A Necropsy Study of Disease and Comorbidity Trends in Morbidity and Mortality in the Koala (Phascolarctos cinereus) in South-East Queensland, Australia

Koalas are an iconic Australian marsupial undergoing precipitous population reduction in South-East Queensland from complex interacting threats. To investigate the causes of death and the interaction of comorbidities with demography in South-East Queensland koalas, a large scale, high-throughput prospective necropsy survey was conducted spanning 2013–2016. During this period, 519 necropsies were conducted in 155 young/subadult koalas, 235 mature, 119 old koalas and 10 of unknown age. Similar numbers of males and females were assessed. Trauma and infectious disease at were the most common single diagnoses. However, comorbidity was frequent, including multicentric infection or infectious disease in combination with trauma or senescence. Female koalas had proportionally more reproductive chlamydiosis compared to males in which the ocular and urinary systems were more commonly affected. Comorbidity and disease were strongly associated with poor body condition, and trauma was associated with good body condition. Animals affected by motor vehicle trauma were often in better body condition than those affected by animal attack, tree fall or other causes of trauma. This study identified a higher frequency of infections and comorbidity then previously reported, confirming the complex nature of interacting threats to the koala population.

. Percentage of diagnoses (N of diagnoses) of the four most common single (N = 220 koalas) and cooccurring (N = 299 koalas) diagnoses made on koalas submitted to wildlife hospitals and receiving necropsies at The University of Queensland between between April 2013 and July 2016 in South-East Queensland, Australia.
www.nature.com/scientificreports www.nature.com/scientificreports/ infectious disease. Of the total of 63.4% (329/519) cases with infectious processes (alone or in combination with other conditions), 92.4% (304/329) had hallmark lesions of Chlamydia spp. infection. The percentage of koalas by sex for each body system (alone or in combination with other body systems) affected by Chlamydia spp. are shown in Fig. 2. There were significant differences between female and male koalas in the body system affected  Table 2. Percentage of koalas (N of koalas) of the four most common diagnoses by sex (stratified by age class; N = 509 koalas with age and sex information) for koalas submitted to South-East Queensland hospitals and receiving necropsies at The University of Queensland from 2013 through to 2016. Koalas may be counted more than once in the diagnoses categories due to co-morbidities. * Statistical differences between sex groups at p < 0.05. Percentage of koalas per sex group within each age class for each diagnosis was calculated as N of koalas per sex and age group for each diagnoses/Total number of koalas per sex and age group.  (p < 0.001), with the reproductive system alone or in combination with other body systems affecting about 90.0% of female koalas, while ocular, urinary or urinary-ocular infections were observed between 65-78% of male koalas. The distribution of body systems affected by chlamydiosis for each sex strata are shown in Supplementary  Fig. 1, highlighting that the urinary tract was the most common body system affected in males (52.5% of all Chlamydiosis cases in males) and the urinary-reproductive tracts the most frequent body system affected in females ( Urinary disease. Cystitis was characterized grossly by thickening and firmness of the urinary bladder wall, with mucosal infolding and occasional petechiae (Fig. 3a). Soiling around the common vestibule colloquially referred to as 'wet bottom' 10,13 , was common in chronic cases (Fig. 3b). Clinical chlamydial scoring was recovered from medical records for 43 of the 176 cystitis cases, with 58.1% (25/43) severe and 18.6% (8/43) moderate grade. The predominant histological lesion was moderate-to-severe fibrosis accompanied by detrusor muscle hypertrophy, transmural lymphoplasmacytic infiltrates, and occasional urothelial ulceration. Chronic cystitis resulted in obstruction of the bladder outlet due fibrosis and myohypertrophy in the bladder wall. Renal complications compatible with chlamydial infection affected 34.8% (106/304) of all koalas and included 9/106 cases of hydronephrosis (Fig. 3c) and 13/106 of renal pyramidal fibrosis from chronic chlamydial pyelonephritis (Fig. 3d). Primary renal disease was also observed without cystitis in 51.8% (43/83) cases. The predominant histological lesion was lymphocytic interstitial nephritis or chronic lymphocytic segmental pyelonephritis. www.nature.com/scientificreports www.nature.com/scientificreports/ Severe renal disease was infrequent, and included ascending urinary tract infection -UTI (4 cases), end stage kidney disease (4 cases), and pyelonephritis (10 cases). Renal crystal precipitation was observed in 21 cases, mostly morphologically consistent with struvite in 47.6% (10/21) cases, and calcium oxalate in 23.8% cases (5/21).
Reproductive disease. Reproductive lesions consistent with chlamydia infection, predominantly cystic dilation of the ovarian bursa (bursitis) (Fig. 3e), occurred in 66.6% (122/183) female cases with Chlamydia spp. infections at necropsy. In 3.8% (7/183) of these cases, chronic bursitis was found along with gross evidence of salpingitis (acute or chronic). Association of bursitis with fibrotic salpingitis was observed but not consistently recorded. Fluid within the bursa varied from clear and straw-colored, to purulent or hemorrhagic. The bursa was bilaterally compromised in 55.7% (68/122) cases. The main histological feature was replacement of ovarian and fallopian structures with mature collagen, occasionally with lymphocytic infiltration. Infection of other reproductive tract sites co-occurred with bursitis in 22.1% (27/122) of cases, primarily fibrosis or suppurative infection of the uteri. In the male, 64.2% (9/14) of cases of urogenital infections were consistent with chlamydiosis, with cystitis and urethritis most common. Ten of fourteen cases had chronic multifocal prostatitis with intraglandular pustules. ocular disease. Ocular lesions consistent with chlamydiosis were diagnosed in 28.9% of koalas (88/304 Chlamydia spp. infections) (Fig. 3f). Chlamydial clinical scoring was recovered for 46.6% (41/88) of cases and frequently was high-grade (85.3%; 35/41; score ≥2). Ocular chlamydiosis was macroscopically a chronic, bilateral (60/88; 68.1%) mucopurulent conjunctivitis/keratoconjunctivitis, with periocular fur matting or alopecia, and palpebral adhesion.
trauma. Of the 36.4% (189/519) koalas diagnosed with trauma (alone or in combination with other conditions), based on hospital records, 57.7% (109/189) cases were from motor vehicle collisions, 21.2% (40/189) were from animal attacks (dogs, livestock), 6.9% (13/189) cases corresponded to tree falls, and in 14.3% (27/189) the cause was unknown (Table 3). Acute bone fractures affected 46.0% (87/189), mostly located in the head in 55.1% cases (104/189) or in long bones in 34.5% cases (65/189). Severe hemoabdomen and 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. A total of 10/19 females with evidence or history of carrying pouch or back young perished to trauma caused by motor vehicles.
In general, koalas with trauma (alone or in combination with other conditions) had better body condition than koalas submitted for other causes than trauma (p < 0.001) (Fig. 4).
Miscellaneous findings. Pulmonary lesions were found in 4.0% (21/519) of koalas including suppurative pneumonias and bronchopneumonias. Focal or interstitial fibrosis indicating chronicity was found in 33.3% (7/21) of koalas with respiratory lesions. Cases with rare or novel pathogens included a Psynchrobacter spp. pleuritis, a pulmonary abscess caused by Mycobacterium abscessus, and a pulmonary phaeohyphomycosis caused by Cladosporium spp., all PCR-confirmed.
Seven koalas of the 519 koalas presented with multinodular masses in the spleen (Fig. 6a), termed by the authors "koala fibromatosis syndrome". Nodules were formed by well-differentiated fibrocytes, which effaced and infiltrated the spleen (Fig. 6b) and liver (3 cases). Demographically, sex distribution was similar (4 males, 3 females) and no specific comorbidity pattern was observed. Lympholysis with moderate-to-marked splenic

Discussion
In the current study, the use of systematic necropsies in passive surveillance increased the sensitivity of disease detection and accurately determined comorbidity patterns compared to a previous analysis using medical records where only brief notes on disease were recorded in some animals. Further, systematic necropsies allowed the characterization of several novel diseases and lower frequency etiologies of infectious, neoplastic or idiopathic disease 10,11,15,16 .
Pathology data was recorded using a previously developed rapid classification scheme that streamlined data analysis 11 . The combination of a standardized data recording methodology with systematic necropsies, alongside medical record analysis and ancillary testing, permitted a high (>90%) proportion of conclusive necropsy diagnoses in this study, an excellent rate for wildlife mortality investigations. Other free-ranging wildlife necropsy studies report 46% to 95% rates of determination of primary cause of death 10,[17][18][19][20] .
The present study interrogated the interplay of concomitant disease conditions in SEQLD koalas, demonstrating a high rate of comorbidities and reinforcing the complexity and variability of threats to the species. As a single entity trauma was the most common diagnosis in the study, followed by infectious disease. Comorbidities were primarily due to multicentric infections, debilitation, co-occurring trauma and infections. Comorbidities were more common in older animals but still surprisingly prevalent in young koalas, where nearly half had more than one presenting problem. At a population level these conditions would limit breeding potential and increase vulnerability to disease or predation 21 .
Concomitant conditions in koalas have been investigated previously [21][22][23] . A higher proportion of Queensland koalas with loss of BC from comorbidities such as chronic infections or neoplastic disease has been reported than southern states 10,13,24 . Direct comparison of comorbidity rates with previous studies is complicated due to the different study designs used. Nonetheless, the determination of variation in comorbidity categories between studies and over time assists in establishing a historical pattern for disease interaction in koalas 10 . Reports from northern koala populations using smaller sample sizes have found comorbidity proportions ranging from 38-60% in QLD 13,24 and 19-53% in NSW 21,22,25 . The proportion of comorbidities found in this study (58%) and our former retrospective epidemiological analysis (38%) is within the documented QLD range. Given the large cohort of koalas assessed and the detailed assessment in the present study, higher accuracy is likely in this analysis.
Historically, trauma has been a major contributor to koala mortality in SEQLD 10,11 . When koalas inhabit forested land in close proximity to urbanized centers, there is high mortality from trauma, particularly by motor vehicles and dog predation, demonstrating poor adaptability as these threats permeate into their habitat 10,26 . Koala-motor vehicle collisions have a high fatality rate (>80%) 10,27 from a combination of factors 22,28 , including comminuted fractures 21,22 and severe soft tissue trauma causing massive hemorrhage and shock. Dog predation also causes severe soft tissue damage, often leaving pathognomonic bite wound lesions 29 . Although most animal attacks on koalas in urban areas come from domestic dogs [30][31][32] , predation threat could also permeate into protected forests or bushland habitat when carried out by wild dogs and dingoes 21,33 . Koalas are crepuscular species 1 thus responsible pet ownership that restricts dog movements on private land during dawn and dusk may be beneficial 10 . At younger ages, both male and female koalas are at risk of trauma during post-weaning dispersal 30 . However, the predominance of mature male koalas in the trauma group could reflect larger home ranges and increased mobility during breeding season 34,35 . Younger males also tend to disperse more frequently and further than female koalas 36 . The removal of subadult males by trauma could disrupt population genetic heterogeneity, as these koalas are the primary inter-group dispersers contributing to gene flow 36 . Notably, in both the medical record and necropsy studies, over half of the animals killed by motor vehicle trauma were healthy. The proportion of healthy animals was also high in animal attack and other trauma causes. Thus, many trauma cases were unnaturally removed from the population predominantly due to anthropogenic factors, potentially causing detrimental impacts on the survival of local populations by reducing healthy breeding stock.
The most frequently reported infectious disease in koalas is chlamydiosis caused by Chlamydia pneumoniae producing conjunctivitis and pneumonia, or the more pathogenic C. pecorum, which causes conjunctivitis, cystitis and reproductive infection resulting in infertility [37][38][39] . Classical presentations of clinical chlamydiosis in this study were typically chronic and severe. This reflects the bias of sick wild koalas coming to attention of the public and wildlife rescuers, as well as the clinical management of severely debilitated koalas by euthanasia. In this study morphologic confirmation of typical chlamydial lesions was considered sufficient for diagnosis. Molecular testing for chlamydiosis is of limited value in our context of interrogating drivers of mortality, as subclinical carriers are common and PCR detection and chlamydial load do not correlate with clinical severity 40,41 . For example, koalas with substantial chronic Chlamydia-associated structural pathology may be PCR-negative and lack chlamydial inclusion bodies, and conversely higher loads can be observed in infected asymptomatic animals then those with clinical disease [40][41][42] . A recent study on Chlamydia pecorum in the male koala reproductive system demonstrated 70% of morphologically normal samples and 89% of abnormal samples were qPCR positive 43 . Typically studies utilizing active surveillance document elevated frequencies of milder chlamydial infections, compared with this study emphasizing bias in sampling methodology 44 . This dataset will overestimate the proportion of severe end-stage conditions, and underestimate subclinical carriers, but was appropriate for identifying the impact of chlamydiosis as a cause of mortality and infertility 10 .
A disproportionate number of female koalas of breeding age were diagnosed with chlamydial reproductive disease alone or as a comorbidity 10,44-46 compared to males, despite the documented similarity in infection rates across sexes 38,47 . This reflects the use of ultrasonography to identify permanent female infertility and clinical management by euthanasia 10,45,46,48 . Additional features influencing female bias in reproductive chlamydiosis include urogenital anatomy, increased rate of opportunistic infections, and lack of detectable sonographic changes in the male unless severely affected 44,49,50 . However, clinical and subclinical male reproductive tract disease was detected with histopathology. The impact of male urogenital chlamydosis on individual and population health warrants further investigation. Specifically diagnostic testing of hospitalized male koalas and evidenced-based evaluation of the value of treatment of asymptomatic infections should be explored 11,44,48 .
Cystitis, the main clinical manifestation compatible with chlamydiosis, was found in similar proportion (34%) to other reports 21,23,51 but higher compared to that described in an previous study by the authors (27%) 10,11 . This difference is likely due to the increased sensitivity of necropsies. Notably, a higher proportion of males had ocular and urinary infections alone or as comorbidities. Previous QLD reports questioned the relationship between low BC and ophthalmic chlamydiosis 52 . In the present study, only 9% of individuals with solely ocular disease had low BC, whilst >60% of koalas with low BC had chronic ocular chlamydiosis associated with comorbidities. Thus, general systemic disease rather than visual impairment causes low BC compromising survivorship 52 . This study has reinforced that chlamydiosis is still the major impactor on koala mortality and morbidity, with a high rate of multicentric disease and comorbidity.
Koalas live on a highly specialized, nutrient-restricted diet, and have limited visceral fat reserves 10,52 . As a result, koalas are prone to debilitating and sometimes irreversible BC loss, known as wasting 10,11,[53][54][55] . Factors causing loss of BC include stressors (prolonged veterinary treatment), or those derived from biotic (social dynamics, resource availability, predation), climatic factors, and dental attrition reducing digestion efficiency [56][57][58][59] . Thus, BC is a relevant parameter to assess in koalas 10,11,60 as it may impact successful rehabilitation, population management, or even interact with comorbid states reflecting disease severity. In this study a strong correlation between the presence of disease, comorbidities, type of trauma and body condition was present. Virtually half of diseased koalas (48%) in the present study had poor BC, and animals with poor BC were more likely to have died from non-traumatic causes. In contrast, 95% of the healthy koalas with trauma had a fair to excellent BC. A larger proportion of koalas with low BC were involved in animal attacks, tree fall and other trauma, confirming an elevated risk with increased ground time due to debilitation in increasingly urbanized landscapes 21 . Conversely, vehicle collisions generally affected animals in good or excellent BC, suggesting these may be active and robust animals moving in search of mates or forage. Notably, male koalas in the young/subadult age class were diagnosed most frequently with idiopathic low BC, exhibiting non-specific clinical histories and lacking necropsy findings. Previous studies reported idiopathic low BC primarily in females, and associated with lymphoid depletion, though this was not consistent in the young males in this study 21,23,24,44 .
Passive surveillance methodology has limitations including minimal control over data quality, lack of denominator values to estimate disease frequency, and overrepresentation of certain demographic groups or diseases (i.e. incurable, terminal) 10,59 . For instance, chronic disease and low BC were observed at high frequency, reflecting the biased selection of terminal animals. Conversely, acute infectious disease was underrepresented, reflecting a bias against individuals which could be successfully rehabilitated 52 . Carcass preservation is often a challenge with wildlife sampling. In koalas, the quality of tissues available for study may have sporadically compromised diagnostic sensitivity due to poor preservation, freezing artefact or eviscerated from previous necropsies, or pap collection for orphaned koalas. In spite of these limitations, utilizing hospital admissions and standardized necropsies and sample collection. Necropsies were conducted at the School of Veterinary Science, The University of Queensland, between April 2013 and July 2016. The majority of carcasses were frozen at −20 °C due to logistical reasons, and thawed between 0-4 °C for 4-7 days prior to examination. Koalas were identified by ear tags, microchip numbers, or by the hospital accession numbers. Aging was done according to premolar and molar tooth wear designating koalas to an age class (young/subadult, mature, and old) 10,56,62 . 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 60,63 .
Histopathology samples were fixed in 10% NBF, sectioned at 4 µm 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 10,64 . The bacterial PCR assay targeted conserved regions U1 and U3 for the 16S rRNA gene using broad range primers 10,65 . Variable portions within the 16S rRNA gene provide unique signatures of any bacterium. Identification was conducted via sequence analysis of the amplified product 10,65 . clinical disease grading criteria. Chlamydial conjunctivitis and cystitis was subjectively scored by hospitals based on the degree of tissue damage and disease progression following the disease grading criteria of the Moggill Koala Rehabilitation Centre Standard Operating Procedures for conjunctivitis and cystitis treatment 2017 (Supplementary Table 2) as well as previous reports 44 . Chlamydial bursitis was not clinically-scored as it causes permanent infertility regardless of the degree, with affected females being unable to be returned to the wild, most being euthanased with barbiturates, as per the discretion of the consulting veterinary surgeon under the Veterinary Surgeons Act of Queensland.
Data analysis. Binomial, or also called exact or Clopper-Pearson confidence intervals were calculated for the overall prevalence of observed co-morbidities and for co-morbidities stratified by sex and age 66 . Exact confidence intervals are more appropriate for small proportions (as for the co-morbidity strata in our study), but as they more conservative, they might be wider than traditional Wald-type intervals 67 .
Analyses of relationships between the various diagnoses made and demographic information of koalas submitted (sex, age, age class, body condition) were conducted using contingency tables. Fisher exact tests were used to compare proportions in the contingency tables 68