During November 2015–March 2017, we collected 460 wild canids (416 coyotes and 44 foxes) from 25 of the 29 southern Ontario PHUs and tested them for the presence of E. multilocularis DNA by using MC-PCR. We collected 205 wild canids (183 coyotes and 22 foxes) in the first collection period and 255 wild canids (233 coyotes and 22 foxes) in the second. During both collection periods, we collected >80% of the wild canids during the months of January, February, and March. Hunters and trappers consistently reported low fox population numbers throughout the duration of the project, resulting in a low number of sampled foxes. No canids were collected from PHUs 3535, 3553, 3555, and 3595 (Figure 1). Overall, 23% (95% CI 20%–27%) of wild canids, from 18 PHUs, tested positive for E. multilocularis (Table). Among coyotes, 24% (95% CI 20%–28%) tested positive; 21% (95% CI 11%–35%) of foxes tested positive. Raw prevalence ranged from 0% to 100% among PHUs (Figure 2). Smoothed prevalence by PHU estimates ranged from 4% to 46% (Table) and varied on a gradient of higher to lower prevalence from the southwestern to northeastern regions of southern Ontario (Figure 2).

The spatial scan test, georeferenced by PHU, detected a significant spatial cluster of high prevalence of infection (relative risk 2.26; p = 0.002) centered in PHU 3534, consisting of 10 contiguous PHUs (3527, 3531, 3534, 3536, 3537, 3544, 3546, 3552, 3554, and 3565) (Figure 2). The prevalence of infection among the 205 wild canids included in the cluster was 34% (95% CI 28%–40%). A second spatial scan test, using data georeferenced to each wild canid’s location of origin, detected a significant spatial cluster of high prevalence of infection (relative risk 2.53; p = 0.001), with a radius of 120 km also centered in PHU 3534. The prevalence of infection among the 216 wild canids included in the cluster was 34% (95% CI 28%–41%). No statistically significant nonoverlapping secondary high-risk spatial clusters were identified at either level of spatial resolution.

This report describes the prevalence of E. multilocularis infection in wild canids in Ontario. Because Ontario was previously considered free of the E. multilocularis tapeworm, we were surprised that 23% of the wild canids (107/460) tested positive for the parasite in our study. This finding is comparable to recent wild canid prevalence data from Edmonton, Alberta, Canada (9), where the E. multilocularis tapeworm has been recognized for decades (32). We anticipated that infection among wild canids would be confined to the region surrounding the western shores of Lake Ontario in southern Ontario, where the aforementioned cases of AE were observed. However, our findings indicate that E. multilocularis infection in wild canids is widely distributed across the western, central, and eastern regions of southern Ontario, with a high prevalence hotspot consisting of 10 PHUs in the western-central region (Figure 2). The combination of the high prevalence and wide geographic distribution of infection suggests that the E. multilocularis tapeworm was not a recent introduction into Ontario. In addition, to the authors’ knowledge, only 1 other study has investigated E. multilocularis tapeworms in the province; a survey of 302 red foxes from southern Ontario during 1979–1980 did not detect evidence of the parasite (33). Therefore, E. multilocularis tapeworms probably were introduced sometime after 1980.

How E. multilocularis tapeworms were introduced into Ontario is unclear. However, the spatial pattern of high infection prevalence among wild canids in the southern PHUs that border the northern shores of Lake Erie (Figure 2) might indicate a natural northeastern expansion from Michigan, a known endemic area (33). Also, the importation of dogs from endemic areas in North America or Europe, without any requirement for cestocide treatment, might have contributed to the introduction of E. multilocularis tapeworms into the province. Notably, molecular characterization of the metacestode stage from 1 of the southern Ontario dogs diagnosed with AE without travel history was consistent with E. multilocularis tapeworms of possible European origin (1), whereas another appeared to be North American in origin (K. Gesy, pers. comm., 2017 Dec 14). These findings strengthen the possibility that an importation event occurred, perhaps in addition to a natural range expansion. However, the meaning of this information remains unclear because data concerning the epidemiologic importance of individual strain variants are limited (34,35).

We measured an infection prevalence of 34% (95% CI 28%–40%) among wild canids within the southern Ontario hotspot. Consequently, a question of public health importance is to what extent the human population in southern Ontario is at risk for human AE. Across the endemic countries in Europe, where the prevalence of E. multilocularis infection in wild canids ranges from <1% to >50% (1), human AE is rare; the overall average annual incidence in these countries ranges from 0.03 to 0.3 cases/100,000 residents (36). However, substantial variation in risk exists across regions. For example, in areas with consistently high prevalence in wild canids (i.e., 35%–65% prevalence), the annual incidence of human AE can be as high as 8.1 cases/100,000 residents (37,38), which is similar to the prevalence estimates among wild canids in the southern Ontario hotspot that we describe. Furthermore, the location of the infection cluster encompasses multiple urban areas with human population densities of up to 1,700 residents/km² (23). Therefore, transmission of E. multilocularis tapeworms should be considered a public health risk.

In areas endemic for E. multilocularis tapeworms, dog ownership has been associated with increased risk for human AE (37,39–41). Dog ownership might entail various human and dog behaviors that might lead to an increased risk for human infection with E. multilocularis tapeworms. These behaviors include leaving dogs outside unattended, walking dogs without a leash, allowing dogs to consume rodents, and inconsistent deworming of dogs (40). As such, monthly treatment with praziquantel is recommended for dogs that consume rodents in AE-endemic areas to prevent patent intestinal infections and therefore mitigate the risk for transmission to humans (36). The same is also recommended for dogs with hepatic AE because such dogs might also have concurrent intestinal infections (42). Thus, even in instances of canine hepatic AE, a follow-up investigation of possible exposure to E. multilocularis tapeworms for in-contact humans is warranted (43).

As of January 1, 2018, E. multilocularis infection was designated a reportable disease in animals in Ontario (44). Veterinarians and diagnostic laboratories are required to report animal cases directly to their local PHUs to minimize potential risks to human and public health. Furthermore, as of May 1, 2018, E. multilocularis infection in humans was designated a disease of public health importance (i.e., a disease that must be reported) in Ontario (45). Although human AE was not reportable before 2018, data from the Canadian Institute for Health Information indicate that >3 cases of human AE have been diagnosed in Ontario since 2014 (46); however, these data do not include information regarding patient travel or exposure histories. Therefore, whether these cases were locally acquired is unknown. Designating E. multilocularis infection as reportable in humans and animals is potentially important because, in AE-endemic areas (i.e., Europe), a large proportion of the economic burden associated with human AE is attributable to patients typically being diagnosed in the late stages of the disease, requiring lifelong chemotherapy and occasionally interventional procedures (e.g., percutaneous biliary and centroparasitic abscess drainage) (5,36). Therefore, the ability to anticipate E. multilocularis exposure and to diagnose early-stage human AE is essential to reduce the need for long-term treatment, thereby minimizing the economic burden associated with the disease. A limitation of having the infection reportable only in humans is that, given the long clinical incubation period of AE in humans, other persons potentially at risk would likely have been infected years earlier. Thus, in areas where E. multilocularis infection is endemic, a One Health surveillance approach that also requires mandatory reporting of E. multilocularis infection in animals to public health authorities could improve rates of prompt investigation of suspected exposure in persons and lead to earlier diagnosis.

Our study has several limitations. First, sample collection depended on carcass submission from hunters, trappers, and the Ontario Ministry of Natural Resources and Forestry. Although this convenience sampling allowed us to achieve a large sample size, it resulted in underrepresentation of parts of the study area. Because this approach resulted in PHUs with low sample sizes and thus potentially unreliable prevalence estimates, we used a Bayesian estimation method to smooth prevalence estimates. Second, the MC-PCR method used to detect E. multilocularis DNA is imperfect, having an overall sensitivity of 88% and specificity of 99% (25,27). However, we chose to employ this method because it is suitable for large-scale screening and its performance is comparable to what is considered the reference standard for diagnosis of infection in wild canids, the sedimentation and counting technique (26).

Our findings underscore the importance for continued surveillance among wild canids within and outside endemic areas in North America to monitor the spread of E. multilocularis tapeworms. In addition, an understanding of the prevalence of intestinal infections among dogs in AE-endemic areas would provide valuable information on potential exposure risk in human populations. Collectively, the data would guide public health and veterinary efforts in the development of targeted prevention strategies in this region.