Mortality Associated with Bushfire Smoke Inhalation in a Captive Population of the Smoky Mouse (Pseudomys fumeus), a Threatened Australian Rodent

Southern and eastern Australia has been subjected to unprecedented bushfires between September 2019 and February 2020 (Nolan et al. 2020; Yu et al. 2020), burning millions of hectares and causing extensive loss of property and human life. At least one billion wild birds, mammals, and reptiles are estimated to have perished in these fires (Loeb 2020), which are thought to have affected 113 threatened animal species (Australian Government 2020). While it is recognized that bushfires are capable of causing significant wildlife mortalities (Russell et al. 2003; Wallis 2013), there is scant knowledge about the pathophysiology of those mortalities. Carcasses of animals killed in bushfires are typically unable to be retrieved or are burnt after death, hindering the assignment of exact cause of death. Burn injuries are conspicuous in surviving wildlife and can be diagnosed more easily than injuries due to smoke inhalation, receiving more attention in veterinary clinical and wildlife rehabilitation literature. In human health, there is increasing concern for the role of hazardous smoke associated with bushfires as a cause of acute and chronic health effects (Yu et al. 2020). Bushfire smoke exposure from the 2019–20 Australian bushfires is thought to have caused the death of at least 417 people (Borchers Arriagada et al. 2020). The role of bushfire smoke as a cause of mortality and morbidity in wildlife is largely unknown.

Nine smoky mice (Pseudomys fumeus), a vulnerable endemic Australian rodent species (Woinarski and Burbidge 2016), died at a captive breeding facility housing 45 smoky mice located 25 km east of Canberra, Australia, between 4–31 January 2020 (Table 1). Seven of seven mice housed in two acclimatization enclosures that were continually open to environmental airflow died (mice 1–4 and 7–9). The remaining two mice were each housed in a separate indoor enclosure, each with one other mouse, with these enclosures together in a room that managed to ensure the internal temperature remained below 30 C on hotter days. Mortalities in the acclimatization enclosures occurred on 4 January (mice 1–4, including two males and a breeding pair), at which time ambient temperature in the enclosure peaked at 42.5 C with a night minimum of 18.4 C, and on 31 January (mice 7–9, which were three independent young from the breeding pair that died on 4 January), at which time ambient temperature in the enclosure reached a maximum of 42.2 C with a minimum that night of 27.3 C (Fig. 1). Mortalities in the indoor enclosures occurred on 11 and 21 January (mice 5 and 6, respectively). All mice had appeared healthy and active, had been eating well, and had healthy weights prior to their deaths. Animals observed by camera in the acclimatization enclosures left their hides in the afternoon and died suddenly while moving around the enclosure.

Extensive bushfires were present to the east and southeast of the captive breeding facility at the time of the mortalities. Two periods of particularly hazardous air quality due to bushfires were observed at the site, on 1 and 5 January, associated with east-southeasterly airflow. An air quality monitoring site in Canberra (Florey Air Quality Site, Australian Capital Territory Government), 32 km west-northwest of the facility, measured extremely high levels of particulate matter more than 50-fold the World Health Organization recommended maximum threshold (2018) and known to be hazardous to human and animal health on these dates (Fig. 1). Although air quality data are not available at the captive breeding facility itself, it is plausible that levels of particulate matter were even higher than those measured in Canberra due to the closer proximity of the facility to the bushfires and its position on an east-facing slope.

Of the mice that died, histologic examination of tissues was performed only for six. All mice were refrigerated at 4 C immediately after death for 24–72 h (Supplementary Material Table S1) and were examined grossly by the submitting veterinarian before formalin fixation of tissues. Gross findings were not provided. Lung, heart, liver, and gut were examined in all six mice, brain and kidney in all except mouse 1, and spleen in all except mouse 8 (Table S1). Trachea was examined in mice 5 and 8 and was normal. Pulmonary lesions were significant in all animals (Table 1) and included moderate to severe pulmonary congestion and edema of airways with a focal to diffuse but distinctly central distribution (Figs. 2A, S1–S4). Peripheral lung tissue was, for the most part, normal. In several animals, bronchiolar epithelial cytoplasmic blebbing was observed (Fig. S2) as well as loss of intercellular epithelial adhesion and sloughing (Fig. S1). Other findings in some mice included severe focal pulmonary hemorrhage and mild, focal interstitial pneumonia. Of the mice that died on 31 January, pulmonary infarction associated with thrombosis (with the presence of occasional microthrombi) was the likely cause of well-demarcated, lobar loss of perfusion in one animal (Figs. S4, S5), while occasional colonies of monomorphic rod bacteria were observed in luminal spaces in that and another mouse, though these were not associated with any inflammatory response and were likely to be a postmortem artifact. In four of the six animals examined histologically, moderate amounts of <0.5–2.5 µm dark brown to black pigmented particulate matter were focally distributed and associated with severe lesions in the lung (Figs. 2B, C, S1–S4). This matter was intracellular in bronchiolar epithelium and alveolar macrophages in most animals.

Other histologic findings in the mice, some of which may be incidental, included focal areas of white matter spongiosis and rare hypereosinophilic neurons and mild to moderate diffuse renal tubular karyorrhexis and pyknosis, possibly associated with autolysis, in four animals (mice 5, 7–9); focal intracellular aggregates of 0.5–1.5 µm brown to black pigmented granules in hepatic Kupffer cells in mouse 1; apoptosis in a moderate proportion of Kupffer cells in mice 7–9; mild to moderate infiltration of hepatic periportal areas with mostly lymphoid cells in mice 6, 8, and 9; and diffuse microvesicular vacuolation of hepatocytes and moderate amounts of aggregated yellow brown pigment, likely hemosiderin, in the splenic periarteriolar lymphatic sheath and red pulp in mouse 5.

All animals had severe pulmonary lesions consistent with smoke inhalation. The distribution of lesions in the central regions of the lungs with distal and peripheral tissue largely unaffected was consistent with inhalation of the insulting agent. The lesions observed were consistent with those seen experimentally in murine models of acute wood smoke inhalation (Matthew et al. 2001). The pigmented particulate matter observed in four of the six examined mice was morphologically consistent with smoke particulate matter, specifically particles 2.5 µm in diameter or less (PM_2.5). In one animal that died acutely, identical particles of up to 1.5 µm in diameter were observed focally in hepatic Kupffer cells. The animals died between three and 30 d after exposure to extremely hazardous levels of bushfire smoke.

Acute smoke inhalation from wood combustion is known to cause severe pathology in people, especially pulmonary edema, which develops because of the toxic components of smoke that damage the alveolar epithelium and capillary endothelium (Witten et al. 1988). These toxins include necrotizing agents such as acrolein and other aldehydes that are present in both the gaseous phase of smoke and the particulate phase consisting of carbon particles with many gas phase compounds, heavy metals, and other toxins adhered (Demling 2008). Effects of the gaseous phase are short-lived, while deposition of particles in the lung leads to both short- and long-term injury. Notably, it is the particulate matter in smoke that is thought to be the major factor affecting alveolar epithelial permeability (Minty and Royston 1985). Pulmonary congestion and severe atelectasis can also occur due to loss of alveolar surfactant production (Witten et al. 1988). Microthrombi associated with regions of likely pulmonary infarction were observed in one of the mice. Interestingly, acrolein has been demonstrated to create a prothrombotic state in mice models (Sithu et al. 2010), though a relationship between wood smoke inhalation and thrombosis has not been conclusively demonstrated in humans (Hunter et al. 2014). The PM_2.5 are known to be the dominant smoke and pollution particle size retained in human lung parenchyma (Churg and Brauer 1997). The timing of deaths did not coincide immediately with periods of the extremely hazardous peak of particles 10 µm in diameter or less (PM₁₀) and PM_2.5 exposure (Fig. 1). Atmospheric conditions, especially ambient temperature, and individual animal traits may be responsible for the timing of mortalities. Between 2 December 2019 and 5 February 2020 ambient temperature in the two affected outdoor acclimatization enclosures exceeded 40 C on five occasions. All seven animals in these enclosures died on the two >40 C days after the first period of extremely hazardous smoke exposure, while no mortalities were observed on the three earlier >40 C days. This important observation provides evidence differentiating smoke versus heat as the primary cause of the observed lung changes in the affected mice. Additional evidence against heat as the primary cause of mortality was the deaths of two affected mice in indoor enclosures (mice 5 and 6) on cooler days (Fig. 1).

The role of ambient temperature on respiratory demand and probable decompensation with respiratory failure in animals with smoke inhalation injury deserves attention, especially in light of a predicted increase in the number of hot days and warm nights annually in southeastern Australia (Garnaut 2008). Additionally, individual susceptibility associated with life history parameters including age and breeding status will likely influence population effects associated with bushfire smoke inhalation. Differing species susceptibility occurred in this event, with little effect observed in numerous species of co-located aviary birds. Factors such as tidal volume influence the site of deposition of smoke particles (Demling 2008) and thus the associated pathology, potentially contributing to different species sensitivity to smoke inhalation. A more nuanced understanding of the varying composition of particulate matter in bushfire smoke and the relationship of that composition to pathologic effect is warranted as different bushfire and other environmental conditions may affect the lethality of smoke.

Our observations are a novel report of mortalities in wildlife associated with exposure to hazardous bushfire smoke. The bushfire events that generated this smoke are unprecedented in Australia but are predicted to occur over a longer season and to increase in intensity due to climate change (Garnaut 2008). Knowledge of the pathophysiology of acute, subacute, and chronic smoke inhalation in different species of wildlife, risk factors for individual mortality after smoke inhalation, biologic and ecologic traits defining species susceptibility to smoke inhalation, and smoke intensity, duration, and composition during bushfires will be needed to assess bushfire smoke inhalation as a cause of mortality in wildlife. Nevertheless, our report suggests two novel risks for conservation. First, severe bushfires may affect wildlife populations beyond the bounds of directly burnt areas, potentially increasing the proportion of species' distributions affected by such events and affecting animal survival in unburnt refugia. Second, conservation interventions involving animal capture, handling, or transport following bushfires may cause animals with smoke inhalation injury to decompensate and die of respiratory failure secondary to oxidative stress. Clearly, further clarity is needed to quantify the likelihood and consequences of such a risk, but a precautionary approach is justified in which the risk of smoke inhalation injury is considered and influencing factors such as ambient temperature are controlled during conservation interventions that are assessed as being worth the risk.

The New South Wales Office of Environment and Heritage Saving Our Species program initiated and funds the captive breeding program described in this article.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-20-00026.