A One Health approach to rabies management in Manitoba, Canada

Abstract

A One Health approach was developed in the province of Manitoba in 2014 to manage human and domestic animal exposures to rabies. Manitoba Rabies Central is a collaboration of 3 provincial departments responsible for animal, human, and environmental health. Since the inception of the program 537 samples from animals suspected of rabies and causing an exposure to a human or domestic animal have been evaluated with 11.3% testing positive, 85.7% testing negative, and 3.0% being unfit for testing. Most of the positive rabies test results came from skunks (52.0%), which accounted for 12.5% of submissions. Dogs and cats accounted for 52.5% of submissions; however, only 18.9% of these animals tested positive for rabies. Domestic animals were more likely to be exposed to a rabid animal (most commonly skunks) than were humans. Humans were more likely to be exposed to dogs and cats (regardless of rabies test result).

Introduction

Rabies is a virus in the genus Lyssavirus in the family Rhabdoviridae, and the causative agent of rabies viral encephalomyelitis affecting mammals. Rabies has the highest case fatality of any known infectious agent (1). Globally most rabies cases are transmitted by domestic dogs, mainly in Asia and Africa (2). In 2017, 2.9% of animals that tested positive for rabies in Canada were dogs (n = 7), and 92.5% were wildlife (n = 221), with the remaining positive samples coming from cats (n = 4) and cattle (n = 7) (3). These results highlight the efforts to mitigate the risk of rabies from dogs in Canada through practices such as vaccination, population control, mandatory observation periods, wildlife vaccine baits, and risk reduction surrounding suspect cases.

In Canada, rabies is a reportable disease, and as of 2014 each province and territory in Canada became responsible for managing suspected and confirmed rabies exposures. Prior to 2014 rabies sampling was managed at the federal level by the Canadian Food Inspection Agency (CFIA). In Manitoba (MB) a One Health approach was developed under the working group of Manitoba Rabies Central (MRC). One Health is a method which guides health policies and practices across multiple sectors through a collaborative approach (4). Manitoba Rabies Central is a collaborative effort of Manitoba Agriculture (MBAg), Manitoba Health, Seniors, and Active Living (MHSAL), and Sustainable Development (SD). They are responsible for domestic animal health, human health, and environmental and wildlife health, respectively. The inclusion of multiple stakeholders ensures that health risks can be accounted for and addressed with consideration for each stakeholder’s concerns. The One Health approach also uses resources more efficiently by ensuring work is not erroneously duplicated, and that goals are synergistic instead of oppositional.

The overall objective of MRC is to maintain zero cases of human rabies while minimizing domestic animal cases in MB through effective risk management of possible exposures of humans and domestic animals to rabies virus. Specifically, this includes implementation of efficient reporting mechanisms for suspect cases, science-based risk assessments, and reliable sample collection and testing. Manitoba Health, Seniors, and Active Living is responsible for managing potential human exposures together with regional health authorities. Manitoba Agriculture is responsible for managing potential domestic animal exposures, and SD is responsible for providing expert advice on wildlife management and the sylvatic rabies cycle. While each group has its specific responsibilities, regular communication about cases occurs, as there is often an overlap of human, animal, and wildlife in a potential exposure. Communication is both informal in regard to individual cases, and formal in monthly meetings to discuss ongoing or new issues. The overall objectives of this study were to: i) describe the function of MRC using a One Health approach; ii) evaluate sample submissions and determine the association of: rabies positive samples, days to obtain test results, and human exposure with the predictors: health region, species, and month/year of submission; and iii) determine the success of the program through efficiency of reporting results, and ability to maintain MB’s record of zero reported human cases of rabies.

Materials and methods

Sample criteria, collection, and testing

Animals suspected of clinical rabies infection to which either a human or a domestic animal has been exposed are considered for testing in MB based on the category of exposure that occurred as defined by the World Health Organization (WHO) (5). These categories are:

• Category I — feeding or touching animals; animal licks healthy intact skin;

• Category II — animal nibbling uncovered skin or causing minor scratches/abrasions without bleeding; and

• Category III — a single or multiple transdermal bites or scratches, licks on broken skin, or contamination of mucous membranes with the animal’s saliva or contact with bats. Animals suspected of clinical rabies and the cause of a Category II or III exposure, and which cannot be quarantined to be observed for symptoms of rabies, are euthanized in order to retrieve the brain and brainstem for testing. In the instance of wildlife Category II or III exposures, a sample is retrieved whenever possible. Domestic animals exposed to rabies virus (and not the cause of an exposure) can be quarantined for a period of time, which is dependent on previous rabies vaccination, and booster vaccination following the exposure.

When a suspected exposure occurs, the initial report may be to a primary care doctor or veterinarian, to the provincial healthcare phone line Health Links, to a public health official, or directly to the MRC phone line. Doctors and veterinarians are advised to contact MRC if they suspect a rabies exposure. Once an exposure is determined and a sample can be obtained, a case identification number is assigned by MHSAL and MBAg coordinates sample collection and submission. Data collected on each sample include: human and/or animal exposure, species of animal suspected of causing rabies exposure, date of exposure, health region in which the exposure occurred (Figure 1), rabies virus test result and variant if positive, and time between generation of the case identification number and results being received (days to receive test result). Churchill, MB is considered part of the Winnipeg Regional Health Authority; however, for the purposes of this study it was considered as part of the Northern region based on its location and likelihood of risk factors being more similar to those of the Northern region than Winnipeg. First Nations and Inuit Health (FNIH) are a separate health region as their healthcare is federally administered, compared with provincially administered healthcare. First Nations and Inuit Health administered healthcare is located in the Northern and Interlake Eastern regions.

Figure 1.

Health regions of Manitoba.

A trained sample collector collects and ships samples for testing. Animals weighing < 500 g are submitted whole. Animals weighing > 500 g but < 100 kg have their entire head submitted, and animals > 100 kg have their brain and brainstem submitted. Samples are packaged to comply with transportation of dangerous goods guidelines for UN3373 — Biological Substance, Category B, and packaged with rigid icepacks based on outside ambient temperature.

Samples are shipped using Purolator air and ground transport to the Lethbridge Laboratory, or Ottawa Laboratory Fallowfield, run by the CFIA. Diagnostic tests are performed with the same techniques at both laboratories and include: fluorescent antibody test, the formalin-fixed assay test, and tissue culture inoculation test. The 2 latter tests are not commonly used. Both laboratories have quality assurance programs and are ISO 17025 accredited. Results are reported as positive, negative, or unfit (unfit due to tissue degradation or incomplete sample). Unfit samples are treated as positive in terms of clinical case management due to the risk assessment already indicating the animal was at risk of having a clinical rabies infection and causing an exposure.

Statistical analysis

Data entry was in Microsoft Excel (Microsoft, Redmond, Washington, USA) in a database maintained by MBAg, and data analyses were performed in Stata 15.1 for Windows (StataCorp, College Station, Texas, USA). Descriptive statistics were evaluated for the variables: species of animal causing exposure, region in which the exposure occurred, year of exposure, month of exposure, rabies test result, rabies variant, human exposure, and time to receive results. The 2 outcomes of interest were the rabies test result and number of days to receive test results. The outcome of rabies test result was evaluated as a 2-category outcome, positive or negative, with unfit test results considered positive.

Chi-squared analyses were used to evaluate if the outcome of the rabies test results varied between species, month/year exposed, human exposure, and region. Chi-square testing was also used to evaluate if species, region, and month and year of submission were significantly associated with human exposure. Analysis of variance (ANOVA) testing was used to evaluate the difference in time to get test results (in days) by rabies test result (to evaluate if unfit samples took longer to get to the laboratory), by region, by species, and by month and year of submission. ANOVA models were evaluated for variance homogeneity, outliers, and normal distribution by use of model residuals. In all statistical tests a significant P-value was ≤ 0.05.

Results

Between April of 2014 and March of 2018, 537 samples were submitted for testing with 11.3% of samples testing positive for rabies virus (n = 61), 85.7% testing negative (n = 460), and 3.0% being unfit for testing due to decomposition, damage, or incomplete sample (n = 16) (Table 1). For the purposes of program evaluation the unfit samples were included in the category of positive test results due to these cases being treated as positive in regards to clinical management (Table 2).

Table 1.

Year and month of suspect rabies sample submissions between April 2014 and March 2018 in Manitoba.

	Fiscal year
Month	2014–2015	2015–2016	2016–2017	2017–2018	Total
January	9	12	15	13	49
February	5	15	6	7	33
March	8	12	9	10	39
April	12	5	14	8	39
May	11	8	9	21	49
June	6	15	8	14	43
July	12	19	7	13	51
August	11	13	22	3	49
September	20	8	16	13	57
October	9	17	13	11	50
November	11	10	13	8	42
December	10	8	11	7	36
Total	124	142	143	128	537

Table 2.

Chi-square analysis of species as a predictor for 537 rabies sample results in Manitoba between April 2014 and March 2018.

Species of animal	Rabies result negative n (%)	Rabies result positivea n (%)
Dog	139 (93.9)	9 (6.1)
Bat	22 (73.3)	8 (26.7)
Cattle	58 (86.6)	9 (13.4)
Cat	129 (96.3)	5 (3.7)
Horse	17 (94.4)	1 (5.6)
Other domestic speciesb	12 (92.3)	1 (7.7)
Other wildlife speciesc	56 (93.3)	4 (6.7)
Skunk	27 (40.3)	40 (59.7)
Total	460 (85.7)	77 (14.3)
Pearson Chi-Square (7) = 140.97, P < 0.005

^aPositive test results include all unfit samples n = 16 (dog n = 5, bat n = 6, cattle n = 1, skunk n = 2, other wildlife [(muskrat and raccoon) n = 2]).

^bOther domestic species that was positive was a goat. Domestic animals submitted were: sheep (n = 5), goat (n = 6), alpaca (n = 1), and pig (n = 1).

^cOther wildlife that were positive were a red fox and an Arctic fox. Wildlife included: raccoon (n = 20), fox (n = 9), muskrat (n = 6), mink (n = 6, 5 were farmed mink), coyote (n = 4), squirrel (n = 3), wolf (n = 2), beaver (n = 2), and 1 of each of the following: groundhog, bobcat, badger, black bear, weasel, rat, vole, and rabbit.

Most samples came from the Prairie Mountain (34.1%), Southern Health (28.8%), and Interlake Eastern (21.6%) regions, with the remainder coming from the Winnipeg region (9.5%), FNHI (4.3%), and Northern region (1.7%). The majority of samples submitted were from dogs (27.6%) and cats (24.9%), and the remainder of samples were from cattle (12.5%), horses (3.3%), other domestic animals (2.4%), skunks (12.5%), bats (5.6%), and other wildlife (11.2%).

The variables significantly associated with the rabies test results were species of animal tested and region of submission (Tables 2, 3). Skunks were more likely to have positive sample submissions; dogs and cats were less likely to have positive sample submissions (Table 2). Variant testing was attempted on 50 of the 61 positive samples, with the reminder still awaiting testing at time of publication. Western skunk variant was identified in 45 samples, 1 sample was an Arctic fox variant (from an Arctic fox sample), 2 samples were unidentifiable due to insufficient tissue (from a dog and a bat), and 2 samples were bat variants. The bat variants were the Lasiurus variant from a red eastern bat (Lasiurus borealis), and a silver-haired variant (Lasionycteris noctivagans), which was identified in a cat. The untested samples by species were: skunks (n = 6), cattle (n = 2), dogs (n = 2), and a red fox (n = 1).

Table 3.

Chi-square analysis of region as a predictor for 537 rabies sample results in Manitoba between April 2014 and March 2018.

Region in Manitoba	Rabies result negative n (%)	Rabies result positivea n (%)
Winnipeg	49 (96.1)	2 (3.9)
First Nations and Inuit Health	22 (95.7)	1 (4.3)
Interlake Eastern	107 (92.2)	9 (7.8)
Northern	6 (66.8)	3 (33.3)
Prairie Mountain	154 (84.2)	29 (15.8)
Southern	122 (78.7)	33 (21.3)
Total	460 (85.7)	77 (14.4)
Pearson Chi-Square (5) = 19.54, P = 0.002

^aPositive test results include all unfit samples, n = 16 (unfit samples by region were SH: n = 6, PM: n = 7, IE: n = 1, FNIH: n = 1, NR: n = 1).

The Southern Health region had 21.3% of their samples test positive, and the Northern region had 33.3%, which were higher than the overall positive of 14.3%; however, it should be noted the Northern region had very few samples in general (Table 3). There was no significant association of year of submission or month of submission with the rabies test result (P = 0.482, P = 0.407, respectively).

Compared to human exposures, domestic animal exposures were more likely to be caused by a rabid animal (Table 4). When considering total exposures regardless of the rabies test result, humans were more likely to be exposed to dogs or cats as opposed to wildlife, whereas domestic animals were more likely to suffer an exposure due to wildlife (Table 5). Approximately 2/3 of exposures in all regions were human exposures, except in FNIH regions, and Northern region where human exposures were more common (Table 6). Humans were also more likely to be exposed in March, June, October, and November compared with domestic animals (Table 7). No humans were reported by MHSAL to have developed clinical rabies infection between April 2014 and March 2018.

Table 4.

Chi-square analysis of human exposure by rabies test result for 537 rabies sample results in Manitoba between April 2014 and March 2018.

Exposed	Rabies result negative n (%)	Rabies result positivea n (%)
Domestic animal exposure onlyb	135 (73.4)	49 (26.6)
Human exposurec	325 (92.1)	28 (7.9)
Total	460 (85.7)	61 (14.4)
Pearson Chi-Square (1) = 34.43, P < 0.005

^aPositive test results include all unfit samples n = 16 (domestic animal only n = 6, human exposure n = 10).

^bDomestic animal exposures caused by rabid animals were caused by skunks (n = 34), cattle (n = 6), and one of each of the following, bat, horse, cat, goat, and fox. The unfit samples categorized as positive causing a domestic exposure were: skunk (n = 2), bat (n = 3), muskrat (n = 1). Domestic animals which were exposed to rabid animals were compromised of: dogs (n = 30), cattle (n = 6), horse (n = 1), goat/sheep (n = 3), cat (n = 3), mink (n = 1), and species not recorded (n = 2). Unfit samples categorized as testing positive, the domestic animals exposed were cat (n = 2), dog (n = 3), not recorded (n = 1).

^c18 cases where a human was exposed to a rabies positive animal, the species were: skunks (n = 6), dogs (n = 4), cats (n = 4), cattle (n = 2), bat (n = 1), and fox (n = 1). The remaining 10 unfit samples where humans were exposed and categorized as positive for rabies test results were: dog (n = 5), bat (n = 3), cattle (n = 1), and raccoon (n = 1).

Table 5.

Chi-square analysis of species as a predictor for human exposure for 537 rabies samples in Manitoba between April 2014 and March 2018.

Species causing exposure	Domestic animal exposed n (%)	Human exposed n (%)
Dog	13 (8.8)	135 (91.2)
Bat	11 (36.7)	19 (63.3)
Cattle	37 (55.2)	30 (44.8)
Cat	15 (11.2)	119 (88.8)
Horse	9 (50.0)	9 (50.0)
Other domestic species	6 (46.2)	7 (53.8)
Other wildlife species	35 (58.3)	25 (41.7)
Skunk	58 (86.6)	9 (13.4)
Total	184 (34.3)	353 (65.7)
Pearson Chi-Square (7) = 187.07, P < 0.005

Table 6.

Chi-square analysis of region as a predictor for human exposure for 537 rabies samples in Manitoba between April 2014 and March 2018.

Region of Manitoba	Domestic animal exposed n (%)	Human exposed n (%)
Winnipeg	15 (29.4)	36 (70.6)
First Nations and Inuit Health	0 (0.0)	23 (100.0)
Interlake eastern	38 (32.8)	78 (67.2)
Northern	5 (55.6)	4 (44.4)
Prairie mountain	68 (37.2)	115 (62.8)
Southern	58 (37.4)	97 (62.6)
Total	184 (34.3)	353 (65.7)
Pearson Chi-Square (5) = 15.82, P = 0.007

Table 7.

Chi-square analysis of month as a predictor for human exposure for 537 rabies samples in Manitoba between April 2014 and March 2018.

Month of exposure	Domestic animal exposed n (%)	Human exposed n (%)
January	20 (40.8)	29 (59.2)
February	10 (30.3)	23 (69.7)
March	6 (15.4)	33 (84.6)
April	21 (53.9)	18 (46.1)
May	19 (38.8)	30 (61.2)
June	9 (20.9)	34 (79.1)
July	21 (41.2)	30 (58.8)
August	21 (42.9)	28 (57.1)
September	18 (31.6)	39 (68.4)
October	14 (28.0)	36 (72.0)
November	11 (26.2)	31 (73.8)
December	14 (38.9)	22 (61.1)
Total	184 (34.3)	353 (65.7)
Pearson Chi-Square (11) = 23.11, P = 0.017

When domestic animals were exposed to rabid animals it was to a rabid skunk 69.4% of the time (n = 34, including 2 unfit samples) (Table 4). Dogs were the most common domestic animal exposed, and most often were exposed to skunks. In only 1 instance was a skunk observed with livestock associated exposure.

The average number of days from generation of a case identification number to receiving results was 3.5 d (median: 3 d, range: 1 to 15 d). Days to test result was natural logarithm (ln) transformed to meet model assumptions. When evaluating the days to receive results by program year, the ANOVA results were significant (P < 0.005, mean square error (MSE) 0.33, degrees of freedom for error (DFE) = 533), where fiscal year 2014–2015 was significantly different from all subsequent fiscal years with 3.8 d to receive results versus 2.8 d for all subsequent years. Region was also significantly associated with days to receive test results (P = 0.010, MSE 0.34, DFE = 531). The Northern region took 5.8 d to receive test results compared with ~3 d from Interlake Eastern, Prairie Mountain, Southern Health, and Winnipeg regions. Samples from FNIH were not significantly different from any other regions as there was an intermediary time of 3.6 d to receive test results. The data were well approximated by ANOVA model assumptions in both aforementioned models. Month and species were not significantly associated with days to receive test results.

Discussion

In MB, between 2014 and 2018, skunks accounted for 12.5% of the samples submitted and 52.0% of the rabies virus positive samples. The western skunk variant accounted for 93.8% of samples that had variant testing results. This suggests that the greatest risk for exposure to rabies virus in MB is through skunk exposure, either directly or through cross species transmission. In 2017, the CFIA reported that out of the 239 positive rabies sample submissions in Canada, 25.9% were from skunks (3). Results were similar in 2016 with 29.3% of positive samples coming from skunks in Canada (6). Manitoba may see higher proportions of skunks accounting for rabies virus positive cases because it has a low prevalence of rabies in raccoons and bats. For example, Ontario reported 86 cases of rabies in raccoons in 2017 (57.7% of positive samples from Ontario), and 20 cases of rabies in bats (13.4% of positive samples from Ontario) (3). Whereas over the study period in MB, 2 bats tested positive for rabies (6 unfit bat samples were classified as positive), and no raccoons tested positive (1 was classified as positive due to unfit test results).

While Ontario differed in the proportion of positive rabies virus samples when compared to MB, Saskatchewan had similar levels of positive proportions of rabies in skunks as MB with 56.4% of positive samples (n = 57) from skunks between 2014 and 2017 (7). From the samples that tested positive in Saskatchewan (n = 101) during this time, 32.7% were from bats (n = 33), and there were no reported cases of rabies in raccoons (7). It is also noteworthy that the CFIA reports positive samples identified within their laboratories that are submitted based on risk-based surveillance. Therefore, other surveillance activities, such as active wildlife surveillance are not necessarily captured in the CFIA reports.

The variants noted in Manitoba were western skunk, Arctic fox, and bat, which are the known variants in the prairie provinces of Canada (1). No raccoon variants were identified, despite the presence of this variant in the neighboring province of Ontario (1). Between 1999 and 2011, 5 raccoons tested positive for rabies in MB, but no variant testing is available for these samples (8). While Ontario has a raccoon rabies variant, an east to west spread of the raccoon variant has not been noted in North America (9). Instead, raccoon rabies variant incursions into Canada (Ontario, Quebec, and New Brunswick), have come from distinct and separate south to north spread from the United States (9). In North Dakota, 7 raccoons tested positive between 1996 and 2017; however, no data are reported on the variant strain (10). Minnesota had 1 raccoon test positive for rabies between 1999 and 2017 (11), but no variant information is available. Based on the low numbers of raccoons with rabies in the southern neighboring States, and lack of raccoon variant data, the spread of raccoon variant rabies from south of MB appears low risk at this time.

While skunks account for most rabies virus positive samples in MB, they accounted for only 2.6% of human exposures overall. In Arizona, USA, 58% of human exposures to rabies positive wildlife over a 6-year period were via skunks (n = 44), indicating that human exposures to skunks are occurring more readily in this region (12). It is difficult to compare human exposure to positive wildlife in MB to those in Arizona, due to the small number which occurred in MB between April 2014 and March 2018. Additionally, the study in Arizona was only able to evaluate positive sample submissions, and the number of skunks and other wildlife samples was not known (12).

In MB, dogs and cats accounted for 18.9% of the rabies positive samples, most overall submissions (52.5%), and the most human exposures (72.0%). In Northern Quebec dogs accounted for 61% of all animals tested for rabies, and only 15% of dog samples were positive for rabies virus, which is similar to that of MB (13). In 2017, dogs and cats accounted for 4.6% of positive rabies samples in Canada as reported by the CFIA, but it is unknown what proportion of all samples submitted were from dogs and cats (3). The proportion of positive samples in dogs and cats in Canada may be lower than that of MB due to the aforementioned reasons regarding the relatively low numbers of wildlife rabies cases in MB, as well as the relatively high number of positive wildlife submissions from Ontario contributing to the overall numbers of positive submissions in Canada. Additionally, inclusion of unfit samples in the positive test result category for the data from MB in this study could have artificially inflated the number of dogs which were truly positive for the rabies virus in this study. It is noteworthy that 11.7% (inclusive of 1 unfit cattle test result) of rabies positive samples came from cattle in MB, and that most livestock are not seen with a skunk or other wildlife reservoir. Therefore, veterinary practitioners should consider rabies infection as a differential diagnosis in livestock with neurologic signs in MB.

The authors hypothesize that humans are more likely to be exposed to dogs and cats due to their proximity as pets and would explain why most human exposures are to these species. Pets also receive veterinary care and vaccinations making them less likely to develop rabies. In northern Quebec most wild animals submitted were positive and did not have human or domestic animal exposure, and submitted dogs mostly tested negative (13). The proportion of vaccinated animals in MB is unknown; however, in the neighboring province of Ontario, 52% of dogs and cats reported to have bitten humans were up-to-date on their rabies vaccines, 32% were not up to date, and the remainder had unknown vaccine history (14). A Canadian communicable disease report for Ottawa, Ontario in 2002 estimated that 61% of dogs and 53% of cats were up-to-date on their rabies vaccinations (15).

It is possible that a significant number of dogs and cats in MB are submitted for sampling due to the serious nature of rabies infections and wanting to mitigate any risk for human health by knowing the rabies status of an animal that bit a person. Anecdotally, MRC notes that unvaccinated, not up-to-date on vaccines, and occasionally fully vaccinated animals are submitted for testing when an observation period could have been observed instead of testing the animal. In a meta-analysis, between 11% to 34% and 10% to 18% of dogs are surrendered or euthanized for behavioral problems (16), and this could be the possible cause for the overrepresentation of dogs and cats being sampled in Manitoba since they were all involved in a biting incident. Interestingly, the use of post-exposure prophylaxis (PEP) is also reported to be higher in more affluent communities due to the ease of travel and availability of PEP, even when the risk of exposure to rabies is low (17), as is the case for humans exposed to dogs and cats in MB.

Bats had a higher proportion of unfit sample results than other species (6/16 unfit samples). Bats can be submitted if there is any suspicion of human or domestic animal exposure, and therefore an old bat carcass may occasionally be submitted, and possibly lead to an unfit sample. There was a trend that in the months of May to July more unfit samples were submitted; however, this was not statistically significant. This could be due to inappropriately packed samples and warm weather causing sample decay. It is possible that rabies in bats is overestimated in MB due to 20% of the bat submissions being unfit for testing and included in the positive test result category. Including unfit samples in the positive test result category allowed for evaluation of the program run by MRC since those samples were treated as positive in regard to clinical management of the exposed individual or animal; however, it is possible the suspected animals were negative for rabies. This potential misclassification could lead to bias in interpretation of results; however, the low number of unfit results across all submissions (n = 16), and the only notable difference being in bats, likely makes the bias minimal across other variables (region, human exposure, time of year, other species of animals). In future years the unfit samples should continue to be evaluated for any variability across species, region, time of year, and human exposure.

In FNIH the only samples submitted were associated with human exposure, and in all but one instance, dogs were the species submitted for testing. While it is possible only humans were involved in exposures, it is also possible that domestic animal exposures and exposures to wildlife were underreported. Most samples came from the southern regions of Interlake Eastern, Prairie Mountain, and Southern Health. Dogs and cats accounted for the majority of samples, which came from more densely populated areas in southern MB (Figure 2). Additionally, skunks predominantly live in the southern region of MB (18,19), and this is consistent with the areas from which skunk samples came during the study period (Figure 2).

Figure 2.

Sampling locations in Manitoba between April 2014 and March 2018.

The number of days to receive sample results varied among regions and years of sample submission, with the first year of the program and more remote communities having longer times to reporting of results. During the first year of MRC there was a learning period for sample collectors, determining the most appropriate shipping method, and the coordination of determining if a sample was required. While the time until sample results are received improved after the first year, it is evident that it still takes longer to obtain results from more remote Northern region, and this is unlikely to change due to the limited methods by which samples can be shipped from these regions. Samples from northern Quebec in remote Inuit communities took a median of 7 d to get diagnostic results from time to exposure (13), indicating issues surrounding diagnostics in remote communities occur in other provinces. It should be noted that the time to receive results is still longer than the 2 d preferred by MHSAL in order to make prompt decisions regarding PEP in humans (20). While time to receive test results is longer than preferred, there were no reported human cases of rabies infections in Manitoba due to collaborative risk assessments by MRC.

In Ontario and Saskatchewan similar programs to MRC are in place where veterinary, public health, agriculture, and natural resources work together (21,22). One significant difference is that Ontario uses wildlife vaccine baits, and has research initiatives regarding rabies in wildlife, whereas MB does not (21). It is often suggested that rabies should be managed through One Health collaborative efforts (12,17,23), and continued collaborative work must be maintained to ensure rabies risk is managed in areas with endemic wildlife populations such as MB.

The purpose of MRC is to manage the risk of rabies exposures in humans and domestic animals; therefore, the animals identified to have rabies are not indicative of the prevalence of rabies in MB as sampling is risk based. In future, wildlife surveillance, especially in bats would help determine the risk for human and domestic animal exposure. Knowing the rabies variant in each positive case would also help in assessing which species is causing the greatest risk for exposure, and if any wildlife management plan would be required in the future.

In MB skunks had the highest proportion of positive samples, with most exposures occurring in southern MB. Humans were less likely to be exposed to rabid animals, as most often their exposures were with dogs and cats which are a lower risk population. Knowing the species, regions, and months of greatest risk will allow for more targeted education to the public and health system about rabies prevention. Through continued collaboration, MRC can offer these educational initiatives, ensuring that a broader One Health approach is taken.