The ‘other’ epidemic: canine ehrlichiosis in Australia

Keywords: canine, dog, Ehrlichia, Ehrlichia canis, Rhipicephalus linnaei, Rhipicephalus sanguineus, tick, vector-borne disease.

Canine monocytic ehrlichiosis (CME) is a tick-transmitted disease of domestic dogs and wild canids caused by Ehrlichia canis (E. canis), a Gram-negative, intracellular bacterium belonging to the family Anaplasmataceae.¹ The organism, initially named Rickettsia canis, was first described in dogs in Algeria,² and later came to prominence during the Vietnam War in the 1960–70s when scores of military working dogs died of what was referred to as ‘tropical canine pancytopenia’.^3–5 Today, CME is recognised as a common illness affecting dogs in tropical and subtropical regions of the world with the exception, until recently, of Australia.

In May 2020, in Kununurra, northern Western Australia (WA), dogs with significant fever, malaise, and haemorrhagic diatheses were diagnosed with E. canis infection, sparking a national veterinary biosecurity alert.⁶ Despite its designation as a notifiable disease, the swift introduction of movement restrictions for dogs, and the unfolding COVID-19 pandemic during 2020 (that imposed travel limitations on the general public, including dog owners), CME was confirmed to have spread throughout northern WA and the Northern Territory, with high morbidity and mortalities reported, especially among dogs at indigenous communities.⁷

The biological and clinical features of E. canis infection have been studied extensively for many years (outside Australia). The vector is the brown dog tick, Rhipicephalus linnaei, previously named Rhipicephalus sanguineus sensu lato (‘tropical strain’).^8,9 As the latter name implies, this tick is enzootic throughout warm and humid climates, including northern and central Australia, and is well adapted to survival in dwellings and kennel environments.¹⁰ Transmission of E. canis to the dog occurs rapidly, within 3 h of tick attachment,¹¹ and pleomorphic organisms may be observed as intracytoplasmic inclusions in circulating monocytes, either singly or in compact colonies (morulae) (Fig. 1).⁵ It is known from experimental studies that E. canis infection in dogs progresses through acute, sub-clinical, and chronic (terminal) phases,¹² and clinical illness develops in dogs approximately 1–3 weeks post-infection. Typical clinical signs reported globally include lethargy, weakness, anorexia, fever, pale membranes, panting, bleeding diathesis, hepatosplenomegaly, and dehydration.^8,13 Despite the severe nature of some cases of acute E. canis infection, including fatalities, many dogs in endemic areas overseas appear to tolerate infection without developing overt clinical disease.⁸

Fig. 1.  Two intracytoplasmic inclusions (morulae) of Ehrlichia canis.

In contrast, clinical data recorded by one of us (JB, unpublished observations) in Broome, WA during 2020–2021, reveals a severe clinical syndrome in dogs diagnosed with CME during the first 6–9 months after the disease was recognised. Epistaxis was often profound (Fig. 2a), and body temperatures as high as 41.5°C were not uncommon. Some dogs presented with dyspnoea and fever, but no bleeding, and in others cachexia was often severe and rapid in onset. Dogs with marked pancytopenia on haematology rapidly progressed to sepsis and uncontrollable bleeding within a few days, and some dogs presented with blindness due to retinal haemorrhages. Another group developed generalised dependant oedema, often months after the acute illness, characterised by marked swelling of the face, neck and limbs (Fig. 2b) and, in some cases, the scrotum (Fig. 2c).

Fig. 2.  (a) Epistaxis associated with severe thrombocytopenia in a dog with E. canis infection. (b) Generalised oedema in a dog with E. canis infection. (c) Scrotal oedema in a dog with acute E. canis infection.

A fascinating historical comparison can be drawn between these severe clinical descriptions in the tropical maritime climate of Broome and those reported by veterinarians in similar climatic conditions managing the disease in military working dogs (MWD) brought into Southeast Asia from Europe and the US during the conflicts of the 1960s and 1970s.¹⁴ Reports then described a highly fatal epizootic tropical canine pancytopenia, the aetiology of which was not immediately apparent, even raising concerns about environmental toxins and ionising radiation.³ Some of the worst clinical manifestations, including the development of oedema, are less commonly reported since the dramatic descriptions from Indochina 50 years ago. We believe that the clinical parallels between Vietnam in the 1960s and Broome in 2020 deserve to be recorded, and presumably reflect the sudden and overwhelming exposure of immunologically naïve dogs to large numbers of ticks infected with E. canis. Systemic inflammation and widespread vasculitis are the underlying pathological processes of acute CME, with the development of myelosuppression in some cases.¹⁵

Due to its close taxonomic/phylogenetic relationship with rickettsial organisms, E. canis was treated originally with tetracyclines,¹⁴ but doxycycline is now the most widely used antimicrobial for CME.^16,17 Evidence for its antimicrobial efficacy comes from in vitro studies that indicate E. canis is highly susceptible to doxycycline, requiring a very low minimum inhibitory concentration (MIC) (0.03 µg/mL).¹⁸ There is generally a good clinical response to a 28 day course of doxycycline (and other supportive care, as indicated) in the acute phase, however some dogs remain bacteraemic (therefore infectious to ticks) for unpredictably variable periods of time during the course of antimicrobial drug treatment, and in a few cases after cessation of doxycycline.^19,20

Evidence for strain variation comes from research into variable genes of E. canis predominantly comprising genetic loci that encode major immunoreactive proteins.²¹ At least four glycoproteins (gp19, gp36, gp140, gp200) have been characterised molecularly and major continuous species-specific epitopes have been mapped to acidic serine-rich tandem repeat proteins (TRP).^22,23 Genetic variations have been described through analysis of these TRPs, demonstrating different amino acid sequences among E. canis strains from different geographical areas.²⁴ Ehrlichia canis TRP36 is a species-specific antigen with different tandem repeat sequences among E. canis strains and variations in the TRP36 gene are used widely to describe strains of E. canis worldwide, and at least four clades/genogroups of TRP36 have been described.²⁵

Recently, the full genome of E. canis in Australia was sequenced,²⁶ providing information about the strain and its phylogenetic relationship with overseas isolates. Ehrlichia canis sequenced from ticks and blood samples collected from dogs with CME were found to be highly conserved across large geographical distances,²⁶ consistent with the assumption of a recent introduction of the disease into the country. Analysis demonstrated that the Australian E. canis belongs to the Taiwan genotype, that is widespread throughout Asia and the Middle East.²⁶

Until the arrival of CME, the most serious tick-associated disorder afflicting dogs in Australia was tick paralysis, resulting from envenomation by the eastern Australian paralysis tick, Ixodes holocyclus. In contrast to the rapid transmission of E. canis from brown dog ticks described above,¹¹ the neurotoxins injected by I. holocyclus do not take effect until 3–4 days after tick attachment. Understanding this difference in timing is crucial with respect to the use of acaricides (tick prevention products). Topically acting synthetic pyrethroids (e.g. flumethrin) exert their protective effect through repellency (and kill), thus preventing tick attachment, and are proven to protect dogs against transmission of E. canis.²⁷ The other major class of acaricides currently in use in companion animal practice, the isoxazolines, are systemically acting, requiring the tick to feed in order to take effect.²⁸ This works well to prevent tick paralysis but does not protect dogs from acquiring CME, although isoxazoline products stop onward transmission of E. canis from an infected dog, and have proven a valuable adjunctive treatment to reduce tick burdens where dogs are housed.

Ehrlichia canis is related to several members of the Anaplasmataceae that are recognised to be zoonotic, including E. chaffeensis, E. ewingii and E. muris eauclairensis,¹ and there is evidence for the zoonotic potential of E. canis from molecular and serological testing in Central and South America.²⁹ However, to date there have been only small numbers of people with confirmed E. canis infections in South America and there are no reports of infection in humans that implicate the Asian/Taiwan genogroup of E. canis. It therefore seems unlikely that E. canis will be of public health concern in Australia.

It is fortunate and, indeed, somewhat surprising, that Australia remained free of CME for so long, considering the abundant population of brown dog ticks across the north. Prior to 2020 there were occasional reports of E. canis-positive serological tests in dogs from northern and central Australia, yet follow-up investigations using PCR testing were negative. Seropositivity in these individuals was attributed to false positive results, likely cross-reactions with the closely related Anaplasma platys, also associated with the same tick vector.³⁰ Mandatory pre-import screening (comprising immunofluorescence antibody test by registered laboratories) was established in September 1995,³⁰ and dogs testing positive were prohibited entry into the country.

Regardless, current surveillance data indicate that infected dogs (or ticks) have been detected in most states and territories, and the disease is now (2022) regarded as endemic. There is no hope of eradication, so veterinarians will need to add CME to the differential diagnosis of canine illnesses in northern Australia and in travelled dogs, and institute prompt treatment when cases are diagnosed.

The data that support this study cannot be publicly shared due to ethical or privacy reasons and may be shared upon reasonable request to the corresponding author if appropriate.

The authors declare no conflicts of interest.

This research did not receive any specific funding.