HISTOPATHOLOGIC FINDINGS IN FREE-RANGING CALIFORNIA HUMMINGBIRDS, 1996–2017

Hummingbirds belong to one of the most diverse avian families (Trochilidae), are widely distributed across the Western Hemisphere, and seven species are reported in California, US (Williamson 2001). These birds are well adapted to both natural and urban environments and occupy a unique environmental niche as both nectivorous and insectivorous feeders. As such, they not only provide insect control and pollination services, but also function as an indicator for increased environmental pollutants such as pesticides and trace elements due to their distinctive feeding habits (Godoy et al. 2014; Mikoni et al. 2017).

Despite their documented importance within an ecosystem, little is known about causes and trends of mortality of free-ranging hummingbird populations. Most reports focus on a specific disease in an individual bird (Meteyer et al. 1992; Snowden et al. 2001; Saidenberg et al. 2007; Godoy et al. 2013). Population-level studies of mortality have not been performed. Such studies are needed to provide baseline data in order to better understand and prospectively track the overall health of hummingbird populations within an ecosystem. As hummingbird population health may potentially serve as a sentinel for environmental pollution, understanding the present causes of mortality and their relative prevalence within the population is of great value, as was true with seabirds (Clatterbuck et al. 2018). The purpose of our study was to identify causes of mortality in free-ranging California hummingbirds through histologic examination of tissues.

We studied histologic lesions in free-ranging California hummingbirds in birds collected from the grounds of the San Diego Zoo (San Diego, California) and San Diego Zoo Safari Park (Escondido, California; San Diego Zoo mentions collectively referred to as SDZ throughout) from 1996 to 2016 and by the Lindsay Wildlife Experience (LWE; Walnut Creek, California) from 2015 to 2017. All of the birds were free-ranging birds that either died (SDZ) or were submitted from the surrounding areas (LWE). We examined tissues from 61 hummingbirds representing five native California species: Anna's Hummingbirds (Calypte anna), Allen's Hummingbirds (Selasphorus sasin), Rufous Hummingbirds (Selasphorus rufus), Costa's Hummingbird (Calypte costae), and Black-chinned Hummingbirds (Archilochus alexandri). All birds were either found dead or discovered in a moribund state and subsequently euthanized; long term–rehabilitated birds were not included in this study.

Cases were categorized by species, sex, age (juvenile or adult), and nutritional status (poor, good, excellent amount of adipose tissue, unknown). All hummingbirds from the SDZ were identified to the species level by experienced prosectors using standard field guides when specimen condition was adequate. Accurate speciation could not be confirmed for cases submitted from the LWE; therefore, these birds were grouped in an “unknown” species category. Sex was determined by microscopic identification of gonadal tissue. Sex was classified as unknown when gonads were not present in the examined sections. Birds that had prominent bursal or thymic tissue were classified as juveniles, and all other birds were considered to be adults. Nutritional status was subjectively assessed based on the amount of adipose tissue seen on microscopic examination of the visceral organs and marrow cavities of tibiotarsi.

For cases from the SDZ, we evaluated case histories, gross necropsy reports, and H&E-stained sections of visceral organs, tibiotarsus, and head including eye and brain from previously necropsied birds. For cases from the LWE, whole birds fixed in formalin following opening of the calvarium and the coelom were submitted to the Michigan State University Veterinary Diagnostic Laboratory (Lansing, Michigan) for gross and histologic examination. Collected tissues were embedded in paraffin, sectioned at 5 µm, and routinely stained with H&E for microscopic examination. Histologic evaluation included examination of all visceral organs, decalcified portions of tibiotarsus, and sections through the decalcified head including eye and brain. Additional histochemical stains, including Gram stain for bacteria, Grocott-Gomori's methenamine silver stain for fungi, Prussian blue stain for iron, and Ziehl-Neelsen acid fast stain for Mycobacterium spp., were performed when indicated based on the initial histologic findings.

Ancillary diagnostic tests were pursued when indicated based on microscopic examination. West Nile virus infection was identified with immunohistochemistry (Smedley et al. 2007). Polymerase chain reaction was utilized to identify multiple etiologic agents including adenovirus, Mycobacterium spp., and avian hemoparasites. All PCR tests were conducted at the SDZ Molecular Diagnostics Laboratory (San Diego, California). We extracted DNA from two 15-µm, formalin-fixed, paraffin-embedded tissue scrolls of each sample using the QIAamp DNA FFPE Tissue Kit (Qiagen, Valencia, California, USA) and following the manufacturer's instructions. We eluted DNA in 60 µL ATE buffer. Three PCR assays were performed. The first assay was a nested PCR targeting the DNA polymerase gene of adenoviruses using primers polFouter/polRouter and polFinner/polRinner and the described cycling parameters (Wellehan et al. 2004). The second assay was a conventional PCR targeting the mitochondrial SSU and LSU rRNA genes of Plasmodium spp. and Haemoproteus spp. using primers 343F/496R and the described cycling parameters (Fallon et al. 2003). The third assay was a conventional PCR targeting the heat shock protein of Mycobacterium spp. using primers Tb11/Tb12 and the described cycling parameters (Telenti et al. 1993). Reaction mixes for each assay were composed of 0.4 µM of each primer, 12.5 µL of MyTaq HS Red Mix, 2X (Bioline, Taunton, Massachusetts, USA), 3 µL extracted DNA, and nuclease-free water to a final volume of 25 µL. The PCR products were visualized on a 1% agarose gel. We purified the DNA from bands of expected size using the Millipore Ultrafree-DA Kit (Merck Millipore Ltd., Tullagreen, Carrigtwohill, Cork, Ireland). Purified DNA was submitted for direct sequencing to Eton Bioscience Inc. (San Diego, California, USA).

Histologic lesions were grouped by organs and subjectively ranked as mild, moderate, or severe. When appropriate, suspected etiologies were further grouped into broad categories of infectious, trauma, or multifactorial. The infectious disease category was further subdivided by causative agent (bacterial, viral, fungal, protozoal, metazoal). Assessment of parasitism included observation of macroparasites whether or not they caused microscopic lesions. Most infectious causes were further characterized with additional histologic stains or ancillary diagnostic testing as listed above. Traumatic lesions were classified based on the primary location of the trauma (head, respiratory, multifocal, or other).

Traumatic lesion incidence was compared among sex, species, age, and sampling location (SDZ or LWE) using Fischer's Exact test using a commercial statistical software program (GraphPad Software, La Jolla, California, USA). We considered P-values less than or equal to 0.05 to be significant.

Sixty-one hummingbirds were examined. Twenty hummingbirds were submitted from the LWE and 41 were from the SDZ. The LWE commonly receives hummingbirds from the public with a mean (SD) of 278 (43) birds from 2015 to 2017. The most common species represented was Anna's Hummingbird (34%, 21/61) followed by Allen's Hummingbird (13%, 8/61), Rufous Hummingbird (13%, 8/61), Costa's Hummingbird (8%, 5/61), and Black-chinned Hummingbird (2%, 1/61). Eighteen birds were not identified to the species level (30%, 18/61). Twenty-eight of the birds were male (46%), 16 were female (26%), and 17 were of unknown sex (28%). Fifteen birds (25%) were classified as juveniles, based on the presence of well-developed thymic or bursal tissue, and the remaining 46 birds (75%) were considered adults.

Twenty-three percent (14/61) of the study birds had lesions attributable to acute trauma (Table 1). This was the most common category of lesions observed in birds followed by viral, bacterial or fungal disease (16%, 10/61), parasitism (11%, 7/61), and poor nutritional status score (5%, 3/61; Table 1). Thirteen percent (8/61) had lesions indicative of multiple concurrent processes (Table 1). The most commonly affected organ was the lung (68%, 40/59), for which the majority of the histologic lesions were diffuse congestion and hemorrhage. The second and third most commonly affected organs were the ingluvies (67%, 34/51) and the liver (54%, 33/61). The most common lesion in the ingluvies was mucosal-associated bacterial overgrowth, and the most common liver lesion was hepatocyte microvacuolation (Table 2).

Acute trauma was the most commonly identified cause of mortality (Table 1). Typical microscopic lesions of trauma consisted of areas of severe acute hemorrhage within visceral organs or subcutaneous tissue. Acute necrosis was occasionally apparent in the affected areas of the liver and heart. The most common organ system with histologic lesions attributable to trauma was the respiratory system (Table 3). There were no statistical differences among sex, species, age, and sampling location (SDZ or LWE) for traumatic lesions.

Twenty-seven birds had lesions of infectious (bacterial, fungal, viral, or parasitic)disease, with a total of nine different etiologies represented (Table 4). Lesions attributable to infectious diseases identified included those associated with avian poxvirus, intestinal adenovirus, disseminated aspergillosis, bacterial septicemia, and acid-fast staining–confirmed mycobacteriosis (Table 4). Parasites identified in these birds included malaria (Haemoproteus spp.), cestodes, and feather mites, none of which appeared to be associated with significant disease (Table 4).

A single adult male Anna's Hummingbird was suspected to have a chronic Mycobacterium spp. infection within the bone marrow of the tibiotarsus. Hematopoietic cell lines were hyperplastic, filling nearly 90% of the marrow cavity, and there were multiple aggregates of vacuolated macrophages that infrequently appeared to contain small, clear, rod-shaped organisms (Fig. 1A). Additional granulomas composed of similar macrophages were noted within the hepatic parenchyma and the dermis of the feathered skin overlying the head. Although the mycobacterial PCR was negative, mycobacteriosis was still suspected based on the histologic lesions and confirmed with presence of acid-fast bacilli using ZielNeelson stain. In this case, the number of organisms may have been too low to detect by PCR, or formalin fixation of tissue may have affected tissue quality and resulted in a false-negative PCR result.

In one adult hummingbird, for which the species was unknown, there were foci of necrosis and inflammation with intralesional bacterial cocci within the myocardium, renal parenchyma, and tibiotarsus, consistent with bacterial septicemia. This bird had evidence of traumatic injury to the soft tissue overlying the tibiotarsus, which was likely the point of origin of the septicemia. Lesions of bacterial bronchopneumonia were identified in two birds. These lesions included pulmonary consolidation and necrosis with accumulation of karyorrhectic debris and multifocal infiltrates of heterophils and histiocytes.

Cases of intestinal adenovirus were characterized by 20–40-µm diameter basophilic intranuclear inclusions that caused marginated chromatin within enterocytes. No associated necrosis or inflammation was observed (Fig. 1B). Of the two cases with microscopic adenovirus lesions, one tested PCR-positive for adenovirus. Both affected birds were adult female Anna's Hummingbirds. Three adult male hummingbirds (species unknown) showed characteristic histologic lesions of poxvirus, evidenced by epidermal hyperplasia with ballooning degeneration of keratinocytes and 20–30-µm diameter eosinophilic intracytoplasmic viral inclusions (Bollinger bodies). These lesions were present on the feathered skin of the face and the eyelid (Fig. 1C).

Three hummingbirds (species unknown) had microscopic evidence of disseminated aspergillosis, characterized by foci of granulomatous inflammation with intralesional 5–7-µm diameter parallel-walled, septate fungal hyphae in the lungs, trachea, corneal epithelium, and vessel lumens.

Brown, granular, refractile pigment (anisotropic pigment) was present within vessel lumens and macrophages of multiple organs in two adult female unspeciated hummingbirds (Fig. 1D). One of these birds additionally had moderate splenic necrosis characterized by multifocal loss of architecture and accumulation of karyorrhectic debris; PCR for hemoparasites was positive for this bird. This infection was thought to be most consistent with Hemoproteus spp. (Fig. 1E).

Intestinal cestodiasis was the most common metazoal parasitic infection in the surveyed population, with 12 hummingbirds affected. Multiple species were affected including seven Anna's Hummingbirds, two Allen's Hummingbirds, and one Costa's Hummingbird. The remaining two affected birds were unspeciated. Seven of the affected birds were males, three were females, and two were of undetermined sex. Three of the affected birds were juveniles. Cestodes were located primarily in the intestines, and were occasionally found in the proventriculus, consistent with postmortem aboral migration. The cestodes were 250–2,500 µm long and 250–400 µm wide, and frequently filled the entire intestinal lumen; they contained both male and female reproductive organs and had basophilic calcareous corpuscles. There was no associated inflammation or necrosis of the intestinal mucosa (Fig. 1F).

The nutritional status for the 61 hummingbirds examined was poor (23%, 14/61), good (69%, 42/61), excellent (5%, 3/61), or unknown (3%, 2/61). Of the 14 birds with poor nutritional status, six were males, five were females, and three were of undetermined sex. The most commonly represented species was Anna's Hummingbird (10 birds), with the remaining four birds being unspeciated (three birds) or an Allen's Hummingbird (one bird). Six of the birds with poor nutritional status had concurrent infectious diseases. All three birds with poxvirus infections had a poor nutritional status, and four birds had severe cases of intestinal cestodiasis. Of the three birds with excellent nutritional status, one was a Rufous Hummingbird and two were not identified to the species level. Two of these birds had histologic lesions of acute trauma, and an obvious cause of mortality was not identified in the remaining bird.

A cystic structure intimately associated with the hepatic parenchyma, which was histologically consistent with a gallbladder, was identified in 21% (13/61) birds. This structure was lined by tall, columnar epithelium and contained variable amounts of intraluminal, homogenous, basophilic, flocculent material. Occasionally, this structure was seen to directly connect to large biliary ducts within the liver (Fig. 2A, B). Four of these birds were Anna's Hummingbirds, and the remaining nine birds were unspeciated. In two unspeciated hummingbirds, multiple, well-defined foci of ossification were identified within the walls of the aorta and pulmonary artery at the level of the heart base, suggestive of an osseous cardiac skeleton (Fig. 2C).

A variety of lesions that we identified in the hummingbirds were associated with infectious disease, acute trauma, or a combination of the two processes. Significant mortality attributable to a single etiology was not observed, and there were no observed sex, species, age, or sampling location predilections for any of the identified traumatic causes of mortality.

Twenty-seven birds had lesions of infectious-parasitic disease, representing nine different etiologies. To our knowledge, intestinal adenovirus has not been previously reported in hummingbirds. The remainder of infectious-parasitic etiologies identified in this study have been previously reported in the literature (Godoy et al. 2013; Godoy et al. 2014; Bradshaw et al. 2017; Yamasaki et al. 2018). While previous studies have reported hemoparasitism more commonly in female birds and poxvirus selectively within Anna's Hummingbirds, the numbers of affected birds in this study was too small to draw significant statistical conclusions regarding the prevalence of these conditions between different sexes or species (Asghar et al. 2011; Godoy et al. 2013).

Twenty-one birds had lesions of acute trauma; 12 of these birds were from the SDZ and nine were from the LWE. According to the histories provided, many of the birds from the SDZ were found either dead or moribund next to a glass exhibit window. Often, these birds were found in pairs at a single location, suggestive of territorial or breeding behavior. A previous study examining the relative vulnerability of different species to bird-building collisions identified hummingbirds as one of the most susceptible avian families to infrastructure-associated mortality (Loss et al. 2014). Loss et al. (2014) postulate several ideas to explain why various species of birds collide with buildings, including attraction to lighted structures at night by migratory species and increased proximity to residences with feeders for birds in general. In our study, the myriad glass exhibit windows within the SDZ likely presented a microenvironment that predisposed to increased window collision events. Exhibits at SDZ with recurring window collisions have been treated with patterned window film (typically using vertical 3.175-mm stripes, 102 mm apart), which has dramatically reduced window collisions. The number of birds with acute traumatic lesions did not vary significantly between the sampling locations in our study (P=0.574), indicating that comparable numbers of hummingbirds experienced lethal acute traumatic events in areas without extensive glass window coverage.

Only three birds in this study were classified as having excellent nutritional status, including one Rufous Hummingbird. While Rufous Hummingbirds have been shown to store large amounts of body fat (Healy and Calder 2006), the number of birds examined in this study was not large enough to draw significant conclusions about species predilections for increased fat storage. Many of the birds assigned a poor nutritional status had concurrent infectious-parasitic disease processes. Birds with chronic, debilitating diseases such as poxvirus tended to be in the worst body condition, consistent with the slowly progressive nature of their afflictions. Interestingly, four of the 12 birds with histologic evidence of intestinal cestodiasis were classified in the “poor” nutritional status category. While there was no evidence of overt damage to the intestinal mucosa by these organisms, this may imply some level of associated chronic nutrient deprivation-metabolic strain imparted by their presence. However, the remainder of the hummingbirds with intestinal cestodes were generally in good body condition. These birds may have been more-recently infected and thus may not have had time to reach a state of malnutrition. Alternatively, intestinal cestodiasis may represent a largely subclinical condition in hummingbirds.

Hummingbirds were previously thought to lack a gallbladder based on gross necropsy findings (Crompton and Nesheim 2016). However, gallbladders were histologically identified in 13 birds in this study, four of which were Anna's Hummingbirds. We concluded that gallbladders are present in the Anna's Hummingbird and likely in other hummingbird species as well. Failure to identify this structure previously was likely to due to its relatively small size that precluded gross visualization.

The hummingbird heart is notable for its relatively large size (compared to total body weight) and ability to maintain an incredibly fast rate of contraction and subsequently high cardiac output (King and McLelland 1984). Despite its physiologic efficiency, overall heart anatomy in hummingbirds was not found to vary significantly compared to other birds (King and McLelland 1984). In this study, multiple, well-defined foci of ossification were identified within the walls of the aorta and pulmonary artery at the level of the heart base, suggestive of an osseous cardiac skeleton. Such foci of ossification in this location have been previously described in quail chicks and shown to be of neural crest origin (Sumida et al. 1989); the structure has not been reported in hummingbirds.

Limitations of our study include the sampling location bias, as all hummingbirds were collected from only two areas in California. Additionally, the number of hummingbirds sampled was too small to make any inferences about the prevalence of different diseases within the free-ranging hummingbird population. Future studies should aim to examine more birds from different locations and compare causes and rates of mortality between captive and free-ranging populations.

We thank the SDZ and its employees and the LWE and its employees and volunteers for submission of these cases and Josephine Braun, Jennifer Burchell, and Asa Preston of the SDZ Molecular Diagnostic Laboratory for PCR testing. This study was partially funded by generous donations from Grant Patrick, the Hunter-Jelks Fund, and The Daniel and Susan Gottlieb Foundation.