Pig-hunting dogs are an at-risk population for canine heartworm (Dirofilaria immitis) infection in eastern Australia

Background Canine heartworm disease, caused by Dirofilaria immitis, has global veterinary importance. In Australia, the prevalence of canine heartworm infection decreased markedly following the introduction of over-the-counter macrocyclic lactones. We aimed to estimate the prevalence of canine heartworm infection in at-risk populations of dogs in eastern Australia and analyse published prevalence data from Australia. Methods In total, 566 dogs from eastern Australia were tested for the presence of D. immitis antigen. Four cohorts were studied: pig-hunting dogs from Queensland (Cohort 1, n = 104), dogs from remote New South Wales (NSW) (Cohort 2, n = 332), urban pets from rural NSW (Cohort 3, n = 45) and ex-racing Greyhounds from Sydney, NSW (Cohort 4, n = 85). Serum samples were screened for D. immitis antigen using a reference laboratory microwell-based assay (DiroChek®) or a point-of-care immunochromatography test kit (Anigen Rapid®). Risk factors associated with the odds of D. immitis antigen seropositivity were identified using binary logistic regression models. Seropositive blood samples were tested for the presence and quantity of D. immitis DNA using a species specific real-time (q)PCR assay. A metanalysis of the Australian canine heartworm literature was conducted. Results The prevalence of dirofilariasis in pig-hunting dogs from Queensland (Cohort 1) was 12.5% (95% CI: 6.5–18.9%), with a subpopulation of dogs from Central Queensland having a prevalence of 21% (95% CI: 12.3–33.4%). Age was significantly associated with D. immitis antigen seropositivity (increased risk with increased age). The odds of being > 5 years versus ≤ 5 years was 3.7-times (95% CI: 1.1–12.5) greater in antigen positive versus antigen negative dogs. No D. immitis antigen positive dogs were detected in dogs from NSW (Cohorts 2–4). The Australian canine heartworm disease literature includes 98 peer-reviewed publications (1901–2019) with 30 studies reporting on D. immitis prevalence in dogs. Throughout the publication peak period (1980s), the primary antemortem diagnostic test was detection of microfilariae. Conclusions Canine heartworm infection in dogs used for pig hunting is a previously unexplored topic in Australia. Pig-hunting dogs are infected with canine heartworm in Queensland, Australia, placing pet dogs and cats at increased risk of infection.

Historically, Australia experienced high rates of canine heartworm disease across several regions in the 1970s and 1980s. During this era, the prevalence of canine heartworm in Queensland, Northern Territory and New South Wales (NSW) was reported to range from 30% to 100%, with the tendency for higher prevalence to occur in the tropical and subtropical regions of northern Australia [5][6][7][8][9][10]. Since then, the prevalence of canine heartworm in Australia has decreased progressively. By the early 1990s, a wide range of once-a-month oral or 'spot-on' macrocyclic lactones (MLs) for canine heartworm prevention had become available in Australia. These preventatives did not require a prescription and could be purchased across the counter from veterinary clinics, pet stores and (later) from online retail outlets and supermarkets [11,12]. In 2016, despite these measures, an apparent re-emergence of D. immitis infection and canine heartworm disease was documented in pet dogs in Central Queensland [13]. Apart from the impact of prophylactic therapy, factors affecting the prevalence of canine heartworm disease in Australia are incompletely understood.
For D. immitis to successfully (re)emerge and sustain its life-cycle, certain pre-requisites need to be met. First, the parasite is mosquito-borne. In Australia, the competent vectors have been shown to be mosquitoes of the genera Aedes, Culex and Anopheles [10,14,15]. Secondly, there needs to be an absence of prophylactic treatment with MLs, either through poor or absent client compliance or inadequate dosing, as this probably represents the most common reason for canine heartworm infection globally [16][17][18]. Finally, there needs to be competent species acting as reservoirs for D. immitis such as wild dogs, dingoes and owned dogs not receiving prophylaxis [19]. Identifying and monitoring canine cohorts that do not receive regular heartworm prevention (at-risk populations) is imperative, particularly in regions with low endemicity. One possible group in Australia is pig-hunting dogs. Hunting feral pigs (known as hogs in the USA) is a popular activity in Australia, with an estimated 156,000 dogs kept specifically for this purpose [20] (Fig. 1). Recent enquiries into the health and welfare of pig-hunting dogs in Australia and New Zealand suggested they were at high risk of exposure to numerous infectious diseases and received minimal veterinary preventive care [20,21].
In this study, we aimed to identify at-risk populations of dogs based on a history of incomplete canine heartworm (D. immitis) prevention. To do so, we used D. immitis antigen testing and positive samples were tested for the presence of D. immitis DNA. Four populations of dogs with moderate or no canine heartworm prevention were tested, including pig-hunting dogs, remote community dogs, owned urban dogs and ex-racing Greyhounds. For the cohort of pig-hunting dogs, reference laboratory D. immitis antigen testing was performed to evaluate the accuracy of a point-of-care immunomigration D. immitis antigen test. The context of canine heartworm in Australia is explained using a historical review (metanalysis) of published data.

Dog cohorts from Queensland and New South Wales, Australia
Four canine cohorts in Australia were sampled to determine the presence of canine heartworm (D. immitis) antigen (Fig. 2, Table 1). In total, 566 dogs were sampled. Whole blood samples were separated into serum and stored at − 20 °C prior to testing at the Sydney School of Veterinary Science (SSVS), University of Sydney.
Cohort 1 [Pig-hunting dogs (n = 104)] were sampled from August to November 2018 from towns across Central and North Queensland, Australia. Locations included Sarina, Clermont, Proserpine, Charters Towers, Innisfail, Tully, Malanda and Atherton. Dogs sampled had been used for hunting feral pigs (hogs), and generally were large, mixed breed dogs living outdoors in kennels or yards. Based on currently unpublished survey data from the owners, only intermittent canine heartworm prophylaxis was used, if at all. Cohort 2 [Remote community dogs (n = 332)], predominantly crossbreeds, were sampled from six communities across remote and very remote western New South Wales between 2016-2018. Dogs were recruited from the participants of Royal Society for the Prevention of Cruelty to Animals (RSPCA) NSW Community Companion Animal Health Programs, a service provided for people on low incomes, many of whom identify as Aboriginal. These dogs have limited or no access to veterinary treatment including canine heartworm prophylaxis.
Cohort 3 [Owned urban dogs (n = 45)] were sampled between June and September 2018 in Goulburn, a small inland city in NSW, Australia. These dogs were all pets visiting a local veterinary clinic, none of which were receiving heartworm prophylaxis from their owners at the time of sampling.
Cohort 4 [Adopted ex-racing Greyhounds (n = 85)] were sampled between January 2018 and November 2019 in the Greater Sydney Region, NSW. These dogs were all part of Greyhound Adoption Program (NSW) Inc, having retired from racing and currently living as pet dogs. More than half of these dogs (46/85, 54%; self-reported by owners) were receiving monthly heartworm preventatives, although their treatment history was not known in any more detail.

Testing for Dirofilaria immitis antigen
Serum samples from Cohorts 1-3 were tested using DiroChek ® Heartworm Antigen Test Kits (Zoetis, Parsippany, USA), a qualitative microwell-based enzyme-linked immunosorbent assay (ELISA) for the detection of adult female D. immitis antigen. DiroChek ® is intended for use by diagnostic laboratories and was used according to manufacturer's instructions; 5 min after the addition of the final solution, the 96-well plate or 8-well strip with manufacturers positive and negative controls were visually examined for a clear to blue color change. All blue colored wells, regardless of intensity, were considered antigen positive or, if no blue color was detected, samples were considered to have 'no detectable antigen' (NDA).
Serum samples from Cohorts 1 and 4 were tested using point-of-care Anigen Rapid ® Canine Heartworm (CHW) Ag 2.0 Test Kits (BioNote, Gyeonggi-do, Korea) a chromatographic immunoassay for the detection of D. immitis antigen. Sensitivity and specificity of Anigen Rapid ® CHW Ag 2.0 for D. immitis antigen have been reported as 99.5% and 94%, respectively [22]. The Anigen Rapid ® kit used in the study by Henry et al. [22] is identical to the one marketed and sold in Australia (Mark Thacker, Life Bioscience, Oakleigh, Australia, personal communication).
Knott's test for the identification of D. immitis microfilariae (Mff ) was not performed due to the logistical constrains, remote sampling locations and lack of laboratory resources in the field.

Dirofilaria immitis specific real-time PCR
Blood clot samples (200 μl) from D. immitis antigen positive dogs and a random selection of NDA D. immitis dogs (Cohort 1) were used for molecular studies. Genomic DNA was isolated using PowerMag ® Blood DNA/RNA Isolation Kits (Qiagen, Chadstone, Australia) optimised for the KingFisher ® Duo Prime Purification System (Thermo-Fisher Scientific, Scoresby, Australia). DNA isolation was performed as a batch of 11 blood samples and an extraction blank. DNA eluted in 100 μl was stored in aliquots at − 20 °C prior to analysis.
In Cohort 1, antigen status was significantly (P = 0.03) associated with age: the odds of being > 5 years versus ≤ 5 years was 3.73-times (95% CI: 1.1-12.5) greater in seropositive versus seronegative dogs (binary logistic regression models). Seropositivity was not associated with sex (P = 0.37; male versus female) or type (P = 0.6; Bull Arab and crossbreds versus others). Location (n = 8) as a fixed effect was not significantly (P = 0.54) associated with antigen status. However, location categorised as coastal versus inland was significantly (P = 0.03) associated with serostatus: the odds of being coastal versus inland was 3.89-times (95% CI: 1.2-12.8) higher in antigen positive versus antigen negative dogs. Finally, both age and location were significantly associated with serostatus when included in the same logistic regression model ( Table 2). This model adequately fitted the data (Hosmer & Lemeshow test, χ 2 = 1082, df = 2, P = 0.58), while the interaction between age and location was not significant (P = 0.30).
To determine if the D. immitis antigen positive pighunting dogs were a source of D. immitis for mosquitoes (i.e. that they were microfilaraemic), we used available blood clots from 11/13 antigen positive dogs and a selection (n = 9) of antigen-negative dogs (Table 3). DNA was successfully purified and canine GAPDH amplified (C q values: 21.8-31.5) from 20 canine blood samples. Blank DNA samples remained negative in the canine GAPDH assay. Seven Cohort 1 samples (7/11, 64%) from D. immitis antigen positive dogs were considered positive for D. immitis microfilariae DNA ( Table 3). The calculated microfilariae concentration ranged from < 1 to 22,050 microfilariae/ml ( Table 3). Four of 11 samples returned negative results (0 to < 5) for D. immitis DNA for samples that were D. immitis antigen positive (Table 3). These negative results are consistent with a concentration of microfilariae < 5/ml of blood, because we used 200 µl of blood for DNA isolation and thus required at least 1 microfilaria for a positive DNA result. All antigennegative samples (n = 9) yielded negative results for the presence of D. immitis microfilariae DNA (Table 3), as expected. For the D. immitis DNA positive samples, DNA sequencing of the cox1 PCR amplicon was 100% identical with the D. immitis cox1 gene reference sequence (Gen-Bank: NC_005305).
Searches to determine the historic prevalence of canine heartworm disease in Australia yielded 98 peer-reviewed publications spanning the years 1901-2019 ( Figure 3). After applying our inclusion criteria, 32 studies remained ( Figure 3, Table 4). Of these, 30 studies reported on D. immitis prevalence in dogs, and two studies included data on D. immitis prevalence in red foxes (Table 4). Most publications reporting D. immitis in canids were published during 1966-2005, with a peak of 12 publications during the 1980s (Figure 3). Throughout this peak period, the primary antemortem diagnostic test was detection of microfilariae (Table 4).

Discussion
The only canine heartworm (D. immitis) antigen positive dogs detected in our survey of 566 dogs were pig-hunting dogs from Queensland. All 13 D. immitis antigen-positive dogs, including one domiciled in Far North Queensland at the time of testing, had spent time in the Mackay and Whitsundays regions of Central Queensland. Recent research by Nguyen et al. [13] demonstrated the presence of D. immitis in this region in dogs visiting veterinary practices. More than one-fifth of the pig-hunting dogs tested in Central Queensland were D. immitis antigen-positive. The current prevalence of D. immitis antigen seropositivity in pig-hunting dogs in this region is comparable to the historical D. immitis prevalence data in pound dogs based on necropsy or microfilaraemia in Queensland during 1960-1990 [6, 7, 10, 25]. A significantly higher prevalence of D. immitis antigen seropositivity in dogs older than five years of age compared to  younger dogs has previously been reported in Australia, and is a consequence of increased likelihood of infection with the cumulative exposure to more mosquito bites throughout life [6,26]. In a large Australian study [10], no dog younger than 9 months showed presence of D. immitis microfilariae, and male dogs were more often positive then females. Possibly due to smaller samples sizes of atrisk dogs in our study, there was no difference observed in the sex of the dog in relation to antigen seropositivity. Pig-hunting dogs are thought to receive minimal veterinary and preventative health care and often travel widely with their owners during hunting expeditions [20,21]. Pig-hunting dogs are generally robust, large breed, shortcoated dogs that live outdoors [20], increasing their risk of being bitten by mosquitoes and thereby developing canine heartworm disease. Since they often reside close to pet dog populations, it will be important to investigate their potential as a reservoir for pet dogs sharing the same geography and receiving incomplete or lax canine heartworm preventative therapy.
In this study, we principally used the reference laboratory test, DiroChek ® Heartworm Antigen Test Kit, for detection of adult D. immitis antigen. Courtney et al. [27] reported DiroChek ® sensitivity to be 86% and specificity 97% in the USA. More recently, Henry et al. [22] re-evaluated the accuracy of DiroChek ® testing in the USA using a cohort of dogs with known D. immitis status, reporting test sensitivity of 99% (95% CI: 96.4-99.9%) and specificity of 96% (95% CI: 86.3-99.5%), including a cohort of 50 dogs with low worm burdens (1-5 female D. immitis; Mff in 58% of dogs). In Australia, the DiroChek ® has been tested extensively for its ability to detect D. immitis during the prepatent period and low heartworm burdens [5,28,29]. The DiroChek ® was validated in 1987 by testing 100 dogs from Sydney, NSW and immediately performing necropsy examination, with the test correctly identifying 15/24 infected dogs (sensitivity 62.5%) and 74/76 uninfected dogs (specificity 97.4%) [29]. Dogs returning a false negative result using DiroChek ® included four dogs with one male heartworm, three dogs with two male worms, one dog with one male and two female worms and one dog with five male worms [29]. In another group of 100 pound dogs from Sydney examined by necropsy, 8 dogs that were negative using DiroChek ® had adult heartworms at necropsy, four with prepatent infections and four with infections of three or less worms, all of one sex [5]. Similar results were obtained in a study that validated DiroChek ® in Brisbane, Queensland using a 1 9 8 6 -1 9 9 0 1 9 9 1 -1 9 9 5    [28] and Sydney [5], respectively. In our study, we applied species-specific PCR to detect the presence of microfilarial DNA, which was absent in four of DiroChek ® antigen positive dogs. These findings confirm that detection of microfilariae by filtration (Difil) or concentration (Knotts) techniques are complimentary to DiroChek ® to identify microfilariae-positive dogs prior to antigen testing and should be performed whenever screening at-risk dogs for D. immitis. Detection of low burdens of D. immitis in dogs is difficult, because DiroChek ® sensitivity can be variable (53% in [28]; 62.5% in [29]; 86% in [27]) due to unknown worm burden distribution in the dog population. In areas of recent heartworm introduction, newly infected dogs will have no Mff due to the prepatent period of at least six months, single sex infections or burdens with less than two female heartworms. Additionally, the formation of immune complexes can further complicate diagnosis and detection of D. immitis antigen [30,31].

Conclusions
The fall in the prevalence of D. immitis in Australia since the introduction of MLs (monthly tablets and chews, top-spots and yearly depot injections) contrasts to the situation in most of the USA and Canada where the prevalence and extent of heartworm disease has increased despite the availability of these drugs [32][33][34]. In Australia, preventative drugs are widely available at veterinary clinics, pet stores, supermarkets and on the internet via online veterinary wholesalers, providing easy access to prophylactic therapy [13,35]. More restricted availability of MLs in North America, reduced competition in the marketplace and possibly higher cost (as a result) may in part explain why heartworm disease has been virtually eradicated from NSW in Australia, in contrast to most of the USA [36]. Additionally, NSW has been in extended drought conditions for three years, which presumably has reduced the availability of mosquito vectors. In contrast to NSW, where we did not demonstrate canine heartworm infection in dog cohorts receiving no or intermittent prevention, Queensland pig-hunting dogs are identified as the at-risk population with 21% D. immitis antigen prevalence in Central Queensland.