Australia is an island continent, with most dogs and people residing in large cities within 70 km of the coast. The population is mainly located in capital cities of the mainland states. This is where most veterinary schools are located, at least until recently when Charles Sturt University (Wagga Wagga, NSW) started a veterinary program. The canine literature is therefore biased towards conditions seen in dogs living in Sydney, Melbourne, Brisbane, Perth, Adelaide and Townsville. For this reason, there is a paucity of information concerning tick-borne disease in Australia, especially diseases transmitted by the brown dog tick, Rhipicephalus sanguineus, which is generally found further inland [1]. Limited recent information concerning tick-borne diseases is largely derived from free-roaming dogs living in close association with a limited number of indigenous Aboriginal communities, mainly in the Northern Territory (NT) [2, 3]. A further study concerns vector-borne diseases of pet dogs from Darwin and Southeast Queensland [4, 5].

The interplay of infectious agents transmitted by ticks is relevant firstly to dogs living in areas where these organisms are biological vectors, such as tropical north Queensland, the far north of Western Australia and NT. The concept of ‘One Health’ incorporates intra- and inter- species disease transmission, taking into account environmental variables. Thus, within the current context, dogs represent unique sentinels of infection for human medicine. But infections of dogs in these regions impacts also on other parts of Australia, because people can take dogs where they travel. This might become an even bigger issue if, hypothetically, Darwin was subjected to a tropical cyclone, with subsequent translocation of dogs to other regions. Such a mechanism was implicated anecdotally for the accelerated spread of dirofilariasis in the 1970s following cyclone Tracy [6].

Babesiosis was the first tick-borne disease characterised in Australia by the Veterinary Tropical Health Division, James Cook University in Townsville [7, 8]. Babesia vogeli (formerly Babesia canis vogeli) is endemic in far north Queensland and the NT. Babesiosis was characterised with the methodologies of the time by Irwin and colleagues [9, 10] and further insights occurred subsequently, as polymerase chain reaction (PCR) testing of venous blood replaced examination of Romanowsky-stained blood films made from capillary blood [11, 12]. PCR simplified testing, being more sensitive and specific than microscopy, with the capacity to detect both clinical and subclinical infections. Canine babesiosis due to B. vogeli continues to be a cause of symptomatic haemolytic, subclinical anaemia and thrombocytopenia in these areas [5]. Investigations by various groups have confirmed babesiosis as a common cause of anaemia in NT ‘camp dogs’ in aboriginal communities [3, 5]. Anaemia can be severe enough to cause pale mucous membranes. In puppies and dogs immunosuppressed by malnutrition, life-threatening haemolytic anaemia can develop [9, 13]. Compared to B. rossi and B. canis, however, B. vogeli is considerably less virulent [14, 15], possibly as a result of its longer association with domesticated dogs [16]. Babesiosis due to Babesia gibsoni has been reported sporadically in Australia in American pit bulls [17, 18], the disease originating from dogs imported from the USA [19] or Asia [20, 21] and inoculated by fighting, rather than the feeding of ticks [22–24].

In 2001, Anaplasma platys was discovered in Australia by Brown and colleagues during an evaluation of the health status of free roaming aboriginal dogs [25]. Further work, including experimental infections, showed A. platys caused mild to moderate cyclic thrombocytopenia [26–28], which was usually subclinical. Anecdotally, dogs with heavy A. platys infections have been observed to bleed more freely from tick attachment sites than uninfected dogs (Dr Graeme Brown, personal communication). Preliminary data showed some dogs were infected with B. vogeli, others with A. platys, while a small percentage were infected by both pathogens [3].

Two research groups demonstrated that three or more species of haemotropic mycoplasmas infected free-roaming camp dogs of Central Australia [2, 4, 5, 29]. Two species were preponderant, Mycoplasma haemocanis (the more virulent ‘large form’) and Candidatus M. haematoparvum with a third unnamed species unable to be characterised further [2], and the fourth found to be M. haemobos, a cattle pathogen [5]. The clinical significance of these erythrocytic parasites is currently unclear, as presence of Mycoplasma sp. has no apparent correlation with anaemia, even when the cycling threshold (C _T ) of the real-time PCR (qPCR) indicates heavy bacteraemia.

The zoonotic pathogen Rickettsia felis has recently been shown to have the dog as a definitive host, with transmission requiring fleas or ticks as biological vectors [30]. Rickettsia felis has been found in dogs in Southeast Queensland and Central Australia [31]. Other canine tick-borne diseases, such as ehrlichiosis and cytauxzoonosis, are not endemic, but have the potential to be imported from overseas.

Bartonella spp. have yet to be reported in the peer-reviewed literature from Australian dogs, even though fleas and ticks are common in many parts of Australia and Bartonella henselae and Bartonella clarridgeiae are known to be present in fleas [32], foxes [33], humans [34] and cats in Australia, especially young cats [32, 35]. Furthermore, many Australian native animals have a variety of Bartonella spp. which could potentially be transmitted to canids [36, 37].

Coxiella burnetii, the agent of ‘Q fever’ in human patients and coxiellosis in animals, is a Gram-negative bacterium. Although theoretically capable of being transmitted by fleas and ticks, it is generally transmitted by aerosols from reproductive secretions, or possibly by the ingestion of infectious forms present in uncooked meat or offal from infected animals [38, 39]. A recent Australian serosurvey found dogs living in Aboriginal communities had the highest seroprevalence of coxiellosis [40].

Work to date has emphasised that individual dogs in certain settings can be infected by a combination of flea- and tick-borne infectious agents. Polymicrobial infections with so-called ‘stealth pathogens’ can have complex effects due to intricate interactions such as molecular mimicry and immune-stimulation that cannot be appreciated by studies of a single pathogen under experimental conditions [41, 42]. Thus, in regions where vectors are common, and especially where prophylactic ectoparasitic measures are not undertaken routinely, polymicrobial infections, in association with genetic, nutritional and environmental factors, are likely to impact on haematologic variables and the overall health of infected dogs. Among individuals, disease expression may be influenced by nutritional and environmental stressors, including insufficient quantity or quality of food, pregnancy, lactation or hard work.

The purpose of the present study was to 1) Determine the prevalence of selected zoonotic pathogens and vector-borne pathogens among free-roaming dogs living in Aboriginal communities in different parts of Australia; 2) Assess potential correlations between anaemia and thrombocytopenia and the qPCR C _T values for B. vogeli, A. platys and haemotropic mycoplasmas; 3) Determine whether dogs from Central Australia were infected by Bartonella spp.

Within the two study cohorts, there were conspicuous differences in the prevalence of tick-borne diseases, with all three genera (Anaplasma, Babesia, haemotropic Mycoplasma) of tick-borne pathogens being less common in NW-NSW than Central Australia. This was correlated with both a much lower prevalence of ticks and also better nutrition and overall health status. The results of this study build on previous work concerning vector-borne and zoonotic infectious diseases of free living dogs in and around Aboriginal communities [2–5, 25, 31, 49, 50].

There was a statistically significant association between the presence of B. vogeli and/or A. platys and the presence of anaemia and/or thrombocytopenia. Having said this, in the great majority of dogs, the deficiency in red cell mass and platelet numbers was minor, and unlikely to give rise to clinical disease. This is in accordance with the literature, where B. vogeli is considered less virulent than its two sibling species and tends to produce subclinical disease, except in young or immunosuppressed dogs [9, 13, 15, 16]. A. platys strains in Australia are considered minimally pathogenic, although in some dogs, especially younger dogs with co-infections with other organisms, thrombocytopenia might be severe enough to impact patients subjected to injury or surgery, e.g. dog fights or surgical neutering [3]. It might be prudent therefore to evaluate a peripheral blood smear in the field to ensure platelet numbers are adequate, or to perform a platelet function test e.g. buccal mucosal bleeding time prior to major surgery [51]. Intra-operative bleeding has not been reported as a clinical problem in the field during large-scale surgical neutering campaigns in these communities. It has been the experience of one of the authors (GB) that removing ticks is a good platelet function test, dogs with A. platys often having prolonged bleeding from tick attachment sites. It was of great interest to observe that there was no positive correlation, and indeed in one instance (B. vogeli and anaemia) there was a negative correlation, between the quantity of bacterial or protozoan DNA in blood and red cell mass or platelet count (Fig. 3). In the instance of B. vogeli, the simplest explanation for this finding would be that when anaemia occurs in the setting of long-standing subclinical infection, anaemia is due to sequestration of infected erythrocytes by the mononuclear phagocytic system, reducing or eliminating parasitised red cells and the organisms they contain from peripheral blood.

Although commonly encountered, haemotropic mycoplasmas did not impact on haematologic values in a clinically meaningful fashion. This is also in accord with the traditional view from the literature, which suggests that in the absence of splenectomy or immunocompromise, neither M. haemocanis nor Candidatus M. haematoparvum causes haemolytic anaemia [4, 5, 52–54].

The most striking observation of the present work was an unequivocal negative finding: PCR testing of whole blood, even after a special pre-incubation step using BAPGM media (in one subgroup), failed to detect Bartonella nucleic acid from any dog in either of the two sites tested. Likewise, there was no serological evidence for previous Bartonella infection. Although dogs around the world can be infected with numerous Bartonella spp., B. henselae, B. vinsonii subsp. berkhoffii and B. koehlerae were most frequently documented in sick dogs in the United States [55]. Following the initial isolation of Bartonella vinsonii subsp. berkhoffii, tick exposure was determined to be a risk factor for prior exposure based upon IFA seroreactivity [56]. Based upon the high prevalence of both Ehrlichia canis and B. vinsonii subsp. berkhoffii antibodies in dogs from the south-eastern United States [56], R. sanguineus was a suspected vector, although this has never been confirmed [57]. Similar to the results of this study, a subsequent investigation from Brazil failed to identify an association between E. canis serology or PCR positivity and Bartonella spp. antibodies [58]. However, as recent research has documented at least four phylogenetically distinct clades (that justify different species designation) within the morphological designation R. sanguineus [59], a role for Bartonella spp. transmission for ‘R. sanguineus’ clades around the world requires further study.

Cats are underrepresented as companion animals in most Aboriginal communities, certainly in the ones sampled in Central Australia. So the likelihood of dogs contacting any of the feline-adapted Bartonella (B. henselae, B. clarridgeiae, B. elizabethiae) via cat fleas would be minimal, especially in a hot, dry, arid environment where fleas are rare to absent, while brown dog ticks are conspicuously common. Furthermore, B. vinsonii subsp. berkhoffii has yet to be reported from any Australian jurisdiction, possibly because Australia does not have wolves or coyotes, the natural reservoir for this organism in North America [60–62]. In the context of sick dogs, both IFA serology and PCR from blood lack diagnostic sensitivity as compared to BAPGM enrichment blood culture [55]. In addition, because B. henselae, B. vinsonii subsp. berkhoffii and B. koehlerae strains in Australia might differ antigenically from North American strains used for testing in this study and would be unlikely to cross react with indigenous Australian marsupial strains [44, 63], negative serology results should be viewed with caution. These factors could account for the absence of published reports of canine bartonellosis from Australia, despite the widespread availability of multiplex PCR assays.¹ It is important that studies such as this be extended to regions where fleas are more common and dogs encounter more than one tick species, such as the coastal parts of tropical north Queensland, because in this environment Bartonella spp. may be more likely to ‘spill over’ from cats and foxes to dogs.

It is quite difficult to directly compare the work presented here with that done previously in Australia, as the communities sampled were somewhat different, as were the methodologies employed, including the specifics of the PCR assays. Without equivocation, Traub’s group has shown dogs from central Australian indigenous communities had much higher prevalence of all tick-borne pathogens than dogs in Darwin, which in turn had higher overall prevalences than south east Queensland and Sydney [4, 5]. Presumably this reflects the better condition of owned dogs in Darwin, with better ectoparasitic control, nutrition and so forth, while in Brisbane and its environs and Sydney, Rhipicephalus is just not sufficiently common to be an effective disease vector, the hard tick Ixodes holocyclus being the preponderant tick species. It should be noted that our ‘healthy’ dogs from the NT living in indigenous communities were, on the whole, more affected by tick-borne pathogens than the ‘affected’ cohort of hospital patients in the 2015 study by Hii et al. [4].

Dogs in indigenous communities sampled were generally judged to be healthy, despite the commonness of tick-borne diseases. The presence of tick-borne pathogens causing subclinical anaemia and thrombocytopenia complicates the diagnosis of diseases common in owned dogs (such as primary immune-mediated haemolytic anaemia [IMHA] and thrombocytopenia [IMT]) in nearby regions such as Alice Springs and Darwin, as there is some flow of dogs between cities in the NT and the surrounding communities. The response to specific anti-infective therapy might be required to dissect out the contribution of various infectious agents.

In dogs living in indigenous communities, the high prevalence of tick-borne diseases requires addressing, even though the majority of dogs in our study appeared in reasonable condition. Our view is that the regular application of acaricides, such as topical moxidectin (especially inexpensive off-label ‘pour on’ cattle formulations), represents a key control measure, perhaps combined with environmental treatment to reduce the number of adult and juvenile ticks, so that immunologically naïve pups are not subjected to sudden inoculation. It is difficult to envisage how imidiocarb or doxycycline could be routinely given in an indigenous setting, because of the requirement to treat dogs for several consecutive days with painful injections or tablets, respectively. If dogs from Aboriginal communities become translocated into pet homes in places like Alice Springs or Darwin, haematologic and multiplex qPCR testing would be prudent, so that subclinical carriers of these blood-borne pathogens could be treated to either eliminate the infection, or more realistically, reduce the extent of the infection so that clinical manifestations are unlikely to develop and transmission to dogs in non-endemic areas is minimised. Likewise, blood donors in such geographies should be tested by multiplex qPCR, as transfusion of blood containing these usually mild pathogens might cause more substantive issues if given to patients with IMHA and/or IMT, where immunosuppressive therapy will be on-going, or after splenectomy.

This study incorporated a ‘One Health’ perspective as bartonellosis, coxiellosis and brucellosis are all important zoonoses, with dogs from Aboriginal communities potentially serving as sentinels for these infections. Bartonellosis has been discussed at length already. In relation to coxiellosis, 4/96 dogs tested sero-positive for C. burnetii infection. None of these was symptomatic, but worryingly two of the dogs were sexually intact females, including a pregnant bitch. Despite the apparently high prevalence of anti-Coxiella antibodies in dogs living around indigenous communities, Q fever is not considered a common or important disease of Aboriginal people, despite the very close relationship they have with their dogs [40]. One of the seropositive dogs in the NW-NSW cohort was actually a kelpie belonging to a drover, rather than residing in an Aboriginal community. The Moree region subpopulation was also tested for serological evidence for subclinical brucellosis. Only one dog tested weak-positive with the Rose Bengal method, but as it was negative on complement fixation testing, it was classed as negative as per laboratory protocol, although the dog could remain subclinically infected or have eliminated the infection resulting in a residual antibody titre. From a ‘One Health’ perspective, this report importantly draws light to both negative and positive dog data which are equally instructive for both canine and human health.

The vector-borne pathogens B. vogeli, A. platys and haemotropic mycoplasmas were prevalent in dogs living in Aboriginal communities in Central Australia, but less common in similar environments in NW-NSW. Although a minimally pathogenic bacterial infection in dogs and humans, the potential for A. platys infection should be investigated among sick human individuals in indigenous communities, due to previously mentioned reports in humans overseas. B. vogeli was associated with anaemia, and to a lesser extent thrombocytopenia, A. platys was strongly associated with thrombocytopenia, whereas neither haemotropic mycoplasma was associated with any haematologic aberration. Similar to the dogs in this study, co-infection with A. platys and Candidatus M. haematoparvum was reported in a veterinarian, who was concurrently infected with B. henselae [64]. Polymicrobial combinations of these infectious agents were common in dogs, yet they had little discernible clinical impact. Serological evidence for coxiellosis was evident in one male and three female dogs, yet spill over into in-contact humans in these geographical regions is unknown. The presence of tick-borne pathogens in dogs residing or sourced from these indigenous communities must be considered when investigating haematologic disorders such as anaemia and thrombocytopenia. It remains problematic to differentiate between primary IMHA or IMT from disease secondary to the immune response to these pathogens, or to the actions of the pathogens themselves. Testing of blood donors in areas endemic for these blood-borne pathogens is mandatory, to prevent infection of naïve patients.