAVANAGH, R., BAXTER, G., GORDON, G., ELLIS, W., DE VILLIERS, D. L., WALTER, M., MOLSHER, R., TODD, C., MOORE, B., MENKHORST, P., CLOSE, R., PHALEN, D. & RHODES, J. 2012. Conserving Koalas in the 21 Century: synthesising the dynamics of Australia's Koala populations., Australia, Terrestrial Ecosystem Research Network.

MCALPINE, C., RHODES, J., CALLAGHAN, J. G., BOWEN, M. E., LUNNEY, D., MITCHELL, D. L., PULLAR, D. V. & POSSINGHAM, H. P. 2006. The importance of forest area and configuration relative to local habitat factors for conserving forest mammals. Biological Conservation, 132, 153-165.

MCALPINE, C., RHODES, J., PETERSON, A., POSSINGHAM, H. P., CALLAGHAN, J., CURRAN, T., MITCHELL, D. R. & LUNNEY, D. 2007. Planning Guidelines for Koala Conservation and Recovery: A guide to best planning practice March 2007, viewed 10 May 2009,.

MCCALLUM, H. 2008. Tasmanian devil facial tumour disease: lessons for conservation biology. Trends in Ecology and Evolution, 23, 631-7.

MCCALLUM, H. 2012. Disease and the dynamics of extinction. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 367, 2828-39.

MCGAVIN, M. D. & ZACHARY, J. F. 2006. General Pathology. In: MCGAVIN, M. D. & ZACHARY, J. F. (eds.) Pathologic Basis of Veterinary Disease. 5th ed. St Louis, Missouri: Mosby Elsevier.

MCINNES, L. M., GILLETT, A., HANGER, J., REID, S. A. & RYAN, U. M. 2011. The potential impact of native Australian trypanosome infections on the health of koalas (Phascolarctos cinereus). Parasitology, 138, 873-883.

MCINNES, L. M., HANGER, J., SIMMONS, G., REID, S. A. & RYAN, U. M. 2011. Novel trypanosome Trypanosoma gilletti sp. (Euglenozoa: Trypanosomatidae) and the extension of the host range of Trypanosoma copemani to include the koala (Phascolarctos cinereus). Parasitology, 138, 59-70.

MCINNES, L. M., GILLETT, A., RYAN, U. M., AUSTEN, J., CAMPBELL, R. S., HANGER, J. & REID, S. A. 2009. Trypanosoma irwini n. sp (Sarcomastigophora:

130

Trypanosomatidae) from the koala (Phascolarctos cinereus). Parasitology, 136, 875-85.

MCKENZIE, R. A. 1981. Observations on diseases of free-living and captive koalas (Phascolarctos cinereus). Australian Veterinary Journal, 57, 243-246.

MCKENZIE, R. A., WOOD, A. D. & BLACKALL, P. J. 1979. Pneumonia associated with Bordetella bronchiseptica in captive koalas. Australian Veterinary Journal, 55, 427-30.

MCORIST, S., JERRETT, I. V., ANDERSON, M. & HAYMAN, J. 1985. Cutaneous and respiratory tract infection with Mycobacterium ulcerans in two koalas (Phascolarctos cinereus). Journal of Wildlife Diseases, 21, 171-3.

MELZER, A. 1995. Aspects of the ecology of the koala, Phascolarctos cinereus (Goldfuss, 1817), in the sub-humid woodlands of central Queensland. Ph.D., The University of Queensland.

MELZER, A., CARRICK, F., MENKHORST, P., LUNNEY, D. & JOHN, B. S. 2000a. Overview, critical assessment, and conservation implications of koala distribution and abundance. Conservation Biology, 14, 619-628.

MELZER, A., CARRICK, F., MENKHORST, P., LUNNEY, D. & ST JOHN, B. 2000b. Overview, critical assessment, and conservation implications of koala distribution and abundance. Conservation Biology, 14, 619-628.

MELZER, A., HODGON, J., ELLIOTT, B. & TUCKER, G. 2003. A note on predation on koalas Phascolarctos cinereus by raptors, including wedge-tailed eagles Aquila audax, in Queensland. The Queensland Naturalist, 41, 38-41.

MENKHORST, P. 2004. Victoria’s koala management strategy [Online]. Department of Sustainability and Environment. Commonwealth of Australia. Melbourne, Australia. Available: http://www.dse.vic.gov.au/CA256F310024B628/0/A133F97A28FF70D6CA2575 62007F85AE/$File/Koala.pdf. [Accessed February 3 2018]

MIFSUD, G. no year. Wild dogs and their impacts on Koala's a very real threat [Online]. Parliament of Australia. Available: https://www.aph.gov.au/Parliamentary_Business/Committees/Senate/Environme nt_and_Communications/Completed_inquiries/2010- 13/koalas/submissionsaccessibilitytrial/sub018 [Accessed 15 March 2018 2018].

MILNER, A. R., WILKS, C. R., SPRATT, D. M. & PRESIDENTE, P. J. 1981. The prevalence of anti-leptospiral agglutinins in sera of wildlife in southeastern Australia. Journal of Wildlife Diseases, 17, 197-202.

MITCHELL, P. 1989. The social organisation of koalas. Ph.D., Monash University.

MITCHELL, P. 1990a. The home ranges and social activity of koalas—A quantitative analysis. In: LEE, A. K., HANDASYDE, K. & SANSON, G. D. (eds.) Biology of the koala. Chipping Norton, New South Wales, Australia: Surrey Beatty and Sons Pty Ltd. 131

MITCHELL, P. 1990b. Social behaviour and communication of koalas. In: LEE, A. K., HANDASYDE, K. & SANSON, G. D. (eds.) Biology of the Koala. Sydney, Australia: Beatty and Sons.

MITCHELL, P. 1991. Diseases of koalas. Victorian Animal Health Newsletter.

MITCHELL, P. J., JERRETT, I. V. & SLEE, K. J. 1984. Skin ulcers caused by Mycobacterium ulcerans in koalas near Bairnsdale, Australia. Pathology, 16, 256- 260.

MITCHELL, P. J., MCORIST, S. & BILNEY, R. 1987. Epidemiology of Mycobacterium ulcerans Infection in Koalas (Phascolarctos cinereus) on Raymond Island, Southeastern Australia. Journal of Wildlife Diseases, 23, 386-390.

MOORE, B. D. & FOLEY, W. J. 2005. Tree use by koalas in a chemically complex landscape. Nature, 435, 488.

MOORE, B. D., LAWLER, I. R., WALLIS, I. R., BEALE, C. M. & FOLEY, W. J. 2010. Palatability mapping: a koala's eye view of spatial variation in habitat quality. Ecology, 91, 3165-76.

MUNDAY, B. 2012. Effects of urban landscape change on the health and mortality of free-ranging koalas (Phascolarctos cinereus): a case study in Southeast Queensland, Australia. Honors Thesis, The University of Queensland.

MUNDAY, B. L. 1978. Marsupial disease. The 1978 JD Stewart Course for Veterinarians. Postgraduate Foundation in Veterinary Science, Sydney, NSW. The University of Sydney, 335-386.

NARAYAN, E. J., WEBSTER, K., NICOLSON, V., MUCCI, A. & HERO, J. M. 2013. Non- invasive evaluation of physiological stress in an iconic Australian marsupial: the Koala (Phascolarctos cinereus). General and Comparative Endocrinology, 187, 39-47.

NARAYAN, E. J. & WILLIAMS, M. 2016. Understanding the dynamics of physiological impacts of environmental stressors on Australian marsupials, focus on the koala (Phascolarctos cinereus). BMC Zoology, 1, 1-13.

NSUBUGA, P., WHITE, M. E., THACKER, S. B., ANDERSON, M. A., BLOUNT, S. B., BROOME, C. V., CHILLER, T. M., ESPITIA, V., IMTIAZ, R., SOSIN, D., STROUP, D. F., TAUXE, R. V., VIJAYARAGHAVAN, M. & TROSTLE, M. 2006. Public Health Surveillance: A Tool for Targeting and Monitoring Interventions. Disease Control Priorities in Developing Countries, 2nd Edition, 997-1015.

NYARI, S., WAUGH, C. A., DONG, J., QUIGLEY, B. L., HANGER, J., LOADER, J., POLKINGHORNE, A. & TIMMS, P. 2017. Epidemiology of chlamydial infection and disease in a free-ranging koala (Phascolarctos cinereus) population. PLoS One, 12, e0190114.

OBENDORF, D. L. 1981. Pathology of the female reproductive tract in the koala, Phascolarctos cinereus (Goldfuss), from Victoria, Australia. Journal of Wildlife Diseases, 17, 587-92. 132

OBENDORF, D. L. 1983. Causes of mortality and mobidity of wild koala, Phascolarctos cinereus (Goldfuss), in Victoria, Australia. Journal of Wildlife Diseases, 19, 123- 131.

OBENDORF, D. L. 1993. Koala urogenital tract disease. Pathology of the reproductive tract. Proceedings of the Australian Society for Veterinary Pathology, Brisbane, Queensland. 61-62.

Organisation Mondiale de la Santé Animale (OIE). 2015. Training manual on surveillance and international reporting of diseases in wild animals. Workshop for OIE National Focal Points for Wildlife. Paris, France: World Organisation for Animal Health.

OLIVEIRA, N. M., SATIJA, H., KOUWENHOVEN, I. A. & EIDEN, M. V. 2007. Changes in viral protein function that accompany retroviral endogenization. Proceedings of the National Academy of Sciences, 104, 17506.

OSAWA, R., BLANSHARD, W. H. & O'CALLAGHAN, P. G. 1992. Microflora of the pouch of the koala (Phascolarctos cinereus). Journal of Wildlife Diseases, 28, 276-80.

OXENFORD, C. J. & CANFIELD, P. J. 1986. Cholecystitis and bronchopneumonia associated with Pseudomonas aeruginosa in a koala. Australian Veterinary Journal, 63, 339.

PALMIERI, C., HULSE, L., PAGLIARANI, S., LARKIN, R., HIGGINS, D.P., BEAGLEY, K., JOHNSTON, S. 2018. Chlamydia pecorum infection in the male reproductive system of koalas (Phascolarctos cinereus). Veterinary Pathology, https://doi.org/10.1177/0300985818806963.

PANNWITZ, G., WOLF, C. & HARDER, T. 2009. Active surveillance for avian influenza virus infection in wild birds by analysis of avian fecal samples from the environment. Journal of Wildlife Diseases, 45, 512-8.

PAYNE, K. L., VITALE, S. D., EDEN, P. A., VAUGHAN, R. J. & KROCKENBERGER, A. Incidence of cryptococcal disease at Perth zoo – an epidemiological investigation. Wildlife Health in a Shrinking World: Ecology, Management and Conservation. Proceedings of the Wildlife Disease Association International Conference, 2005 Cairns, QLD. Wildlife Disease Association, 282.

PENN, A. M., SHERWIN, W. B., GORDON, G., LUNNEY, D., MELZER, A. & LACY, R. C. 2000. Demographic forecasting in koala conservation. Conservation Biology, 14, 629-638.

PHILLIPS, B. 1990. Koalas: the little Australians we’d all hate to lose., Canberra, Australian Capital Territory, Australian Government Publishing Service, Australian National Parks and Wildlife Service.

PHILLIPS, S. S. 2000. Population Trends and the Koala Conservation Debate. Conservation Biology, 14, 650–659.

PICARD-MEYER, E., DUBOURG-SAVAGE, M. J., ARTHUR, L., BARATAUD, M., BECU, D., BRACCO, S., BOREL, C., LARCHER, G., MEME-LAFOND, B., 133

MOINET, M., ROBARDET, E., WASNIEWSKI, M. & CLIQUET, F. 2011. Active surveillance of bat rabies in France: a 5-year study (2004-2009). Veterinary Microbiology, 151, 390-5.

POLKINGHORNE, A., HANGER, J. & TIMMS, P. 2013. Recent advances in understanding the biology, epidemiology and control of chlamydial infections in koalas. Veterinary Microbiology, 165, 214-23.

PRATT, A. 1937. The Call of the Koala, Melbourne, Robertson and Mullens.

PREECE, H. [no year]. Identifying Hotspots for Threats to Koalas Using Spatial Analysis Proceedings of the MODSIM 2007 international congress on modelling and simulation. Modelling and Simulation Society of Australia and New Zealand. Queensland, Australia: Queensland Parks and Wildlife Service. Available: https://www.mssanz.org.au/MODSIM07/papers/21_s46/IdentifyingHotspots_s46 _Preece_.pdf

PREECE, N. D., ABELL, S. E., GROGAN, L. F., WAYNE, A. F., SKERRATT, L. F., VAN OOSTERZEE, P., SHIMA, A. L., DASZAK, P., FIELD, H., REISS, A., BERGER, L., RYMER, T. L., FISHER, D. O., LAWES, M. J., LAURANCE, S. G., MC, C. H., ESSON, C. & EPSTEIN, J. H. 2017. A guide for ecologists: Detecting the role of disease in faunal declines and managing population recovery. Biological Conservation, 2014, 136-146.

Queensland Government (QLD Government). 1992. Nature Conservation Act 1992. Brisbane, Australia. Available: https://www.legislation.qld.gov.au/view/pdf/2013- 11-07/act-1992-020. [Accessed August 20, 2017].

Queensland Government (QLD Government). 2010. South East Queensland Koala State Planning Regulatory Provisions, Brisbane, Australia. Available: http://www.parliament.qld.gov.au/Documents/TableOffice/TabledPapers/2009/53 09T1016.pdf. [Accessed November 21, 2017].

Queensland Government (QLD Government). 2015. Koala mapping. Department of Environment and Heritage Protection. Brisbane, Australia. Available: https://ehp.qld.gov.au/wildlife/koalas/mapping/. [Accessed August 18, 2017].

QUIGLEY, B.L., CARVER, S., HANGER, J., VIDGEN, M.E., & TIMMS, P. 2018. The relative contribution of causal factors in the transision ffrom infection to clinical chlamydial disease. Scientific Reports, 8, 8893.

RELMAN, D. A. 1993. Universal Bacterial 16S rRNA Amplification and Sequencing. In: PERSING, D. H., SMITH, T. F., TENOVER, F. C. & WHITE, T. J. (eds.) Diagnostic Molecular Biology, Principles and Applications. American Society for Microbiology.

RHODES, J., BEYER, H., PREECE, H. & MCALPINE, C. 2015. South East Queensland Koala Population Modelling Study. In: UNIQUEST (ed.). Brisbane.

RHODES, J., NG, C. F., DE VILLIERS, D., PREECE, H., MCALPINE, C. & POSSINGHAM, H. P. 2011. Using integrated population modelling to quantify the

134

implications of multiple threatening processes for a rapidly declining population. Biological Conservation, 144, 1081-1088.

RHODES, J. R., LUNNEY, D., CALLAGHAN, J. & MCALPINE, C. 2014. A Few Large Roads or Many Small Ones? How to Accommodate Growth in Vehicle Numbers to Minimise Impacts on Wildlife. PLoS ONE, 9, e91093.

RHODES, J. R., WIEGAND, T., MCALPINE, C., CALLAGHAN, J., LUNNEY, D., BOWEN, M. & POSSINGHAM, H. 2006. Modeling Species’ Distributions to Improve Conservation in Semiurban Landscapes: Koala Case Study. Conservation Biology, 2, 449-459.

RICHARDS, J., LICKEY, A. & SLEEMAN, J. M. 2005. Decreasing prevalence and seasonal variation of gunshot trauma in raptors admitted to the wildlife center of Virginia: 1993-2002. Journal of Zoo and Wildlife Medicine, 36, 485-8.

RUSSELL, R. E. & FRANSON, J. C. 2014. Causes of Mortality in Eagles Submitted to The National Wildlife Health Center 1975-2013. Wildlife Society Bulletin, 38, 697- 704.

SACK, A., BUTLER, E., COWEN, P. & LEWBART, G. A. 2017. Morbidity and Mortality of Wild Turtles at a North Carolina Wildlife Clinic: A 10-Year Retrospective. Journal of Zoo and Wildlife Medicine, 48, 716-724.

SALMAN, M. D. 2003a. Animal disease surveillance and survey systems : methods and applications / edited by M.D. Salman, Ames, Iowa, Ames, Iowa : Iowa State Press.

SALMAN, M. D. 2003b. Surveillance and monitoring systems for animal health programs and disease surveys. In: Salman, M. D. (ed.) Animal Disease Surveillance and Survey Systems. Ames, Iowa: Iowa State Press.

SANTOS, S. M., CARVALHO, F. & MIRA, A. 2011. How Long Do the Dead Survive on the Road? Carcass Persistence Probability and Implications for Road-Kill Monitoring Surveys (Persistence Probability of Road-Killed Vertebrates). PLoS ONE, 6, e25383.

SCHEELINGS, T. F. 2016. Morbidity and mortality of monotremes admitted to the Australian Wildlife Health Centre, Healesville Sanctuary, Australia, 2000-2014. Australian Veterinary Journal, 94, 121-4.

SCHIPPER, J., CHANSON, J. S., CHIOZZA, F., COX, N. A., HOFFMANN, M., KATARIYA, V., LAMOREUX, J., RODRIGUES, A. S., STUART, S. N., TEMPLE, H. J., BAILLIE, J., BOITANI, L., LACHER, T. E., JR., MITTERMEIER, R. A., SMITH, A. T., ABSOLON, D., AGUIAR, J. M., AMORI, G., BAKKOUR, N., BALDI, R., BERRIDGE, R. J., BIELBY, J., BLACK, P. A., BLANC, J. J., BROOKS, T. M., BURTON, J. A., BUTYNSKI, T. M., CATULLO, G., CHAPMAN, R., COKELISS, Z., COLLEN, B., CONROY, J., COOKE, J. G., DA FONSECA, G. A., DEROCHER, A. E., DUBLIN, H. T., DUCKWORTH, J. W., EMMONS, L., EMSLIE, R. H., FESTA-BIANCHET, M., FOSTER, M., FOSTER, S., GARSHELIS, D. L., GATES, C., GIMENEZ-DIXON, M., GONZALEZ, S., GONZALEZ-MAYA, J. F., GOOD, T. C., HAMMERSON, G., HAMMOND, P. S., HAPPOLD, D., HAPPOLD,

135

M., HARE, J., HARRIS, R. B., HAWKINS, C. E., HAYWOOD, M., HEANEY, L. R., HEDGES, S., HELGEN, K. M., HILTON-TAYLOR, C., HUSSAIN, S. A., ISHII, N., JEFFERSON, T. A., JENKINS, R. K., JOHNSTON, C. H., KEITH, M., KINGDON, J., KNOX, D. H., KOVACS, K. M., LANGHAMMER, P., LEUS, K., LEWISON, R., LICHTENSTEIN, G., LOWRY, L. F., , et al., 2008. The status of the world's land and marine mammals: diversity, threat, and knowledge. Science, 322, 225-30.

SCHMERTMANN, L.J., STALDER, K., HUDSON, D., MARTIN, P., MAKARA, M., MEYER, W., MALIK, R., & KROCKENBERGER, M.B. 2017. Cryptococcosis in the koala (Phascolarctos cinereus): pathogenesis and treatment in the context of two atypical cases. Medical Mycology, 0, 1-10.

SCHOBER, A., PARLATO, R., HUBER, K., KINSCHERF, R., HARTLEBEN, B., HUBER, T. B., SCHUTZ, G. & UNSICKER, K. 2013. Cell loss and autophagy in the extra- adrenal chromaffin organ of Zuckerkandl are regulated by glucocorticoid signalling. Journal of Neuroendocrinology, 25, 34-47.

SEABROOK, L., MCALPINE, C., BAXTER, G., RHODES, J., BRADLEY, A. & LUNNEY, D. 2011. Drought-driven change in wildlife distribution and numbers: a case study of koalas in south west Queensland. Wildlife Research, 38, 509-524.

SELYE, H. 1950. Stress and the general adaptation syndrome. British Medical Journal, 1, 1383-92.

SERGEANT, E. & PERKINS, N. 2015. Epidemiology for Field Veterinarians: An Introduction, Oxfordshire, UK, CABI International.

SHEA, T. J., BECK, C. A., BONDE, R. K., KOCHMAN, H. I. & ODELL, D. K. 1985. An Analysis of Manatee Mortality Patterns in Florida, 1976-81. The Journal of Wildlife Management, 49, 1-11.

SHERWIN, W. B., TIMMS, P., WILCKEN, J. & HOULDEN, B. A. 2000. Analysis and Conservation Implications of Koala Genetics. Conservation Biology, 14, 639-649.

SIMMONS, G. S., YOUNG, P. R., HANGER, J. J., JONES, K., CLARKE, D., MCKEE, J. J. & MEERS, J. 2012. Prevalence of koala retrovirus in geographically diverse populations in Australia. Australian Veterinary Journal, 90, 404-9.

SMITH, P. Urban koalas and local government. In: LUNNEY, D., URQUHART, C. A. & REED, P., eds. Koala Summit - Managing Koalas in New South Wales, 1990 Sydney. National Parks and Wildlife, 130-136.

SPEIGHT, K. N., POLKINGHORNE, A., PENN, R., BOARDMAN, W., TIMMS, P., FRASER, T., JOHNSON, K., FAULL, R., BATE, S. & WOOLFORD, L. 2016. Prevalence and Pathologic Features of Chlamydia Pecorum Infections in South Australian Koalas (Phascolarctos cinereus). Journal of Wildlife Diseases, 52, 301- 6.

SPEIGHT, N., HICKS, P., GRAHAM, C., BOARDMAN, W. & WOOLFORD, L. 2014. Causes of morbidity and mortality in South Australian koalas. Proceedings of the 60th meeting of the Australian Mammal Society., 7-10 July 2014 2014, Melbourne Zoo. 136

SPRATT, D. M. & BEVERIDGE, I. 2016. Helminth parasites of Australasian monotremes and marsupials. Zootaxa, 4123, 1-198.

STALDER, K., ALLEN, G. & KROCKENBERGER, M. 2003. Further investigations into diseases of the koala (Phascolarctos cinereus). Proceedings of Wildlife Disease Assocation (Australasian Section) Annual Conference, Healesville, Victoria, Australia, 68.

STEPHENS, R. S., MYERS, G., EPPINGER, M. & BAVOIL, P. M. 2009. Divergence without difference: phylogenetics and taxonomy of Chlamydia resolved. FEMS Immunology and Medical Microbiology, 55, 115-119.

STEWART, M. 2012. Tasman Brodie. Gum Tips. NSW: Koala Preservation Society of NSW Inc.

STEWART, P. M. 2003. The adrenal cortex. In: LARSEN, P. R., KRONENBERG, H. M., MELMED, S. & POLONSKY, K. S. (eds.) Williams Texbook of Endocrinology. Philadelphia: Saunders.

STONE, B. F. & CARRICK, F. N. Some effects of ticks on koalas, especially tick paralysis. In: GORDON, G., ed. Koalas: Research for Management. Proceedings of the Brisbane Koala Symposium, 22-23 September 1996,Corinda, Queensland: Australia. World Koala Research, 65-69.

STOUT, J. E., KOH, W. J. & YEW, W. W. 2016. Update on pulmonary disease due to non-tuberculous mycobacteria. International Journal of Infectious Diseases, 45, 123-134.

Department of Environment, Water, Heritage and the Arts (DEWHA) 2014. National Koala Conservation and Management Strategy 2009–2014 [Online]. Commonwealth of Australia. Canberra: Commonwealth of Australia. Available: https://www.environment.gov.au/system/files/resources/165139fc-3ab5-4c96- 8b15-d11a1ad882ab/files/koala-strategy.pdf. [Accessed April 4, 2017].

STUDDERT, V. P. 2011. Saunders Comprehensive Veterinary Dictionary E-Book (4), London, London, GB: Elsevier Health Sciences.

TARLINTON, R. E., MEERS, J. & YOUNG, P. R. 2006. Retroviral invasion of the koala genome. Nature, 442, 79-81.

TARLINTON, R. E., YOUNG, P., MEERS, J. & HANGER, J. 2005. Real-time reverse transcriptase PCR for the endogenous koala retrovirus reveals an association between plasma viral load and neoplastic disease in koalas. Journal of General Virology, 86, 783-787.

THOMPSON, J. A. 2006. The comparative ecology and population dynamics of koalas in the koala coast region of south-east Queensland. Ph.D. , The University of Queensland.

THRUSFIELD, M. 2007. Surveillance. Veterinary Epidemiology. 3rd ed.: John Wiley & Sons, Inc.

137

Threatened Species Scientific Committee (TSSC). 2011. Assessment of the sensitivity of estimates of the trend in the national koala population to uncertainty in estimates of the populations at state level [Online]. Commonwealth of Australia. Available: http://www.environment.gov.au/biodiversity/threatened/species/pubs/197- sensitivity-analysis.pdf. [Accessed December 18, 2017].

Threatened Species Scientific Committee (TSSC). 2012. Advice to the Minister for Sustainability, Environment, Water, Population and Communities from the Threatened Species Scientific Committee (the Committee) on Amendment to the list of Threatened Species under the Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act). Commonwealth of Australia. Available: https://www.environment.gov.au/biodiversity/threatened/species/pubs/197- listing-advice.pdf. [Accessed April 12, 2017].

TUBARO, A., MARIANI, S., DE NUNZIO, C. & MIANO, R. 2010. Bladder weight and detrusor thickness as parameters of progression of benign prostatic hyperplasia. Current Opinion in Urology, 20, 37-42.

TUCKER, G., MELZER, A. & ELLIS, W. 2007. The development of habitat selection by subadult koalas. Australian Journal of Zoology, 55, 285-289.

VAINAS, T., STASSEN, F.R., SCHURINK, G.W., TORDOIR, J.H., WELTEN, R.J., VAN DEN AKKER, L.H., KURVERS, H.A., BRUGGEMAN, C.A., & KITSLAAR, P.J. 2005. Secondary prevention of atherosclerosis through Chlamydia pneumoniae eradication (SPACE Trial): a randomised clinical trial in patients with peripheral arterial disease. European Journal of Vascular and Endovascular Surgery. 29: 403-411.

VELA, A. I., COLLINS, M. D., LATRE, M. V., MATEOS, A., MORENO, M. A., HUTSON, R., DOMINGUEZ, L. & FERNANDEZ-GARAYZABAL, J. F. 2003. Psychrobacter pulmonis sp. nov., isolated from the lungs of lambs. International Journal of Systematic and Evolutionary Microbiology, 53, 415-9.

VITOUSEK, P. M., MOONEY, H. A., LUBCHENCO, J. & MELILLO, J. M. 1997. Human domination of Earth's ecosystems. Science, 277, 494.

WALLIS, R. L. 2013. Koalas Phascolarctos cinereus in Framlingham Forest, south-west Victoria: introduction, translocation and the effects of a bushfire. Victorian Naturalist, 130, 37-40.

WAN, C., LOADER, J., HANGER, J., BEAGLEY, K., TIMMS, P. & POLKINGHORNE, A. 2011. Using quantitative polymerase chain reaction to correlate Chlamydia pecorum infectious load with ocular, urinary and reproductive tract disease in the koala (Phascolarctos cinereus). Australian Veterinary Journal, 89, 409-412.

WAUGH, C., GILLETT, A., POLKINGHORNE, A. & TIMMS, P. 2016. Serum Antibody Response to Koala Retrovirus Antigens Varies in Free-Ranging Koalas (Phascolarctos cinereus ) in Australia: Implications for Vaccine Design. Journal of Wildlife Diseases, 52, 422-5.

138

WAUGH, C., KHAN, S. A., CARVER, S., HANGER, J., LOADER, J., POLKINGHORNE, A., BEAGLEY, K. & TIMMS, P. 2016. A Prototype Recombinant-Protein Based Chlamydia pecorum Vaccine Results in Reduced Chlamydial Burden and Less Clinical Disease in Free-Ranging Koalas (Phascolarctos cinereus). PLoS One, 11, e0146934.

WAUGH, C. A., HANGER, J., LOADER, J., KING, A., HOBBS, M., JOHNSON, R. & TIMMS, P. 2017. Infection with koala retrovirus subgroup B (KoRV-B), but not KoRV-A, is associated with chlamydial disease in free-ranging koalas (Phascolarctos cinereus). Scientific Reports, 7, 134.

WEIGLER, B. J., BOOTH, R. J., OSAWA, R. & CARRICK, F. N. 1987. Causes of morbidity and mortality in 75 free-ranging and captive koalas in south east Queensland, Australia. Veterinary Record, 121, 571-2.

WEIGLER, B. J., GIRJES, A. A., WHITE, N., KUNST, N. D., CARRICK, F. & LAVIN, M. F. 1988. Aspects of the epidemiology of Chlamydia psittaci infection in a population of koalas (Phascolarctos cinereus) in southeastern Queensland, Australia. Journal of Wildlife Diseases, 24, 282-291.

WHITE, N. A. 1999. Ecology of the koala (Phascolarctos cinereus) in rural South-East Queensland, Australia. Wildlife Research, 26, 731-744.

WHITE, N. A. & TIMMS, P. 1994. Chlamydia psittaci in a koala (Phascolarctos cinereus) population in south-east Queensland. Wildlife Research, 21, 41-47.

WHITE, N.A. & KUNST, N.D. 1990. Aspects of the ecology of the koala in south-eastern Queensland. In: Biology of the Koala (Eds A. K. Lee, K. A. Handasyde and G. D. Sanson.) pp. 109–116. (Surrey Beatty and Sons: Chipping Norton, NSW)

WIGNEY, D. I., GEE, D. R. & CANFIELD, P. J. 1989. Pyogranulomatous pneumonias due to Nocardia asteroides and Staphylococcus epidermidis in two koalas (Phascolarctos cinereus). Journal of Wildlife Diseases, 25, 592-6.

WOBESER, G. A. 2005. Essentials of Disease in Wild Animals, Arnes, AI, United States, Iowa State University Press.

WOBESER, G. A. 2007. Disease in Wild Animals: Investigation and Management, New York, Springer-Verlag Berlin Heidelberg.

WOINARSKI, J. & BURBIDGE, A. A. 2016. The IUCN Red List of Threatened Species 2016 [Online]. Available: http://www.iucnredlist.org/details/16892/0 [Accessed].

WOOD, A. D. 1978. The diseases of the captive koala. Proceedings of the Taronga Symposium on Koala Biology, Management and Medicine, Sydney, NSW. Zoological Parks Board of NSW/Royal Zoological Society of NSW/NSW National Parks and Wildlife Service, 158-165.

WORLEY, M., RIDEOUT, B., SHIMA, A. & JANSSEN, D. 1993. Opportunistic infections, cancer and haematologic disorders associated with retrovirus infection in the koala. Proceedings of the American Association of Zoo Veterinarians. 181-182.

139

XU, W., STADLER, C. K., GORMAN, K., JENSEN, N., KIM, D., ZHENG, H., TANG, S., SWITZER, W. M., PYE, G. W. & EIDEN, M. V. 2013. An exogenous retrovirus isolated from koalas with malignant nesoplasias in a US zoo. Proceedings of the National Academy of Sciences U S A, 110, 11547-52.

YANAI, T., TERANISHI, M., TAKAOKA, M. & YOSHIKAWA, Y. 2003. Retroperitoneal Fibromatosis in a Cynomolgus Monkey (Macaca fascicularis). Journal of Toxicologic Pathology, 16, 117-121.

140

Appendix 1 Copy of Animal Ethics Approval Certificate ANRFA/SVS/193/13/EHP by UQ Native and Exotic Wildlife and Marine Mammals Animal Ethics Committee.

141

Appendix 2 Copy of Approval of Scientific Purposes Permit WISP 13247813 by the Department of Environment and Heritage Protection.

142

Dead on Arrival Euthanized N koalas Wald Percentage RRR (CI 95%) p Percentage RRR (CI 95%) p Wald test Risk Factor submitted test Year Period 1997-2001 1,548 69.6 1 <0.001 16.3 1 <0.001 2002-2005 1,115 73.4 1.27 (1.00-1.60) 0.050 14.9 1.11 (0.83-1.49) 0.473 2006-2009 1,259 69.7 0.76 (0.62-0.93) 0.008 11.7 0.54 (0.41-0.71) 0.000 2010-2013 1,125 70.0 0.66 (0.54-0.81) <0.001 8.6 0.34 (0.25-0.45) <0.001 Age Class Young 1,466 73.3 1 - 9 1 Adult 3,581 69.5 1.07 (0.91-1.27) 0.392 14.7 1.83 (1.44-2.33) <0.001 RRR – Relative Risk Ratio CI – Confidence Intervals

143

Dead on Arrival Euthanized N koalas Percentage RRR (CI 95%) p Wald test Percentage RRR (CI 95%) p Wald test Risk Factor submitted Year 1997-2001 1,474 20.0 1 <0.001 48.5 1 <0.001 2002-2005 684 13.0 0.52 (0.39-0.69) <0.001 52.0 0.83 (0.66-1.03) 0.096 2006-2009 826 25.4 0.90 (0.71-1.14) 0.394 36.8 0.51 (0.41-0.64) <0.001 2010-2013 746 20.6 0.41 (0.32-0.52) <0.001 20.1 0.14 (0.11-0.18) <0.001 Sex Female 2,068 21.3 1 55.1 1 Male 1,642 18.5 0.31 (0.26-0.38) <0.001 23.3 0.14 (0.11-0.16) <0.001 Age Young 575 24.0 1 25.7 1 Adult 3,155 19.5 1.22 (0.96-1.54) 0.099 44.0 2.67 (2.11-3.39) <0.001

144

Dead on Arrival Euthanized N koalas Percentage RRR (CI 95%) p Wald test Percentage RRR (CI 95%) p Wald test Risk Factor submitted Year 1997-2001 610 32.0 1 <0.001 66.6 1 <0.001 2002-2005 781 24.8 0.60 (0.25-1.45) 0.256 73.5 0.83 (0.35-1.97) 0.674 2006-2009 981 36.2 0.93 (0.40-2.17) 0.861 62.3 0.74 (0.32-1.71) 0.475 2010-2013 1,105 47.8 0.22 (0.11-0.44) <0.001 44.5 0.09 (0.04-0.18) <0.001 Sex Female 1,927 35.1 1 62.9 1 Male 1,543 38.4 0.32 (0.21-0.48) <0.001 56.2 0.23 (0.15-0.34) <0.001 Age Young 203 43.3 1 49.7 1 Adult 3,274 36.2 2.17 (1.17-4.02) 0.014 60.5 3.42 (1.83-6.39) <0.001

145

Dead on Arrival Euthanized N koalas Wald Wald Risk Factor Percentage RRR (CI 95%) p Percentage RRR (CI 95%) p submitted test test Year 1997-2001 546 64.1 1 <0.001 21.4 1 <0.001 2002-2005 329 62.6 0.94 (0.63-1.40) 0.773 22.5 1.03 (0.65-1.64) 0.905 2006-2009 361 56.2 0.42 (0.30-0.59) <0.001 14.1 0.32 (0.21-0.50) <0.001 2010-2013 301 55.8 0.40 (0.28-0.57) <0.001 14.6 0.29 (0.18-0.47) <0.001 Age Young 606 60.2 1 14.7 1 Adult 931 60.4 1.42 (1.09-1.85) 0.01 21.2 2.01 (1.42-2.83) <0.001 Sex Female 689 61.2 1 16.1 1 Male 838 59.4 1.13 (0.87-1.46) 0.376 20.6 1.45 (1.04-2.02) 0.03

146

N Dead on Arrival Euthanized koalas Risk Factor submitted Percentage RRR (CI 95%) p Wald test Percentage RRR (CI 95%) p Wald test Sex Female 466 60.3 1 <0.0001 37.77 1 <0.0001 Male 338 60.06 0.2 (0.09-0.43) <0.0001 30.47 0.16 (0.08-0.36) <0.0001

147

N koalas Dead on Arrival Euthanized Risk Factor submitted Percentage RRR (CI 95%) p Wald test Percentage RRR (CI 95%) p Wald test Year 1997-2001 197 30 1 0.0203 35.6 1 0.0013 2001-2005 109 17.7 1.37 (0.77-2.44) 0.287 23 1.55 (0.73-3.27) 0.254 2005-2009 169 25.1 0.72 (0.46-1.14) 0.161 13.8 0.33 (0.15-0.71) 0.004 2009-2013 204 27.3 0.66 (0.42-1.02) 0.061 27.6 0.5 (0.26-0.95) 0.035 Age Young 324 62.35 1 8.02 1 Adult 355 48.45 0.7 (0.5-0.98) 0.039 17.18 1.96 (1.14-3.37) 0.014

148

N koalas Dead on Arrival Risk Factor submitted Percentage RRR (CI 95%) p Wald test Year 1997-2001 95 18.95 1 <0.0001 2001-2005 102 16.67 0.88 (0.42-1.85) 0.732 2005-2009 113 28.32 1.6 (0.82-3.13) 0.168 2009-2013 207 51.69 4.92 (2.69-8.99) <0.0001 Season Jan-Mar 134 23.88 1 0.0009 Apr-Jun 115 40 2.8 (1.55-5.06) 0.001 Jul-Sep 114 43.86 3.03 (1.69-5.45) <0.0001 Oct-Dec 154 29.87 1.68 (0.96-2.95) 0.071 Sex Female 363 36.64 1 0.3163 Male 152 26.97 0.81 (0.52-1.28) 0.372

149

N koalas Dead on Arrival Euthanized Risk Factor submitted Percentage RRR (CI 95%) p Wald test Percentage RRR (CI 95%) p Wald test Year 1997-2001 114 51.75 1 0.0178 45.61 1 0.0019 2001-2005 61 50.82 0.79 (0.13-4.97) 0.8 45.9 0.81 (0.13-5.12) 0.821 2005-2009 55 54.55 0.15 (0.04-0.6) 0.007 27.27 0.09 (0.02-0.36) 0.001 2009-2013 38 63.16 0.61 (0.1-3.88) 0.601 31.58 0.35 (0.05-2.31) 0.273

150

N koalas Dead on Arrival Euthanized Wald Risk Factor submitted Percentage RRR (CI 95%) p Wald test Percentage RRR (CI 95%) p test Age Young 65 46.15 1 - 13.85 1 - Adult 194 63.4 4.23 (2.12-8.42) <0.0001 23.71 5.29 (2.13-13.11) <0.0001

151

.2 .76 .16 .74 .14 .18 .12 .72 .16 Probability Released Probability Euthanised .1 Probability Dead on Arrival .7 .14 .08

.68 Young Adult Young Adult Young Adult Age Class Age Class Age Class

152

Appendix 13 Chlamydia-like signs: Predicted probabilities (with 95% confidence intervals) derived from multinominal regression models of risk factors associated with the diagnostic outcomes (dead on arrival, euthanized, released) for koalas submitted to wildlife hospitals in SEQLD from 1997-2013.

.6 .24 .28 .26 .5 .22 .24 .2 .4 .22 Probability Euthanised Probability Dead on Arrival Probability Dead on Arrival .3 .18 .2 .18 .2 Young Adult .16 Age Class Female Male Female Male Sex Sex .6 .55 .45 .5 .5 .4 .4 .35 .45 Probability Released Probabilities Released .3 Probability Euthanised .3 .4 .25 .2 .35 Female Male Young Adult Sex Young Adult Age Class Age Class

153

.42 .08 .65 .4 .06 .38 .6 .04 .36 Probability Released Probability Dead on Arrival Probability Euthanised .55 .02 .34 0 .32 .5 Female Male Female Male Sex Female Male Sex Sex .5 .12 .65 .1 .6 .45 .08 .55 .06 .4 .5 Probability Released Probability Dead on Arrival Probability Euthanised .04 .45 .35 .02 Young Adult .4 Young Adult Age Class Young Adult Age Class Age Class

154

1997-2001 2001-2005 2005-2009 2009-2013 1997-2001 2001-2005 2005-2009 2009-2013 1997-2001 2001-2005 2005-2009 2009-2013 Year Period Year Period Year Period .24 .26 .64 .22 .24 .62 .2 .22 .6 .18 Probability Released .2 Probability Euthanised Probability Dead on Arrival .58 .16 .18 .14 .56 Female Male Female Male Female Male Sex Sex Sex .25 .64 .3 .62 .2 .25 .6 .15 .2 Probability Euthanised Probability Released Probability Dead on Arrival .58 .1 .56 .15 Young Adult Young Adult Young Adult Age Class Age Class Age Class

155

Appendix 16 Pathology of Chlamydiosis: Ocular disease with clinical grading

156

Appendix 17 Pathology of Chlamydiosis: Lower urinary tract disease with clinical grading

157

Appendix 18 Pathology of Chlamydiosis: Upper urinary tract disease

158

Appendix 19 Pathology of Neoplasms

159

160

Appendix 20 Pathology of Parasitosis

161

Appendix 21 Pathology of Putative AIDS-like Syndrome

162

Appendix 22 Pathology of Idiopathic Diseases

163

Appendix 23 Pathology of Congenital Diseases

164

Appendix 24 Pathology of Degenerative Diseases

165

Abscessation 0.458 (0.434-0.483) 0.622 (0.561-0.682) 0.515 (0.495-0.538) 0.428 (0.366-0.491) 0.542 (0.522-0.57) 0.504 (0.462-0.537) Anaemia 0.452 (0.425-0.479) 0.647 (0.58-0.711) 0.519 (0.495-0.548) 0.428 (0.368-0.486) 0.529 (0.512-0.554) 0.515 (0.473-0.55) Blind 0.462 (0.427-0.503) 0.703 (0.623-0.771) 0.545 (0.509-0.584) 0.45 (0.393-0.502) 0.511 (0.499-0.523) 0.547 (0.503-0.592) Bone marrow disease 0.458 (0.432-0.485) 0.596 (0.549-0.649) 0.513 (0.497-0.53) 0.415 (0.342-0.496) 0.561 (0.533-0.598) 0.498 (0.454-0.53) Cardiac/pericardial disease 0.445 (0.42-0.471) 0.609 (0.55-0.669) 0.499 (0.481-0.52) 0.415 (0.346-0.488) 0.556 (0.53-0.591) 0.487 (0.442-0.521) Chlamydia 0.48 (0.441-0.528) 0.733 (0.653-0.799) 0.57 (0.528-0.612) 0.473 (0.416-0.527) 0.513 (0.501-0.526) 0.575 (0.527-0.622) Chlamydial conjunctivitis 0.499 (0.452-0.555) 0.766 (0.683-0.831) 0.599 (0.549-0.646) 0.496 (0.437-0.557) 0.509 (0.493-0.533) 0.607 (0.55-0.66) Chlamydial cystitis/urogenital 0.531 (0.477-0.592) 0.801 (0.722-0.86) 0.637 (0.582-0.687) 0.532 (0.47-0.597) 0.522 (0.497-0.556) 0.649 (0.588-0.704) Chlamydial female reproductive 0.521 (0.469-0.58) 0.791 (0.71-0.852) 0.625 (0.572-0.675) 0.52 (0.459-0.585) 0.518 (0.496-0.548) 0.636 (0.577-0.691) Chlamydial male reproductive 0.445 (0.421-0.471) 0.607 (0.548-0.667) 0.498 (0.48-0.519) 0.414 (0.347-0.487) 0.552 (0.526-0.587) 0.486 (0.441-0.52) Chlamydial nephritis/pyelonephritis 0.465 (0.427-0.511) 0.718 (0.635-0.788) 0.557 (0.515-0.6) 0.457 (0.399-0.512) 0.501 (0.489-0.513) 0.563 (0.515-0.612) Dermatitis 0.462 (0.43-0.499) 0.689 (0.614-0.755) 0.537 (0.503-0.572) 0.446 (0.39-0.496) 0.519 (0.506-0.535) 0.537 (0.494-0.578) Dermatosis 0.447 (0.422-0.472) 0.6 (0.545-0.658) 0.495 (0.477-0.514) 0.418 (0.347-0.495) 0.565 (0.536-0.6) 0.482 (0.435-0.518) Diabetes 0.464 (0.438-0.49) 0.642 (0.579-0.703) 0.525 (0.502-0.551) 0.438 (0.378-0.498) 0.538 (0.521-0.563) 0.519 (0.479-0.554) Disease 0.457 (0.433-0.482) 0.598 (0.547-0.653) 0.503 (0.486-0.52) 0.42 (0.35-0.495) 0.563 (0.536-0.598) 0.486 (0.44-0.521) Ectoparasitosis (mite, tick) 0.438 (0.412-0.464) 0.586 (0.53-0.647) 0.482 (0.465-0.499) 0.404 (0.332-0.483) 0.56 (0.527-0.599) 0.466 (0.419-0.503) Emaciation/poor body condition 0.536 (0.481-0.597) 0.806 (0.727-0.864) 0.643 (0.588-0.693) 0.537 (0.476-0.603) 0.525 (0.498-0.559) 0.655 (0.594-0.71) End stage kidney disease 0.454 (0.428-0.48) 0.63 (0.568-0.689) 0.512 (0.492-0.536) 0.428 (0.363-0.495) 0.553 (0.531-0.582) 0.502 (0.458-0.535) Facial oedema 0.462 (0.436-0.488) 0.64 (0.578-0.7) 0.521 (0.499-0.547) 0.436 (0.376-0.496) 0.543 (0.525-0.568) 0.514 (0.472-0.547) Flebitis 0.454 (0.429-0.479) 0.618 (0.558-0.678) 0.505 (0.486-0.527) 0.425 (0.36-0.491) 0.55 (0.528-0.581) 0.493 (0.448-0.526) 166

UQ-WH1 UQ-WH2 UQ-WH3 WH1-WH2 WH1-WH3 WH2-WH3 Probability of Probability of Probability of Probability of Probability of Probability of Diagnostic Category agreement agreement agreement agreement agreement agreement (95%) (95%) (95%) (95%) (95%) (95%) Fungal infection 0.465 (0.439-0.491) 0.639 (0.577-0.699) 0.528 (0.506-0.553) 0.435 (0.376-0.495) 0.539 (0.522-0.564) 0.52 (0.48-0.555) Gastrointestinal - inflammatory 0.446 (0.42-0.474) 0.642 (0.574-0.708) 0.513 (0.488-0.541) 0.424 (0.364-0.483) 0.526 (0.509-0.553) 0.507 (0.466-0.543) Gastrointestinal - noninflammatory 0.473 (0.437-0.517) 0.711 (0.632-0.779) 0.555 (0.517-0.596) 0.463 (0.407-0.517) 0.511 (0.5-0.522) 0.559 (0.514-0.605) Heat stress 0.463 (0.437-0.49) 0.648 (0.586-0.707) 0.53 (0.506-0.556) 0.438 (0.378-0.498) 0.541 (0.524-0.565) 0.523 (0.483-0.557) Hydronephrosis/hydroureter 0.455 (0.428-0.482) 0.642 (0.576-0.704) 0.517 (0.494-0.543) 0.433 (0.371-0.497) 0.548 (0.529-0.575) 0.509 (0.466-0.542) Hypoglycaemia 0.468 (0.439-0.498) 0.663 (0.598-0.724) 0.534 (0.508-0.563) 0.446 (0.387-0.502) 0.537 (0.522-0.557) 0.529 (0.487-0.565) Infection 0.483 (0.452-0.517) 0.686 (0.618-0.747) 0.554 (0.524-0.587) 0.465 (0.409-0.514) 0.532 (0.521-0.545) 0.554 (0.513-0.593) Intersex 0.466 (0.438-0.495) 0.661 (0.596-0.722) 0.536 (0.51-0.565) 0.445 (0.385-0.501) 0.538 (0.522-0.558) 0.531 (0.49-0.566) Koala fibromatosis syndrome 0.452 (0.428-0.478) 0.606 (0.551-0.663) 0.505 (0.488-0.525) 0.417 (0.348-0.491) 0.556 (0.53-0.59) 0.492 (0.447-0.526) Lung non-inflammatory 0.448 (0.422-0.475) 0.586 (0.536-0.641) 0.491 (0.474-0.507) 0.412 (0.337-0.495) 0.573 (0.541-0.612) 0.473 (0.425-0.512) Lymphadenitis/lymphamegaly 0.463 (0.438-0.488) 0.628 (0.57-0.686) 0.517 (0.497-0.539) 0.432 (0.368-0.497) 0.554 (0.533-0.582) 0.506 (0.461-0.539) Lymphatic - other 0.45 (0.424-0.477) 0.579 (0.532-0.633) 0.492 (0.475-0.507) 0.413 (0.337-0.497) 0.573 (0.539-0.613) 0.474 (0.426-0.513) Mammary gland mass 0.464 (0.437-0.492) 0.655 (0.59-0.716) 0.537 (0.511-0.566) 0.441 (0.381-0.497) 0.533 (0.517-0.554) 0.534 (0.493-0.568) Membrane injury 0.468 (0.444-0.492) 0.61 (0.558-0.664) 0.519 (0.503-0.538) 0.428 (0.363-0.496) 0.551 (0.529-0.581) 0.506 (0.463-0.537) Mouth - inflammatory 0.444 (0.417-0.472) 0.644 (0.575-0.71) 0.512 (0.487-0.541) 0.422 (0.362-0.482) 0.526 (0.508-0.553) 0.507 (0.466-0.543) Mouth - non-inflammatory 0.501 (0.463-0.545) 0.736 (0.664-0.797) 0.591 (0.552-0.63) 0.491 (0.435-0.545) 0.532 (0.519-0.547) 0.596 (0.551-0.64) Musculoskeletal disease-other 0.456 (0.424-0.493) 0.679 (0.603-0.748) 0.533 (0.5-0.57) 0.44 (0.384-0.49) 0.511 (0.497-0.527) 0.535 (0.493-0.578) Neoplasia craniofacial tumour 0.458 (0.433-0.483) 0.625 (0.567-0.682) 0.516 (0.496-0.537) 0.43 (0.364-0.497) 0.557 (0.535-0.587) 0.504 (0.46-0.537) Neoplasia leukaemia 0.458 (0.435-0.482) 0.597 (0.543-0.654) 0.499 (0.482-0.517) 0.426 (0.36-0.495) 0.555 (0.529-0.587) 0.484 (0.439-0.518) Neoplasia lymphoma 0.474 (0.439-0.515) 0.704 (0.628-0.77) 0.556 (0.52-0.594) 0.461 (0.404-0.512) 0.517 (0.505-0.528) 0.559 (0.515-0.603) Neoplasia other 0.439 (0.413-0.466) 0.628 (0.56-0.695) 0.501 (0.478-0.527) 0.413 (0.353-0.475) 0.529 (0.508-0.559) 0.494 (0.452-0.529)

167

Nephrosis 0.444 (0.417-0.472) 0.641 (0.571-0.708) 0.51 (0.485-0.539) 0.423 (0.363-0.482) 0.524 (0.507-0.551) 0.505 (0.464-0.541) Omentum fibrosis 0.459 (0.433-0.485) 0.638 (0.576-0.697) 0.52 (0.499-0.545) 0.434 (0.371-0.498) 0.55 (0.531-0.577) 0.511 (0.468-0.544) Organ aplasia/hypoplasia 0.5 (0.46-0.547) 0.74 (0.663-0.803) 0.583 (0.543-0.624) 0.494 (0.438-0.548) 0.53 (0.519-0.545) 0.587 (0.542-0.632) Parasitosis other 0.463 (0.433-0.496) 0.672 (0.6-0.738) 0.535 (0.506-0.568) 0.447 (0.389-0.501) 0.53 (0.516-0.549) 0.532 (0.49-0.568) Peritonitis 0.46 (0.434-0.485) 0.628 (0.568-0.686) 0.519 (0.498-0.542) 0.432 (0.369-0.496) 0.55 (0.53-0.578) 0.509 (0.465-0.541) Pneumonia/bronchopneumonia 0.442 (0.415-0.469) 0.634 (0.566-0.701) 0.505 (0.481-0.533) 0.417 (0.357-0.477) 0.526 (0.506-0.554) 0.499 (0.458-0.535) Prolapse 0.463 (0.435-0.491) 0.655 (0.59-0.716) 0.536 (0.51-0.565) 0.44 (0.38-0.497) 0.532 (0.516-0.554) 0.533 (0.492-0.568) Pulmonary fibrosis 0.444 (0.418-0.471) 0.636 (0.571-0.697) 0.508 (0.486-0.532) 0.423 (0.357-0.491) 0.557 (0.534-0.588) 0.499 (0.454-0.533) Renal complication 0.458 (0.428-0.49) 0.549 (0.516-0.596) 0.484 (0.467-0.502) 0.413 (0.323-0.512) 0.602 (0.562-0.648) 0.458 (0.399-0.515) Senescence 0.466 (0.436-0.5) 0.678 (0.607-0.743) 0.547 (0.515-0.581) 0.449 (0.391-0.5) 0.522 (0.508-0.537) 0.547 (0.507-0.586) Septicaemia 0.454 (0.427-0.482) 0.569 (0.526-0.622) 0.49 (0.474-0.504) 0.41 (0.332-0.495) 0.576 (0.541-0.617) 0.468 (0.418-0.51) Snake bite 0.477 (0.446-0.511) 0.685 (0.617-0.747) 0.551 (0.521-0.583) 0.461 (0.404-0.511) 0.536 (0.523-0.549) 0.55 (0.507-0.587) Suspected retrovirus 0.461 (0.432-0.492) 0.667 (0.597-0.73) 0.544 (0.515-0.577) 0.437 (0.378-0.493) 0.519 (0.504-0.536) 0.545 (0.505-0.585) Trauma animal attack 0.609 (0.547-0.673) 0.861 (0.794-0.907) 0.714 (0.659-0.762) 0.613 (0.549-0.679) 0.575 (0.534-0.622) 0.726 (0.667-0.777) Trauma hit by car 0.73 (0.668-0.789) 0.929 (0.886-0.957) 0.824 (0.777-0.862) 0.737 (0.673-0.795) 0.683 (0.627-0.737) 0.833 (0.785-0.872) Trauma other 0.465 (0.441-0.488) 0.609 (0.556-0.666) 0.512 (0.495-0.532) 0.43 (0.365-0.495) 0.549 (0.528-0.579) 0.498 (0.454-0.531) Trauma tree fall 0.461 (0.43-0.496) 0.676 (0.602-0.742) 0.535 (0.504-0.57) 0.444 (0.387-0.494) 0.517 (0.504-0.534) 0.536 (0.494-0.576) Trauma unknown 0.45 (0.419-0.485) 0.674 (0.597-0.743) 0.525 (0.493-0.561) 0.433 (0.377-0.483) 0.51 (0.497-0.529) 0.525 (0.483-0.567) Urinary male/female non-chlamydia 0.448 (0.421-0.475) 0.64 (0.574-0.703) 0.512 (0.489-0.539) 0.424 (0.364-0.486) 0.539 (0.519-0.566) 0.505 (0.463-0.539)

168

Appendix 26 Excerpt from Basic Field Autopsy Guide for Koalas

169

Appendix 27 Excerpt from Basic Field Autopsy Guide for Koalas: Heart Dissection

170