Zoonotic Infections in Native Communities of James Bay,Canada

Introduction

C hisasibi and Waskaganish are Cree communities of approximately 4000 and 2000 residents, respectively. They are located in the James Bay region of Northern Quebec, Canada, south of the Inuit territory of Nunavik (Fig. 1). The latter is a vast region, but is sparsely populated. It has about 11,000 permanent residents living in 14 communities, and nearly 90% of them are Inuit (Régie Régionale de la Santé et des Services Sociaux du Nunavik 2003). The eating and cooking habits of the Cree are different from those of the Inuit. The Cree eat their food well cooked, while the Inuit regularly consume raw meat (Messier et al. 2009).

OPEN IN VIEWER

FIG. 1.

Cree and Inuit communities of Quebec. (Source: Graphics Services, Centre de recherche, Centre hospitalier universitaire de Québec [CHUQ]).

Both communities are located on the eastern shore of James Bay. Waskaganish is a little further south than Chisasibi. The two communities are accessible by road and airplane. Chisasibi is a vibrant young community that has continued to grow since its relocation in the eighties. Waskaganish, one of the oldest permanent settlements in all of Canada, was the focal point of the early fur trade in the region. The main economic activity in Waskaganish throughout its history has been the harvesting of furs, and this is still an important sector of its economy (Cree Nation of Waskaganish 2010). The Cree are entitled to the same health services as other Quebecers and Canadians. They manage their own health services in their own language in a way that incorporates traditional values and ethics. There is a health center in each community.

Traditional lifestyles remain common among the Cree. As a result, many residents of these communities have close contacts with wildlife and may therefore be exposed to zoonotic pathogens (Tanner et al. 1987; Levesque et al. 1995; Levesque et al. 2007; Messier et al. 2009; Campagna et al. 2011). However, there is currently relatively little information on the prevalence of zoonoses among the Cree. These infections are generally underreported (Carlin 2007) for a variety of reasons, including limited access to testing and the non-specificity of symptoms (Agesilas et al. 2005; Merien et al. 2005). Consequently, they often remain undetected by health surveillance systems. It is important to determine their prevalence to assess the need for appropriate preventive measures.

The Nituuchischaayihtitaau Aschii Multi-Community Environment and Health Longitudinal Study in Iiyiyiu Aschii is a survey that collects data on social and health conditions, including zoonotic infections, among the Cree population of northern Quebec. Data for the survey were collected in Mistissini in 2005 (Bonnier-Viger et al. 2007), Eastmain and Wemindji in 2007 (Bonnier-Viger et al. 2011), and Chisasibi and Waskaganish in 2008 (Fig. 1). The main objective of the current work was to determine the seroprevalence of seven zoonotic infections among the general adult population of Chisasibi and Waskaganish. Specifically, risk factors for seropositivity were sought, and the medical records of seropositive subjects were reviewed for symptoms or signs compatible with the suggested exposures.

Materials and Methods

Study subjects were randomly selected from the general population of Chisasibi and Waskaganish and stratified by age and sex based on the census lists of each community. The following age categories were considered: children between 0 and 7 years old, children between 8 and 14 years old, teenagers and adults between 15 and 39 years old, and adults over 40 years old. Weights were assigned to each stratum to ensure adequate population representation (Bonnier-Viger et al. 2011). However, only residents aged 18 years and over were eligible for inclusion in this zoonoses study. They received a small remuneration for their participation. The protocol was approved by the Research Committee of the Cree Board of Health and the Ethics Committee of the Centre de recherche, Centre hospitalier de l'Université Laval. After providing written informed consent, participants completed three questionnaires, and blood samples were drawn.

The questionnaires were adapted from those already used in other Cree communities (Levesque et al. 2007; Campagna et al. 2011). The information gathered focused on demographic characteristics and lifestyle (age, sex, education, community of residence, occupation, and the presence of pets at home), and on hunting and trapping activities (fur and meat handling), wearing gloves while handling animals and birds (gutting, cleaning, skinning, or plucking), the number of years spent hunting and trapping, forest activities (hunting, fishing, trapping, and keeping camp) during the last year, and species and numbers of animals killed or captured during the last year.

Blood samples were permitted to clot, and serum aliquots were maintained at −20°C until tested for antibodies to a range of zoonotic pathogens. The zoonotic pathogens targeted included three bacteria (Coxiella burnetii, Francisella tularensis, and Leptospira spp.), and four parasites (Echinococcus granulosus, Toxocara canis, Toxoplasma gondii, and Trichinella spp.). Immunoenzymatic methods were undertaken to detect IgG antibodies against Trichinella spp., T. canis, and E. granulosus (SCI Inc., Carlsbad, CA), T. gondii (AxSYM; Abbott Diagnostics, Abbott Park, IL), Leptospira spp., and C. burnetii (Virion\Serion; Serion Immundiagnostica GmbH, Würzburg, Germany). Antibodies against F. tularensis were detected by the tube agglutination test (Snyder 1980; Stewart 1981). The test specifications are presented in Table 1.

The medical records from the health centers of Chisasibi and Waskaganish for all subjects with one or more positive results were reviewed to identify clinical manifestations compatible with the suggested zoonotic exposures. They were audited according to the predicted period of antibody persistence, the last 5 years for persons testing positive for C. burnetii, Leptospira spp., and Trichinella spp., and 10 years for those positive for F. tularensis, T. canis, T. gondii, and E. granulosus.

We compared the proportions of selected characteristics from the questionnaires and seroprevalence rates between the two communities by chi-square testing or Fisher's exact test, as appropriate. The frequency distribution of the seroprevalence data was stratified by community. Zoonoses with seroprevalence ≥5% in both communities were kept for analysis. When seropositivity for a given infection was statistically significantly different between the two communities (p≤0.05), they were analyzed separately. For zoonoses with seroprevalence <5% overall, but >5% in one community, analysis was conducted only in the latter. Univariable logistic regression was performed to assess relationships between seropositivity for the seven zoonotic infections and both socio-demographic and wildlife exposure variables. Age was examined as a continuous variable. Variables with p≤0.1 values were included in a multivariable logistic regression model. The statistical significance threshold for multivariable models was set at 0.05. The results are reported as odds ratios. Data analyses were conducted with SAS version 9.2 (SAS Institute, Cary, NC).

Results

A total of 267 participants 18 years or older provided blood samples for this study, 166 from Chisasibi (98 women, 68 men), and 101 from Waskaganish (59 women, 42 men). Participation rates, defined as the number of participants over the number of people invited to participate, were 32.5% (166/510) in Chisasibi, and 31.1% (101/325) in Waskaganish, representing, respectively, 6.9% (166/2388) and 8.7% (101/1168) of the total population aged 18 years or older. A breakdown of the data by age, sex, and community is presented in Table 2.

Table 3 describes the frequency of key variables in both communities. Overall, it is noteworthy that in Chisasibi the study population was slightly older (p=0.05). There were very few cat owners, but owning a cat was more common in the Waskaganish (p<0.01). Several variables suggested that participants from Chisasibi had greater contacts with wildlife (handling animals other than herbivores, and trapping and hunting in autumn) than those from Waskaganish (Table 3).

The results of serological analyses are presented in Table 4. Overall, 47% of the participants presented serological evidence of exposure to at least one zoonosis. A total of 96 people were positive on one serology, 21 were positive on two, and six were positive on three. There was no statistical relationship between the number of positive serologies and sex (p=0.37), community (p=0.35), or hunting or not (p=0.48). However, people aged 40 years or older were more prone to be positive on two or three serologies (p=0.04). The most prevalent pathogens were Leptospira spp. (23%), F. tularensis (18%), and T. gondii (9%). Seroprevalence rates for the bacterium C. burnetii and the parasite T. canis were both below 5%. Less than 1% (0.7%) showed evidence of exposure to E. granulosus, and no subject tested positive for Trichinella spp. in either community. There were statistically significant differences in seroprevalence rates for Leptospira spp., T. canis, and T. gondii between the two communities (Table 4). Seropositivity for Leptospira spp. and T. gondii was higher in Chisasibi, while participants from Waskaganish appeared to have been more exposed to T. canis.

Data from the medical records of participants with serology positive for one or more pathogens are summarized in Table 5. Clinical manifestations consistent with exposures suggested by the serological results were documented for C. burnetii, Leptospira, F. tularensis, and T. canis in a small number of subjects. In contrast, the medical records of patients seropositive for E. granulosus and T. gondii did not reveal any history of symptoms or signs suggestive of infection. Comparisons of proportions of the seroprevalence data by sex and age group in each community indicated greater exposure to F. tularensis among women in Waskaganish (women: 31%, men: 10%) (p=0.02). There was a statistically significant interaction between age and community for seropositivity to Leptospira spp. (p<0.01) and F. tularensis (p=0.02). Leptospira spp. seroprevalence was higher among people over 40 years old (<40 years: 22%, ≥40 years: 32%) in Chisasibi, while it was lower in the same age group in Waskaganish (<40 years: 22%, ≥40 years: 5%). The observed relationship was reversed for F. tularensis seropositivity; the prevalence rate was higher among those aged 40 years or older (<40 years: 18%, ≥40 years: 29%) in Waskaganish, while it was lower in the same age group in Chisasibi (<40 years: 19%, ≥40 years: 12%).

Table 6 presents the significant predictive variables found by multivariate analysis for the two villages combined or stratified by village based on criteria in analysis design.

Discussion

The study sample was selected to be representative of the general populations of Chisasibi and Waskaganish, including children. Representativeness could have been partly hampered considering that this study was restricted to the adult population. Moreover, the participation rates of both communities were around 30%. We cannot explain these low participation rates. The Nituuchischaayihtitaau Aschii Multi-Community Environment and Health Longitudinal Study in Iiyiyiu Aschii visited Eastmain and Wemindji before Waskaganish and Chisasibi. Participation rates were around 60% in these communities (Campagna et al. 2011), where the same methods of recruitment and the same incentives were applied. A response rate of around 70% is usually targeted for interview surveys (Tolonen 2005). Therefore, we should question the external validity of the results presented here, and they should therefore be interpreted with caution. Nevertheless, we believe that this study provides interesting data on the populations of both Chisasibi and Waskaganish regarding their exposure to zoonoses associated with their lifestyle.

Samples from the two villages differed slightly in the age of participants and their overall relationship with wildlife; those of Chisasibi were older and more exposed to wildlife (Table 3). These differences may explain some variations in seroprevalence rates for different pathogens between the two communities.

Nearly half of all participants (47%) were seropositive for at least one of the zoonotic organisms studied. Among them, Leptospira spp. and F. tularensis were the most prevalent. Probably due to the persistence of some antibodies, people aged 40 years or older were more prone to be positive on more than one serology.

Several seroprevalence investigations have been conducted in the Cree communities of Quebec (Levesque et al. 2007; Campagna et al. 2011), and in the 14 villages of Nunavik (the Inuit territory of Quebec; Fig. 1; Messier et al. 2007) in recent years. The principal results of these studies are included in Table 7. Except for the Mistissini study, which focused on a small sample of hunters and their wives, the other studies were based on samples of the general population. The seroprevalence rates documented in Chisasibi and Waskaganish are comparable to those reported in Mistissini, with some exceptions. For trappers in Mistissini, seroprevalence of C. burnetii was significantly higher, and even if the prevalences were not statistically significantly different, it seems also to be the case for F. tularensis, indicating an environment more favorable to these bacteria in Mistissini, and/or higher risk activities among these hunters. Lévesque and associates (1995) reported a similar seroprevalence rate for C. burnetii (15%) among 165 trappers and an equal number of rural controls in the Quebec City area of southern Quebec, but the latter population is more exposed to domestic and farm animals, known vectors of this infection.

The Inuit from Nunavik are exposed significantly less to Leptospira spp. and C. burnetii, but more to E. granulosus and T. gondii. For the latter pathogen, differences between prevalence estimates in the Nunavik and the Cree communities may be explained by their dissimilar eating and culinary habits. As already stated, unlike the Cree, the Inuit consume raw meat regularly, which increases their risk of exposure (Messier et al. 2009). Overall, the seroprevalence rates reported previously for the neighboring Cree communities of Eastmain and Wemindji were not statistically significantly different from the results of the present study and validate its findings.

In Quebec, leptospirosis, tularemia, Q fever, and trichinellosis are notifiable diseases. Our results show that several study participants from Chisasibi and Waskaganish were exposed to some of these pathogens (Leptospira spp., F. tularensis, and C. burnetii). Nevertheless, no such infections were reported in the Cree territories between 1990 and 2006 (Carlin 2007), indicating that they were either mildly symptomatic (asymptomatic), they did not seek care, and/or they were not diagnosed. Our review of the medical records of these participants occasionally revealed a clinical illness consistent with the implicated infection. Unfortunately, it is impossible to assess or assign causality with certainty in retrospect. Nevertheless, our results strongly support advising local health care providers to be more alert for zoonotic infections.

Unlike Levesque and associates (1995 and 2007), who showed links between F. tularensis seropositivity and exposure to wildlife, we could not demonstrate such an association. Overall, gender (female) was the only variable linked with exposure to the bacteria (p<0.01) (Table 6). This relationship may be linked to culinary activities (food preparation and handling game meat). The significant interaction between age and community found in this study is difficult to explain, although participants from the two villages were different with respect to age and exposure to wildlife (Table 3). The relatively benign clinical histories of participants seropositive for F. tularensis suggest that the subspecies holartica (type B), known to be less virulent, could be responsible (Eigelsbach and McGann 1984; Tarnvik 1989).

Leptospira spp. seropositivity also showed a significant interaction between age and community (p<0.01) of unclear etiology. Moreover, this study identified rabbit handling as a risk factor associated with seropositivity for Leptospira spp. (Table 6). Shotts and colleagues (1971) reported a seroprevalence of 77% for leptospirosis in a population of 50 rabbits captured in Mississippi (Shotts et al. 1971). Community was also a significant variable in Leptospira spp. seroprevalence (p=0.01). When stratified by village, our analysis indicates that wearing gloves when handling carcasses has a protective effect against infection in Chisasibi. In Waskaganish, younger participants seem to be more exposed to Leptospira spp. infection. In the absence of good data on antibody persistence, this finding is hard to interpret in a cross-sectional study. A history of spring hunting was also associated with infection in Waskaganish. Hunting activities in spring focus on birds and involve frequent contact with water, possibly explaining the linkage with leptospirosis. Indeed, it is well-known that leptospirosis is connected with aquatic activities (Mumford 1989).

The current study identified older age as a risk factor for toxoplasmosis (Table 6). This association with age is consistent with the lifelong persistence of circulating antibodies in people exposed to T. gondii (Tenter et al. 2000; Bouhamdan et al. 2010). We also documented a tendency towards stronger serological evidence of T. gondii exposure among those who had handled ducks during the last year. Some indicators suggest that ducks may be potential sources of infection (Boehringer et al. 1962; Bartova et al. 2009).

We identified older age as a risk factor for Q fever in Chisasibi (Table 6). Such an association has also been reported in France (Tissot Dupont et al. 1992; Raoult et al. 2000). Pneumonia is a common clinical manifestation of Q fever (Marrie et al. 1985). Interestingly, nearly half of 11 seropositive participants had a history of documented pneumonia.

The proportion of people exposed to T. canis was clearly higher in Waskaganish than in Chisasibi (Table 4). The dog is the main reservoir of T. canis (Glickman and Schantz 1981; Barriga 1988; Wolfe and Wright 2003; Mandarino-Pereira et al. 2010), but only 4 of 11 participants seropositive for T. canis declared owning a dog. It is possible that the presence of stray dogs in the villages may increase the risk of exposure. Although our study also identified handling of caribou as a risk factor for T. canis exposure in Waskaganish, the relative lack of statistical power must be considered. Two of these seropositive individuals had histories of unexplained eosinophilia without other signs or symptoms indicative of the disease (Table 5).

Only two people had serology positive for E. granulosus, one in Chisasibi and the other in Waskaganish (Table 4). Even though a recent study showed a higher incidence of hospitalization for echinococcosis among people living north of the 55th parallel (2.9 per million per year) (Gilbert et al. 2010), we encountered only one case of echinococcosis in the literature on Cree in Quebec Cree territories in 1955 (Begin et al. 1956). However, 3% seroprevalence was documented by Tanner and associates (1987) during the eighties in a sample of hospitalized patients from different Cree communities. We did not find either signs or symptoms suggestive of echinococcosis in the medical records of the two seropositive individuals in Chisasibi and Waskaganish.

As in Mistissini, no evidence of antibodies against Trichinella spp. was detected in the populations of Chisasibi and Waskaganish. In the Eastmain and Wemindji study, only one individual was seropositive (Campagna et al. 2011). These data indicate infrequent exposure to this parasite among the Cree.

Despite our inability to retrospectively identify clinical episodes compatible with most of these zoonoses, the serology data suggest that large proportions of the populations of Chisasibi and Waskaganish have been exposed to at least one of the targeted zoonotic agents. The Cree population as well as the medical staff living in these regions should be on the alert for these diseases and implement measures to prevent their transmission. Enhanced awareness among the Cree of simple preventive measures, such as cooking game meat thoroughly and wearing gloves when manipulating animals, and raising the level of vigilance among health care workers, would help to decrease exposure as well as increase the chances that appropriate diagnostic testing will be performed. Our data indicate that active pursuit of a diagnosis may be particularly fruitful in subjects with unexplained eosinophilia, high and prolonged fever, and/or atypical pneumonia.