A 9‐year retrospective mortality review of captive orange‐bellied parrots ( N eophema chrysogaster) in Australia

Abstract

The orange‐bellied parrot (Neophema chrysogaster) (OBP) is a critically endangered migratory parrot found in south‐eastern Australia. A long‐standing conservation recovery programme includes monitoring and support of the wild population and a multi‐site captive insurance population. While disease represents a threatening process to the OBP, minimal information about causes of mortality in this species has been published. A systematic review of necropsy reports from the captive OBP metapopulation over a 9‐year period (2010–2019) was undertaken to quantify and analyse the causes of mortality of captive OBPs. Infectious disease was the most common cause of death (35%), including respiratory aspergillosis and a range of bacterial infections, followed by trauma (7%). Confirmed or suspected Psittacine Beak and Feather Disease caused few deaths overall (1.4%) but caused an outbreak of juvenile and subadult mortality at one facility in 2019. Necropsy records alone were insufficient to determine the contribution of this infection to captive population health beyond these cases. Disease occurrence varied with site and age class; in particular, aspergillosis and mycobacteriosis were each more common at two separate sites, offering opportunities for further investigation of disease risk factors that could include environmental, husbandry and pathogen factors. A high proportion of cases (46%) could not be assigned a cause of death, though diagnostic success increased with age class, year of study and when histology was performed in younger age classes. These findings are relevant to future conservation decisions for the OBP, particularly resource allocation for research and diagnostics, and have broader interest to veterinarians, pathologists, ecologists and captive management specialists. Our findings also reinforce the importance of consistency in necropsy investigations and pathological terminology and the need for accessible and standardised mortality records for conservation programmes.

Abbreviations

BFDV

Beak and feather disease virus

OBP

Orange‐bellied parrot

OBPRT

Orange‐bellied parrot recovery team

PBFD

Psittacine beak and feather disease

VTRG

Veterinary technical reference group

WDRA

Wildlife disease risk assessment

ZIMS

Zoological Information Management System

The orange‐bellied parrot (Neophema chrysogaster, OBP) is a critically endangered migratory parrot with a restricted wild distribution in southern mainland Australia and a single breeding site in southwestern Tasmania. The species has suffered severe population decline, partly due to loss, fragmentation and degradation of habitat, but also due to high mortality in the first year of life. ¹ , ² , ³ The 2025/26 summer census counted 86 returning individuals. ⁴

A multi‐institutional species recovery program has been underway since 1984 ¹ and is coordinated by the OBP Recovery Team (OBPRT). As part of recovery actions, the program has established a secure captive breeding and insurance population across multiple institutions. Suitable captive‐bred individuals from this population have been translocated between captive sites for breeding or seasonally released to supplement the wild flock. Other recovery actions include monitoring and management of the wild population, actions to secure and improve habitat quality, and ongoing population tracking and modelling.

Disease was listed as a high‐risk threatening process in the most recent report of OBP recovery activities. ⁵ Additionally, loss of diversity in innate immune function genes has been reported in captive and wild OBP, potentially making the population more vulnerable to disease outbreaks and less adaptable to new pathogens. Diseases may impact the health, breeding success and translocation success of captive individuals, and risk introducing novel disease threats to the wild population, with potentially adverse conservation outcomes. ⁶ The vulnerability of small populations to stochastic factors such as disease, and the potential for further loss of genetic diversity through avoidable mortality means that disease control is of vital importance for successful conservation outcomes in endangered species. ⁷

Morbidity and mortality reviews are an integral part of captive wildlife management and disease control, allowing accurate identification of pathogens and patterns of disease, timely control actions and tracking of interventions. ⁸ Reviews also provide valuable data to stakeholders and can aggregate disconnected health information across a metapopulation. The cost of diagnostics and disease investigation is significant and must be directed towards the highest potential conservation and welfare value, such as diseases which affect nestling mortality or reproductive success. Opportunity costs can therefore arise from incomplete knowledge of disease threats, due to misdirection of resources at the cost of other beneficial recovery actions and delayed response in the face of disease incursions.

Despite its critically endangered status and high public profile, disease threats to the OBP are sparsely documented in the peer‐reviewed literature. A 1989 conference paper about OBP management briefly noted Psittacine Beak and Feather Disease (PBFD) as the main cause of captive deaths from 1985 to 1989. ⁹ An investigation of 77 deaths over a 5‐year period at one facility was published in 2000; this identified elevated tissue zinc in a subset of birds, and zinc intoxication from galvanised aviary wire was considered a likely cause of death and reduced fertility. ¹⁰ This report also noted several other causes of death including chlamydiosis, aspergillosis and acute trauma. Other studies are generally limited to assessment for specific viral infections, particularly psittacine beak and feather disease virus (BFDV). ¹¹ , ¹² , ¹³ Quantification of the OBP viral microbiome has also recently been reported as an initial step at identifying endemic viruses in the captive population. ¹⁴

While there are also limited, informal reports of diseases in recovery program reports and threat assessments, ⁵ , ¹⁵ there are no published systematic reviews of mortality and morbidity in the captive or wild population. Institution‐specific health and disease records remain in‐house or are consolidated within disparate unpublished reports to the OBPRT. This compartmentalisation of health information makes it difficult to accurately assess disease threats across the captive metapopulation and consequent potential risks to the wild population.

A Veterinary Technical Reference Group (VTRG) was established in 2016 to advise the recovery programme on disease management and biosecurity. ¹⁶ The group recommended a Wildlife Disease Risk Assessment (WDRA) be conducted to assess the risks involved in OBP translocation, which was undertaken in 2020. ¹⁷ Wildlife disease risk assessment provides a formal critical review process that can guide the direction of resources toward reducing disease risks associated with wildlife interventions. ¹⁸ To aggregate all known causes of mortality within the contemporaneous captive metapopulation and better inform the OBP WDRA, a detailed mortality review was conducted. This report presents the results of that review and includes a discussion of the diseases of importance, the challenges encountered in the review process, the usefulness of the review not only to the WDRA process but also more broadly, and suggestions for improving the mortality review process when dealing with a dispersed metapopulation.

Methods

Case definition and discovery

A nine‐year study period from 1 July 2010 to 30 June 2019 was nominated with consideration of the mean captive OBP lifespan of 8–10 years, ⁵ the limited time available to collate the dataset for the WDRA and the need to capture diseases that posed a risk to the contemporaneous captive OBP population.

Cases were defined as any OBP that died in a captive facility within the study period and had a discoverable necropsy record. Case records were identified by contacting facilities housing captive OBPs and requesting necropsy and associated records stored in their in‐house electronic animal management systems, including the Species 360 Zoological Information Management System (ZIMS), ¹⁹ or bespoke databases.

Six facilities were included in this study, all housing birds from the captive insurance population and spread across four states (New South Wales, South Australia, Victoria and Tasmania). Two facilities were not included due to limited availability of relevant data at the time the study was undertaken. The OBP Studbook allowed cross‐referencing to obtain birth and death dates. ²⁰ ZIMS inventory summaries were obtained from each site to allow reference to total captive population numbers and the number of bird deaths over the study period.

Animal ethics statement

The OBP captive breeding and insurance population is managed as a regional species program with the support of the Zoo and Aquarium Association. ²¹ All case material was derived from records of routine diagnostic investigations performed as part of normal captive species management.

Data handling

Information from case records was manually entered into an Excel database. Data included ZIMS identification number or site‐specific unique animal identifier, date of birth, date of death, age at death, sex, captive site at the time of death, clinical history, circumstances of death, date of laboratory report, laboratory number, laboratory, histological lesions and morphological diagnoses. Cause of death was assigned by consensus between two veterinary pathologists (CB and ED) using information from the histopathology report, gross post‐mortem report, clinical history or a combination of these. Where a cause of death could not be decided, a classification of ‘Not Determined’ was assigned.

Assumptions and definitions

Age classes were determined in consultation with OBPRT and VTRG staff (S Troy and V Peric, personal comment) to reflect captive management as follows: Hatchling (0–5 days), Nestling (6–28 days), Fledgling (29 days to 6 weeks), Juvenile (6–12 weeks), Subadult (12 weeks to 1 year), Adult (1 to 9 years) and Geriatric (older than 9 years). The geriatric class was included as an effort to highlight any disease processes specific to aged birds. In the absence of published definitions of geriatric age in OBPs, this was confirmed with OBPRT staff to reflect an aged disposition (V Peric, personal comment). ²² Age was calculated from birth and death dates where this data was available. If not, age class was assigned based on information in the clinical or pathology record.

Circumstances of death were assigned as Found Dead (bird found dead with no premonitory signs of ill health or therapy noted in the history), Euthanased (euthanased after a period of veterinary care), Died Under Veterinary Care (died after clinical signs had been noted, either during diagnostic procedures or after therapy started), GA Death (death during or immediately after general anaesthetic for any reason) or Missing (nestlings missing from nest).

Terminology and reporting of pathological findings varied across necropsy records as expected with the use of multiple clinicians, laboratories and pathologists. To allow pathological findings to be collated and compared, two veterinary pathologists (ED and CB) developed standardised cause‐of‐death terminology. Multiple levels of categorisation were applied to allow each cause of death to be grouped with mechanistically similar diseases while still being able to be differentiated as a distinct process where relevant. Each cause of death was categorised as infectious or non‐infectious. Infectious diseases were further categorised by the main organ system and, where possible, microbial aetiology, while non‐infectious diseases were grouped into categories with common causal features including trauma, nutritional and developmental disease. Where multiple pathological processes were reported, the process considered the most likely driver of death was assigned.

Results

A total of 641 cases were identified within the study period and are summarised by location in Table 1. Necropsy record format comprised ZIMS necropsy case reports (Facilities A, C, E and F), a ZIMS activity report (Facility D) and an Excel database (Facility B). The proportion of deaths with a discoverable necropsy record was at or close to 100% for most facilities. A lower proportion of deaths had a discoverable necropsy record at one site (Facility B, 43%). The two sites with the largest OBP population (A and B) together contributed the majority (67%) of cases.

Table 1.

Records of captive orange‐bellied parrot mortality between July 2010 and June 2019. Data are shown as the number of necropsy records discovered over the number of deaths reported in ZIMS population summaries

Captive facility	Number of necropsy records discovered/number of deaths		Proportion of all necropsy records in study
Facility A	265/272	(97%)	41%
Facility B	167/388	(43%)	26%
Facility C	120/122	(98%)	19%
Facility D	57/57	(100%)	9%
Facility E	17/17	(100%)	3%
Facility F	15/16	(94%)	2%
Total	641/872	(74%)	100%

Case demographics

Necropsy records covered the full range of age classes, as shown in Figure 1. Birth and death dates were available for 603 cases. Of the remaining 38 cases, 13 could not be classified to an age class using other information. A high proportion of hatchlings (88%, 120/136), nestlings (77%, 84/109) and fledglings (88%, 28/32) did not have sex recorded in the case record, meaning sex could not be determined in 41% (266/641) of cases. Where available, cases included approximately equal proportions of male and female birds (186 males and 189 females).

Figure 1.

Age class at death in captive orange‐bellied parrots that died between July 2010 and June 2019 and had a discoverable necropsy report. Exact age could not be determined for 38 cases. Age classes defined as Hatchling (0–5 days), Nestling (6–28 days), Fledgling (29 days–6 weeks), Juvenile (6–12 weeks), Subadult (12 weeks–1 year), Adult (1 year–9 years) and Geriatric (older than 9 years).

The age at death ranged from 1 day to 15 years, 8 months. More than half (59%, 377/641) of the necropsy records were from birds that died at less than 1 year of age, with the two youngest age classes (hatchlings and nestlings) representing more than one‐third of all deaths. More frequent deaths in this age class led to a seasonal pattern where the highest number of mortalities was seen over the austral summer months. Relatively few records were from birds that had died at 5 years of age or older (14%, 82/603) or older than 10 years (4%, 24/603).

Circumstances of death

Approximately half of all necropsy reports (52%, 333/641) indicated the bird was found dead with no premonitory signs of illness (see Figure 2). There were fewer cases where birds had been euthanased after a period of clinical illness or care (26%, 167/641) or died in care (14%, 87/641). Hatchlings were more frequently found dead than other age classes.

Figure 2.

Circumstances of death for different age classes of captive orange‐bellied parrots that died between July 2010 and June 2019 and had a discoverable necropsy record. Age class could not be determined in 13 cases. GA, general anaesthetic.

Necropsy report quality and diagnostic success rate

The extent of necropsy investigation documented in each necropsy record varied between sites (Table 2). A histopathology report was included in 76% (490/641) of all cases, but this ranged from 5% to 100% of cases at different sites. Where histology was not performed or had not been recorded in the discoverable necropsy record, a cause of death was assigned based on the observations and conclusions of the facility staff as documented in the necropsy record. Reasons for histology not being performed were generally not documented in the necropsy record. Results of ancillary testing such as microbial culture, serology or PCR, if performed, were typically found in the necropsy records as additional notes or comments on a case‐by‐case basis. Gross and histological pathology reports and the conduct of ancillary testing did not follow a standardised format across institutions.

Table 2.

Diagnostic success based on age class and whether histology was performed in captive orange‐bellied parrots that died between July 2010 and June 2019. Diagnostic success was defined as the proportion of cases where a cause of death could be assigned. Cases where age class could not be determined are not included (n = 13)

Age class	n	Proportion of cases with a histology report	Diagnostic success rate
			In cases with histology	In cases without histology	In age class overall
Hatchling	136	63%	38%	12%	29%
Nestling	109	71%	55%	35%	50%
Fledgling	32	72%	75%	38%	66%
Juvenile	34	86%	87%	20%	77%
Subadult	64	77%	61%	67%	63%
Adult	221	85%	63%	52%	61%
Geriatric	32	77%	71%	75%	72%
	628	76%	59%	36%	54%

The overall diagnostic success rate, being the proportion of all cases that could be assigned a cause of death based on the necropsy record, was 54% (344/641). Diagnostic success was higher where histology was reported (59%, 289/491) and lower where histology was not performed or not recorded in the discoverable history (36%, 55/151). Diagnostic success improved with age class, particularly between the younger age classes. Diagnostic success was lowest in hatchlings at 29% but was consistently over 60% for fledglings and adults. Diagnostic success in hatchlings increased to 38% if histology was reported but dropped to just 12% without histology. The same pattern was seen in adults, with a success rate of 63% with histology and 52% without. The highest diagnostic success was seen in fledglings and juveniles where histology was performed (75% and 87% respectively).

Cause of death

Infectious and non‐infectious causes of death in captive OBPs within the study period are categorised by age class in Tables 3 and 4. The causes of death of most concern, as identified by the WDRA or due to their potential to cause population effects, are also shown by captive site in Table 5. Infectious disease was the most common cause of death identified (35%, 226/641), followed by trauma (7%, 46/641), degenerative conditions (4%, 23/641), nutritional or feed‐associated conditions (4%, 28/641), death after anaesthesia or blood collection (1%, 9/641), and visceral or articular gout (less than 1%, 5/641). There were very low numbers of developmental abnormalities, neoplasia and deaths due to adverse environmental conditions such as exposure from extreme temperatures.

Table 3.

Infectious and inflammatory causes of death in captive orange‐bellied parrots that died between July 2010 and June 2019 and had a discoverable necropsy record. Figures in parentheses represent percentage of all cases in that age class

Infectious cause of death	Hatchling	Nestling	Fledgling	Juvenile	Subadult	Adult	Geriatric	NS	Total cases
Cases	134	111	32	34	65	221	31	13	641
Respiratory (sinusitis, pneumonia, air sacculitis)	6 (4.5%)	32 (28.8%)	6 (18.8%)	11 (32.4%)	11 (16.9%)	27 (12.2%)	4 (12.9%)	3 (23.1%)	100 (15.6%)
Aspergillosis	3 (2.2%)	27 (24.3%)	2 (6.3%)	7 (20.6%)	6 (9.2%)	19 (8.6%)	3 (9.7%)	3 (23.1%)	70 (10.9%)
Other fungal	–	1 (0.9%)	1 (3.1%)	–	–	–	–	–	2 (0.3%)
Bacterial	2 (1.5%)	2 (1.8%)	1 (3.1%)	1 (2.9%)	–	4 (1.8%)	1 (3.2%)	–	11 (1.7%)
Pathogen not identified	1 (0.7%)	2 (1.8%)	2 (6.3%)	3 (8.8%)	5 (7.7%)	4 (1.8%)	–	–	17 (2.7%)
Systemic	5 (3.7%)	3 (2.7%)	4 (12.5%)	5 (14.7%)	10 (15.4%)	47 (21.3%)	–	–	74 (11.5%)
Pseudomonas aeruginosa septicaemia	3 (2.2%)	1 (0.9%)	2 (6.3%)	1 (2.9%)	1 (1.5%)	10 (4.5%)	–	–	18 (2.8%)
Other bacterial septicaemias	2 (1.5%)	2 (1.8%)	2 (6.3%)	1 (2.9%)	–	8 (3.6%)	–	–	15 (2.3%)
Mycobacteriosis	–	–	–	–	–	15 (6.8%)	–	–	15 (2.3%)
BFDV‐associated	–	–	–	3 (8.8%)	4 (6.2%)	2 (0.9%)	–	–	9 (1.4%)
Chlamydiosis	–	–	–	–	–	4 (1.8%)	–	–	4 (0.6%)
Gastrointestinal	1 (0.7%)	1 (0.9%)	–	–	7 (10.8%)	19 (8.6%)	2 (6.5%)	–	30 (4.7%)
Regurgitation syndrome, cause not identified	–	–	–	–	5 (7.7%)	8 (3.6%)	–	–	13 (2%)
Parasitism	–	–	–	–	–	6 (2.7%)	1 (3.2%)	–	7 (1.1%)
Bacterial ingluvitis, ventriculitis, cloacitis	1 (0.7%)	–	–	–	1 (1.5%)	1 (0.5%)	–	–	3 (0.5%)
Avian gastric yeast	–	1 (0.9%)	–	–	1 (1.5%)	2 (0.9%)	–	–	4 (0.6%)
Viral enteritis, suspected	–	–	–	–	–	1 (0.5%)	–	–	1 (0.2%)
Enteritis, cause not identified	–	–	–	–	–	1 (0.5%)	1 (3.2%)	–	2 (0.3%)
Coelomic	4 (3%)	–	1 (3.1%)	–	3 (4.6%)	4 (1.8%)	–	–	12 (1.9%)
Yolk coelomitis	2 (1.5%)	–		–		3 (1.4%)	–	–	5 (0.8%)
Coelomitis, cause not identified	–	–	1 (3.1%)	–	1 (1.5%)		–	–	2 (0.3%)
Bacterial coelomitis	2 (1.5%)	–		–	2 (3.1%)	1 (0.5%)	–	–	5 (0.8%)
Hepatic	–	–	2 (6.3%)	–	–	–	–	–	8 (1.2%)
Hepatitis, cause not identified	–	–	1 (3.1%)	–	–	5 (2.3%)	–	–	6 (0.9%)
Bacterial	–	–		–	–	1 (0.5%)	–	–	1 (0.2%)
Fungal	–	–	1 (3.1%)	–	–		–	–	1 (0.2%)
Integumentary	4 (3%)	–	–	–	–	2 (0.9%)	–	–	6 (0.9%)
Bacterial dermatitis, omphalitis	4 (3%)	–	–	–	–		–	–	4 (0.6%)
Dermatitis, cause not identified	–	–	–	–	–	2 (0.9%)	–	–	2 (0.3%)
Renal	–	–	–	–	–	2 (0.9%)	1 (3.2%)	–	3 (0.5%)
Glomerulonephritis	–	–	–	–	–	2 (0.9%)	1 (3.2%)	–	3 (0.5%)
Ocular	–	1 (0.9%)	–	–	–	1 (0.5%)	–	–	2 (0.3%)
Musculoskeletal	–	–	–	–	1 (1.5%)	1 (0.5%)	–	–	2 (0.3%)
Bacterial osteomyelitis	–	–	–	–	1 (1.5%)	1 (0.5%)	–	–	2 (0.3%)
Cardiovascular	–	–	–	–	–	–	1 (3.2%)	–	1 (0.2%)
Bacterial pericarditis	–	–	–	–	–	–	1 (3.2%)	–	1 (0.2%)
Splenic	–	1 (0.9%)	–	–	–	–	–	–	1 (0.2%)
Bacterial splenitis	–	1 (0.9%)	–	–	–	–	–	–	1 (0.2%)
Total infectious deaths	20 (14.9%)	38 (34.2%)	13 (40.6%)	16 (47.1%)	27 (41.5%)	101 (45.7%)	8 (25.8%)	3 (23.1%)	226 (35.3%)

BFDV, Psittacine beak and feather disease virus; NS, not stated.

Bolded data are subtotals for each subcategory.

Table 4.

Non‐infectious causes of death in captive orange‐bellied parrots that died between July 2010 and June 2019 and had a discoverable necropsy record. Data in parentheses are percentage of all cases in that age class

Non‐infectious cause of death	Hatchlings	Nestlings	Fledgling	Juvenile	Subadult	Adult	Geriatric	NS	Total cases
Cases	134	111	32	34	65	221	31	13	641
Trauma	4 (3%)	7 (6.3%)	3 (9.4%)	3 (8.6%)	11 (16.9%)	15 (6.8%)	2 (6.5%)	1 (7.7%)	46 (7.2%)
Degenerative diseases	2 (1.5%)	1 (0.9%)	–	2 (5.7%)	1 (1.5%)	9 (4.1%)	11 (35.5%)	2 (15.4%)	28 (4.4%)
Nutrition or feed‐associated	7 (5.2%)	5 (4.5%)	3 (9.4%)	2 (5.7%)	–	3 (1.4%)	–	–	20 (3.1%)
Metabolic bone disease	2 (1.5%)	4 (3.6%)	–	–	–	1 (0.5%)	–	–	7 (1.1%)
Vitamin D intoxication	5 (3.7%)	–	1 (3.1%)	–	–	–	–	–	6 (0.9%)
GIT perforation by plant matter	–	–	1 (3.1%)	1 (2.9%)	–	2 (0.9%)	–	–	4 (0.6%)
Food aspiration with suffocation	–	–	1 (3.1%)	–	–	–	–	–	1 (0.2%)
Crop impaction	–	1 (0.9%)	–	–	–	–	–	–	1 (0.2%)
Osteopaenia	–	–	–	1 (2.9%)	–	–	–	–	1 (0.2%)
Veterinary intervention	–	–	–	3 (8.6%)	–	4 (1.8%)	2 (6.5%)	–	9 (1.4%)
Inappropriate parenting	5 (3.7%)	3 (2.7%)	–	–	–	–	–	–	8 (1.2%)
Developmental	–	1 (0.9%)	1 (3.1%)	–	1 (1.5%)	–	–	–	3 (0.5%)
Congenital deformity (gross)	–	–	–	–	1 (1.5%)	–	–	–	1 (0.2%)
Retained yolk sac	–	1 (0.9%)	–	–	–	–	–	–	1 (0.2%)
Malformation (scissor beak)	–	–	1 (3.1%)	–	–	–	–	–	1 (0.2%)
Neoplasia	–	–	–	–	–	2 (0.9%)	–	–	2 (0.3%)
Environmental	–	–	–	–	–	1 (0.5%)	–	–	1 (0.2%)
Adverse environmental conditions	–	–	–	–	–	1 (0.5%)	–	–	1 (0.2%)
Intoxication	–	–	–	–	1 (1.5%)	–	–	–	1 (0.2%)
Lead toxicosis	–	–	–	–	1 (1.5%)	–	–	–	1 (0.2%)
Total non‐infectious deaths	18 (13.4%)	17 (15.3%)	7 (21.9%)	10 (28.6%)	14 (21.5%)	34 (15.4%)	15 (48.4%)	3 (23.1%)	118 (18.4%)

GIT, gastrointestinal; NS, not stated.

Table 5.

Diseases of most concern causing mortality, listed by captive facility (A to F), in captive orange‐bellied parrots that died between July 2010 and June 2019 and had a discoverable necropsy record. Data in parentheses are percentage of cases at that site

	Captive facility						Total cases
	A	B	C	D	E	F
Cases at each site	265	167	120	57	17	15	641
Respiratory infections, total	72 (27.2%)	12 (7.1%)	12 (10%)	–	3 (17.6%)	1 (6.7%)	100 (15.6%)
Respiratory aspergillosis	54 (20.4%)	6 (3.6%)	7 (5.8%)	–	3 (17.6%)	–	70 (10.9%)
Pseudomonas aeruginosa sepsis	–	18 (10.7%)	–	–	–	–	18 (2.8%)
Mycobacteriosis	1 (0.4%)	11 (6.5%)	1 (0.8%)	–	2 (11.8%)	–	15 (2.3%)
GIT syndrome, cause unknown	13 (4.9%)	–	–	–	–	–	13 (2%)
BFDV‐associated	–	7 (4.8%)	–	–	2 (11.8%)	–	9 (1.4%)
Trauma	19 (7.2%)	5 (3%)	5 (4.2%)	10 (17.5%)	3 (17.6%)	4 (26.7%)	46 (7.2%)
Chlamydiosis	–	–	–	3 (5.3%)	–	1 (6.7%)	4 (0.6%)
Cause of death not determined	105 (39.6%)	79 (47%)	75 (62.5%)	32 (56.1%)	2 (11.8%)	4 (26.7%)	297 (46.3%)

BFDV, Psittacine beak and feather disease virus; GIT, gastrointestinal.

Infectious causes of death

Respiratory infections (including pneumonia, air sacculitis and sinusitis) were the most common cause of death (16%, 100/641) and were dominated by aspergillosis, which caused 70% of respiratory infections and 11% of deaths overall (see Figure 3). The majority of cases of respiratory aspergillosis were seen at Facility A (77%, 54/70) with relatively few at other sites and no cases reported from two sites (Facility F and Facility D). Aspergillosis was more commonly reported in the younger age classes, particularly nestlings, where it accounted for 24% of all nestling deaths, compared with 9% of adult deaths. Other respiratory infections included bacterial pneumonia and respiratory inflammation without an identifiable infectious agent. Culture results were rarely reported, though two cases of pneumonia due to Pseudomonas aeruginosa were noted in 2015, one in combination with aspergillosis. There were rare cases of other fungal infections including non‐aspergillus sinusitis and two cases of avian gastric yeast (Macrorhabdus ornithogaster) infection that were deemed severe enough to cause death.

Figure 3.

Lung from a captive orange‐bellied parrot showing marked congestion and a granuloma with a rim of pale histiocytes (arrow) and a necrotic centre (*). Haematoxylin and eosin, 100× magnification. Inset: Same case, showing numerous pale basophilic fungal hyphae within the granuloma core amid haemorrhage and necrotic granulocytic and nuclear debris. Haematoxylin and eosin, 400× magnification.

Other bacterial infections caused 11% (70/641) of deaths and included septicaemia and internal organ infections such as hepatitis, coelomitis and splenitis. Culture was rarely reported, but included single reports of Streptococcus spp., Enterococcus faecalis and Nocardia spp. each causing disseminated infection. Bacterial septicaemia due to Pseudomonas aeruginosa was notable due to an outbreak at Facility B in the 2016/17 season, which accounted for at least 33% (17/52) of all infections at that site during the 2016/17 year. This may be an underestimate, as classification was primarily based on positive liver culture for P. aeruginosa, and not all mortalities within the outbreak period could be conclusively assigned a cause of death based on the necropsy record alone. This mortality event was traced to contaminated sprouted seed fed to wild and captive OBPs. ²³

Mycobacteriosis was similarly an uncommon cause of death overall (2%, 15/641) but almost all cases were seen at one site (Facility B), with sporadic cases seen at three other sites. All mycobacteriosis cases had granulomatous hepatitis and all but one also had granulomatous splenitis (see Figure 4). Lesional mycobacterial PCR results were included with the necropsy record in six cases, with all results showing Mycobacterium avium. An additional case was identified as M. branderi by culture (personal comment, D McLelland). Of the four cases of chlamydiosis, one from Facility F was confirmed by PCR of the liver and had typical histological lesions of lymphohistiocytic hepatitis. The other three cases were from Facility D and did not have histology performed; rather, the diagnosis was noted in the clinical history without elaboration of any gross lesions or ancillary testing. Yolk coelomitis was included in the infectious classification as two of the five cases had visible bacteria histologically.

Figure 4.

Liver from a captive orange‐bellied parrot showing a granuloma with a rim of pale histiocytes (arrow) and a necrotic centre (*). Haematoxylin and eosin, 100× magnification. Inset: Acid‐fast staining of the same lesion, showing numerous red‐staining acid‐fast bacilli within the necrotic core and adjacent histiocyte cytoplasm. Ziehl‐Neelsen, 1000× magnification.

Viral infections were an uncommon cause of death (2%, 10/641) and nine of these cases were birds that were euthanased due to clinical or suspected BFDV infection. This included seven juvenile or subadult birds from Facility B in 2019, with no clinical notes provided, and two adult birds from Facility E who were suspected to be chronic BFDV carriers due to serological testing. Ancillary BFDV serology or molecular testing was not included in the necropsy records. Histology was performed in the Facility E cases and three of the Facility B cases, the latter also showing positive BFDV immunohistochemistry, and two of which had folliculitis. Immunohistochemistry was not performed or was not recorded in the necropsy record for the other cases. The suspected BFDV carriers from Facility E had nonspecific lymphocytic inflammation including dermatitis, nephritis, pancreatitis and hepatitis, and one had minimal heterophilic feather pulpitis; the necropsy report had no record of immunohistochemistry to confirm if these lesions were directly associated with BFDV infection. One of these cases also had ascariasis and renal collecting duct inclusions as detailed below.

A single case of suspected viral enteritis was seen in an adult bird that was found dead in 2019. Histopathology showed a moderately severe segmental necrotising enteritis with reparative glandular hyperplasia and lymphoplasmacytic inflammation. Inclusion bodies were not seen. Lymphohistiocytic hepatitis and pancreatitis were also noted. While the histological pattern was considered most consistent with viral injury, next generation sequencing at the Australian Centre for Disease Preparedness (formerly the Australian Animal Health Laboratory) excluded known viral pathogens and viral infection could not be confirmed.

Intranuclear inclusion bodies were noted in the renal collecting duct epithelium and/or hepatocytes in 19 birds out of the 490 for which histology was performed, being identified at all but one of the study sites (Table 6). These were large, glassy, lightly eosinophilic, intranuclear inclusion bodies (see Figure 5). Histologically identical inclusions have been conclusively identified by others as Psittacine Siadenovirus F (previously known as Psittacid Adenovirus‐2). ²⁴ In the absence of routinely available testing for this virus, these inclusions were presumed to represent Psittacine Siadenovirus F infection based in their consistent appearance. Inclusions were renal in 16/19, hepatocellular in 2/19, and both renal and hepatic in 1/19 cases. These inclusions did not have associated renal or hepatic necrosis or inflammation and thus there was no direct evidence of viral infection contributing to death, which was instead attributed to a range of infectious causes, most commonly mycobacteriosis. All the mycobacteriosis cases with inclusions were from one site (Facility B) between 2017 and 2019. Five additional cases of mycobacteriosis at Facility B and the four reported across other facilities did not have inclusions reported. In the two cases with inclusions where the cause of death was undetermined, both cases were found dead and showed histological lesions of mild, nonspecific hepatic necrosis. One was a young adult with mild heterophilic proventriculitis and hepatic vacuolar degeneration, and the other was a juvenile with mild splenic lymphoid hyperplasia and renal collecting duct epithelial mineralisation; no lesion was of sufficient severity to account for the sudden death or attribute directly to viral infection.

Table 6.

Cause of death in cases with histological renal or hepatic Psittacine Siadenovirus F‐consistent inclusions, in captive orange‐bellied parrots that died between 2010 and 2019 and had a discoverable necropsy record

Cause of death where adenoviral‐like inclusions were noted at histology	Juvenile	Subadult	Adult	Geriatric	Total
Mycobacteriosis	‐	‐	6	‐	6
Aspergillosis	‐	‐	3	‐	3
Bacterial septicaemia, ingluvitis or pneumonia	‐	‐	3	‐	3
Bacterial septicaemia due to Pseudomonas aeruginosa	‐	1	‐	‐	1
Cestodiasis	‐	‐	‐	1	1
Euthanasia (chronic BFDV carrier)	‐	‐	1	‐	1
Visceral gout	‐	‐	1	‐	1
Pulmonary oedema, cause not identified	‐	‐	1	‐	1
Cause of death not determined	1	‐	1	‐	2
Total deaths	1	1	16	1	19

Bolded data are subtotals for each subcategory.

Figure 5.

Kidney from a captive orange‐bellied parrot showing numerous eosinophilic intranuclear inclusion bodies (black arrows) within collecting duct epithelial cells. Haematoxylin and eosin, 400× magnification.

Parasitic diseases were an uncommon cause of death (less than 1%, 7/641). Most cases were due to heavy burdens of ascarids, including those causing intestinal perforation; all cases occurred in 2017 or prior. An ascarid burden was noted to be present but not assigned as a cause of death in a further 10 cases. A single case died due to cestodiasis (1/641).

A syndrome of regurgitation and diarrhoea was noted at Facility A in November 2015, which accounted for 5% (13/265) of deaths from that site. Subadults and adults were affected, with a typical clinical history of either sudden death or acute signs of regurgitation and diarrhoea followed by death within days despite supportive treatment. Histological lesions were mild and nonspecific, including splenic reticuloendothelial cell hyperplasia, mild hepatic necrosis and a range of other mild changes. One case with the typical clinical syndrome had bacterial sepsis, with Proteus mirabilis cultured from a liver abscess. Another had a coelomic bacterial abscess; culture was not performed. Two birds showed heavy growth of E. coli with faecal culture. The serotype was O169:H8, and no known avian pathogenic E. coli virulence factors were identified. Five birds had negative chlamydial PCR and one was PCR‐positive for BFDV. Moderate numbers of these cases (8/13) showed mild to moderate hepatocellular anisokaryosis and anisocytosis (see Figure 6). The ultimate cause of the mortality event was not determined.

Figure 6.

Liver from an orange‐bellied parrot that died during a mortality event of unknown cause associated with regurgitation and sudden death. Hepatocytes show mild cytoplasmic vacuolar degeneration, moderate anisokaryosis and occasional karyomegaly (black arrow), without significant other degenerative or inflammatory changes. This degree of anisokaryosis was not uncommon in OBPs affected by a range of disease processes and is of unknown significance. Haematoxylin and eosin, 400× magnification.

Non‐infectious causes of death

The most common non‐infectious cause of death was trauma, which accounted for 7% (46/641) of deaths and was commonly due to head trauma and limb fractures (91%, 42/46), but also included snake predation (9%, 4/46). Degenerative conditions accounted for a low proportion (4%, 28/641) of cases and included hepatic necrosis without identified cause, hepatopathies, pulmonary oedema of unknown cause, renal disease and osteoarthritis. Nutritional, feed‐associated and behavioural conditions accounted for 4% (28/641) of cases and included inappropriate parenting, gastrointestinal perforation by ingested bedding fibres or seeds (see Figure 7), metabolic bone disease, gout and vitamin D intoxication. Deaths associated with veterinary interventions (anaesthesia and blood collection) were very rare (1%, 9/641), as were developmental conditions such as growth deformities (less than 1%, 3/641).

Figure 7.

Gross specimen of ventriculus from a captive juvenile orange‐bellied parrot. The ventriculus has been cut in half to show perforation of the wall by an ingested plant fibre presumed to be bedding material (white arrows). The ventricular submucosa and tunica muscularis are expanded by localised haemorrhage and inflammation. The background of the image has been digitally removed for clarity. Bar = 2 mm.

A single case of lead toxicosis was found in a 9‐month‐old subadult bird from Facility B with no clinical history recorded. Histology showed renal tubular necrosis and the necropsy record included tissue testing showing markedly elevated kidney lead levels.

Only two cases of neoplasia were noted, one being lymphoproliferative disease in a 7‐year‐old bird and one case of suspected renal neoplasia in a 7‐year‐old bird noted at gross necropsy, where the histology report was not included in the necropsy record.

Discussion

Mortality reviews are a key methodology to aggregate and interrogate health data for captive animals and have been valuable in other threatened species recovery programs. ⁸ In this review, infectious disease was the most common cause of death in all ages except geriatric birds, but a substantial proportion of cases in some age classes could not be assigned a cause of death.

The most common pathogen‐specific cause of death over the study period was respiratory aspergillosis. Aspergillosis is an environmentally acquired fungal infection, with organic bedding and litter a source of spores for aerosolisation and inhalation. ²⁵ Seabirds and waterfowl are particularly susceptible, with aspergillosis the cause of death of 2%–61% of captive and 3%–31% of free‐living birds across a range of species in a recent review of 30 mortality studies. ²⁵ Two reviews focusing on captive psittacines or psittaformes identified aspergillosis as the primary cause of death in 1.5% and 14.2% of captive psittaformes deaths due to aspergillosis. ²⁶

Aspergillosis has historically received only brief attention in OBP disease risk advice. ⁵ These results help to further define the nature of this disease threat to the captive population, with overall aspergillosis mortality accounting for around 10% of deaths, but reaching over 20% in nestlings and juveniles, and cases highly concentrated at one site during the study period. This is a higher proportion than the only other reported rate of aspergillosis in captive OBPs, which caused 6.5% of deaths over a 5‐year period at one facility.¹⁰ This suggests site‐specific and age‐specific factors are likely to influence disease expression. Aspergillosis risk factors include conditions that promote environmental fungal spore load such as moisture in bedding and litter; warm, humid conditions; poor ventilation; husbandry factors leading to stress or immunocompromise; and possibly ground foraging behaviour. ²⁵ , ²⁷ Additionally, it is unknown how site‐specific factors might interact with host factors to promote aspergillosis, such as the markedly reduced diversity in immune function genes in the OBP. ²⁸ , ²⁹ While most avian aspergillosis involves multiple strains, ²⁷ it is possible for a single strain to cause a mortality outbreak, as shown by a significant wild mortality event in the critically endangered kākāpō parrot in 2019, attributed to a single strain of Aspergillus fumigatus by whole genome sequencing. ³⁰ Investigation of pathogen‐specific and site‐specific risk factors for clinical aspergillosis in captive OBPs represents a potential avenue for future work, though the genetic determinants of Aspergillus spp. virulence are incompletely understood. ³⁰

Captive sites should ensure that housing and husbandry protocols consider aspergillosis‐specific risk factors. Given that respiratory disease was a common cause of death overall in this study period, protocols and housing design to avoid aspergillosis risk factors are likely to improve the overall respiratory disease burden.

Mycobacteriosis, which causes a progressive wasting syndrome, has been identified as a disease of concern in the captive OBP population. ⁵ Avian infections are most often due to M. avium subsp. avium or M. genavense, both of which have been reported in captive OBPs. ¹⁷ , ³¹ Transmission is primarily faeco‐oral and organisms can persist in diverse environmental niches, meaning infection can persist in captive facilities despite best efforts at environmental decontamination and identifying carriers. Infection can be clinically silent until lesions are well progressed, making timely eradication difficult and costly.

Mycobacteriosis caused a low proportion of deaths in the current study, but disease was concentrated at one site over the study period. The reasons behind this distribution remain unknown and represent an area of future study to disentangle host, environment and pathogen risk factors. The insidious clinical course and difficulty of eradication of mycobacteriosis mean that early identification of infected birds and bird‐to‐bird transmission should be a priority, though the lack of highly sensitive and specific pre‐mortem diagnostic tests makes this challenging, representing another avenue for more research.

Early in the captive‐breeding program, PBFD was reported to be the main cause of mortality, accounting for 14 of 24 deaths (58%) from 1985 to 1989. ⁹ This review found otherwise: of 641 records, only seven young birds were euthanased at one site due to an outbreak of PBFD, while another two birds were euthanased as test results indicated a carrier state for BFDV. Beyond this, there are limited published reports of overt clinical disease in BFDV‐infected OBPs (mortality of wild nestlings in 2007–2008). Viral phylogenetic analysis has found a host‐adapted strain of BFDV in captive OBPs, as opposed to viral spillover from other hosts as noted in the wild population. ¹¹ , ¹⁵ , ³² , ³³ , ³⁴ BFDV is known to cause immunosuppression, which may impact health and productivity through increased susceptibility to secondary infections, ³⁵ however, this has been poorly studied. There are no published reports where histology has been performed on clinically BFDV‐affected or BFDV‐positive dead OBPs to screen for concurrent illnesses. The cases seen here had a range of histological lesions including viral antigen‐associated folliculitis in two birds, but histology was inconsistently performed; the necropsy records available did not include relevant information about clinical history, and only 3 of the 9 cases recorded within the necropsy record how the diagnoses was confirmed. In cases where further testing to localise viral antigen to lesions was not performed, it is not possible to know if the histological lesions reflected BFDV infection. Any further study of the significance of BFDV‐related illness in the metapopulation, including secondary infections, would therefore require more extensive data discovery than necropsy reports alone.

There are currently no publications which detail any relationship between BFDV infection prevalence and long‐term breeding success or current mortality trends in the captive OBP population. Genetic differences between BFDV strains in captive and wild OBPs indicate that infection in wild OBPs results from spillover from other parrot species rather than endemicity or introduction from the captive population. ¹¹ , ³³ The current impact of BFDV infection on overall mortality in the captive metapopulation and potential risk of future introduction to the wild therefore remains to be adequately quantified, though the potential for BFDV‐associated mortality events remains a recognised threat. It is also notable that loss of endemicity of OBP‐associated BFDV in the captive population could lead to negative conservation outcomes where immunologically naïve captive birds are exposed to BFDV spillover in the wild or through exposure to wild birds while in captivity, adding to the threat. Correlation of long‐term serological and molecular testing with detailed cause of death data, such as that presented here, may further refine the role of BFDV as a risk factor for captive mortality.

The number of chlamydiosis deaths found here (less than 1% of deaths) is lower than that noted in a previous report of OBP mortality over a 5‐year period at one facility, where it caused 6/77 (7.7%) of deaths. ¹⁰ Chlamydiosis has been identified as a potential threat to captive and wild OBPs and is an important zoonosis. ⁵ Avian chlamydiosis is typically caused by Chlamydia psittaci and affects wild and captive birds of many species, with psittacines particularly susceptible. ³⁶ Chlamydia psittaci prevalence in wild Australian birds has been reported as 0.7% and 3%. ³⁷ , ³⁸ Transmission can be direct by aerosolisation of bacteria in faeces, oculonasal and respiratory secretions. ³⁹ Captive OBP housing includes open‐air aviaries as a necessary component of environmental enrichment, natural behaviour promotion and to preclude immune naivete on release. ⁵ As such, chlamydiosis remains a threat to captive OBP populations despite the low numbers of deaths reported in our study.

The low rate of chlamydiosis seen here may reflect the continuous improvements to biosecurity as the recovery program has evolved. It is also possible that cases of chlamydiosis could be undetected if necropsy investigations do not include histology or proactive chlamydial PCR screening. Typical gross lesions of hepatomegaly and splenomegaly are not always present, and disease may be present as a subclinical chronic or carrier state. ³⁶ Importantly, diagnosis of chlamydiosis in three of the four cases noted here was not clearly supported by documentation of ancillary testing within the necropsy record. Some caution is therefore warranted in accepting these as true cases. Gross and histological suspicion for chlamydiosis should be confirmed by PCR testing, as organisms are not always visible microscopically. As recent studies have demonstrated a range of chlamydial species and strains in Australian wild birds detectable by PCR, ³⁸ , ⁴⁰ sequencing and strain typing would allow a deeper understanding of the origin of future chlamydial infections in OBPs.

Viral inclusions histologically typical of a previously identified avian adenovirus were seen in a low proportion of birds. Identical inclusions have been characterised by electron microscopy and molecular studies as Psittacine Siadenovirus F (previously known as Psittacid Adenovirus‐2). ²⁴ This virus has been confirmed by faecal or cloacal PCR testing as endemic in all OBP captive‐breeding facilities, with shedding rates up to 77% and has also been found in post‐release captive‐bred birds. ²⁴ , ⁴¹ Testing for this virus is not routinely available, so identification currently relies on histological detection of inclusions. While these inclusions bear some histological similarity with avian polyomaviral and herpesviral inclusions, previous work by others has failed to demonstrate those viruses in the OBP population. ⁴² , ⁴³

Despite the reported high prevalence, in the present study less than 4% of captive birds that had histology performed were reported to have Psittacine Siadenovirus F‐consistent renal or hepatic viral inclusions. In two separate captive OBP mortality events, the presence of this virus was not thought to directly contribute to disease. ²⁴ Pathogenicity of this adenovirus is thus yet to be demonstrated. In our study, the presence of inclusions was not associated with a consistent histological pattern of lesions, and affected birds died from a range of seemingly unrelated causes, though a high proportion had bacterial infection or aspergillosis. An interesting finding was the relatively high proportion of mycobacteriosis cases that also had Psittacine Siadenovirus F‐consistent renal inclusions, which were clustered at one site. Whether this reflects coincidence, the higher prevalence of mycobacteriosis at that time, viral‐associated immunosuppression, or some other interaction between viral and mycobacterial infection would need further research to determine. Additionally, the sensitivity of inclusion detection at histology for identifying adenovirus‐infected birds is unknown, meaning that other birds in the dataset could have been infected but not had detectable inclusions. These results present an opportunity for retrospective or observational studies to compare the morbidity or mortality in adenovirus‐positive and uninfected birds, though the sensitivity and specificity of screening testing by faecal or cloacal PCR is also yet to be quantified.

This review shows a diverse range of diseases that are broadly consistent with existing assessments of causes of mortality in captive orange‐bellied parrots, which have identified PDFD, mycobacteriosis, chlamydiosis, adenoviral infection, aspergillosis, intestinal helminth impaction, trauma and handling accidents, predation, zinc toxicity and Pseudomonas septicaemia. Three previously reported causes of death were not noted here. Zinc intoxication was reported as a likely cause of mortality and reduced fertility in a previous review of 5 years of deaths at one facility. ¹⁰ No cases of zinc toxicity were noted in this study, and the standout lesion in the historical cases (renal tubular necrosis) was rarely reported as a cause of death; this suggests that zinc intoxication is not currently a mortality risk to the captive population.

The second previously reported disease threat is a mortality event due to aflatoxicosis at Facility A in 2015, reported in a 2020 stakeholder report. ⁵ These cases are the same reported here with a syndrome of acute regurgitation and diarrhoea, where an ultimate cause was not identified. In the absence of a clear cause, the presence of mild to moderate hepatocellular anisocytosis and anisokaryosis in some affected birds had raised initial thought for a hepatotoxic agent such as aflatoxin; however, the degree of change was mild, not present in all birds, and the typical lesions of acute aflatoxicosis including severe liver necrosis, icterus and tissue haemorrhage ⁴⁴ were not present, meaning this possibility was ultimately discarded by the original pathologists (ED and CB). Hepatocellular anisokaryosis and anisocytosis have been observed by the authors as a notable but nonspecific finding in a range of unrelated pathological syndromes in OBPs. Ancillary testing, including testing of food for aflatoxins, infectious disease screening and heavy metal testing, was not able to diagnose a specific cause of this event (P Eden, pers. Comm.) ⁵

The third previously reported disease is an informal report of high captive OBP nestling mortality reported in 2005 that could not be definitively attributed to a particular cause, though a non‐Pacheco's herpesvirus of uncertain significance was identified by molecular testing. ¹⁵ No lesions suggestive of herpesvirus were found in our study; where intranuclear inclusion bodies were seen, they were not directly associated with typical herpesviral pathology, and were reasonably attributable to Psittacine Siadenovirus F, as discussed above. Serological surveys, including a survey within the study period, have not identified herpesviruses in the wild or captive OBP population. ⁴² , ⁴³ Herpesviral infection has thus not been demonstrated to be a current contributor to mortality in the captive OBP population.

Diagnostic success is not always explicitly reported in mortality reviews. We believe this metric is a valuable aid to interpreting mortality surveys, as it helps not only to contextualise the relevance of the findings, but also to support initiatives to improve diagnostic investigations and manage expectations where resources are limited. The diagnostic success rate reported here (54%) is lower than similar mortality surveys in eagles, loris and pottos, reptiles and Hawaiian geese, which have rates of unknown cause of death ranging from 9.7% to 25%. ⁴⁵ , ⁴⁶ , ⁴⁷ In the only other published OBP mortality investigation, diagnostic success was not clearly stated. ¹⁰ Cause of death was determined in at least 11 of 77 birds examined (14%) and otherwise proposed to be linked to zinc intoxication, and the high frequency of birds showing no specific gross or histological lesions was noted. ¹⁰

The relatively low rate of diagnostic success in OBPs may be explained by the high proportion of neonates and young animals, the number of animals found dead and the low rates of inclusion of histological examination in some sites. Inclusion of histopathology in diagnostic investigations was associated with an improvement in diagnostic success across most age classes. Histology allows the characterisation of distinct pathological processes that may share gross features, and reveals lesions that are only apparent microscopically, including background or subclinical abnormalities that may be contributing to disease. Histology adds a significant cost to mortality investigations and may not always be economically feasible, particularly where there are large numbers of animals, funding is scarce, or the seasonality of mortality spikes leads to uneven distribution of annual costs. Given the benefits to diagnostic sensitivity, we encourage providers of veterinary services to wildlife facilities to recognise the value of histopathology as a standard component of necropsy investigations, where possible.

Younger birds had a particularly low rate of diagnostic success, and a high proportion of hatchlings and nestlings were found dead with minimal lesions. Sudden death can suggest peracute disease where gross and histological lesions may be absent or have not had time to develop to diagnostic thresholds. ⁴⁸ Differentials in very young birds could include environmental stressors such as excess heat or cold; peracute and undetected infection such as sinusitis; undetected trauma; inadequate yolk or other undetected developmental or incubation disruption; genetic issues; or mismothering leading to inadequate nutrition or exposure. ⁴⁹ Additionally, rapid autolysis in birds can mask subtle lesions of peracute illness. ⁵⁰ These factors can make it challenging to determine the cause of death for younger age classes and likely contribute to why sites with a breeding focus rather than housing or ranching have a lower overall diagnostic success rate, as they have more birds in these age groups. Health in these younger age classes is particularly important to the conservation effort. Nestling body condition, which is assumed to correlate with health, is positively associated with wild survivorship in released captive‐bred birds, ⁵¹ and increased nestling survival leads to more birds for potential breeding or release. Additionally, previous studies of factors affecting captive neonatal OBP survival have been confounded by husbandry changes following disease events ⁵¹ and the effects of the disease event itself, for example, on body weight.

The differences in disease occurrence and diagnostic success between captive sites reinforces the need for a standardised approach to necropsy investigations across the metapopulation. This would promote a similar level of diagnostic sensitivity and specificity and allow an accurate interrogation of risk factors. Recommendations for a standardised necropsy protocol with input from the VTRG and supporting diagnosticians would minimally include a baseline gross necropsy protocol, standardised tissue collection for histopathologic examination (including any observed lesions), and appropriate collection and storage of tissue samples for ancillary testing or archiving. Protocols may offer options to consider factors such as resourcing and the age at death (e.g., hatchling or nestling versus fully grown OBP). Standardised reporting could also be provided with a template necropsy record in paper or digital form and include collection of relevant clinical and animal husbandry information. Acknowledging that access to prompt necropsy by a veterinarian may vary across sites, an abbreviated protocol that prioritises selected critical sample collection and then prompt formalin‐fixation of an opened carcass for histology would also be of value.

Future mortality reviews would also benefit from standardisation of necropsy reports to include an explicit cause of death in standardised terminology, and standardisation of records management with a search and collation function for the cause of death field. A recent review of captive animal mortality survey methodology recommended using the ZIMS pathology module to record disease information in a standardised manner, ⁸ and our experience here supports the utility of such a standardised recording methodology. Additionally, incorporating cause of death with existing centralised databases for bird records would be the natural consequence of standardised mortality reporting.

This mortality review provided useful data for the OBP Recovery Program WDRA process. The diseases nominated by the WDRA as priority infectious disease risks posed by bird translocations (mycobacteriosis, chlamydiosis and BFDV) were able to be quantified and contextualised by the review findings. While the most common causes of death (aspergillosis and diverse bacterial infections) were not as relevant to the WDRA's translocation focus, collation and review provide invaluable site‐specific information that can prompt management changes and stimulate new research to reduce disease burdens and ultimately improve survival in captivity. Biosecurity and preventative health initiatives can then be justified and refined on the basis of these findings. This is a key reason for publication of the mortality reviews that underpin disease risk assessments in captive and wild animals.

The clustering of diseases within sites highlights the importance of a multi‐site approach to captive breeding for managing impacts of disease risks. Despite best efforts, disease incursions can never be completely prevented. A multi‐site approach to managing mortality risks can buffer the metapopulation against disease mortality in any particular site, while also offering the potential for analysis of site‐specific risk factors to help determine their role in disease development and breeding success.

This study has several limitations. Our data do not capture all deaths in the captive population over the study period, as the number of discoverable necropsy records was fewer than the recorded deaths in ZIMS population summaries. Given the limitations on data collection, deaths that occurred prior to 1 July 2010 are not captured in this study. This amounts to 626 birds across the entire captive population since 1988, including those housed at additional facilities. ²⁰ It is therefore possible that some mortality threats are not captured in the current data, including historical threats that could still pose a risk to the contemporaneous population. Opportunities to compare current and past mortality patterns are also limited, aside from with reference to the few published disease reports as noted in the above discussion. Interrogation of pre‐2010 necropsy records remains a potential avenue for future work, and standardisation of necropsy record keeping to facilitate rapid collation of cause‐of‐death data would allow more efficient mortality review in the future.

Additionally, the extent of necropsy investigation and data recording was inconsistent across sites. Variation between clinicians and pathologists in lesion interpretation is an inherent limitation, particularly where pathogen‐specific ancillary testing is limited. In our experience, a lack of standardised necropsy sampling protocol can also lead to variation in diagnostic sensitivity. For example, failure to submit the proventriculus‐ventriculus isthmus can limit the detection of avian gastric yeast, ⁵² and failure to include the head in histopathology samples may exclude lesions of fatal sinusitis. Necropsy records where histology or ancillary testing have been performed but not added to the record limited the ability to accurately assign a cause of death in some cases, as did cases where the cause of death was not explicitly assigned by the submitting veterinarian or reporting pathologist. Without integration of other clinical records, the necropsy records alone may therefore be insufficient at capturing all contributors to death at both an individual and population level.

As our dataset will naturally be biased towards diseases that cause death, subclinical or sublethal illness affecting fertility and fitness for release may not be well‐represented here. For example, diseases such as lead intoxication, represented here by a single confirmed case, are difficult to accurately contextualise as risk factors without further clinical details about the circumstances of intoxication and any follow‐up testing of at‐risk birds.

As the results are dominated by sites with more birds, diseases that disproportionately affected specific sites, such as aspergillosis and mycobacteriosis, will skew the results and care should be taken when interpreting overall disease frequency for this reason. Comparing these results to overall and site‐specific population numbers would yield a baseline mortality rate for each process that could be useful for more accurate tracking of changes in disease prevalence over time, but is beyond the scope of this review. Additionally, while this historical data confirms susceptibility of captive OBPs to the disease processes listed, these may not predict specific rates of disease in the contemporary OBP captive population.

Mortality results were presented by age class to allow age‐specific differences to be captured. Published captive OBP resources typically do not explicitly define age classes in days of age, ⁵ , ²² so the scheme here was based on direct advice from the captive program. It is possible that alternative age classifications based on different features such as anatomy, behaviour or natural ecology could alter these findings, though it would be unlikely to change the general finding that a majority of deaths occurred within the first year of life.

Our study provides observational data that demonstrates the value and limitations of passive disease surveillance in captive‐breeding programs. Mortality review datasets are suitable for further interrogation to explore potential relationships between diseases and a broad range of potential risk factors. Implementation of standardised and thorough necropsy investigations would provide a mechanism to track disease prevalence and detect new and emerging diseases affecting captive survival. Such considerations go hand‐in‐hand with the ongoing ecological, genetic, husbandry and breeding recovery activities, and help contextualise survival data in the captive metapopulation. Program managers continue to work with the OBP Veterinary Technical Reference Group and Captive Management Group to improve program operations and outputs.

Conflicts of interest

Paul Eden is a current member of the OBP VTRG. The opinions stated here belong to the authors and are not the opinions of the OBPRT, Zoos Victoria or the VTRG.

Data availability statement

Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the author(s) with the permission of orange‐bellied parrot recovery team.